ho journal club pav vs psv - jsepticrally adjusted ventilatory assist ventilation modes are...

47
聖マリアンナ・東京ベイ合同 Journal Club PAVは呼吸器離脱に適したモードか? 東京ベイ・浦安市川医療センター 集中治療科 ⽚岡 2016.03.22

Upload: others

Post on 28-Jun-2020

4 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

聖マリアンナ東京ベイ合同Journal Club

PAVは呼吸器離脱に適したモードか

東京ベイ浦安市川医療センター集中治療科 岡 惇

20160322

本の論

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Critical Care Medicine wwwccmjournalorg 1

Objectives Despite protocols incorporating spontaneous breath-ing trials 31 of ICU patients experience difficult or prolonged weaning from mechanical ventilation Nonfatiguing modes such as pressure support ventilation are recommended Proportional assist ventilation provides assistance in proportion to patient effort which may optimize weaning However it is not known how proportional assist ventilation performs relative to pressure sup-port ventilation over a prolonged period in the complex ICU set-ting The purpose of this study was to compare the physiologic and clinical performance (failure rate) safety and feasibility of protocols using daily spontaneous breathing trial plus pressure support ventilation versus proportional assist ventilation until ven-tilation discontinuationDesign Single-center unblinded pilot randomized controlled trialSetting Medical-surgical ICU of a tertiary-care hospitalPatients Adult patients intubated greater than 36 hours were ran-domized if they met eligibility criteria for partial ventilatory support

tolerated pressure support ventilation greater than or equal to 30 minutes and either failed or did not meet criteria for a sponta-neous breathing trialInterventions Patients were randomized to the pressure support ventilation or proportional assist ventilation protocol (PAV+ Puri-tan Bennett 840 Covidien Boulder CO) Both protocols used progressive decreases in level of assistance as tolerated coupled with daily assessment for spontaneous breathing trialsMeasurements and Main Results Of 54 patients randomized outcome data are available for 50 patients 27 were randomized to receive proportional assist ventilation and 23 to receive pres-sure support ventilation There were no adverse events linked to the study interventions and protocol violations were infrequent Recruitment was slower than projected (13 patients per month) The median (interquartile range) time from randomization to suc-cessful extubation was 39 days (28ndash84 d) on proportional assist ventilation versus 49 days (29ndash263 d) on pressure support ventilation (p = 039) Time to live ICU discharge was 73 days (52ndash114 d) on proportional assist ventilation versus 124 days (75ndash308 d) on pressure support ventilation (p = 003)Conclusion This pilot study demonstrates the utility safety and feasibility of the weaning protocols and provides important infor-mation to guide the design of a future randomized controlled trial comparing weaning from mechanical ventilation on pressure sup-port ventilation versus proportional assist ventilation (Crit Care Med 2016 XX00ndash00)Key Words assisted mechanical ventilation asynchrony index proportional assist ventilation respiratory muscles ventilator weaning

Difficult or prolonged weaning from mechanical ventilation (MV) affects 31 (range 26ndash42) of mechanically ventilated patients (1) and increases

patient morbidity mortality ICU length of stay and cost to the healthcare system (2) When initial weaning attempts fail to liberate patients from MV ideally clinicians should choose

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights ReservedDOI 101097CCM0000000000001600

1London Health Sciences Centre London ON Canada2Department of Medicine The University of Western Ontario London ON Canada

This study was performed at London Health Sciences Centre University Hospital London CanadaSupplemental digital content is available for this article Direct URL cita-tions appear in the printed text and are provided in the HTML and PDF versions of this article on the journalrsquos website (httpjournalslwwcomccmjournal)Supported in part by Critical Care Western Department of Medicine The University of Western Ontario Program of Experimental Medi-cine (POEM) Department of Medicine University of Western Ontario Lawson Health Research Institute London Health Sciences Centre London Canada and Covidien Boulder USADr Bosma has received honourarium for lectures and travel reimburse-ments from Covidien Dr Bosmarsquos institution has received grant support from Covidien for Dr Bosmarsquos research studies The remaining authors have disclosed that they do not have any potential conflicts of interestFor information regarding this article E-mail KarenJBosmalhsconca

A Pilot Randomized Trial Comparing Weaning From Mechanical Ventilation on Pressure Support Versus Proportional Assist Ventilation

Karen J Bosma MD FRCPC12 Brooke A Read MHS BSc RRT1 Mohammad J Bahrgard Nikoo MD FRCPC12 Philip M Jones MD FRCPC MSc12 Fran A Priestap MSc1 James F Lewis MD FRCPC12

Crit Care Med 2016 Jan 20[Epub ahead of print]

呼吸器離脱の定義分類 定義

Simple weaning 最初のSBTで呼吸器離脱する患者Difficult weaning 最3回までのSBTあるいは1回

のSBTから呼吸器離脱までに最7間かかる患者

Prolonged weaning 4回以上のSBTあるいは最初のSBTから呼吸器離脱までに8間以上かかる患者

Eur Respir J 2007 29 1033-1056

呼吸器離脱困難bull Difficult weaning prolonged weaningは

呼吸器装着患者の31に起こる

bull それらは患者のICU滞在数の延死亡率の増加コスト増につながる

bull PSVがSBTに失敗した患者がweaningをう際に最も使われているモードである

bull PSVはトリガーした後は受動的であり横隔膜の萎縮につながる可能性同調が多い可能性がある

EurRespirJ2007291033ndash1056

JAMA2002287345ndash355

Chest19941061188ndash1193

JApplPhysiol(1985)199681426ndash436

IntensiveCareMed2008341477ndash1486

呼吸器離脱困難患者に適したモードはあるのか

従来の呼吸器モード

医師 呼吸器 患者モード

1回換気量呼吸数

Closed-loop Ventilation

医師 呼吸器 患者 サポート

気道抵抗コンプライアンスを測定患者が求める1回換気量流速を計測する

Closed-loop Ventilationbull 患者の吸気努に例したサポートをうモード

Proportional Assist Ventilation(PAV)rarrPAV+(PB840980) PPS(Evita)

Neurally Adjusted Ventilatory Assist(NAVA)rarrServo i

bull 動weaningシステムにより離脱を指すモードSmartCare

rarrEvita XLinfinity V500Adaptive Support Ventilation(ASV)

rarrHammilton G5

患者の吸気努に例してサポートする

conventional ventilation The patient still has to generatesufficient flow to trigger the ventilator However as dis-cussed later conventional modes may force a larger tidalvolume than normal causing air trapping and thus increas-ing the number of missed triggers

Neurally Adjusted Ventilatory Assist

NAVA accomplishes the same goals as PAV but uti-lizes measurement of the diaphragmatic EMG signal tocontrol gas delivery6 This is accomplished by the place-ment of a specifically designed nasogastric tube that has aseries of EMG electrodes near its distal end positionedacross the diaphragm As illustrated in Figure 1 support-ing ventilation based on the diaphragmatic EMG signalshould greatly improve the response of the ventilator andsynchrony since the signal recognized is high up on theneural pathway controlling ventilation7 What is set by theclinician is the pressure applied for each millivolt of EMGactivity Thus similar to PAV a portion of the ventilatoryeffort is proportionally provided by the ventilator and theremainder by the patient As EMG activity increases pres-sure is applied during the inspiratory phase and as thediaphragm relaxes airway pressure decreases Inspirationends at a specific percentage of the peak EMG activityContrary to PAV NAVA greatly improves triggering since

gas delivery begins when the diaphragm is simulated notas a result of flow in the airway Thus even in the pres-ence of severe air trapping or large system leaks trigger-ing is not compromised

PAV and NAVA Versus Traditional Gas Delivery

Figure 2 nicely illustrates the relationship between andthe patientrsquos effort or work of breathing and the ventilatorpressure during various ventilation modes1 During a spe-cific breath with all forms of volume ventilation there isan indirect relationship between patient effort (Pmus) andthe pressure applied by the ventilator Essentially the morework assumed by the patient the less work applied by theventilator This relationship creates difficult problems re-garding synchrony which can easily be illustrated with alung model with proper setting of inspiratory demand Ifdemand is high and the lung model tidal volume is greaterthan the ventilator set tidal volume no pressure abovebaseline will be applied to the lung

In pressure-targeted ventilation regardless of mode ven-tilator work or applied pressure should be independent ofPmus That is during a specific breath a constant pressureshould be applied to the airway during inspiration nomatter how much work or Pmus is generated by the patientOf course in patients with very high ventilatory demandpressure may not be sustained since the ventilator may notalways be able to meet patient flow demand

In PAV and NAVA there is a direct relationship be-tween the patientrsquos effort (Pmus) and ventilator pressure

Fig 1 Steps in the process of activating a ventilator breath Neu-rally adjusted ventilatory assist ventilation modes are activatedand controlled by diaphragm excitation Proportional assist ven-tilation modes are activated by a change in airway pressure flowand volume (Adapted from Reference 6)

Fig 2 Relationship between patient effort and ventilator pressureduring various ventilation modes During volume ventilation pa-tient effort and ventilator pressure are inversely related Duringpressure ventilation ventilator pressure in unaffected by patienteffort During proportional assist ventilation (PAV) and neurally ad-justed ventilatory assist (NAVA) patient effort and ventilator pres-sure are directly related (Adapted from Reference 5)

PROPORTIONAL ASSIST VENTILATION AND NEURALLY ADJUSTED VENTILATORY ASSIST

142 RESPIRATORY CARE bull FEBRUARY 2011 VOL 56 NO 2

RespirCare201156(2)140ndash148

Proportional Assist Ventilation

bull コンプライアンス(C)とレジスタンス(R)を測定して患者の吸気努に例したサポートをう

bull Paw = VC + RtimesFlowbull WOB = ʃ PtimesFlow dt

bull PB840では8-15呼吸ごとに300msecの吸気ポーズをいCとRを測定

設定するのはサポート率のみ

患者の呼吸仕事量(WOBpt)が03〜08JLになるようにサポート率を設定する

PAVのメリットデメリット

bull メリット患者-呼吸器の同調性が良い呼吸筋を適度に使うことができる

bull デメリット患者の呼吸努が弱い場合はサポートが減ってしまうリークがある場合過度にサポート圧が上がってしまう可能性(runaway現象)

PAVのエビデンスbull PAVとPSVによるウィーニングをい48

時間観察したRCTではPAVのが調節呼吸に戻る割合が低かった(PAV 11 vsPSV 22 p=004)

bull 夜間PAVにすることで睡眠の質が改善したIntensiveCareMed2008342026ndash2034

Crit CareMed2007351048ndash1054

PAVは呼吸器離脱困難患者のウィーニングにおいて

優れたモードか

本の論

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Critical Care Medicine wwwccmjournalorg 1

Objectives Despite protocols incorporating spontaneous breath-ing trials 31 of ICU patients experience difficult or prolonged weaning from mechanical ventilation Nonfatiguing modes such as pressure support ventilation are recommended Proportional assist ventilation provides assistance in proportion to patient effort which may optimize weaning However it is not known how proportional assist ventilation performs relative to pressure sup-port ventilation over a prolonged period in the complex ICU set-ting The purpose of this study was to compare the physiologic and clinical performance (failure rate) safety and feasibility of protocols using daily spontaneous breathing trial plus pressure support ventilation versus proportional assist ventilation until ven-tilation discontinuationDesign Single-center unblinded pilot randomized controlled trialSetting Medical-surgical ICU of a tertiary-care hospitalPatients Adult patients intubated greater than 36 hours were ran-domized if they met eligibility criteria for partial ventilatory support

tolerated pressure support ventilation greater than or equal to 30 minutes and either failed or did not meet criteria for a sponta-neous breathing trialInterventions Patients were randomized to the pressure support ventilation or proportional assist ventilation protocol (PAV+ Puri-tan Bennett 840 Covidien Boulder CO) Both protocols used progressive decreases in level of assistance as tolerated coupled with daily assessment for spontaneous breathing trialsMeasurements and Main Results Of 54 patients randomized outcome data are available for 50 patients 27 were randomized to receive proportional assist ventilation and 23 to receive pres-sure support ventilation There were no adverse events linked to the study interventions and protocol violations were infrequent Recruitment was slower than projected (13 patients per month) The median (interquartile range) time from randomization to suc-cessful extubation was 39 days (28ndash84 d) on proportional assist ventilation versus 49 days (29ndash263 d) on pressure support ventilation (p = 039) Time to live ICU discharge was 73 days (52ndash114 d) on proportional assist ventilation versus 124 days (75ndash308 d) on pressure support ventilation (p = 003)Conclusion This pilot study demonstrates the utility safety and feasibility of the weaning protocols and provides important infor-mation to guide the design of a future randomized controlled trial comparing weaning from mechanical ventilation on pressure sup-port ventilation versus proportional assist ventilation (Crit Care Med 2016 XX00ndash00)Key Words assisted mechanical ventilation asynchrony index proportional assist ventilation respiratory muscles ventilator weaning

Difficult or prolonged weaning from mechanical ventilation (MV) affects 31 (range 26ndash42) of mechanically ventilated patients (1) and increases

patient morbidity mortality ICU length of stay and cost to the healthcare system (2) When initial weaning attempts fail to liberate patients from MV ideally clinicians should choose

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights ReservedDOI 101097CCM0000000000001600

1London Health Sciences Centre London ON Canada2Department of Medicine The University of Western Ontario London ON Canada

This study was performed at London Health Sciences Centre University Hospital London CanadaSupplemental digital content is available for this article Direct URL cita-tions appear in the printed text and are provided in the HTML and PDF versions of this article on the journalrsquos website (httpjournalslwwcomccmjournal)Supported in part by Critical Care Western Department of Medicine The University of Western Ontario Program of Experimental Medi-cine (POEM) Department of Medicine University of Western Ontario Lawson Health Research Institute London Health Sciences Centre London Canada and Covidien Boulder USADr Bosma has received honourarium for lectures and travel reimburse-ments from Covidien Dr Bosmarsquos institution has received grant support from Covidien for Dr Bosmarsquos research studies The remaining authors have disclosed that they do not have any potential conflicts of interestFor information regarding this article E-mail KarenJBosmalhsconca

A Pilot Randomized Trial Comparing Weaning From Mechanical Ventilation on Pressure Support Versus Proportional Assist Ventilation

Karen J Bosma MD FRCPC12 Brooke A Read MHS BSc RRT1 Mohammad J Bahrgard Nikoo MD FRCPC12 Philip M Jones MD FRCPC MSc12 Fran A Priestap MSc1 James F Lewis MD FRCPC12

Crit Care Med 2016 Jan 20[Epub ahead of print]

論のPICO

P 呼吸器離脱困難な患者(simple weaningではない)

I PAVを使

C PSVを使

O 呼吸器離脱までの時間

Study Designbull 単施設盲検前向きランダム化較

試験(pilot study)bull カナダロンドンにあるUniversity

Hospital-London Health Science Centre の内科外科ICU(20床)

bull ランダム割付はcomputer-generated random number sequence によってわれ隠蔽化されている

PatientsInclusion criteria

bull 18歳以上の挿管患者bull 36時間以上呼吸器管理をされている

PatientsExclusion criteria

1) 48時間以内にwithdrawが考慮される患者2) 位の脊髄障害進性の神経筋疾患の

よって慢性的な呼吸器依存となる可能性があるまたは脳外科的介が必要である患者

Enrolment Procedure

Supplemental Digital Content Figure 1 Enrolment Procedure

Daily Screening of all ICU patients gt18 yrs of age Intubated gt36 hours

Exclude if Exclusion Criteria

Successful SBT awaiting extubation Withdrawal of life support considered High spinal cord injury Neurosurgical patient Neuromuscular disease

Follow using daily eligibility checklist until meet all Clinical Stability Criteria to trial mode of partial ventilatory support (PAV or PSV)

partialcomplete reversal of cause of respiratory failure

body temperature between 36ordmC and 39ordmC metabolic disorders corrected pH gt732 hemoglobin gt70 gL no ongoing bleeding stable hemodynamic status Able to trigger PaO2 ge 60 mmHg or SaO2 ge 90 on FiO2le

060 and PEEP le 15 cm H2O ACVC Pplatle30cmH2O with Vt 6-8 mlkg

OR ACPC PC+PEEPle30cmH2O

Exclude if Exclusion Criteria Has met all Criteria for Switching to

Spontaneous Mode of Ventilation gt24 hours

Pre-inclusion Test on PSV 15 cmH2O for 30 minutes

Patient fails Pre-inclusion test Return to AC ventilation and reassess within 24 hour

Patient passes Pre-inclusion test Assess for Criteria to Initiate Weaning

PaO2FiO2gt200 on FiO2 le050 and PEEP le8 cmH2O No inotropesvasopressors except low dose dopamine

RSBI lt105 Patient passes Pre-SBT readiness assessment Proceed to SBT Place patient on CPAP 5 cmH2O and FiO2 040 for 30 minutes

Patient does not meet Criteria to Initiate Weaning

Patient meets Criteria to Initiate Weaning Proceed to Pre-SBT readiness assessment Place patient on CPAP for 1-2 min measure RSBI

RSBIge 105 Patient fails Pre-SBT readiness assessment

Patient passes SBT Patient fails SBT

Exclude if Exclusion Criteria

Passed SBT and tolerating PSVle7cmH2O and PEEPle5 cmH2O while awaiting extubation

Obtain informed consent from Patient or SDM as soon as available (deferred consent permitted)

Randomize

Exclude if Exclusion Criteria

Patient SDM refuse consent

bull 呼吸不全の原因に改善が認められてきている

bull 体温は36〜39度であるbull pH>732であるbull Hb>7gdLで現在出なしbull 動態が安定しているbull トリガーできるbull FiO2≦06 PEEP≦15cmH2Oで

PaO2≧60mmHg or SaO2≧90bull ACVCでVt 6-8mlkgで

Pplat≦30cmH2OもしくはACPCでPC+PEEP≦30cmH2O

上記をすべて満たせばプレテストとしてPSV 15cmH2Oに変更

Enrolment Procedure

Supplemental Digital Content Figure 1 Enrolment Procedure

Daily Screening of all ICU patients gt18 yrs of age Intubated gt36 hours

Exclude if Exclusion Criteria

Successful SBT awaiting extubation Withdrawal of life support considered High spinal cord injury Neurosurgical patient Neuromuscular disease

Follow using daily eligibility checklist until meet all Clinical Stability Criteria to trial mode of partial ventilatory support (PAV or PSV)

partialcomplete reversal of cause of respiratory failure

body temperature between 36ordmC and 39ordmC metabolic disorders corrected pH gt732 hemoglobin gt70 gL no ongoing bleeding stable hemodynamic status Able to trigger PaO2 ge 60 mmHg or SaO2 ge 90 on FiO2le

060 and PEEP le 15 cm H2O ACVC Pplatle30cmH2O with Vt 6-8 mlkg

OR ACPC PC+PEEPle30cmH2O

Exclude if Exclusion Criteria Has met all Criteria for Switching to

Spontaneous Mode of Ventilation gt24 hours

Pre-inclusion Test on PSV 15 cmH2O for 30 minutes

Patient fails Pre-inclusion test Return to AC ventilation and reassess within 24 hour

Patient passes Pre-inclusion test Assess for Criteria to Initiate Weaning

PaO2FiO2gt200 on FiO2 le050 and PEEP le8 cmH2O No inotropesvasopressors except low dose dopamine

RSBI lt105 Patient passes Pre-SBT readiness assessment Proceed to SBT Place patient on CPAP 5 cmH2O and FiO2 040 for 30 minutes

Patient does not meet Criteria to Initiate Weaning

Patient meets Criteria to Initiate Weaning Proceed to Pre-SBT readiness assessment Place patient on CPAP for 1-2 min measure RSBI

RSBIge 105 Patient fails Pre-SBT readiness assessment

Patient passes SBT Patient fails SBT

Exclude if Exclusion Criteria

Passed SBT and tolerating PSVle7cmH2O and PEEPle5 cmH2O while awaiting extubation

Obtain informed consent from Patient or SDM as soon as available (deferred consent permitted)

Randomize

Exclude if Exclusion Criteria

Patient SDM refuse consent

呼吸窮迫がなければプレテストクリアbull FiO2≦05 PEEP≦8cmH2OでPF≧200bull 強薬昇圧剤の使なしであればSBTへ

まず1-2分CPAP 5cmH2OにしてRSBIを測定

RSBI<105であればSBTへPSV 5cmH2Oにして30-120分

酸素化の条件を満たさないRSBI≧105SBT失敗であれば割り付けへ

Ventilation Protocols bull 呼吸器はPuritan Bennett 840 を使bull 後述するPAVプロトコールPSVプロトコー

ルに従って呼吸療法(RRT)が設定を変更(respiratory distress 呼吸窮迫がないようにできるだけサポートを下げていく)

bull 設定したmaxのサポートが必要になった場合はACに戻す

bull PAV群はPAV+とACPSV群はPSVとACしか使しない

Ventilation Protocols bull RCTの開始前に15名の患者をいて観察

研究をいRRTにプロトコルの確認をってもらった

bull プロトコルはすぐにられるようにラミネートしてベッドサイドに置いた

bull RRTは毎SBTをえるかどうか後述するチェックリストをいて確認除外項を満たさない場合は毎RSBIのアセスメントとSBTをう

Respiratory Distress bull 呼吸数>35回minbull 脈拍>140min もしくはベースラインより20以上の

上昇bull 収縮期圧>180mmHg もしくは<90mmHg ベース

ラインより30以上の変化bull 不安bull 発汗bull 副呼吸筋の使bull 奇異呼吸bull 呼吸困難の訴え

PAV protocolSupplemental Digital Content Figure 2a Proportional Assist Ventilation Protocol

bull 開始の設定はサポート率70

bull 2-3時間おきにRRTがアセスメントし問題なければ10-20ずつPAVサポートを下げる

サポート率90まで上げてもダメな場合はACに戻す

PSV protocolSupplemental Digital Content Figure 2b Pressure Support Ventilation Protocol

bull 開始の設定はPS 15cmH2O

bull まずRRlt35 VTgt5mLkgで耐えられるPSまで下げる

bull 2-3時間おきにRRTがアセスメントしてPS2-3cmH2Pずつ下げる

Data collection form for RRTsSupplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

SBTの開始基準bull PF≧200bull FiO2≦05bull pH≧732bull トリガーできるbull 24時間以内に筋虚イベントなしbull 強薬昇圧剤の使なし

Data collection form for RRTs

Supplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

CPAPにしてRSBIを測定RSBI<105であればSBT施

PAV群はサポート率20PSV群はPS 5cmH2OでSBTを施

PEEP≦8cmH2OでSaO2≧90呼吸窮迫なしであればSBTクリア

抜管へ

Endpoints 呼吸器離脱までの時間ICU病院滞在期間

呼吸器離脱成功の定義抜管後48時間再挿管なし侵襲的呼吸(NIV)の使もなし気管切開使時は呼吸器から最終的に外れた時点1のうち12時間以上NIVを使していない

Statistical Analysis bull Pilot studyのためサンプルサイズの計

算はしていないbull P値<005を統計学的有意としたbull Intention to treat approachをった

RESULTS

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial

エンロール期間は42か80名が基準を満たしが最終的に割付されたのは54名

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial最終的に解析がわれたのはPAV群27名PSV群23名

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

ベースラインは両群間で変わりなし重症度はAPACHEⅡ 27点程度患者は肺炎が3割もともとCOPDを持っている患者は2割程度

Feasibility Outcomes bull 呼吸窮迫でRRTが呼吸器サポートを上昇

させた回数はPAV群で76回232daysPSV群で124回295daysと有意にPSV群で多かった(p=0002)

bull ACに戻した回数はPAV群で66回PSV群で90回と変わりなし(p=061)

bull 全くACに戻さなかった患者はPAV群で627名PSV群で623名(p=10)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

ベースラインは両群間で変わりなしPAVにすることでΔP圧Peak圧が低く吸気時間呼吸時間がくなるPSV群と較するとPAV群のがPaoAreaPaoAreaVE(呼吸器のサポートの強さをす)が低くVTPaoAreaがくsecond post-randomizationの解析においてΔP圧Peak圧が低い

Asynchrony index

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 7

B635) documenting compliance with the daily screening for SBT noncompletion of either the CPAP screening test or SBT occurred on 5 of 292 of patient study days (17) and this violation involved two patients in each study arm The second audit examining for nine different potential deviations found complete adherence to study protocol on 254 of 292 of patient study days (900) There were no adverse events (pneumo-thorax arrhythmia or death) resulting from any of the study interventions

Physiologic Response to Ventilation ProtocolsRecordings of flow and airway pressure were available for analysis for all 50 patients prerandomization 49 patients immediately (median 1 interquartile range [IQR] 0ndash15 hr) postrandomization and for 26 patients during subsequent days (median 27 IQR 20ndash70-hr postrandomization) Approximately 15000 breaths in 125 recordings were analyzed Table 2 compares the respiratory parameters in both groups pre- and postrandomization There were no statistically sig-nificant differences between the PAV and PSV groups at base-line prerandomization (p values not shown) In the PAV group there was a significant decrease in peak airway pressure delta pressure (measured as peak airway opening pressure minus the PEEP) and area under the inspiratory pressure curve immediately postrandomization and an increase in the per-cent of total cycle time spent in inspiration (TiTtot) whereas there were no significant differences immediately postran-domization in the PSV group The second measurement of postrandomization respiratory parameters revealed signifi-cant decreases in peak airway pressure and delta pressure in both PAV and PSV groups Adjusting for baseline measure-ments as a covariate the PAV group received 7 cm H2

O∙s less ventilator assistance (PaoArea) per liter of minute ventilation (p lt 00001) and increased their VT by 115 mLcm H

2O∙s

relative to the PSV group (p lt 0001)PAV significantly reduced AI postrandomization (Fig 2A)

In total nine patients (18) had a high AI (gt 10 of breaths) on PSV or on AC mode at baseline No patients had a high AI on PAV In the PSV group six patients (26) had a high AI

at baseline four (17) immediately postrandomization and one patient (7) after further adjustment of ventilator settings according to the PSV protocol (Fig 2B) Adjusting for baseline differences patients in the PAV group tended to have a lower AI than patients in the PSV group immediately postrandom-ization but this did not reach statistical significance (adjusted difference ndash45 [95 CI ndash95 to 04] p = 007)

Clinical OutcomesTable 3 shows the results for the main clinical endpoints Weaning duration and time to successful SBT did not dif-fer significantly among survivors between the groups Among ICU survivors median time to ICU discharge was shorter in the group randomized to PAV 73 days (IQR 52ndash114 d) ver-sus 124 days (IQR 75ndash308 d) in the PSV group (p = 003) Kaplan-Meier plots of the proportion of at-risk patients achiev-ing successful extubation and live ICU discharge are shown in Figure 3 A and B No difference was found for hospital length of stay ICU mortality or hospital mortality Four patients (15) in the PAV group and three patients (13) in the PSV group died in ICU (respiratory rate [RR] of PAV 114 95 CI 028 to 46 p = 086) Overall in-hospital mortality was nine patients (33) in the PAV group and six patients (26) in the PSV group (RR of PAV 128 95 CI 054 to 31 p = 058)

Cointerventions Sedating MedicationsTable 4 compares the use of medications that may prolong weaning (by causing sedation or neuromuscular weakness) in the two groups pre- and postrandomization Type and dose of sedative medication were highly variable among individual patients and therefore we analyzed pre- and postrandomiza-tion changes to determine the impact of the ventilation pro-tocol on sedative use within each patient In the PAV group benzodiazepine narcotic and propofol use measured as both cumulative dose and average daily dose all decreased signifi-cantly postrandomization as did neuromuscular blocker use In the PSV group average daily dose of lorazepam equiva-lent and average daily dose of propofol decreased signifi-cantly postrandomization whereas the proportion of patients

Figure 2 A Asynchrony index pre- and postrandomization on proportional assist ventilation B Asynchrony index pre- and postrandomization on pressure support ventilationPAV群 PSV群

Asynchrony Index () = 同調イベントの回数全呼吸回数 times 100

IntensiveCareMed(2006)321515ndash1522

AI>10は3名いたがPAVに設定変更後0名となった

AI>10は6名いたがPSVに設定変更後4名となった

Clinical Outcomes

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

成功したSBTまでの数抜管成功までの数に変わりなしICU退室までの数はPAV群で有意に短い

Kaplan-Meier curves

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

抜管成功までの時間 ICU退室までの時間

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 9

TABLE 4 Comparison of Medications That May Alter Weaning Duration Pre- and Postrandomization

Variables

PAV Group (n = 27)

PSV Group (n = 23) Postrandomization

Pre randomization Pos trandomization Pre randomization Pos trandomizationAdjusted Difference PSV ndash PAV (95 CI)

Lorazepam equivalents

Cumulative dose mg

308 (207) 48 (257)a 137 (440) 54 (258) 237 (ndash580 to 1055)

Average daily dose mg

44 (255) 09 (43)a 23 (85) 06 (40)a 24 (ndash62 to 110)

Single-day dose mgb

08 (72) 12 (50) 16 (70) 10 (64) 32 (ndash13 to 78)

Morphine equivalents

Cumulative dose mg

5003 (6188) 1140 (3140)b 1710 (13310) 1610 (4860) 591 (ndash2515 to 3696)

Average daily dose mg

858 (1638) 170 (578)c 428 (1724) 136 (320)d 108 (ndash316 to 532)

Single-day dose mga

180 (1240) 187 (77) 100 (1920) 50 (630)d 334 (ndash026 to 670)d

Propofol

Cumulative dose mg

7400 (55500) 0 (700)a 29450 (85800) 0 (27660) 226 (ndash42156 to 42607)

Average daily dose mg

529 (11690) 0 (53)a 4879 (15980) 0 (2738)a 651 (ndash3211 to 4514)

Single-day dose mga

0 (0) 0 (0) 0 (1000) 0 (0) 344 (ndash3781 to 4469)

Neuroleptics

Days receiving neuroleptics n

0 (2) 0 (4) 0 (0) 1 (4)a ndash16 (ndash34 to 03)

Proportion of patients who received a neuroleptic

407 407 174 565b

NMBs

Days receiving NMBs n

1 (1) 0 (0)a 0 (1) 0 (0) ndash01 (ndash05 to 03)

Proportion of patients who received an NMB

518 148a 348 174

Corticosteroids

Days receiving corticosteroids n

0 (0) 0 (0) 0 (2) 0 (2) 02 (ndash06 to 10)

Proportion of patients who received corticosteroids

185 148 348 261

PAV = proportional assist ventilation PSV = pressure support ventilation NMB = neuromuscular blockera p lt 005b Single-day dose = total dose given day prior to randomization or total dose give day after randomizationc p lt 0005d p lt 00005e p le 007All data points except percentage of patients receiving neuroleptics NMBs or corticosteroids are presented as median (interquartile range)

鎮静剤鎮痛剤についてはPAV群とPSV群で変わりなし

その他bull 不整脈気胸プロトコル施中の死亡

は両群とも起こらなかった

筆者らの考察bull PSVはweaningに最も使われているモードだ

がそれと較しても遜ないweaningができることがわかった

bull アシストが少ない傾向であったがPAV群のが呼吸窮迫となる回数が少なかった

rarr同調性が良くなった結果の可能性

ICU退室までの期間はPAV群で短いという結果であったがpowerが少ない

筆者らの考察bull 単施設での結果であり他の施設にこの

新しいプロトコルそのものが受けれられるかは不明である

bull SBTに失敗した患者を主に対象としておりICU患者すべてに適応はできない

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる

Page 2: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

本の論

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Critical Care Medicine wwwccmjournalorg 1

Objectives Despite protocols incorporating spontaneous breath-ing trials 31 of ICU patients experience difficult or prolonged weaning from mechanical ventilation Nonfatiguing modes such as pressure support ventilation are recommended Proportional assist ventilation provides assistance in proportion to patient effort which may optimize weaning However it is not known how proportional assist ventilation performs relative to pressure sup-port ventilation over a prolonged period in the complex ICU set-ting The purpose of this study was to compare the physiologic and clinical performance (failure rate) safety and feasibility of protocols using daily spontaneous breathing trial plus pressure support ventilation versus proportional assist ventilation until ven-tilation discontinuationDesign Single-center unblinded pilot randomized controlled trialSetting Medical-surgical ICU of a tertiary-care hospitalPatients Adult patients intubated greater than 36 hours were ran-domized if they met eligibility criteria for partial ventilatory support

tolerated pressure support ventilation greater than or equal to 30 minutes and either failed or did not meet criteria for a sponta-neous breathing trialInterventions Patients were randomized to the pressure support ventilation or proportional assist ventilation protocol (PAV+ Puri-tan Bennett 840 Covidien Boulder CO) Both protocols used progressive decreases in level of assistance as tolerated coupled with daily assessment for spontaneous breathing trialsMeasurements and Main Results Of 54 patients randomized outcome data are available for 50 patients 27 were randomized to receive proportional assist ventilation and 23 to receive pres-sure support ventilation There were no adverse events linked to the study interventions and protocol violations were infrequent Recruitment was slower than projected (13 patients per month) The median (interquartile range) time from randomization to suc-cessful extubation was 39 days (28ndash84 d) on proportional assist ventilation versus 49 days (29ndash263 d) on pressure support ventilation (p = 039) Time to live ICU discharge was 73 days (52ndash114 d) on proportional assist ventilation versus 124 days (75ndash308 d) on pressure support ventilation (p = 003)Conclusion This pilot study demonstrates the utility safety and feasibility of the weaning protocols and provides important infor-mation to guide the design of a future randomized controlled trial comparing weaning from mechanical ventilation on pressure sup-port ventilation versus proportional assist ventilation (Crit Care Med 2016 XX00ndash00)Key Words assisted mechanical ventilation asynchrony index proportional assist ventilation respiratory muscles ventilator weaning

Difficult or prolonged weaning from mechanical ventilation (MV) affects 31 (range 26ndash42) of mechanically ventilated patients (1) and increases

patient morbidity mortality ICU length of stay and cost to the healthcare system (2) When initial weaning attempts fail to liberate patients from MV ideally clinicians should choose

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights ReservedDOI 101097CCM0000000000001600

1London Health Sciences Centre London ON Canada2Department of Medicine The University of Western Ontario London ON Canada

This study was performed at London Health Sciences Centre University Hospital London CanadaSupplemental digital content is available for this article Direct URL cita-tions appear in the printed text and are provided in the HTML and PDF versions of this article on the journalrsquos website (httpjournalslwwcomccmjournal)Supported in part by Critical Care Western Department of Medicine The University of Western Ontario Program of Experimental Medi-cine (POEM) Department of Medicine University of Western Ontario Lawson Health Research Institute London Health Sciences Centre London Canada and Covidien Boulder USADr Bosma has received honourarium for lectures and travel reimburse-ments from Covidien Dr Bosmarsquos institution has received grant support from Covidien for Dr Bosmarsquos research studies The remaining authors have disclosed that they do not have any potential conflicts of interestFor information regarding this article E-mail KarenJBosmalhsconca

A Pilot Randomized Trial Comparing Weaning From Mechanical Ventilation on Pressure Support Versus Proportional Assist Ventilation

Karen J Bosma MD FRCPC12 Brooke A Read MHS BSc RRT1 Mohammad J Bahrgard Nikoo MD FRCPC12 Philip M Jones MD FRCPC MSc12 Fran A Priestap MSc1 James F Lewis MD FRCPC12

Crit Care Med 2016 Jan 20[Epub ahead of print]

呼吸器離脱の定義分類 定義

Simple weaning 最初のSBTで呼吸器離脱する患者Difficult weaning 最3回までのSBTあるいは1回

のSBTから呼吸器離脱までに最7間かかる患者

Prolonged weaning 4回以上のSBTあるいは最初のSBTから呼吸器離脱までに8間以上かかる患者

Eur Respir J 2007 29 1033-1056

呼吸器離脱困難bull Difficult weaning prolonged weaningは

呼吸器装着患者の31に起こる

bull それらは患者のICU滞在数の延死亡率の増加コスト増につながる

bull PSVがSBTに失敗した患者がweaningをう際に最も使われているモードである

bull PSVはトリガーした後は受動的であり横隔膜の萎縮につながる可能性同調が多い可能性がある

EurRespirJ2007291033ndash1056

JAMA2002287345ndash355

Chest19941061188ndash1193

JApplPhysiol(1985)199681426ndash436

IntensiveCareMed2008341477ndash1486

呼吸器離脱困難患者に適したモードはあるのか

従来の呼吸器モード

医師 呼吸器 患者モード

1回換気量呼吸数

Closed-loop Ventilation

医師 呼吸器 患者 サポート

気道抵抗コンプライアンスを測定患者が求める1回換気量流速を計測する

Closed-loop Ventilationbull 患者の吸気努に例したサポートをうモード

Proportional Assist Ventilation(PAV)rarrPAV+(PB840980) PPS(Evita)

Neurally Adjusted Ventilatory Assist(NAVA)rarrServo i

bull 動weaningシステムにより離脱を指すモードSmartCare

rarrEvita XLinfinity V500Adaptive Support Ventilation(ASV)

rarrHammilton G5

患者の吸気努に例してサポートする

conventional ventilation The patient still has to generatesufficient flow to trigger the ventilator However as dis-cussed later conventional modes may force a larger tidalvolume than normal causing air trapping and thus increas-ing the number of missed triggers

Neurally Adjusted Ventilatory Assist

NAVA accomplishes the same goals as PAV but uti-lizes measurement of the diaphragmatic EMG signal tocontrol gas delivery6 This is accomplished by the place-ment of a specifically designed nasogastric tube that has aseries of EMG electrodes near its distal end positionedacross the diaphragm As illustrated in Figure 1 support-ing ventilation based on the diaphragmatic EMG signalshould greatly improve the response of the ventilator andsynchrony since the signal recognized is high up on theneural pathway controlling ventilation7 What is set by theclinician is the pressure applied for each millivolt of EMGactivity Thus similar to PAV a portion of the ventilatoryeffort is proportionally provided by the ventilator and theremainder by the patient As EMG activity increases pres-sure is applied during the inspiratory phase and as thediaphragm relaxes airway pressure decreases Inspirationends at a specific percentage of the peak EMG activityContrary to PAV NAVA greatly improves triggering since

gas delivery begins when the diaphragm is simulated notas a result of flow in the airway Thus even in the pres-ence of severe air trapping or large system leaks trigger-ing is not compromised

PAV and NAVA Versus Traditional Gas Delivery

Figure 2 nicely illustrates the relationship between andthe patientrsquos effort or work of breathing and the ventilatorpressure during various ventilation modes1 During a spe-cific breath with all forms of volume ventilation there isan indirect relationship between patient effort (Pmus) andthe pressure applied by the ventilator Essentially the morework assumed by the patient the less work applied by theventilator This relationship creates difficult problems re-garding synchrony which can easily be illustrated with alung model with proper setting of inspiratory demand Ifdemand is high and the lung model tidal volume is greaterthan the ventilator set tidal volume no pressure abovebaseline will be applied to the lung

In pressure-targeted ventilation regardless of mode ven-tilator work or applied pressure should be independent ofPmus That is during a specific breath a constant pressureshould be applied to the airway during inspiration nomatter how much work or Pmus is generated by the patientOf course in patients with very high ventilatory demandpressure may not be sustained since the ventilator may notalways be able to meet patient flow demand

In PAV and NAVA there is a direct relationship be-tween the patientrsquos effort (Pmus) and ventilator pressure

Fig 1 Steps in the process of activating a ventilator breath Neu-rally adjusted ventilatory assist ventilation modes are activatedand controlled by diaphragm excitation Proportional assist ven-tilation modes are activated by a change in airway pressure flowand volume (Adapted from Reference 6)

Fig 2 Relationship between patient effort and ventilator pressureduring various ventilation modes During volume ventilation pa-tient effort and ventilator pressure are inversely related Duringpressure ventilation ventilator pressure in unaffected by patienteffort During proportional assist ventilation (PAV) and neurally ad-justed ventilatory assist (NAVA) patient effort and ventilator pres-sure are directly related (Adapted from Reference 5)

PROPORTIONAL ASSIST VENTILATION AND NEURALLY ADJUSTED VENTILATORY ASSIST

142 RESPIRATORY CARE bull FEBRUARY 2011 VOL 56 NO 2

RespirCare201156(2)140ndash148

Proportional Assist Ventilation

bull コンプライアンス(C)とレジスタンス(R)を測定して患者の吸気努に例したサポートをう

bull Paw = VC + RtimesFlowbull WOB = ʃ PtimesFlow dt

bull PB840では8-15呼吸ごとに300msecの吸気ポーズをいCとRを測定

設定するのはサポート率のみ

患者の呼吸仕事量(WOBpt)が03〜08JLになるようにサポート率を設定する

PAVのメリットデメリット

bull メリット患者-呼吸器の同調性が良い呼吸筋を適度に使うことができる

bull デメリット患者の呼吸努が弱い場合はサポートが減ってしまうリークがある場合過度にサポート圧が上がってしまう可能性(runaway現象)

PAVのエビデンスbull PAVとPSVによるウィーニングをい48

時間観察したRCTではPAVのが調節呼吸に戻る割合が低かった(PAV 11 vsPSV 22 p=004)

bull 夜間PAVにすることで睡眠の質が改善したIntensiveCareMed2008342026ndash2034

Crit CareMed2007351048ndash1054

PAVは呼吸器離脱困難患者のウィーニングにおいて

優れたモードか

本の論

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Critical Care Medicine wwwccmjournalorg 1

Objectives Despite protocols incorporating spontaneous breath-ing trials 31 of ICU patients experience difficult or prolonged weaning from mechanical ventilation Nonfatiguing modes such as pressure support ventilation are recommended Proportional assist ventilation provides assistance in proportion to patient effort which may optimize weaning However it is not known how proportional assist ventilation performs relative to pressure sup-port ventilation over a prolonged period in the complex ICU set-ting The purpose of this study was to compare the physiologic and clinical performance (failure rate) safety and feasibility of protocols using daily spontaneous breathing trial plus pressure support ventilation versus proportional assist ventilation until ven-tilation discontinuationDesign Single-center unblinded pilot randomized controlled trialSetting Medical-surgical ICU of a tertiary-care hospitalPatients Adult patients intubated greater than 36 hours were ran-domized if they met eligibility criteria for partial ventilatory support

tolerated pressure support ventilation greater than or equal to 30 minutes and either failed or did not meet criteria for a sponta-neous breathing trialInterventions Patients were randomized to the pressure support ventilation or proportional assist ventilation protocol (PAV+ Puri-tan Bennett 840 Covidien Boulder CO) Both protocols used progressive decreases in level of assistance as tolerated coupled with daily assessment for spontaneous breathing trialsMeasurements and Main Results Of 54 patients randomized outcome data are available for 50 patients 27 were randomized to receive proportional assist ventilation and 23 to receive pres-sure support ventilation There were no adverse events linked to the study interventions and protocol violations were infrequent Recruitment was slower than projected (13 patients per month) The median (interquartile range) time from randomization to suc-cessful extubation was 39 days (28ndash84 d) on proportional assist ventilation versus 49 days (29ndash263 d) on pressure support ventilation (p = 039) Time to live ICU discharge was 73 days (52ndash114 d) on proportional assist ventilation versus 124 days (75ndash308 d) on pressure support ventilation (p = 003)Conclusion This pilot study demonstrates the utility safety and feasibility of the weaning protocols and provides important infor-mation to guide the design of a future randomized controlled trial comparing weaning from mechanical ventilation on pressure sup-port ventilation versus proportional assist ventilation (Crit Care Med 2016 XX00ndash00)Key Words assisted mechanical ventilation asynchrony index proportional assist ventilation respiratory muscles ventilator weaning

Difficult or prolonged weaning from mechanical ventilation (MV) affects 31 (range 26ndash42) of mechanically ventilated patients (1) and increases

patient morbidity mortality ICU length of stay and cost to the healthcare system (2) When initial weaning attempts fail to liberate patients from MV ideally clinicians should choose

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights ReservedDOI 101097CCM0000000000001600

1London Health Sciences Centre London ON Canada2Department of Medicine The University of Western Ontario London ON Canada

This study was performed at London Health Sciences Centre University Hospital London CanadaSupplemental digital content is available for this article Direct URL cita-tions appear in the printed text and are provided in the HTML and PDF versions of this article on the journalrsquos website (httpjournalslwwcomccmjournal)Supported in part by Critical Care Western Department of Medicine The University of Western Ontario Program of Experimental Medi-cine (POEM) Department of Medicine University of Western Ontario Lawson Health Research Institute London Health Sciences Centre London Canada and Covidien Boulder USADr Bosma has received honourarium for lectures and travel reimburse-ments from Covidien Dr Bosmarsquos institution has received grant support from Covidien for Dr Bosmarsquos research studies The remaining authors have disclosed that they do not have any potential conflicts of interestFor information regarding this article E-mail KarenJBosmalhsconca

A Pilot Randomized Trial Comparing Weaning From Mechanical Ventilation on Pressure Support Versus Proportional Assist Ventilation

Karen J Bosma MD FRCPC12 Brooke A Read MHS BSc RRT1 Mohammad J Bahrgard Nikoo MD FRCPC12 Philip M Jones MD FRCPC MSc12 Fran A Priestap MSc1 James F Lewis MD FRCPC12

Crit Care Med 2016 Jan 20[Epub ahead of print]

論のPICO

P 呼吸器離脱困難な患者(simple weaningではない)

I PAVを使

C PSVを使

O 呼吸器離脱までの時間

Study Designbull 単施設盲検前向きランダム化較

試験(pilot study)bull カナダロンドンにあるUniversity

Hospital-London Health Science Centre の内科外科ICU(20床)

bull ランダム割付はcomputer-generated random number sequence によってわれ隠蔽化されている

PatientsInclusion criteria

bull 18歳以上の挿管患者bull 36時間以上呼吸器管理をされている

PatientsExclusion criteria

1) 48時間以内にwithdrawが考慮される患者2) 位の脊髄障害進性の神経筋疾患の

よって慢性的な呼吸器依存となる可能性があるまたは脳外科的介が必要である患者

Enrolment Procedure

Supplemental Digital Content Figure 1 Enrolment Procedure

Daily Screening of all ICU patients gt18 yrs of age Intubated gt36 hours

Exclude if Exclusion Criteria

Successful SBT awaiting extubation Withdrawal of life support considered High spinal cord injury Neurosurgical patient Neuromuscular disease

Follow using daily eligibility checklist until meet all Clinical Stability Criteria to trial mode of partial ventilatory support (PAV or PSV)

partialcomplete reversal of cause of respiratory failure

body temperature between 36ordmC and 39ordmC metabolic disorders corrected pH gt732 hemoglobin gt70 gL no ongoing bleeding stable hemodynamic status Able to trigger PaO2 ge 60 mmHg or SaO2 ge 90 on FiO2le

060 and PEEP le 15 cm H2O ACVC Pplatle30cmH2O with Vt 6-8 mlkg

OR ACPC PC+PEEPle30cmH2O

Exclude if Exclusion Criteria Has met all Criteria for Switching to

Spontaneous Mode of Ventilation gt24 hours

Pre-inclusion Test on PSV 15 cmH2O for 30 minutes

Patient fails Pre-inclusion test Return to AC ventilation and reassess within 24 hour

Patient passes Pre-inclusion test Assess for Criteria to Initiate Weaning

PaO2FiO2gt200 on FiO2 le050 and PEEP le8 cmH2O No inotropesvasopressors except low dose dopamine

RSBI lt105 Patient passes Pre-SBT readiness assessment Proceed to SBT Place patient on CPAP 5 cmH2O and FiO2 040 for 30 minutes

Patient does not meet Criteria to Initiate Weaning

Patient meets Criteria to Initiate Weaning Proceed to Pre-SBT readiness assessment Place patient on CPAP for 1-2 min measure RSBI

RSBIge 105 Patient fails Pre-SBT readiness assessment

Patient passes SBT Patient fails SBT

Exclude if Exclusion Criteria

Passed SBT and tolerating PSVle7cmH2O and PEEPle5 cmH2O while awaiting extubation

Obtain informed consent from Patient or SDM as soon as available (deferred consent permitted)

Randomize

Exclude if Exclusion Criteria

Patient SDM refuse consent

bull 呼吸不全の原因に改善が認められてきている

bull 体温は36〜39度であるbull pH>732であるbull Hb>7gdLで現在出なしbull 動態が安定しているbull トリガーできるbull FiO2≦06 PEEP≦15cmH2Oで

PaO2≧60mmHg or SaO2≧90bull ACVCでVt 6-8mlkgで

Pplat≦30cmH2OもしくはACPCでPC+PEEP≦30cmH2O

上記をすべて満たせばプレテストとしてPSV 15cmH2Oに変更

Enrolment Procedure

Supplemental Digital Content Figure 1 Enrolment Procedure

Daily Screening of all ICU patients gt18 yrs of age Intubated gt36 hours

Exclude if Exclusion Criteria

Successful SBT awaiting extubation Withdrawal of life support considered High spinal cord injury Neurosurgical patient Neuromuscular disease

Follow using daily eligibility checklist until meet all Clinical Stability Criteria to trial mode of partial ventilatory support (PAV or PSV)

partialcomplete reversal of cause of respiratory failure

body temperature between 36ordmC and 39ordmC metabolic disorders corrected pH gt732 hemoglobin gt70 gL no ongoing bleeding stable hemodynamic status Able to trigger PaO2 ge 60 mmHg or SaO2 ge 90 on FiO2le

060 and PEEP le 15 cm H2O ACVC Pplatle30cmH2O with Vt 6-8 mlkg

OR ACPC PC+PEEPle30cmH2O

Exclude if Exclusion Criteria Has met all Criteria for Switching to

Spontaneous Mode of Ventilation gt24 hours

Pre-inclusion Test on PSV 15 cmH2O for 30 minutes

Patient fails Pre-inclusion test Return to AC ventilation and reassess within 24 hour

Patient passes Pre-inclusion test Assess for Criteria to Initiate Weaning

PaO2FiO2gt200 on FiO2 le050 and PEEP le8 cmH2O No inotropesvasopressors except low dose dopamine

RSBI lt105 Patient passes Pre-SBT readiness assessment Proceed to SBT Place patient on CPAP 5 cmH2O and FiO2 040 for 30 minutes

Patient does not meet Criteria to Initiate Weaning

Patient meets Criteria to Initiate Weaning Proceed to Pre-SBT readiness assessment Place patient on CPAP for 1-2 min measure RSBI

RSBIge 105 Patient fails Pre-SBT readiness assessment

Patient passes SBT Patient fails SBT

Exclude if Exclusion Criteria

Passed SBT and tolerating PSVle7cmH2O and PEEPle5 cmH2O while awaiting extubation

Obtain informed consent from Patient or SDM as soon as available (deferred consent permitted)

Randomize

Exclude if Exclusion Criteria

Patient SDM refuse consent

呼吸窮迫がなければプレテストクリアbull FiO2≦05 PEEP≦8cmH2OでPF≧200bull 強薬昇圧剤の使なしであればSBTへ

まず1-2分CPAP 5cmH2OにしてRSBIを測定

RSBI<105であればSBTへPSV 5cmH2Oにして30-120分

酸素化の条件を満たさないRSBI≧105SBT失敗であれば割り付けへ

Ventilation Protocols bull 呼吸器はPuritan Bennett 840 を使bull 後述するPAVプロトコールPSVプロトコー

ルに従って呼吸療法(RRT)が設定を変更(respiratory distress 呼吸窮迫がないようにできるだけサポートを下げていく)

bull 設定したmaxのサポートが必要になった場合はACに戻す

bull PAV群はPAV+とACPSV群はPSVとACしか使しない

Ventilation Protocols bull RCTの開始前に15名の患者をいて観察

研究をいRRTにプロトコルの確認をってもらった

bull プロトコルはすぐにられるようにラミネートしてベッドサイドに置いた

bull RRTは毎SBTをえるかどうか後述するチェックリストをいて確認除外項を満たさない場合は毎RSBIのアセスメントとSBTをう

Respiratory Distress bull 呼吸数>35回minbull 脈拍>140min もしくはベースラインより20以上の

上昇bull 収縮期圧>180mmHg もしくは<90mmHg ベース

ラインより30以上の変化bull 不安bull 発汗bull 副呼吸筋の使bull 奇異呼吸bull 呼吸困難の訴え

PAV protocolSupplemental Digital Content Figure 2a Proportional Assist Ventilation Protocol

bull 開始の設定はサポート率70

bull 2-3時間おきにRRTがアセスメントし問題なければ10-20ずつPAVサポートを下げる

サポート率90まで上げてもダメな場合はACに戻す

PSV protocolSupplemental Digital Content Figure 2b Pressure Support Ventilation Protocol

bull 開始の設定はPS 15cmH2O

bull まずRRlt35 VTgt5mLkgで耐えられるPSまで下げる

bull 2-3時間おきにRRTがアセスメントしてPS2-3cmH2Pずつ下げる

Data collection form for RRTsSupplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

SBTの開始基準bull PF≧200bull FiO2≦05bull pH≧732bull トリガーできるbull 24時間以内に筋虚イベントなしbull 強薬昇圧剤の使なし

Data collection form for RRTs

Supplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

CPAPにしてRSBIを測定RSBI<105であればSBT施

PAV群はサポート率20PSV群はPS 5cmH2OでSBTを施

PEEP≦8cmH2OでSaO2≧90呼吸窮迫なしであればSBTクリア

抜管へ

Endpoints 呼吸器離脱までの時間ICU病院滞在期間

呼吸器離脱成功の定義抜管後48時間再挿管なし侵襲的呼吸(NIV)の使もなし気管切開使時は呼吸器から最終的に外れた時点1のうち12時間以上NIVを使していない

Statistical Analysis bull Pilot studyのためサンプルサイズの計

算はしていないbull P値<005を統計学的有意としたbull Intention to treat approachをった

RESULTS

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial

エンロール期間は42か80名が基準を満たしが最終的に割付されたのは54名

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial最終的に解析がわれたのはPAV群27名PSV群23名

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

ベースラインは両群間で変わりなし重症度はAPACHEⅡ 27点程度患者は肺炎が3割もともとCOPDを持っている患者は2割程度

Feasibility Outcomes bull 呼吸窮迫でRRTが呼吸器サポートを上昇

させた回数はPAV群で76回232daysPSV群で124回295daysと有意にPSV群で多かった(p=0002)

bull ACに戻した回数はPAV群で66回PSV群で90回と変わりなし(p=061)

bull 全くACに戻さなかった患者はPAV群で627名PSV群で623名(p=10)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

ベースラインは両群間で変わりなしPAVにすることでΔP圧Peak圧が低く吸気時間呼吸時間がくなるPSV群と較するとPAV群のがPaoAreaPaoAreaVE(呼吸器のサポートの強さをす)が低くVTPaoAreaがくsecond post-randomizationの解析においてΔP圧Peak圧が低い

Asynchrony index

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 7

B635) documenting compliance with the daily screening for SBT noncompletion of either the CPAP screening test or SBT occurred on 5 of 292 of patient study days (17) and this violation involved two patients in each study arm The second audit examining for nine different potential deviations found complete adherence to study protocol on 254 of 292 of patient study days (900) There were no adverse events (pneumo-thorax arrhythmia or death) resulting from any of the study interventions

Physiologic Response to Ventilation ProtocolsRecordings of flow and airway pressure were available for analysis for all 50 patients prerandomization 49 patients immediately (median 1 interquartile range [IQR] 0ndash15 hr) postrandomization and for 26 patients during subsequent days (median 27 IQR 20ndash70-hr postrandomization) Approximately 15000 breaths in 125 recordings were analyzed Table 2 compares the respiratory parameters in both groups pre- and postrandomization There were no statistically sig-nificant differences between the PAV and PSV groups at base-line prerandomization (p values not shown) In the PAV group there was a significant decrease in peak airway pressure delta pressure (measured as peak airway opening pressure minus the PEEP) and area under the inspiratory pressure curve immediately postrandomization and an increase in the per-cent of total cycle time spent in inspiration (TiTtot) whereas there were no significant differences immediately postran-domization in the PSV group The second measurement of postrandomization respiratory parameters revealed signifi-cant decreases in peak airway pressure and delta pressure in both PAV and PSV groups Adjusting for baseline measure-ments as a covariate the PAV group received 7 cm H2

O∙s less ventilator assistance (PaoArea) per liter of minute ventilation (p lt 00001) and increased their VT by 115 mLcm H

2O∙s

relative to the PSV group (p lt 0001)PAV significantly reduced AI postrandomization (Fig 2A)

In total nine patients (18) had a high AI (gt 10 of breaths) on PSV or on AC mode at baseline No patients had a high AI on PAV In the PSV group six patients (26) had a high AI

at baseline four (17) immediately postrandomization and one patient (7) after further adjustment of ventilator settings according to the PSV protocol (Fig 2B) Adjusting for baseline differences patients in the PAV group tended to have a lower AI than patients in the PSV group immediately postrandom-ization but this did not reach statistical significance (adjusted difference ndash45 [95 CI ndash95 to 04] p = 007)

Clinical OutcomesTable 3 shows the results for the main clinical endpoints Weaning duration and time to successful SBT did not dif-fer significantly among survivors between the groups Among ICU survivors median time to ICU discharge was shorter in the group randomized to PAV 73 days (IQR 52ndash114 d) ver-sus 124 days (IQR 75ndash308 d) in the PSV group (p = 003) Kaplan-Meier plots of the proportion of at-risk patients achiev-ing successful extubation and live ICU discharge are shown in Figure 3 A and B No difference was found for hospital length of stay ICU mortality or hospital mortality Four patients (15) in the PAV group and three patients (13) in the PSV group died in ICU (respiratory rate [RR] of PAV 114 95 CI 028 to 46 p = 086) Overall in-hospital mortality was nine patients (33) in the PAV group and six patients (26) in the PSV group (RR of PAV 128 95 CI 054 to 31 p = 058)

Cointerventions Sedating MedicationsTable 4 compares the use of medications that may prolong weaning (by causing sedation or neuromuscular weakness) in the two groups pre- and postrandomization Type and dose of sedative medication were highly variable among individual patients and therefore we analyzed pre- and postrandomiza-tion changes to determine the impact of the ventilation pro-tocol on sedative use within each patient In the PAV group benzodiazepine narcotic and propofol use measured as both cumulative dose and average daily dose all decreased signifi-cantly postrandomization as did neuromuscular blocker use In the PSV group average daily dose of lorazepam equiva-lent and average daily dose of propofol decreased signifi-cantly postrandomization whereas the proportion of patients

Figure 2 A Asynchrony index pre- and postrandomization on proportional assist ventilation B Asynchrony index pre- and postrandomization on pressure support ventilationPAV群 PSV群

Asynchrony Index () = 同調イベントの回数全呼吸回数 times 100

IntensiveCareMed(2006)321515ndash1522

AI>10は3名いたがPAVに設定変更後0名となった

AI>10は6名いたがPSVに設定変更後4名となった

Clinical Outcomes

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

成功したSBTまでの数抜管成功までの数に変わりなしICU退室までの数はPAV群で有意に短い

Kaplan-Meier curves

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

抜管成功までの時間 ICU退室までの時間

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 9

TABLE 4 Comparison of Medications That May Alter Weaning Duration Pre- and Postrandomization

Variables

PAV Group (n = 27)

PSV Group (n = 23) Postrandomization

Pre randomization Pos trandomization Pre randomization Pos trandomizationAdjusted Difference PSV ndash PAV (95 CI)

Lorazepam equivalents

Cumulative dose mg

308 (207) 48 (257)a 137 (440) 54 (258) 237 (ndash580 to 1055)

Average daily dose mg

44 (255) 09 (43)a 23 (85) 06 (40)a 24 (ndash62 to 110)

Single-day dose mgb

08 (72) 12 (50) 16 (70) 10 (64) 32 (ndash13 to 78)

Morphine equivalents

Cumulative dose mg

5003 (6188) 1140 (3140)b 1710 (13310) 1610 (4860) 591 (ndash2515 to 3696)

Average daily dose mg

858 (1638) 170 (578)c 428 (1724) 136 (320)d 108 (ndash316 to 532)

Single-day dose mga

180 (1240) 187 (77) 100 (1920) 50 (630)d 334 (ndash026 to 670)d

Propofol

Cumulative dose mg

7400 (55500) 0 (700)a 29450 (85800) 0 (27660) 226 (ndash42156 to 42607)

Average daily dose mg

529 (11690) 0 (53)a 4879 (15980) 0 (2738)a 651 (ndash3211 to 4514)

Single-day dose mga

0 (0) 0 (0) 0 (1000) 0 (0) 344 (ndash3781 to 4469)

Neuroleptics

Days receiving neuroleptics n

0 (2) 0 (4) 0 (0) 1 (4)a ndash16 (ndash34 to 03)

Proportion of patients who received a neuroleptic

407 407 174 565b

NMBs

Days receiving NMBs n

1 (1) 0 (0)a 0 (1) 0 (0) ndash01 (ndash05 to 03)

Proportion of patients who received an NMB

518 148a 348 174

Corticosteroids

Days receiving corticosteroids n

0 (0) 0 (0) 0 (2) 0 (2) 02 (ndash06 to 10)

Proportion of patients who received corticosteroids

185 148 348 261

PAV = proportional assist ventilation PSV = pressure support ventilation NMB = neuromuscular blockera p lt 005b Single-day dose = total dose given day prior to randomization or total dose give day after randomizationc p lt 0005d p lt 00005e p le 007All data points except percentage of patients receiving neuroleptics NMBs or corticosteroids are presented as median (interquartile range)

鎮静剤鎮痛剤についてはPAV群とPSV群で変わりなし

その他bull 不整脈気胸プロトコル施中の死亡

は両群とも起こらなかった

筆者らの考察bull PSVはweaningに最も使われているモードだ

がそれと較しても遜ないweaningができることがわかった

bull アシストが少ない傾向であったがPAV群のが呼吸窮迫となる回数が少なかった

rarr同調性が良くなった結果の可能性

ICU退室までの期間はPAV群で短いという結果であったがpowerが少ない

筆者らの考察bull 単施設での結果であり他の施設にこの

新しいプロトコルそのものが受けれられるかは不明である

bull SBTに失敗した患者を主に対象としておりICU患者すべてに適応はできない

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる

Page 3: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

呼吸器離脱の定義分類 定義

Simple weaning 最初のSBTで呼吸器離脱する患者Difficult weaning 最3回までのSBTあるいは1回

のSBTから呼吸器離脱までに最7間かかる患者

Prolonged weaning 4回以上のSBTあるいは最初のSBTから呼吸器離脱までに8間以上かかる患者

Eur Respir J 2007 29 1033-1056

呼吸器離脱困難bull Difficult weaning prolonged weaningは

呼吸器装着患者の31に起こる

bull それらは患者のICU滞在数の延死亡率の増加コスト増につながる

bull PSVがSBTに失敗した患者がweaningをう際に最も使われているモードである

bull PSVはトリガーした後は受動的であり横隔膜の萎縮につながる可能性同調が多い可能性がある

EurRespirJ2007291033ndash1056

JAMA2002287345ndash355

Chest19941061188ndash1193

JApplPhysiol(1985)199681426ndash436

IntensiveCareMed2008341477ndash1486

呼吸器離脱困難患者に適したモードはあるのか

従来の呼吸器モード

医師 呼吸器 患者モード

1回換気量呼吸数

Closed-loop Ventilation

医師 呼吸器 患者 サポート

気道抵抗コンプライアンスを測定患者が求める1回換気量流速を計測する

Closed-loop Ventilationbull 患者の吸気努に例したサポートをうモード

Proportional Assist Ventilation(PAV)rarrPAV+(PB840980) PPS(Evita)

Neurally Adjusted Ventilatory Assist(NAVA)rarrServo i

bull 動weaningシステムにより離脱を指すモードSmartCare

rarrEvita XLinfinity V500Adaptive Support Ventilation(ASV)

rarrHammilton G5

患者の吸気努に例してサポートする

conventional ventilation The patient still has to generatesufficient flow to trigger the ventilator However as dis-cussed later conventional modes may force a larger tidalvolume than normal causing air trapping and thus increas-ing the number of missed triggers

Neurally Adjusted Ventilatory Assist

NAVA accomplishes the same goals as PAV but uti-lizes measurement of the diaphragmatic EMG signal tocontrol gas delivery6 This is accomplished by the place-ment of a specifically designed nasogastric tube that has aseries of EMG electrodes near its distal end positionedacross the diaphragm As illustrated in Figure 1 support-ing ventilation based on the diaphragmatic EMG signalshould greatly improve the response of the ventilator andsynchrony since the signal recognized is high up on theneural pathway controlling ventilation7 What is set by theclinician is the pressure applied for each millivolt of EMGactivity Thus similar to PAV a portion of the ventilatoryeffort is proportionally provided by the ventilator and theremainder by the patient As EMG activity increases pres-sure is applied during the inspiratory phase and as thediaphragm relaxes airway pressure decreases Inspirationends at a specific percentage of the peak EMG activityContrary to PAV NAVA greatly improves triggering since

gas delivery begins when the diaphragm is simulated notas a result of flow in the airway Thus even in the pres-ence of severe air trapping or large system leaks trigger-ing is not compromised

PAV and NAVA Versus Traditional Gas Delivery

Figure 2 nicely illustrates the relationship between andthe patientrsquos effort or work of breathing and the ventilatorpressure during various ventilation modes1 During a spe-cific breath with all forms of volume ventilation there isan indirect relationship between patient effort (Pmus) andthe pressure applied by the ventilator Essentially the morework assumed by the patient the less work applied by theventilator This relationship creates difficult problems re-garding synchrony which can easily be illustrated with alung model with proper setting of inspiratory demand Ifdemand is high and the lung model tidal volume is greaterthan the ventilator set tidal volume no pressure abovebaseline will be applied to the lung

In pressure-targeted ventilation regardless of mode ven-tilator work or applied pressure should be independent ofPmus That is during a specific breath a constant pressureshould be applied to the airway during inspiration nomatter how much work or Pmus is generated by the patientOf course in patients with very high ventilatory demandpressure may not be sustained since the ventilator may notalways be able to meet patient flow demand

In PAV and NAVA there is a direct relationship be-tween the patientrsquos effort (Pmus) and ventilator pressure

Fig 1 Steps in the process of activating a ventilator breath Neu-rally adjusted ventilatory assist ventilation modes are activatedand controlled by diaphragm excitation Proportional assist ven-tilation modes are activated by a change in airway pressure flowand volume (Adapted from Reference 6)

Fig 2 Relationship between patient effort and ventilator pressureduring various ventilation modes During volume ventilation pa-tient effort and ventilator pressure are inversely related Duringpressure ventilation ventilator pressure in unaffected by patienteffort During proportional assist ventilation (PAV) and neurally ad-justed ventilatory assist (NAVA) patient effort and ventilator pres-sure are directly related (Adapted from Reference 5)

PROPORTIONAL ASSIST VENTILATION AND NEURALLY ADJUSTED VENTILATORY ASSIST

142 RESPIRATORY CARE bull FEBRUARY 2011 VOL 56 NO 2

RespirCare201156(2)140ndash148

Proportional Assist Ventilation

bull コンプライアンス(C)とレジスタンス(R)を測定して患者の吸気努に例したサポートをう

bull Paw = VC + RtimesFlowbull WOB = ʃ PtimesFlow dt

bull PB840では8-15呼吸ごとに300msecの吸気ポーズをいCとRを測定

設定するのはサポート率のみ

患者の呼吸仕事量(WOBpt)が03〜08JLになるようにサポート率を設定する

PAVのメリットデメリット

bull メリット患者-呼吸器の同調性が良い呼吸筋を適度に使うことができる

bull デメリット患者の呼吸努が弱い場合はサポートが減ってしまうリークがある場合過度にサポート圧が上がってしまう可能性(runaway現象)

PAVのエビデンスbull PAVとPSVによるウィーニングをい48

時間観察したRCTではPAVのが調節呼吸に戻る割合が低かった(PAV 11 vsPSV 22 p=004)

bull 夜間PAVにすることで睡眠の質が改善したIntensiveCareMed2008342026ndash2034

Crit CareMed2007351048ndash1054

PAVは呼吸器離脱困難患者のウィーニングにおいて

優れたモードか

本の論

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Critical Care Medicine wwwccmjournalorg 1

Objectives Despite protocols incorporating spontaneous breath-ing trials 31 of ICU patients experience difficult or prolonged weaning from mechanical ventilation Nonfatiguing modes such as pressure support ventilation are recommended Proportional assist ventilation provides assistance in proportion to patient effort which may optimize weaning However it is not known how proportional assist ventilation performs relative to pressure sup-port ventilation over a prolonged period in the complex ICU set-ting The purpose of this study was to compare the physiologic and clinical performance (failure rate) safety and feasibility of protocols using daily spontaneous breathing trial plus pressure support ventilation versus proportional assist ventilation until ven-tilation discontinuationDesign Single-center unblinded pilot randomized controlled trialSetting Medical-surgical ICU of a tertiary-care hospitalPatients Adult patients intubated greater than 36 hours were ran-domized if they met eligibility criteria for partial ventilatory support

tolerated pressure support ventilation greater than or equal to 30 minutes and either failed or did not meet criteria for a sponta-neous breathing trialInterventions Patients were randomized to the pressure support ventilation or proportional assist ventilation protocol (PAV+ Puri-tan Bennett 840 Covidien Boulder CO) Both protocols used progressive decreases in level of assistance as tolerated coupled with daily assessment for spontaneous breathing trialsMeasurements and Main Results Of 54 patients randomized outcome data are available for 50 patients 27 were randomized to receive proportional assist ventilation and 23 to receive pres-sure support ventilation There were no adverse events linked to the study interventions and protocol violations were infrequent Recruitment was slower than projected (13 patients per month) The median (interquartile range) time from randomization to suc-cessful extubation was 39 days (28ndash84 d) on proportional assist ventilation versus 49 days (29ndash263 d) on pressure support ventilation (p = 039) Time to live ICU discharge was 73 days (52ndash114 d) on proportional assist ventilation versus 124 days (75ndash308 d) on pressure support ventilation (p = 003)Conclusion This pilot study demonstrates the utility safety and feasibility of the weaning protocols and provides important infor-mation to guide the design of a future randomized controlled trial comparing weaning from mechanical ventilation on pressure sup-port ventilation versus proportional assist ventilation (Crit Care Med 2016 XX00ndash00)Key Words assisted mechanical ventilation asynchrony index proportional assist ventilation respiratory muscles ventilator weaning

Difficult or prolonged weaning from mechanical ventilation (MV) affects 31 (range 26ndash42) of mechanically ventilated patients (1) and increases

patient morbidity mortality ICU length of stay and cost to the healthcare system (2) When initial weaning attempts fail to liberate patients from MV ideally clinicians should choose

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights ReservedDOI 101097CCM0000000000001600

1London Health Sciences Centre London ON Canada2Department of Medicine The University of Western Ontario London ON Canada

This study was performed at London Health Sciences Centre University Hospital London CanadaSupplemental digital content is available for this article Direct URL cita-tions appear in the printed text and are provided in the HTML and PDF versions of this article on the journalrsquos website (httpjournalslwwcomccmjournal)Supported in part by Critical Care Western Department of Medicine The University of Western Ontario Program of Experimental Medi-cine (POEM) Department of Medicine University of Western Ontario Lawson Health Research Institute London Health Sciences Centre London Canada and Covidien Boulder USADr Bosma has received honourarium for lectures and travel reimburse-ments from Covidien Dr Bosmarsquos institution has received grant support from Covidien for Dr Bosmarsquos research studies The remaining authors have disclosed that they do not have any potential conflicts of interestFor information regarding this article E-mail KarenJBosmalhsconca

A Pilot Randomized Trial Comparing Weaning From Mechanical Ventilation on Pressure Support Versus Proportional Assist Ventilation

Karen J Bosma MD FRCPC12 Brooke A Read MHS BSc RRT1 Mohammad J Bahrgard Nikoo MD FRCPC12 Philip M Jones MD FRCPC MSc12 Fran A Priestap MSc1 James F Lewis MD FRCPC12

Crit Care Med 2016 Jan 20[Epub ahead of print]

論のPICO

P 呼吸器離脱困難な患者(simple weaningではない)

I PAVを使

C PSVを使

O 呼吸器離脱までの時間

Study Designbull 単施設盲検前向きランダム化較

試験(pilot study)bull カナダロンドンにあるUniversity

Hospital-London Health Science Centre の内科外科ICU(20床)

bull ランダム割付はcomputer-generated random number sequence によってわれ隠蔽化されている

PatientsInclusion criteria

bull 18歳以上の挿管患者bull 36時間以上呼吸器管理をされている

PatientsExclusion criteria

1) 48時間以内にwithdrawが考慮される患者2) 位の脊髄障害進性の神経筋疾患の

よって慢性的な呼吸器依存となる可能性があるまたは脳外科的介が必要である患者

Enrolment Procedure

Supplemental Digital Content Figure 1 Enrolment Procedure

Daily Screening of all ICU patients gt18 yrs of age Intubated gt36 hours

Exclude if Exclusion Criteria

Successful SBT awaiting extubation Withdrawal of life support considered High spinal cord injury Neurosurgical patient Neuromuscular disease

Follow using daily eligibility checklist until meet all Clinical Stability Criteria to trial mode of partial ventilatory support (PAV or PSV)

partialcomplete reversal of cause of respiratory failure

body temperature between 36ordmC and 39ordmC metabolic disorders corrected pH gt732 hemoglobin gt70 gL no ongoing bleeding stable hemodynamic status Able to trigger PaO2 ge 60 mmHg or SaO2 ge 90 on FiO2le

060 and PEEP le 15 cm H2O ACVC Pplatle30cmH2O with Vt 6-8 mlkg

OR ACPC PC+PEEPle30cmH2O

Exclude if Exclusion Criteria Has met all Criteria for Switching to

Spontaneous Mode of Ventilation gt24 hours

Pre-inclusion Test on PSV 15 cmH2O for 30 minutes

Patient fails Pre-inclusion test Return to AC ventilation and reassess within 24 hour

Patient passes Pre-inclusion test Assess for Criteria to Initiate Weaning

PaO2FiO2gt200 on FiO2 le050 and PEEP le8 cmH2O No inotropesvasopressors except low dose dopamine

RSBI lt105 Patient passes Pre-SBT readiness assessment Proceed to SBT Place patient on CPAP 5 cmH2O and FiO2 040 for 30 minutes

Patient does not meet Criteria to Initiate Weaning

Patient meets Criteria to Initiate Weaning Proceed to Pre-SBT readiness assessment Place patient on CPAP for 1-2 min measure RSBI

RSBIge 105 Patient fails Pre-SBT readiness assessment

Patient passes SBT Patient fails SBT

Exclude if Exclusion Criteria

Passed SBT and tolerating PSVle7cmH2O and PEEPle5 cmH2O while awaiting extubation

Obtain informed consent from Patient or SDM as soon as available (deferred consent permitted)

Randomize

Exclude if Exclusion Criteria

Patient SDM refuse consent

bull 呼吸不全の原因に改善が認められてきている

bull 体温は36〜39度であるbull pH>732であるbull Hb>7gdLで現在出なしbull 動態が安定しているbull トリガーできるbull FiO2≦06 PEEP≦15cmH2Oで

PaO2≧60mmHg or SaO2≧90bull ACVCでVt 6-8mlkgで

Pplat≦30cmH2OもしくはACPCでPC+PEEP≦30cmH2O

上記をすべて満たせばプレテストとしてPSV 15cmH2Oに変更

Enrolment Procedure

Supplemental Digital Content Figure 1 Enrolment Procedure

Daily Screening of all ICU patients gt18 yrs of age Intubated gt36 hours

Exclude if Exclusion Criteria

Successful SBT awaiting extubation Withdrawal of life support considered High spinal cord injury Neurosurgical patient Neuromuscular disease

Follow using daily eligibility checklist until meet all Clinical Stability Criteria to trial mode of partial ventilatory support (PAV or PSV)

partialcomplete reversal of cause of respiratory failure

body temperature between 36ordmC and 39ordmC metabolic disorders corrected pH gt732 hemoglobin gt70 gL no ongoing bleeding stable hemodynamic status Able to trigger PaO2 ge 60 mmHg or SaO2 ge 90 on FiO2le

060 and PEEP le 15 cm H2O ACVC Pplatle30cmH2O with Vt 6-8 mlkg

OR ACPC PC+PEEPle30cmH2O

Exclude if Exclusion Criteria Has met all Criteria for Switching to

Spontaneous Mode of Ventilation gt24 hours

Pre-inclusion Test on PSV 15 cmH2O for 30 minutes

Patient fails Pre-inclusion test Return to AC ventilation and reassess within 24 hour

Patient passes Pre-inclusion test Assess for Criteria to Initiate Weaning

PaO2FiO2gt200 on FiO2 le050 and PEEP le8 cmH2O No inotropesvasopressors except low dose dopamine

RSBI lt105 Patient passes Pre-SBT readiness assessment Proceed to SBT Place patient on CPAP 5 cmH2O and FiO2 040 for 30 minutes

Patient does not meet Criteria to Initiate Weaning

Patient meets Criteria to Initiate Weaning Proceed to Pre-SBT readiness assessment Place patient on CPAP for 1-2 min measure RSBI

RSBIge 105 Patient fails Pre-SBT readiness assessment

Patient passes SBT Patient fails SBT

Exclude if Exclusion Criteria

Passed SBT and tolerating PSVle7cmH2O and PEEPle5 cmH2O while awaiting extubation

Obtain informed consent from Patient or SDM as soon as available (deferred consent permitted)

Randomize

Exclude if Exclusion Criteria

Patient SDM refuse consent

呼吸窮迫がなければプレテストクリアbull FiO2≦05 PEEP≦8cmH2OでPF≧200bull 強薬昇圧剤の使なしであればSBTへ

まず1-2分CPAP 5cmH2OにしてRSBIを測定

RSBI<105であればSBTへPSV 5cmH2Oにして30-120分

酸素化の条件を満たさないRSBI≧105SBT失敗であれば割り付けへ

Ventilation Protocols bull 呼吸器はPuritan Bennett 840 を使bull 後述するPAVプロトコールPSVプロトコー

ルに従って呼吸療法(RRT)が設定を変更(respiratory distress 呼吸窮迫がないようにできるだけサポートを下げていく)

bull 設定したmaxのサポートが必要になった場合はACに戻す

bull PAV群はPAV+とACPSV群はPSVとACしか使しない

Ventilation Protocols bull RCTの開始前に15名の患者をいて観察

研究をいRRTにプロトコルの確認をってもらった

bull プロトコルはすぐにられるようにラミネートしてベッドサイドに置いた

bull RRTは毎SBTをえるかどうか後述するチェックリストをいて確認除外項を満たさない場合は毎RSBIのアセスメントとSBTをう

Respiratory Distress bull 呼吸数>35回minbull 脈拍>140min もしくはベースラインより20以上の

上昇bull 収縮期圧>180mmHg もしくは<90mmHg ベース

ラインより30以上の変化bull 不安bull 発汗bull 副呼吸筋の使bull 奇異呼吸bull 呼吸困難の訴え

PAV protocolSupplemental Digital Content Figure 2a Proportional Assist Ventilation Protocol

bull 開始の設定はサポート率70

bull 2-3時間おきにRRTがアセスメントし問題なければ10-20ずつPAVサポートを下げる

サポート率90まで上げてもダメな場合はACに戻す

PSV protocolSupplemental Digital Content Figure 2b Pressure Support Ventilation Protocol

bull 開始の設定はPS 15cmH2O

bull まずRRlt35 VTgt5mLkgで耐えられるPSまで下げる

bull 2-3時間おきにRRTがアセスメントしてPS2-3cmH2Pずつ下げる

Data collection form for RRTsSupplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

SBTの開始基準bull PF≧200bull FiO2≦05bull pH≧732bull トリガーできるbull 24時間以内に筋虚イベントなしbull 強薬昇圧剤の使なし

Data collection form for RRTs

Supplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

CPAPにしてRSBIを測定RSBI<105であればSBT施

PAV群はサポート率20PSV群はPS 5cmH2OでSBTを施

PEEP≦8cmH2OでSaO2≧90呼吸窮迫なしであればSBTクリア

抜管へ

Endpoints 呼吸器離脱までの時間ICU病院滞在期間

呼吸器離脱成功の定義抜管後48時間再挿管なし侵襲的呼吸(NIV)の使もなし気管切開使時は呼吸器から最終的に外れた時点1のうち12時間以上NIVを使していない

Statistical Analysis bull Pilot studyのためサンプルサイズの計

算はしていないbull P値<005を統計学的有意としたbull Intention to treat approachをった

RESULTS

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial

エンロール期間は42か80名が基準を満たしが最終的に割付されたのは54名

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial最終的に解析がわれたのはPAV群27名PSV群23名

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

ベースラインは両群間で変わりなし重症度はAPACHEⅡ 27点程度患者は肺炎が3割もともとCOPDを持っている患者は2割程度

Feasibility Outcomes bull 呼吸窮迫でRRTが呼吸器サポートを上昇

させた回数はPAV群で76回232daysPSV群で124回295daysと有意にPSV群で多かった(p=0002)

bull ACに戻した回数はPAV群で66回PSV群で90回と変わりなし(p=061)

bull 全くACに戻さなかった患者はPAV群で627名PSV群で623名(p=10)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

ベースラインは両群間で変わりなしPAVにすることでΔP圧Peak圧が低く吸気時間呼吸時間がくなるPSV群と較するとPAV群のがPaoAreaPaoAreaVE(呼吸器のサポートの強さをす)が低くVTPaoAreaがくsecond post-randomizationの解析においてΔP圧Peak圧が低い

Asynchrony index

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 7

B635) documenting compliance with the daily screening for SBT noncompletion of either the CPAP screening test or SBT occurred on 5 of 292 of patient study days (17) and this violation involved two patients in each study arm The second audit examining for nine different potential deviations found complete adherence to study protocol on 254 of 292 of patient study days (900) There were no adverse events (pneumo-thorax arrhythmia or death) resulting from any of the study interventions

Physiologic Response to Ventilation ProtocolsRecordings of flow and airway pressure were available for analysis for all 50 patients prerandomization 49 patients immediately (median 1 interquartile range [IQR] 0ndash15 hr) postrandomization and for 26 patients during subsequent days (median 27 IQR 20ndash70-hr postrandomization) Approximately 15000 breaths in 125 recordings were analyzed Table 2 compares the respiratory parameters in both groups pre- and postrandomization There were no statistically sig-nificant differences between the PAV and PSV groups at base-line prerandomization (p values not shown) In the PAV group there was a significant decrease in peak airway pressure delta pressure (measured as peak airway opening pressure minus the PEEP) and area under the inspiratory pressure curve immediately postrandomization and an increase in the per-cent of total cycle time spent in inspiration (TiTtot) whereas there were no significant differences immediately postran-domization in the PSV group The second measurement of postrandomization respiratory parameters revealed signifi-cant decreases in peak airway pressure and delta pressure in both PAV and PSV groups Adjusting for baseline measure-ments as a covariate the PAV group received 7 cm H2

O∙s less ventilator assistance (PaoArea) per liter of minute ventilation (p lt 00001) and increased their VT by 115 mLcm H

2O∙s

relative to the PSV group (p lt 0001)PAV significantly reduced AI postrandomization (Fig 2A)

In total nine patients (18) had a high AI (gt 10 of breaths) on PSV or on AC mode at baseline No patients had a high AI on PAV In the PSV group six patients (26) had a high AI

at baseline four (17) immediately postrandomization and one patient (7) after further adjustment of ventilator settings according to the PSV protocol (Fig 2B) Adjusting for baseline differences patients in the PAV group tended to have a lower AI than patients in the PSV group immediately postrandom-ization but this did not reach statistical significance (adjusted difference ndash45 [95 CI ndash95 to 04] p = 007)

Clinical OutcomesTable 3 shows the results for the main clinical endpoints Weaning duration and time to successful SBT did not dif-fer significantly among survivors between the groups Among ICU survivors median time to ICU discharge was shorter in the group randomized to PAV 73 days (IQR 52ndash114 d) ver-sus 124 days (IQR 75ndash308 d) in the PSV group (p = 003) Kaplan-Meier plots of the proportion of at-risk patients achiev-ing successful extubation and live ICU discharge are shown in Figure 3 A and B No difference was found for hospital length of stay ICU mortality or hospital mortality Four patients (15) in the PAV group and three patients (13) in the PSV group died in ICU (respiratory rate [RR] of PAV 114 95 CI 028 to 46 p = 086) Overall in-hospital mortality was nine patients (33) in the PAV group and six patients (26) in the PSV group (RR of PAV 128 95 CI 054 to 31 p = 058)

Cointerventions Sedating MedicationsTable 4 compares the use of medications that may prolong weaning (by causing sedation or neuromuscular weakness) in the two groups pre- and postrandomization Type and dose of sedative medication were highly variable among individual patients and therefore we analyzed pre- and postrandomiza-tion changes to determine the impact of the ventilation pro-tocol on sedative use within each patient In the PAV group benzodiazepine narcotic and propofol use measured as both cumulative dose and average daily dose all decreased signifi-cantly postrandomization as did neuromuscular blocker use In the PSV group average daily dose of lorazepam equiva-lent and average daily dose of propofol decreased signifi-cantly postrandomization whereas the proportion of patients

Figure 2 A Asynchrony index pre- and postrandomization on proportional assist ventilation B Asynchrony index pre- and postrandomization on pressure support ventilationPAV群 PSV群

Asynchrony Index () = 同調イベントの回数全呼吸回数 times 100

IntensiveCareMed(2006)321515ndash1522

AI>10は3名いたがPAVに設定変更後0名となった

AI>10は6名いたがPSVに設定変更後4名となった

Clinical Outcomes

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

成功したSBTまでの数抜管成功までの数に変わりなしICU退室までの数はPAV群で有意に短い

Kaplan-Meier curves

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

抜管成功までの時間 ICU退室までの時間

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 9

TABLE 4 Comparison of Medications That May Alter Weaning Duration Pre- and Postrandomization

Variables

PAV Group (n = 27)

PSV Group (n = 23) Postrandomization

Pre randomization Pos trandomization Pre randomization Pos trandomizationAdjusted Difference PSV ndash PAV (95 CI)

Lorazepam equivalents

Cumulative dose mg

308 (207) 48 (257)a 137 (440) 54 (258) 237 (ndash580 to 1055)

Average daily dose mg

44 (255) 09 (43)a 23 (85) 06 (40)a 24 (ndash62 to 110)

Single-day dose mgb

08 (72) 12 (50) 16 (70) 10 (64) 32 (ndash13 to 78)

Morphine equivalents

Cumulative dose mg

5003 (6188) 1140 (3140)b 1710 (13310) 1610 (4860) 591 (ndash2515 to 3696)

Average daily dose mg

858 (1638) 170 (578)c 428 (1724) 136 (320)d 108 (ndash316 to 532)

Single-day dose mga

180 (1240) 187 (77) 100 (1920) 50 (630)d 334 (ndash026 to 670)d

Propofol

Cumulative dose mg

7400 (55500) 0 (700)a 29450 (85800) 0 (27660) 226 (ndash42156 to 42607)

Average daily dose mg

529 (11690) 0 (53)a 4879 (15980) 0 (2738)a 651 (ndash3211 to 4514)

Single-day dose mga

0 (0) 0 (0) 0 (1000) 0 (0) 344 (ndash3781 to 4469)

Neuroleptics

Days receiving neuroleptics n

0 (2) 0 (4) 0 (0) 1 (4)a ndash16 (ndash34 to 03)

Proportion of patients who received a neuroleptic

407 407 174 565b

NMBs

Days receiving NMBs n

1 (1) 0 (0)a 0 (1) 0 (0) ndash01 (ndash05 to 03)

Proportion of patients who received an NMB

518 148a 348 174

Corticosteroids

Days receiving corticosteroids n

0 (0) 0 (0) 0 (2) 0 (2) 02 (ndash06 to 10)

Proportion of patients who received corticosteroids

185 148 348 261

PAV = proportional assist ventilation PSV = pressure support ventilation NMB = neuromuscular blockera p lt 005b Single-day dose = total dose given day prior to randomization or total dose give day after randomizationc p lt 0005d p lt 00005e p le 007All data points except percentage of patients receiving neuroleptics NMBs or corticosteroids are presented as median (interquartile range)

鎮静剤鎮痛剤についてはPAV群とPSV群で変わりなし

その他bull 不整脈気胸プロトコル施中の死亡

は両群とも起こらなかった

筆者らの考察bull PSVはweaningに最も使われているモードだ

がそれと較しても遜ないweaningができることがわかった

bull アシストが少ない傾向であったがPAV群のが呼吸窮迫となる回数が少なかった

rarr同調性が良くなった結果の可能性

ICU退室までの期間はPAV群で短いという結果であったがpowerが少ない

筆者らの考察bull 単施設での結果であり他の施設にこの

新しいプロトコルそのものが受けれられるかは不明である

bull SBTに失敗した患者を主に対象としておりICU患者すべてに適応はできない

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる

Page 4: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

呼吸器離脱困難bull Difficult weaning prolonged weaningは

呼吸器装着患者の31に起こる

bull それらは患者のICU滞在数の延死亡率の増加コスト増につながる

bull PSVがSBTに失敗した患者がweaningをう際に最も使われているモードである

bull PSVはトリガーした後は受動的であり横隔膜の萎縮につながる可能性同調が多い可能性がある

EurRespirJ2007291033ndash1056

JAMA2002287345ndash355

Chest19941061188ndash1193

JApplPhysiol(1985)199681426ndash436

IntensiveCareMed2008341477ndash1486

呼吸器離脱困難患者に適したモードはあるのか

従来の呼吸器モード

医師 呼吸器 患者モード

1回換気量呼吸数

Closed-loop Ventilation

医師 呼吸器 患者 サポート

気道抵抗コンプライアンスを測定患者が求める1回換気量流速を計測する

Closed-loop Ventilationbull 患者の吸気努に例したサポートをうモード

Proportional Assist Ventilation(PAV)rarrPAV+(PB840980) PPS(Evita)

Neurally Adjusted Ventilatory Assist(NAVA)rarrServo i

bull 動weaningシステムにより離脱を指すモードSmartCare

rarrEvita XLinfinity V500Adaptive Support Ventilation(ASV)

rarrHammilton G5

患者の吸気努に例してサポートする

conventional ventilation The patient still has to generatesufficient flow to trigger the ventilator However as dis-cussed later conventional modes may force a larger tidalvolume than normal causing air trapping and thus increas-ing the number of missed triggers

Neurally Adjusted Ventilatory Assist

NAVA accomplishes the same goals as PAV but uti-lizes measurement of the diaphragmatic EMG signal tocontrol gas delivery6 This is accomplished by the place-ment of a specifically designed nasogastric tube that has aseries of EMG electrodes near its distal end positionedacross the diaphragm As illustrated in Figure 1 support-ing ventilation based on the diaphragmatic EMG signalshould greatly improve the response of the ventilator andsynchrony since the signal recognized is high up on theneural pathway controlling ventilation7 What is set by theclinician is the pressure applied for each millivolt of EMGactivity Thus similar to PAV a portion of the ventilatoryeffort is proportionally provided by the ventilator and theremainder by the patient As EMG activity increases pres-sure is applied during the inspiratory phase and as thediaphragm relaxes airway pressure decreases Inspirationends at a specific percentage of the peak EMG activityContrary to PAV NAVA greatly improves triggering since

gas delivery begins when the diaphragm is simulated notas a result of flow in the airway Thus even in the pres-ence of severe air trapping or large system leaks trigger-ing is not compromised

PAV and NAVA Versus Traditional Gas Delivery

Figure 2 nicely illustrates the relationship between andthe patientrsquos effort or work of breathing and the ventilatorpressure during various ventilation modes1 During a spe-cific breath with all forms of volume ventilation there isan indirect relationship between patient effort (Pmus) andthe pressure applied by the ventilator Essentially the morework assumed by the patient the less work applied by theventilator This relationship creates difficult problems re-garding synchrony which can easily be illustrated with alung model with proper setting of inspiratory demand Ifdemand is high and the lung model tidal volume is greaterthan the ventilator set tidal volume no pressure abovebaseline will be applied to the lung

In pressure-targeted ventilation regardless of mode ven-tilator work or applied pressure should be independent ofPmus That is during a specific breath a constant pressureshould be applied to the airway during inspiration nomatter how much work or Pmus is generated by the patientOf course in patients with very high ventilatory demandpressure may not be sustained since the ventilator may notalways be able to meet patient flow demand

In PAV and NAVA there is a direct relationship be-tween the patientrsquos effort (Pmus) and ventilator pressure

Fig 1 Steps in the process of activating a ventilator breath Neu-rally adjusted ventilatory assist ventilation modes are activatedand controlled by diaphragm excitation Proportional assist ven-tilation modes are activated by a change in airway pressure flowand volume (Adapted from Reference 6)

Fig 2 Relationship between patient effort and ventilator pressureduring various ventilation modes During volume ventilation pa-tient effort and ventilator pressure are inversely related Duringpressure ventilation ventilator pressure in unaffected by patienteffort During proportional assist ventilation (PAV) and neurally ad-justed ventilatory assist (NAVA) patient effort and ventilator pres-sure are directly related (Adapted from Reference 5)

PROPORTIONAL ASSIST VENTILATION AND NEURALLY ADJUSTED VENTILATORY ASSIST

142 RESPIRATORY CARE bull FEBRUARY 2011 VOL 56 NO 2

RespirCare201156(2)140ndash148

Proportional Assist Ventilation

bull コンプライアンス(C)とレジスタンス(R)を測定して患者の吸気努に例したサポートをう

bull Paw = VC + RtimesFlowbull WOB = ʃ PtimesFlow dt

bull PB840では8-15呼吸ごとに300msecの吸気ポーズをいCとRを測定

設定するのはサポート率のみ

患者の呼吸仕事量(WOBpt)が03〜08JLになるようにサポート率を設定する

PAVのメリットデメリット

bull メリット患者-呼吸器の同調性が良い呼吸筋を適度に使うことができる

bull デメリット患者の呼吸努が弱い場合はサポートが減ってしまうリークがある場合過度にサポート圧が上がってしまう可能性(runaway現象)

PAVのエビデンスbull PAVとPSVによるウィーニングをい48

時間観察したRCTではPAVのが調節呼吸に戻る割合が低かった(PAV 11 vsPSV 22 p=004)

bull 夜間PAVにすることで睡眠の質が改善したIntensiveCareMed2008342026ndash2034

Crit CareMed2007351048ndash1054

PAVは呼吸器離脱困難患者のウィーニングにおいて

優れたモードか

本の論

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Critical Care Medicine wwwccmjournalorg 1

Objectives Despite protocols incorporating spontaneous breath-ing trials 31 of ICU patients experience difficult or prolonged weaning from mechanical ventilation Nonfatiguing modes such as pressure support ventilation are recommended Proportional assist ventilation provides assistance in proportion to patient effort which may optimize weaning However it is not known how proportional assist ventilation performs relative to pressure sup-port ventilation over a prolonged period in the complex ICU set-ting The purpose of this study was to compare the physiologic and clinical performance (failure rate) safety and feasibility of protocols using daily spontaneous breathing trial plus pressure support ventilation versus proportional assist ventilation until ven-tilation discontinuationDesign Single-center unblinded pilot randomized controlled trialSetting Medical-surgical ICU of a tertiary-care hospitalPatients Adult patients intubated greater than 36 hours were ran-domized if they met eligibility criteria for partial ventilatory support

tolerated pressure support ventilation greater than or equal to 30 minutes and either failed or did not meet criteria for a sponta-neous breathing trialInterventions Patients were randomized to the pressure support ventilation or proportional assist ventilation protocol (PAV+ Puri-tan Bennett 840 Covidien Boulder CO) Both protocols used progressive decreases in level of assistance as tolerated coupled with daily assessment for spontaneous breathing trialsMeasurements and Main Results Of 54 patients randomized outcome data are available for 50 patients 27 were randomized to receive proportional assist ventilation and 23 to receive pres-sure support ventilation There were no adverse events linked to the study interventions and protocol violations were infrequent Recruitment was slower than projected (13 patients per month) The median (interquartile range) time from randomization to suc-cessful extubation was 39 days (28ndash84 d) on proportional assist ventilation versus 49 days (29ndash263 d) on pressure support ventilation (p = 039) Time to live ICU discharge was 73 days (52ndash114 d) on proportional assist ventilation versus 124 days (75ndash308 d) on pressure support ventilation (p = 003)Conclusion This pilot study demonstrates the utility safety and feasibility of the weaning protocols and provides important infor-mation to guide the design of a future randomized controlled trial comparing weaning from mechanical ventilation on pressure sup-port ventilation versus proportional assist ventilation (Crit Care Med 2016 XX00ndash00)Key Words assisted mechanical ventilation asynchrony index proportional assist ventilation respiratory muscles ventilator weaning

Difficult or prolonged weaning from mechanical ventilation (MV) affects 31 (range 26ndash42) of mechanically ventilated patients (1) and increases

patient morbidity mortality ICU length of stay and cost to the healthcare system (2) When initial weaning attempts fail to liberate patients from MV ideally clinicians should choose

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights ReservedDOI 101097CCM0000000000001600

1London Health Sciences Centre London ON Canada2Department of Medicine The University of Western Ontario London ON Canada

This study was performed at London Health Sciences Centre University Hospital London CanadaSupplemental digital content is available for this article Direct URL cita-tions appear in the printed text and are provided in the HTML and PDF versions of this article on the journalrsquos website (httpjournalslwwcomccmjournal)Supported in part by Critical Care Western Department of Medicine The University of Western Ontario Program of Experimental Medi-cine (POEM) Department of Medicine University of Western Ontario Lawson Health Research Institute London Health Sciences Centre London Canada and Covidien Boulder USADr Bosma has received honourarium for lectures and travel reimburse-ments from Covidien Dr Bosmarsquos institution has received grant support from Covidien for Dr Bosmarsquos research studies The remaining authors have disclosed that they do not have any potential conflicts of interestFor information regarding this article E-mail KarenJBosmalhsconca

A Pilot Randomized Trial Comparing Weaning From Mechanical Ventilation on Pressure Support Versus Proportional Assist Ventilation

Karen J Bosma MD FRCPC12 Brooke A Read MHS BSc RRT1 Mohammad J Bahrgard Nikoo MD FRCPC12 Philip M Jones MD FRCPC MSc12 Fran A Priestap MSc1 James F Lewis MD FRCPC12

Crit Care Med 2016 Jan 20[Epub ahead of print]

論のPICO

P 呼吸器離脱困難な患者(simple weaningではない)

I PAVを使

C PSVを使

O 呼吸器離脱までの時間

Study Designbull 単施設盲検前向きランダム化較

試験(pilot study)bull カナダロンドンにあるUniversity

Hospital-London Health Science Centre の内科外科ICU(20床)

bull ランダム割付はcomputer-generated random number sequence によってわれ隠蔽化されている

PatientsInclusion criteria

bull 18歳以上の挿管患者bull 36時間以上呼吸器管理をされている

PatientsExclusion criteria

1) 48時間以内にwithdrawが考慮される患者2) 位の脊髄障害進性の神経筋疾患の

よって慢性的な呼吸器依存となる可能性があるまたは脳外科的介が必要である患者

Enrolment Procedure

Supplemental Digital Content Figure 1 Enrolment Procedure

Daily Screening of all ICU patients gt18 yrs of age Intubated gt36 hours

Exclude if Exclusion Criteria

Successful SBT awaiting extubation Withdrawal of life support considered High spinal cord injury Neurosurgical patient Neuromuscular disease

Follow using daily eligibility checklist until meet all Clinical Stability Criteria to trial mode of partial ventilatory support (PAV or PSV)

partialcomplete reversal of cause of respiratory failure

body temperature between 36ordmC and 39ordmC metabolic disorders corrected pH gt732 hemoglobin gt70 gL no ongoing bleeding stable hemodynamic status Able to trigger PaO2 ge 60 mmHg or SaO2 ge 90 on FiO2le

060 and PEEP le 15 cm H2O ACVC Pplatle30cmH2O with Vt 6-8 mlkg

OR ACPC PC+PEEPle30cmH2O

Exclude if Exclusion Criteria Has met all Criteria for Switching to

Spontaneous Mode of Ventilation gt24 hours

Pre-inclusion Test on PSV 15 cmH2O for 30 minutes

Patient fails Pre-inclusion test Return to AC ventilation and reassess within 24 hour

Patient passes Pre-inclusion test Assess for Criteria to Initiate Weaning

PaO2FiO2gt200 on FiO2 le050 and PEEP le8 cmH2O No inotropesvasopressors except low dose dopamine

RSBI lt105 Patient passes Pre-SBT readiness assessment Proceed to SBT Place patient on CPAP 5 cmH2O and FiO2 040 for 30 minutes

Patient does not meet Criteria to Initiate Weaning

Patient meets Criteria to Initiate Weaning Proceed to Pre-SBT readiness assessment Place patient on CPAP for 1-2 min measure RSBI

RSBIge 105 Patient fails Pre-SBT readiness assessment

Patient passes SBT Patient fails SBT

Exclude if Exclusion Criteria

Passed SBT and tolerating PSVle7cmH2O and PEEPle5 cmH2O while awaiting extubation

Obtain informed consent from Patient or SDM as soon as available (deferred consent permitted)

Randomize

Exclude if Exclusion Criteria

Patient SDM refuse consent

bull 呼吸不全の原因に改善が認められてきている

bull 体温は36〜39度であるbull pH>732であるbull Hb>7gdLで現在出なしbull 動態が安定しているbull トリガーできるbull FiO2≦06 PEEP≦15cmH2Oで

PaO2≧60mmHg or SaO2≧90bull ACVCでVt 6-8mlkgで

Pplat≦30cmH2OもしくはACPCでPC+PEEP≦30cmH2O

上記をすべて満たせばプレテストとしてPSV 15cmH2Oに変更

Enrolment Procedure

Supplemental Digital Content Figure 1 Enrolment Procedure

Daily Screening of all ICU patients gt18 yrs of age Intubated gt36 hours

Exclude if Exclusion Criteria

Successful SBT awaiting extubation Withdrawal of life support considered High spinal cord injury Neurosurgical patient Neuromuscular disease

Follow using daily eligibility checklist until meet all Clinical Stability Criteria to trial mode of partial ventilatory support (PAV or PSV)

partialcomplete reversal of cause of respiratory failure

body temperature between 36ordmC and 39ordmC metabolic disorders corrected pH gt732 hemoglobin gt70 gL no ongoing bleeding stable hemodynamic status Able to trigger PaO2 ge 60 mmHg or SaO2 ge 90 on FiO2le

060 and PEEP le 15 cm H2O ACVC Pplatle30cmH2O with Vt 6-8 mlkg

OR ACPC PC+PEEPle30cmH2O

Exclude if Exclusion Criteria Has met all Criteria for Switching to

Spontaneous Mode of Ventilation gt24 hours

Pre-inclusion Test on PSV 15 cmH2O for 30 minutes

Patient fails Pre-inclusion test Return to AC ventilation and reassess within 24 hour

Patient passes Pre-inclusion test Assess for Criteria to Initiate Weaning

PaO2FiO2gt200 on FiO2 le050 and PEEP le8 cmH2O No inotropesvasopressors except low dose dopamine

RSBI lt105 Patient passes Pre-SBT readiness assessment Proceed to SBT Place patient on CPAP 5 cmH2O and FiO2 040 for 30 minutes

Patient does not meet Criteria to Initiate Weaning

Patient meets Criteria to Initiate Weaning Proceed to Pre-SBT readiness assessment Place patient on CPAP for 1-2 min measure RSBI

RSBIge 105 Patient fails Pre-SBT readiness assessment

Patient passes SBT Patient fails SBT

Exclude if Exclusion Criteria

Passed SBT and tolerating PSVle7cmH2O and PEEPle5 cmH2O while awaiting extubation

Obtain informed consent from Patient or SDM as soon as available (deferred consent permitted)

Randomize

Exclude if Exclusion Criteria

Patient SDM refuse consent

呼吸窮迫がなければプレテストクリアbull FiO2≦05 PEEP≦8cmH2OでPF≧200bull 強薬昇圧剤の使なしであればSBTへ

まず1-2分CPAP 5cmH2OにしてRSBIを測定

RSBI<105であればSBTへPSV 5cmH2Oにして30-120分

酸素化の条件を満たさないRSBI≧105SBT失敗であれば割り付けへ

Ventilation Protocols bull 呼吸器はPuritan Bennett 840 を使bull 後述するPAVプロトコールPSVプロトコー

ルに従って呼吸療法(RRT)が設定を変更(respiratory distress 呼吸窮迫がないようにできるだけサポートを下げていく)

bull 設定したmaxのサポートが必要になった場合はACに戻す

bull PAV群はPAV+とACPSV群はPSVとACしか使しない

Ventilation Protocols bull RCTの開始前に15名の患者をいて観察

研究をいRRTにプロトコルの確認をってもらった

bull プロトコルはすぐにられるようにラミネートしてベッドサイドに置いた

bull RRTは毎SBTをえるかどうか後述するチェックリストをいて確認除外項を満たさない場合は毎RSBIのアセスメントとSBTをう

Respiratory Distress bull 呼吸数>35回minbull 脈拍>140min もしくはベースラインより20以上の

上昇bull 収縮期圧>180mmHg もしくは<90mmHg ベース

ラインより30以上の変化bull 不安bull 発汗bull 副呼吸筋の使bull 奇異呼吸bull 呼吸困難の訴え

PAV protocolSupplemental Digital Content Figure 2a Proportional Assist Ventilation Protocol

bull 開始の設定はサポート率70

bull 2-3時間おきにRRTがアセスメントし問題なければ10-20ずつPAVサポートを下げる

サポート率90まで上げてもダメな場合はACに戻す

PSV protocolSupplemental Digital Content Figure 2b Pressure Support Ventilation Protocol

bull 開始の設定はPS 15cmH2O

bull まずRRlt35 VTgt5mLkgで耐えられるPSまで下げる

bull 2-3時間おきにRRTがアセスメントしてPS2-3cmH2Pずつ下げる

Data collection form for RRTsSupplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

SBTの開始基準bull PF≧200bull FiO2≦05bull pH≧732bull トリガーできるbull 24時間以内に筋虚イベントなしbull 強薬昇圧剤の使なし

Data collection form for RRTs

Supplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

CPAPにしてRSBIを測定RSBI<105であればSBT施

PAV群はサポート率20PSV群はPS 5cmH2OでSBTを施

PEEP≦8cmH2OでSaO2≧90呼吸窮迫なしであればSBTクリア

抜管へ

Endpoints 呼吸器離脱までの時間ICU病院滞在期間

呼吸器離脱成功の定義抜管後48時間再挿管なし侵襲的呼吸(NIV)の使もなし気管切開使時は呼吸器から最終的に外れた時点1のうち12時間以上NIVを使していない

Statistical Analysis bull Pilot studyのためサンプルサイズの計

算はしていないbull P値<005を統計学的有意としたbull Intention to treat approachをった

RESULTS

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial

エンロール期間は42か80名が基準を満たしが最終的に割付されたのは54名

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial最終的に解析がわれたのはPAV群27名PSV群23名

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

ベースラインは両群間で変わりなし重症度はAPACHEⅡ 27点程度患者は肺炎が3割もともとCOPDを持っている患者は2割程度

Feasibility Outcomes bull 呼吸窮迫でRRTが呼吸器サポートを上昇

させた回数はPAV群で76回232daysPSV群で124回295daysと有意にPSV群で多かった(p=0002)

bull ACに戻した回数はPAV群で66回PSV群で90回と変わりなし(p=061)

bull 全くACに戻さなかった患者はPAV群で627名PSV群で623名(p=10)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

ベースラインは両群間で変わりなしPAVにすることでΔP圧Peak圧が低く吸気時間呼吸時間がくなるPSV群と較するとPAV群のがPaoAreaPaoAreaVE(呼吸器のサポートの強さをす)が低くVTPaoAreaがくsecond post-randomizationの解析においてΔP圧Peak圧が低い

Asynchrony index

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 7

B635) documenting compliance with the daily screening for SBT noncompletion of either the CPAP screening test or SBT occurred on 5 of 292 of patient study days (17) and this violation involved two patients in each study arm The second audit examining for nine different potential deviations found complete adherence to study protocol on 254 of 292 of patient study days (900) There were no adverse events (pneumo-thorax arrhythmia or death) resulting from any of the study interventions

Physiologic Response to Ventilation ProtocolsRecordings of flow and airway pressure were available for analysis for all 50 patients prerandomization 49 patients immediately (median 1 interquartile range [IQR] 0ndash15 hr) postrandomization and for 26 patients during subsequent days (median 27 IQR 20ndash70-hr postrandomization) Approximately 15000 breaths in 125 recordings were analyzed Table 2 compares the respiratory parameters in both groups pre- and postrandomization There were no statistically sig-nificant differences between the PAV and PSV groups at base-line prerandomization (p values not shown) In the PAV group there was a significant decrease in peak airway pressure delta pressure (measured as peak airway opening pressure minus the PEEP) and area under the inspiratory pressure curve immediately postrandomization and an increase in the per-cent of total cycle time spent in inspiration (TiTtot) whereas there were no significant differences immediately postran-domization in the PSV group The second measurement of postrandomization respiratory parameters revealed signifi-cant decreases in peak airway pressure and delta pressure in both PAV and PSV groups Adjusting for baseline measure-ments as a covariate the PAV group received 7 cm H2

O∙s less ventilator assistance (PaoArea) per liter of minute ventilation (p lt 00001) and increased their VT by 115 mLcm H

2O∙s

relative to the PSV group (p lt 0001)PAV significantly reduced AI postrandomization (Fig 2A)

In total nine patients (18) had a high AI (gt 10 of breaths) on PSV or on AC mode at baseline No patients had a high AI on PAV In the PSV group six patients (26) had a high AI

at baseline four (17) immediately postrandomization and one patient (7) after further adjustment of ventilator settings according to the PSV protocol (Fig 2B) Adjusting for baseline differences patients in the PAV group tended to have a lower AI than patients in the PSV group immediately postrandom-ization but this did not reach statistical significance (adjusted difference ndash45 [95 CI ndash95 to 04] p = 007)

Clinical OutcomesTable 3 shows the results for the main clinical endpoints Weaning duration and time to successful SBT did not dif-fer significantly among survivors between the groups Among ICU survivors median time to ICU discharge was shorter in the group randomized to PAV 73 days (IQR 52ndash114 d) ver-sus 124 days (IQR 75ndash308 d) in the PSV group (p = 003) Kaplan-Meier plots of the proportion of at-risk patients achiev-ing successful extubation and live ICU discharge are shown in Figure 3 A and B No difference was found for hospital length of stay ICU mortality or hospital mortality Four patients (15) in the PAV group and three patients (13) in the PSV group died in ICU (respiratory rate [RR] of PAV 114 95 CI 028 to 46 p = 086) Overall in-hospital mortality was nine patients (33) in the PAV group and six patients (26) in the PSV group (RR of PAV 128 95 CI 054 to 31 p = 058)

Cointerventions Sedating MedicationsTable 4 compares the use of medications that may prolong weaning (by causing sedation or neuromuscular weakness) in the two groups pre- and postrandomization Type and dose of sedative medication were highly variable among individual patients and therefore we analyzed pre- and postrandomiza-tion changes to determine the impact of the ventilation pro-tocol on sedative use within each patient In the PAV group benzodiazepine narcotic and propofol use measured as both cumulative dose and average daily dose all decreased signifi-cantly postrandomization as did neuromuscular blocker use In the PSV group average daily dose of lorazepam equiva-lent and average daily dose of propofol decreased signifi-cantly postrandomization whereas the proportion of patients

Figure 2 A Asynchrony index pre- and postrandomization on proportional assist ventilation B Asynchrony index pre- and postrandomization on pressure support ventilationPAV群 PSV群

Asynchrony Index () = 同調イベントの回数全呼吸回数 times 100

IntensiveCareMed(2006)321515ndash1522

AI>10は3名いたがPAVに設定変更後0名となった

AI>10は6名いたがPSVに設定変更後4名となった

Clinical Outcomes

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

成功したSBTまでの数抜管成功までの数に変わりなしICU退室までの数はPAV群で有意に短い

Kaplan-Meier curves

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

抜管成功までの時間 ICU退室までの時間

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 9

TABLE 4 Comparison of Medications That May Alter Weaning Duration Pre- and Postrandomization

Variables

PAV Group (n = 27)

PSV Group (n = 23) Postrandomization

Pre randomization Pos trandomization Pre randomization Pos trandomizationAdjusted Difference PSV ndash PAV (95 CI)

Lorazepam equivalents

Cumulative dose mg

308 (207) 48 (257)a 137 (440) 54 (258) 237 (ndash580 to 1055)

Average daily dose mg

44 (255) 09 (43)a 23 (85) 06 (40)a 24 (ndash62 to 110)

Single-day dose mgb

08 (72) 12 (50) 16 (70) 10 (64) 32 (ndash13 to 78)

Morphine equivalents

Cumulative dose mg

5003 (6188) 1140 (3140)b 1710 (13310) 1610 (4860) 591 (ndash2515 to 3696)

Average daily dose mg

858 (1638) 170 (578)c 428 (1724) 136 (320)d 108 (ndash316 to 532)

Single-day dose mga

180 (1240) 187 (77) 100 (1920) 50 (630)d 334 (ndash026 to 670)d

Propofol

Cumulative dose mg

7400 (55500) 0 (700)a 29450 (85800) 0 (27660) 226 (ndash42156 to 42607)

Average daily dose mg

529 (11690) 0 (53)a 4879 (15980) 0 (2738)a 651 (ndash3211 to 4514)

Single-day dose mga

0 (0) 0 (0) 0 (1000) 0 (0) 344 (ndash3781 to 4469)

Neuroleptics

Days receiving neuroleptics n

0 (2) 0 (4) 0 (0) 1 (4)a ndash16 (ndash34 to 03)

Proportion of patients who received a neuroleptic

407 407 174 565b

NMBs

Days receiving NMBs n

1 (1) 0 (0)a 0 (1) 0 (0) ndash01 (ndash05 to 03)

Proportion of patients who received an NMB

518 148a 348 174

Corticosteroids

Days receiving corticosteroids n

0 (0) 0 (0) 0 (2) 0 (2) 02 (ndash06 to 10)

Proportion of patients who received corticosteroids

185 148 348 261

PAV = proportional assist ventilation PSV = pressure support ventilation NMB = neuromuscular blockera p lt 005b Single-day dose = total dose given day prior to randomization or total dose give day after randomizationc p lt 0005d p lt 00005e p le 007All data points except percentage of patients receiving neuroleptics NMBs or corticosteroids are presented as median (interquartile range)

鎮静剤鎮痛剤についてはPAV群とPSV群で変わりなし

その他bull 不整脈気胸プロトコル施中の死亡

は両群とも起こらなかった

筆者らの考察bull PSVはweaningに最も使われているモードだ

がそれと較しても遜ないweaningができることがわかった

bull アシストが少ない傾向であったがPAV群のが呼吸窮迫となる回数が少なかった

rarr同調性が良くなった結果の可能性

ICU退室までの期間はPAV群で短いという結果であったがpowerが少ない

筆者らの考察bull 単施設での結果であり他の施設にこの

新しいプロトコルそのものが受けれられるかは不明である

bull SBTに失敗した患者を主に対象としておりICU患者すべてに適応はできない

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる

Page 5: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

呼吸器離脱困難患者に適したモードはあるのか

従来の呼吸器モード

医師 呼吸器 患者モード

1回換気量呼吸数

Closed-loop Ventilation

医師 呼吸器 患者 サポート

気道抵抗コンプライアンスを測定患者が求める1回換気量流速を計測する

Closed-loop Ventilationbull 患者の吸気努に例したサポートをうモード

Proportional Assist Ventilation(PAV)rarrPAV+(PB840980) PPS(Evita)

Neurally Adjusted Ventilatory Assist(NAVA)rarrServo i

bull 動weaningシステムにより離脱を指すモードSmartCare

rarrEvita XLinfinity V500Adaptive Support Ventilation(ASV)

rarrHammilton G5

患者の吸気努に例してサポートする

conventional ventilation The patient still has to generatesufficient flow to trigger the ventilator However as dis-cussed later conventional modes may force a larger tidalvolume than normal causing air trapping and thus increas-ing the number of missed triggers

Neurally Adjusted Ventilatory Assist

NAVA accomplishes the same goals as PAV but uti-lizes measurement of the diaphragmatic EMG signal tocontrol gas delivery6 This is accomplished by the place-ment of a specifically designed nasogastric tube that has aseries of EMG electrodes near its distal end positionedacross the diaphragm As illustrated in Figure 1 support-ing ventilation based on the diaphragmatic EMG signalshould greatly improve the response of the ventilator andsynchrony since the signal recognized is high up on theneural pathway controlling ventilation7 What is set by theclinician is the pressure applied for each millivolt of EMGactivity Thus similar to PAV a portion of the ventilatoryeffort is proportionally provided by the ventilator and theremainder by the patient As EMG activity increases pres-sure is applied during the inspiratory phase and as thediaphragm relaxes airway pressure decreases Inspirationends at a specific percentage of the peak EMG activityContrary to PAV NAVA greatly improves triggering since

gas delivery begins when the diaphragm is simulated notas a result of flow in the airway Thus even in the pres-ence of severe air trapping or large system leaks trigger-ing is not compromised

PAV and NAVA Versus Traditional Gas Delivery

Figure 2 nicely illustrates the relationship between andthe patientrsquos effort or work of breathing and the ventilatorpressure during various ventilation modes1 During a spe-cific breath with all forms of volume ventilation there isan indirect relationship between patient effort (Pmus) andthe pressure applied by the ventilator Essentially the morework assumed by the patient the less work applied by theventilator This relationship creates difficult problems re-garding synchrony which can easily be illustrated with alung model with proper setting of inspiratory demand Ifdemand is high and the lung model tidal volume is greaterthan the ventilator set tidal volume no pressure abovebaseline will be applied to the lung

In pressure-targeted ventilation regardless of mode ven-tilator work or applied pressure should be independent ofPmus That is during a specific breath a constant pressureshould be applied to the airway during inspiration nomatter how much work or Pmus is generated by the patientOf course in patients with very high ventilatory demandpressure may not be sustained since the ventilator may notalways be able to meet patient flow demand

In PAV and NAVA there is a direct relationship be-tween the patientrsquos effort (Pmus) and ventilator pressure

Fig 1 Steps in the process of activating a ventilator breath Neu-rally adjusted ventilatory assist ventilation modes are activatedand controlled by diaphragm excitation Proportional assist ven-tilation modes are activated by a change in airway pressure flowand volume (Adapted from Reference 6)

Fig 2 Relationship between patient effort and ventilator pressureduring various ventilation modes During volume ventilation pa-tient effort and ventilator pressure are inversely related Duringpressure ventilation ventilator pressure in unaffected by patienteffort During proportional assist ventilation (PAV) and neurally ad-justed ventilatory assist (NAVA) patient effort and ventilator pres-sure are directly related (Adapted from Reference 5)

PROPORTIONAL ASSIST VENTILATION AND NEURALLY ADJUSTED VENTILATORY ASSIST

142 RESPIRATORY CARE bull FEBRUARY 2011 VOL 56 NO 2

RespirCare201156(2)140ndash148

Proportional Assist Ventilation

bull コンプライアンス(C)とレジスタンス(R)を測定して患者の吸気努に例したサポートをう

bull Paw = VC + RtimesFlowbull WOB = ʃ PtimesFlow dt

bull PB840では8-15呼吸ごとに300msecの吸気ポーズをいCとRを測定

設定するのはサポート率のみ

患者の呼吸仕事量(WOBpt)が03〜08JLになるようにサポート率を設定する

PAVのメリットデメリット

bull メリット患者-呼吸器の同調性が良い呼吸筋を適度に使うことができる

bull デメリット患者の呼吸努が弱い場合はサポートが減ってしまうリークがある場合過度にサポート圧が上がってしまう可能性(runaway現象)

PAVのエビデンスbull PAVとPSVによるウィーニングをい48

時間観察したRCTではPAVのが調節呼吸に戻る割合が低かった(PAV 11 vsPSV 22 p=004)

bull 夜間PAVにすることで睡眠の質が改善したIntensiveCareMed2008342026ndash2034

Crit CareMed2007351048ndash1054

PAVは呼吸器離脱困難患者のウィーニングにおいて

優れたモードか

本の論

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Critical Care Medicine wwwccmjournalorg 1

Objectives Despite protocols incorporating spontaneous breath-ing trials 31 of ICU patients experience difficult or prolonged weaning from mechanical ventilation Nonfatiguing modes such as pressure support ventilation are recommended Proportional assist ventilation provides assistance in proportion to patient effort which may optimize weaning However it is not known how proportional assist ventilation performs relative to pressure sup-port ventilation over a prolonged period in the complex ICU set-ting The purpose of this study was to compare the physiologic and clinical performance (failure rate) safety and feasibility of protocols using daily spontaneous breathing trial plus pressure support ventilation versus proportional assist ventilation until ven-tilation discontinuationDesign Single-center unblinded pilot randomized controlled trialSetting Medical-surgical ICU of a tertiary-care hospitalPatients Adult patients intubated greater than 36 hours were ran-domized if they met eligibility criteria for partial ventilatory support

tolerated pressure support ventilation greater than or equal to 30 minutes and either failed or did not meet criteria for a sponta-neous breathing trialInterventions Patients were randomized to the pressure support ventilation or proportional assist ventilation protocol (PAV+ Puri-tan Bennett 840 Covidien Boulder CO) Both protocols used progressive decreases in level of assistance as tolerated coupled with daily assessment for spontaneous breathing trialsMeasurements and Main Results Of 54 patients randomized outcome data are available for 50 patients 27 were randomized to receive proportional assist ventilation and 23 to receive pres-sure support ventilation There were no adverse events linked to the study interventions and protocol violations were infrequent Recruitment was slower than projected (13 patients per month) The median (interquartile range) time from randomization to suc-cessful extubation was 39 days (28ndash84 d) on proportional assist ventilation versus 49 days (29ndash263 d) on pressure support ventilation (p = 039) Time to live ICU discharge was 73 days (52ndash114 d) on proportional assist ventilation versus 124 days (75ndash308 d) on pressure support ventilation (p = 003)Conclusion This pilot study demonstrates the utility safety and feasibility of the weaning protocols and provides important infor-mation to guide the design of a future randomized controlled trial comparing weaning from mechanical ventilation on pressure sup-port ventilation versus proportional assist ventilation (Crit Care Med 2016 XX00ndash00)Key Words assisted mechanical ventilation asynchrony index proportional assist ventilation respiratory muscles ventilator weaning

Difficult or prolonged weaning from mechanical ventilation (MV) affects 31 (range 26ndash42) of mechanically ventilated patients (1) and increases

patient morbidity mortality ICU length of stay and cost to the healthcare system (2) When initial weaning attempts fail to liberate patients from MV ideally clinicians should choose

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights ReservedDOI 101097CCM0000000000001600

1London Health Sciences Centre London ON Canada2Department of Medicine The University of Western Ontario London ON Canada

This study was performed at London Health Sciences Centre University Hospital London CanadaSupplemental digital content is available for this article Direct URL cita-tions appear in the printed text and are provided in the HTML and PDF versions of this article on the journalrsquos website (httpjournalslwwcomccmjournal)Supported in part by Critical Care Western Department of Medicine The University of Western Ontario Program of Experimental Medi-cine (POEM) Department of Medicine University of Western Ontario Lawson Health Research Institute London Health Sciences Centre London Canada and Covidien Boulder USADr Bosma has received honourarium for lectures and travel reimburse-ments from Covidien Dr Bosmarsquos institution has received grant support from Covidien for Dr Bosmarsquos research studies The remaining authors have disclosed that they do not have any potential conflicts of interestFor information regarding this article E-mail KarenJBosmalhsconca

A Pilot Randomized Trial Comparing Weaning From Mechanical Ventilation on Pressure Support Versus Proportional Assist Ventilation

Karen J Bosma MD FRCPC12 Brooke A Read MHS BSc RRT1 Mohammad J Bahrgard Nikoo MD FRCPC12 Philip M Jones MD FRCPC MSc12 Fran A Priestap MSc1 James F Lewis MD FRCPC12

Crit Care Med 2016 Jan 20[Epub ahead of print]

論のPICO

P 呼吸器離脱困難な患者(simple weaningではない)

I PAVを使

C PSVを使

O 呼吸器離脱までの時間

Study Designbull 単施設盲検前向きランダム化較

試験(pilot study)bull カナダロンドンにあるUniversity

Hospital-London Health Science Centre の内科外科ICU(20床)

bull ランダム割付はcomputer-generated random number sequence によってわれ隠蔽化されている

PatientsInclusion criteria

bull 18歳以上の挿管患者bull 36時間以上呼吸器管理をされている

PatientsExclusion criteria

1) 48時間以内にwithdrawが考慮される患者2) 位の脊髄障害進性の神経筋疾患の

よって慢性的な呼吸器依存となる可能性があるまたは脳外科的介が必要である患者

Enrolment Procedure

Supplemental Digital Content Figure 1 Enrolment Procedure

Daily Screening of all ICU patients gt18 yrs of age Intubated gt36 hours

Exclude if Exclusion Criteria

Successful SBT awaiting extubation Withdrawal of life support considered High spinal cord injury Neurosurgical patient Neuromuscular disease

Follow using daily eligibility checklist until meet all Clinical Stability Criteria to trial mode of partial ventilatory support (PAV or PSV)

partialcomplete reversal of cause of respiratory failure

body temperature between 36ordmC and 39ordmC metabolic disorders corrected pH gt732 hemoglobin gt70 gL no ongoing bleeding stable hemodynamic status Able to trigger PaO2 ge 60 mmHg or SaO2 ge 90 on FiO2le

060 and PEEP le 15 cm H2O ACVC Pplatle30cmH2O with Vt 6-8 mlkg

OR ACPC PC+PEEPle30cmH2O

Exclude if Exclusion Criteria Has met all Criteria for Switching to

Spontaneous Mode of Ventilation gt24 hours

Pre-inclusion Test on PSV 15 cmH2O for 30 minutes

Patient fails Pre-inclusion test Return to AC ventilation and reassess within 24 hour

Patient passes Pre-inclusion test Assess for Criteria to Initiate Weaning

PaO2FiO2gt200 on FiO2 le050 and PEEP le8 cmH2O No inotropesvasopressors except low dose dopamine

RSBI lt105 Patient passes Pre-SBT readiness assessment Proceed to SBT Place patient on CPAP 5 cmH2O and FiO2 040 for 30 minutes

Patient does not meet Criteria to Initiate Weaning

Patient meets Criteria to Initiate Weaning Proceed to Pre-SBT readiness assessment Place patient on CPAP for 1-2 min measure RSBI

RSBIge 105 Patient fails Pre-SBT readiness assessment

Patient passes SBT Patient fails SBT

Exclude if Exclusion Criteria

Passed SBT and tolerating PSVle7cmH2O and PEEPle5 cmH2O while awaiting extubation

Obtain informed consent from Patient or SDM as soon as available (deferred consent permitted)

Randomize

Exclude if Exclusion Criteria

Patient SDM refuse consent

bull 呼吸不全の原因に改善が認められてきている

bull 体温は36〜39度であるbull pH>732であるbull Hb>7gdLで現在出なしbull 動態が安定しているbull トリガーできるbull FiO2≦06 PEEP≦15cmH2Oで

PaO2≧60mmHg or SaO2≧90bull ACVCでVt 6-8mlkgで

Pplat≦30cmH2OもしくはACPCでPC+PEEP≦30cmH2O

上記をすべて満たせばプレテストとしてPSV 15cmH2Oに変更

Enrolment Procedure

Supplemental Digital Content Figure 1 Enrolment Procedure

Daily Screening of all ICU patients gt18 yrs of age Intubated gt36 hours

Exclude if Exclusion Criteria

Successful SBT awaiting extubation Withdrawal of life support considered High spinal cord injury Neurosurgical patient Neuromuscular disease

Follow using daily eligibility checklist until meet all Clinical Stability Criteria to trial mode of partial ventilatory support (PAV or PSV)

partialcomplete reversal of cause of respiratory failure

body temperature between 36ordmC and 39ordmC metabolic disorders corrected pH gt732 hemoglobin gt70 gL no ongoing bleeding stable hemodynamic status Able to trigger PaO2 ge 60 mmHg or SaO2 ge 90 on FiO2le

060 and PEEP le 15 cm H2O ACVC Pplatle30cmH2O with Vt 6-8 mlkg

OR ACPC PC+PEEPle30cmH2O

Exclude if Exclusion Criteria Has met all Criteria for Switching to

Spontaneous Mode of Ventilation gt24 hours

Pre-inclusion Test on PSV 15 cmH2O for 30 minutes

Patient fails Pre-inclusion test Return to AC ventilation and reassess within 24 hour

Patient passes Pre-inclusion test Assess for Criteria to Initiate Weaning

PaO2FiO2gt200 on FiO2 le050 and PEEP le8 cmH2O No inotropesvasopressors except low dose dopamine

RSBI lt105 Patient passes Pre-SBT readiness assessment Proceed to SBT Place patient on CPAP 5 cmH2O and FiO2 040 for 30 minutes

Patient does not meet Criteria to Initiate Weaning

Patient meets Criteria to Initiate Weaning Proceed to Pre-SBT readiness assessment Place patient on CPAP for 1-2 min measure RSBI

RSBIge 105 Patient fails Pre-SBT readiness assessment

Patient passes SBT Patient fails SBT

Exclude if Exclusion Criteria

Passed SBT and tolerating PSVle7cmH2O and PEEPle5 cmH2O while awaiting extubation

Obtain informed consent from Patient or SDM as soon as available (deferred consent permitted)

Randomize

Exclude if Exclusion Criteria

Patient SDM refuse consent

呼吸窮迫がなければプレテストクリアbull FiO2≦05 PEEP≦8cmH2OでPF≧200bull 強薬昇圧剤の使なしであればSBTへ

まず1-2分CPAP 5cmH2OにしてRSBIを測定

RSBI<105であればSBTへPSV 5cmH2Oにして30-120分

酸素化の条件を満たさないRSBI≧105SBT失敗であれば割り付けへ

Ventilation Protocols bull 呼吸器はPuritan Bennett 840 を使bull 後述するPAVプロトコールPSVプロトコー

ルに従って呼吸療法(RRT)が設定を変更(respiratory distress 呼吸窮迫がないようにできるだけサポートを下げていく)

bull 設定したmaxのサポートが必要になった場合はACに戻す

bull PAV群はPAV+とACPSV群はPSVとACしか使しない

Ventilation Protocols bull RCTの開始前に15名の患者をいて観察

研究をいRRTにプロトコルの確認をってもらった

bull プロトコルはすぐにられるようにラミネートしてベッドサイドに置いた

bull RRTは毎SBTをえるかどうか後述するチェックリストをいて確認除外項を満たさない場合は毎RSBIのアセスメントとSBTをう

Respiratory Distress bull 呼吸数>35回minbull 脈拍>140min もしくはベースラインより20以上の

上昇bull 収縮期圧>180mmHg もしくは<90mmHg ベース

ラインより30以上の変化bull 不安bull 発汗bull 副呼吸筋の使bull 奇異呼吸bull 呼吸困難の訴え

PAV protocolSupplemental Digital Content Figure 2a Proportional Assist Ventilation Protocol

bull 開始の設定はサポート率70

bull 2-3時間おきにRRTがアセスメントし問題なければ10-20ずつPAVサポートを下げる

サポート率90まで上げてもダメな場合はACに戻す

PSV protocolSupplemental Digital Content Figure 2b Pressure Support Ventilation Protocol

bull 開始の設定はPS 15cmH2O

bull まずRRlt35 VTgt5mLkgで耐えられるPSまで下げる

bull 2-3時間おきにRRTがアセスメントしてPS2-3cmH2Pずつ下げる

Data collection form for RRTsSupplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

SBTの開始基準bull PF≧200bull FiO2≦05bull pH≧732bull トリガーできるbull 24時間以内に筋虚イベントなしbull 強薬昇圧剤の使なし

Data collection form for RRTs

Supplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

CPAPにしてRSBIを測定RSBI<105であればSBT施

PAV群はサポート率20PSV群はPS 5cmH2OでSBTを施

PEEP≦8cmH2OでSaO2≧90呼吸窮迫なしであればSBTクリア

抜管へ

Endpoints 呼吸器離脱までの時間ICU病院滞在期間

呼吸器離脱成功の定義抜管後48時間再挿管なし侵襲的呼吸(NIV)の使もなし気管切開使時は呼吸器から最終的に外れた時点1のうち12時間以上NIVを使していない

Statistical Analysis bull Pilot studyのためサンプルサイズの計

算はしていないbull P値<005を統計学的有意としたbull Intention to treat approachをった

RESULTS

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial

エンロール期間は42か80名が基準を満たしが最終的に割付されたのは54名

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial最終的に解析がわれたのはPAV群27名PSV群23名

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

ベースラインは両群間で変わりなし重症度はAPACHEⅡ 27点程度患者は肺炎が3割もともとCOPDを持っている患者は2割程度

Feasibility Outcomes bull 呼吸窮迫でRRTが呼吸器サポートを上昇

させた回数はPAV群で76回232daysPSV群で124回295daysと有意にPSV群で多かった(p=0002)

bull ACに戻した回数はPAV群で66回PSV群で90回と変わりなし(p=061)

bull 全くACに戻さなかった患者はPAV群で627名PSV群で623名(p=10)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

ベースラインは両群間で変わりなしPAVにすることでΔP圧Peak圧が低く吸気時間呼吸時間がくなるPSV群と較するとPAV群のがPaoAreaPaoAreaVE(呼吸器のサポートの強さをす)が低くVTPaoAreaがくsecond post-randomizationの解析においてΔP圧Peak圧が低い

Asynchrony index

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 7

B635) documenting compliance with the daily screening for SBT noncompletion of either the CPAP screening test or SBT occurred on 5 of 292 of patient study days (17) and this violation involved two patients in each study arm The second audit examining for nine different potential deviations found complete adherence to study protocol on 254 of 292 of patient study days (900) There were no adverse events (pneumo-thorax arrhythmia or death) resulting from any of the study interventions

Physiologic Response to Ventilation ProtocolsRecordings of flow and airway pressure were available for analysis for all 50 patients prerandomization 49 patients immediately (median 1 interquartile range [IQR] 0ndash15 hr) postrandomization and for 26 patients during subsequent days (median 27 IQR 20ndash70-hr postrandomization) Approximately 15000 breaths in 125 recordings were analyzed Table 2 compares the respiratory parameters in both groups pre- and postrandomization There were no statistically sig-nificant differences between the PAV and PSV groups at base-line prerandomization (p values not shown) In the PAV group there was a significant decrease in peak airway pressure delta pressure (measured as peak airway opening pressure minus the PEEP) and area under the inspiratory pressure curve immediately postrandomization and an increase in the per-cent of total cycle time spent in inspiration (TiTtot) whereas there were no significant differences immediately postran-domization in the PSV group The second measurement of postrandomization respiratory parameters revealed signifi-cant decreases in peak airway pressure and delta pressure in both PAV and PSV groups Adjusting for baseline measure-ments as a covariate the PAV group received 7 cm H2

O∙s less ventilator assistance (PaoArea) per liter of minute ventilation (p lt 00001) and increased their VT by 115 mLcm H

2O∙s

relative to the PSV group (p lt 0001)PAV significantly reduced AI postrandomization (Fig 2A)

In total nine patients (18) had a high AI (gt 10 of breaths) on PSV or on AC mode at baseline No patients had a high AI on PAV In the PSV group six patients (26) had a high AI

at baseline four (17) immediately postrandomization and one patient (7) after further adjustment of ventilator settings according to the PSV protocol (Fig 2B) Adjusting for baseline differences patients in the PAV group tended to have a lower AI than patients in the PSV group immediately postrandom-ization but this did not reach statistical significance (adjusted difference ndash45 [95 CI ndash95 to 04] p = 007)

Clinical OutcomesTable 3 shows the results for the main clinical endpoints Weaning duration and time to successful SBT did not dif-fer significantly among survivors between the groups Among ICU survivors median time to ICU discharge was shorter in the group randomized to PAV 73 days (IQR 52ndash114 d) ver-sus 124 days (IQR 75ndash308 d) in the PSV group (p = 003) Kaplan-Meier plots of the proportion of at-risk patients achiev-ing successful extubation and live ICU discharge are shown in Figure 3 A and B No difference was found for hospital length of stay ICU mortality or hospital mortality Four patients (15) in the PAV group and three patients (13) in the PSV group died in ICU (respiratory rate [RR] of PAV 114 95 CI 028 to 46 p = 086) Overall in-hospital mortality was nine patients (33) in the PAV group and six patients (26) in the PSV group (RR of PAV 128 95 CI 054 to 31 p = 058)

Cointerventions Sedating MedicationsTable 4 compares the use of medications that may prolong weaning (by causing sedation or neuromuscular weakness) in the two groups pre- and postrandomization Type and dose of sedative medication were highly variable among individual patients and therefore we analyzed pre- and postrandomiza-tion changes to determine the impact of the ventilation pro-tocol on sedative use within each patient In the PAV group benzodiazepine narcotic and propofol use measured as both cumulative dose and average daily dose all decreased signifi-cantly postrandomization as did neuromuscular blocker use In the PSV group average daily dose of lorazepam equiva-lent and average daily dose of propofol decreased signifi-cantly postrandomization whereas the proportion of patients

Figure 2 A Asynchrony index pre- and postrandomization on proportional assist ventilation B Asynchrony index pre- and postrandomization on pressure support ventilationPAV群 PSV群

Asynchrony Index () = 同調イベントの回数全呼吸回数 times 100

IntensiveCareMed(2006)321515ndash1522

AI>10は3名いたがPAVに設定変更後0名となった

AI>10は6名いたがPSVに設定変更後4名となった

Clinical Outcomes

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

成功したSBTまでの数抜管成功までの数に変わりなしICU退室までの数はPAV群で有意に短い

Kaplan-Meier curves

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

抜管成功までの時間 ICU退室までの時間

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 9

TABLE 4 Comparison of Medications That May Alter Weaning Duration Pre- and Postrandomization

Variables

PAV Group (n = 27)

PSV Group (n = 23) Postrandomization

Pre randomization Pos trandomization Pre randomization Pos trandomizationAdjusted Difference PSV ndash PAV (95 CI)

Lorazepam equivalents

Cumulative dose mg

308 (207) 48 (257)a 137 (440) 54 (258) 237 (ndash580 to 1055)

Average daily dose mg

44 (255) 09 (43)a 23 (85) 06 (40)a 24 (ndash62 to 110)

Single-day dose mgb

08 (72) 12 (50) 16 (70) 10 (64) 32 (ndash13 to 78)

Morphine equivalents

Cumulative dose mg

5003 (6188) 1140 (3140)b 1710 (13310) 1610 (4860) 591 (ndash2515 to 3696)

Average daily dose mg

858 (1638) 170 (578)c 428 (1724) 136 (320)d 108 (ndash316 to 532)

Single-day dose mga

180 (1240) 187 (77) 100 (1920) 50 (630)d 334 (ndash026 to 670)d

Propofol

Cumulative dose mg

7400 (55500) 0 (700)a 29450 (85800) 0 (27660) 226 (ndash42156 to 42607)

Average daily dose mg

529 (11690) 0 (53)a 4879 (15980) 0 (2738)a 651 (ndash3211 to 4514)

Single-day dose mga

0 (0) 0 (0) 0 (1000) 0 (0) 344 (ndash3781 to 4469)

Neuroleptics

Days receiving neuroleptics n

0 (2) 0 (4) 0 (0) 1 (4)a ndash16 (ndash34 to 03)

Proportion of patients who received a neuroleptic

407 407 174 565b

NMBs

Days receiving NMBs n

1 (1) 0 (0)a 0 (1) 0 (0) ndash01 (ndash05 to 03)

Proportion of patients who received an NMB

518 148a 348 174

Corticosteroids

Days receiving corticosteroids n

0 (0) 0 (0) 0 (2) 0 (2) 02 (ndash06 to 10)

Proportion of patients who received corticosteroids

185 148 348 261

PAV = proportional assist ventilation PSV = pressure support ventilation NMB = neuromuscular blockera p lt 005b Single-day dose = total dose given day prior to randomization or total dose give day after randomizationc p lt 0005d p lt 00005e p le 007All data points except percentage of patients receiving neuroleptics NMBs or corticosteroids are presented as median (interquartile range)

鎮静剤鎮痛剤についてはPAV群とPSV群で変わりなし

その他bull 不整脈気胸プロトコル施中の死亡

は両群とも起こらなかった

筆者らの考察bull PSVはweaningに最も使われているモードだ

がそれと較しても遜ないweaningができることがわかった

bull アシストが少ない傾向であったがPAV群のが呼吸窮迫となる回数が少なかった

rarr同調性が良くなった結果の可能性

ICU退室までの期間はPAV群で短いという結果であったがpowerが少ない

筆者らの考察bull 単施設での結果であり他の施設にこの

新しいプロトコルそのものが受けれられるかは不明である

bull SBTに失敗した患者を主に対象としておりICU患者すべてに適応はできない

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる

Page 6: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

従来の呼吸器モード

医師 呼吸器 患者モード

1回換気量呼吸数

Closed-loop Ventilation

医師 呼吸器 患者 サポート

気道抵抗コンプライアンスを測定患者が求める1回換気量流速を計測する

Closed-loop Ventilationbull 患者の吸気努に例したサポートをうモード

Proportional Assist Ventilation(PAV)rarrPAV+(PB840980) PPS(Evita)

Neurally Adjusted Ventilatory Assist(NAVA)rarrServo i

bull 動weaningシステムにより離脱を指すモードSmartCare

rarrEvita XLinfinity V500Adaptive Support Ventilation(ASV)

rarrHammilton G5

患者の吸気努に例してサポートする

conventional ventilation The patient still has to generatesufficient flow to trigger the ventilator However as dis-cussed later conventional modes may force a larger tidalvolume than normal causing air trapping and thus increas-ing the number of missed triggers

Neurally Adjusted Ventilatory Assist

NAVA accomplishes the same goals as PAV but uti-lizes measurement of the diaphragmatic EMG signal tocontrol gas delivery6 This is accomplished by the place-ment of a specifically designed nasogastric tube that has aseries of EMG electrodes near its distal end positionedacross the diaphragm As illustrated in Figure 1 support-ing ventilation based on the diaphragmatic EMG signalshould greatly improve the response of the ventilator andsynchrony since the signal recognized is high up on theneural pathway controlling ventilation7 What is set by theclinician is the pressure applied for each millivolt of EMGactivity Thus similar to PAV a portion of the ventilatoryeffort is proportionally provided by the ventilator and theremainder by the patient As EMG activity increases pres-sure is applied during the inspiratory phase and as thediaphragm relaxes airway pressure decreases Inspirationends at a specific percentage of the peak EMG activityContrary to PAV NAVA greatly improves triggering since

gas delivery begins when the diaphragm is simulated notas a result of flow in the airway Thus even in the pres-ence of severe air trapping or large system leaks trigger-ing is not compromised

PAV and NAVA Versus Traditional Gas Delivery

Figure 2 nicely illustrates the relationship between andthe patientrsquos effort or work of breathing and the ventilatorpressure during various ventilation modes1 During a spe-cific breath with all forms of volume ventilation there isan indirect relationship between patient effort (Pmus) andthe pressure applied by the ventilator Essentially the morework assumed by the patient the less work applied by theventilator This relationship creates difficult problems re-garding synchrony which can easily be illustrated with alung model with proper setting of inspiratory demand Ifdemand is high and the lung model tidal volume is greaterthan the ventilator set tidal volume no pressure abovebaseline will be applied to the lung

In pressure-targeted ventilation regardless of mode ven-tilator work or applied pressure should be independent ofPmus That is during a specific breath a constant pressureshould be applied to the airway during inspiration nomatter how much work or Pmus is generated by the patientOf course in patients with very high ventilatory demandpressure may not be sustained since the ventilator may notalways be able to meet patient flow demand

In PAV and NAVA there is a direct relationship be-tween the patientrsquos effort (Pmus) and ventilator pressure

Fig 1 Steps in the process of activating a ventilator breath Neu-rally adjusted ventilatory assist ventilation modes are activatedand controlled by diaphragm excitation Proportional assist ven-tilation modes are activated by a change in airway pressure flowand volume (Adapted from Reference 6)

Fig 2 Relationship between patient effort and ventilator pressureduring various ventilation modes During volume ventilation pa-tient effort and ventilator pressure are inversely related Duringpressure ventilation ventilator pressure in unaffected by patienteffort During proportional assist ventilation (PAV) and neurally ad-justed ventilatory assist (NAVA) patient effort and ventilator pres-sure are directly related (Adapted from Reference 5)

PROPORTIONAL ASSIST VENTILATION AND NEURALLY ADJUSTED VENTILATORY ASSIST

142 RESPIRATORY CARE bull FEBRUARY 2011 VOL 56 NO 2

RespirCare201156(2)140ndash148

Proportional Assist Ventilation

bull コンプライアンス(C)とレジスタンス(R)を測定して患者の吸気努に例したサポートをう

bull Paw = VC + RtimesFlowbull WOB = ʃ PtimesFlow dt

bull PB840では8-15呼吸ごとに300msecの吸気ポーズをいCとRを測定

設定するのはサポート率のみ

患者の呼吸仕事量(WOBpt)が03〜08JLになるようにサポート率を設定する

PAVのメリットデメリット

bull メリット患者-呼吸器の同調性が良い呼吸筋を適度に使うことができる

bull デメリット患者の呼吸努が弱い場合はサポートが減ってしまうリークがある場合過度にサポート圧が上がってしまう可能性(runaway現象)

PAVのエビデンスbull PAVとPSVによるウィーニングをい48

時間観察したRCTではPAVのが調節呼吸に戻る割合が低かった(PAV 11 vsPSV 22 p=004)

bull 夜間PAVにすることで睡眠の質が改善したIntensiveCareMed2008342026ndash2034

Crit CareMed2007351048ndash1054

PAVは呼吸器離脱困難患者のウィーニングにおいて

優れたモードか

本の論

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Critical Care Medicine wwwccmjournalorg 1

Objectives Despite protocols incorporating spontaneous breath-ing trials 31 of ICU patients experience difficult or prolonged weaning from mechanical ventilation Nonfatiguing modes such as pressure support ventilation are recommended Proportional assist ventilation provides assistance in proportion to patient effort which may optimize weaning However it is not known how proportional assist ventilation performs relative to pressure sup-port ventilation over a prolonged period in the complex ICU set-ting The purpose of this study was to compare the physiologic and clinical performance (failure rate) safety and feasibility of protocols using daily spontaneous breathing trial plus pressure support ventilation versus proportional assist ventilation until ven-tilation discontinuationDesign Single-center unblinded pilot randomized controlled trialSetting Medical-surgical ICU of a tertiary-care hospitalPatients Adult patients intubated greater than 36 hours were ran-domized if they met eligibility criteria for partial ventilatory support

tolerated pressure support ventilation greater than or equal to 30 minutes and either failed or did not meet criteria for a sponta-neous breathing trialInterventions Patients were randomized to the pressure support ventilation or proportional assist ventilation protocol (PAV+ Puri-tan Bennett 840 Covidien Boulder CO) Both protocols used progressive decreases in level of assistance as tolerated coupled with daily assessment for spontaneous breathing trialsMeasurements and Main Results Of 54 patients randomized outcome data are available for 50 patients 27 were randomized to receive proportional assist ventilation and 23 to receive pres-sure support ventilation There were no adverse events linked to the study interventions and protocol violations were infrequent Recruitment was slower than projected (13 patients per month) The median (interquartile range) time from randomization to suc-cessful extubation was 39 days (28ndash84 d) on proportional assist ventilation versus 49 days (29ndash263 d) on pressure support ventilation (p = 039) Time to live ICU discharge was 73 days (52ndash114 d) on proportional assist ventilation versus 124 days (75ndash308 d) on pressure support ventilation (p = 003)Conclusion This pilot study demonstrates the utility safety and feasibility of the weaning protocols and provides important infor-mation to guide the design of a future randomized controlled trial comparing weaning from mechanical ventilation on pressure sup-port ventilation versus proportional assist ventilation (Crit Care Med 2016 XX00ndash00)Key Words assisted mechanical ventilation asynchrony index proportional assist ventilation respiratory muscles ventilator weaning

Difficult or prolonged weaning from mechanical ventilation (MV) affects 31 (range 26ndash42) of mechanically ventilated patients (1) and increases

patient morbidity mortality ICU length of stay and cost to the healthcare system (2) When initial weaning attempts fail to liberate patients from MV ideally clinicians should choose

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights ReservedDOI 101097CCM0000000000001600

1London Health Sciences Centre London ON Canada2Department of Medicine The University of Western Ontario London ON Canada

This study was performed at London Health Sciences Centre University Hospital London CanadaSupplemental digital content is available for this article Direct URL cita-tions appear in the printed text and are provided in the HTML and PDF versions of this article on the journalrsquos website (httpjournalslwwcomccmjournal)Supported in part by Critical Care Western Department of Medicine The University of Western Ontario Program of Experimental Medi-cine (POEM) Department of Medicine University of Western Ontario Lawson Health Research Institute London Health Sciences Centre London Canada and Covidien Boulder USADr Bosma has received honourarium for lectures and travel reimburse-ments from Covidien Dr Bosmarsquos institution has received grant support from Covidien for Dr Bosmarsquos research studies The remaining authors have disclosed that they do not have any potential conflicts of interestFor information regarding this article E-mail KarenJBosmalhsconca

A Pilot Randomized Trial Comparing Weaning From Mechanical Ventilation on Pressure Support Versus Proportional Assist Ventilation

Karen J Bosma MD FRCPC12 Brooke A Read MHS BSc RRT1 Mohammad J Bahrgard Nikoo MD FRCPC12 Philip M Jones MD FRCPC MSc12 Fran A Priestap MSc1 James F Lewis MD FRCPC12

Crit Care Med 2016 Jan 20[Epub ahead of print]

論のPICO

P 呼吸器離脱困難な患者(simple weaningではない)

I PAVを使

C PSVを使

O 呼吸器離脱までの時間

Study Designbull 単施設盲検前向きランダム化較

試験(pilot study)bull カナダロンドンにあるUniversity

Hospital-London Health Science Centre の内科外科ICU(20床)

bull ランダム割付はcomputer-generated random number sequence によってわれ隠蔽化されている

PatientsInclusion criteria

bull 18歳以上の挿管患者bull 36時間以上呼吸器管理をされている

PatientsExclusion criteria

1) 48時間以内にwithdrawが考慮される患者2) 位の脊髄障害進性の神経筋疾患の

よって慢性的な呼吸器依存となる可能性があるまたは脳外科的介が必要である患者

Enrolment Procedure

Supplemental Digital Content Figure 1 Enrolment Procedure

Daily Screening of all ICU patients gt18 yrs of age Intubated gt36 hours

Exclude if Exclusion Criteria

Successful SBT awaiting extubation Withdrawal of life support considered High spinal cord injury Neurosurgical patient Neuromuscular disease

Follow using daily eligibility checklist until meet all Clinical Stability Criteria to trial mode of partial ventilatory support (PAV or PSV)

partialcomplete reversal of cause of respiratory failure

body temperature between 36ordmC and 39ordmC metabolic disorders corrected pH gt732 hemoglobin gt70 gL no ongoing bleeding stable hemodynamic status Able to trigger PaO2 ge 60 mmHg or SaO2 ge 90 on FiO2le

060 and PEEP le 15 cm H2O ACVC Pplatle30cmH2O with Vt 6-8 mlkg

OR ACPC PC+PEEPle30cmH2O

Exclude if Exclusion Criteria Has met all Criteria for Switching to

Spontaneous Mode of Ventilation gt24 hours

Pre-inclusion Test on PSV 15 cmH2O for 30 minutes

Patient fails Pre-inclusion test Return to AC ventilation and reassess within 24 hour

Patient passes Pre-inclusion test Assess for Criteria to Initiate Weaning

PaO2FiO2gt200 on FiO2 le050 and PEEP le8 cmH2O No inotropesvasopressors except low dose dopamine

RSBI lt105 Patient passes Pre-SBT readiness assessment Proceed to SBT Place patient on CPAP 5 cmH2O and FiO2 040 for 30 minutes

Patient does not meet Criteria to Initiate Weaning

Patient meets Criteria to Initiate Weaning Proceed to Pre-SBT readiness assessment Place patient on CPAP for 1-2 min measure RSBI

RSBIge 105 Patient fails Pre-SBT readiness assessment

Patient passes SBT Patient fails SBT

Exclude if Exclusion Criteria

Passed SBT and tolerating PSVle7cmH2O and PEEPle5 cmH2O while awaiting extubation

Obtain informed consent from Patient or SDM as soon as available (deferred consent permitted)

Randomize

Exclude if Exclusion Criteria

Patient SDM refuse consent

bull 呼吸不全の原因に改善が認められてきている

bull 体温は36〜39度であるbull pH>732であるbull Hb>7gdLで現在出なしbull 動態が安定しているbull トリガーできるbull FiO2≦06 PEEP≦15cmH2Oで

PaO2≧60mmHg or SaO2≧90bull ACVCでVt 6-8mlkgで

Pplat≦30cmH2OもしくはACPCでPC+PEEP≦30cmH2O

上記をすべて満たせばプレテストとしてPSV 15cmH2Oに変更

Enrolment Procedure

Supplemental Digital Content Figure 1 Enrolment Procedure

Daily Screening of all ICU patients gt18 yrs of age Intubated gt36 hours

Exclude if Exclusion Criteria

Successful SBT awaiting extubation Withdrawal of life support considered High spinal cord injury Neurosurgical patient Neuromuscular disease

Follow using daily eligibility checklist until meet all Clinical Stability Criteria to trial mode of partial ventilatory support (PAV or PSV)

partialcomplete reversal of cause of respiratory failure

body temperature between 36ordmC and 39ordmC metabolic disorders corrected pH gt732 hemoglobin gt70 gL no ongoing bleeding stable hemodynamic status Able to trigger PaO2 ge 60 mmHg or SaO2 ge 90 on FiO2le

060 and PEEP le 15 cm H2O ACVC Pplatle30cmH2O with Vt 6-8 mlkg

OR ACPC PC+PEEPle30cmH2O

Exclude if Exclusion Criteria Has met all Criteria for Switching to

Spontaneous Mode of Ventilation gt24 hours

Pre-inclusion Test on PSV 15 cmH2O for 30 minutes

Patient fails Pre-inclusion test Return to AC ventilation and reassess within 24 hour

Patient passes Pre-inclusion test Assess for Criteria to Initiate Weaning

PaO2FiO2gt200 on FiO2 le050 and PEEP le8 cmH2O No inotropesvasopressors except low dose dopamine

RSBI lt105 Patient passes Pre-SBT readiness assessment Proceed to SBT Place patient on CPAP 5 cmH2O and FiO2 040 for 30 minutes

Patient does not meet Criteria to Initiate Weaning

Patient meets Criteria to Initiate Weaning Proceed to Pre-SBT readiness assessment Place patient on CPAP for 1-2 min measure RSBI

RSBIge 105 Patient fails Pre-SBT readiness assessment

Patient passes SBT Patient fails SBT

Exclude if Exclusion Criteria

Passed SBT and tolerating PSVle7cmH2O and PEEPle5 cmH2O while awaiting extubation

Obtain informed consent from Patient or SDM as soon as available (deferred consent permitted)

Randomize

Exclude if Exclusion Criteria

Patient SDM refuse consent

呼吸窮迫がなければプレテストクリアbull FiO2≦05 PEEP≦8cmH2OでPF≧200bull 強薬昇圧剤の使なしであればSBTへ

まず1-2分CPAP 5cmH2OにしてRSBIを測定

RSBI<105であればSBTへPSV 5cmH2Oにして30-120分

酸素化の条件を満たさないRSBI≧105SBT失敗であれば割り付けへ

Ventilation Protocols bull 呼吸器はPuritan Bennett 840 を使bull 後述するPAVプロトコールPSVプロトコー

ルに従って呼吸療法(RRT)が設定を変更(respiratory distress 呼吸窮迫がないようにできるだけサポートを下げていく)

bull 設定したmaxのサポートが必要になった場合はACに戻す

bull PAV群はPAV+とACPSV群はPSVとACしか使しない

Ventilation Protocols bull RCTの開始前に15名の患者をいて観察

研究をいRRTにプロトコルの確認をってもらった

bull プロトコルはすぐにられるようにラミネートしてベッドサイドに置いた

bull RRTは毎SBTをえるかどうか後述するチェックリストをいて確認除外項を満たさない場合は毎RSBIのアセスメントとSBTをう

Respiratory Distress bull 呼吸数>35回minbull 脈拍>140min もしくはベースラインより20以上の

上昇bull 収縮期圧>180mmHg もしくは<90mmHg ベース

ラインより30以上の変化bull 不安bull 発汗bull 副呼吸筋の使bull 奇異呼吸bull 呼吸困難の訴え

PAV protocolSupplemental Digital Content Figure 2a Proportional Assist Ventilation Protocol

bull 開始の設定はサポート率70

bull 2-3時間おきにRRTがアセスメントし問題なければ10-20ずつPAVサポートを下げる

サポート率90まで上げてもダメな場合はACに戻す

PSV protocolSupplemental Digital Content Figure 2b Pressure Support Ventilation Protocol

bull 開始の設定はPS 15cmH2O

bull まずRRlt35 VTgt5mLkgで耐えられるPSまで下げる

bull 2-3時間おきにRRTがアセスメントしてPS2-3cmH2Pずつ下げる

Data collection form for RRTsSupplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

SBTの開始基準bull PF≧200bull FiO2≦05bull pH≧732bull トリガーできるbull 24時間以内に筋虚イベントなしbull 強薬昇圧剤の使なし

Data collection form for RRTs

Supplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

CPAPにしてRSBIを測定RSBI<105であればSBT施

PAV群はサポート率20PSV群はPS 5cmH2OでSBTを施

PEEP≦8cmH2OでSaO2≧90呼吸窮迫なしであればSBTクリア

抜管へ

Endpoints 呼吸器離脱までの時間ICU病院滞在期間

呼吸器離脱成功の定義抜管後48時間再挿管なし侵襲的呼吸(NIV)の使もなし気管切開使時は呼吸器から最終的に外れた時点1のうち12時間以上NIVを使していない

Statistical Analysis bull Pilot studyのためサンプルサイズの計

算はしていないbull P値<005を統計学的有意としたbull Intention to treat approachをった

RESULTS

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial

エンロール期間は42か80名が基準を満たしが最終的に割付されたのは54名

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial最終的に解析がわれたのはPAV群27名PSV群23名

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

ベースラインは両群間で変わりなし重症度はAPACHEⅡ 27点程度患者は肺炎が3割もともとCOPDを持っている患者は2割程度

Feasibility Outcomes bull 呼吸窮迫でRRTが呼吸器サポートを上昇

させた回数はPAV群で76回232daysPSV群で124回295daysと有意にPSV群で多かった(p=0002)

bull ACに戻した回数はPAV群で66回PSV群で90回と変わりなし(p=061)

bull 全くACに戻さなかった患者はPAV群で627名PSV群で623名(p=10)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

ベースラインは両群間で変わりなしPAVにすることでΔP圧Peak圧が低く吸気時間呼吸時間がくなるPSV群と較するとPAV群のがPaoAreaPaoAreaVE(呼吸器のサポートの強さをす)が低くVTPaoAreaがくsecond post-randomizationの解析においてΔP圧Peak圧が低い

Asynchrony index

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 7

B635) documenting compliance with the daily screening for SBT noncompletion of either the CPAP screening test or SBT occurred on 5 of 292 of patient study days (17) and this violation involved two patients in each study arm The second audit examining for nine different potential deviations found complete adherence to study protocol on 254 of 292 of patient study days (900) There were no adverse events (pneumo-thorax arrhythmia or death) resulting from any of the study interventions

Physiologic Response to Ventilation ProtocolsRecordings of flow and airway pressure were available for analysis for all 50 patients prerandomization 49 patients immediately (median 1 interquartile range [IQR] 0ndash15 hr) postrandomization and for 26 patients during subsequent days (median 27 IQR 20ndash70-hr postrandomization) Approximately 15000 breaths in 125 recordings were analyzed Table 2 compares the respiratory parameters in both groups pre- and postrandomization There were no statistically sig-nificant differences between the PAV and PSV groups at base-line prerandomization (p values not shown) In the PAV group there was a significant decrease in peak airway pressure delta pressure (measured as peak airway opening pressure minus the PEEP) and area under the inspiratory pressure curve immediately postrandomization and an increase in the per-cent of total cycle time spent in inspiration (TiTtot) whereas there were no significant differences immediately postran-domization in the PSV group The second measurement of postrandomization respiratory parameters revealed signifi-cant decreases in peak airway pressure and delta pressure in both PAV and PSV groups Adjusting for baseline measure-ments as a covariate the PAV group received 7 cm H2

O∙s less ventilator assistance (PaoArea) per liter of minute ventilation (p lt 00001) and increased their VT by 115 mLcm H

2O∙s

relative to the PSV group (p lt 0001)PAV significantly reduced AI postrandomization (Fig 2A)

In total nine patients (18) had a high AI (gt 10 of breaths) on PSV or on AC mode at baseline No patients had a high AI on PAV In the PSV group six patients (26) had a high AI

at baseline four (17) immediately postrandomization and one patient (7) after further adjustment of ventilator settings according to the PSV protocol (Fig 2B) Adjusting for baseline differences patients in the PAV group tended to have a lower AI than patients in the PSV group immediately postrandom-ization but this did not reach statistical significance (adjusted difference ndash45 [95 CI ndash95 to 04] p = 007)

Clinical OutcomesTable 3 shows the results for the main clinical endpoints Weaning duration and time to successful SBT did not dif-fer significantly among survivors between the groups Among ICU survivors median time to ICU discharge was shorter in the group randomized to PAV 73 days (IQR 52ndash114 d) ver-sus 124 days (IQR 75ndash308 d) in the PSV group (p = 003) Kaplan-Meier plots of the proportion of at-risk patients achiev-ing successful extubation and live ICU discharge are shown in Figure 3 A and B No difference was found for hospital length of stay ICU mortality or hospital mortality Four patients (15) in the PAV group and three patients (13) in the PSV group died in ICU (respiratory rate [RR] of PAV 114 95 CI 028 to 46 p = 086) Overall in-hospital mortality was nine patients (33) in the PAV group and six patients (26) in the PSV group (RR of PAV 128 95 CI 054 to 31 p = 058)

Cointerventions Sedating MedicationsTable 4 compares the use of medications that may prolong weaning (by causing sedation or neuromuscular weakness) in the two groups pre- and postrandomization Type and dose of sedative medication were highly variable among individual patients and therefore we analyzed pre- and postrandomiza-tion changes to determine the impact of the ventilation pro-tocol on sedative use within each patient In the PAV group benzodiazepine narcotic and propofol use measured as both cumulative dose and average daily dose all decreased signifi-cantly postrandomization as did neuromuscular blocker use In the PSV group average daily dose of lorazepam equiva-lent and average daily dose of propofol decreased signifi-cantly postrandomization whereas the proportion of patients

Figure 2 A Asynchrony index pre- and postrandomization on proportional assist ventilation B Asynchrony index pre- and postrandomization on pressure support ventilationPAV群 PSV群

Asynchrony Index () = 同調イベントの回数全呼吸回数 times 100

IntensiveCareMed(2006)321515ndash1522

AI>10は3名いたがPAVに設定変更後0名となった

AI>10は6名いたがPSVに設定変更後4名となった

Clinical Outcomes

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

成功したSBTまでの数抜管成功までの数に変わりなしICU退室までの数はPAV群で有意に短い

Kaplan-Meier curves

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

抜管成功までの時間 ICU退室までの時間

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 9

TABLE 4 Comparison of Medications That May Alter Weaning Duration Pre- and Postrandomization

Variables

PAV Group (n = 27)

PSV Group (n = 23) Postrandomization

Pre randomization Pos trandomization Pre randomization Pos trandomizationAdjusted Difference PSV ndash PAV (95 CI)

Lorazepam equivalents

Cumulative dose mg

308 (207) 48 (257)a 137 (440) 54 (258) 237 (ndash580 to 1055)

Average daily dose mg

44 (255) 09 (43)a 23 (85) 06 (40)a 24 (ndash62 to 110)

Single-day dose mgb

08 (72) 12 (50) 16 (70) 10 (64) 32 (ndash13 to 78)

Morphine equivalents

Cumulative dose mg

5003 (6188) 1140 (3140)b 1710 (13310) 1610 (4860) 591 (ndash2515 to 3696)

Average daily dose mg

858 (1638) 170 (578)c 428 (1724) 136 (320)d 108 (ndash316 to 532)

Single-day dose mga

180 (1240) 187 (77) 100 (1920) 50 (630)d 334 (ndash026 to 670)d

Propofol

Cumulative dose mg

7400 (55500) 0 (700)a 29450 (85800) 0 (27660) 226 (ndash42156 to 42607)

Average daily dose mg

529 (11690) 0 (53)a 4879 (15980) 0 (2738)a 651 (ndash3211 to 4514)

Single-day dose mga

0 (0) 0 (0) 0 (1000) 0 (0) 344 (ndash3781 to 4469)

Neuroleptics

Days receiving neuroleptics n

0 (2) 0 (4) 0 (0) 1 (4)a ndash16 (ndash34 to 03)

Proportion of patients who received a neuroleptic

407 407 174 565b

NMBs

Days receiving NMBs n

1 (1) 0 (0)a 0 (1) 0 (0) ndash01 (ndash05 to 03)

Proportion of patients who received an NMB

518 148a 348 174

Corticosteroids

Days receiving corticosteroids n

0 (0) 0 (0) 0 (2) 0 (2) 02 (ndash06 to 10)

Proportion of patients who received corticosteroids

185 148 348 261

PAV = proportional assist ventilation PSV = pressure support ventilation NMB = neuromuscular blockera p lt 005b Single-day dose = total dose given day prior to randomization or total dose give day after randomizationc p lt 0005d p lt 00005e p le 007All data points except percentage of patients receiving neuroleptics NMBs or corticosteroids are presented as median (interquartile range)

鎮静剤鎮痛剤についてはPAV群とPSV群で変わりなし

その他bull 不整脈気胸プロトコル施中の死亡

は両群とも起こらなかった

筆者らの考察bull PSVはweaningに最も使われているモードだ

がそれと較しても遜ないweaningができることがわかった

bull アシストが少ない傾向であったがPAV群のが呼吸窮迫となる回数が少なかった

rarr同調性が良くなった結果の可能性

ICU退室までの期間はPAV群で短いという結果であったがpowerが少ない

筆者らの考察bull 単施設での結果であり他の施設にこの

新しいプロトコルそのものが受けれられるかは不明である

bull SBTに失敗した患者を主に対象としておりICU患者すべてに適応はできない

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる

Page 7: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

Closed-loop Ventilation

医師 呼吸器 患者 サポート

気道抵抗コンプライアンスを測定患者が求める1回換気量流速を計測する

Closed-loop Ventilationbull 患者の吸気努に例したサポートをうモード

Proportional Assist Ventilation(PAV)rarrPAV+(PB840980) PPS(Evita)

Neurally Adjusted Ventilatory Assist(NAVA)rarrServo i

bull 動weaningシステムにより離脱を指すモードSmartCare

rarrEvita XLinfinity V500Adaptive Support Ventilation(ASV)

rarrHammilton G5

患者の吸気努に例してサポートする

conventional ventilation The patient still has to generatesufficient flow to trigger the ventilator However as dis-cussed later conventional modes may force a larger tidalvolume than normal causing air trapping and thus increas-ing the number of missed triggers

Neurally Adjusted Ventilatory Assist

NAVA accomplishes the same goals as PAV but uti-lizes measurement of the diaphragmatic EMG signal tocontrol gas delivery6 This is accomplished by the place-ment of a specifically designed nasogastric tube that has aseries of EMG electrodes near its distal end positionedacross the diaphragm As illustrated in Figure 1 support-ing ventilation based on the diaphragmatic EMG signalshould greatly improve the response of the ventilator andsynchrony since the signal recognized is high up on theneural pathway controlling ventilation7 What is set by theclinician is the pressure applied for each millivolt of EMGactivity Thus similar to PAV a portion of the ventilatoryeffort is proportionally provided by the ventilator and theremainder by the patient As EMG activity increases pres-sure is applied during the inspiratory phase and as thediaphragm relaxes airway pressure decreases Inspirationends at a specific percentage of the peak EMG activityContrary to PAV NAVA greatly improves triggering since

gas delivery begins when the diaphragm is simulated notas a result of flow in the airway Thus even in the pres-ence of severe air trapping or large system leaks trigger-ing is not compromised

PAV and NAVA Versus Traditional Gas Delivery

Figure 2 nicely illustrates the relationship between andthe patientrsquos effort or work of breathing and the ventilatorpressure during various ventilation modes1 During a spe-cific breath with all forms of volume ventilation there isan indirect relationship between patient effort (Pmus) andthe pressure applied by the ventilator Essentially the morework assumed by the patient the less work applied by theventilator This relationship creates difficult problems re-garding synchrony which can easily be illustrated with alung model with proper setting of inspiratory demand Ifdemand is high and the lung model tidal volume is greaterthan the ventilator set tidal volume no pressure abovebaseline will be applied to the lung

In pressure-targeted ventilation regardless of mode ven-tilator work or applied pressure should be independent ofPmus That is during a specific breath a constant pressureshould be applied to the airway during inspiration nomatter how much work or Pmus is generated by the patientOf course in patients with very high ventilatory demandpressure may not be sustained since the ventilator may notalways be able to meet patient flow demand

In PAV and NAVA there is a direct relationship be-tween the patientrsquos effort (Pmus) and ventilator pressure

Fig 1 Steps in the process of activating a ventilator breath Neu-rally adjusted ventilatory assist ventilation modes are activatedand controlled by diaphragm excitation Proportional assist ven-tilation modes are activated by a change in airway pressure flowand volume (Adapted from Reference 6)

Fig 2 Relationship between patient effort and ventilator pressureduring various ventilation modes During volume ventilation pa-tient effort and ventilator pressure are inversely related Duringpressure ventilation ventilator pressure in unaffected by patienteffort During proportional assist ventilation (PAV) and neurally ad-justed ventilatory assist (NAVA) patient effort and ventilator pres-sure are directly related (Adapted from Reference 5)

PROPORTIONAL ASSIST VENTILATION AND NEURALLY ADJUSTED VENTILATORY ASSIST

142 RESPIRATORY CARE bull FEBRUARY 2011 VOL 56 NO 2

RespirCare201156(2)140ndash148

Proportional Assist Ventilation

bull コンプライアンス(C)とレジスタンス(R)を測定して患者の吸気努に例したサポートをう

bull Paw = VC + RtimesFlowbull WOB = ʃ PtimesFlow dt

bull PB840では8-15呼吸ごとに300msecの吸気ポーズをいCとRを測定

設定するのはサポート率のみ

患者の呼吸仕事量(WOBpt)が03〜08JLになるようにサポート率を設定する

PAVのメリットデメリット

bull メリット患者-呼吸器の同調性が良い呼吸筋を適度に使うことができる

bull デメリット患者の呼吸努が弱い場合はサポートが減ってしまうリークがある場合過度にサポート圧が上がってしまう可能性(runaway現象)

PAVのエビデンスbull PAVとPSVによるウィーニングをい48

時間観察したRCTではPAVのが調節呼吸に戻る割合が低かった(PAV 11 vsPSV 22 p=004)

bull 夜間PAVにすることで睡眠の質が改善したIntensiveCareMed2008342026ndash2034

Crit CareMed2007351048ndash1054

PAVは呼吸器離脱困難患者のウィーニングにおいて

優れたモードか

本の論

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Critical Care Medicine wwwccmjournalorg 1

Objectives Despite protocols incorporating spontaneous breath-ing trials 31 of ICU patients experience difficult or prolonged weaning from mechanical ventilation Nonfatiguing modes such as pressure support ventilation are recommended Proportional assist ventilation provides assistance in proportion to patient effort which may optimize weaning However it is not known how proportional assist ventilation performs relative to pressure sup-port ventilation over a prolonged period in the complex ICU set-ting The purpose of this study was to compare the physiologic and clinical performance (failure rate) safety and feasibility of protocols using daily spontaneous breathing trial plus pressure support ventilation versus proportional assist ventilation until ven-tilation discontinuationDesign Single-center unblinded pilot randomized controlled trialSetting Medical-surgical ICU of a tertiary-care hospitalPatients Adult patients intubated greater than 36 hours were ran-domized if they met eligibility criteria for partial ventilatory support

tolerated pressure support ventilation greater than or equal to 30 minutes and either failed or did not meet criteria for a sponta-neous breathing trialInterventions Patients were randomized to the pressure support ventilation or proportional assist ventilation protocol (PAV+ Puri-tan Bennett 840 Covidien Boulder CO) Both protocols used progressive decreases in level of assistance as tolerated coupled with daily assessment for spontaneous breathing trialsMeasurements and Main Results Of 54 patients randomized outcome data are available for 50 patients 27 were randomized to receive proportional assist ventilation and 23 to receive pres-sure support ventilation There were no adverse events linked to the study interventions and protocol violations were infrequent Recruitment was slower than projected (13 patients per month) The median (interquartile range) time from randomization to suc-cessful extubation was 39 days (28ndash84 d) on proportional assist ventilation versus 49 days (29ndash263 d) on pressure support ventilation (p = 039) Time to live ICU discharge was 73 days (52ndash114 d) on proportional assist ventilation versus 124 days (75ndash308 d) on pressure support ventilation (p = 003)Conclusion This pilot study demonstrates the utility safety and feasibility of the weaning protocols and provides important infor-mation to guide the design of a future randomized controlled trial comparing weaning from mechanical ventilation on pressure sup-port ventilation versus proportional assist ventilation (Crit Care Med 2016 XX00ndash00)Key Words assisted mechanical ventilation asynchrony index proportional assist ventilation respiratory muscles ventilator weaning

Difficult or prolonged weaning from mechanical ventilation (MV) affects 31 (range 26ndash42) of mechanically ventilated patients (1) and increases

patient morbidity mortality ICU length of stay and cost to the healthcare system (2) When initial weaning attempts fail to liberate patients from MV ideally clinicians should choose

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights ReservedDOI 101097CCM0000000000001600

1London Health Sciences Centre London ON Canada2Department of Medicine The University of Western Ontario London ON Canada

This study was performed at London Health Sciences Centre University Hospital London CanadaSupplemental digital content is available for this article Direct URL cita-tions appear in the printed text and are provided in the HTML and PDF versions of this article on the journalrsquos website (httpjournalslwwcomccmjournal)Supported in part by Critical Care Western Department of Medicine The University of Western Ontario Program of Experimental Medi-cine (POEM) Department of Medicine University of Western Ontario Lawson Health Research Institute London Health Sciences Centre London Canada and Covidien Boulder USADr Bosma has received honourarium for lectures and travel reimburse-ments from Covidien Dr Bosmarsquos institution has received grant support from Covidien for Dr Bosmarsquos research studies The remaining authors have disclosed that they do not have any potential conflicts of interestFor information regarding this article E-mail KarenJBosmalhsconca

A Pilot Randomized Trial Comparing Weaning From Mechanical Ventilation on Pressure Support Versus Proportional Assist Ventilation

Karen J Bosma MD FRCPC12 Brooke A Read MHS BSc RRT1 Mohammad J Bahrgard Nikoo MD FRCPC12 Philip M Jones MD FRCPC MSc12 Fran A Priestap MSc1 James F Lewis MD FRCPC12

Crit Care Med 2016 Jan 20[Epub ahead of print]

論のPICO

P 呼吸器離脱困難な患者(simple weaningではない)

I PAVを使

C PSVを使

O 呼吸器離脱までの時間

Study Designbull 単施設盲検前向きランダム化較

試験(pilot study)bull カナダロンドンにあるUniversity

Hospital-London Health Science Centre の内科外科ICU(20床)

bull ランダム割付はcomputer-generated random number sequence によってわれ隠蔽化されている

PatientsInclusion criteria

bull 18歳以上の挿管患者bull 36時間以上呼吸器管理をされている

PatientsExclusion criteria

1) 48時間以内にwithdrawが考慮される患者2) 位の脊髄障害進性の神経筋疾患の

よって慢性的な呼吸器依存となる可能性があるまたは脳外科的介が必要である患者

Enrolment Procedure

Supplemental Digital Content Figure 1 Enrolment Procedure

Daily Screening of all ICU patients gt18 yrs of age Intubated gt36 hours

Exclude if Exclusion Criteria

Successful SBT awaiting extubation Withdrawal of life support considered High spinal cord injury Neurosurgical patient Neuromuscular disease

Follow using daily eligibility checklist until meet all Clinical Stability Criteria to trial mode of partial ventilatory support (PAV or PSV)

partialcomplete reversal of cause of respiratory failure

body temperature between 36ordmC and 39ordmC metabolic disorders corrected pH gt732 hemoglobin gt70 gL no ongoing bleeding stable hemodynamic status Able to trigger PaO2 ge 60 mmHg or SaO2 ge 90 on FiO2le

060 and PEEP le 15 cm H2O ACVC Pplatle30cmH2O with Vt 6-8 mlkg

OR ACPC PC+PEEPle30cmH2O

Exclude if Exclusion Criteria Has met all Criteria for Switching to

Spontaneous Mode of Ventilation gt24 hours

Pre-inclusion Test on PSV 15 cmH2O for 30 minutes

Patient fails Pre-inclusion test Return to AC ventilation and reassess within 24 hour

Patient passes Pre-inclusion test Assess for Criteria to Initiate Weaning

PaO2FiO2gt200 on FiO2 le050 and PEEP le8 cmH2O No inotropesvasopressors except low dose dopamine

RSBI lt105 Patient passes Pre-SBT readiness assessment Proceed to SBT Place patient on CPAP 5 cmH2O and FiO2 040 for 30 minutes

Patient does not meet Criteria to Initiate Weaning

Patient meets Criteria to Initiate Weaning Proceed to Pre-SBT readiness assessment Place patient on CPAP for 1-2 min measure RSBI

RSBIge 105 Patient fails Pre-SBT readiness assessment

Patient passes SBT Patient fails SBT

Exclude if Exclusion Criteria

Passed SBT and tolerating PSVle7cmH2O and PEEPle5 cmH2O while awaiting extubation

Obtain informed consent from Patient or SDM as soon as available (deferred consent permitted)

Randomize

Exclude if Exclusion Criteria

Patient SDM refuse consent

bull 呼吸不全の原因に改善が認められてきている

bull 体温は36〜39度であるbull pH>732であるbull Hb>7gdLで現在出なしbull 動態が安定しているbull トリガーできるbull FiO2≦06 PEEP≦15cmH2Oで

PaO2≧60mmHg or SaO2≧90bull ACVCでVt 6-8mlkgで

Pplat≦30cmH2OもしくはACPCでPC+PEEP≦30cmH2O

上記をすべて満たせばプレテストとしてPSV 15cmH2Oに変更

Enrolment Procedure

Supplemental Digital Content Figure 1 Enrolment Procedure

Daily Screening of all ICU patients gt18 yrs of age Intubated gt36 hours

Exclude if Exclusion Criteria

Successful SBT awaiting extubation Withdrawal of life support considered High spinal cord injury Neurosurgical patient Neuromuscular disease

Follow using daily eligibility checklist until meet all Clinical Stability Criteria to trial mode of partial ventilatory support (PAV or PSV)

partialcomplete reversal of cause of respiratory failure

body temperature between 36ordmC and 39ordmC metabolic disorders corrected pH gt732 hemoglobin gt70 gL no ongoing bleeding stable hemodynamic status Able to trigger PaO2 ge 60 mmHg or SaO2 ge 90 on FiO2le

060 and PEEP le 15 cm H2O ACVC Pplatle30cmH2O with Vt 6-8 mlkg

OR ACPC PC+PEEPle30cmH2O

Exclude if Exclusion Criteria Has met all Criteria for Switching to

Spontaneous Mode of Ventilation gt24 hours

Pre-inclusion Test on PSV 15 cmH2O for 30 minutes

Patient fails Pre-inclusion test Return to AC ventilation and reassess within 24 hour

Patient passes Pre-inclusion test Assess for Criteria to Initiate Weaning

PaO2FiO2gt200 on FiO2 le050 and PEEP le8 cmH2O No inotropesvasopressors except low dose dopamine

RSBI lt105 Patient passes Pre-SBT readiness assessment Proceed to SBT Place patient on CPAP 5 cmH2O and FiO2 040 for 30 minutes

Patient does not meet Criteria to Initiate Weaning

Patient meets Criteria to Initiate Weaning Proceed to Pre-SBT readiness assessment Place patient on CPAP for 1-2 min measure RSBI

RSBIge 105 Patient fails Pre-SBT readiness assessment

Patient passes SBT Patient fails SBT

Exclude if Exclusion Criteria

Passed SBT and tolerating PSVle7cmH2O and PEEPle5 cmH2O while awaiting extubation

Obtain informed consent from Patient or SDM as soon as available (deferred consent permitted)

Randomize

Exclude if Exclusion Criteria

Patient SDM refuse consent

呼吸窮迫がなければプレテストクリアbull FiO2≦05 PEEP≦8cmH2OでPF≧200bull 強薬昇圧剤の使なしであればSBTへ

まず1-2分CPAP 5cmH2OにしてRSBIを測定

RSBI<105であればSBTへPSV 5cmH2Oにして30-120分

酸素化の条件を満たさないRSBI≧105SBT失敗であれば割り付けへ

Ventilation Protocols bull 呼吸器はPuritan Bennett 840 を使bull 後述するPAVプロトコールPSVプロトコー

ルに従って呼吸療法(RRT)が設定を変更(respiratory distress 呼吸窮迫がないようにできるだけサポートを下げていく)

bull 設定したmaxのサポートが必要になった場合はACに戻す

bull PAV群はPAV+とACPSV群はPSVとACしか使しない

Ventilation Protocols bull RCTの開始前に15名の患者をいて観察

研究をいRRTにプロトコルの確認をってもらった

bull プロトコルはすぐにられるようにラミネートしてベッドサイドに置いた

bull RRTは毎SBTをえるかどうか後述するチェックリストをいて確認除外項を満たさない場合は毎RSBIのアセスメントとSBTをう

Respiratory Distress bull 呼吸数>35回minbull 脈拍>140min もしくはベースラインより20以上の

上昇bull 収縮期圧>180mmHg もしくは<90mmHg ベース

ラインより30以上の変化bull 不安bull 発汗bull 副呼吸筋の使bull 奇異呼吸bull 呼吸困難の訴え

PAV protocolSupplemental Digital Content Figure 2a Proportional Assist Ventilation Protocol

bull 開始の設定はサポート率70

bull 2-3時間おきにRRTがアセスメントし問題なければ10-20ずつPAVサポートを下げる

サポート率90まで上げてもダメな場合はACに戻す

PSV protocolSupplemental Digital Content Figure 2b Pressure Support Ventilation Protocol

bull 開始の設定はPS 15cmH2O

bull まずRRlt35 VTgt5mLkgで耐えられるPSまで下げる

bull 2-3時間おきにRRTがアセスメントしてPS2-3cmH2Pずつ下げる

Data collection form for RRTsSupplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

SBTの開始基準bull PF≧200bull FiO2≦05bull pH≧732bull トリガーできるbull 24時間以内に筋虚イベントなしbull 強薬昇圧剤の使なし

Data collection form for RRTs

Supplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

CPAPにしてRSBIを測定RSBI<105であればSBT施

PAV群はサポート率20PSV群はPS 5cmH2OでSBTを施

PEEP≦8cmH2OでSaO2≧90呼吸窮迫なしであればSBTクリア

抜管へ

Endpoints 呼吸器離脱までの時間ICU病院滞在期間

呼吸器離脱成功の定義抜管後48時間再挿管なし侵襲的呼吸(NIV)の使もなし気管切開使時は呼吸器から最終的に外れた時点1のうち12時間以上NIVを使していない

Statistical Analysis bull Pilot studyのためサンプルサイズの計

算はしていないbull P値<005を統計学的有意としたbull Intention to treat approachをった

RESULTS

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial

エンロール期間は42か80名が基準を満たしが最終的に割付されたのは54名

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial最終的に解析がわれたのはPAV群27名PSV群23名

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

ベースラインは両群間で変わりなし重症度はAPACHEⅡ 27点程度患者は肺炎が3割もともとCOPDを持っている患者は2割程度

Feasibility Outcomes bull 呼吸窮迫でRRTが呼吸器サポートを上昇

させた回数はPAV群で76回232daysPSV群で124回295daysと有意にPSV群で多かった(p=0002)

bull ACに戻した回数はPAV群で66回PSV群で90回と変わりなし(p=061)

bull 全くACに戻さなかった患者はPAV群で627名PSV群で623名(p=10)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

ベースラインは両群間で変わりなしPAVにすることでΔP圧Peak圧が低く吸気時間呼吸時間がくなるPSV群と較するとPAV群のがPaoAreaPaoAreaVE(呼吸器のサポートの強さをす)が低くVTPaoAreaがくsecond post-randomizationの解析においてΔP圧Peak圧が低い

Asynchrony index

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 7

B635) documenting compliance with the daily screening for SBT noncompletion of either the CPAP screening test or SBT occurred on 5 of 292 of patient study days (17) and this violation involved two patients in each study arm The second audit examining for nine different potential deviations found complete adherence to study protocol on 254 of 292 of patient study days (900) There were no adverse events (pneumo-thorax arrhythmia or death) resulting from any of the study interventions

Physiologic Response to Ventilation ProtocolsRecordings of flow and airway pressure were available for analysis for all 50 patients prerandomization 49 patients immediately (median 1 interquartile range [IQR] 0ndash15 hr) postrandomization and for 26 patients during subsequent days (median 27 IQR 20ndash70-hr postrandomization) Approximately 15000 breaths in 125 recordings were analyzed Table 2 compares the respiratory parameters in both groups pre- and postrandomization There were no statistically sig-nificant differences between the PAV and PSV groups at base-line prerandomization (p values not shown) In the PAV group there was a significant decrease in peak airway pressure delta pressure (measured as peak airway opening pressure minus the PEEP) and area under the inspiratory pressure curve immediately postrandomization and an increase in the per-cent of total cycle time spent in inspiration (TiTtot) whereas there were no significant differences immediately postran-domization in the PSV group The second measurement of postrandomization respiratory parameters revealed signifi-cant decreases in peak airway pressure and delta pressure in both PAV and PSV groups Adjusting for baseline measure-ments as a covariate the PAV group received 7 cm H2

O∙s less ventilator assistance (PaoArea) per liter of minute ventilation (p lt 00001) and increased their VT by 115 mLcm H

2O∙s

relative to the PSV group (p lt 0001)PAV significantly reduced AI postrandomization (Fig 2A)

In total nine patients (18) had a high AI (gt 10 of breaths) on PSV or on AC mode at baseline No patients had a high AI on PAV In the PSV group six patients (26) had a high AI

at baseline four (17) immediately postrandomization and one patient (7) after further adjustment of ventilator settings according to the PSV protocol (Fig 2B) Adjusting for baseline differences patients in the PAV group tended to have a lower AI than patients in the PSV group immediately postrandom-ization but this did not reach statistical significance (adjusted difference ndash45 [95 CI ndash95 to 04] p = 007)

Clinical OutcomesTable 3 shows the results for the main clinical endpoints Weaning duration and time to successful SBT did not dif-fer significantly among survivors between the groups Among ICU survivors median time to ICU discharge was shorter in the group randomized to PAV 73 days (IQR 52ndash114 d) ver-sus 124 days (IQR 75ndash308 d) in the PSV group (p = 003) Kaplan-Meier plots of the proportion of at-risk patients achiev-ing successful extubation and live ICU discharge are shown in Figure 3 A and B No difference was found for hospital length of stay ICU mortality or hospital mortality Four patients (15) in the PAV group and three patients (13) in the PSV group died in ICU (respiratory rate [RR] of PAV 114 95 CI 028 to 46 p = 086) Overall in-hospital mortality was nine patients (33) in the PAV group and six patients (26) in the PSV group (RR of PAV 128 95 CI 054 to 31 p = 058)

Cointerventions Sedating MedicationsTable 4 compares the use of medications that may prolong weaning (by causing sedation or neuromuscular weakness) in the two groups pre- and postrandomization Type and dose of sedative medication were highly variable among individual patients and therefore we analyzed pre- and postrandomiza-tion changes to determine the impact of the ventilation pro-tocol on sedative use within each patient In the PAV group benzodiazepine narcotic and propofol use measured as both cumulative dose and average daily dose all decreased signifi-cantly postrandomization as did neuromuscular blocker use In the PSV group average daily dose of lorazepam equiva-lent and average daily dose of propofol decreased signifi-cantly postrandomization whereas the proportion of patients

Figure 2 A Asynchrony index pre- and postrandomization on proportional assist ventilation B Asynchrony index pre- and postrandomization on pressure support ventilationPAV群 PSV群

Asynchrony Index () = 同調イベントの回数全呼吸回数 times 100

IntensiveCareMed(2006)321515ndash1522

AI>10は3名いたがPAVに設定変更後0名となった

AI>10は6名いたがPSVに設定変更後4名となった

Clinical Outcomes

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

成功したSBTまでの数抜管成功までの数に変わりなしICU退室までの数はPAV群で有意に短い

Kaplan-Meier curves

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

抜管成功までの時間 ICU退室までの時間

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 9

TABLE 4 Comparison of Medications That May Alter Weaning Duration Pre- and Postrandomization

Variables

PAV Group (n = 27)

PSV Group (n = 23) Postrandomization

Pre randomization Pos trandomization Pre randomization Pos trandomizationAdjusted Difference PSV ndash PAV (95 CI)

Lorazepam equivalents

Cumulative dose mg

308 (207) 48 (257)a 137 (440) 54 (258) 237 (ndash580 to 1055)

Average daily dose mg

44 (255) 09 (43)a 23 (85) 06 (40)a 24 (ndash62 to 110)

Single-day dose mgb

08 (72) 12 (50) 16 (70) 10 (64) 32 (ndash13 to 78)

Morphine equivalents

Cumulative dose mg

5003 (6188) 1140 (3140)b 1710 (13310) 1610 (4860) 591 (ndash2515 to 3696)

Average daily dose mg

858 (1638) 170 (578)c 428 (1724) 136 (320)d 108 (ndash316 to 532)

Single-day dose mga

180 (1240) 187 (77) 100 (1920) 50 (630)d 334 (ndash026 to 670)d

Propofol

Cumulative dose mg

7400 (55500) 0 (700)a 29450 (85800) 0 (27660) 226 (ndash42156 to 42607)

Average daily dose mg

529 (11690) 0 (53)a 4879 (15980) 0 (2738)a 651 (ndash3211 to 4514)

Single-day dose mga

0 (0) 0 (0) 0 (1000) 0 (0) 344 (ndash3781 to 4469)

Neuroleptics

Days receiving neuroleptics n

0 (2) 0 (4) 0 (0) 1 (4)a ndash16 (ndash34 to 03)

Proportion of patients who received a neuroleptic

407 407 174 565b

NMBs

Days receiving NMBs n

1 (1) 0 (0)a 0 (1) 0 (0) ndash01 (ndash05 to 03)

Proportion of patients who received an NMB

518 148a 348 174

Corticosteroids

Days receiving corticosteroids n

0 (0) 0 (0) 0 (2) 0 (2) 02 (ndash06 to 10)

Proportion of patients who received corticosteroids

185 148 348 261

PAV = proportional assist ventilation PSV = pressure support ventilation NMB = neuromuscular blockera p lt 005b Single-day dose = total dose given day prior to randomization or total dose give day after randomizationc p lt 0005d p lt 00005e p le 007All data points except percentage of patients receiving neuroleptics NMBs or corticosteroids are presented as median (interquartile range)

鎮静剤鎮痛剤についてはPAV群とPSV群で変わりなし

その他bull 不整脈気胸プロトコル施中の死亡

は両群とも起こらなかった

筆者らの考察bull PSVはweaningに最も使われているモードだ

がそれと較しても遜ないweaningができることがわかった

bull アシストが少ない傾向であったがPAV群のが呼吸窮迫となる回数が少なかった

rarr同調性が良くなった結果の可能性

ICU退室までの期間はPAV群で短いという結果であったがpowerが少ない

筆者らの考察bull 単施設での結果であり他の施設にこの

新しいプロトコルそのものが受けれられるかは不明である

bull SBTに失敗した患者を主に対象としておりICU患者すべてに適応はできない

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる

Page 8: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

Closed-loop Ventilationbull 患者の吸気努に例したサポートをうモード

Proportional Assist Ventilation(PAV)rarrPAV+(PB840980) PPS(Evita)

Neurally Adjusted Ventilatory Assist(NAVA)rarrServo i

bull 動weaningシステムにより離脱を指すモードSmartCare

rarrEvita XLinfinity V500Adaptive Support Ventilation(ASV)

rarrHammilton G5

患者の吸気努に例してサポートする

conventional ventilation The patient still has to generatesufficient flow to trigger the ventilator However as dis-cussed later conventional modes may force a larger tidalvolume than normal causing air trapping and thus increas-ing the number of missed triggers

Neurally Adjusted Ventilatory Assist

NAVA accomplishes the same goals as PAV but uti-lizes measurement of the diaphragmatic EMG signal tocontrol gas delivery6 This is accomplished by the place-ment of a specifically designed nasogastric tube that has aseries of EMG electrodes near its distal end positionedacross the diaphragm As illustrated in Figure 1 support-ing ventilation based on the diaphragmatic EMG signalshould greatly improve the response of the ventilator andsynchrony since the signal recognized is high up on theneural pathway controlling ventilation7 What is set by theclinician is the pressure applied for each millivolt of EMGactivity Thus similar to PAV a portion of the ventilatoryeffort is proportionally provided by the ventilator and theremainder by the patient As EMG activity increases pres-sure is applied during the inspiratory phase and as thediaphragm relaxes airway pressure decreases Inspirationends at a specific percentage of the peak EMG activityContrary to PAV NAVA greatly improves triggering since

gas delivery begins when the diaphragm is simulated notas a result of flow in the airway Thus even in the pres-ence of severe air trapping or large system leaks trigger-ing is not compromised

PAV and NAVA Versus Traditional Gas Delivery

Figure 2 nicely illustrates the relationship between andthe patientrsquos effort or work of breathing and the ventilatorpressure during various ventilation modes1 During a spe-cific breath with all forms of volume ventilation there isan indirect relationship between patient effort (Pmus) andthe pressure applied by the ventilator Essentially the morework assumed by the patient the less work applied by theventilator This relationship creates difficult problems re-garding synchrony which can easily be illustrated with alung model with proper setting of inspiratory demand Ifdemand is high and the lung model tidal volume is greaterthan the ventilator set tidal volume no pressure abovebaseline will be applied to the lung

In pressure-targeted ventilation regardless of mode ven-tilator work or applied pressure should be independent ofPmus That is during a specific breath a constant pressureshould be applied to the airway during inspiration nomatter how much work or Pmus is generated by the patientOf course in patients with very high ventilatory demandpressure may not be sustained since the ventilator may notalways be able to meet patient flow demand

In PAV and NAVA there is a direct relationship be-tween the patientrsquos effort (Pmus) and ventilator pressure

Fig 1 Steps in the process of activating a ventilator breath Neu-rally adjusted ventilatory assist ventilation modes are activatedand controlled by diaphragm excitation Proportional assist ven-tilation modes are activated by a change in airway pressure flowand volume (Adapted from Reference 6)

Fig 2 Relationship between patient effort and ventilator pressureduring various ventilation modes During volume ventilation pa-tient effort and ventilator pressure are inversely related Duringpressure ventilation ventilator pressure in unaffected by patienteffort During proportional assist ventilation (PAV) and neurally ad-justed ventilatory assist (NAVA) patient effort and ventilator pres-sure are directly related (Adapted from Reference 5)

PROPORTIONAL ASSIST VENTILATION AND NEURALLY ADJUSTED VENTILATORY ASSIST

142 RESPIRATORY CARE bull FEBRUARY 2011 VOL 56 NO 2

RespirCare201156(2)140ndash148

Proportional Assist Ventilation

bull コンプライアンス(C)とレジスタンス(R)を測定して患者の吸気努に例したサポートをう

bull Paw = VC + RtimesFlowbull WOB = ʃ PtimesFlow dt

bull PB840では8-15呼吸ごとに300msecの吸気ポーズをいCとRを測定

設定するのはサポート率のみ

患者の呼吸仕事量(WOBpt)が03〜08JLになるようにサポート率を設定する

PAVのメリットデメリット

bull メリット患者-呼吸器の同調性が良い呼吸筋を適度に使うことができる

bull デメリット患者の呼吸努が弱い場合はサポートが減ってしまうリークがある場合過度にサポート圧が上がってしまう可能性(runaway現象)

PAVのエビデンスbull PAVとPSVによるウィーニングをい48

時間観察したRCTではPAVのが調節呼吸に戻る割合が低かった(PAV 11 vsPSV 22 p=004)

bull 夜間PAVにすることで睡眠の質が改善したIntensiveCareMed2008342026ndash2034

Crit CareMed2007351048ndash1054

PAVは呼吸器離脱困難患者のウィーニングにおいて

優れたモードか

本の論

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Critical Care Medicine wwwccmjournalorg 1

Objectives Despite protocols incorporating spontaneous breath-ing trials 31 of ICU patients experience difficult or prolonged weaning from mechanical ventilation Nonfatiguing modes such as pressure support ventilation are recommended Proportional assist ventilation provides assistance in proportion to patient effort which may optimize weaning However it is not known how proportional assist ventilation performs relative to pressure sup-port ventilation over a prolonged period in the complex ICU set-ting The purpose of this study was to compare the physiologic and clinical performance (failure rate) safety and feasibility of protocols using daily spontaneous breathing trial plus pressure support ventilation versus proportional assist ventilation until ven-tilation discontinuationDesign Single-center unblinded pilot randomized controlled trialSetting Medical-surgical ICU of a tertiary-care hospitalPatients Adult patients intubated greater than 36 hours were ran-domized if they met eligibility criteria for partial ventilatory support

tolerated pressure support ventilation greater than or equal to 30 minutes and either failed or did not meet criteria for a sponta-neous breathing trialInterventions Patients were randomized to the pressure support ventilation or proportional assist ventilation protocol (PAV+ Puri-tan Bennett 840 Covidien Boulder CO) Both protocols used progressive decreases in level of assistance as tolerated coupled with daily assessment for spontaneous breathing trialsMeasurements and Main Results Of 54 patients randomized outcome data are available for 50 patients 27 were randomized to receive proportional assist ventilation and 23 to receive pres-sure support ventilation There were no adverse events linked to the study interventions and protocol violations were infrequent Recruitment was slower than projected (13 patients per month) The median (interquartile range) time from randomization to suc-cessful extubation was 39 days (28ndash84 d) on proportional assist ventilation versus 49 days (29ndash263 d) on pressure support ventilation (p = 039) Time to live ICU discharge was 73 days (52ndash114 d) on proportional assist ventilation versus 124 days (75ndash308 d) on pressure support ventilation (p = 003)Conclusion This pilot study demonstrates the utility safety and feasibility of the weaning protocols and provides important infor-mation to guide the design of a future randomized controlled trial comparing weaning from mechanical ventilation on pressure sup-port ventilation versus proportional assist ventilation (Crit Care Med 2016 XX00ndash00)Key Words assisted mechanical ventilation asynchrony index proportional assist ventilation respiratory muscles ventilator weaning

Difficult or prolonged weaning from mechanical ventilation (MV) affects 31 (range 26ndash42) of mechanically ventilated patients (1) and increases

patient morbidity mortality ICU length of stay and cost to the healthcare system (2) When initial weaning attempts fail to liberate patients from MV ideally clinicians should choose

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights ReservedDOI 101097CCM0000000000001600

1London Health Sciences Centre London ON Canada2Department of Medicine The University of Western Ontario London ON Canada

This study was performed at London Health Sciences Centre University Hospital London CanadaSupplemental digital content is available for this article Direct URL cita-tions appear in the printed text and are provided in the HTML and PDF versions of this article on the journalrsquos website (httpjournalslwwcomccmjournal)Supported in part by Critical Care Western Department of Medicine The University of Western Ontario Program of Experimental Medi-cine (POEM) Department of Medicine University of Western Ontario Lawson Health Research Institute London Health Sciences Centre London Canada and Covidien Boulder USADr Bosma has received honourarium for lectures and travel reimburse-ments from Covidien Dr Bosmarsquos institution has received grant support from Covidien for Dr Bosmarsquos research studies The remaining authors have disclosed that they do not have any potential conflicts of interestFor information regarding this article E-mail KarenJBosmalhsconca

A Pilot Randomized Trial Comparing Weaning From Mechanical Ventilation on Pressure Support Versus Proportional Assist Ventilation

Karen J Bosma MD FRCPC12 Brooke A Read MHS BSc RRT1 Mohammad J Bahrgard Nikoo MD FRCPC12 Philip M Jones MD FRCPC MSc12 Fran A Priestap MSc1 James F Lewis MD FRCPC12

Crit Care Med 2016 Jan 20[Epub ahead of print]

論のPICO

P 呼吸器離脱困難な患者(simple weaningではない)

I PAVを使

C PSVを使

O 呼吸器離脱までの時間

Study Designbull 単施設盲検前向きランダム化較

試験(pilot study)bull カナダロンドンにあるUniversity

Hospital-London Health Science Centre の内科外科ICU(20床)

bull ランダム割付はcomputer-generated random number sequence によってわれ隠蔽化されている

PatientsInclusion criteria

bull 18歳以上の挿管患者bull 36時間以上呼吸器管理をされている

PatientsExclusion criteria

1) 48時間以内にwithdrawが考慮される患者2) 位の脊髄障害進性の神経筋疾患の

よって慢性的な呼吸器依存となる可能性があるまたは脳外科的介が必要である患者

Enrolment Procedure

Supplemental Digital Content Figure 1 Enrolment Procedure

Daily Screening of all ICU patients gt18 yrs of age Intubated gt36 hours

Exclude if Exclusion Criteria

Successful SBT awaiting extubation Withdrawal of life support considered High spinal cord injury Neurosurgical patient Neuromuscular disease

Follow using daily eligibility checklist until meet all Clinical Stability Criteria to trial mode of partial ventilatory support (PAV or PSV)

partialcomplete reversal of cause of respiratory failure

body temperature between 36ordmC and 39ordmC metabolic disorders corrected pH gt732 hemoglobin gt70 gL no ongoing bleeding stable hemodynamic status Able to trigger PaO2 ge 60 mmHg or SaO2 ge 90 on FiO2le

060 and PEEP le 15 cm H2O ACVC Pplatle30cmH2O with Vt 6-8 mlkg

OR ACPC PC+PEEPle30cmH2O

Exclude if Exclusion Criteria Has met all Criteria for Switching to

Spontaneous Mode of Ventilation gt24 hours

Pre-inclusion Test on PSV 15 cmH2O for 30 minutes

Patient fails Pre-inclusion test Return to AC ventilation and reassess within 24 hour

Patient passes Pre-inclusion test Assess for Criteria to Initiate Weaning

PaO2FiO2gt200 on FiO2 le050 and PEEP le8 cmH2O No inotropesvasopressors except low dose dopamine

RSBI lt105 Patient passes Pre-SBT readiness assessment Proceed to SBT Place patient on CPAP 5 cmH2O and FiO2 040 for 30 minutes

Patient does not meet Criteria to Initiate Weaning

Patient meets Criteria to Initiate Weaning Proceed to Pre-SBT readiness assessment Place patient on CPAP for 1-2 min measure RSBI

RSBIge 105 Patient fails Pre-SBT readiness assessment

Patient passes SBT Patient fails SBT

Exclude if Exclusion Criteria

Passed SBT and tolerating PSVle7cmH2O and PEEPle5 cmH2O while awaiting extubation

Obtain informed consent from Patient or SDM as soon as available (deferred consent permitted)

Randomize

Exclude if Exclusion Criteria

Patient SDM refuse consent

bull 呼吸不全の原因に改善が認められてきている

bull 体温は36〜39度であるbull pH>732であるbull Hb>7gdLで現在出なしbull 動態が安定しているbull トリガーできるbull FiO2≦06 PEEP≦15cmH2Oで

PaO2≧60mmHg or SaO2≧90bull ACVCでVt 6-8mlkgで

Pplat≦30cmH2OもしくはACPCでPC+PEEP≦30cmH2O

上記をすべて満たせばプレテストとしてPSV 15cmH2Oに変更

Enrolment Procedure

Supplemental Digital Content Figure 1 Enrolment Procedure

Daily Screening of all ICU patients gt18 yrs of age Intubated gt36 hours

Exclude if Exclusion Criteria

Successful SBT awaiting extubation Withdrawal of life support considered High spinal cord injury Neurosurgical patient Neuromuscular disease

Follow using daily eligibility checklist until meet all Clinical Stability Criteria to trial mode of partial ventilatory support (PAV or PSV)

partialcomplete reversal of cause of respiratory failure

body temperature between 36ordmC and 39ordmC metabolic disorders corrected pH gt732 hemoglobin gt70 gL no ongoing bleeding stable hemodynamic status Able to trigger PaO2 ge 60 mmHg or SaO2 ge 90 on FiO2le

060 and PEEP le 15 cm H2O ACVC Pplatle30cmH2O with Vt 6-8 mlkg

OR ACPC PC+PEEPle30cmH2O

Exclude if Exclusion Criteria Has met all Criteria for Switching to

Spontaneous Mode of Ventilation gt24 hours

Pre-inclusion Test on PSV 15 cmH2O for 30 minutes

Patient fails Pre-inclusion test Return to AC ventilation and reassess within 24 hour

Patient passes Pre-inclusion test Assess for Criteria to Initiate Weaning

PaO2FiO2gt200 on FiO2 le050 and PEEP le8 cmH2O No inotropesvasopressors except low dose dopamine

RSBI lt105 Patient passes Pre-SBT readiness assessment Proceed to SBT Place patient on CPAP 5 cmH2O and FiO2 040 for 30 minutes

Patient does not meet Criteria to Initiate Weaning

Patient meets Criteria to Initiate Weaning Proceed to Pre-SBT readiness assessment Place patient on CPAP for 1-2 min measure RSBI

RSBIge 105 Patient fails Pre-SBT readiness assessment

Patient passes SBT Patient fails SBT

Exclude if Exclusion Criteria

Passed SBT and tolerating PSVle7cmH2O and PEEPle5 cmH2O while awaiting extubation

Obtain informed consent from Patient or SDM as soon as available (deferred consent permitted)

Randomize

Exclude if Exclusion Criteria

Patient SDM refuse consent

呼吸窮迫がなければプレテストクリアbull FiO2≦05 PEEP≦8cmH2OでPF≧200bull 強薬昇圧剤の使なしであればSBTへ

まず1-2分CPAP 5cmH2OにしてRSBIを測定

RSBI<105であればSBTへPSV 5cmH2Oにして30-120分

酸素化の条件を満たさないRSBI≧105SBT失敗であれば割り付けへ

Ventilation Protocols bull 呼吸器はPuritan Bennett 840 を使bull 後述するPAVプロトコールPSVプロトコー

ルに従って呼吸療法(RRT)が設定を変更(respiratory distress 呼吸窮迫がないようにできるだけサポートを下げていく)

bull 設定したmaxのサポートが必要になった場合はACに戻す

bull PAV群はPAV+とACPSV群はPSVとACしか使しない

Ventilation Protocols bull RCTの開始前に15名の患者をいて観察

研究をいRRTにプロトコルの確認をってもらった

bull プロトコルはすぐにられるようにラミネートしてベッドサイドに置いた

bull RRTは毎SBTをえるかどうか後述するチェックリストをいて確認除外項を満たさない場合は毎RSBIのアセスメントとSBTをう

Respiratory Distress bull 呼吸数>35回minbull 脈拍>140min もしくはベースラインより20以上の

上昇bull 収縮期圧>180mmHg もしくは<90mmHg ベース

ラインより30以上の変化bull 不安bull 発汗bull 副呼吸筋の使bull 奇異呼吸bull 呼吸困難の訴え

PAV protocolSupplemental Digital Content Figure 2a Proportional Assist Ventilation Protocol

bull 開始の設定はサポート率70

bull 2-3時間おきにRRTがアセスメントし問題なければ10-20ずつPAVサポートを下げる

サポート率90まで上げてもダメな場合はACに戻す

PSV protocolSupplemental Digital Content Figure 2b Pressure Support Ventilation Protocol

bull 開始の設定はPS 15cmH2O

bull まずRRlt35 VTgt5mLkgで耐えられるPSまで下げる

bull 2-3時間おきにRRTがアセスメントしてPS2-3cmH2Pずつ下げる

Data collection form for RRTsSupplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

SBTの開始基準bull PF≧200bull FiO2≦05bull pH≧732bull トリガーできるbull 24時間以内に筋虚イベントなしbull 強薬昇圧剤の使なし

Data collection form for RRTs

Supplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

CPAPにしてRSBIを測定RSBI<105であればSBT施

PAV群はサポート率20PSV群はPS 5cmH2OでSBTを施

PEEP≦8cmH2OでSaO2≧90呼吸窮迫なしであればSBTクリア

抜管へ

Endpoints 呼吸器離脱までの時間ICU病院滞在期間

呼吸器離脱成功の定義抜管後48時間再挿管なし侵襲的呼吸(NIV)の使もなし気管切開使時は呼吸器から最終的に外れた時点1のうち12時間以上NIVを使していない

Statistical Analysis bull Pilot studyのためサンプルサイズの計

算はしていないbull P値<005を統計学的有意としたbull Intention to treat approachをった

RESULTS

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial

エンロール期間は42か80名が基準を満たしが最終的に割付されたのは54名

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial最終的に解析がわれたのはPAV群27名PSV群23名

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

ベースラインは両群間で変わりなし重症度はAPACHEⅡ 27点程度患者は肺炎が3割もともとCOPDを持っている患者は2割程度

Feasibility Outcomes bull 呼吸窮迫でRRTが呼吸器サポートを上昇

させた回数はPAV群で76回232daysPSV群で124回295daysと有意にPSV群で多かった(p=0002)

bull ACに戻した回数はPAV群で66回PSV群で90回と変わりなし(p=061)

bull 全くACに戻さなかった患者はPAV群で627名PSV群で623名(p=10)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

ベースラインは両群間で変わりなしPAVにすることでΔP圧Peak圧が低く吸気時間呼吸時間がくなるPSV群と較するとPAV群のがPaoAreaPaoAreaVE(呼吸器のサポートの強さをす)が低くVTPaoAreaがくsecond post-randomizationの解析においてΔP圧Peak圧が低い

Asynchrony index

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 7

B635) documenting compliance with the daily screening for SBT noncompletion of either the CPAP screening test or SBT occurred on 5 of 292 of patient study days (17) and this violation involved two patients in each study arm The second audit examining for nine different potential deviations found complete adherence to study protocol on 254 of 292 of patient study days (900) There were no adverse events (pneumo-thorax arrhythmia or death) resulting from any of the study interventions

Physiologic Response to Ventilation ProtocolsRecordings of flow and airway pressure were available for analysis for all 50 patients prerandomization 49 patients immediately (median 1 interquartile range [IQR] 0ndash15 hr) postrandomization and for 26 patients during subsequent days (median 27 IQR 20ndash70-hr postrandomization) Approximately 15000 breaths in 125 recordings were analyzed Table 2 compares the respiratory parameters in both groups pre- and postrandomization There were no statistically sig-nificant differences between the PAV and PSV groups at base-line prerandomization (p values not shown) In the PAV group there was a significant decrease in peak airway pressure delta pressure (measured as peak airway opening pressure minus the PEEP) and area under the inspiratory pressure curve immediately postrandomization and an increase in the per-cent of total cycle time spent in inspiration (TiTtot) whereas there were no significant differences immediately postran-domization in the PSV group The second measurement of postrandomization respiratory parameters revealed signifi-cant decreases in peak airway pressure and delta pressure in both PAV and PSV groups Adjusting for baseline measure-ments as a covariate the PAV group received 7 cm H2

O∙s less ventilator assistance (PaoArea) per liter of minute ventilation (p lt 00001) and increased their VT by 115 mLcm H

2O∙s

relative to the PSV group (p lt 0001)PAV significantly reduced AI postrandomization (Fig 2A)

In total nine patients (18) had a high AI (gt 10 of breaths) on PSV or on AC mode at baseline No patients had a high AI on PAV In the PSV group six patients (26) had a high AI

at baseline four (17) immediately postrandomization and one patient (7) after further adjustment of ventilator settings according to the PSV protocol (Fig 2B) Adjusting for baseline differences patients in the PAV group tended to have a lower AI than patients in the PSV group immediately postrandom-ization but this did not reach statistical significance (adjusted difference ndash45 [95 CI ndash95 to 04] p = 007)

Clinical OutcomesTable 3 shows the results for the main clinical endpoints Weaning duration and time to successful SBT did not dif-fer significantly among survivors between the groups Among ICU survivors median time to ICU discharge was shorter in the group randomized to PAV 73 days (IQR 52ndash114 d) ver-sus 124 days (IQR 75ndash308 d) in the PSV group (p = 003) Kaplan-Meier plots of the proportion of at-risk patients achiev-ing successful extubation and live ICU discharge are shown in Figure 3 A and B No difference was found for hospital length of stay ICU mortality or hospital mortality Four patients (15) in the PAV group and three patients (13) in the PSV group died in ICU (respiratory rate [RR] of PAV 114 95 CI 028 to 46 p = 086) Overall in-hospital mortality was nine patients (33) in the PAV group and six patients (26) in the PSV group (RR of PAV 128 95 CI 054 to 31 p = 058)

Cointerventions Sedating MedicationsTable 4 compares the use of medications that may prolong weaning (by causing sedation or neuromuscular weakness) in the two groups pre- and postrandomization Type and dose of sedative medication were highly variable among individual patients and therefore we analyzed pre- and postrandomiza-tion changes to determine the impact of the ventilation pro-tocol on sedative use within each patient In the PAV group benzodiazepine narcotic and propofol use measured as both cumulative dose and average daily dose all decreased signifi-cantly postrandomization as did neuromuscular blocker use In the PSV group average daily dose of lorazepam equiva-lent and average daily dose of propofol decreased signifi-cantly postrandomization whereas the proportion of patients

Figure 2 A Asynchrony index pre- and postrandomization on proportional assist ventilation B Asynchrony index pre- and postrandomization on pressure support ventilationPAV群 PSV群

Asynchrony Index () = 同調イベントの回数全呼吸回数 times 100

IntensiveCareMed(2006)321515ndash1522

AI>10は3名いたがPAVに設定変更後0名となった

AI>10は6名いたがPSVに設定変更後4名となった

Clinical Outcomes

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

成功したSBTまでの数抜管成功までの数に変わりなしICU退室までの数はPAV群で有意に短い

Kaplan-Meier curves

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

抜管成功までの時間 ICU退室までの時間

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 9

TABLE 4 Comparison of Medications That May Alter Weaning Duration Pre- and Postrandomization

Variables

PAV Group (n = 27)

PSV Group (n = 23) Postrandomization

Pre randomization Pos trandomization Pre randomization Pos trandomizationAdjusted Difference PSV ndash PAV (95 CI)

Lorazepam equivalents

Cumulative dose mg

308 (207) 48 (257)a 137 (440) 54 (258) 237 (ndash580 to 1055)

Average daily dose mg

44 (255) 09 (43)a 23 (85) 06 (40)a 24 (ndash62 to 110)

Single-day dose mgb

08 (72) 12 (50) 16 (70) 10 (64) 32 (ndash13 to 78)

Morphine equivalents

Cumulative dose mg

5003 (6188) 1140 (3140)b 1710 (13310) 1610 (4860) 591 (ndash2515 to 3696)

Average daily dose mg

858 (1638) 170 (578)c 428 (1724) 136 (320)d 108 (ndash316 to 532)

Single-day dose mga

180 (1240) 187 (77) 100 (1920) 50 (630)d 334 (ndash026 to 670)d

Propofol

Cumulative dose mg

7400 (55500) 0 (700)a 29450 (85800) 0 (27660) 226 (ndash42156 to 42607)

Average daily dose mg

529 (11690) 0 (53)a 4879 (15980) 0 (2738)a 651 (ndash3211 to 4514)

Single-day dose mga

0 (0) 0 (0) 0 (1000) 0 (0) 344 (ndash3781 to 4469)

Neuroleptics

Days receiving neuroleptics n

0 (2) 0 (4) 0 (0) 1 (4)a ndash16 (ndash34 to 03)

Proportion of patients who received a neuroleptic

407 407 174 565b

NMBs

Days receiving NMBs n

1 (1) 0 (0)a 0 (1) 0 (0) ndash01 (ndash05 to 03)

Proportion of patients who received an NMB

518 148a 348 174

Corticosteroids

Days receiving corticosteroids n

0 (0) 0 (0) 0 (2) 0 (2) 02 (ndash06 to 10)

Proportion of patients who received corticosteroids

185 148 348 261

PAV = proportional assist ventilation PSV = pressure support ventilation NMB = neuromuscular blockera p lt 005b Single-day dose = total dose given day prior to randomization or total dose give day after randomizationc p lt 0005d p lt 00005e p le 007All data points except percentage of patients receiving neuroleptics NMBs or corticosteroids are presented as median (interquartile range)

鎮静剤鎮痛剤についてはPAV群とPSV群で変わりなし

その他bull 不整脈気胸プロトコル施中の死亡

は両群とも起こらなかった

筆者らの考察bull PSVはweaningに最も使われているモードだ

がそれと較しても遜ないweaningができることがわかった

bull アシストが少ない傾向であったがPAV群のが呼吸窮迫となる回数が少なかった

rarr同調性が良くなった結果の可能性

ICU退室までの期間はPAV群で短いという結果であったがpowerが少ない

筆者らの考察bull 単施設での結果であり他の施設にこの

新しいプロトコルそのものが受けれられるかは不明である

bull SBTに失敗した患者を主に対象としておりICU患者すべてに適応はできない

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる

Page 9: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

患者の吸気努に例してサポートする

conventional ventilation The patient still has to generatesufficient flow to trigger the ventilator However as dis-cussed later conventional modes may force a larger tidalvolume than normal causing air trapping and thus increas-ing the number of missed triggers

Neurally Adjusted Ventilatory Assist

NAVA accomplishes the same goals as PAV but uti-lizes measurement of the diaphragmatic EMG signal tocontrol gas delivery6 This is accomplished by the place-ment of a specifically designed nasogastric tube that has aseries of EMG electrodes near its distal end positionedacross the diaphragm As illustrated in Figure 1 support-ing ventilation based on the diaphragmatic EMG signalshould greatly improve the response of the ventilator andsynchrony since the signal recognized is high up on theneural pathway controlling ventilation7 What is set by theclinician is the pressure applied for each millivolt of EMGactivity Thus similar to PAV a portion of the ventilatoryeffort is proportionally provided by the ventilator and theremainder by the patient As EMG activity increases pres-sure is applied during the inspiratory phase and as thediaphragm relaxes airway pressure decreases Inspirationends at a specific percentage of the peak EMG activityContrary to PAV NAVA greatly improves triggering since

gas delivery begins when the diaphragm is simulated notas a result of flow in the airway Thus even in the pres-ence of severe air trapping or large system leaks trigger-ing is not compromised

PAV and NAVA Versus Traditional Gas Delivery

Figure 2 nicely illustrates the relationship between andthe patientrsquos effort or work of breathing and the ventilatorpressure during various ventilation modes1 During a spe-cific breath with all forms of volume ventilation there isan indirect relationship between patient effort (Pmus) andthe pressure applied by the ventilator Essentially the morework assumed by the patient the less work applied by theventilator This relationship creates difficult problems re-garding synchrony which can easily be illustrated with alung model with proper setting of inspiratory demand Ifdemand is high and the lung model tidal volume is greaterthan the ventilator set tidal volume no pressure abovebaseline will be applied to the lung

In pressure-targeted ventilation regardless of mode ven-tilator work or applied pressure should be independent ofPmus That is during a specific breath a constant pressureshould be applied to the airway during inspiration nomatter how much work or Pmus is generated by the patientOf course in patients with very high ventilatory demandpressure may not be sustained since the ventilator may notalways be able to meet patient flow demand

In PAV and NAVA there is a direct relationship be-tween the patientrsquos effort (Pmus) and ventilator pressure

Fig 1 Steps in the process of activating a ventilator breath Neu-rally adjusted ventilatory assist ventilation modes are activatedand controlled by diaphragm excitation Proportional assist ven-tilation modes are activated by a change in airway pressure flowand volume (Adapted from Reference 6)

Fig 2 Relationship between patient effort and ventilator pressureduring various ventilation modes During volume ventilation pa-tient effort and ventilator pressure are inversely related Duringpressure ventilation ventilator pressure in unaffected by patienteffort During proportional assist ventilation (PAV) and neurally ad-justed ventilatory assist (NAVA) patient effort and ventilator pres-sure are directly related (Adapted from Reference 5)

PROPORTIONAL ASSIST VENTILATION AND NEURALLY ADJUSTED VENTILATORY ASSIST

142 RESPIRATORY CARE bull FEBRUARY 2011 VOL 56 NO 2

RespirCare201156(2)140ndash148

Proportional Assist Ventilation

bull コンプライアンス(C)とレジスタンス(R)を測定して患者の吸気努に例したサポートをう

bull Paw = VC + RtimesFlowbull WOB = ʃ PtimesFlow dt

bull PB840では8-15呼吸ごとに300msecの吸気ポーズをいCとRを測定

設定するのはサポート率のみ

患者の呼吸仕事量(WOBpt)が03〜08JLになるようにサポート率を設定する

PAVのメリットデメリット

bull メリット患者-呼吸器の同調性が良い呼吸筋を適度に使うことができる

bull デメリット患者の呼吸努が弱い場合はサポートが減ってしまうリークがある場合過度にサポート圧が上がってしまう可能性(runaway現象)

PAVのエビデンスbull PAVとPSVによるウィーニングをい48

時間観察したRCTではPAVのが調節呼吸に戻る割合が低かった(PAV 11 vsPSV 22 p=004)

bull 夜間PAVにすることで睡眠の質が改善したIntensiveCareMed2008342026ndash2034

Crit CareMed2007351048ndash1054

PAVは呼吸器離脱困難患者のウィーニングにおいて

優れたモードか

本の論

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Critical Care Medicine wwwccmjournalorg 1

Objectives Despite protocols incorporating spontaneous breath-ing trials 31 of ICU patients experience difficult or prolonged weaning from mechanical ventilation Nonfatiguing modes such as pressure support ventilation are recommended Proportional assist ventilation provides assistance in proportion to patient effort which may optimize weaning However it is not known how proportional assist ventilation performs relative to pressure sup-port ventilation over a prolonged period in the complex ICU set-ting The purpose of this study was to compare the physiologic and clinical performance (failure rate) safety and feasibility of protocols using daily spontaneous breathing trial plus pressure support ventilation versus proportional assist ventilation until ven-tilation discontinuationDesign Single-center unblinded pilot randomized controlled trialSetting Medical-surgical ICU of a tertiary-care hospitalPatients Adult patients intubated greater than 36 hours were ran-domized if they met eligibility criteria for partial ventilatory support

tolerated pressure support ventilation greater than or equal to 30 minutes and either failed or did not meet criteria for a sponta-neous breathing trialInterventions Patients were randomized to the pressure support ventilation or proportional assist ventilation protocol (PAV+ Puri-tan Bennett 840 Covidien Boulder CO) Both protocols used progressive decreases in level of assistance as tolerated coupled with daily assessment for spontaneous breathing trialsMeasurements and Main Results Of 54 patients randomized outcome data are available for 50 patients 27 were randomized to receive proportional assist ventilation and 23 to receive pres-sure support ventilation There were no adverse events linked to the study interventions and protocol violations were infrequent Recruitment was slower than projected (13 patients per month) The median (interquartile range) time from randomization to suc-cessful extubation was 39 days (28ndash84 d) on proportional assist ventilation versus 49 days (29ndash263 d) on pressure support ventilation (p = 039) Time to live ICU discharge was 73 days (52ndash114 d) on proportional assist ventilation versus 124 days (75ndash308 d) on pressure support ventilation (p = 003)Conclusion This pilot study demonstrates the utility safety and feasibility of the weaning protocols and provides important infor-mation to guide the design of a future randomized controlled trial comparing weaning from mechanical ventilation on pressure sup-port ventilation versus proportional assist ventilation (Crit Care Med 2016 XX00ndash00)Key Words assisted mechanical ventilation asynchrony index proportional assist ventilation respiratory muscles ventilator weaning

Difficult or prolonged weaning from mechanical ventilation (MV) affects 31 (range 26ndash42) of mechanically ventilated patients (1) and increases

patient morbidity mortality ICU length of stay and cost to the healthcare system (2) When initial weaning attempts fail to liberate patients from MV ideally clinicians should choose

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights ReservedDOI 101097CCM0000000000001600

1London Health Sciences Centre London ON Canada2Department of Medicine The University of Western Ontario London ON Canada

This study was performed at London Health Sciences Centre University Hospital London CanadaSupplemental digital content is available for this article Direct URL cita-tions appear in the printed text and are provided in the HTML and PDF versions of this article on the journalrsquos website (httpjournalslwwcomccmjournal)Supported in part by Critical Care Western Department of Medicine The University of Western Ontario Program of Experimental Medi-cine (POEM) Department of Medicine University of Western Ontario Lawson Health Research Institute London Health Sciences Centre London Canada and Covidien Boulder USADr Bosma has received honourarium for lectures and travel reimburse-ments from Covidien Dr Bosmarsquos institution has received grant support from Covidien for Dr Bosmarsquos research studies The remaining authors have disclosed that they do not have any potential conflicts of interestFor information regarding this article E-mail KarenJBosmalhsconca

A Pilot Randomized Trial Comparing Weaning From Mechanical Ventilation on Pressure Support Versus Proportional Assist Ventilation

Karen J Bosma MD FRCPC12 Brooke A Read MHS BSc RRT1 Mohammad J Bahrgard Nikoo MD FRCPC12 Philip M Jones MD FRCPC MSc12 Fran A Priestap MSc1 James F Lewis MD FRCPC12

Crit Care Med 2016 Jan 20[Epub ahead of print]

論のPICO

P 呼吸器離脱困難な患者(simple weaningではない)

I PAVを使

C PSVを使

O 呼吸器離脱までの時間

Study Designbull 単施設盲検前向きランダム化較

試験(pilot study)bull カナダロンドンにあるUniversity

Hospital-London Health Science Centre の内科外科ICU(20床)

bull ランダム割付はcomputer-generated random number sequence によってわれ隠蔽化されている

PatientsInclusion criteria

bull 18歳以上の挿管患者bull 36時間以上呼吸器管理をされている

PatientsExclusion criteria

1) 48時間以内にwithdrawが考慮される患者2) 位の脊髄障害進性の神経筋疾患の

よって慢性的な呼吸器依存となる可能性があるまたは脳外科的介が必要である患者

Enrolment Procedure

Supplemental Digital Content Figure 1 Enrolment Procedure

Daily Screening of all ICU patients gt18 yrs of age Intubated gt36 hours

Exclude if Exclusion Criteria

Successful SBT awaiting extubation Withdrawal of life support considered High spinal cord injury Neurosurgical patient Neuromuscular disease

Follow using daily eligibility checklist until meet all Clinical Stability Criteria to trial mode of partial ventilatory support (PAV or PSV)

partialcomplete reversal of cause of respiratory failure

body temperature between 36ordmC and 39ordmC metabolic disorders corrected pH gt732 hemoglobin gt70 gL no ongoing bleeding stable hemodynamic status Able to trigger PaO2 ge 60 mmHg or SaO2 ge 90 on FiO2le

060 and PEEP le 15 cm H2O ACVC Pplatle30cmH2O with Vt 6-8 mlkg

OR ACPC PC+PEEPle30cmH2O

Exclude if Exclusion Criteria Has met all Criteria for Switching to

Spontaneous Mode of Ventilation gt24 hours

Pre-inclusion Test on PSV 15 cmH2O for 30 minutes

Patient fails Pre-inclusion test Return to AC ventilation and reassess within 24 hour

Patient passes Pre-inclusion test Assess for Criteria to Initiate Weaning

PaO2FiO2gt200 on FiO2 le050 and PEEP le8 cmH2O No inotropesvasopressors except low dose dopamine

RSBI lt105 Patient passes Pre-SBT readiness assessment Proceed to SBT Place patient on CPAP 5 cmH2O and FiO2 040 for 30 minutes

Patient does not meet Criteria to Initiate Weaning

Patient meets Criteria to Initiate Weaning Proceed to Pre-SBT readiness assessment Place patient on CPAP for 1-2 min measure RSBI

RSBIge 105 Patient fails Pre-SBT readiness assessment

Patient passes SBT Patient fails SBT

Exclude if Exclusion Criteria

Passed SBT and tolerating PSVle7cmH2O and PEEPle5 cmH2O while awaiting extubation

Obtain informed consent from Patient or SDM as soon as available (deferred consent permitted)

Randomize

Exclude if Exclusion Criteria

Patient SDM refuse consent

bull 呼吸不全の原因に改善が認められてきている

bull 体温は36〜39度であるbull pH>732であるbull Hb>7gdLで現在出なしbull 動態が安定しているbull トリガーできるbull FiO2≦06 PEEP≦15cmH2Oで

PaO2≧60mmHg or SaO2≧90bull ACVCでVt 6-8mlkgで

Pplat≦30cmH2OもしくはACPCでPC+PEEP≦30cmH2O

上記をすべて満たせばプレテストとしてPSV 15cmH2Oに変更

Enrolment Procedure

Supplemental Digital Content Figure 1 Enrolment Procedure

Daily Screening of all ICU patients gt18 yrs of age Intubated gt36 hours

Exclude if Exclusion Criteria

Successful SBT awaiting extubation Withdrawal of life support considered High spinal cord injury Neurosurgical patient Neuromuscular disease

Follow using daily eligibility checklist until meet all Clinical Stability Criteria to trial mode of partial ventilatory support (PAV or PSV)

partialcomplete reversal of cause of respiratory failure

body temperature between 36ordmC and 39ordmC metabolic disorders corrected pH gt732 hemoglobin gt70 gL no ongoing bleeding stable hemodynamic status Able to trigger PaO2 ge 60 mmHg or SaO2 ge 90 on FiO2le

060 and PEEP le 15 cm H2O ACVC Pplatle30cmH2O with Vt 6-8 mlkg

OR ACPC PC+PEEPle30cmH2O

Exclude if Exclusion Criteria Has met all Criteria for Switching to

Spontaneous Mode of Ventilation gt24 hours

Pre-inclusion Test on PSV 15 cmH2O for 30 minutes

Patient fails Pre-inclusion test Return to AC ventilation and reassess within 24 hour

Patient passes Pre-inclusion test Assess for Criteria to Initiate Weaning

PaO2FiO2gt200 on FiO2 le050 and PEEP le8 cmH2O No inotropesvasopressors except low dose dopamine

RSBI lt105 Patient passes Pre-SBT readiness assessment Proceed to SBT Place patient on CPAP 5 cmH2O and FiO2 040 for 30 minutes

Patient does not meet Criteria to Initiate Weaning

Patient meets Criteria to Initiate Weaning Proceed to Pre-SBT readiness assessment Place patient on CPAP for 1-2 min measure RSBI

RSBIge 105 Patient fails Pre-SBT readiness assessment

Patient passes SBT Patient fails SBT

Exclude if Exclusion Criteria

Passed SBT and tolerating PSVle7cmH2O and PEEPle5 cmH2O while awaiting extubation

Obtain informed consent from Patient or SDM as soon as available (deferred consent permitted)

Randomize

Exclude if Exclusion Criteria

Patient SDM refuse consent

呼吸窮迫がなければプレテストクリアbull FiO2≦05 PEEP≦8cmH2OでPF≧200bull 強薬昇圧剤の使なしであればSBTへ

まず1-2分CPAP 5cmH2OにしてRSBIを測定

RSBI<105であればSBTへPSV 5cmH2Oにして30-120分

酸素化の条件を満たさないRSBI≧105SBT失敗であれば割り付けへ

Ventilation Protocols bull 呼吸器はPuritan Bennett 840 を使bull 後述するPAVプロトコールPSVプロトコー

ルに従って呼吸療法(RRT)が設定を変更(respiratory distress 呼吸窮迫がないようにできるだけサポートを下げていく)

bull 設定したmaxのサポートが必要になった場合はACに戻す

bull PAV群はPAV+とACPSV群はPSVとACしか使しない

Ventilation Protocols bull RCTの開始前に15名の患者をいて観察

研究をいRRTにプロトコルの確認をってもらった

bull プロトコルはすぐにられるようにラミネートしてベッドサイドに置いた

bull RRTは毎SBTをえるかどうか後述するチェックリストをいて確認除外項を満たさない場合は毎RSBIのアセスメントとSBTをう

Respiratory Distress bull 呼吸数>35回minbull 脈拍>140min もしくはベースラインより20以上の

上昇bull 収縮期圧>180mmHg もしくは<90mmHg ベース

ラインより30以上の変化bull 不安bull 発汗bull 副呼吸筋の使bull 奇異呼吸bull 呼吸困難の訴え

PAV protocolSupplemental Digital Content Figure 2a Proportional Assist Ventilation Protocol

bull 開始の設定はサポート率70

bull 2-3時間おきにRRTがアセスメントし問題なければ10-20ずつPAVサポートを下げる

サポート率90まで上げてもダメな場合はACに戻す

PSV protocolSupplemental Digital Content Figure 2b Pressure Support Ventilation Protocol

bull 開始の設定はPS 15cmH2O

bull まずRRlt35 VTgt5mLkgで耐えられるPSまで下げる

bull 2-3時間おきにRRTがアセスメントしてPS2-3cmH2Pずつ下げる

Data collection form for RRTsSupplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

SBTの開始基準bull PF≧200bull FiO2≦05bull pH≧732bull トリガーできるbull 24時間以内に筋虚イベントなしbull 強薬昇圧剤の使なし

Data collection form for RRTs

Supplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

CPAPにしてRSBIを測定RSBI<105であればSBT施

PAV群はサポート率20PSV群はPS 5cmH2OでSBTを施

PEEP≦8cmH2OでSaO2≧90呼吸窮迫なしであればSBTクリア

抜管へ

Endpoints 呼吸器離脱までの時間ICU病院滞在期間

呼吸器離脱成功の定義抜管後48時間再挿管なし侵襲的呼吸(NIV)の使もなし気管切開使時は呼吸器から最終的に外れた時点1のうち12時間以上NIVを使していない

Statistical Analysis bull Pilot studyのためサンプルサイズの計

算はしていないbull P値<005を統計学的有意としたbull Intention to treat approachをった

RESULTS

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial

エンロール期間は42か80名が基準を満たしが最終的に割付されたのは54名

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial最終的に解析がわれたのはPAV群27名PSV群23名

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

ベースラインは両群間で変わりなし重症度はAPACHEⅡ 27点程度患者は肺炎が3割もともとCOPDを持っている患者は2割程度

Feasibility Outcomes bull 呼吸窮迫でRRTが呼吸器サポートを上昇

させた回数はPAV群で76回232daysPSV群で124回295daysと有意にPSV群で多かった(p=0002)

bull ACに戻した回数はPAV群で66回PSV群で90回と変わりなし(p=061)

bull 全くACに戻さなかった患者はPAV群で627名PSV群で623名(p=10)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

ベースラインは両群間で変わりなしPAVにすることでΔP圧Peak圧が低く吸気時間呼吸時間がくなるPSV群と較するとPAV群のがPaoAreaPaoAreaVE(呼吸器のサポートの強さをす)が低くVTPaoAreaがくsecond post-randomizationの解析においてΔP圧Peak圧が低い

Asynchrony index

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 7

B635) documenting compliance with the daily screening for SBT noncompletion of either the CPAP screening test or SBT occurred on 5 of 292 of patient study days (17) and this violation involved two patients in each study arm The second audit examining for nine different potential deviations found complete adherence to study protocol on 254 of 292 of patient study days (900) There were no adverse events (pneumo-thorax arrhythmia or death) resulting from any of the study interventions

Physiologic Response to Ventilation ProtocolsRecordings of flow and airway pressure were available for analysis for all 50 patients prerandomization 49 patients immediately (median 1 interquartile range [IQR] 0ndash15 hr) postrandomization and for 26 patients during subsequent days (median 27 IQR 20ndash70-hr postrandomization) Approximately 15000 breaths in 125 recordings were analyzed Table 2 compares the respiratory parameters in both groups pre- and postrandomization There were no statistically sig-nificant differences between the PAV and PSV groups at base-line prerandomization (p values not shown) In the PAV group there was a significant decrease in peak airway pressure delta pressure (measured as peak airway opening pressure minus the PEEP) and area under the inspiratory pressure curve immediately postrandomization and an increase in the per-cent of total cycle time spent in inspiration (TiTtot) whereas there were no significant differences immediately postran-domization in the PSV group The second measurement of postrandomization respiratory parameters revealed signifi-cant decreases in peak airway pressure and delta pressure in both PAV and PSV groups Adjusting for baseline measure-ments as a covariate the PAV group received 7 cm H2

O∙s less ventilator assistance (PaoArea) per liter of minute ventilation (p lt 00001) and increased their VT by 115 mLcm H

2O∙s

relative to the PSV group (p lt 0001)PAV significantly reduced AI postrandomization (Fig 2A)

In total nine patients (18) had a high AI (gt 10 of breaths) on PSV or on AC mode at baseline No patients had a high AI on PAV In the PSV group six patients (26) had a high AI

at baseline four (17) immediately postrandomization and one patient (7) after further adjustment of ventilator settings according to the PSV protocol (Fig 2B) Adjusting for baseline differences patients in the PAV group tended to have a lower AI than patients in the PSV group immediately postrandom-ization but this did not reach statistical significance (adjusted difference ndash45 [95 CI ndash95 to 04] p = 007)

Clinical OutcomesTable 3 shows the results for the main clinical endpoints Weaning duration and time to successful SBT did not dif-fer significantly among survivors between the groups Among ICU survivors median time to ICU discharge was shorter in the group randomized to PAV 73 days (IQR 52ndash114 d) ver-sus 124 days (IQR 75ndash308 d) in the PSV group (p = 003) Kaplan-Meier plots of the proportion of at-risk patients achiev-ing successful extubation and live ICU discharge are shown in Figure 3 A and B No difference was found for hospital length of stay ICU mortality or hospital mortality Four patients (15) in the PAV group and three patients (13) in the PSV group died in ICU (respiratory rate [RR] of PAV 114 95 CI 028 to 46 p = 086) Overall in-hospital mortality was nine patients (33) in the PAV group and six patients (26) in the PSV group (RR of PAV 128 95 CI 054 to 31 p = 058)

Cointerventions Sedating MedicationsTable 4 compares the use of medications that may prolong weaning (by causing sedation or neuromuscular weakness) in the two groups pre- and postrandomization Type and dose of sedative medication were highly variable among individual patients and therefore we analyzed pre- and postrandomiza-tion changes to determine the impact of the ventilation pro-tocol on sedative use within each patient In the PAV group benzodiazepine narcotic and propofol use measured as both cumulative dose and average daily dose all decreased signifi-cantly postrandomization as did neuromuscular blocker use In the PSV group average daily dose of lorazepam equiva-lent and average daily dose of propofol decreased signifi-cantly postrandomization whereas the proportion of patients

Figure 2 A Asynchrony index pre- and postrandomization on proportional assist ventilation B Asynchrony index pre- and postrandomization on pressure support ventilationPAV群 PSV群

Asynchrony Index () = 同調イベントの回数全呼吸回数 times 100

IntensiveCareMed(2006)321515ndash1522

AI>10は3名いたがPAVに設定変更後0名となった

AI>10は6名いたがPSVに設定変更後4名となった

Clinical Outcomes

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

成功したSBTまでの数抜管成功までの数に変わりなしICU退室までの数はPAV群で有意に短い

Kaplan-Meier curves

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

抜管成功までの時間 ICU退室までの時間

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 9

TABLE 4 Comparison of Medications That May Alter Weaning Duration Pre- and Postrandomization

Variables

PAV Group (n = 27)

PSV Group (n = 23) Postrandomization

Pre randomization Pos trandomization Pre randomization Pos trandomizationAdjusted Difference PSV ndash PAV (95 CI)

Lorazepam equivalents

Cumulative dose mg

308 (207) 48 (257)a 137 (440) 54 (258) 237 (ndash580 to 1055)

Average daily dose mg

44 (255) 09 (43)a 23 (85) 06 (40)a 24 (ndash62 to 110)

Single-day dose mgb

08 (72) 12 (50) 16 (70) 10 (64) 32 (ndash13 to 78)

Morphine equivalents

Cumulative dose mg

5003 (6188) 1140 (3140)b 1710 (13310) 1610 (4860) 591 (ndash2515 to 3696)

Average daily dose mg

858 (1638) 170 (578)c 428 (1724) 136 (320)d 108 (ndash316 to 532)

Single-day dose mga

180 (1240) 187 (77) 100 (1920) 50 (630)d 334 (ndash026 to 670)d

Propofol

Cumulative dose mg

7400 (55500) 0 (700)a 29450 (85800) 0 (27660) 226 (ndash42156 to 42607)

Average daily dose mg

529 (11690) 0 (53)a 4879 (15980) 0 (2738)a 651 (ndash3211 to 4514)

Single-day dose mga

0 (0) 0 (0) 0 (1000) 0 (0) 344 (ndash3781 to 4469)

Neuroleptics

Days receiving neuroleptics n

0 (2) 0 (4) 0 (0) 1 (4)a ndash16 (ndash34 to 03)

Proportion of patients who received a neuroleptic

407 407 174 565b

NMBs

Days receiving NMBs n

1 (1) 0 (0)a 0 (1) 0 (0) ndash01 (ndash05 to 03)

Proportion of patients who received an NMB

518 148a 348 174

Corticosteroids

Days receiving corticosteroids n

0 (0) 0 (0) 0 (2) 0 (2) 02 (ndash06 to 10)

Proportion of patients who received corticosteroids

185 148 348 261

PAV = proportional assist ventilation PSV = pressure support ventilation NMB = neuromuscular blockera p lt 005b Single-day dose = total dose given day prior to randomization or total dose give day after randomizationc p lt 0005d p lt 00005e p le 007All data points except percentage of patients receiving neuroleptics NMBs or corticosteroids are presented as median (interquartile range)

鎮静剤鎮痛剤についてはPAV群とPSV群で変わりなし

その他bull 不整脈気胸プロトコル施中の死亡

は両群とも起こらなかった

筆者らの考察bull PSVはweaningに最も使われているモードだ

がそれと較しても遜ないweaningができることがわかった

bull アシストが少ない傾向であったがPAV群のが呼吸窮迫となる回数が少なかった

rarr同調性が良くなった結果の可能性

ICU退室までの期間はPAV群で短いという結果であったがpowerが少ない

筆者らの考察bull 単施設での結果であり他の施設にこの

新しいプロトコルそのものが受けれられるかは不明である

bull SBTに失敗した患者を主に対象としておりICU患者すべてに適応はできない

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる

Page 10: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

Proportional Assist Ventilation

bull コンプライアンス(C)とレジスタンス(R)を測定して患者の吸気努に例したサポートをう

bull Paw = VC + RtimesFlowbull WOB = ʃ PtimesFlow dt

bull PB840では8-15呼吸ごとに300msecの吸気ポーズをいCとRを測定

設定するのはサポート率のみ

患者の呼吸仕事量(WOBpt)が03〜08JLになるようにサポート率を設定する

PAVのメリットデメリット

bull メリット患者-呼吸器の同調性が良い呼吸筋を適度に使うことができる

bull デメリット患者の呼吸努が弱い場合はサポートが減ってしまうリークがある場合過度にサポート圧が上がってしまう可能性(runaway現象)

PAVのエビデンスbull PAVとPSVによるウィーニングをい48

時間観察したRCTではPAVのが調節呼吸に戻る割合が低かった(PAV 11 vsPSV 22 p=004)

bull 夜間PAVにすることで睡眠の質が改善したIntensiveCareMed2008342026ndash2034

Crit CareMed2007351048ndash1054

PAVは呼吸器離脱困難患者のウィーニングにおいて

優れたモードか

本の論

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Critical Care Medicine wwwccmjournalorg 1

Objectives Despite protocols incorporating spontaneous breath-ing trials 31 of ICU patients experience difficult or prolonged weaning from mechanical ventilation Nonfatiguing modes such as pressure support ventilation are recommended Proportional assist ventilation provides assistance in proportion to patient effort which may optimize weaning However it is not known how proportional assist ventilation performs relative to pressure sup-port ventilation over a prolonged period in the complex ICU set-ting The purpose of this study was to compare the physiologic and clinical performance (failure rate) safety and feasibility of protocols using daily spontaneous breathing trial plus pressure support ventilation versus proportional assist ventilation until ven-tilation discontinuationDesign Single-center unblinded pilot randomized controlled trialSetting Medical-surgical ICU of a tertiary-care hospitalPatients Adult patients intubated greater than 36 hours were ran-domized if they met eligibility criteria for partial ventilatory support

tolerated pressure support ventilation greater than or equal to 30 minutes and either failed or did not meet criteria for a sponta-neous breathing trialInterventions Patients were randomized to the pressure support ventilation or proportional assist ventilation protocol (PAV+ Puri-tan Bennett 840 Covidien Boulder CO) Both protocols used progressive decreases in level of assistance as tolerated coupled with daily assessment for spontaneous breathing trialsMeasurements and Main Results Of 54 patients randomized outcome data are available for 50 patients 27 were randomized to receive proportional assist ventilation and 23 to receive pres-sure support ventilation There were no adverse events linked to the study interventions and protocol violations were infrequent Recruitment was slower than projected (13 patients per month) The median (interquartile range) time from randomization to suc-cessful extubation was 39 days (28ndash84 d) on proportional assist ventilation versus 49 days (29ndash263 d) on pressure support ventilation (p = 039) Time to live ICU discharge was 73 days (52ndash114 d) on proportional assist ventilation versus 124 days (75ndash308 d) on pressure support ventilation (p = 003)Conclusion This pilot study demonstrates the utility safety and feasibility of the weaning protocols and provides important infor-mation to guide the design of a future randomized controlled trial comparing weaning from mechanical ventilation on pressure sup-port ventilation versus proportional assist ventilation (Crit Care Med 2016 XX00ndash00)Key Words assisted mechanical ventilation asynchrony index proportional assist ventilation respiratory muscles ventilator weaning

Difficult or prolonged weaning from mechanical ventilation (MV) affects 31 (range 26ndash42) of mechanically ventilated patients (1) and increases

patient morbidity mortality ICU length of stay and cost to the healthcare system (2) When initial weaning attempts fail to liberate patients from MV ideally clinicians should choose

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights ReservedDOI 101097CCM0000000000001600

1London Health Sciences Centre London ON Canada2Department of Medicine The University of Western Ontario London ON Canada

This study was performed at London Health Sciences Centre University Hospital London CanadaSupplemental digital content is available for this article Direct URL cita-tions appear in the printed text and are provided in the HTML and PDF versions of this article on the journalrsquos website (httpjournalslwwcomccmjournal)Supported in part by Critical Care Western Department of Medicine The University of Western Ontario Program of Experimental Medi-cine (POEM) Department of Medicine University of Western Ontario Lawson Health Research Institute London Health Sciences Centre London Canada and Covidien Boulder USADr Bosma has received honourarium for lectures and travel reimburse-ments from Covidien Dr Bosmarsquos institution has received grant support from Covidien for Dr Bosmarsquos research studies The remaining authors have disclosed that they do not have any potential conflicts of interestFor information regarding this article E-mail KarenJBosmalhsconca

A Pilot Randomized Trial Comparing Weaning From Mechanical Ventilation on Pressure Support Versus Proportional Assist Ventilation

Karen J Bosma MD FRCPC12 Brooke A Read MHS BSc RRT1 Mohammad J Bahrgard Nikoo MD FRCPC12 Philip M Jones MD FRCPC MSc12 Fran A Priestap MSc1 James F Lewis MD FRCPC12

Crit Care Med 2016 Jan 20[Epub ahead of print]

論のPICO

P 呼吸器離脱困難な患者(simple weaningではない)

I PAVを使

C PSVを使

O 呼吸器離脱までの時間

Study Designbull 単施設盲検前向きランダム化較

試験(pilot study)bull カナダロンドンにあるUniversity

Hospital-London Health Science Centre の内科外科ICU(20床)

bull ランダム割付はcomputer-generated random number sequence によってわれ隠蔽化されている

PatientsInclusion criteria

bull 18歳以上の挿管患者bull 36時間以上呼吸器管理をされている

PatientsExclusion criteria

1) 48時間以内にwithdrawが考慮される患者2) 位の脊髄障害進性の神経筋疾患の

よって慢性的な呼吸器依存となる可能性があるまたは脳外科的介が必要である患者

Enrolment Procedure

Supplemental Digital Content Figure 1 Enrolment Procedure

Daily Screening of all ICU patients gt18 yrs of age Intubated gt36 hours

Exclude if Exclusion Criteria

Successful SBT awaiting extubation Withdrawal of life support considered High spinal cord injury Neurosurgical patient Neuromuscular disease

Follow using daily eligibility checklist until meet all Clinical Stability Criteria to trial mode of partial ventilatory support (PAV or PSV)

partialcomplete reversal of cause of respiratory failure

body temperature between 36ordmC and 39ordmC metabolic disorders corrected pH gt732 hemoglobin gt70 gL no ongoing bleeding stable hemodynamic status Able to trigger PaO2 ge 60 mmHg or SaO2 ge 90 on FiO2le

060 and PEEP le 15 cm H2O ACVC Pplatle30cmH2O with Vt 6-8 mlkg

OR ACPC PC+PEEPle30cmH2O

Exclude if Exclusion Criteria Has met all Criteria for Switching to

Spontaneous Mode of Ventilation gt24 hours

Pre-inclusion Test on PSV 15 cmH2O for 30 minutes

Patient fails Pre-inclusion test Return to AC ventilation and reassess within 24 hour

Patient passes Pre-inclusion test Assess for Criteria to Initiate Weaning

PaO2FiO2gt200 on FiO2 le050 and PEEP le8 cmH2O No inotropesvasopressors except low dose dopamine

RSBI lt105 Patient passes Pre-SBT readiness assessment Proceed to SBT Place patient on CPAP 5 cmH2O and FiO2 040 for 30 minutes

Patient does not meet Criteria to Initiate Weaning

Patient meets Criteria to Initiate Weaning Proceed to Pre-SBT readiness assessment Place patient on CPAP for 1-2 min measure RSBI

RSBIge 105 Patient fails Pre-SBT readiness assessment

Patient passes SBT Patient fails SBT

Exclude if Exclusion Criteria

Passed SBT and tolerating PSVle7cmH2O and PEEPle5 cmH2O while awaiting extubation

Obtain informed consent from Patient or SDM as soon as available (deferred consent permitted)

Randomize

Exclude if Exclusion Criteria

Patient SDM refuse consent

bull 呼吸不全の原因に改善が認められてきている

bull 体温は36〜39度であるbull pH>732であるbull Hb>7gdLで現在出なしbull 動態が安定しているbull トリガーできるbull FiO2≦06 PEEP≦15cmH2Oで

PaO2≧60mmHg or SaO2≧90bull ACVCでVt 6-8mlkgで

Pplat≦30cmH2OもしくはACPCでPC+PEEP≦30cmH2O

上記をすべて満たせばプレテストとしてPSV 15cmH2Oに変更

Enrolment Procedure

Supplemental Digital Content Figure 1 Enrolment Procedure

Daily Screening of all ICU patients gt18 yrs of age Intubated gt36 hours

Exclude if Exclusion Criteria

Successful SBT awaiting extubation Withdrawal of life support considered High spinal cord injury Neurosurgical patient Neuromuscular disease

Follow using daily eligibility checklist until meet all Clinical Stability Criteria to trial mode of partial ventilatory support (PAV or PSV)

partialcomplete reversal of cause of respiratory failure

body temperature between 36ordmC and 39ordmC metabolic disorders corrected pH gt732 hemoglobin gt70 gL no ongoing bleeding stable hemodynamic status Able to trigger PaO2 ge 60 mmHg or SaO2 ge 90 on FiO2le

060 and PEEP le 15 cm H2O ACVC Pplatle30cmH2O with Vt 6-8 mlkg

OR ACPC PC+PEEPle30cmH2O

Exclude if Exclusion Criteria Has met all Criteria for Switching to

Spontaneous Mode of Ventilation gt24 hours

Pre-inclusion Test on PSV 15 cmH2O for 30 minutes

Patient fails Pre-inclusion test Return to AC ventilation and reassess within 24 hour

Patient passes Pre-inclusion test Assess for Criteria to Initiate Weaning

PaO2FiO2gt200 on FiO2 le050 and PEEP le8 cmH2O No inotropesvasopressors except low dose dopamine

RSBI lt105 Patient passes Pre-SBT readiness assessment Proceed to SBT Place patient on CPAP 5 cmH2O and FiO2 040 for 30 minutes

Patient does not meet Criteria to Initiate Weaning

Patient meets Criteria to Initiate Weaning Proceed to Pre-SBT readiness assessment Place patient on CPAP for 1-2 min measure RSBI

RSBIge 105 Patient fails Pre-SBT readiness assessment

Patient passes SBT Patient fails SBT

Exclude if Exclusion Criteria

Passed SBT and tolerating PSVle7cmH2O and PEEPle5 cmH2O while awaiting extubation

Obtain informed consent from Patient or SDM as soon as available (deferred consent permitted)

Randomize

Exclude if Exclusion Criteria

Patient SDM refuse consent

呼吸窮迫がなければプレテストクリアbull FiO2≦05 PEEP≦8cmH2OでPF≧200bull 強薬昇圧剤の使なしであればSBTへ

まず1-2分CPAP 5cmH2OにしてRSBIを測定

RSBI<105であればSBTへPSV 5cmH2Oにして30-120分

酸素化の条件を満たさないRSBI≧105SBT失敗であれば割り付けへ

Ventilation Protocols bull 呼吸器はPuritan Bennett 840 を使bull 後述するPAVプロトコールPSVプロトコー

ルに従って呼吸療法(RRT)が設定を変更(respiratory distress 呼吸窮迫がないようにできるだけサポートを下げていく)

bull 設定したmaxのサポートが必要になった場合はACに戻す

bull PAV群はPAV+とACPSV群はPSVとACしか使しない

Ventilation Protocols bull RCTの開始前に15名の患者をいて観察

研究をいRRTにプロトコルの確認をってもらった

bull プロトコルはすぐにられるようにラミネートしてベッドサイドに置いた

bull RRTは毎SBTをえるかどうか後述するチェックリストをいて確認除外項を満たさない場合は毎RSBIのアセスメントとSBTをう

Respiratory Distress bull 呼吸数>35回minbull 脈拍>140min もしくはベースラインより20以上の

上昇bull 収縮期圧>180mmHg もしくは<90mmHg ベース

ラインより30以上の変化bull 不安bull 発汗bull 副呼吸筋の使bull 奇異呼吸bull 呼吸困難の訴え

PAV protocolSupplemental Digital Content Figure 2a Proportional Assist Ventilation Protocol

bull 開始の設定はサポート率70

bull 2-3時間おきにRRTがアセスメントし問題なければ10-20ずつPAVサポートを下げる

サポート率90まで上げてもダメな場合はACに戻す

PSV protocolSupplemental Digital Content Figure 2b Pressure Support Ventilation Protocol

bull 開始の設定はPS 15cmH2O

bull まずRRlt35 VTgt5mLkgで耐えられるPSまで下げる

bull 2-3時間おきにRRTがアセスメントしてPS2-3cmH2Pずつ下げる

Data collection form for RRTsSupplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

SBTの開始基準bull PF≧200bull FiO2≦05bull pH≧732bull トリガーできるbull 24時間以内に筋虚イベントなしbull 強薬昇圧剤の使なし

Data collection form for RRTs

Supplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

CPAPにしてRSBIを測定RSBI<105であればSBT施

PAV群はサポート率20PSV群はPS 5cmH2OでSBTを施

PEEP≦8cmH2OでSaO2≧90呼吸窮迫なしであればSBTクリア

抜管へ

Endpoints 呼吸器離脱までの時間ICU病院滞在期間

呼吸器離脱成功の定義抜管後48時間再挿管なし侵襲的呼吸(NIV)の使もなし気管切開使時は呼吸器から最終的に外れた時点1のうち12時間以上NIVを使していない

Statistical Analysis bull Pilot studyのためサンプルサイズの計

算はしていないbull P値<005を統計学的有意としたbull Intention to treat approachをった

RESULTS

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial

エンロール期間は42か80名が基準を満たしが最終的に割付されたのは54名

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial最終的に解析がわれたのはPAV群27名PSV群23名

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

ベースラインは両群間で変わりなし重症度はAPACHEⅡ 27点程度患者は肺炎が3割もともとCOPDを持っている患者は2割程度

Feasibility Outcomes bull 呼吸窮迫でRRTが呼吸器サポートを上昇

させた回数はPAV群で76回232daysPSV群で124回295daysと有意にPSV群で多かった(p=0002)

bull ACに戻した回数はPAV群で66回PSV群で90回と変わりなし(p=061)

bull 全くACに戻さなかった患者はPAV群で627名PSV群で623名(p=10)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

ベースラインは両群間で変わりなしPAVにすることでΔP圧Peak圧が低く吸気時間呼吸時間がくなるPSV群と較するとPAV群のがPaoAreaPaoAreaVE(呼吸器のサポートの強さをす)が低くVTPaoAreaがくsecond post-randomizationの解析においてΔP圧Peak圧が低い

Asynchrony index

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 7

B635) documenting compliance with the daily screening for SBT noncompletion of either the CPAP screening test or SBT occurred on 5 of 292 of patient study days (17) and this violation involved two patients in each study arm The second audit examining for nine different potential deviations found complete adherence to study protocol on 254 of 292 of patient study days (900) There were no adverse events (pneumo-thorax arrhythmia or death) resulting from any of the study interventions

Physiologic Response to Ventilation ProtocolsRecordings of flow and airway pressure were available for analysis for all 50 patients prerandomization 49 patients immediately (median 1 interquartile range [IQR] 0ndash15 hr) postrandomization and for 26 patients during subsequent days (median 27 IQR 20ndash70-hr postrandomization) Approximately 15000 breaths in 125 recordings were analyzed Table 2 compares the respiratory parameters in both groups pre- and postrandomization There were no statistically sig-nificant differences between the PAV and PSV groups at base-line prerandomization (p values not shown) In the PAV group there was a significant decrease in peak airway pressure delta pressure (measured as peak airway opening pressure minus the PEEP) and area under the inspiratory pressure curve immediately postrandomization and an increase in the per-cent of total cycle time spent in inspiration (TiTtot) whereas there were no significant differences immediately postran-domization in the PSV group The second measurement of postrandomization respiratory parameters revealed signifi-cant decreases in peak airway pressure and delta pressure in both PAV and PSV groups Adjusting for baseline measure-ments as a covariate the PAV group received 7 cm H2

O∙s less ventilator assistance (PaoArea) per liter of minute ventilation (p lt 00001) and increased their VT by 115 mLcm H

2O∙s

relative to the PSV group (p lt 0001)PAV significantly reduced AI postrandomization (Fig 2A)

In total nine patients (18) had a high AI (gt 10 of breaths) on PSV or on AC mode at baseline No patients had a high AI on PAV In the PSV group six patients (26) had a high AI

at baseline four (17) immediately postrandomization and one patient (7) after further adjustment of ventilator settings according to the PSV protocol (Fig 2B) Adjusting for baseline differences patients in the PAV group tended to have a lower AI than patients in the PSV group immediately postrandom-ization but this did not reach statistical significance (adjusted difference ndash45 [95 CI ndash95 to 04] p = 007)

Clinical OutcomesTable 3 shows the results for the main clinical endpoints Weaning duration and time to successful SBT did not dif-fer significantly among survivors between the groups Among ICU survivors median time to ICU discharge was shorter in the group randomized to PAV 73 days (IQR 52ndash114 d) ver-sus 124 days (IQR 75ndash308 d) in the PSV group (p = 003) Kaplan-Meier plots of the proportion of at-risk patients achiev-ing successful extubation and live ICU discharge are shown in Figure 3 A and B No difference was found for hospital length of stay ICU mortality or hospital mortality Four patients (15) in the PAV group and three patients (13) in the PSV group died in ICU (respiratory rate [RR] of PAV 114 95 CI 028 to 46 p = 086) Overall in-hospital mortality was nine patients (33) in the PAV group and six patients (26) in the PSV group (RR of PAV 128 95 CI 054 to 31 p = 058)

Cointerventions Sedating MedicationsTable 4 compares the use of medications that may prolong weaning (by causing sedation or neuromuscular weakness) in the two groups pre- and postrandomization Type and dose of sedative medication were highly variable among individual patients and therefore we analyzed pre- and postrandomiza-tion changes to determine the impact of the ventilation pro-tocol on sedative use within each patient In the PAV group benzodiazepine narcotic and propofol use measured as both cumulative dose and average daily dose all decreased signifi-cantly postrandomization as did neuromuscular blocker use In the PSV group average daily dose of lorazepam equiva-lent and average daily dose of propofol decreased signifi-cantly postrandomization whereas the proportion of patients

Figure 2 A Asynchrony index pre- and postrandomization on proportional assist ventilation B Asynchrony index pre- and postrandomization on pressure support ventilationPAV群 PSV群

Asynchrony Index () = 同調イベントの回数全呼吸回数 times 100

IntensiveCareMed(2006)321515ndash1522

AI>10は3名いたがPAVに設定変更後0名となった

AI>10は6名いたがPSVに設定変更後4名となった

Clinical Outcomes

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

成功したSBTまでの数抜管成功までの数に変わりなしICU退室までの数はPAV群で有意に短い

Kaplan-Meier curves

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

抜管成功までの時間 ICU退室までの時間

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 9

TABLE 4 Comparison of Medications That May Alter Weaning Duration Pre- and Postrandomization

Variables

PAV Group (n = 27)

PSV Group (n = 23) Postrandomization

Pre randomization Pos trandomization Pre randomization Pos trandomizationAdjusted Difference PSV ndash PAV (95 CI)

Lorazepam equivalents

Cumulative dose mg

308 (207) 48 (257)a 137 (440) 54 (258) 237 (ndash580 to 1055)

Average daily dose mg

44 (255) 09 (43)a 23 (85) 06 (40)a 24 (ndash62 to 110)

Single-day dose mgb

08 (72) 12 (50) 16 (70) 10 (64) 32 (ndash13 to 78)

Morphine equivalents

Cumulative dose mg

5003 (6188) 1140 (3140)b 1710 (13310) 1610 (4860) 591 (ndash2515 to 3696)

Average daily dose mg

858 (1638) 170 (578)c 428 (1724) 136 (320)d 108 (ndash316 to 532)

Single-day dose mga

180 (1240) 187 (77) 100 (1920) 50 (630)d 334 (ndash026 to 670)d

Propofol

Cumulative dose mg

7400 (55500) 0 (700)a 29450 (85800) 0 (27660) 226 (ndash42156 to 42607)

Average daily dose mg

529 (11690) 0 (53)a 4879 (15980) 0 (2738)a 651 (ndash3211 to 4514)

Single-day dose mga

0 (0) 0 (0) 0 (1000) 0 (0) 344 (ndash3781 to 4469)

Neuroleptics

Days receiving neuroleptics n

0 (2) 0 (4) 0 (0) 1 (4)a ndash16 (ndash34 to 03)

Proportion of patients who received a neuroleptic

407 407 174 565b

NMBs

Days receiving NMBs n

1 (1) 0 (0)a 0 (1) 0 (0) ndash01 (ndash05 to 03)

Proportion of patients who received an NMB

518 148a 348 174

Corticosteroids

Days receiving corticosteroids n

0 (0) 0 (0) 0 (2) 0 (2) 02 (ndash06 to 10)

Proportion of patients who received corticosteroids

185 148 348 261

PAV = proportional assist ventilation PSV = pressure support ventilation NMB = neuromuscular blockera p lt 005b Single-day dose = total dose given day prior to randomization or total dose give day after randomizationc p lt 0005d p lt 00005e p le 007All data points except percentage of patients receiving neuroleptics NMBs or corticosteroids are presented as median (interquartile range)

鎮静剤鎮痛剤についてはPAV群とPSV群で変わりなし

その他bull 不整脈気胸プロトコル施中の死亡

は両群とも起こらなかった

筆者らの考察bull PSVはweaningに最も使われているモードだ

がそれと較しても遜ないweaningができることがわかった

bull アシストが少ない傾向であったがPAV群のが呼吸窮迫となる回数が少なかった

rarr同調性が良くなった結果の可能性

ICU退室までの期間はPAV群で短いという結果であったがpowerが少ない

筆者らの考察bull 単施設での結果であり他の施設にこの

新しいプロトコルそのものが受けれられるかは不明である

bull SBTに失敗した患者を主に対象としておりICU患者すべてに適応はできない

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる

Page 11: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

設定するのはサポート率のみ

患者の呼吸仕事量(WOBpt)が03〜08JLになるようにサポート率を設定する

PAVのメリットデメリット

bull メリット患者-呼吸器の同調性が良い呼吸筋を適度に使うことができる

bull デメリット患者の呼吸努が弱い場合はサポートが減ってしまうリークがある場合過度にサポート圧が上がってしまう可能性(runaway現象)

PAVのエビデンスbull PAVとPSVによるウィーニングをい48

時間観察したRCTではPAVのが調節呼吸に戻る割合が低かった(PAV 11 vsPSV 22 p=004)

bull 夜間PAVにすることで睡眠の質が改善したIntensiveCareMed2008342026ndash2034

Crit CareMed2007351048ndash1054

PAVは呼吸器離脱困難患者のウィーニングにおいて

優れたモードか

本の論

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Critical Care Medicine wwwccmjournalorg 1

Objectives Despite protocols incorporating spontaneous breath-ing trials 31 of ICU patients experience difficult or prolonged weaning from mechanical ventilation Nonfatiguing modes such as pressure support ventilation are recommended Proportional assist ventilation provides assistance in proportion to patient effort which may optimize weaning However it is not known how proportional assist ventilation performs relative to pressure sup-port ventilation over a prolonged period in the complex ICU set-ting The purpose of this study was to compare the physiologic and clinical performance (failure rate) safety and feasibility of protocols using daily spontaneous breathing trial plus pressure support ventilation versus proportional assist ventilation until ven-tilation discontinuationDesign Single-center unblinded pilot randomized controlled trialSetting Medical-surgical ICU of a tertiary-care hospitalPatients Adult patients intubated greater than 36 hours were ran-domized if they met eligibility criteria for partial ventilatory support

tolerated pressure support ventilation greater than or equal to 30 minutes and either failed or did not meet criteria for a sponta-neous breathing trialInterventions Patients were randomized to the pressure support ventilation or proportional assist ventilation protocol (PAV+ Puri-tan Bennett 840 Covidien Boulder CO) Both protocols used progressive decreases in level of assistance as tolerated coupled with daily assessment for spontaneous breathing trialsMeasurements and Main Results Of 54 patients randomized outcome data are available for 50 patients 27 were randomized to receive proportional assist ventilation and 23 to receive pres-sure support ventilation There were no adverse events linked to the study interventions and protocol violations were infrequent Recruitment was slower than projected (13 patients per month) The median (interquartile range) time from randomization to suc-cessful extubation was 39 days (28ndash84 d) on proportional assist ventilation versus 49 days (29ndash263 d) on pressure support ventilation (p = 039) Time to live ICU discharge was 73 days (52ndash114 d) on proportional assist ventilation versus 124 days (75ndash308 d) on pressure support ventilation (p = 003)Conclusion This pilot study demonstrates the utility safety and feasibility of the weaning protocols and provides important infor-mation to guide the design of a future randomized controlled trial comparing weaning from mechanical ventilation on pressure sup-port ventilation versus proportional assist ventilation (Crit Care Med 2016 XX00ndash00)Key Words assisted mechanical ventilation asynchrony index proportional assist ventilation respiratory muscles ventilator weaning

Difficult or prolonged weaning from mechanical ventilation (MV) affects 31 (range 26ndash42) of mechanically ventilated patients (1) and increases

patient morbidity mortality ICU length of stay and cost to the healthcare system (2) When initial weaning attempts fail to liberate patients from MV ideally clinicians should choose

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights ReservedDOI 101097CCM0000000000001600

1London Health Sciences Centre London ON Canada2Department of Medicine The University of Western Ontario London ON Canada

This study was performed at London Health Sciences Centre University Hospital London CanadaSupplemental digital content is available for this article Direct URL cita-tions appear in the printed text and are provided in the HTML and PDF versions of this article on the journalrsquos website (httpjournalslwwcomccmjournal)Supported in part by Critical Care Western Department of Medicine The University of Western Ontario Program of Experimental Medi-cine (POEM) Department of Medicine University of Western Ontario Lawson Health Research Institute London Health Sciences Centre London Canada and Covidien Boulder USADr Bosma has received honourarium for lectures and travel reimburse-ments from Covidien Dr Bosmarsquos institution has received grant support from Covidien for Dr Bosmarsquos research studies The remaining authors have disclosed that they do not have any potential conflicts of interestFor information regarding this article E-mail KarenJBosmalhsconca

A Pilot Randomized Trial Comparing Weaning From Mechanical Ventilation on Pressure Support Versus Proportional Assist Ventilation

Karen J Bosma MD FRCPC12 Brooke A Read MHS BSc RRT1 Mohammad J Bahrgard Nikoo MD FRCPC12 Philip M Jones MD FRCPC MSc12 Fran A Priestap MSc1 James F Lewis MD FRCPC12

Crit Care Med 2016 Jan 20[Epub ahead of print]

論のPICO

P 呼吸器離脱困難な患者(simple weaningではない)

I PAVを使

C PSVを使

O 呼吸器離脱までの時間

Study Designbull 単施設盲検前向きランダム化較

試験(pilot study)bull カナダロンドンにあるUniversity

Hospital-London Health Science Centre の内科外科ICU(20床)

bull ランダム割付はcomputer-generated random number sequence によってわれ隠蔽化されている

PatientsInclusion criteria

bull 18歳以上の挿管患者bull 36時間以上呼吸器管理をされている

PatientsExclusion criteria

1) 48時間以内にwithdrawが考慮される患者2) 位の脊髄障害進性の神経筋疾患の

よって慢性的な呼吸器依存となる可能性があるまたは脳外科的介が必要である患者

Enrolment Procedure

Supplemental Digital Content Figure 1 Enrolment Procedure

Daily Screening of all ICU patients gt18 yrs of age Intubated gt36 hours

Exclude if Exclusion Criteria

Successful SBT awaiting extubation Withdrawal of life support considered High spinal cord injury Neurosurgical patient Neuromuscular disease

Follow using daily eligibility checklist until meet all Clinical Stability Criteria to trial mode of partial ventilatory support (PAV or PSV)

partialcomplete reversal of cause of respiratory failure

body temperature between 36ordmC and 39ordmC metabolic disorders corrected pH gt732 hemoglobin gt70 gL no ongoing bleeding stable hemodynamic status Able to trigger PaO2 ge 60 mmHg or SaO2 ge 90 on FiO2le

060 and PEEP le 15 cm H2O ACVC Pplatle30cmH2O with Vt 6-8 mlkg

OR ACPC PC+PEEPle30cmH2O

Exclude if Exclusion Criteria Has met all Criteria for Switching to

Spontaneous Mode of Ventilation gt24 hours

Pre-inclusion Test on PSV 15 cmH2O for 30 minutes

Patient fails Pre-inclusion test Return to AC ventilation and reassess within 24 hour

Patient passes Pre-inclusion test Assess for Criteria to Initiate Weaning

PaO2FiO2gt200 on FiO2 le050 and PEEP le8 cmH2O No inotropesvasopressors except low dose dopamine

RSBI lt105 Patient passes Pre-SBT readiness assessment Proceed to SBT Place patient on CPAP 5 cmH2O and FiO2 040 for 30 minutes

Patient does not meet Criteria to Initiate Weaning

Patient meets Criteria to Initiate Weaning Proceed to Pre-SBT readiness assessment Place patient on CPAP for 1-2 min measure RSBI

RSBIge 105 Patient fails Pre-SBT readiness assessment

Patient passes SBT Patient fails SBT

Exclude if Exclusion Criteria

Passed SBT and tolerating PSVle7cmH2O and PEEPle5 cmH2O while awaiting extubation

Obtain informed consent from Patient or SDM as soon as available (deferred consent permitted)

Randomize

Exclude if Exclusion Criteria

Patient SDM refuse consent

bull 呼吸不全の原因に改善が認められてきている

bull 体温は36〜39度であるbull pH>732であるbull Hb>7gdLで現在出なしbull 動態が安定しているbull トリガーできるbull FiO2≦06 PEEP≦15cmH2Oで

PaO2≧60mmHg or SaO2≧90bull ACVCでVt 6-8mlkgで

Pplat≦30cmH2OもしくはACPCでPC+PEEP≦30cmH2O

上記をすべて満たせばプレテストとしてPSV 15cmH2Oに変更

Enrolment Procedure

Supplemental Digital Content Figure 1 Enrolment Procedure

Daily Screening of all ICU patients gt18 yrs of age Intubated gt36 hours

Exclude if Exclusion Criteria

Successful SBT awaiting extubation Withdrawal of life support considered High spinal cord injury Neurosurgical patient Neuromuscular disease

Follow using daily eligibility checklist until meet all Clinical Stability Criteria to trial mode of partial ventilatory support (PAV or PSV)

partialcomplete reversal of cause of respiratory failure

body temperature between 36ordmC and 39ordmC metabolic disorders corrected pH gt732 hemoglobin gt70 gL no ongoing bleeding stable hemodynamic status Able to trigger PaO2 ge 60 mmHg or SaO2 ge 90 on FiO2le

060 and PEEP le 15 cm H2O ACVC Pplatle30cmH2O with Vt 6-8 mlkg

OR ACPC PC+PEEPle30cmH2O

Exclude if Exclusion Criteria Has met all Criteria for Switching to

Spontaneous Mode of Ventilation gt24 hours

Pre-inclusion Test on PSV 15 cmH2O for 30 minutes

Patient fails Pre-inclusion test Return to AC ventilation and reassess within 24 hour

Patient passes Pre-inclusion test Assess for Criteria to Initiate Weaning

PaO2FiO2gt200 on FiO2 le050 and PEEP le8 cmH2O No inotropesvasopressors except low dose dopamine

RSBI lt105 Patient passes Pre-SBT readiness assessment Proceed to SBT Place patient on CPAP 5 cmH2O and FiO2 040 for 30 minutes

Patient does not meet Criteria to Initiate Weaning

Patient meets Criteria to Initiate Weaning Proceed to Pre-SBT readiness assessment Place patient on CPAP for 1-2 min measure RSBI

RSBIge 105 Patient fails Pre-SBT readiness assessment

Patient passes SBT Patient fails SBT

Exclude if Exclusion Criteria

Passed SBT and tolerating PSVle7cmH2O and PEEPle5 cmH2O while awaiting extubation

Obtain informed consent from Patient or SDM as soon as available (deferred consent permitted)

Randomize

Exclude if Exclusion Criteria

Patient SDM refuse consent

呼吸窮迫がなければプレテストクリアbull FiO2≦05 PEEP≦8cmH2OでPF≧200bull 強薬昇圧剤の使なしであればSBTへ

まず1-2分CPAP 5cmH2OにしてRSBIを測定

RSBI<105であればSBTへPSV 5cmH2Oにして30-120分

酸素化の条件を満たさないRSBI≧105SBT失敗であれば割り付けへ

Ventilation Protocols bull 呼吸器はPuritan Bennett 840 を使bull 後述するPAVプロトコールPSVプロトコー

ルに従って呼吸療法(RRT)が設定を変更(respiratory distress 呼吸窮迫がないようにできるだけサポートを下げていく)

bull 設定したmaxのサポートが必要になった場合はACに戻す

bull PAV群はPAV+とACPSV群はPSVとACしか使しない

Ventilation Protocols bull RCTの開始前に15名の患者をいて観察

研究をいRRTにプロトコルの確認をってもらった

bull プロトコルはすぐにられるようにラミネートしてベッドサイドに置いた

bull RRTは毎SBTをえるかどうか後述するチェックリストをいて確認除外項を満たさない場合は毎RSBIのアセスメントとSBTをう

Respiratory Distress bull 呼吸数>35回minbull 脈拍>140min もしくはベースラインより20以上の

上昇bull 収縮期圧>180mmHg もしくは<90mmHg ベース

ラインより30以上の変化bull 不安bull 発汗bull 副呼吸筋の使bull 奇異呼吸bull 呼吸困難の訴え

PAV protocolSupplemental Digital Content Figure 2a Proportional Assist Ventilation Protocol

bull 開始の設定はサポート率70

bull 2-3時間おきにRRTがアセスメントし問題なければ10-20ずつPAVサポートを下げる

サポート率90まで上げてもダメな場合はACに戻す

PSV protocolSupplemental Digital Content Figure 2b Pressure Support Ventilation Protocol

bull 開始の設定はPS 15cmH2O

bull まずRRlt35 VTgt5mLkgで耐えられるPSまで下げる

bull 2-3時間おきにRRTがアセスメントしてPS2-3cmH2Pずつ下げる

Data collection form for RRTsSupplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

SBTの開始基準bull PF≧200bull FiO2≦05bull pH≧732bull トリガーできるbull 24時間以内に筋虚イベントなしbull 強薬昇圧剤の使なし

Data collection form for RRTs

Supplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

CPAPにしてRSBIを測定RSBI<105であればSBT施

PAV群はサポート率20PSV群はPS 5cmH2OでSBTを施

PEEP≦8cmH2OでSaO2≧90呼吸窮迫なしであればSBTクリア

抜管へ

Endpoints 呼吸器離脱までの時間ICU病院滞在期間

呼吸器離脱成功の定義抜管後48時間再挿管なし侵襲的呼吸(NIV)の使もなし気管切開使時は呼吸器から最終的に外れた時点1のうち12時間以上NIVを使していない

Statistical Analysis bull Pilot studyのためサンプルサイズの計

算はしていないbull P値<005を統計学的有意としたbull Intention to treat approachをった

RESULTS

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial

エンロール期間は42か80名が基準を満たしが最終的に割付されたのは54名

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial最終的に解析がわれたのはPAV群27名PSV群23名

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

ベースラインは両群間で変わりなし重症度はAPACHEⅡ 27点程度患者は肺炎が3割もともとCOPDを持っている患者は2割程度

Feasibility Outcomes bull 呼吸窮迫でRRTが呼吸器サポートを上昇

させた回数はPAV群で76回232daysPSV群で124回295daysと有意にPSV群で多かった(p=0002)

bull ACに戻した回数はPAV群で66回PSV群で90回と変わりなし(p=061)

bull 全くACに戻さなかった患者はPAV群で627名PSV群で623名(p=10)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

ベースラインは両群間で変わりなしPAVにすることでΔP圧Peak圧が低く吸気時間呼吸時間がくなるPSV群と較するとPAV群のがPaoAreaPaoAreaVE(呼吸器のサポートの強さをす)が低くVTPaoAreaがくsecond post-randomizationの解析においてΔP圧Peak圧が低い

Asynchrony index

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 7

B635) documenting compliance with the daily screening for SBT noncompletion of either the CPAP screening test or SBT occurred on 5 of 292 of patient study days (17) and this violation involved two patients in each study arm The second audit examining for nine different potential deviations found complete adherence to study protocol on 254 of 292 of patient study days (900) There were no adverse events (pneumo-thorax arrhythmia or death) resulting from any of the study interventions

Physiologic Response to Ventilation ProtocolsRecordings of flow and airway pressure were available for analysis for all 50 patients prerandomization 49 patients immediately (median 1 interquartile range [IQR] 0ndash15 hr) postrandomization and for 26 patients during subsequent days (median 27 IQR 20ndash70-hr postrandomization) Approximately 15000 breaths in 125 recordings were analyzed Table 2 compares the respiratory parameters in both groups pre- and postrandomization There were no statistically sig-nificant differences between the PAV and PSV groups at base-line prerandomization (p values not shown) In the PAV group there was a significant decrease in peak airway pressure delta pressure (measured as peak airway opening pressure minus the PEEP) and area under the inspiratory pressure curve immediately postrandomization and an increase in the per-cent of total cycle time spent in inspiration (TiTtot) whereas there were no significant differences immediately postran-domization in the PSV group The second measurement of postrandomization respiratory parameters revealed signifi-cant decreases in peak airway pressure and delta pressure in both PAV and PSV groups Adjusting for baseline measure-ments as a covariate the PAV group received 7 cm H2

O∙s less ventilator assistance (PaoArea) per liter of minute ventilation (p lt 00001) and increased their VT by 115 mLcm H

2O∙s

relative to the PSV group (p lt 0001)PAV significantly reduced AI postrandomization (Fig 2A)

In total nine patients (18) had a high AI (gt 10 of breaths) on PSV or on AC mode at baseline No patients had a high AI on PAV In the PSV group six patients (26) had a high AI

at baseline four (17) immediately postrandomization and one patient (7) after further adjustment of ventilator settings according to the PSV protocol (Fig 2B) Adjusting for baseline differences patients in the PAV group tended to have a lower AI than patients in the PSV group immediately postrandom-ization but this did not reach statistical significance (adjusted difference ndash45 [95 CI ndash95 to 04] p = 007)

Clinical OutcomesTable 3 shows the results for the main clinical endpoints Weaning duration and time to successful SBT did not dif-fer significantly among survivors between the groups Among ICU survivors median time to ICU discharge was shorter in the group randomized to PAV 73 days (IQR 52ndash114 d) ver-sus 124 days (IQR 75ndash308 d) in the PSV group (p = 003) Kaplan-Meier plots of the proportion of at-risk patients achiev-ing successful extubation and live ICU discharge are shown in Figure 3 A and B No difference was found for hospital length of stay ICU mortality or hospital mortality Four patients (15) in the PAV group and three patients (13) in the PSV group died in ICU (respiratory rate [RR] of PAV 114 95 CI 028 to 46 p = 086) Overall in-hospital mortality was nine patients (33) in the PAV group and six patients (26) in the PSV group (RR of PAV 128 95 CI 054 to 31 p = 058)

Cointerventions Sedating MedicationsTable 4 compares the use of medications that may prolong weaning (by causing sedation or neuromuscular weakness) in the two groups pre- and postrandomization Type and dose of sedative medication were highly variable among individual patients and therefore we analyzed pre- and postrandomiza-tion changes to determine the impact of the ventilation pro-tocol on sedative use within each patient In the PAV group benzodiazepine narcotic and propofol use measured as both cumulative dose and average daily dose all decreased signifi-cantly postrandomization as did neuromuscular blocker use In the PSV group average daily dose of lorazepam equiva-lent and average daily dose of propofol decreased signifi-cantly postrandomization whereas the proportion of patients

Figure 2 A Asynchrony index pre- and postrandomization on proportional assist ventilation B Asynchrony index pre- and postrandomization on pressure support ventilationPAV群 PSV群

Asynchrony Index () = 同調イベントの回数全呼吸回数 times 100

IntensiveCareMed(2006)321515ndash1522

AI>10は3名いたがPAVに設定変更後0名となった

AI>10は6名いたがPSVに設定変更後4名となった

Clinical Outcomes

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

成功したSBTまでの数抜管成功までの数に変わりなしICU退室までの数はPAV群で有意に短い

Kaplan-Meier curves

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

抜管成功までの時間 ICU退室までの時間

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 9

TABLE 4 Comparison of Medications That May Alter Weaning Duration Pre- and Postrandomization

Variables

PAV Group (n = 27)

PSV Group (n = 23) Postrandomization

Pre randomization Pos trandomization Pre randomization Pos trandomizationAdjusted Difference PSV ndash PAV (95 CI)

Lorazepam equivalents

Cumulative dose mg

308 (207) 48 (257)a 137 (440) 54 (258) 237 (ndash580 to 1055)

Average daily dose mg

44 (255) 09 (43)a 23 (85) 06 (40)a 24 (ndash62 to 110)

Single-day dose mgb

08 (72) 12 (50) 16 (70) 10 (64) 32 (ndash13 to 78)

Morphine equivalents

Cumulative dose mg

5003 (6188) 1140 (3140)b 1710 (13310) 1610 (4860) 591 (ndash2515 to 3696)

Average daily dose mg

858 (1638) 170 (578)c 428 (1724) 136 (320)d 108 (ndash316 to 532)

Single-day dose mga

180 (1240) 187 (77) 100 (1920) 50 (630)d 334 (ndash026 to 670)d

Propofol

Cumulative dose mg

7400 (55500) 0 (700)a 29450 (85800) 0 (27660) 226 (ndash42156 to 42607)

Average daily dose mg

529 (11690) 0 (53)a 4879 (15980) 0 (2738)a 651 (ndash3211 to 4514)

Single-day dose mga

0 (0) 0 (0) 0 (1000) 0 (0) 344 (ndash3781 to 4469)

Neuroleptics

Days receiving neuroleptics n

0 (2) 0 (4) 0 (0) 1 (4)a ndash16 (ndash34 to 03)

Proportion of patients who received a neuroleptic

407 407 174 565b

NMBs

Days receiving NMBs n

1 (1) 0 (0)a 0 (1) 0 (0) ndash01 (ndash05 to 03)

Proportion of patients who received an NMB

518 148a 348 174

Corticosteroids

Days receiving corticosteroids n

0 (0) 0 (0) 0 (2) 0 (2) 02 (ndash06 to 10)

Proportion of patients who received corticosteroids

185 148 348 261

PAV = proportional assist ventilation PSV = pressure support ventilation NMB = neuromuscular blockera p lt 005b Single-day dose = total dose given day prior to randomization or total dose give day after randomizationc p lt 0005d p lt 00005e p le 007All data points except percentage of patients receiving neuroleptics NMBs or corticosteroids are presented as median (interquartile range)

鎮静剤鎮痛剤についてはPAV群とPSV群で変わりなし

その他bull 不整脈気胸プロトコル施中の死亡

は両群とも起こらなかった

筆者らの考察bull PSVはweaningに最も使われているモードだ

がそれと較しても遜ないweaningができることがわかった

bull アシストが少ない傾向であったがPAV群のが呼吸窮迫となる回数が少なかった

rarr同調性が良くなった結果の可能性

ICU退室までの期間はPAV群で短いという結果であったがpowerが少ない

筆者らの考察bull 単施設での結果であり他の施設にこの

新しいプロトコルそのものが受けれられるかは不明である

bull SBTに失敗した患者を主に対象としておりICU患者すべてに適応はできない

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる

Page 12: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

患者の呼吸仕事量(WOBpt)が03〜08JLになるようにサポート率を設定する

PAVのメリットデメリット

bull メリット患者-呼吸器の同調性が良い呼吸筋を適度に使うことができる

bull デメリット患者の呼吸努が弱い場合はサポートが減ってしまうリークがある場合過度にサポート圧が上がってしまう可能性(runaway現象)

PAVのエビデンスbull PAVとPSVによるウィーニングをい48

時間観察したRCTではPAVのが調節呼吸に戻る割合が低かった(PAV 11 vsPSV 22 p=004)

bull 夜間PAVにすることで睡眠の質が改善したIntensiveCareMed2008342026ndash2034

Crit CareMed2007351048ndash1054

PAVは呼吸器離脱困難患者のウィーニングにおいて

優れたモードか

本の論

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Critical Care Medicine wwwccmjournalorg 1

Objectives Despite protocols incorporating spontaneous breath-ing trials 31 of ICU patients experience difficult or prolonged weaning from mechanical ventilation Nonfatiguing modes such as pressure support ventilation are recommended Proportional assist ventilation provides assistance in proportion to patient effort which may optimize weaning However it is not known how proportional assist ventilation performs relative to pressure sup-port ventilation over a prolonged period in the complex ICU set-ting The purpose of this study was to compare the physiologic and clinical performance (failure rate) safety and feasibility of protocols using daily spontaneous breathing trial plus pressure support ventilation versus proportional assist ventilation until ven-tilation discontinuationDesign Single-center unblinded pilot randomized controlled trialSetting Medical-surgical ICU of a tertiary-care hospitalPatients Adult patients intubated greater than 36 hours were ran-domized if they met eligibility criteria for partial ventilatory support

tolerated pressure support ventilation greater than or equal to 30 minutes and either failed or did not meet criteria for a sponta-neous breathing trialInterventions Patients were randomized to the pressure support ventilation or proportional assist ventilation protocol (PAV+ Puri-tan Bennett 840 Covidien Boulder CO) Both protocols used progressive decreases in level of assistance as tolerated coupled with daily assessment for spontaneous breathing trialsMeasurements and Main Results Of 54 patients randomized outcome data are available for 50 patients 27 were randomized to receive proportional assist ventilation and 23 to receive pres-sure support ventilation There were no adverse events linked to the study interventions and protocol violations were infrequent Recruitment was slower than projected (13 patients per month) The median (interquartile range) time from randomization to suc-cessful extubation was 39 days (28ndash84 d) on proportional assist ventilation versus 49 days (29ndash263 d) on pressure support ventilation (p = 039) Time to live ICU discharge was 73 days (52ndash114 d) on proportional assist ventilation versus 124 days (75ndash308 d) on pressure support ventilation (p = 003)Conclusion This pilot study demonstrates the utility safety and feasibility of the weaning protocols and provides important infor-mation to guide the design of a future randomized controlled trial comparing weaning from mechanical ventilation on pressure sup-port ventilation versus proportional assist ventilation (Crit Care Med 2016 XX00ndash00)Key Words assisted mechanical ventilation asynchrony index proportional assist ventilation respiratory muscles ventilator weaning

Difficult or prolonged weaning from mechanical ventilation (MV) affects 31 (range 26ndash42) of mechanically ventilated patients (1) and increases

patient morbidity mortality ICU length of stay and cost to the healthcare system (2) When initial weaning attempts fail to liberate patients from MV ideally clinicians should choose

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights ReservedDOI 101097CCM0000000000001600

1London Health Sciences Centre London ON Canada2Department of Medicine The University of Western Ontario London ON Canada

This study was performed at London Health Sciences Centre University Hospital London CanadaSupplemental digital content is available for this article Direct URL cita-tions appear in the printed text and are provided in the HTML and PDF versions of this article on the journalrsquos website (httpjournalslwwcomccmjournal)Supported in part by Critical Care Western Department of Medicine The University of Western Ontario Program of Experimental Medi-cine (POEM) Department of Medicine University of Western Ontario Lawson Health Research Institute London Health Sciences Centre London Canada and Covidien Boulder USADr Bosma has received honourarium for lectures and travel reimburse-ments from Covidien Dr Bosmarsquos institution has received grant support from Covidien for Dr Bosmarsquos research studies The remaining authors have disclosed that they do not have any potential conflicts of interestFor information regarding this article E-mail KarenJBosmalhsconca

A Pilot Randomized Trial Comparing Weaning From Mechanical Ventilation on Pressure Support Versus Proportional Assist Ventilation

Karen J Bosma MD FRCPC12 Brooke A Read MHS BSc RRT1 Mohammad J Bahrgard Nikoo MD FRCPC12 Philip M Jones MD FRCPC MSc12 Fran A Priestap MSc1 James F Lewis MD FRCPC12

Crit Care Med 2016 Jan 20[Epub ahead of print]

論のPICO

P 呼吸器離脱困難な患者(simple weaningではない)

I PAVを使

C PSVを使

O 呼吸器離脱までの時間

Study Designbull 単施設盲検前向きランダム化較

試験(pilot study)bull カナダロンドンにあるUniversity

Hospital-London Health Science Centre の内科外科ICU(20床)

bull ランダム割付はcomputer-generated random number sequence によってわれ隠蔽化されている

PatientsInclusion criteria

bull 18歳以上の挿管患者bull 36時間以上呼吸器管理をされている

PatientsExclusion criteria

1) 48時間以内にwithdrawが考慮される患者2) 位の脊髄障害進性の神経筋疾患の

よって慢性的な呼吸器依存となる可能性があるまたは脳外科的介が必要である患者

Enrolment Procedure

Supplemental Digital Content Figure 1 Enrolment Procedure

Daily Screening of all ICU patients gt18 yrs of age Intubated gt36 hours

Exclude if Exclusion Criteria

Successful SBT awaiting extubation Withdrawal of life support considered High spinal cord injury Neurosurgical patient Neuromuscular disease

Follow using daily eligibility checklist until meet all Clinical Stability Criteria to trial mode of partial ventilatory support (PAV or PSV)

partialcomplete reversal of cause of respiratory failure

body temperature between 36ordmC and 39ordmC metabolic disorders corrected pH gt732 hemoglobin gt70 gL no ongoing bleeding stable hemodynamic status Able to trigger PaO2 ge 60 mmHg or SaO2 ge 90 on FiO2le

060 and PEEP le 15 cm H2O ACVC Pplatle30cmH2O with Vt 6-8 mlkg

OR ACPC PC+PEEPle30cmH2O

Exclude if Exclusion Criteria Has met all Criteria for Switching to

Spontaneous Mode of Ventilation gt24 hours

Pre-inclusion Test on PSV 15 cmH2O for 30 minutes

Patient fails Pre-inclusion test Return to AC ventilation and reassess within 24 hour

Patient passes Pre-inclusion test Assess for Criteria to Initiate Weaning

PaO2FiO2gt200 on FiO2 le050 and PEEP le8 cmH2O No inotropesvasopressors except low dose dopamine

RSBI lt105 Patient passes Pre-SBT readiness assessment Proceed to SBT Place patient on CPAP 5 cmH2O and FiO2 040 for 30 minutes

Patient does not meet Criteria to Initiate Weaning

Patient meets Criteria to Initiate Weaning Proceed to Pre-SBT readiness assessment Place patient on CPAP for 1-2 min measure RSBI

RSBIge 105 Patient fails Pre-SBT readiness assessment

Patient passes SBT Patient fails SBT

Exclude if Exclusion Criteria

Passed SBT and tolerating PSVle7cmH2O and PEEPle5 cmH2O while awaiting extubation

Obtain informed consent from Patient or SDM as soon as available (deferred consent permitted)

Randomize

Exclude if Exclusion Criteria

Patient SDM refuse consent

bull 呼吸不全の原因に改善が認められてきている

bull 体温は36〜39度であるbull pH>732であるbull Hb>7gdLで現在出なしbull 動態が安定しているbull トリガーできるbull FiO2≦06 PEEP≦15cmH2Oで

PaO2≧60mmHg or SaO2≧90bull ACVCでVt 6-8mlkgで

Pplat≦30cmH2OもしくはACPCでPC+PEEP≦30cmH2O

上記をすべて満たせばプレテストとしてPSV 15cmH2Oに変更

Enrolment Procedure

Supplemental Digital Content Figure 1 Enrolment Procedure

Daily Screening of all ICU patients gt18 yrs of age Intubated gt36 hours

Exclude if Exclusion Criteria

Successful SBT awaiting extubation Withdrawal of life support considered High spinal cord injury Neurosurgical patient Neuromuscular disease

Follow using daily eligibility checklist until meet all Clinical Stability Criteria to trial mode of partial ventilatory support (PAV or PSV)

partialcomplete reversal of cause of respiratory failure

body temperature between 36ordmC and 39ordmC metabolic disorders corrected pH gt732 hemoglobin gt70 gL no ongoing bleeding stable hemodynamic status Able to trigger PaO2 ge 60 mmHg or SaO2 ge 90 on FiO2le

060 and PEEP le 15 cm H2O ACVC Pplatle30cmH2O with Vt 6-8 mlkg

OR ACPC PC+PEEPle30cmH2O

Exclude if Exclusion Criteria Has met all Criteria for Switching to

Spontaneous Mode of Ventilation gt24 hours

Pre-inclusion Test on PSV 15 cmH2O for 30 minutes

Patient fails Pre-inclusion test Return to AC ventilation and reassess within 24 hour

Patient passes Pre-inclusion test Assess for Criteria to Initiate Weaning

PaO2FiO2gt200 on FiO2 le050 and PEEP le8 cmH2O No inotropesvasopressors except low dose dopamine

RSBI lt105 Patient passes Pre-SBT readiness assessment Proceed to SBT Place patient on CPAP 5 cmH2O and FiO2 040 for 30 minutes

Patient does not meet Criteria to Initiate Weaning

Patient meets Criteria to Initiate Weaning Proceed to Pre-SBT readiness assessment Place patient on CPAP for 1-2 min measure RSBI

RSBIge 105 Patient fails Pre-SBT readiness assessment

Patient passes SBT Patient fails SBT

Exclude if Exclusion Criteria

Passed SBT and tolerating PSVle7cmH2O and PEEPle5 cmH2O while awaiting extubation

Obtain informed consent from Patient or SDM as soon as available (deferred consent permitted)

Randomize

Exclude if Exclusion Criteria

Patient SDM refuse consent

呼吸窮迫がなければプレテストクリアbull FiO2≦05 PEEP≦8cmH2OでPF≧200bull 強薬昇圧剤の使なしであればSBTへ

まず1-2分CPAP 5cmH2OにしてRSBIを測定

RSBI<105であればSBTへPSV 5cmH2Oにして30-120分

酸素化の条件を満たさないRSBI≧105SBT失敗であれば割り付けへ

Ventilation Protocols bull 呼吸器はPuritan Bennett 840 を使bull 後述するPAVプロトコールPSVプロトコー

ルに従って呼吸療法(RRT)が設定を変更(respiratory distress 呼吸窮迫がないようにできるだけサポートを下げていく)

bull 設定したmaxのサポートが必要になった場合はACに戻す

bull PAV群はPAV+とACPSV群はPSVとACしか使しない

Ventilation Protocols bull RCTの開始前に15名の患者をいて観察

研究をいRRTにプロトコルの確認をってもらった

bull プロトコルはすぐにられるようにラミネートしてベッドサイドに置いた

bull RRTは毎SBTをえるかどうか後述するチェックリストをいて確認除外項を満たさない場合は毎RSBIのアセスメントとSBTをう

Respiratory Distress bull 呼吸数>35回minbull 脈拍>140min もしくはベースラインより20以上の

上昇bull 収縮期圧>180mmHg もしくは<90mmHg ベース

ラインより30以上の変化bull 不安bull 発汗bull 副呼吸筋の使bull 奇異呼吸bull 呼吸困難の訴え

PAV protocolSupplemental Digital Content Figure 2a Proportional Assist Ventilation Protocol

bull 開始の設定はサポート率70

bull 2-3時間おきにRRTがアセスメントし問題なければ10-20ずつPAVサポートを下げる

サポート率90まで上げてもダメな場合はACに戻す

PSV protocolSupplemental Digital Content Figure 2b Pressure Support Ventilation Protocol

bull 開始の設定はPS 15cmH2O

bull まずRRlt35 VTgt5mLkgで耐えられるPSまで下げる

bull 2-3時間おきにRRTがアセスメントしてPS2-3cmH2Pずつ下げる

Data collection form for RRTsSupplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

SBTの開始基準bull PF≧200bull FiO2≦05bull pH≧732bull トリガーできるbull 24時間以内に筋虚イベントなしbull 強薬昇圧剤の使なし

Data collection form for RRTs

Supplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

CPAPにしてRSBIを測定RSBI<105であればSBT施

PAV群はサポート率20PSV群はPS 5cmH2OでSBTを施

PEEP≦8cmH2OでSaO2≧90呼吸窮迫なしであればSBTクリア

抜管へ

Endpoints 呼吸器離脱までの時間ICU病院滞在期間

呼吸器離脱成功の定義抜管後48時間再挿管なし侵襲的呼吸(NIV)の使もなし気管切開使時は呼吸器から最終的に外れた時点1のうち12時間以上NIVを使していない

Statistical Analysis bull Pilot studyのためサンプルサイズの計

算はしていないbull P値<005を統計学的有意としたbull Intention to treat approachをった

RESULTS

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial

エンロール期間は42か80名が基準を満たしが最終的に割付されたのは54名

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial最終的に解析がわれたのはPAV群27名PSV群23名

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

ベースラインは両群間で変わりなし重症度はAPACHEⅡ 27点程度患者は肺炎が3割もともとCOPDを持っている患者は2割程度

Feasibility Outcomes bull 呼吸窮迫でRRTが呼吸器サポートを上昇

させた回数はPAV群で76回232daysPSV群で124回295daysと有意にPSV群で多かった(p=0002)

bull ACに戻した回数はPAV群で66回PSV群で90回と変わりなし(p=061)

bull 全くACに戻さなかった患者はPAV群で627名PSV群で623名(p=10)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

ベースラインは両群間で変わりなしPAVにすることでΔP圧Peak圧が低く吸気時間呼吸時間がくなるPSV群と較するとPAV群のがPaoAreaPaoAreaVE(呼吸器のサポートの強さをす)が低くVTPaoAreaがくsecond post-randomizationの解析においてΔP圧Peak圧が低い

Asynchrony index

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 7

B635) documenting compliance with the daily screening for SBT noncompletion of either the CPAP screening test or SBT occurred on 5 of 292 of patient study days (17) and this violation involved two patients in each study arm The second audit examining for nine different potential deviations found complete adherence to study protocol on 254 of 292 of patient study days (900) There were no adverse events (pneumo-thorax arrhythmia or death) resulting from any of the study interventions

Physiologic Response to Ventilation ProtocolsRecordings of flow and airway pressure were available for analysis for all 50 patients prerandomization 49 patients immediately (median 1 interquartile range [IQR] 0ndash15 hr) postrandomization and for 26 patients during subsequent days (median 27 IQR 20ndash70-hr postrandomization) Approximately 15000 breaths in 125 recordings were analyzed Table 2 compares the respiratory parameters in both groups pre- and postrandomization There were no statistically sig-nificant differences between the PAV and PSV groups at base-line prerandomization (p values not shown) In the PAV group there was a significant decrease in peak airway pressure delta pressure (measured as peak airway opening pressure minus the PEEP) and area under the inspiratory pressure curve immediately postrandomization and an increase in the per-cent of total cycle time spent in inspiration (TiTtot) whereas there were no significant differences immediately postran-domization in the PSV group The second measurement of postrandomization respiratory parameters revealed signifi-cant decreases in peak airway pressure and delta pressure in both PAV and PSV groups Adjusting for baseline measure-ments as a covariate the PAV group received 7 cm H2

O∙s less ventilator assistance (PaoArea) per liter of minute ventilation (p lt 00001) and increased their VT by 115 mLcm H

2O∙s

relative to the PSV group (p lt 0001)PAV significantly reduced AI postrandomization (Fig 2A)

In total nine patients (18) had a high AI (gt 10 of breaths) on PSV or on AC mode at baseline No patients had a high AI on PAV In the PSV group six patients (26) had a high AI

at baseline four (17) immediately postrandomization and one patient (7) after further adjustment of ventilator settings according to the PSV protocol (Fig 2B) Adjusting for baseline differences patients in the PAV group tended to have a lower AI than patients in the PSV group immediately postrandom-ization but this did not reach statistical significance (adjusted difference ndash45 [95 CI ndash95 to 04] p = 007)

Clinical OutcomesTable 3 shows the results for the main clinical endpoints Weaning duration and time to successful SBT did not dif-fer significantly among survivors between the groups Among ICU survivors median time to ICU discharge was shorter in the group randomized to PAV 73 days (IQR 52ndash114 d) ver-sus 124 days (IQR 75ndash308 d) in the PSV group (p = 003) Kaplan-Meier plots of the proportion of at-risk patients achiev-ing successful extubation and live ICU discharge are shown in Figure 3 A and B No difference was found for hospital length of stay ICU mortality or hospital mortality Four patients (15) in the PAV group and three patients (13) in the PSV group died in ICU (respiratory rate [RR] of PAV 114 95 CI 028 to 46 p = 086) Overall in-hospital mortality was nine patients (33) in the PAV group and six patients (26) in the PSV group (RR of PAV 128 95 CI 054 to 31 p = 058)

Cointerventions Sedating MedicationsTable 4 compares the use of medications that may prolong weaning (by causing sedation or neuromuscular weakness) in the two groups pre- and postrandomization Type and dose of sedative medication were highly variable among individual patients and therefore we analyzed pre- and postrandomiza-tion changes to determine the impact of the ventilation pro-tocol on sedative use within each patient In the PAV group benzodiazepine narcotic and propofol use measured as both cumulative dose and average daily dose all decreased signifi-cantly postrandomization as did neuromuscular blocker use In the PSV group average daily dose of lorazepam equiva-lent and average daily dose of propofol decreased signifi-cantly postrandomization whereas the proportion of patients

Figure 2 A Asynchrony index pre- and postrandomization on proportional assist ventilation B Asynchrony index pre- and postrandomization on pressure support ventilationPAV群 PSV群

Asynchrony Index () = 同調イベントの回数全呼吸回数 times 100

IntensiveCareMed(2006)321515ndash1522

AI>10は3名いたがPAVに設定変更後0名となった

AI>10は6名いたがPSVに設定変更後4名となった

Clinical Outcomes

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

成功したSBTまでの数抜管成功までの数に変わりなしICU退室までの数はPAV群で有意に短い

Kaplan-Meier curves

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

抜管成功までの時間 ICU退室までの時間

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 9

TABLE 4 Comparison of Medications That May Alter Weaning Duration Pre- and Postrandomization

Variables

PAV Group (n = 27)

PSV Group (n = 23) Postrandomization

Pre randomization Pos trandomization Pre randomization Pos trandomizationAdjusted Difference PSV ndash PAV (95 CI)

Lorazepam equivalents

Cumulative dose mg

308 (207) 48 (257)a 137 (440) 54 (258) 237 (ndash580 to 1055)

Average daily dose mg

44 (255) 09 (43)a 23 (85) 06 (40)a 24 (ndash62 to 110)

Single-day dose mgb

08 (72) 12 (50) 16 (70) 10 (64) 32 (ndash13 to 78)

Morphine equivalents

Cumulative dose mg

5003 (6188) 1140 (3140)b 1710 (13310) 1610 (4860) 591 (ndash2515 to 3696)

Average daily dose mg

858 (1638) 170 (578)c 428 (1724) 136 (320)d 108 (ndash316 to 532)

Single-day dose mga

180 (1240) 187 (77) 100 (1920) 50 (630)d 334 (ndash026 to 670)d

Propofol

Cumulative dose mg

7400 (55500) 0 (700)a 29450 (85800) 0 (27660) 226 (ndash42156 to 42607)

Average daily dose mg

529 (11690) 0 (53)a 4879 (15980) 0 (2738)a 651 (ndash3211 to 4514)

Single-day dose mga

0 (0) 0 (0) 0 (1000) 0 (0) 344 (ndash3781 to 4469)

Neuroleptics

Days receiving neuroleptics n

0 (2) 0 (4) 0 (0) 1 (4)a ndash16 (ndash34 to 03)

Proportion of patients who received a neuroleptic

407 407 174 565b

NMBs

Days receiving NMBs n

1 (1) 0 (0)a 0 (1) 0 (0) ndash01 (ndash05 to 03)

Proportion of patients who received an NMB

518 148a 348 174

Corticosteroids

Days receiving corticosteroids n

0 (0) 0 (0) 0 (2) 0 (2) 02 (ndash06 to 10)

Proportion of patients who received corticosteroids

185 148 348 261

PAV = proportional assist ventilation PSV = pressure support ventilation NMB = neuromuscular blockera p lt 005b Single-day dose = total dose given day prior to randomization or total dose give day after randomizationc p lt 0005d p lt 00005e p le 007All data points except percentage of patients receiving neuroleptics NMBs or corticosteroids are presented as median (interquartile range)

鎮静剤鎮痛剤についてはPAV群とPSV群で変わりなし

その他bull 不整脈気胸プロトコル施中の死亡

は両群とも起こらなかった

筆者らの考察bull PSVはweaningに最も使われているモードだ

がそれと較しても遜ないweaningができることがわかった

bull アシストが少ない傾向であったがPAV群のが呼吸窮迫となる回数が少なかった

rarr同調性が良くなった結果の可能性

ICU退室までの期間はPAV群で短いという結果であったがpowerが少ない

筆者らの考察bull 単施設での結果であり他の施設にこの

新しいプロトコルそのものが受けれられるかは不明である

bull SBTに失敗した患者を主に対象としておりICU患者すべてに適応はできない

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる

Page 13: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

PAVのメリットデメリット

bull メリット患者-呼吸器の同調性が良い呼吸筋を適度に使うことができる

bull デメリット患者の呼吸努が弱い場合はサポートが減ってしまうリークがある場合過度にサポート圧が上がってしまう可能性(runaway現象)

PAVのエビデンスbull PAVとPSVによるウィーニングをい48

時間観察したRCTではPAVのが調節呼吸に戻る割合が低かった(PAV 11 vsPSV 22 p=004)

bull 夜間PAVにすることで睡眠の質が改善したIntensiveCareMed2008342026ndash2034

Crit CareMed2007351048ndash1054

PAVは呼吸器離脱困難患者のウィーニングにおいて

優れたモードか

本の論

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Critical Care Medicine wwwccmjournalorg 1

Objectives Despite protocols incorporating spontaneous breath-ing trials 31 of ICU patients experience difficult or prolonged weaning from mechanical ventilation Nonfatiguing modes such as pressure support ventilation are recommended Proportional assist ventilation provides assistance in proportion to patient effort which may optimize weaning However it is not known how proportional assist ventilation performs relative to pressure sup-port ventilation over a prolonged period in the complex ICU set-ting The purpose of this study was to compare the physiologic and clinical performance (failure rate) safety and feasibility of protocols using daily spontaneous breathing trial plus pressure support ventilation versus proportional assist ventilation until ven-tilation discontinuationDesign Single-center unblinded pilot randomized controlled trialSetting Medical-surgical ICU of a tertiary-care hospitalPatients Adult patients intubated greater than 36 hours were ran-domized if they met eligibility criteria for partial ventilatory support

tolerated pressure support ventilation greater than or equal to 30 minutes and either failed or did not meet criteria for a sponta-neous breathing trialInterventions Patients were randomized to the pressure support ventilation or proportional assist ventilation protocol (PAV+ Puri-tan Bennett 840 Covidien Boulder CO) Both protocols used progressive decreases in level of assistance as tolerated coupled with daily assessment for spontaneous breathing trialsMeasurements and Main Results Of 54 patients randomized outcome data are available for 50 patients 27 were randomized to receive proportional assist ventilation and 23 to receive pres-sure support ventilation There were no adverse events linked to the study interventions and protocol violations were infrequent Recruitment was slower than projected (13 patients per month) The median (interquartile range) time from randomization to suc-cessful extubation was 39 days (28ndash84 d) on proportional assist ventilation versus 49 days (29ndash263 d) on pressure support ventilation (p = 039) Time to live ICU discharge was 73 days (52ndash114 d) on proportional assist ventilation versus 124 days (75ndash308 d) on pressure support ventilation (p = 003)Conclusion This pilot study demonstrates the utility safety and feasibility of the weaning protocols and provides important infor-mation to guide the design of a future randomized controlled trial comparing weaning from mechanical ventilation on pressure sup-port ventilation versus proportional assist ventilation (Crit Care Med 2016 XX00ndash00)Key Words assisted mechanical ventilation asynchrony index proportional assist ventilation respiratory muscles ventilator weaning

Difficult or prolonged weaning from mechanical ventilation (MV) affects 31 (range 26ndash42) of mechanically ventilated patients (1) and increases

patient morbidity mortality ICU length of stay and cost to the healthcare system (2) When initial weaning attempts fail to liberate patients from MV ideally clinicians should choose

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights ReservedDOI 101097CCM0000000000001600

1London Health Sciences Centre London ON Canada2Department of Medicine The University of Western Ontario London ON Canada

This study was performed at London Health Sciences Centre University Hospital London CanadaSupplemental digital content is available for this article Direct URL cita-tions appear in the printed text and are provided in the HTML and PDF versions of this article on the journalrsquos website (httpjournalslwwcomccmjournal)Supported in part by Critical Care Western Department of Medicine The University of Western Ontario Program of Experimental Medi-cine (POEM) Department of Medicine University of Western Ontario Lawson Health Research Institute London Health Sciences Centre London Canada and Covidien Boulder USADr Bosma has received honourarium for lectures and travel reimburse-ments from Covidien Dr Bosmarsquos institution has received grant support from Covidien for Dr Bosmarsquos research studies The remaining authors have disclosed that they do not have any potential conflicts of interestFor information regarding this article E-mail KarenJBosmalhsconca

A Pilot Randomized Trial Comparing Weaning From Mechanical Ventilation on Pressure Support Versus Proportional Assist Ventilation

Karen J Bosma MD FRCPC12 Brooke A Read MHS BSc RRT1 Mohammad J Bahrgard Nikoo MD FRCPC12 Philip M Jones MD FRCPC MSc12 Fran A Priestap MSc1 James F Lewis MD FRCPC12

Crit Care Med 2016 Jan 20[Epub ahead of print]

論のPICO

P 呼吸器離脱困難な患者(simple weaningではない)

I PAVを使

C PSVを使

O 呼吸器離脱までの時間

Study Designbull 単施設盲検前向きランダム化較

試験(pilot study)bull カナダロンドンにあるUniversity

Hospital-London Health Science Centre の内科外科ICU(20床)

bull ランダム割付はcomputer-generated random number sequence によってわれ隠蔽化されている

PatientsInclusion criteria

bull 18歳以上の挿管患者bull 36時間以上呼吸器管理をされている

PatientsExclusion criteria

1) 48時間以内にwithdrawが考慮される患者2) 位の脊髄障害進性の神経筋疾患の

よって慢性的な呼吸器依存となる可能性があるまたは脳外科的介が必要である患者

Enrolment Procedure

Supplemental Digital Content Figure 1 Enrolment Procedure

Daily Screening of all ICU patients gt18 yrs of age Intubated gt36 hours

Exclude if Exclusion Criteria

Successful SBT awaiting extubation Withdrawal of life support considered High spinal cord injury Neurosurgical patient Neuromuscular disease

Follow using daily eligibility checklist until meet all Clinical Stability Criteria to trial mode of partial ventilatory support (PAV or PSV)

partialcomplete reversal of cause of respiratory failure

body temperature between 36ordmC and 39ordmC metabolic disorders corrected pH gt732 hemoglobin gt70 gL no ongoing bleeding stable hemodynamic status Able to trigger PaO2 ge 60 mmHg or SaO2 ge 90 on FiO2le

060 and PEEP le 15 cm H2O ACVC Pplatle30cmH2O with Vt 6-8 mlkg

OR ACPC PC+PEEPle30cmH2O

Exclude if Exclusion Criteria Has met all Criteria for Switching to

Spontaneous Mode of Ventilation gt24 hours

Pre-inclusion Test on PSV 15 cmH2O for 30 minutes

Patient fails Pre-inclusion test Return to AC ventilation and reassess within 24 hour

Patient passes Pre-inclusion test Assess for Criteria to Initiate Weaning

PaO2FiO2gt200 on FiO2 le050 and PEEP le8 cmH2O No inotropesvasopressors except low dose dopamine

RSBI lt105 Patient passes Pre-SBT readiness assessment Proceed to SBT Place patient on CPAP 5 cmH2O and FiO2 040 for 30 minutes

Patient does not meet Criteria to Initiate Weaning

Patient meets Criteria to Initiate Weaning Proceed to Pre-SBT readiness assessment Place patient on CPAP for 1-2 min measure RSBI

RSBIge 105 Patient fails Pre-SBT readiness assessment

Patient passes SBT Patient fails SBT

Exclude if Exclusion Criteria

Passed SBT and tolerating PSVle7cmH2O and PEEPle5 cmH2O while awaiting extubation

Obtain informed consent from Patient or SDM as soon as available (deferred consent permitted)

Randomize

Exclude if Exclusion Criteria

Patient SDM refuse consent

bull 呼吸不全の原因に改善が認められてきている

bull 体温は36〜39度であるbull pH>732であるbull Hb>7gdLで現在出なしbull 動態が安定しているbull トリガーできるbull FiO2≦06 PEEP≦15cmH2Oで

PaO2≧60mmHg or SaO2≧90bull ACVCでVt 6-8mlkgで

Pplat≦30cmH2OもしくはACPCでPC+PEEP≦30cmH2O

上記をすべて満たせばプレテストとしてPSV 15cmH2Oに変更

Enrolment Procedure

Supplemental Digital Content Figure 1 Enrolment Procedure

Daily Screening of all ICU patients gt18 yrs of age Intubated gt36 hours

Exclude if Exclusion Criteria

Successful SBT awaiting extubation Withdrawal of life support considered High spinal cord injury Neurosurgical patient Neuromuscular disease

Follow using daily eligibility checklist until meet all Clinical Stability Criteria to trial mode of partial ventilatory support (PAV or PSV)

partialcomplete reversal of cause of respiratory failure

body temperature between 36ordmC and 39ordmC metabolic disorders corrected pH gt732 hemoglobin gt70 gL no ongoing bleeding stable hemodynamic status Able to trigger PaO2 ge 60 mmHg or SaO2 ge 90 on FiO2le

060 and PEEP le 15 cm H2O ACVC Pplatle30cmH2O with Vt 6-8 mlkg

OR ACPC PC+PEEPle30cmH2O

Exclude if Exclusion Criteria Has met all Criteria for Switching to

Spontaneous Mode of Ventilation gt24 hours

Pre-inclusion Test on PSV 15 cmH2O for 30 minutes

Patient fails Pre-inclusion test Return to AC ventilation and reassess within 24 hour

Patient passes Pre-inclusion test Assess for Criteria to Initiate Weaning

PaO2FiO2gt200 on FiO2 le050 and PEEP le8 cmH2O No inotropesvasopressors except low dose dopamine

RSBI lt105 Patient passes Pre-SBT readiness assessment Proceed to SBT Place patient on CPAP 5 cmH2O and FiO2 040 for 30 minutes

Patient does not meet Criteria to Initiate Weaning

Patient meets Criteria to Initiate Weaning Proceed to Pre-SBT readiness assessment Place patient on CPAP for 1-2 min measure RSBI

RSBIge 105 Patient fails Pre-SBT readiness assessment

Patient passes SBT Patient fails SBT

Exclude if Exclusion Criteria

Passed SBT and tolerating PSVle7cmH2O and PEEPle5 cmH2O while awaiting extubation

Obtain informed consent from Patient or SDM as soon as available (deferred consent permitted)

Randomize

Exclude if Exclusion Criteria

Patient SDM refuse consent

呼吸窮迫がなければプレテストクリアbull FiO2≦05 PEEP≦8cmH2OでPF≧200bull 強薬昇圧剤の使なしであればSBTへ

まず1-2分CPAP 5cmH2OにしてRSBIを測定

RSBI<105であればSBTへPSV 5cmH2Oにして30-120分

酸素化の条件を満たさないRSBI≧105SBT失敗であれば割り付けへ

Ventilation Protocols bull 呼吸器はPuritan Bennett 840 を使bull 後述するPAVプロトコールPSVプロトコー

ルに従って呼吸療法(RRT)が設定を変更(respiratory distress 呼吸窮迫がないようにできるだけサポートを下げていく)

bull 設定したmaxのサポートが必要になった場合はACに戻す

bull PAV群はPAV+とACPSV群はPSVとACしか使しない

Ventilation Protocols bull RCTの開始前に15名の患者をいて観察

研究をいRRTにプロトコルの確認をってもらった

bull プロトコルはすぐにられるようにラミネートしてベッドサイドに置いた

bull RRTは毎SBTをえるかどうか後述するチェックリストをいて確認除外項を満たさない場合は毎RSBIのアセスメントとSBTをう

Respiratory Distress bull 呼吸数>35回minbull 脈拍>140min もしくはベースラインより20以上の

上昇bull 収縮期圧>180mmHg もしくは<90mmHg ベース

ラインより30以上の変化bull 不安bull 発汗bull 副呼吸筋の使bull 奇異呼吸bull 呼吸困難の訴え

PAV protocolSupplemental Digital Content Figure 2a Proportional Assist Ventilation Protocol

bull 開始の設定はサポート率70

bull 2-3時間おきにRRTがアセスメントし問題なければ10-20ずつPAVサポートを下げる

サポート率90まで上げてもダメな場合はACに戻す

PSV protocolSupplemental Digital Content Figure 2b Pressure Support Ventilation Protocol

bull 開始の設定はPS 15cmH2O

bull まずRRlt35 VTgt5mLkgで耐えられるPSまで下げる

bull 2-3時間おきにRRTがアセスメントしてPS2-3cmH2Pずつ下げる

Data collection form for RRTsSupplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

SBTの開始基準bull PF≧200bull FiO2≦05bull pH≧732bull トリガーできるbull 24時間以内に筋虚イベントなしbull 強薬昇圧剤の使なし

Data collection form for RRTs

Supplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

CPAPにしてRSBIを測定RSBI<105であればSBT施

PAV群はサポート率20PSV群はPS 5cmH2OでSBTを施

PEEP≦8cmH2OでSaO2≧90呼吸窮迫なしであればSBTクリア

抜管へ

Endpoints 呼吸器離脱までの時間ICU病院滞在期間

呼吸器離脱成功の定義抜管後48時間再挿管なし侵襲的呼吸(NIV)の使もなし気管切開使時は呼吸器から最終的に外れた時点1のうち12時間以上NIVを使していない

Statistical Analysis bull Pilot studyのためサンプルサイズの計

算はしていないbull P値<005を統計学的有意としたbull Intention to treat approachをった

RESULTS

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial

エンロール期間は42か80名が基準を満たしが最終的に割付されたのは54名

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial最終的に解析がわれたのはPAV群27名PSV群23名

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

ベースラインは両群間で変わりなし重症度はAPACHEⅡ 27点程度患者は肺炎が3割もともとCOPDを持っている患者は2割程度

Feasibility Outcomes bull 呼吸窮迫でRRTが呼吸器サポートを上昇

させた回数はPAV群で76回232daysPSV群で124回295daysと有意にPSV群で多かった(p=0002)

bull ACに戻した回数はPAV群で66回PSV群で90回と変わりなし(p=061)

bull 全くACに戻さなかった患者はPAV群で627名PSV群で623名(p=10)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

ベースラインは両群間で変わりなしPAVにすることでΔP圧Peak圧が低く吸気時間呼吸時間がくなるPSV群と較するとPAV群のがPaoAreaPaoAreaVE(呼吸器のサポートの強さをす)が低くVTPaoAreaがくsecond post-randomizationの解析においてΔP圧Peak圧が低い

Asynchrony index

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 7

B635) documenting compliance with the daily screening for SBT noncompletion of either the CPAP screening test or SBT occurred on 5 of 292 of patient study days (17) and this violation involved two patients in each study arm The second audit examining for nine different potential deviations found complete adherence to study protocol on 254 of 292 of patient study days (900) There were no adverse events (pneumo-thorax arrhythmia or death) resulting from any of the study interventions

Physiologic Response to Ventilation ProtocolsRecordings of flow and airway pressure were available for analysis for all 50 patients prerandomization 49 patients immediately (median 1 interquartile range [IQR] 0ndash15 hr) postrandomization and for 26 patients during subsequent days (median 27 IQR 20ndash70-hr postrandomization) Approximately 15000 breaths in 125 recordings were analyzed Table 2 compares the respiratory parameters in both groups pre- and postrandomization There were no statistically sig-nificant differences between the PAV and PSV groups at base-line prerandomization (p values not shown) In the PAV group there was a significant decrease in peak airway pressure delta pressure (measured as peak airway opening pressure minus the PEEP) and area under the inspiratory pressure curve immediately postrandomization and an increase in the per-cent of total cycle time spent in inspiration (TiTtot) whereas there were no significant differences immediately postran-domization in the PSV group The second measurement of postrandomization respiratory parameters revealed signifi-cant decreases in peak airway pressure and delta pressure in both PAV and PSV groups Adjusting for baseline measure-ments as a covariate the PAV group received 7 cm H2

O∙s less ventilator assistance (PaoArea) per liter of minute ventilation (p lt 00001) and increased their VT by 115 mLcm H

2O∙s

relative to the PSV group (p lt 0001)PAV significantly reduced AI postrandomization (Fig 2A)

In total nine patients (18) had a high AI (gt 10 of breaths) on PSV or on AC mode at baseline No patients had a high AI on PAV In the PSV group six patients (26) had a high AI

at baseline four (17) immediately postrandomization and one patient (7) after further adjustment of ventilator settings according to the PSV protocol (Fig 2B) Adjusting for baseline differences patients in the PAV group tended to have a lower AI than patients in the PSV group immediately postrandom-ization but this did not reach statistical significance (adjusted difference ndash45 [95 CI ndash95 to 04] p = 007)

Clinical OutcomesTable 3 shows the results for the main clinical endpoints Weaning duration and time to successful SBT did not dif-fer significantly among survivors between the groups Among ICU survivors median time to ICU discharge was shorter in the group randomized to PAV 73 days (IQR 52ndash114 d) ver-sus 124 days (IQR 75ndash308 d) in the PSV group (p = 003) Kaplan-Meier plots of the proportion of at-risk patients achiev-ing successful extubation and live ICU discharge are shown in Figure 3 A and B No difference was found for hospital length of stay ICU mortality or hospital mortality Four patients (15) in the PAV group and three patients (13) in the PSV group died in ICU (respiratory rate [RR] of PAV 114 95 CI 028 to 46 p = 086) Overall in-hospital mortality was nine patients (33) in the PAV group and six patients (26) in the PSV group (RR of PAV 128 95 CI 054 to 31 p = 058)

Cointerventions Sedating MedicationsTable 4 compares the use of medications that may prolong weaning (by causing sedation or neuromuscular weakness) in the two groups pre- and postrandomization Type and dose of sedative medication were highly variable among individual patients and therefore we analyzed pre- and postrandomiza-tion changes to determine the impact of the ventilation pro-tocol on sedative use within each patient In the PAV group benzodiazepine narcotic and propofol use measured as both cumulative dose and average daily dose all decreased signifi-cantly postrandomization as did neuromuscular blocker use In the PSV group average daily dose of lorazepam equiva-lent and average daily dose of propofol decreased signifi-cantly postrandomization whereas the proportion of patients

Figure 2 A Asynchrony index pre- and postrandomization on proportional assist ventilation B Asynchrony index pre- and postrandomization on pressure support ventilationPAV群 PSV群

Asynchrony Index () = 同調イベントの回数全呼吸回数 times 100

IntensiveCareMed(2006)321515ndash1522

AI>10は3名いたがPAVに設定変更後0名となった

AI>10は6名いたがPSVに設定変更後4名となった

Clinical Outcomes

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

成功したSBTまでの数抜管成功までの数に変わりなしICU退室までの数はPAV群で有意に短い

Kaplan-Meier curves

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

抜管成功までの時間 ICU退室までの時間

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 9

TABLE 4 Comparison of Medications That May Alter Weaning Duration Pre- and Postrandomization

Variables

PAV Group (n = 27)

PSV Group (n = 23) Postrandomization

Pre randomization Pos trandomization Pre randomization Pos trandomizationAdjusted Difference PSV ndash PAV (95 CI)

Lorazepam equivalents

Cumulative dose mg

308 (207) 48 (257)a 137 (440) 54 (258) 237 (ndash580 to 1055)

Average daily dose mg

44 (255) 09 (43)a 23 (85) 06 (40)a 24 (ndash62 to 110)

Single-day dose mgb

08 (72) 12 (50) 16 (70) 10 (64) 32 (ndash13 to 78)

Morphine equivalents

Cumulative dose mg

5003 (6188) 1140 (3140)b 1710 (13310) 1610 (4860) 591 (ndash2515 to 3696)

Average daily dose mg

858 (1638) 170 (578)c 428 (1724) 136 (320)d 108 (ndash316 to 532)

Single-day dose mga

180 (1240) 187 (77) 100 (1920) 50 (630)d 334 (ndash026 to 670)d

Propofol

Cumulative dose mg

7400 (55500) 0 (700)a 29450 (85800) 0 (27660) 226 (ndash42156 to 42607)

Average daily dose mg

529 (11690) 0 (53)a 4879 (15980) 0 (2738)a 651 (ndash3211 to 4514)

Single-day dose mga

0 (0) 0 (0) 0 (1000) 0 (0) 344 (ndash3781 to 4469)

Neuroleptics

Days receiving neuroleptics n

0 (2) 0 (4) 0 (0) 1 (4)a ndash16 (ndash34 to 03)

Proportion of patients who received a neuroleptic

407 407 174 565b

NMBs

Days receiving NMBs n

1 (1) 0 (0)a 0 (1) 0 (0) ndash01 (ndash05 to 03)

Proportion of patients who received an NMB

518 148a 348 174

Corticosteroids

Days receiving corticosteroids n

0 (0) 0 (0) 0 (2) 0 (2) 02 (ndash06 to 10)

Proportion of patients who received corticosteroids

185 148 348 261

PAV = proportional assist ventilation PSV = pressure support ventilation NMB = neuromuscular blockera p lt 005b Single-day dose = total dose given day prior to randomization or total dose give day after randomizationc p lt 0005d p lt 00005e p le 007All data points except percentage of patients receiving neuroleptics NMBs or corticosteroids are presented as median (interquartile range)

鎮静剤鎮痛剤についてはPAV群とPSV群で変わりなし

その他bull 不整脈気胸プロトコル施中の死亡

は両群とも起こらなかった

筆者らの考察bull PSVはweaningに最も使われているモードだ

がそれと較しても遜ないweaningができることがわかった

bull アシストが少ない傾向であったがPAV群のが呼吸窮迫となる回数が少なかった

rarr同調性が良くなった結果の可能性

ICU退室までの期間はPAV群で短いという結果であったがpowerが少ない

筆者らの考察bull 単施設での結果であり他の施設にこの

新しいプロトコルそのものが受けれられるかは不明である

bull SBTに失敗した患者を主に対象としておりICU患者すべてに適応はできない

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる

Page 14: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

PAVのエビデンスbull PAVとPSVによるウィーニングをい48

時間観察したRCTではPAVのが調節呼吸に戻る割合が低かった(PAV 11 vsPSV 22 p=004)

bull 夜間PAVにすることで睡眠の質が改善したIntensiveCareMed2008342026ndash2034

Crit CareMed2007351048ndash1054

PAVは呼吸器離脱困難患者のウィーニングにおいて

優れたモードか

本の論

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Critical Care Medicine wwwccmjournalorg 1

Objectives Despite protocols incorporating spontaneous breath-ing trials 31 of ICU patients experience difficult or prolonged weaning from mechanical ventilation Nonfatiguing modes such as pressure support ventilation are recommended Proportional assist ventilation provides assistance in proportion to patient effort which may optimize weaning However it is not known how proportional assist ventilation performs relative to pressure sup-port ventilation over a prolonged period in the complex ICU set-ting The purpose of this study was to compare the physiologic and clinical performance (failure rate) safety and feasibility of protocols using daily spontaneous breathing trial plus pressure support ventilation versus proportional assist ventilation until ven-tilation discontinuationDesign Single-center unblinded pilot randomized controlled trialSetting Medical-surgical ICU of a tertiary-care hospitalPatients Adult patients intubated greater than 36 hours were ran-domized if they met eligibility criteria for partial ventilatory support

tolerated pressure support ventilation greater than or equal to 30 minutes and either failed or did not meet criteria for a sponta-neous breathing trialInterventions Patients were randomized to the pressure support ventilation or proportional assist ventilation protocol (PAV+ Puri-tan Bennett 840 Covidien Boulder CO) Both protocols used progressive decreases in level of assistance as tolerated coupled with daily assessment for spontaneous breathing trialsMeasurements and Main Results Of 54 patients randomized outcome data are available for 50 patients 27 were randomized to receive proportional assist ventilation and 23 to receive pres-sure support ventilation There were no adverse events linked to the study interventions and protocol violations were infrequent Recruitment was slower than projected (13 patients per month) The median (interquartile range) time from randomization to suc-cessful extubation was 39 days (28ndash84 d) on proportional assist ventilation versus 49 days (29ndash263 d) on pressure support ventilation (p = 039) Time to live ICU discharge was 73 days (52ndash114 d) on proportional assist ventilation versus 124 days (75ndash308 d) on pressure support ventilation (p = 003)Conclusion This pilot study demonstrates the utility safety and feasibility of the weaning protocols and provides important infor-mation to guide the design of a future randomized controlled trial comparing weaning from mechanical ventilation on pressure sup-port ventilation versus proportional assist ventilation (Crit Care Med 2016 XX00ndash00)Key Words assisted mechanical ventilation asynchrony index proportional assist ventilation respiratory muscles ventilator weaning

Difficult or prolonged weaning from mechanical ventilation (MV) affects 31 (range 26ndash42) of mechanically ventilated patients (1) and increases

patient morbidity mortality ICU length of stay and cost to the healthcare system (2) When initial weaning attempts fail to liberate patients from MV ideally clinicians should choose

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights ReservedDOI 101097CCM0000000000001600

1London Health Sciences Centre London ON Canada2Department of Medicine The University of Western Ontario London ON Canada

This study was performed at London Health Sciences Centre University Hospital London CanadaSupplemental digital content is available for this article Direct URL cita-tions appear in the printed text and are provided in the HTML and PDF versions of this article on the journalrsquos website (httpjournalslwwcomccmjournal)Supported in part by Critical Care Western Department of Medicine The University of Western Ontario Program of Experimental Medi-cine (POEM) Department of Medicine University of Western Ontario Lawson Health Research Institute London Health Sciences Centre London Canada and Covidien Boulder USADr Bosma has received honourarium for lectures and travel reimburse-ments from Covidien Dr Bosmarsquos institution has received grant support from Covidien for Dr Bosmarsquos research studies The remaining authors have disclosed that they do not have any potential conflicts of interestFor information regarding this article E-mail KarenJBosmalhsconca

A Pilot Randomized Trial Comparing Weaning From Mechanical Ventilation on Pressure Support Versus Proportional Assist Ventilation

Karen J Bosma MD FRCPC12 Brooke A Read MHS BSc RRT1 Mohammad J Bahrgard Nikoo MD FRCPC12 Philip M Jones MD FRCPC MSc12 Fran A Priestap MSc1 James F Lewis MD FRCPC12

Crit Care Med 2016 Jan 20[Epub ahead of print]

論のPICO

P 呼吸器離脱困難な患者(simple weaningではない)

I PAVを使

C PSVを使

O 呼吸器離脱までの時間

Study Designbull 単施設盲検前向きランダム化較

試験(pilot study)bull カナダロンドンにあるUniversity

Hospital-London Health Science Centre の内科外科ICU(20床)

bull ランダム割付はcomputer-generated random number sequence によってわれ隠蔽化されている

PatientsInclusion criteria

bull 18歳以上の挿管患者bull 36時間以上呼吸器管理をされている

PatientsExclusion criteria

1) 48時間以内にwithdrawが考慮される患者2) 位の脊髄障害進性の神経筋疾患の

よって慢性的な呼吸器依存となる可能性があるまたは脳外科的介が必要である患者

Enrolment Procedure

Supplemental Digital Content Figure 1 Enrolment Procedure

Daily Screening of all ICU patients gt18 yrs of age Intubated gt36 hours

Exclude if Exclusion Criteria

Successful SBT awaiting extubation Withdrawal of life support considered High spinal cord injury Neurosurgical patient Neuromuscular disease

Follow using daily eligibility checklist until meet all Clinical Stability Criteria to trial mode of partial ventilatory support (PAV or PSV)

partialcomplete reversal of cause of respiratory failure

body temperature between 36ordmC and 39ordmC metabolic disorders corrected pH gt732 hemoglobin gt70 gL no ongoing bleeding stable hemodynamic status Able to trigger PaO2 ge 60 mmHg or SaO2 ge 90 on FiO2le

060 and PEEP le 15 cm H2O ACVC Pplatle30cmH2O with Vt 6-8 mlkg

OR ACPC PC+PEEPle30cmH2O

Exclude if Exclusion Criteria Has met all Criteria for Switching to

Spontaneous Mode of Ventilation gt24 hours

Pre-inclusion Test on PSV 15 cmH2O for 30 minutes

Patient fails Pre-inclusion test Return to AC ventilation and reassess within 24 hour

Patient passes Pre-inclusion test Assess for Criteria to Initiate Weaning

PaO2FiO2gt200 on FiO2 le050 and PEEP le8 cmH2O No inotropesvasopressors except low dose dopamine

RSBI lt105 Patient passes Pre-SBT readiness assessment Proceed to SBT Place patient on CPAP 5 cmH2O and FiO2 040 for 30 minutes

Patient does not meet Criteria to Initiate Weaning

Patient meets Criteria to Initiate Weaning Proceed to Pre-SBT readiness assessment Place patient on CPAP for 1-2 min measure RSBI

RSBIge 105 Patient fails Pre-SBT readiness assessment

Patient passes SBT Patient fails SBT

Exclude if Exclusion Criteria

Passed SBT and tolerating PSVle7cmH2O and PEEPle5 cmH2O while awaiting extubation

Obtain informed consent from Patient or SDM as soon as available (deferred consent permitted)

Randomize

Exclude if Exclusion Criteria

Patient SDM refuse consent

bull 呼吸不全の原因に改善が認められてきている

bull 体温は36〜39度であるbull pH>732であるbull Hb>7gdLで現在出なしbull 動態が安定しているbull トリガーできるbull FiO2≦06 PEEP≦15cmH2Oで

PaO2≧60mmHg or SaO2≧90bull ACVCでVt 6-8mlkgで

Pplat≦30cmH2OもしくはACPCでPC+PEEP≦30cmH2O

上記をすべて満たせばプレテストとしてPSV 15cmH2Oに変更

Enrolment Procedure

Supplemental Digital Content Figure 1 Enrolment Procedure

Daily Screening of all ICU patients gt18 yrs of age Intubated gt36 hours

Exclude if Exclusion Criteria

Successful SBT awaiting extubation Withdrawal of life support considered High spinal cord injury Neurosurgical patient Neuromuscular disease

Follow using daily eligibility checklist until meet all Clinical Stability Criteria to trial mode of partial ventilatory support (PAV or PSV)

partialcomplete reversal of cause of respiratory failure

body temperature between 36ordmC and 39ordmC metabolic disorders corrected pH gt732 hemoglobin gt70 gL no ongoing bleeding stable hemodynamic status Able to trigger PaO2 ge 60 mmHg or SaO2 ge 90 on FiO2le

060 and PEEP le 15 cm H2O ACVC Pplatle30cmH2O with Vt 6-8 mlkg

OR ACPC PC+PEEPle30cmH2O

Exclude if Exclusion Criteria Has met all Criteria for Switching to

Spontaneous Mode of Ventilation gt24 hours

Pre-inclusion Test on PSV 15 cmH2O for 30 minutes

Patient fails Pre-inclusion test Return to AC ventilation and reassess within 24 hour

Patient passes Pre-inclusion test Assess for Criteria to Initiate Weaning

PaO2FiO2gt200 on FiO2 le050 and PEEP le8 cmH2O No inotropesvasopressors except low dose dopamine

RSBI lt105 Patient passes Pre-SBT readiness assessment Proceed to SBT Place patient on CPAP 5 cmH2O and FiO2 040 for 30 minutes

Patient does not meet Criteria to Initiate Weaning

Patient meets Criteria to Initiate Weaning Proceed to Pre-SBT readiness assessment Place patient on CPAP for 1-2 min measure RSBI

RSBIge 105 Patient fails Pre-SBT readiness assessment

Patient passes SBT Patient fails SBT

Exclude if Exclusion Criteria

Passed SBT and tolerating PSVle7cmH2O and PEEPle5 cmH2O while awaiting extubation

Obtain informed consent from Patient or SDM as soon as available (deferred consent permitted)

Randomize

Exclude if Exclusion Criteria

Patient SDM refuse consent

呼吸窮迫がなければプレテストクリアbull FiO2≦05 PEEP≦8cmH2OでPF≧200bull 強薬昇圧剤の使なしであればSBTへ

まず1-2分CPAP 5cmH2OにしてRSBIを測定

RSBI<105であればSBTへPSV 5cmH2Oにして30-120分

酸素化の条件を満たさないRSBI≧105SBT失敗であれば割り付けへ

Ventilation Protocols bull 呼吸器はPuritan Bennett 840 を使bull 後述するPAVプロトコールPSVプロトコー

ルに従って呼吸療法(RRT)が設定を変更(respiratory distress 呼吸窮迫がないようにできるだけサポートを下げていく)

bull 設定したmaxのサポートが必要になった場合はACに戻す

bull PAV群はPAV+とACPSV群はPSVとACしか使しない

Ventilation Protocols bull RCTの開始前に15名の患者をいて観察

研究をいRRTにプロトコルの確認をってもらった

bull プロトコルはすぐにられるようにラミネートしてベッドサイドに置いた

bull RRTは毎SBTをえるかどうか後述するチェックリストをいて確認除外項を満たさない場合は毎RSBIのアセスメントとSBTをう

Respiratory Distress bull 呼吸数>35回minbull 脈拍>140min もしくはベースラインより20以上の

上昇bull 収縮期圧>180mmHg もしくは<90mmHg ベース

ラインより30以上の変化bull 不安bull 発汗bull 副呼吸筋の使bull 奇異呼吸bull 呼吸困難の訴え

PAV protocolSupplemental Digital Content Figure 2a Proportional Assist Ventilation Protocol

bull 開始の設定はサポート率70

bull 2-3時間おきにRRTがアセスメントし問題なければ10-20ずつPAVサポートを下げる

サポート率90まで上げてもダメな場合はACに戻す

PSV protocolSupplemental Digital Content Figure 2b Pressure Support Ventilation Protocol

bull 開始の設定はPS 15cmH2O

bull まずRRlt35 VTgt5mLkgで耐えられるPSまで下げる

bull 2-3時間おきにRRTがアセスメントしてPS2-3cmH2Pずつ下げる

Data collection form for RRTsSupplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

SBTの開始基準bull PF≧200bull FiO2≦05bull pH≧732bull トリガーできるbull 24時間以内に筋虚イベントなしbull 強薬昇圧剤の使なし

Data collection form for RRTs

Supplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

CPAPにしてRSBIを測定RSBI<105であればSBT施

PAV群はサポート率20PSV群はPS 5cmH2OでSBTを施

PEEP≦8cmH2OでSaO2≧90呼吸窮迫なしであればSBTクリア

抜管へ

Endpoints 呼吸器離脱までの時間ICU病院滞在期間

呼吸器離脱成功の定義抜管後48時間再挿管なし侵襲的呼吸(NIV)の使もなし気管切開使時は呼吸器から最終的に外れた時点1のうち12時間以上NIVを使していない

Statistical Analysis bull Pilot studyのためサンプルサイズの計

算はしていないbull P値<005を統計学的有意としたbull Intention to treat approachをった

RESULTS

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial

エンロール期間は42か80名が基準を満たしが最終的に割付されたのは54名

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial最終的に解析がわれたのはPAV群27名PSV群23名

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

ベースラインは両群間で変わりなし重症度はAPACHEⅡ 27点程度患者は肺炎が3割もともとCOPDを持っている患者は2割程度

Feasibility Outcomes bull 呼吸窮迫でRRTが呼吸器サポートを上昇

させた回数はPAV群で76回232daysPSV群で124回295daysと有意にPSV群で多かった(p=0002)

bull ACに戻した回数はPAV群で66回PSV群で90回と変わりなし(p=061)

bull 全くACに戻さなかった患者はPAV群で627名PSV群で623名(p=10)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

ベースラインは両群間で変わりなしPAVにすることでΔP圧Peak圧が低く吸気時間呼吸時間がくなるPSV群と較するとPAV群のがPaoAreaPaoAreaVE(呼吸器のサポートの強さをす)が低くVTPaoAreaがくsecond post-randomizationの解析においてΔP圧Peak圧が低い

Asynchrony index

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 7

B635) documenting compliance with the daily screening for SBT noncompletion of either the CPAP screening test or SBT occurred on 5 of 292 of patient study days (17) and this violation involved two patients in each study arm The second audit examining for nine different potential deviations found complete adherence to study protocol on 254 of 292 of patient study days (900) There were no adverse events (pneumo-thorax arrhythmia or death) resulting from any of the study interventions

Physiologic Response to Ventilation ProtocolsRecordings of flow and airway pressure were available for analysis for all 50 patients prerandomization 49 patients immediately (median 1 interquartile range [IQR] 0ndash15 hr) postrandomization and for 26 patients during subsequent days (median 27 IQR 20ndash70-hr postrandomization) Approximately 15000 breaths in 125 recordings were analyzed Table 2 compares the respiratory parameters in both groups pre- and postrandomization There were no statistically sig-nificant differences between the PAV and PSV groups at base-line prerandomization (p values not shown) In the PAV group there was a significant decrease in peak airway pressure delta pressure (measured as peak airway opening pressure minus the PEEP) and area under the inspiratory pressure curve immediately postrandomization and an increase in the per-cent of total cycle time spent in inspiration (TiTtot) whereas there were no significant differences immediately postran-domization in the PSV group The second measurement of postrandomization respiratory parameters revealed signifi-cant decreases in peak airway pressure and delta pressure in both PAV and PSV groups Adjusting for baseline measure-ments as a covariate the PAV group received 7 cm H2

O∙s less ventilator assistance (PaoArea) per liter of minute ventilation (p lt 00001) and increased their VT by 115 mLcm H

2O∙s

relative to the PSV group (p lt 0001)PAV significantly reduced AI postrandomization (Fig 2A)

In total nine patients (18) had a high AI (gt 10 of breaths) on PSV or on AC mode at baseline No patients had a high AI on PAV In the PSV group six patients (26) had a high AI

at baseline four (17) immediately postrandomization and one patient (7) after further adjustment of ventilator settings according to the PSV protocol (Fig 2B) Adjusting for baseline differences patients in the PAV group tended to have a lower AI than patients in the PSV group immediately postrandom-ization but this did not reach statistical significance (adjusted difference ndash45 [95 CI ndash95 to 04] p = 007)

Clinical OutcomesTable 3 shows the results for the main clinical endpoints Weaning duration and time to successful SBT did not dif-fer significantly among survivors between the groups Among ICU survivors median time to ICU discharge was shorter in the group randomized to PAV 73 days (IQR 52ndash114 d) ver-sus 124 days (IQR 75ndash308 d) in the PSV group (p = 003) Kaplan-Meier plots of the proportion of at-risk patients achiev-ing successful extubation and live ICU discharge are shown in Figure 3 A and B No difference was found for hospital length of stay ICU mortality or hospital mortality Four patients (15) in the PAV group and three patients (13) in the PSV group died in ICU (respiratory rate [RR] of PAV 114 95 CI 028 to 46 p = 086) Overall in-hospital mortality was nine patients (33) in the PAV group and six patients (26) in the PSV group (RR of PAV 128 95 CI 054 to 31 p = 058)

Cointerventions Sedating MedicationsTable 4 compares the use of medications that may prolong weaning (by causing sedation or neuromuscular weakness) in the two groups pre- and postrandomization Type and dose of sedative medication were highly variable among individual patients and therefore we analyzed pre- and postrandomiza-tion changes to determine the impact of the ventilation pro-tocol on sedative use within each patient In the PAV group benzodiazepine narcotic and propofol use measured as both cumulative dose and average daily dose all decreased signifi-cantly postrandomization as did neuromuscular blocker use In the PSV group average daily dose of lorazepam equiva-lent and average daily dose of propofol decreased signifi-cantly postrandomization whereas the proportion of patients

Figure 2 A Asynchrony index pre- and postrandomization on proportional assist ventilation B Asynchrony index pre- and postrandomization on pressure support ventilationPAV群 PSV群

Asynchrony Index () = 同調イベントの回数全呼吸回数 times 100

IntensiveCareMed(2006)321515ndash1522

AI>10は3名いたがPAVに設定変更後0名となった

AI>10は6名いたがPSVに設定変更後4名となった

Clinical Outcomes

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

成功したSBTまでの数抜管成功までの数に変わりなしICU退室までの数はPAV群で有意に短い

Kaplan-Meier curves

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

抜管成功までの時間 ICU退室までの時間

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 9

TABLE 4 Comparison of Medications That May Alter Weaning Duration Pre- and Postrandomization

Variables

PAV Group (n = 27)

PSV Group (n = 23) Postrandomization

Pre randomization Pos trandomization Pre randomization Pos trandomizationAdjusted Difference PSV ndash PAV (95 CI)

Lorazepam equivalents

Cumulative dose mg

308 (207) 48 (257)a 137 (440) 54 (258) 237 (ndash580 to 1055)

Average daily dose mg

44 (255) 09 (43)a 23 (85) 06 (40)a 24 (ndash62 to 110)

Single-day dose mgb

08 (72) 12 (50) 16 (70) 10 (64) 32 (ndash13 to 78)

Morphine equivalents

Cumulative dose mg

5003 (6188) 1140 (3140)b 1710 (13310) 1610 (4860) 591 (ndash2515 to 3696)

Average daily dose mg

858 (1638) 170 (578)c 428 (1724) 136 (320)d 108 (ndash316 to 532)

Single-day dose mga

180 (1240) 187 (77) 100 (1920) 50 (630)d 334 (ndash026 to 670)d

Propofol

Cumulative dose mg

7400 (55500) 0 (700)a 29450 (85800) 0 (27660) 226 (ndash42156 to 42607)

Average daily dose mg

529 (11690) 0 (53)a 4879 (15980) 0 (2738)a 651 (ndash3211 to 4514)

Single-day dose mga

0 (0) 0 (0) 0 (1000) 0 (0) 344 (ndash3781 to 4469)

Neuroleptics

Days receiving neuroleptics n

0 (2) 0 (4) 0 (0) 1 (4)a ndash16 (ndash34 to 03)

Proportion of patients who received a neuroleptic

407 407 174 565b

NMBs

Days receiving NMBs n

1 (1) 0 (0)a 0 (1) 0 (0) ndash01 (ndash05 to 03)

Proportion of patients who received an NMB

518 148a 348 174

Corticosteroids

Days receiving corticosteroids n

0 (0) 0 (0) 0 (2) 0 (2) 02 (ndash06 to 10)

Proportion of patients who received corticosteroids

185 148 348 261

PAV = proportional assist ventilation PSV = pressure support ventilation NMB = neuromuscular blockera p lt 005b Single-day dose = total dose given day prior to randomization or total dose give day after randomizationc p lt 0005d p lt 00005e p le 007All data points except percentage of patients receiving neuroleptics NMBs or corticosteroids are presented as median (interquartile range)

鎮静剤鎮痛剤についてはPAV群とPSV群で変わりなし

その他bull 不整脈気胸プロトコル施中の死亡

は両群とも起こらなかった

筆者らの考察bull PSVはweaningに最も使われているモードだ

がそれと較しても遜ないweaningができることがわかった

bull アシストが少ない傾向であったがPAV群のが呼吸窮迫となる回数が少なかった

rarr同調性が良くなった結果の可能性

ICU退室までの期間はPAV群で短いという結果であったがpowerが少ない

筆者らの考察bull 単施設での結果であり他の施設にこの

新しいプロトコルそのものが受けれられるかは不明である

bull SBTに失敗した患者を主に対象としておりICU患者すべてに適応はできない

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる

Page 15: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

PAVは呼吸器離脱困難患者のウィーニングにおいて

優れたモードか

本の論

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Critical Care Medicine wwwccmjournalorg 1

Objectives Despite protocols incorporating spontaneous breath-ing trials 31 of ICU patients experience difficult or prolonged weaning from mechanical ventilation Nonfatiguing modes such as pressure support ventilation are recommended Proportional assist ventilation provides assistance in proportion to patient effort which may optimize weaning However it is not known how proportional assist ventilation performs relative to pressure sup-port ventilation over a prolonged period in the complex ICU set-ting The purpose of this study was to compare the physiologic and clinical performance (failure rate) safety and feasibility of protocols using daily spontaneous breathing trial plus pressure support ventilation versus proportional assist ventilation until ven-tilation discontinuationDesign Single-center unblinded pilot randomized controlled trialSetting Medical-surgical ICU of a tertiary-care hospitalPatients Adult patients intubated greater than 36 hours were ran-domized if they met eligibility criteria for partial ventilatory support

tolerated pressure support ventilation greater than or equal to 30 minutes and either failed or did not meet criteria for a sponta-neous breathing trialInterventions Patients were randomized to the pressure support ventilation or proportional assist ventilation protocol (PAV+ Puri-tan Bennett 840 Covidien Boulder CO) Both protocols used progressive decreases in level of assistance as tolerated coupled with daily assessment for spontaneous breathing trialsMeasurements and Main Results Of 54 patients randomized outcome data are available for 50 patients 27 were randomized to receive proportional assist ventilation and 23 to receive pres-sure support ventilation There were no adverse events linked to the study interventions and protocol violations were infrequent Recruitment was slower than projected (13 patients per month) The median (interquartile range) time from randomization to suc-cessful extubation was 39 days (28ndash84 d) on proportional assist ventilation versus 49 days (29ndash263 d) on pressure support ventilation (p = 039) Time to live ICU discharge was 73 days (52ndash114 d) on proportional assist ventilation versus 124 days (75ndash308 d) on pressure support ventilation (p = 003)Conclusion This pilot study demonstrates the utility safety and feasibility of the weaning protocols and provides important infor-mation to guide the design of a future randomized controlled trial comparing weaning from mechanical ventilation on pressure sup-port ventilation versus proportional assist ventilation (Crit Care Med 2016 XX00ndash00)Key Words assisted mechanical ventilation asynchrony index proportional assist ventilation respiratory muscles ventilator weaning

Difficult or prolonged weaning from mechanical ventilation (MV) affects 31 (range 26ndash42) of mechanically ventilated patients (1) and increases

patient morbidity mortality ICU length of stay and cost to the healthcare system (2) When initial weaning attempts fail to liberate patients from MV ideally clinicians should choose

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights ReservedDOI 101097CCM0000000000001600

1London Health Sciences Centre London ON Canada2Department of Medicine The University of Western Ontario London ON Canada

This study was performed at London Health Sciences Centre University Hospital London CanadaSupplemental digital content is available for this article Direct URL cita-tions appear in the printed text and are provided in the HTML and PDF versions of this article on the journalrsquos website (httpjournalslwwcomccmjournal)Supported in part by Critical Care Western Department of Medicine The University of Western Ontario Program of Experimental Medi-cine (POEM) Department of Medicine University of Western Ontario Lawson Health Research Institute London Health Sciences Centre London Canada and Covidien Boulder USADr Bosma has received honourarium for lectures and travel reimburse-ments from Covidien Dr Bosmarsquos institution has received grant support from Covidien for Dr Bosmarsquos research studies The remaining authors have disclosed that they do not have any potential conflicts of interestFor information regarding this article E-mail KarenJBosmalhsconca

A Pilot Randomized Trial Comparing Weaning From Mechanical Ventilation on Pressure Support Versus Proportional Assist Ventilation

Karen J Bosma MD FRCPC12 Brooke A Read MHS BSc RRT1 Mohammad J Bahrgard Nikoo MD FRCPC12 Philip M Jones MD FRCPC MSc12 Fran A Priestap MSc1 James F Lewis MD FRCPC12

Crit Care Med 2016 Jan 20[Epub ahead of print]

論のPICO

P 呼吸器離脱困難な患者(simple weaningではない)

I PAVを使

C PSVを使

O 呼吸器離脱までの時間

Study Designbull 単施設盲検前向きランダム化較

試験(pilot study)bull カナダロンドンにあるUniversity

Hospital-London Health Science Centre の内科外科ICU(20床)

bull ランダム割付はcomputer-generated random number sequence によってわれ隠蔽化されている

PatientsInclusion criteria

bull 18歳以上の挿管患者bull 36時間以上呼吸器管理をされている

PatientsExclusion criteria

1) 48時間以内にwithdrawが考慮される患者2) 位の脊髄障害進性の神経筋疾患の

よって慢性的な呼吸器依存となる可能性があるまたは脳外科的介が必要である患者

Enrolment Procedure

Supplemental Digital Content Figure 1 Enrolment Procedure

Daily Screening of all ICU patients gt18 yrs of age Intubated gt36 hours

Exclude if Exclusion Criteria

Successful SBT awaiting extubation Withdrawal of life support considered High spinal cord injury Neurosurgical patient Neuromuscular disease

Follow using daily eligibility checklist until meet all Clinical Stability Criteria to trial mode of partial ventilatory support (PAV or PSV)

partialcomplete reversal of cause of respiratory failure

body temperature between 36ordmC and 39ordmC metabolic disorders corrected pH gt732 hemoglobin gt70 gL no ongoing bleeding stable hemodynamic status Able to trigger PaO2 ge 60 mmHg or SaO2 ge 90 on FiO2le

060 and PEEP le 15 cm H2O ACVC Pplatle30cmH2O with Vt 6-8 mlkg

OR ACPC PC+PEEPle30cmH2O

Exclude if Exclusion Criteria Has met all Criteria for Switching to

Spontaneous Mode of Ventilation gt24 hours

Pre-inclusion Test on PSV 15 cmH2O for 30 minutes

Patient fails Pre-inclusion test Return to AC ventilation and reassess within 24 hour

Patient passes Pre-inclusion test Assess for Criteria to Initiate Weaning

PaO2FiO2gt200 on FiO2 le050 and PEEP le8 cmH2O No inotropesvasopressors except low dose dopamine

RSBI lt105 Patient passes Pre-SBT readiness assessment Proceed to SBT Place patient on CPAP 5 cmH2O and FiO2 040 for 30 minutes

Patient does not meet Criteria to Initiate Weaning

Patient meets Criteria to Initiate Weaning Proceed to Pre-SBT readiness assessment Place patient on CPAP for 1-2 min measure RSBI

RSBIge 105 Patient fails Pre-SBT readiness assessment

Patient passes SBT Patient fails SBT

Exclude if Exclusion Criteria

Passed SBT and tolerating PSVle7cmH2O and PEEPle5 cmH2O while awaiting extubation

Obtain informed consent from Patient or SDM as soon as available (deferred consent permitted)

Randomize

Exclude if Exclusion Criteria

Patient SDM refuse consent

bull 呼吸不全の原因に改善が認められてきている

bull 体温は36〜39度であるbull pH>732であるbull Hb>7gdLで現在出なしbull 動態が安定しているbull トリガーできるbull FiO2≦06 PEEP≦15cmH2Oで

PaO2≧60mmHg or SaO2≧90bull ACVCでVt 6-8mlkgで

Pplat≦30cmH2OもしくはACPCでPC+PEEP≦30cmH2O

上記をすべて満たせばプレテストとしてPSV 15cmH2Oに変更

Enrolment Procedure

Supplemental Digital Content Figure 1 Enrolment Procedure

Daily Screening of all ICU patients gt18 yrs of age Intubated gt36 hours

Exclude if Exclusion Criteria

Successful SBT awaiting extubation Withdrawal of life support considered High spinal cord injury Neurosurgical patient Neuromuscular disease

Follow using daily eligibility checklist until meet all Clinical Stability Criteria to trial mode of partial ventilatory support (PAV or PSV)

partialcomplete reversal of cause of respiratory failure

body temperature between 36ordmC and 39ordmC metabolic disorders corrected pH gt732 hemoglobin gt70 gL no ongoing bleeding stable hemodynamic status Able to trigger PaO2 ge 60 mmHg or SaO2 ge 90 on FiO2le

060 and PEEP le 15 cm H2O ACVC Pplatle30cmH2O with Vt 6-8 mlkg

OR ACPC PC+PEEPle30cmH2O

Exclude if Exclusion Criteria Has met all Criteria for Switching to

Spontaneous Mode of Ventilation gt24 hours

Pre-inclusion Test on PSV 15 cmH2O for 30 minutes

Patient fails Pre-inclusion test Return to AC ventilation and reassess within 24 hour

Patient passes Pre-inclusion test Assess for Criteria to Initiate Weaning

PaO2FiO2gt200 on FiO2 le050 and PEEP le8 cmH2O No inotropesvasopressors except low dose dopamine

RSBI lt105 Patient passes Pre-SBT readiness assessment Proceed to SBT Place patient on CPAP 5 cmH2O and FiO2 040 for 30 minutes

Patient does not meet Criteria to Initiate Weaning

Patient meets Criteria to Initiate Weaning Proceed to Pre-SBT readiness assessment Place patient on CPAP for 1-2 min measure RSBI

RSBIge 105 Patient fails Pre-SBT readiness assessment

Patient passes SBT Patient fails SBT

Exclude if Exclusion Criteria

Passed SBT and tolerating PSVle7cmH2O and PEEPle5 cmH2O while awaiting extubation

Obtain informed consent from Patient or SDM as soon as available (deferred consent permitted)

Randomize

Exclude if Exclusion Criteria

Patient SDM refuse consent

呼吸窮迫がなければプレテストクリアbull FiO2≦05 PEEP≦8cmH2OでPF≧200bull 強薬昇圧剤の使なしであればSBTへ

まず1-2分CPAP 5cmH2OにしてRSBIを測定

RSBI<105であればSBTへPSV 5cmH2Oにして30-120分

酸素化の条件を満たさないRSBI≧105SBT失敗であれば割り付けへ

Ventilation Protocols bull 呼吸器はPuritan Bennett 840 を使bull 後述するPAVプロトコールPSVプロトコー

ルに従って呼吸療法(RRT)が設定を変更(respiratory distress 呼吸窮迫がないようにできるだけサポートを下げていく)

bull 設定したmaxのサポートが必要になった場合はACに戻す

bull PAV群はPAV+とACPSV群はPSVとACしか使しない

Ventilation Protocols bull RCTの開始前に15名の患者をいて観察

研究をいRRTにプロトコルの確認をってもらった

bull プロトコルはすぐにられるようにラミネートしてベッドサイドに置いた

bull RRTは毎SBTをえるかどうか後述するチェックリストをいて確認除外項を満たさない場合は毎RSBIのアセスメントとSBTをう

Respiratory Distress bull 呼吸数>35回minbull 脈拍>140min もしくはベースラインより20以上の

上昇bull 収縮期圧>180mmHg もしくは<90mmHg ベース

ラインより30以上の変化bull 不安bull 発汗bull 副呼吸筋の使bull 奇異呼吸bull 呼吸困難の訴え

PAV protocolSupplemental Digital Content Figure 2a Proportional Assist Ventilation Protocol

bull 開始の設定はサポート率70

bull 2-3時間おきにRRTがアセスメントし問題なければ10-20ずつPAVサポートを下げる

サポート率90まで上げてもダメな場合はACに戻す

PSV protocolSupplemental Digital Content Figure 2b Pressure Support Ventilation Protocol

bull 開始の設定はPS 15cmH2O

bull まずRRlt35 VTgt5mLkgで耐えられるPSまで下げる

bull 2-3時間おきにRRTがアセスメントしてPS2-3cmH2Pずつ下げる

Data collection form for RRTsSupplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

SBTの開始基準bull PF≧200bull FiO2≦05bull pH≧732bull トリガーできるbull 24時間以内に筋虚イベントなしbull 強薬昇圧剤の使なし

Data collection form for RRTs

Supplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

CPAPにしてRSBIを測定RSBI<105であればSBT施

PAV群はサポート率20PSV群はPS 5cmH2OでSBTを施

PEEP≦8cmH2OでSaO2≧90呼吸窮迫なしであればSBTクリア

抜管へ

Endpoints 呼吸器離脱までの時間ICU病院滞在期間

呼吸器離脱成功の定義抜管後48時間再挿管なし侵襲的呼吸(NIV)の使もなし気管切開使時は呼吸器から最終的に外れた時点1のうち12時間以上NIVを使していない

Statistical Analysis bull Pilot studyのためサンプルサイズの計

算はしていないbull P値<005を統計学的有意としたbull Intention to treat approachをった

RESULTS

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial

エンロール期間は42か80名が基準を満たしが最終的に割付されたのは54名

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial最終的に解析がわれたのはPAV群27名PSV群23名

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

ベースラインは両群間で変わりなし重症度はAPACHEⅡ 27点程度患者は肺炎が3割もともとCOPDを持っている患者は2割程度

Feasibility Outcomes bull 呼吸窮迫でRRTが呼吸器サポートを上昇

させた回数はPAV群で76回232daysPSV群で124回295daysと有意にPSV群で多かった(p=0002)

bull ACに戻した回数はPAV群で66回PSV群で90回と変わりなし(p=061)

bull 全くACに戻さなかった患者はPAV群で627名PSV群で623名(p=10)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

ベースラインは両群間で変わりなしPAVにすることでΔP圧Peak圧が低く吸気時間呼吸時間がくなるPSV群と較するとPAV群のがPaoAreaPaoAreaVE(呼吸器のサポートの強さをす)が低くVTPaoAreaがくsecond post-randomizationの解析においてΔP圧Peak圧が低い

Asynchrony index

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 7

B635) documenting compliance with the daily screening for SBT noncompletion of either the CPAP screening test or SBT occurred on 5 of 292 of patient study days (17) and this violation involved two patients in each study arm The second audit examining for nine different potential deviations found complete adherence to study protocol on 254 of 292 of patient study days (900) There were no adverse events (pneumo-thorax arrhythmia or death) resulting from any of the study interventions

Physiologic Response to Ventilation ProtocolsRecordings of flow and airway pressure were available for analysis for all 50 patients prerandomization 49 patients immediately (median 1 interquartile range [IQR] 0ndash15 hr) postrandomization and for 26 patients during subsequent days (median 27 IQR 20ndash70-hr postrandomization) Approximately 15000 breaths in 125 recordings were analyzed Table 2 compares the respiratory parameters in both groups pre- and postrandomization There were no statistically sig-nificant differences between the PAV and PSV groups at base-line prerandomization (p values not shown) In the PAV group there was a significant decrease in peak airway pressure delta pressure (measured as peak airway opening pressure minus the PEEP) and area under the inspiratory pressure curve immediately postrandomization and an increase in the per-cent of total cycle time spent in inspiration (TiTtot) whereas there were no significant differences immediately postran-domization in the PSV group The second measurement of postrandomization respiratory parameters revealed signifi-cant decreases in peak airway pressure and delta pressure in both PAV and PSV groups Adjusting for baseline measure-ments as a covariate the PAV group received 7 cm H2

O∙s less ventilator assistance (PaoArea) per liter of minute ventilation (p lt 00001) and increased their VT by 115 mLcm H

2O∙s

relative to the PSV group (p lt 0001)PAV significantly reduced AI postrandomization (Fig 2A)

In total nine patients (18) had a high AI (gt 10 of breaths) on PSV or on AC mode at baseline No patients had a high AI on PAV In the PSV group six patients (26) had a high AI

at baseline four (17) immediately postrandomization and one patient (7) after further adjustment of ventilator settings according to the PSV protocol (Fig 2B) Adjusting for baseline differences patients in the PAV group tended to have a lower AI than patients in the PSV group immediately postrandom-ization but this did not reach statistical significance (adjusted difference ndash45 [95 CI ndash95 to 04] p = 007)

Clinical OutcomesTable 3 shows the results for the main clinical endpoints Weaning duration and time to successful SBT did not dif-fer significantly among survivors between the groups Among ICU survivors median time to ICU discharge was shorter in the group randomized to PAV 73 days (IQR 52ndash114 d) ver-sus 124 days (IQR 75ndash308 d) in the PSV group (p = 003) Kaplan-Meier plots of the proportion of at-risk patients achiev-ing successful extubation and live ICU discharge are shown in Figure 3 A and B No difference was found for hospital length of stay ICU mortality or hospital mortality Four patients (15) in the PAV group and three patients (13) in the PSV group died in ICU (respiratory rate [RR] of PAV 114 95 CI 028 to 46 p = 086) Overall in-hospital mortality was nine patients (33) in the PAV group and six patients (26) in the PSV group (RR of PAV 128 95 CI 054 to 31 p = 058)

Cointerventions Sedating MedicationsTable 4 compares the use of medications that may prolong weaning (by causing sedation or neuromuscular weakness) in the two groups pre- and postrandomization Type and dose of sedative medication were highly variable among individual patients and therefore we analyzed pre- and postrandomiza-tion changes to determine the impact of the ventilation pro-tocol on sedative use within each patient In the PAV group benzodiazepine narcotic and propofol use measured as both cumulative dose and average daily dose all decreased signifi-cantly postrandomization as did neuromuscular blocker use In the PSV group average daily dose of lorazepam equiva-lent and average daily dose of propofol decreased signifi-cantly postrandomization whereas the proportion of patients

Figure 2 A Asynchrony index pre- and postrandomization on proportional assist ventilation B Asynchrony index pre- and postrandomization on pressure support ventilationPAV群 PSV群

Asynchrony Index () = 同調イベントの回数全呼吸回数 times 100

IntensiveCareMed(2006)321515ndash1522

AI>10は3名いたがPAVに設定変更後0名となった

AI>10は6名いたがPSVに設定変更後4名となった

Clinical Outcomes

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

成功したSBTまでの数抜管成功までの数に変わりなしICU退室までの数はPAV群で有意に短い

Kaplan-Meier curves

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

抜管成功までの時間 ICU退室までの時間

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 9

TABLE 4 Comparison of Medications That May Alter Weaning Duration Pre- and Postrandomization

Variables

PAV Group (n = 27)

PSV Group (n = 23) Postrandomization

Pre randomization Pos trandomization Pre randomization Pos trandomizationAdjusted Difference PSV ndash PAV (95 CI)

Lorazepam equivalents

Cumulative dose mg

308 (207) 48 (257)a 137 (440) 54 (258) 237 (ndash580 to 1055)

Average daily dose mg

44 (255) 09 (43)a 23 (85) 06 (40)a 24 (ndash62 to 110)

Single-day dose mgb

08 (72) 12 (50) 16 (70) 10 (64) 32 (ndash13 to 78)

Morphine equivalents

Cumulative dose mg

5003 (6188) 1140 (3140)b 1710 (13310) 1610 (4860) 591 (ndash2515 to 3696)

Average daily dose mg

858 (1638) 170 (578)c 428 (1724) 136 (320)d 108 (ndash316 to 532)

Single-day dose mga

180 (1240) 187 (77) 100 (1920) 50 (630)d 334 (ndash026 to 670)d

Propofol

Cumulative dose mg

7400 (55500) 0 (700)a 29450 (85800) 0 (27660) 226 (ndash42156 to 42607)

Average daily dose mg

529 (11690) 0 (53)a 4879 (15980) 0 (2738)a 651 (ndash3211 to 4514)

Single-day dose mga

0 (0) 0 (0) 0 (1000) 0 (0) 344 (ndash3781 to 4469)

Neuroleptics

Days receiving neuroleptics n

0 (2) 0 (4) 0 (0) 1 (4)a ndash16 (ndash34 to 03)

Proportion of patients who received a neuroleptic

407 407 174 565b

NMBs

Days receiving NMBs n

1 (1) 0 (0)a 0 (1) 0 (0) ndash01 (ndash05 to 03)

Proportion of patients who received an NMB

518 148a 348 174

Corticosteroids

Days receiving corticosteroids n

0 (0) 0 (0) 0 (2) 0 (2) 02 (ndash06 to 10)

Proportion of patients who received corticosteroids

185 148 348 261

PAV = proportional assist ventilation PSV = pressure support ventilation NMB = neuromuscular blockera p lt 005b Single-day dose = total dose given day prior to randomization or total dose give day after randomizationc p lt 0005d p lt 00005e p le 007All data points except percentage of patients receiving neuroleptics NMBs or corticosteroids are presented as median (interquartile range)

鎮静剤鎮痛剤についてはPAV群とPSV群で変わりなし

その他bull 不整脈気胸プロトコル施中の死亡

は両群とも起こらなかった

筆者らの考察bull PSVはweaningに最も使われているモードだ

がそれと較しても遜ないweaningができることがわかった

bull アシストが少ない傾向であったがPAV群のが呼吸窮迫となる回数が少なかった

rarr同調性が良くなった結果の可能性

ICU退室までの期間はPAV群で短いという結果であったがpowerが少ない

筆者らの考察bull 単施設での結果であり他の施設にこの

新しいプロトコルそのものが受けれられるかは不明である

bull SBTに失敗した患者を主に対象としておりICU患者すべてに適応はできない

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる

Page 16: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

本の論

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Critical Care Medicine wwwccmjournalorg 1

Objectives Despite protocols incorporating spontaneous breath-ing trials 31 of ICU patients experience difficult or prolonged weaning from mechanical ventilation Nonfatiguing modes such as pressure support ventilation are recommended Proportional assist ventilation provides assistance in proportion to patient effort which may optimize weaning However it is not known how proportional assist ventilation performs relative to pressure sup-port ventilation over a prolonged period in the complex ICU set-ting The purpose of this study was to compare the physiologic and clinical performance (failure rate) safety and feasibility of protocols using daily spontaneous breathing trial plus pressure support ventilation versus proportional assist ventilation until ven-tilation discontinuationDesign Single-center unblinded pilot randomized controlled trialSetting Medical-surgical ICU of a tertiary-care hospitalPatients Adult patients intubated greater than 36 hours were ran-domized if they met eligibility criteria for partial ventilatory support

tolerated pressure support ventilation greater than or equal to 30 minutes and either failed or did not meet criteria for a sponta-neous breathing trialInterventions Patients were randomized to the pressure support ventilation or proportional assist ventilation protocol (PAV+ Puri-tan Bennett 840 Covidien Boulder CO) Both protocols used progressive decreases in level of assistance as tolerated coupled with daily assessment for spontaneous breathing trialsMeasurements and Main Results Of 54 patients randomized outcome data are available for 50 patients 27 were randomized to receive proportional assist ventilation and 23 to receive pres-sure support ventilation There were no adverse events linked to the study interventions and protocol violations were infrequent Recruitment was slower than projected (13 patients per month) The median (interquartile range) time from randomization to suc-cessful extubation was 39 days (28ndash84 d) on proportional assist ventilation versus 49 days (29ndash263 d) on pressure support ventilation (p = 039) Time to live ICU discharge was 73 days (52ndash114 d) on proportional assist ventilation versus 124 days (75ndash308 d) on pressure support ventilation (p = 003)Conclusion This pilot study demonstrates the utility safety and feasibility of the weaning protocols and provides important infor-mation to guide the design of a future randomized controlled trial comparing weaning from mechanical ventilation on pressure sup-port ventilation versus proportional assist ventilation (Crit Care Med 2016 XX00ndash00)Key Words assisted mechanical ventilation asynchrony index proportional assist ventilation respiratory muscles ventilator weaning

Difficult or prolonged weaning from mechanical ventilation (MV) affects 31 (range 26ndash42) of mechanically ventilated patients (1) and increases

patient morbidity mortality ICU length of stay and cost to the healthcare system (2) When initial weaning attempts fail to liberate patients from MV ideally clinicians should choose

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights ReservedDOI 101097CCM0000000000001600

1London Health Sciences Centre London ON Canada2Department of Medicine The University of Western Ontario London ON Canada

This study was performed at London Health Sciences Centre University Hospital London CanadaSupplemental digital content is available for this article Direct URL cita-tions appear in the printed text and are provided in the HTML and PDF versions of this article on the journalrsquos website (httpjournalslwwcomccmjournal)Supported in part by Critical Care Western Department of Medicine The University of Western Ontario Program of Experimental Medi-cine (POEM) Department of Medicine University of Western Ontario Lawson Health Research Institute London Health Sciences Centre London Canada and Covidien Boulder USADr Bosma has received honourarium for lectures and travel reimburse-ments from Covidien Dr Bosmarsquos institution has received grant support from Covidien for Dr Bosmarsquos research studies The remaining authors have disclosed that they do not have any potential conflicts of interestFor information regarding this article E-mail KarenJBosmalhsconca

A Pilot Randomized Trial Comparing Weaning From Mechanical Ventilation on Pressure Support Versus Proportional Assist Ventilation

Karen J Bosma MD FRCPC12 Brooke A Read MHS BSc RRT1 Mohammad J Bahrgard Nikoo MD FRCPC12 Philip M Jones MD FRCPC MSc12 Fran A Priestap MSc1 James F Lewis MD FRCPC12

Crit Care Med 2016 Jan 20[Epub ahead of print]

論のPICO

P 呼吸器離脱困難な患者(simple weaningではない)

I PAVを使

C PSVを使

O 呼吸器離脱までの時間

Study Designbull 単施設盲検前向きランダム化較

試験(pilot study)bull カナダロンドンにあるUniversity

Hospital-London Health Science Centre の内科外科ICU(20床)

bull ランダム割付はcomputer-generated random number sequence によってわれ隠蔽化されている

PatientsInclusion criteria

bull 18歳以上の挿管患者bull 36時間以上呼吸器管理をされている

PatientsExclusion criteria

1) 48時間以内にwithdrawが考慮される患者2) 位の脊髄障害進性の神経筋疾患の

よって慢性的な呼吸器依存となる可能性があるまたは脳外科的介が必要である患者

Enrolment Procedure

Supplemental Digital Content Figure 1 Enrolment Procedure

Daily Screening of all ICU patients gt18 yrs of age Intubated gt36 hours

Exclude if Exclusion Criteria

Successful SBT awaiting extubation Withdrawal of life support considered High spinal cord injury Neurosurgical patient Neuromuscular disease

Follow using daily eligibility checklist until meet all Clinical Stability Criteria to trial mode of partial ventilatory support (PAV or PSV)

partialcomplete reversal of cause of respiratory failure

body temperature between 36ordmC and 39ordmC metabolic disorders corrected pH gt732 hemoglobin gt70 gL no ongoing bleeding stable hemodynamic status Able to trigger PaO2 ge 60 mmHg or SaO2 ge 90 on FiO2le

060 and PEEP le 15 cm H2O ACVC Pplatle30cmH2O with Vt 6-8 mlkg

OR ACPC PC+PEEPle30cmH2O

Exclude if Exclusion Criteria Has met all Criteria for Switching to

Spontaneous Mode of Ventilation gt24 hours

Pre-inclusion Test on PSV 15 cmH2O for 30 minutes

Patient fails Pre-inclusion test Return to AC ventilation and reassess within 24 hour

Patient passes Pre-inclusion test Assess for Criteria to Initiate Weaning

PaO2FiO2gt200 on FiO2 le050 and PEEP le8 cmH2O No inotropesvasopressors except low dose dopamine

RSBI lt105 Patient passes Pre-SBT readiness assessment Proceed to SBT Place patient on CPAP 5 cmH2O and FiO2 040 for 30 minutes

Patient does not meet Criteria to Initiate Weaning

Patient meets Criteria to Initiate Weaning Proceed to Pre-SBT readiness assessment Place patient on CPAP for 1-2 min measure RSBI

RSBIge 105 Patient fails Pre-SBT readiness assessment

Patient passes SBT Patient fails SBT

Exclude if Exclusion Criteria

Passed SBT and tolerating PSVle7cmH2O and PEEPle5 cmH2O while awaiting extubation

Obtain informed consent from Patient or SDM as soon as available (deferred consent permitted)

Randomize

Exclude if Exclusion Criteria

Patient SDM refuse consent

bull 呼吸不全の原因に改善が認められてきている

bull 体温は36〜39度であるbull pH>732であるbull Hb>7gdLで現在出なしbull 動態が安定しているbull トリガーできるbull FiO2≦06 PEEP≦15cmH2Oで

PaO2≧60mmHg or SaO2≧90bull ACVCでVt 6-8mlkgで

Pplat≦30cmH2OもしくはACPCでPC+PEEP≦30cmH2O

上記をすべて満たせばプレテストとしてPSV 15cmH2Oに変更

Enrolment Procedure

Supplemental Digital Content Figure 1 Enrolment Procedure

Daily Screening of all ICU patients gt18 yrs of age Intubated gt36 hours

Exclude if Exclusion Criteria

Successful SBT awaiting extubation Withdrawal of life support considered High spinal cord injury Neurosurgical patient Neuromuscular disease

Follow using daily eligibility checklist until meet all Clinical Stability Criteria to trial mode of partial ventilatory support (PAV or PSV)

partialcomplete reversal of cause of respiratory failure

body temperature between 36ordmC and 39ordmC metabolic disorders corrected pH gt732 hemoglobin gt70 gL no ongoing bleeding stable hemodynamic status Able to trigger PaO2 ge 60 mmHg or SaO2 ge 90 on FiO2le

060 and PEEP le 15 cm H2O ACVC Pplatle30cmH2O with Vt 6-8 mlkg

OR ACPC PC+PEEPle30cmH2O

Exclude if Exclusion Criteria Has met all Criteria for Switching to

Spontaneous Mode of Ventilation gt24 hours

Pre-inclusion Test on PSV 15 cmH2O for 30 minutes

Patient fails Pre-inclusion test Return to AC ventilation and reassess within 24 hour

Patient passes Pre-inclusion test Assess for Criteria to Initiate Weaning

PaO2FiO2gt200 on FiO2 le050 and PEEP le8 cmH2O No inotropesvasopressors except low dose dopamine

RSBI lt105 Patient passes Pre-SBT readiness assessment Proceed to SBT Place patient on CPAP 5 cmH2O and FiO2 040 for 30 minutes

Patient does not meet Criteria to Initiate Weaning

Patient meets Criteria to Initiate Weaning Proceed to Pre-SBT readiness assessment Place patient on CPAP for 1-2 min measure RSBI

RSBIge 105 Patient fails Pre-SBT readiness assessment

Patient passes SBT Patient fails SBT

Exclude if Exclusion Criteria

Passed SBT and tolerating PSVle7cmH2O and PEEPle5 cmH2O while awaiting extubation

Obtain informed consent from Patient or SDM as soon as available (deferred consent permitted)

Randomize

Exclude if Exclusion Criteria

Patient SDM refuse consent

呼吸窮迫がなければプレテストクリアbull FiO2≦05 PEEP≦8cmH2OでPF≧200bull 強薬昇圧剤の使なしであればSBTへ

まず1-2分CPAP 5cmH2OにしてRSBIを測定

RSBI<105であればSBTへPSV 5cmH2Oにして30-120分

酸素化の条件を満たさないRSBI≧105SBT失敗であれば割り付けへ

Ventilation Protocols bull 呼吸器はPuritan Bennett 840 を使bull 後述するPAVプロトコールPSVプロトコー

ルに従って呼吸療法(RRT)が設定を変更(respiratory distress 呼吸窮迫がないようにできるだけサポートを下げていく)

bull 設定したmaxのサポートが必要になった場合はACに戻す

bull PAV群はPAV+とACPSV群はPSVとACしか使しない

Ventilation Protocols bull RCTの開始前に15名の患者をいて観察

研究をいRRTにプロトコルの確認をってもらった

bull プロトコルはすぐにられるようにラミネートしてベッドサイドに置いた

bull RRTは毎SBTをえるかどうか後述するチェックリストをいて確認除外項を満たさない場合は毎RSBIのアセスメントとSBTをう

Respiratory Distress bull 呼吸数>35回minbull 脈拍>140min もしくはベースラインより20以上の

上昇bull 収縮期圧>180mmHg もしくは<90mmHg ベース

ラインより30以上の変化bull 不安bull 発汗bull 副呼吸筋の使bull 奇異呼吸bull 呼吸困難の訴え

PAV protocolSupplemental Digital Content Figure 2a Proportional Assist Ventilation Protocol

bull 開始の設定はサポート率70

bull 2-3時間おきにRRTがアセスメントし問題なければ10-20ずつPAVサポートを下げる

サポート率90まで上げてもダメな場合はACに戻す

PSV protocolSupplemental Digital Content Figure 2b Pressure Support Ventilation Protocol

bull 開始の設定はPS 15cmH2O

bull まずRRlt35 VTgt5mLkgで耐えられるPSまで下げる

bull 2-3時間おきにRRTがアセスメントしてPS2-3cmH2Pずつ下げる

Data collection form for RRTsSupplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

SBTの開始基準bull PF≧200bull FiO2≦05bull pH≧732bull トリガーできるbull 24時間以内に筋虚イベントなしbull 強薬昇圧剤の使なし

Data collection form for RRTs

Supplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

CPAPにしてRSBIを測定RSBI<105であればSBT施

PAV群はサポート率20PSV群はPS 5cmH2OでSBTを施

PEEP≦8cmH2OでSaO2≧90呼吸窮迫なしであればSBTクリア

抜管へ

Endpoints 呼吸器離脱までの時間ICU病院滞在期間

呼吸器離脱成功の定義抜管後48時間再挿管なし侵襲的呼吸(NIV)の使もなし気管切開使時は呼吸器から最終的に外れた時点1のうち12時間以上NIVを使していない

Statistical Analysis bull Pilot studyのためサンプルサイズの計

算はしていないbull P値<005を統計学的有意としたbull Intention to treat approachをった

RESULTS

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial

エンロール期間は42か80名が基準を満たしが最終的に割付されたのは54名

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial最終的に解析がわれたのはPAV群27名PSV群23名

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

ベースラインは両群間で変わりなし重症度はAPACHEⅡ 27点程度患者は肺炎が3割もともとCOPDを持っている患者は2割程度

Feasibility Outcomes bull 呼吸窮迫でRRTが呼吸器サポートを上昇

させた回数はPAV群で76回232daysPSV群で124回295daysと有意にPSV群で多かった(p=0002)

bull ACに戻した回数はPAV群で66回PSV群で90回と変わりなし(p=061)

bull 全くACに戻さなかった患者はPAV群で627名PSV群で623名(p=10)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

ベースラインは両群間で変わりなしPAVにすることでΔP圧Peak圧が低く吸気時間呼吸時間がくなるPSV群と較するとPAV群のがPaoAreaPaoAreaVE(呼吸器のサポートの強さをす)が低くVTPaoAreaがくsecond post-randomizationの解析においてΔP圧Peak圧が低い

Asynchrony index

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 7

B635) documenting compliance with the daily screening for SBT noncompletion of either the CPAP screening test or SBT occurred on 5 of 292 of patient study days (17) and this violation involved two patients in each study arm The second audit examining for nine different potential deviations found complete adherence to study protocol on 254 of 292 of patient study days (900) There were no adverse events (pneumo-thorax arrhythmia or death) resulting from any of the study interventions

Physiologic Response to Ventilation ProtocolsRecordings of flow and airway pressure were available for analysis for all 50 patients prerandomization 49 patients immediately (median 1 interquartile range [IQR] 0ndash15 hr) postrandomization and for 26 patients during subsequent days (median 27 IQR 20ndash70-hr postrandomization) Approximately 15000 breaths in 125 recordings were analyzed Table 2 compares the respiratory parameters in both groups pre- and postrandomization There were no statistically sig-nificant differences between the PAV and PSV groups at base-line prerandomization (p values not shown) In the PAV group there was a significant decrease in peak airway pressure delta pressure (measured as peak airway opening pressure minus the PEEP) and area under the inspiratory pressure curve immediately postrandomization and an increase in the per-cent of total cycle time spent in inspiration (TiTtot) whereas there were no significant differences immediately postran-domization in the PSV group The second measurement of postrandomization respiratory parameters revealed signifi-cant decreases in peak airway pressure and delta pressure in both PAV and PSV groups Adjusting for baseline measure-ments as a covariate the PAV group received 7 cm H2

O∙s less ventilator assistance (PaoArea) per liter of minute ventilation (p lt 00001) and increased their VT by 115 mLcm H

2O∙s

relative to the PSV group (p lt 0001)PAV significantly reduced AI postrandomization (Fig 2A)

In total nine patients (18) had a high AI (gt 10 of breaths) on PSV or on AC mode at baseline No patients had a high AI on PAV In the PSV group six patients (26) had a high AI

at baseline four (17) immediately postrandomization and one patient (7) after further adjustment of ventilator settings according to the PSV protocol (Fig 2B) Adjusting for baseline differences patients in the PAV group tended to have a lower AI than patients in the PSV group immediately postrandom-ization but this did not reach statistical significance (adjusted difference ndash45 [95 CI ndash95 to 04] p = 007)

Clinical OutcomesTable 3 shows the results for the main clinical endpoints Weaning duration and time to successful SBT did not dif-fer significantly among survivors between the groups Among ICU survivors median time to ICU discharge was shorter in the group randomized to PAV 73 days (IQR 52ndash114 d) ver-sus 124 days (IQR 75ndash308 d) in the PSV group (p = 003) Kaplan-Meier plots of the proportion of at-risk patients achiev-ing successful extubation and live ICU discharge are shown in Figure 3 A and B No difference was found for hospital length of stay ICU mortality or hospital mortality Four patients (15) in the PAV group and three patients (13) in the PSV group died in ICU (respiratory rate [RR] of PAV 114 95 CI 028 to 46 p = 086) Overall in-hospital mortality was nine patients (33) in the PAV group and six patients (26) in the PSV group (RR of PAV 128 95 CI 054 to 31 p = 058)

Cointerventions Sedating MedicationsTable 4 compares the use of medications that may prolong weaning (by causing sedation or neuromuscular weakness) in the two groups pre- and postrandomization Type and dose of sedative medication were highly variable among individual patients and therefore we analyzed pre- and postrandomiza-tion changes to determine the impact of the ventilation pro-tocol on sedative use within each patient In the PAV group benzodiazepine narcotic and propofol use measured as both cumulative dose and average daily dose all decreased signifi-cantly postrandomization as did neuromuscular blocker use In the PSV group average daily dose of lorazepam equiva-lent and average daily dose of propofol decreased signifi-cantly postrandomization whereas the proportion of patients

Figure 2 A Asynchrony index pre- and postrandomization on proportional assist ventilation B Asynchrony index pre- and postrandomization on pressure support ventilationPAV群 PSV群

Asynchrony Index () = 同調イベントの回数全呼吸回数 times 100

IntensiveCareMed(2006)321515ndash1522

AI>10は3名いたがPAVに設定変更後0名となった

AI>10は6名いたがPSVに設定変更後4名となった

Clinical Outcomes

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

成功したSBTまでの数抜管成功までの数に変わりなしICU退室までの数はPAV群で有意に短い

Kaplan-Meier curves

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

抜管成功までの時間 ICU退室までの時間

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 9

TABLE 4 Comparison of Medications That May Alter Weaning Duration Pre- and Postrandomization

Variables

PAV Group (n = 27)

PSV Group (n = 23) Postrandomization

Pre randomization Pos trandomization Pre randomization Pos trandomizationAdjusted Difference PSV ndash PAV (95 CI)

Lorazepam equivalents

Cumulative dose mg

308 (207) 48 (257)a 137 (440) 54 (258) 237 (ndash580 to 1055)

Average daily dose mg

44 (255) 09 (43)a 23 (85) 06 (40)a 24 (ndash62 to 110)

Single-day dose mgb

08 (72) 12 (50) 16 (70) 10 (64) 32 (ndash13 to 78)

Morphine equivalents

Cumulative dose mg

5003 (6188) 1140 (3140)b 1710 (13310) 1610 (4860) 591 (ndash2515 to 3696)

Average daily dose mg

858 (1638) 170 (578)c 428 (1724) 136 (320)d 108 (ndash316 to 532)

Single-day dose mga

180 (1240) 187 (77) 100 (1920) 50 (630)d 334 (ndash026 to 670)d

Propofol

Cumulative dose mg

7400 (55500) 0 (700)a 29450 (85800) 0 (27660) 226 (ndash42156 to 42607)

Average daily dose mg

529 (11690) 0 (53)a 4879 (15980) 0 (2738)a 651 (ndash3211 to 4514)

Single-day dose mga

0 (0) 0 (0) 0 (1000) 0 (0) 344 (ndash3781 to 4469)

Neuroleptics

Days receiving neuroleptics n

0 (2) 0 (4) 0 (0) 1 (4)a ndash16 (ndash34 to 03)

Proportion of patients who received a neuroleptic

407 407 174 565b

NMBs

Days receiving NMBs n

1 (1) 0 (0)a 0 (1) 0 (0) ndash01 (ndash05 to 03)

Proportion of patients who received an NMB

518 148a 348 174

Corticosteroids

Days receiving corticosteroids n

0 (0) 0 (0) 0 (2) 0 (2) 02 (ndash06 to 10)

Proportion of patients who received corticosteroids

185 148 348 261

PAV = proportional assist ventilation PSV = pressure support ventilation NMB = neuromuscular blockera p lt 005b Single-day dose = total dose given day prior to randomization or total dose give day after randomizationc p lt 0005d p lt 00005e p le 007All data points except percentage of patients receiving neuroleptics NMBs or corticosteroids are presented as median (interquartile range)

鎮静剤鎮痛剤についてはPAV群とPSV群で変わりなし

その他bull 不整脈気胸プロトコル施中の死亡

は両群とも起こらなかった

筆者らの考察bull PSVはweaningに最も使われているモードだ

がそれと較しても遜ないweaningができることがわかった

bull アシストが少ない傾向であったがPAV群のが呼吸窮迫となる回数が少なかった

rarr同調性が良くなった結果の可能性

ICU退室までの期間はPAV群で短いという結果であったがpowerが少ない

筆者らの考察bull 単施設での結果であり他の施設にこの

新しいプロトコルそのものが受けれられるかは不明である

bull SBTに失敗した患者を主に対象としておりICU患者すべてに適応はできない

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる

Page 17: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

論のPICO

P 呼吸器離脱困難な患者(simple weaningではない)

I PAVを使

C PSVを使

O 呼吸器離脱までの時間

Study Designbull 単施設盲検前向きランダム化較

試験(pilot study)bull カナダロンドンにあるUniversity

Hospital-London Health Science Centre の内科外科ICU(20床)

bull ランダム割付はcomputer-generated random number sequence によってわれ隠蔽化されている

PatientsInclusion criteria

bull 18歳以上の挿管患者bull 36時間以上呼吸器管理をされている

PatientsExclusion criteria

1) 48時間以内にwithdrawが考慮される患者2) 位の脊髄障害進性の神経筋疾患の

よって慢性的な呼吸器依存となる可能性があるまたは脳外科的介が必要である患者

Enrolment Procedure

Supplemental Digital Content Figure 1 Enrolment Procedure

Daily Screening of all ICU patients gt18 yrs of age Intubated gt36 hours

Exclude if Exclusion Criteria

Successful SBT awaiting extubation Withdrawal of life support considered High spinal cord injury Neurosurgical patient Neuromuscular disease

Follow using daily eligibility checklist until meet all Clinical Stability Criteria to trial mode of partial ventilatory support (PAV or PSV)

partialcomplete reversal of cause of respiratory failure

body temperature between 36ordmC and 39ordmC metabolic disorders corrected pH gt732 hemoglobin gt70 gL no ongoing bleeding stable hemodynamic status Able to trigger PaO2 ge 60 mmHg or SaO2 ge 90 on FiO2le

060 and PEEP le 15 cm H2O ACVC Pplatle30cmH2O with Vt 6-8 mlkg

OR ACPC PC+PEEPle30cmH2O

Exclude if Exclusion Criteria Has met all Criteria for Switching to

Spontaneous Mode of Ventilation gt24 hours

Pre-inclusion Test on PSV 15 cmH2O for 30 minutes

Patient fails Pre-inclusion test Return to AC ventilation and reassess within 24 hour

Patient passes Pre-inclusion test Assess for Criteria to Initiate Weaning

PaO2FiO2gt200 on FiO2 le050 and PEEP le8 cmH2O No inotropesvasopressors except low dose dopamine

RSBI lt105 Patient passes Pre-SBT readiness assessment Proceed to SBT Place patient on CPAP 5 cmH2O and FiO2 040 for 30 minutes

Patient does not meet Criteria to Initiate Weaning

Patient meets Criteria to Initiate Weaning Proceed to Pre-SBT readiness assessment Place patient on CPAP for 1-2 min measure RSBI

RSBIge 105 Patient fails Pre-SBT readiness assessment

Patient passes SBT Patient fails SBT

Exclude if Exclusion Criteria

Passed SBT and tolerating PSVle7cmH2O and PEEPle5 cmH2O while awaiting extubation

Obtain informed consent from Patient or SDM as soon as available (deferred consent permitted)

Randomize

Exclude if Exclusion Criteria

Patient SDM refuse consent

bull 呼吸不全の原因に改善が認められてきている

bull 体温は36〜39度であるbull pH>732であるbull Hb>7gdLで現在出なしbull 動態が安定しているbull トリガーできるbull FiO2≦06 PEEP≦15cmH2Oで

PaO2≧60mmHg or SaO2≧90bull ACVCでVt 6-8mlkgで

Pplat≦30cmH2OもしくはACPCでPC+PEEP≦30cmH2O

上記をすべて満たせばプレテストとしてPSV 15cmH2Oに変更

Enrolment Procedure

Supplemental Digital Content Figure 1 Enrolment Procedure

Daily Screening of all ICU patients gt18 yrs of age Intubated gt36 hours

Exclude if Exclusion Criteria

Successful SBT awaiting extubation Withdrawal of life support considered High spinal cord injury Neurosurgical patient Neuromuscular disease

Follow using daily eligibility checklist until meet all Clinical Stability Criteria to trial mode of partial ventilatory support (PAV or PSV)

partialcomplete reversal of cause of respiratory failure

body temperature between 36ordmC and 39ordmC metabolic disorders corrected pH gt732 hemoglobin gt70 gL no ongoing bleeding stable hemodynamic status Able to trigger PaO2 ge 60 mmHg or SaO2 ge 90 on FiO2le

060 and PEEP le 15 cm H2O ACVC Pplatle30cmH2O with Vt 6-8 mlkg

OR ACPC PC+PEEPle30cmH2O

Exclude if Exclusion Criteria Has met all Criteria for Switching to

Spontaneous Mode of Ventilation gt24 hours

Pre-inclusion Test on PSV 15 cmH2O for 30 minutes

Patient fails Pre-inclusion test Return to AC ventilation and reassess within 24 hour

Patient passes Pre-inclusion test Assess for Criteria to Initiate Weaning

PaO2FiO2gt200 on FiO2 le050 and PEEP le8 cmH2O No inotropesvasopressors except low dose dopamine

RSBI lt105 Patient passes Pre-SBT readiness assessment Proceed to SBT Place patient on CPAP 5 cmH2O and FiO2 040 for 30 minutes

Patient does not meet Criteria to Initiate Weaning

Patient meets Criteria to Initiate Weaning Proceed to Pre-SBT readiness assessment Place patient on CPAP for 1-2 min measure RSBI

RSBIge 105 Patient fails Pre-SBT readiness assessment

Patient passes SBT Patient fails SBT

Exclude if Exclusion Criteria

Passed SBT and tolerating PSVle7cmH2O and PEEPle5 cmH2O while awaiting extubation

Obtain informed consent from Patient or SDM as soon as available (deferred consent permitted)

Randomize

Exclude if Exclusion Criteria

Patient SDM refuse consent

呼吸窮迫がなければプレテストクリアbull FiO2≦05 PEEP≦8cmH2OでPF≧200bull 強薬昇圧剤の使なしであればSBTへ

まず1-2分CPAP 5cmH2OにしてRSBIを測定

RSBI<105であればSBTへPSV 5cmH2Oにして30-120分

酸素化の条件を満たさないRSBI≧105SBT失敗であれば割り付けへ

Ventilation Protocols bull 呼吸器はPuritan Bennett 840 を使bull 後述するPAVプロトコールPSVプロトコー

ルに従って呼吸療法(RRT)が設定を変更(respiratory distress 呼吸窮迫がないようにできるだけサポートを下げていく)

bull 設定したmaxのサポートが必要になった場合はACに戻す

bull PAV群はPAV+とACPSV群はPSVとACしか使しない

Ventilation Protocols bull RCTの開始前に15名の患者をいて観察

研究をいRRTにプロトコルの確認をってもらった

bull プロトコルはすぐにられるようにラミネートしてベッドサイドに置いた

bull RRTは毎SBTをえるかどうか後述するチェックリストをいて確認除外項を満たさない場合は毎RSBIのアセスメントとSBTをう

Respiratory Distress bull 呼吸数>35回minbull 脈拍>140min もしくはベースラインより20以上の

上昇bull 収縮期圧>180mmHg もしくは<90mmHg ベース

ラインより30以上の変化bull 不安bull 発汗bull 副呼吸筋の使bull 奇異呼吸bull 呼吸困難の訴え

PAV protocolSupplemental Digital Content Figure 2a Proportional Assist Ventilation Protocol

bull 開始の設定はサポート率70

bull 2-3時間おきにRRTがアセスメントし問題なければ10-20ずつPAVサポートを下げる

サポート率90まで上げてもダメな場合はACに戻す

PSV protocolSupplemental Digital Content Figure 2b Pressure Support Ventilation Protocol

bull 開始の設定はPS 15cmH2O

bull まずRRlt35 VTgt5mLkgで耐えられるPSまで下げる

bull 2-3時間おきにRRTがアセスメントしてPS2-3cmH2Pずつ下げる

Data collection form for RRTsSupplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

SBTの開始基準bull PF≧200bull FiO2≦05bull pH≧732bull トリガーできるbull 24時間以内に筋虚イベントなしbull 強薬昇圧剤の使なし

Data collection form for RRTs

Supplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

CPAPにしてRSBIを測定RSBI<105であればSBT施

PAV群はサポート率20PSV群はPS 5cmH2OでSBTを施

PEEP≦8cmH2OでSaO2≧90呼吸窮迫なしであればSBTクリア

抜管へ

Endpoints 呼吸器離脱までの時間ICU病院滞在期間

呼吸器離脱成功の定義抜管後48時間再挿管なし侵襲的呼吸(NIV)の使もなし気管切開使時は呼吸器から最終的に外れた時点1のうち12時間以上NIVを使していない

Statistical Analysis bull Pilot studyのためサンプルサイズの計

算はしていないbull P値<005を統計学的有意としたbull Intention to treat approachをった

RESULTS

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial

エンロール期間は42か80名が基準を満たしが最終的に割付されたのは54名

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial最終的に解析がわれたのはPAV群27名PSV群23名

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

ベースラインは両群間で変わりなし重症度はAPACHEⅡ 27点程度患者は肺炎が3割もともとCOPDを持っている患者は2割程度

Feasibility Outcomes bull 呼吸窮迫でRRTが呼吸器サポートを上昇

させた回数はPAV群で76回232daysPSV群で124回295daysと有意にPSV群で多かった(p=0002)

bull ACに戻した回数はPAV群で66回PSV群で90回と変わりなし(p=061)

bull 全くACに戻さなかった患者はPAV群で627名PSV群で623名(p=10)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

ベースラインは両群間で変わりなしPAVにすることでΔP圧Peak圧が低く吸気時間呼吸時間がくなるPSV群と較するとPAV群のがPaoAreaPaoAreaVE(呼吸器のサポートの強さをす)が低くVTPaoAreaがくsecond post-randomizationの解析においてΔP圧Peak圧が低い

Asynchrony index

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 7

B635) documenting compliance with the daily screening for SBT noncompletion of either the CPAP screening test or SBT occurred on 5 of 292 of patient study days (17) and this violation involved two patients in each study arm The second audit examining for nine different potential deviations found complete adherence to study protocol on 254 of 292 of patient study days (900) There were no adverse events (pneumo-thorax arrhythmia or death) resulting from any of the study interventions

Physiologic Response to Ventilation ProtocolsRecordings of flow and airway pressure were available for analysis for all 50 patients prerandomization 49 patients immediately (median 1 interquartile range [IQR] 0ndash15 hr) postrandomization and for 26 patients during subsequent days (median 27 IQR 20ndash70-hr postrandomization) Approximately 15000 breaths in 125 recordings were analyzed Table 2 compares the respiratory parameters in both groups pre- and postrandomization There were no statistically sig-nificant differences between the PAV and PSV groups at base-line prerandomization (p values not shown) In the PAV group there was a significant decrease in peak airway pressure delta pressure (measured as peak airway opening pressure minus the PEEP) and area under the inspiratory pressure curve immediately postrandomization and an increase in the per-cent of total cycle time spent in inspiration (TiTtot) whereas there were no significant differences immediately postran-domization in the PSV group The second measurement of postrandomization respiratory parameters revealed signifi-cant decreases in peak airway pressure and delta pressure in both PAV and PSV groups Adjusting for baseline measure-ments as a covariate the PAV group received 7 cm H2

O∙s less ventilator assistance (PaoArea) per liter of minute ventilation (p lt 00001) and increased their VT by 115 mLcm H

2O∙s

relative to the PSV group (p lt 0001)PAV significantly reduced AI postrandomization (Fig 2A)

In total nine patients (18) had a high AI (gt 10 of breaths) on PSV or on AC mode at baseline No patients had a high AI on PAV In the PSV group six patients (26) had a high AI

at baseline four (17) immediately postrandomization and one patient (7) after further adjustment of ventilator settings according to the PSV protocol (Fig 2B) Adjusting for baseline differences patients in the PAV group tended to have a lower AI than patients in the PSV group immediately postrandom-ization but this did not reach statistical significance (adjusted difference ndash45 [95 CI ndash95 to 04] p = 007)

Clinical OutcomesTable 3 shows the results for the main clinical endpoints Weaning duration and time to successful SBT did not dif-fer significantly among survivors between the groups Among ICU survivors median time to ICU discharge was shorter in the group randomized to PAV 73 days (IQR 52ndash114 d) ver-sus 124 days (IQR 75ndash308 d) in the PSV group (p = 003) Kaplan-Meier plots of the proportion of at-risk patients achiev-ing successful extubation and live ICU discharge are shown in Figure 3 A and B No difference was found for hospital length of stay ICU mortality or hospital mortality Four patients (15) in the PAV group and three patients (13) in the PSV group died in ICU (respiratory rate [RR] of PAV 114 95 CI 028 to 46 p = 086) Overall in-hospital mortality was nine patients (33) in the PAV group and six patients (26) in the PSV group (RR of PAV 128 95 CI 054 to 31 p = 058)

Cointerventions Sedating MedicationsTable 4 compares the use of medications that may prolong weaning (by causing sedation or neuromuscular weakness) in the two groups pre- and postrandomization Type and dose of sedative medication were highly variable among individual patients and therefore we analyzed pre- and postrandomiza-tion changes to determine the impact of the ventilation pro-tocol on sedative use within each patient In the PAV group benzodiazepine narcotic and propofol use measured as both cumulative dose and average daily dose all decreased signifi-cantly postrandomization as did neuromuscular blocker use In the PSV group average daily dose of lorazepam equiva-lent and average daily dose of propofol decreased signifi-cantly postrandomization whereas the proportion of patients

Figure 2 A Asynchrony index pre- and postrandomization on proportional assist ventilation B Asynchrony index pre- and postrandomization on pressure support ventilationPAV群 PSV群

Asynchrony Index () = 同調イベントの回数全呼吸回数 times 100

IntensiveCareMed(2006)321515ndash1522

AI>10は3名いたがPAVに設定変更後0名となった

AI>10は6名いたがPSVに設定変更後4名となった

Clinical Outcomes

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

成功したSBTまでの数抜管成功までの数に変わりなしICU退室までの数はPAV群で有意に短い

Kaplan-Meier curves

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

抜管成功までの時間 ICU退室までの時間

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 9

TABLE 4 Comparison of Medications That May Alter Weaning Duration Pre- and Postrandomization

Variables

PAV Group (n = 27)

PSV Group (n = 23) Postrandomization

Pre randomization Pos trandomization Pre randomization Pos trandomizationAdjusted Difference PSV ndash PAV (95 CI)

Lorazepam equivalents

Cumulative dose mg

308 (207) 48 (257)a 137 (440) 54 (258) 237 (ndash580 to 1055)

Average daily dose mg

44 (255) 09 (43)a 23 (85) 06 (40)a 24 (ndash62 to 110)

Single-day dose mgb

08 (72) 12 (50) 16 (70) 10 (64) 32 (ndash13 to 78)

Morphine equivalents

Cumulative dose mg

5003 (6188) 1140 (3140)b 1710 (13310) 1610 (4860) 591 (ndash2515 to 3696)

Average daily dose mg

858 (1638) 170 (578)c 428 (1724) 136 (320)d 108 (ndash316 to 532)

Single-day dose mga

180 (1240) 187 (77) 100 (1920) 50 (630)d 334 (ndash026 to 670)d

Propofol

Cumulative dose mg

7400 (55500) 0 (700)a 29450 (85800) 0 (27660) 226 (ndash42156 to 42607)

Average daily dose mg

529 (11690) 0 (53)a 4879 (15980) 0 (2738)a 651 (ndash3211 to 4514)

Single-day dose mga

0 (0) 0 (0) 0 (1000) 0 (0) 344 (ndash3781 to 4469)

Neuroleptics

Days receiving neuroleptics n

0 (2) 0 (4) 0 (0) 1 (4)a ndash16 (ndash34 to 03)

Proportion of patients who received a neuroleptic

407 407 174 565b

NMBs

Days receiving NMBs n

1 (1) 0 (0)a 0 (1) 0 (0) ndash01 (ndash05 to 03)

Proportion of patients who received an NMB

518 148a 348 174

Corticosteroids

Days receiving corticosteroids n

0 (0) 0 (0) 0 (2) 0 (2) 02 (ndash06 to 10)

Proportion of patients who received corticosteroids

185 148 348 261

PAV = proportional assist ventilation PSV = pressure support ventilation NMB = neuromuscular blockera p lt 005b Single-day dose = total dose given day prior to randomization or total dose give day after randomizationc p lt 0005d p lt 00005e p le 007All data points except percentage of patients receiving neuroleptics NMBs or corticosteroids are presented as median (interquartile range)

鎮静剤鎮痛剤についてはPAV群とPSV群で変わりなし

その他bull 不整脈気胸プロトコル施中の死亡

は両群とも起こらなかった

筆者らの考察bull PSVはweaningに最も使われているモードだ

がそれと較しても遜ないweaningができることがわかった

bull アシストが少ない傾向であったがPAV群のが呼吸窮迫となる回数が少なかった

rarr同調性が良くなった結果の可能性

ICU退室までの期間はPAV群で短いという結果であったがpowerが少ない

筆者らの考察bull 単施設での結果であり他の施設にこの

新しいプロトコルそのものが受けれられるかは不明である

bull SBTに失敗した患者を主に対象としておりICU患者すべてに適応はできない

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる

Page 18: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

Study Designbull 単施設盲検前向きランダム化較

試験(pilot study)bull カナダロンドンにあるUniversity

Hospital-London Health Science Centre の内科外科ICU(20床)

bull ランダム割付はcomputer-generated random number sequence によってわれ隠蔽化されている

PatientsInclusion criteria

bull 18歳以上の挿管患者bull 36時間以上呼吸器管理をされている

PatientsExclusion criteria

1) 48時間以内にwithdrawが考慮される患者2) 位の脊髄障害進性の神経筋疾患の

よって慢性的な呼吸器依存となる可能性があるまたは脳外科的介が必要である患者

Enrolment Procedure

Supplemental Digital Content Figure 1 Enrolment Procedure

Daily Screening of all ICU patients gt18 yrs of age Intubated gt36 hours

Exclude if Exclusion Criteria

Successful SBT awaiting extubation Withdrawal of life support considered High spinal cord injury Neurosurgical patient Neuromuscular disease

Follow using daily eligibility checklist until meet all Clinical Stability Criteria to trial mode of partial ventilatory support (PAV or PSV)

partialcomplete reversal of cause of respiratory failure

body temperature between 36ordmC and 39ordmC metabolic disorders corrected pH gt732 hemoglobin gt70 gL no ongoing bleeding stable hemodynamic status Able to trigger PaO2 ge 60 mmHg or SaO2 ge 90 on FiO2le

060 and PEEP le 15 cm H2O ACVC Pplatle30cmH2O with Vt 6-8 mlkg

OR ACPC PC+PEEPle30cmH2O

Exclude if Exclusion Criteria Has met all Criteria for Switching to

Spontaneous Mode of Ventilation gt24 hours

Pre-inclusion Test on PSV 15 cmH2O for 30 minutes

Patient fails Pre-inclusion test Return to AC ventilation and reassess within 24 hour

Patient passes Pre-inclusion test Assess for Criteria to Initiate Weaning

PaO2FiO2gt200 on FiO2 le050 and PEEP le8 cmH2O No inotropesvasopressors except low dose dopamine

RSBI lt105 Patient passes Pre-SBT readiness assessment Proceed to SBT Place patient on CPAP 5 cmH2O and FiO2 040 for 30 minutes

Patient does not meet Criteria to Initiate Weaning

Patient meets Criteria to Initiate Weaning Proceed to Pre-SBT readiness assessment Place patient on CPAP for 1-2 min measure RSBI

RSBIge 105 Patient fails Pre-SBT readiness assessment

Patient passes SBT Patient fails SBT

Exclude if Exclusion Criteria

Passed SBT and tolerating PSVle7cmH2O and PEEPle5 cmH2O while awaiting extubation

Obtain informed consent from Patient or SDM as soon as available (deferred consent permitted)

Randomize

Exclude if Exclusion Criteria

Patient SDM refuse consent

bull 呼吸不全の原因に改善が認められてきている

bull 体温は36〜39度であるbull pH>732であるbull Hb>7gdLで現在出なしbull 動態が安定しているbull トリガーできるbull FiO2≦06 PEEP≦15cmH2Oで

PaO2≧60mmHg or SaO2≧90bull ACVCでVt 6-8mlkgで

Pplat≦30cmH2OもしくはACPCでPC+PEEP≦30cmH2O

上記をすべて満たせばプレテストとしてPSV 15cmH2Oに変更

Enrolment Procedure

Supplemental Digital Content Figure 1 Enrolment Procedure

Daily Screening of all ICU patients gt18 yrs of age Intubated gt36 hours

Exclude if Exclusion Criteria

Successful SBT awaiting extubation Withdrawal of life support considered High spinal cord injury Neurosurgical patient Neuromuscular disease

Follow using daily eligibility checklist until meet all Clinical Stability Criteria to trial mode of partial ventilatory support (PAV or PSV)

partialcomplete reversal of cause of respiratory failure

body temperature between 36ordmC and 39ordmC metabolic disorders corrected pH gt732 hemoglobin gt70 gL no ongoing bleeding stable hemodynamic status Able to trigger PaO2 ge 60 mmHg or SaO2 ge 90 on FiO2le

060 and PEEP le 15 cm H2O ACVC Pplatle30cmH2O with Vt 6-8 mlkg

OR ACPC PC+PEEPle30cmH2O

Exclude if Exclusion Criteria Has met all Criteria for Switching to

Spontaneous Mode of Ventilation gt24 hours

Pre-inclusion Test on PSV 15 cmH2O for 30 minutes

Patient fails Pre-inclusion test Return to AC ventilation and reassess within 24 hour

Patient passes Pre-inclusion test Assess for Criteria to Initiate Weaning

PaO2FiO2gt200 on FiO2 le050 and PEEP le8 cmH2O No inotropesvasopressors except low dose dopamine

RSBI lt105 Patient passes Pre-SBT readiness assessment Proceed to SBT Place patient on CPAP 5 cmH2O and FiO2 040 for 30 minutes

Patient does not meet Criteria to Initiate Weaning

Patient meets Criteria to Initiate Weaning Proceed to Pre-SBT readiness assessment Place patient on CPAP for 1-2 min measure RSBI

RSBIge 105 Patient fails Pre-SBT readiness assessment

Patient passes SBT Patient fails SBT

Exclude if Exclusion Criteria

Passed SBT and tolerating PSVle7cmH2O and PEEPle5 cmH2O while awaiting extubation

Obtain informed consent from Patient or SDM as soon as available (deferred consent permitted)

Randomize

Exclude if Exclusion Criteria

Patient SDM refuse consent

呼吸窮迫がなければプレテストクリアbull FiO2≦05 PEEP≦8cmH2OでPF≧200bull 強薬昇圧剤の使なしであればSBTへ

まず1-2分CPAP 5cmH2OにしてRSBIを測定

RSBI<105であればSBTへPSV 5cmH2Oにして30-120分

酸素化の条件を満たさないRSBI≧105SBT失敗であれば割り付けへ

Ventilation Protocols bull 呼吸器はPuritan Bennett 840 を使bull 後述するPAVプロトコールPSVプロトコー

ルに従って呼吸療法(RRT)が設定を変更(respiratory distress 呼吸窮迫がないようにできるだけサポートを下げていく)

bull 設定したmaxのサポートが必要になった場合はACに戻す

bull PAV群はPAV+とACPSV群はPSVとACしか使しない

Ventilation Protocols bull RCTの開始前に15名の患者をいて観察

研究をいRRTにプロトコルの確認をってもらった

bull プロトコルはすぐにられるようにラミネートしてベッドサイドに置いた

bull RRTは毎SBTをえるかどうか後述するチェックリストをいて確認除外項を満たさない場合は毎RSBIのアセスメントとSBTをう

Respiratory Distress bull 呼吸数>35回minbull 脈拍>140min もしくはベースラインより20以上の

上昇bull 収縮期圧>180mmHg もしくは<90mmHg ベース

ラインより30以上の変化bull 不安bull 発汗bull 副呼吸筋の使bull 奇異呼吸bull 呼吸困難の訴え

PAV protocolSupplemental Digital Content Figure 2a Proportional Assist Ventilation Protocol

bull 開始の設定はサポート率70

bull 2-3時間おきにRRTがアセスメントし問題なければ10-20ずつPAVサポートを下げる

サポート率90まで上げてもダメな場合はACに戻す

PSV protocolSupplemental Digital Content Figure 2b Pressure Support Ventilation Protocol

bull 開始の設定はPS 15cmH2O

bull まずRRlt35 VTgt5mLkgで耐えられるPSまで下げる

bull 2-3時間おきにRRTがアセスメントしてPS2-3cmH2Pずつ下げる

Data collection form for RRTsSupplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

SBTの開始基準bull PF≧200bull FiO2≦05bull pH≧732bull トリガーできるbull 24時間以内に筋虚イベントなしbull 強薬昇圧剤の使なし

Data collection form for RRTs

Supplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

CPAPにしてRSBIを測定RSBI<105であればSBT施

PAV群はサポート率20PSV群はPS 5cmH2OでSBTを施

PEEP≦8cmH2OでSaO2≧90呼吸窮迫なしであればSBTクリア

抜管へ

Endpoints 呼吸器離脱までの時間ICU病院滞在期間

呼吸器離脱成功の定義抜管後48時間再挿管なし侵襲的呼吸(NIV)の使もなし気管切開使時は呼吸器から最終的に外れた時点1のうち12時間以上NIVを使していない

Statistical Analysis bull Pilot studyのためサンプルサイズの計

算はしていないbull P値<005を統計学的有意としたbull Intention to treat approachをった

RESULTS

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial

エンロール期間は42か80名が基準を満たしが最終的に割付されたのは54名

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial最終的に解析がわれたのはPAV群27名PSV群23名

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

ベースラインは両群間で変わりなし重症度はAPACHEⅡ 27点程度患者は肺炎が3割もともとCOPDを持っている患者は2割程度

Feasibility Outcomes bull 呼吸窮迫でRRTが呼吸器サポートを上昇

させた回数はPAV群で76回232daysPSV群で124回295daysと有意にPSV群で多かった(p=0002)

bull ACに戻した回数はPAV群で66回PSV群で90回と変わりなし(p=061)

bull 全くACに戻さなかった患者はPAV群で627名PSV群で623名(p=10)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

ベースラインは両群間で変わりなしPAVにすることでΔP圧Peak圧が低く吸気時間呼吸時間がくなるPSV群と較するとPAV群のがPaoAreaPaoAreaVE(呼吸器のサポートの強さをす)が低くVTPaoAreaがくsecond post-randomizationの解析においてΔP圧Peak圧が低い

Asynchrony index

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 7

B635) documenting compliance with the daily screening for SBT noncompletion of either the CPAP screening test or SBT occurred on 5 of 292 of patient study days (17) and this violation involved two patients in each study arm The second audit examining for nine different potential deviations found complete adherence to study protocol on 254 of 292 of patient study days (900) There were no adverse events (pneumo-thorax arrhythmia or death) resulting from any of the study interventions

Physiologic Response to Ventilation ProtocolsRecordings of flow and airway pressure were available for analysis for all 50 patients prerandomization 49 patients immediately (median 1 interquartile range [IQR] 0ndash15 hr) postrandomization and for 26 patients during subsequent days (median 27 IQR 20ndash70-hr postrandomization) Approximately 15000 breaths in 125 recordings were analyzed Table 2 compares the respiratory parameters in both groups pre- and postrandomization There were no statistically sig-nificant differences between the PAV and PSV groups at base-line prerandomization (p values not shown) In the PAV group there was a significant decrease in peak airway pressure delta pressure (measured as peak airway opening pressure minus the PEEP) and area under the inspiratory pressure curve immediately postrandomization and an increase in the per-cent of total cycle time spent in inspiration (TiTtot) whereas there were no significant differences immediately postran-domization in the PSV group The second measurement of postrandomization respiratory parameters revealed signifi-cant decreases in peak airway pressure and delta pressure in both PAV and PSV groups Adjusting for baseline measure-ments as a covariate the PAV group received 7 cm H2

O∙s less ventilator assistance (PaoArea) per liter of minute ventilation (p lt 00001) and increased their VT by 115 mLcm H

2O∙s

relative to the PSV group (p lt 0001)PAV significantly reduced AI postrandomization (Fig 2A)

In total nine patients (18) had a high AI (gt 10 of breaths) on PSV or on AC mode at baseline No patients had a high AI on PAV In the PSV group six patients (26) had a high AI

at baseline four (17) immediately postrandomization and one patient (7) after further adjustment of ventilator settings according to the PSV protocol (Fig 2B) Adjusting for baseline differences patients in the PAV group tended to have a lower AI than patients in the PSV group immediately postrandom-ization but this did not reach statistical significance (adjusted difference ndash45 [95 CI ndash95 to 04] p = 007)

Clinical OutcomesTable 3 shows the results for the main clinical endpoints Weaning duration and time to successful SBT did not dif-fer significantly among survivors between the groups Among ICU survivors median time to ICU discharge was shorter in the group randomized to PAV 73 days (IQR 52ndash114 d) ver-sus 124 days (IQR 75ndash308 d) in the PSV group (p = 003) Kaplan-Meier plots of the proportion of at-risk patients achiev-ing successful extubation and live ICU discharge are shown in Figure 3 A and B No difference was found for hospital length of stay ICU mortality or hospital mortality Four patients (15) in the PAV group and three patients (13) in the PSV group died in ICU (respiratory rate [RR] of PAV 114 95 CI 028 to 46 p = 086) Overall in-hospital mortality was nine patients (33) in the PAV group and six patients (26) in the PSV group (RR of PAV 128 95 CI 054 to 31 p = 058)

Cointerventions Sedating MedicationsTable 4 compares the use of medications that may prolong weaning (by causing sedation or neuromuscular weakness) in the two groups pre- and postrandomization Type and dose of sedative medication were highly variable among individual patients and therefore we analyzed pre- and postrandomiza-tion changes to determine the impact of the ventilation pro-tocol on sedative use within each patient In the PAV group benzodiazepine narcotic and propofol use measured as both cumulative dose and average daily dose all decreased signifi-cantly postrandomization as did neuromuscular blocker use In the PSV group average daily dose of lorazepam equiva-lent and average daily dose of propofol decreased signifi-cantly postrandomization whereas the proportion of patients

Figure 2 A Asynchrony index pre- and postrandomization on proportional assist ventilation B Asynchrony index pre- and postrandomization on pressure support ventilationPAV群 PSV群

Asynchrony Index () = 同調イベントの回数全呼吸回数 times 100

IntensiveCareMed(2006)321515ndash1522

AI>10は3名いたがPAVに設定変更後0名となった

AI>10は6名いたがPSVに設定変更後4名となった

Clinical Outcomes

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

成功したSBTまでの数抜管成功までの数に変わりなしICU退室までの数はPAV群で有意に短い

Kaplan-Meier curves

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

抜管成功までの時間 ICU退室までの時間

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 9

TABLE 4 Comparison of Medications That May Alter Weaning Duration Pre- and Postrandomization

Variables

PAV Group (n = 27)

PSV Group (n = 23) Postrandomization

Pre randomization Pos trandomization Pre randomization Pos trandomizationAdjusted Difference PSV ndash PAV (95 CI)

Lorazepam equivalents

Cumulative dose mg

308 (207) 48 (257)a 137 (440) 54 (258) 237 (ndash580 to 1055)

Average daily dose mg

44 (255) 09 (43)a 23 (85) 06 (40)a 24 (ndash62 to 110)

Single-day dose mgb

08 (72) 12 (50) 16 (70) 10 (64) 32 (ndash13 to 78)

Morphine equivalents

Cumulative dose mg

5003 (6188) 1140 (3140)b 1710 (13310) 1610 (4860) 591 (ndash2515 to 3696)

Average daily dose mg

858 (1638) 170 (578)c 428 (1724) 136 (320)d 108 (ndash316 to 532)

Single-day dose mga

180 (1240) 187 (77) 100 (1920) 50 (630)d 334 (ndash026 to 670)d

Propofol

Cumulative dose mg

7400 (55500) 0 (700)a 29450 (85800) 0 (27660) 226 (ndash42156 to 42607)

Average daily dose mg

529 (11690) 0 (53)a 4879 (15980) 0 (2738)a 651 (ndash3211 to 4514)

Single-day dose mga

0 (0) 0 (0) 0 (1000) 0 (0) 344 (ndash3781 to 4469)

Neuroleptics

Days receiving neuroleptics n

0 (2) 0 (4) 0 (0) 1 (4)a ndash16 (ndash34 to 03)

Proportion of patients who received a neuroleptic

407 407 174 565b

NMBs

Days receiving NMBs n

1 (1) 0 (0)a 0 (1) 0 (0) ndash01 (ndash05 to 03)

Proportion of patients who received an NMB

518 148a 348 174

Corticosteroids

Days receiving corticosteroids n

0 (0) 0 (0) 0 (2) 0 (2) 02 (ndash06 to 10)

Proportion of patients who received corticosteroids

185 148 348 261

PAV = proportional assist ventilation PSV = pressure support ventilation NMB = neuromuscular blockera p lt 005b Single-day dose = total dose given day prior to randomization or total dose give day after randomizationc p lt 0005d p lt 00005e p le 007All data points except percentage of patients receiving neuroleptics NMBs or corticosteroids are presented as median (interquartile range)

鎮静剤鎮痛剤についてはPAV群とPSV群で変わりなし

その他bull 不整脈気胸プロトコル施中の死亡

は両群とも起こらなかった

筆者らの考察bull PSVはweaningに最も使われているモードだ

がそれと較しても遜ないweaningができることがわかった

bull アシストが少ない傾向であったがPAV群のが呼吸窮迫となる回数が少なかった

rarr同調性が良くなった結果の可能性

ICU退室までの期間はPAV群で短いという結果であったがpowerが少ない

筆者らの考察bull 単施設での結果であり他の施設にこの

新しいプロトコルそのものが受けれられるかは不明である

bull SBTに失敗した患者を主に対象としておりICU患者すべてに適応はできない

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる

Page 19: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

PatientsInclusion criteria

bull 18歳以上の挿管患者bull 36時間以上呼吸器管理をされている

PatientsExclusion criteria

1) 48時間以内にwithdrawが考慮される患者2) 位の脊髄障害進性の神経筋疾患の

よって慢性的な呼吸器依存となる可能性があるまたは脳外科的介が必要である患者

Enrolment Procedure

Supplemental Digital Content Figure 1 Enrolment Procedure

Daily Screening of all ICU patients gt18 yrs of age Intubated gt36 hours

Exclude if Exclusion Criteria

Successful SBT awaiting extubation Withdrawal of life support considered High spinal cord injury Neurosurgical patient Neuromuscular disease

Follow using daily eligibility checklist until meet all Clinical Stability Criteria to trial mode of partial ventilatory support (PAV or PSV)

partialcomplete reversal of cause of respiratory failure

body temperature between 36ordmC and 39ordmC metabolic disorders corrected pH gt732 hemoglobin gt70 gL no ongoing bleeding stable hemodynamic status Able to trigger PaO2 ge 60 mmHg or SaO2 ge 90 on FiO2le

060 and PEEP le 15 cm H2O ACVC Pplatle30cmH2O with Vt 6-8 mlkg

OR ACPC PC+PEEPle30cmH2O

Exclude if Exclusion Criteria Has met all Criteria for Switching to

Spontaneous Mode of Ventilation gt24 hours

Pre-inclusion Test on PSV 15 cmH2O for 30 minutes

Patient fails Pre-inclusion test Return to AC ventilation and reassess within 24 hour

Patient passes Pre-inclusion test Assess for Criteria to Initiate Weaning

PaO2FiO2gt200 on FiO2 le050 and PEEP le8 cmH2O No inotropesvasopressors except low dose dopamine

RSBI lt105 Patient passes Pre-SBT readiness assessment Proceed to SBT Place patient on CPAP 5 cmH2O and FiO2 040 for 30 minutes

Patient does not meet Criteria to Initiate Weaning

Patient meets Criteria to Initiate Weaning Proceed to Pre-SBT readiness assessment Place patient on CPAP for 1-2 min measure RSBI

RSBIge 105 Patient fails Pre-SBT readiness assessment

Patient passes SBT Patient fails SBT

Exclude if Exclusion Criteria

Passed SBT and tolerating PSVle7cmH2O and PEEPle5 cmH2O while awaiting extubation

Obtain informed consent from Patient or SDM as soon as available (deferred consent permitted)

Randomize

Exclude if Exclusion Criteria

Patient SDM refuse consent

bull 呼吸不全の原因に改善が認められてきている

bull 体温は36〜39度であるbull pH>732であるbull Hb>7gdLで現在出なしbull 動態が安定しているbull トリガーできるbull FiO2≦06 PEEP≦15cmH2Oで

PaO2≧60mmHg or SaO2≧90bull ACVCでVt 6-8mlkgで

Pplat≦30cmH2OもしくはACPCでPC+PEEP≦30cmH2O

上記をすべて満たせばプレテストとしてPSV 15cmH2Oに変更

Enrolment Procedure

Supplemental Digital Content Figure 1 Enrolment Procedure

Daily Screening of all ICU patients gt18 yrs of age Intubated gt36 hours

Exclude if Exclusion Criteria

Successful SBT awaiting extubation Withdrawal of life support considered High spinal cord injury Neurosurgical patient Neuromuscular disease

Follow using daily eligibility checklist until meet all Clinical Stability Criteria to trial mode of partial ventilatory support (PAV or PSV)

partialcomplete reversal of cause of respiratory failure

body temperature between 36ordmC and 39ordmC metabolic disorders corrected pH gt732 hemoglobin gt70 gL no ongoing bleeding stable hemodynamic status Able to trigger PaO2 ge 60 mmHg or SaO2 ge 90 on FiO2le

060 and PEEP le 15 cm H2O ACVC Pplatle30cmH2O with Vt 6-8 mlkg

OR ACPC PC+PEEPle30cmH2O

Exclude if Exclusion Criteria Has met all Criteria for Switching to

Spontaneous Mode of Ventilation gt24 hours

Pre-inclusion Test on PSV 15 cmH2O for 30 minutes

Patient fails Pre-inclusion test Return to AC ventilation and reassess within 24 hour

Patient passes Pre-inclusion test Assess for Criteria to Initiate Weaning

PaO2FiO2gt200 on FiO2 le050 and PEEP le8 cmH2O No inotropesvasopressors except low dose dopamine

RSBI lt105 Patient passes Pre-SBT readiness assessment Proceed to SBT Place patient on CPAP 5 cmH2O and FiO2 040 for 30 minutes

Patient does not meet Criteria to Initiate Weaning

Patient meets Criteria to Initiate Weaning Proceed to Pre-SBT readiness assessment Place patient on CPAP for 1-2 min measure RSBI

RSBIge 105 Patient fails Pre-SBT readiness assessment

Patient passes SBT Patient fails SBT

Exclude if Exclusion Criteria

Passed SBT and tolerating PSVle7cmH2O and PEEPle5 cmH2O while awaiting extubation

Obtain informed consent from Patient or SDM as soon as available (deferred consent permitted)

Randomize

Exclude if Exclusion Criteria

Patient SDM refuse consent

呼吸窮迫がなければプレテストクリアbull FiO2≦05 PEEP≦8cmH2OでPF≧200bull 強薬昇圧剤の使なしであればSBTへ

まず1-2分CPAP 5cmH2OにしてRSBIを測定

RSBI<105であればSBTへPSV 5cmH2Oにして30-120分

酸素化の条件を満たさないRSBI≧105SBT失敗であれば割り付けへ

Ventilation Protocols bull 呼吸器はPuritan Bennett 840 を使bull 後述するPAVプロトコールPSVプロトコー

ルに従って呼吸療法(RRT)が設定を変更(respiratory distress 呼吸窮迫がないようにできるだけサポートを下げていく)

bull 設定したmaxのサポートが必要になった場合はACに戻す

bull PAV群はPAV+とACPSV群はPSVとACしか使しない

Ventilation Protocols bull RCTの開始前に15名の患者をいて観察

研究をいRRTにプロトコルの確認をってもらった

bull プロトコルはすぐにられるようにラミネートしてベッドサイドに置いた

bull RRTは毎SBTをえるかどうか後述するチェックリストをいて確認除外項を満たさない場合は毎RSBIのアセスメントとSBTをう

Respiratory Distress bull 呼吸数>35回minbull 脈拍>140min もしくはベースラインより20以上の

上昇bull 収縮期圧>180mmHg もしくは<90mmHg ベース

ラインより30以上の変化bull 不安bull 発汗bull 副呼吸筋の使bull 奇異呼吸bull 呼吸困難の訴え

PAV protocolSupplemental Digital Content Figure 2a Proportional Assist Ventilation Protocol

bull 開始の設定はサポート率70

bull 2-3時間おきにRRTがアセスメントし問題なければ10-20ずつPAVサポートを下げる

サポート率90まで上げてもダメな場合はACに戻す

PSV protocolSupplemental Digital Content Figure 2b Pressure Support Ventilation Protocol

bull 開始の設定はPS 15cmH2O

bull まずRRlt35 VTgt5mLkgで耐えられるPSまで下げる

bull 2-3時間おきにRRTがアセスメントしてPS2-3cmH2Pずつ下げる

Data collection form for RRTsSupplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

SBTの開始基準bull PF≧200bull FiO2≦05bull pH≧732bull トリガーできるbull 24時間以内に筋虚イベントなしbull 強薬昇圧剤の使なし

Data collection form for RRTs

Supplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

CPAPにしてRSBIを測定RSBI<105であればSBT施

PAV群はサポート率20PSV群はPS 5cmH2OでSBTを施

PEEP≦8cmH2OでSaO2≧90呼吸窮迫なしであればSBTクリア

抜管へ

Endpoints 呼吸器離脱までの時間ICU病院滞在期間

呼吸器離脱成功の定義抜管後48時間再挿管なし侵襲的呼吸(NIV)の使もなし気管切開使時は呼吸器から最終的に外れた時点1のうち12時間以上NIVを使していない

Statistical Analysis bull Pilot studyのためサンプルサイズの計

算はしていないbull P値<005を統計学的有意としたbull Intention to treat approachをった

RESULTS

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial

エンロール期間は42か80名が基準を満たしが最終的に割付されたのは54名

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial最終的に解析がわれたのはPAV群27名PSV群23名

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

ベースラインは両群間で変わりなし重症度はAPACHEⅡ 27点程度患者は肺炎が3割もともとCOPDを持っている患者は2割程度

Feasibility Outcomes bull 呼吸窮迫でRRTが呼吸器サポートを上昇

させた回数はPAV群で76回232daysPSV群で124回295daysと有意にPSV群で多かった(p=0002)

bull ACに戻した回数はPAV群で66回PSV群で90回と変わりなし(p=061)

bull 全くACに戻さなかった患者はPAV群で627名PSV群で623名(p=10)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

ベースラインは両群間で変わりなしPAVにすることでΔP圧Peak圧が低く吸気時間呼吸時間がくなるPSV群と較するとPAV群のがPaoAreaPaoAreaVE(呼吸器のサポートの強さをす)が低くVTPaoAreaがくsecond post-randomizationの解析においてΔP圧Peak圧が低い

Asynchrony index

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 7

B635) documenting compliance with the daily screening for SBT noncompletion of either the CPAP screening test or SBT occurred on 5 of 292 of patient study days (17) and this violation involved two patients in each study arm The second audit examining for nine different potential deviations found complete adherence to study protocol on 254 of 292 of patient study days (900) There were no adverse events (pneumo-thorax arrhythmia or death) resulting from any of the study interventions

Physiologic Response to Ventilation ProtocolsRecordings of flow and airway pressure were available for analysis for all 50 patients prerandomization 49 patients immediately (median 1 interquartile range [IQR] 0ndash15 hr) postrandomization and for 26 patients during subsequent days (median 27 IQR 20ndash70-hr postrandomization) Approximately 15000 breaths in 125 recordings were analyzed Table 2 compares the respiratory parameters in both groups pre- and postrandomization There were no statistically sig-nificant differences between the PAV and PSV groups at base-line prerandomization (p values not shown) In the PAV group there was a significant decrease in peak airway pressure delta pressure (measured as peak airway opening pressure minus the PEEP) and area under the inspiratory pressure curve immediately postrandomization and an increase in the per-cent of total cycle time spent in inspiration (TiTtot) whereas there were no significant differences immediately postran-domization in the PSV group The second measurement of postrandomization respiratory parameters revealed signifi-cant decreases in peak airway pressure and delta pressure in both PAV and PSV groups Adjusting for baseline measure-ments as a covariate the PAV group received 7 cm H2

O∙s less ventilator assistance (PaoArea) per liter of minute ventilation (p lt 00001) and increased their VT by 115 mLcm H

2O∙s

relative to the PSV group (p lt 0001)PAV significantly reduced AI postrandomization (Fig 2A)

In total nine patients (18) had a high AI (gt 10 of breaths) on PSV or on AC mode at baseline No patients had a high AI on PAV In the PSV group six patients (26) had a high AI

at baseline four (17) immediately postrandomization and one patient (7) after further adjustment of ventilator settings according to the PSV protocol (Fig 2B) Adjusting for baseline differences patients in the PAV group tended to have a lower AI than patients in the PSV group immediately postrandom-ization but this did not reach statistical significance (adjusted difference ndash45 [95 CI ndash95 to 04] p = 007)

Clinical OutcomesTable 3 shows the results for the main clinical endpoints Weaning duration and time to successful SBT did not dif-fer significantly among survivors between the groups Among ICU survivors median time to ICU discharge was shorter in the group randomized to PAV 73 days (IQR 52ndash114 d) ver-sus 124 days (IQR 75ndash308 d) in the PSV group (p = 003) Kaplan-Meier plots of the proportion of at-risk patients achiev-ing successful extubation and live ICU discharge are shown in Figure 3 A and B No difference was found for hospital length of stay ICU mortality or hospital mortality Four patients (15) in the PAV group and three patients (13) in the PSV group died in ICU (respiratory rate [RR] of PAV 114 95 CI 028 to 46 p = 086) Overall in-hospital mortality was nine patients (33) in the PAV group and six patients (26) in the PSV group (RR of PAV 128 95 CI 054 to 31 p = 058)

Cointerventions Sedating MedicationsTable 4 compares the use of medications that may prolong weaning (by causing sedation or neuromuscular weakness) in the two groups pre- and postrandomization Type and dose of sedative medication were highly variable among individual patients and therefore we analyzed pre- and postrandomiza-tion changes to determine the impact of the ventilation pro-tocol on sedative use within each patient In the PAV group benzodiazepine narcotic and propofol use measured as both cumulative dose and average daily dose all decreased signifi-cantly postrandomization as did neuromuscular blocker use In the PSV group average daily dose of lorazepam equiva-lent and average daily dose of propofol decreased signifi-cantly postrandomization whereas the proportion of patients

Figure 2 A Asynchrony index pre- and postrandomization on proportional assist ventilation B Asynchrony index pre- and postrandomization on pressure support ventilationPAV群 PSV群

Asynchrony Index () = 同調イベントの回数全呼吸回数 times 100

IntensiveCareMed(2006)321515ndash1522

AI>10は3名いたがPAVに設定変更後0名となった

AI>10は6名いたがPSVに設定変更後4名となった

Clinical Outcomes

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

成功したSBTまでの数抜管成功までの数に変わりなしICU退室までの数はPAV群で有意に短い

Kaplan-Meier curves

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

抜管成功までの時間 ICU退室までの時間

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 9

TABLE 4 Comparison of Medications That May Alter Weaning Duration Pre- and Postrandomization

Variables

PAV Group (n = 27)

PSV Group (n = 23) Postrandomization

Pre randomization Pos trandomization Pre randomization Pos trandomizationAdjusted Difference PSV ndash PAV (95 CI)

Lorazepam equivalents

Cumulative dose mg

308 (207) 48 (257)a 137 (440) 54 (258) 237 (ndash580 to 1055)

Average daily dose mg

44 (255) 09 (43)a 23 (85) 06 (40)a 24 (ndash62 to 110)

Single-day dose mgb

08 (72) 12 (50) 16 (70) 10 (64) 32 (ndash13 to 78)

Morphine equivalents

Cumulative dose mg

5003 (6188) 1140 (3140)b 1710 (13310) 1610 (4860) 591 (ndash2515 to 3696)

Average daily dose mg

858 (1638) 170 (578)c 428 (1724) 136 (320)d 108 (ndash316 to 532)

Single-day dose mga

180 (1240) 187 (77) 100 (1920) 50 (630)d 334 (ndash026 to 670)d

Propofol

Cumulative dose mg

7400 (55500) 0 (700)a 29450 (85800) 0 (27660) 226 (ndash42156 to 42607)

Average daily dose mg

529 (11690) 0 (53)a 4879 (15980) 0 (2738)a 651 (ndash3211 to 4514)

Single-day dose mga

0 (0) 0 (0) 0 (1000) 0 (0) 344 (ndash3781 to 4469)

Neuroleptics

Days receiving neuroleptics n

0 (2) 0 (4) 0 (0) 1 (4)a ndash16 (ndash34 to 03)

Proportion of patients who received a neuroleptic

407 407 174 565b

NMBs

Days receiving NMBs n

1 (1) 0 (0)a 0 (1) 0 (0) ndash01 (ndash05 to 03)

Proportion of patients who received an NMB

518 148a 348 174

Corticosteroids

Days receiving corticosteroids n

0 (0) 0 (0) 0 (2) 0 (2) 02 (ndash06 to 10)

Proportion of patients who received corticosteroids

185 148 348 261

PAV = proportional assist ventilation PSV = pressure support ventilation NMB = neuromuscular blockera p lt 005b Single-day dose = total dose given day prior to randomization or total dose give day after randomizationc p lt 0005d p lt 00005e p le 007All data points except percentage of patients receiving neuroleptics NMBs or corticosteroids are presented as median (interquartile range)

鎮静剤鎮痛剤についてはPAV群とPSV群で変わりなし

その他bull 不整脈気胸プロトコル施中の死亡

は両群とも起こらなかった

筆者らの考察bull PSVはweaningに最も使われているモードだ

がそれと較しても遜ないweaningができることがわかった

bull アシストが少ない傾向であったがPAV群のが呼吸窮迫となる回数が少なかった

rarr同調性が良くなった結果の可能性

ICU退室までの期間はPAV群で短いという結果であったがpowerが少ない

筆者らの考察bull 単施設での結果であり他の施設にこの

新しいプロトコルそのものが受けれられるかは不明である

bull SBTに失敗した患者を主に対象としておりICU患者すべてに適応はできない

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる

Page 20: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

PatientsExclusion criteria

1) 48時間以内にwithdrawが考慮される患者2) 位の脊髄障害進性の神経筋疾患の

よって慢性的な呼吸器依存となる可能性があるまたは脳外科的介が必要である患者

Enrolment Procedure

Supplemental Digital Content Figure 1 Enrolment Procedure

Daily Screening of all ICU patients gt18 yrs of age Intubated gt36 hours

Exclude if Exclusion Criteria

Successful SBT awaiting extubation Withdrawal of life support considered High spinal cord injury Neurosurgical patient Neuromuscular disease

Follow using daily eligibility checklist until meet all Clinical Stability Criteria to trial mode of partial ventilatory support (PAV or PSV)

partialcomplete reversal of cause of respiratory failure

body temperature between 36ordmC and 39ordmC metabolic disorders corrected pH gt732 hemoglobin gt70 gL no ongoing bleeding stable hemodynamic status Able to trigger PaO2 ge 60 mmHg or SaO2 ge 90 on FiO2le

060 and PEEP le 15 cm H2O ACVC Pplatle30cmH2O with Vt 6-8 mlkg

OR ACPC PC+PEEPle30cmH2O

Exclude if Exclusion Criteria Has met all Criteria for Switching to

Spontaneous Mode of Ventilation gt24 hours

Pre-inclusion Test on PSV 15 cmH2O for 30 minutes

Patient fails Pre-inclusion test Return to AC ventilation and reassess within 24 hour

Patient passes Pre-inclusion test Assess for Criteria to Initiate Weaning

PaO2FiO2gt200 on FiO2 le050 and PEEP le8 cmH2O No inotropesvasopressors except low dose dopamine

RSBI lt105 Patient passes Pre-SBT readiness assessment Proceed to SBT Place patient on CPAP 5 cmH2O and FiO2 040 for 30 minutes

Patient does not meet Criteria to Initiate Weaning

Patient meets Criteria to Initiate Weaning Proceed to Pre-SBT readiness assessment Place patient on CPAP for 1-2 min measure RSBI

RSBIge 105 Patient fails Pre-SBT readiness assessment

Patient passes SBT Patient fails SBT

Exclude if Exclusion Criteria

Passed SBT and tolerating PSVle7cmH2O and PEEPle5 cmH2O while awaiting extubation

Obtain informed consent from Patient or SDM as soon as available (deferred consent permitted)

Randomize

Exclude if Exclusion Criteria

Patient SDM refuse consent

bull 呼吸不全の原因に改善が認められてきている

bull 体温は36〜39度であるbull pH>732であるbull Hb>7gdLで現在出なしbull 動態が安定しているbull トリガーできるbull FiO2≦06 PEEP≦15cmH2Oで

PaO2≧60mmHg or SaO2≧90bull ACVCでVt 6-8mlkgで

Pplat≦30cmH2OもしくはACPCでPC+PEEP≦30cmH2O

上記をすべて満たせばプレテストとしてPSV 15cmH2Oに変更

Enrolment Procedure

Supplemental Digital Content Figure 1 Enrolment Procedure

Daily Screening of all ICU patients gt18 yrs of age Intubated gt36 hours

Exclude if Exclusion Criteria

Successful SBT awaiting extubation Withdrawal of life support considered High spinal cord injury Neurosurgical patient Neuromuscular disease

Follow using daily eligibility checklist until meet all Clinical Stability Criteria to trial mode of partial ventilatory support (PAV or PSV)

partialcomplete reversal of cause of respiratory failure

body temperature between 36ordmC and 39ordmC metabolic disorders corrected pH gt732 hemoglobin gt70 gL no ongoing bleeding stable hemodynamic status Able to trigger PaO2 ge 60 mmHg or SaO2 ge 90 on FiO2le

060 and PEEP le 15 cm H2O ACVC Pplatle30cmH2O with Vt 6-8 mlkg

OR ACPC PC+PEEPle30cmH2O

Exclude if Exclusion Criteria Has met all Criteria for Switching to

Spontaneous Mode of Ventilation gt24 hours

Pre-inclusion Test on PSV 15 cmH2O for 30 minutes

Patient fails Pre-inclusion test Return to AC ventilation and reassess within 24 hour

Patient passes Pre-inclusion test Assess for Criteria to Initiate Weaning

PaO2FiO2gt200 on FiO2 le050 and PEEP le8 cmH2O No inotropesvasopressors except low dose dopamine

RSBI lt105 Patient passes Pre-SBT readiness assessment Proceed to SBT Place patient on CPAP 5 cmH2O and FiO2 040 for 30 minutes

Patient does not meet Criteria to Initiate Weaning

Patient meets Criteria to Initiate Weaning Proceed to Pre-SBT readiness assessment Place patient on CPAP for 1-2 min measure RSBI

RSBIge 105 Patient fails Pre-SBT readiness assessment

Patient passes SBT Patient fails SBT

Exclude if Exclusion Criteria

Passed SBT and tolerating PSVle7cmH2O and PEEPle5 cmH2O while awaiting extubation

Obtain informed consent from Patient or SDM as soon as available (deferred consent permitted)

Randomize

Exclude if Exclusion Criteria

Patient SDM refuse consent

呼吸窮迫がなければプレテストクリアbull FiO2≦05 PEEP≦8cmH2OでPF≧200bull 強薬昇圧剤の使なしであればSBTへ

まず1-2分CPAP 5cmH2OにしてRSBIを測定

RSBI<105であればSBTへPSV 5cmH2Oにして30-120分

酸素化の条件を満たさないRSBI≧105SBT失敗であれば割り付けへ

Ventilation Protocols bull 呼吸器はPuritan Bennett 840 を使bull 後述するPAVプロトコールPSVプロトコー

ルに従って呼吸療法(RRT)が設定を変更(respiratory distress 呼吸窮迫がないようにできるだけサポートを下げていく)

bull 設定したmaxのサポートが必要になった場合はACに戻す

bull PAV群はPAV+とACPSV群はPSVとACしか使しない

Ventilation Protocols bull RCTの開始前に15名の患者をいて観察

研究をいRRTにプロトコルの確認をってもらった

bull プロトコルはすぐにられるようにラミネートしてベッドサイドに置いた

bull RRTは毎SBTをえるかどうか後述するチェックリストをいて確認除外項を満たさない場合は毎RSBIのアセスメントとSBTをう

Respiratory Distress bull 呼吸数>35回minbull 脈拍>140min もしくはベースラインより20以上の

上昇bull 収縮期圧>180mmHg もしくは<90mmHg ベース

ラインより30以上の変化bull 不安bull 発汗bull 副呼吸筋の使bull 奇異呼吸bull 呼吸困難の訴え

PAV protocolSupplemental Digital Content Figure 2a Proportional Assist Ventilation Protocol

bull 開始の設定はサポート率70

bull 2-3時間おきにRRTがアセスメントし問題なければ10-20ずつPAVサポートを下げる

サポート率90まで上げてもダメな場合はACに戻す

PSV protocolSupplemental Digital Content Figure 2b Pressure Support Ventilation Protocol

bull 開始の設定はPS 15cmH2O

bull まずRRlt35 VTgt5mLkgで耐えられるPSまで下げる

bull 2-3時間おきにRRTがアセスメントしてPS2-3cmH2Pずつ下げる

Data collection form for RRTsSupplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

SBTの開始基準bull PF≧200bull FiO2≦05bull pH≧732bull トリガーできるbull 24時間以内に筋虚イベントなしbull 強薬昇圧剤の使なし

Data collection form for RRTs

Supplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

CPAPにしてRSBIを測定RSBI<105であればSBT施

PAV群はサポート率20PSV群はPS 5cmH2OでSBTを施

PEEP≦8cmH2OでSaO2≧90呼吸窮迫なしであればSBTクリア

抜管へ

Endpoints 呼吸器離脱までの時間ICU病院滞在期間

呼吸器離脱成功の定義抜管後48時間再挿管なし侵襲的呼吸(NIV)の使もなし気管切開使時は呼吸器から最終的に外れた時点1のうち12時間以上NIVを使していない

Statistical Analysis bull Pilot studyのためサンプルサイズの計

算はしていないbull P値<005を統計学的有意としたbull Intention to treat approachをった

RESULTS

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial

エンロール期間は42か80名が基準を満たしが最終的に割付されたのは54名

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial最終的に解析がわれたのはPAV群27名PSV群23名

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

ベースラインは両群間で変わりなし重症度はAPACHEⅡ 27点程度患者は肺炎が3割もともとCOPDを持っている患者は2割程度

Feasibility Outcomes bull 呼吸窮迫でRRTが呼吸器サポートを上昇

させた回数はPAV群で76回232daysPSV群で124回295daysと有意にPSV群で多かった(p=0002)

bull ACに戻した回数はPAV群で66回PSV群で90回と変わりなし(p=061)

bull 全くACに戻さなかった患者はPAV群で627名PSV群で623名(p=10)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

ベースラインは両群間で変わりなしPAVにすることでΔP圧Peak圧が低く吸気時間呼吸時間がくなるPSV群と較するとPAV群のがPaoAreaPaoAreaVE(呼吸器のサポートの強さをす)が低くVTPaoAreaがくsecond post-randomizationの解析においてΔP圧Peak圧が低い

Asynchrony index

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 7

B635) documenting compliance with the daily screening for SBT noncompletion of either the CPAP screening test or SBT occurred on 5 of 292 of patient study days (17) and this violation involved two patients in each study arm The second audit examining for nine different potential deviations found complete adherence to study protocol on 254 of 292 of patient study days (900) There were no adverse events (pneumo-thorax arrhythmia or death) resulting from any of the study interventions

Physiologic Response to Ventilation ProtocolsRecordings of flow and airway pressure were available for analysis for all 50 patients prerandomization 49 patients immediately (median 1 interquartile range [IQR] 0ndash15 hr) postrandomization and for 26 patients during subsequent days (median 27 IQR 20ndash70-hr postrandomization) Approximately 15000 breaths in 125 recordings were analyzed Table 2 compares the respiratory parameters in both groups pre- and postrandomization There were no statistically sig-nificant differences between the PAV and PSV groups at base-line prerandomization (p values not shown) In the PAV group there was a significant decrease in peak airway pressure delta pressure (measured as peak airway opening pressure minus the PEEP) and area under the inspiratory pressure curve immediately postrandomization and an increase in the per-cent of total cycle time spent in inspiration (TiTtot) whereas there were no significant differences immediately postran-domization in the PSV group The second measurement of postrandomization respiratory parameters revealed signifi-cant decreases in peak airway pressure and delta pressure in both PAV and PSV groups Adjusting for baseline measure-ments as a covariate the PAV group received 7 cm H2

O∙s less ventilator assistance (PaoArea) per liter of minute ventilation (p lt 00001) and increased their VT by 115 mLcm H

2O∙s

relative to the PSV group (p lt 0001)PAV significantly reduced AI postrandomization (Fig 2A)

In total nine patients (18) had a high AI (gt 10 of breaths) on PSV or on AC mode at baseline No patients had a high AI on PAV In the PSV group six patients (26) had a high AI

at baseline four (17) immediately postrandomization and one patient (7) after further adjustment of ventilator settings according to the PSV protocol (Fig 2B) Adjusting for baseline differences patients in the PAV group tended to have a lower AI than patients in the PSV group immediately postrandom-ization but this did not reach statistical significance (adjusted difference ndash45 [95 CI ndash95 to 04] p = 007)

Clinical OutcomesTable 3 shows the results for the main clinical endpoints Weaning duration and time to successful SBT did not dif-fer significantly among survivors between the groups Among ICU survivors median time to ICU discharge was shorter in the group randomized to PAV 73 days (IQR 52ndash114 d) ver-sus 124 days (IQR 75ndash308 d) in the PSV group (p = 003) Kaplan-Meier plots of the proportion of at-risk patients achiev-ing successful extubation and live ICU discharge are shown in Figure 3 A and B No difference was found for hospital length of stay ICU mortality or hospital mortality Four patients (15) in the PAV group and three patients (13) in the PSV group died in ICU (respiratory rate [RR] of PAV 114 95 CI 028 to 46 p = 086) Overall in-hospital mortality was nine patients (33) in the PAV group and six patients (26) in the PSV group (RR of PAV 128 95 CI 054 to 31 p = 058)

Cointerventions Sedating MedicationsTable 4 compares the use of medications that may prolong weaning (by causing sedation or neuromuscular weakness) in the two groups pre- and postrandomization Type and dose of sedative medication were highly variable among individual patients and therefore we analyzed pre- and postrandomiza-tion changes to determine the impact of the ventilation pro-tocol on sedative use within each patient In the PAV group benzodiazepine narcotic and propofol use measured as both cumulative dose and average daily dose all decreased signifi-cantly postrandomization as did neuromuscular blocker use In the PSV group average daily dose of lorazepam equiva-lent and average daily dose of propofol decreased signifi-cantly postrandomization whereas the proportion of patients

Figure 2 A Asynchrony index pre- and postrandomization on proportional assist ventilation B Asynchrony index pre- and postrandomization on pressure support ventilationPAV群 PSV群

Asynchrony Index () = 同調イベントの回数全呼吸回数 times 100

IntensiveCareMed(2006)321515ndash1522

AI>10は3名いたがPAVに設定変更後0名となった

AI>10は6名いたがPSVに設定変更後4名となった

Clinical Outcomes

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

成功したSBTまでの数抜管成功までの数に変わりなしICU退室までの数はPAV群で有意に短い

Kaplan-Meier curves

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

抜管成功までの時間 ICU退室までの時間

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 9

TABLE 4 Comparison of Medications That May Alter Weaning Duration Pre- and Postrandomization

Variables

PAV Group (n = 27)

PSV Group (n = 23) Postrandomization

Pre randomization Pos trandomization Pre randomization Pos trandomizationAdjusted Difference PSV ndash PAV (95 CI)

Lorazepam equivalents

Cumulative dose mg

308 (207) 48 (257)a 137 (440) 54 (258) 237 (ndash580 to 1055)

Average daily dose mg

44 (255) 09 (43)a 23 (85) 06 (40)a 24 (ndash62 to 110)

Single-day dose mgb

08 (72) 12 (50) 16 (70) 10 (64) 32 (ndash13 to 78)

Morphine equivalents

Cumulative dose mg

5003 (6188) 1140 (3140)b 1710 (13310) 1610 (4860) 591 (ndash2515 to 3696)

Average daily dose mg

858 (1638) 170 (578)c 428 (1724) 136 (320)d 108 (ndash316 to 532)

Single-day dose mga

180 (1240) 187 (77) 100 (1920) 50 (630)d 334 (ndash026 to 670)d

Propofol

Cumulative dose mg

7400 (55500) 0 (700)a 29450 (85800) 0 (27660) 226 (ndash42156 to 42607)

Average daily dose mg

529 (11690) 0 (53)a 4879 (15980) 0 (2738)a 651 (ndash3211 to 4514)

Single-day dose mga

0 (0) 0 (0) 0 (1000) 0 (0) 344 (ndash3781 to 4469)

Neuroleptics

Days receiving neuroleptics n

0 (2) 0 (4) 0 (0) 1 (4)a ndash16 (ndash34 to 03)

Proportion of patients who received a neuroleptic

407 407 174 565b

NMBs

Days receiving NMBs n

1 (1) 0 (0)a 0 (1) 0 (0) ndash01 (ndash05 to 03)

Proportion of patients who received an NMB

518 148a 348 174

Corticosteroids

Days receiving corticosteroids n

0 (0) 0 (0) 0 (2) 0 (2) 02 (ndash06 to 10)

Proportion of patients who received corticosteroids

185 148 348 261

PAV = proportional assist ventilation PSV = pressure support ventilation NMB = neuromuscular blockera p lt 005b Single-day dose = total dose given day prior to randomization or total dose give day after randomizationc p lt 0005d p lt 00005e p le 007All data points except percentage of patients receiving neuroleptics NMBs or corticosteroids are presented as median (interquartile range)

鎮静剤鎮痛剤についてはPAV群とPSV群で変わりなし

その他bull 不整脈気胸プロトコル施中の死亡

は両群とも起こらなかった

筆者らの考察bull PSVはweaningに最も使われているモードだ

がそれと較しても遜ないweaningができることがわかった

bull アシストが少ない傾向であったがPAV群のが呼吸窮迫となる回数が少なかった

rarr同調性が良くなった結果の可能性

ICU退室までの期間はPAV群で短いという結果であったがpowerが少ない

筆者らの考察bull 単施設での結果であり他の施設にこの

新しいプロトコルそのものが受けれられるかは不明である

bull SBTに失敗した患者を主に対象としておりICU患者すべてに適応はできない

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる

Page 21: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

Enrolment Procedure

Supplemental Digital Content Figure 1 Enrolment Procedure

Daily Screening of all ICU patients gt18 yrs of age Intubated gt36 hours

Exclude if Exclusion Criteria

Successful SBT awaiting extubation Withdrawal of life support considered High spinal cord injury Neurosurgical patient Neuromuscular disease

Follow using daily eligibility checklist until meet all Clinical Stability Criteria to trial mode of partial ventilatory support (PAV or PSV)

partialcomplete reversal of cause of respiratory failure

body temperature between 36ordmC and 39ordmC metabolic disorders corrected pH gt732 hemoglobin gt70 gL no ongoing bleeding stable hemodynamic status Able to trigger PaO2 ge 60 mmHg or SaO2 ge 90 on FiO2le

060 and PEEP le 15 cm H2O ACVC Pplatle30cmH2O with Vt 6-8 mlkg

OR ACPC PC+PEEPle30cmH2O

Exclude if Exclusion Criteria Has met all Criteria for Switching to

Spontaneous Mode of Ventilation gt24 hours

Pre-inclusion Test on PSV 15 cmH2O for 30 minutes

Patient fails Pre-inclusion test Return to AC ventilation and reassess within 24 hour

Patient passes Pre-inclusion test Assess for Criteria to Initiate Weaning

PaO2FiO2gt200 on FiO2 le050 and PEEP le8 cmH2O No inotropesvasopressors except low dose dopamine

RSBI lt105 Patient passes Pre-SBT readiness assessment Proceed to SBT Place patient on CPAP 5 cmH2O and FiO2 040 for 30 minutes

Patient does not meet Criteria to Initiate Weaning

Patient meets Criteria to Initiate Weaning Proceed to Pre-SBT readiness assessment Place patient on CPAP for 1-2 min measure RSBI

RSBIge 105 Patient fails Pre-SBT readiness assessment

Patient passes SBT Patient fails SBT

Exclude if Exclusion Criteria

Passed SBT and tolerating PSVle7cmH2O and PEEPle5 cmH2O while awaiting extubation

Obtain informed consent from Patient or SDM as soon as available (deferred consent permitted)

Randomize

Exclude if Exclusion Criteria

Patient SDM refuse consent

bull 呼吸不全の原因に改善が認められてきている

bull 体温は36〜39度であるbull pH>732であるbull Hb>7gdLで現在出なしbull 動態が安定しているbull トリガーできるbull FiO2≦06 PEEP≦15cmH2Oで

PaO2≧60mmHg or SaO2≧90bull ACVCでVt 6-8mlkgで

Pplat≦30cmH2OもしくはACPCでPC+PEEP≦30cmH2O

上記をすべて満たせばプレテストとしてPSV 15cmH2Oに変更

Enrolment Procedure

Supplemental Digital Content Figure 1 Enrolment Procedure

Daily Screening of all ICU patients gt18 yrs of age Intubated gt36 hours

Exclude if Exclusion Criteria

Successful SBT awaiting extubation Withdrawal of life support considered High spinal cord injury Neurosurgical patient Neuromuscular disease

Follow using daily eligibility checklist until meet all Clinical Stability Criteria to trial mode of partial ventilatory support (PAV or PSV)

partialcomplete reversal of cause of respiratory failure

body temperature between 36ordmC and 39ordmC metabolic disorders corrected pH gt732 hemoglobin gt70 gL no ongoing bleeding stable hemodynamic status Able to trigger PaO2 ge 60 mmHg or SaO2 ge 90 on FiO2le

060 and PEEP le 15 cm H2O ACVC Pplatle30cmH2O with Vt 6-8 mlkg

OR ACPC PC+PEEPle30cmH2O

Exclude if Exclusion Criteria Has met all Criteria for Switching to

Spontaneous Mode of Ventilation gt24 hours

Pre-inclusion Test on PSV 15 cmH2O for 30 minutes

Patient fails Pre-inclusion test Return to AC ventilation and reassess within 24 hour

Patient passes Pre-inclusion test Assess for Criteria to Initiate Weaning

PaO2FiO2gt200 on FiO2 le050 and PEEP le8 cmH2O No inotropesvasopressors except low dose dopamine

RSBI lt105 Patient passes Pre-SBT readiness assessment Proceed to SBT Place patient on CPAP 5 cmH2O and FiO2 040 for 30 minutes

Patient does not meet Criteria to Initiate Weaning

Patient meets Criteria to Initiate Weaning Proceed to Pre-SBT readiness assessment Place patient on CPAP for 1-2 min measure RSBI

RSBIge 105 Patient fails Pre-SBT readiness assessment

Patient passes SBT Patient fails SBT

Exclude if Exclusion Criteria

Passed SBT and tolerating PSVle7cmH2O and PEEPle5 cmH2O while awaiting extubation

Obtain informed consent from Patient or SDM as soon as available (deferred consent permitted)

Randomize

Exclude if Exclusion Criteria

Patient SDM refuse consent

呼吸窮迫がなければプレテストクリアbull FiO2≦05 PEEP≦8cmH2OでPF≧200bull 強薬昇圧剤の使なしであればSBTへ

まず1-2分CPAP 5cmH2OにしてRSBIを測定

RSBI<105であればSBTへPSV 5cmH2Oにして30-120分

酸素化の条件を満たさないRSBI≧105SBT失敗であれば割り付けへ

Ventilation Protocols bull 呼吸器はPuritan Bennett 840 を使bull 後述するPAVプロトコールPSVプロトコー

ルに従って呼吸療法(RRT)が設定を変更(respiratory distress 呼吸窮迫がないようにできるだけサポートを下げていく)

bull 設定したmaxのサポートが必要になった場合はACに戻す

bull PAV群はPAV+とACPSV群はPSVとACしか使しない

Ventilation Protocols bull RCTの開始前に15名の患者をいて観察

研究をいRRTにプロトコルの確認をってもらった

bull プロトコルはすぐにられるようにラミネートしてベッドサイドに置いた

bull RRTは毎SBTをえるかどうか後述するチェックリストをいて確認除外項を満たさない場合は毎RSBIのアセスメントとSBTをう

Respiratory Distress bull 呼吸数>35回minbull 脈拍>140min もしくはベースラインより20以上の

上昇bull 収縮期圧>180mmHg もしくは<90mmHg ベース

ラインより30以上の変化bull 不安bull 発汗bull 副呼吸筋の使bull 奇異呼吸bull 呼吸困難の訴え

PAV protocolSupplemental Digital Content Figure 2a Proportional Assist Ventilation Protocol

bull 開始の設定はサポート率70

bull 2-3時間おきにRRTがアセスメントし問題なければ10-20ずつPAVサポートを下げる

サポート率90まで上げてもダメな場合はACに戻す

PSV protocolSupplemental Digital Content Figure 2b Pressure Support Ventilation Protocol

bull 開始の設定はPS 15cmH2O

bull まずRRlt35 VTgt5mLkgで耐えられるPSまで下げる

bull 2-3時間おきにRRTがアセスメントしてPS2-3cmH2Pずつ下げる

Data collection form for RRTsSupplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

SBTの開始基準bull PF≧200bull FiO2≦05bull pH≧732bull トリガーできるbull 24時間以内に筋虚イベントなしbull 強薬昇圧剤の使なし

Data collection form for RRTs

Supplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

CPAPにしてRSBIを測定RSBI<105であればSBT施

PAV群はサポート率20PSV群はPS 5cmH2OでSBTを施

PEEP≦8cmH2OでSaO2≧90呼吸窮迫なしであればSBTクリア

抜管へ

Endpoints 呼吸器離脱までの時間ICU病院滞在期間

呼吸器離脱成功の定義抜管後48時間再挿管なし侵襲的呼吸(NIV)の使もなし気管切開使時は呼吸器から最終的に外れた時点1のうち12時間以上NIVを使していない

Statistical Analysis bull Pilot studyのためサンプルサイズの計

算はしていないbull P値<005を統計学的有意としたbull Intention to treat approachをった

RESULTS

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial

エンロール期間は42か80名が基準を満たしが最終的に割付されたのは54名

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial最終的に解析がわれたのはPAV群27名PSV群23名

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

ベースラインは両群間で変わりなし重症度はAPACHEⅡ 27点程度患者は肺炎が3割もともとCOPDを持っている患者は2割程度

Feasibility Outcomes bull 呼吸窮迫でRRTが呼吸器サポートを上昇

させた回数はPAV群で76回232daysPSV群で124回295daysと有意にPSV群で多かった(p=0002)

bull ACに戻した回数はPAV群で66回PSV群で90回と変わりなし(p=061)

bull 全くACに戻さなかった患者はPAV群で627名PSV群で623名(p=10)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

ベースラインは両群間で変わりなしPAVにすることでΔP圧Peak圧が低く吸気時間呼吸時間がくなるPSV群と較するとPAV群のがPaoAreaPaoAreaVE(呼吸器のサポートの強さをす)が低くVTPaoAreaがくsecond post-randomizationの解析においてΔP圧Peak圧が低い

Asynchrony index

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 7

B635) documenting compliance with the daily screening for SBT noncompletion of either the CPAP screening test or SBT occurred on 5 of 292 of patient study days (17) and this violation involved two patients in each study arm The second audit examining for nine different potential deviations found complete adherence to study protocol on 254 of 292 of patient study days (900) There were no adverse events (pneumo-thorax arrhythmia or death) resulting from any of the study interventions

Physiologic Response to Ventilation ProtocolsRecordings of flow and airway pressure were available for analysis for all 50 patients prerandomization 49 patients immediately (median 1 interquartile range [IQR] 0ndash15 hr) postrandomization and for 26 patients during subsequent days (median 27 IQR 20ndash70-hr postrandomization) Approximately 15000 breaths in 125 recordings were analyzed Table 2 compares the respiratory parameters in both groups pre- and postrandomization There were no statistically sig-nificant differences between the PAV and PSV groups at base-line prerandomization (p values not shown) In the PAV group there was a significant decrease in peak airway pressure delta pressure (measured as peak airway opening pressure minus the PEEP) and area under the inspiratory pressure curve immediately postrandomization and an increase in the per-cent of total cycle time spent in inspiration (TiTtot) whereas there were no significant differences immediately postran-domization in the PSV group The second measurement of postrandomization respiratory parameters revealed signifi-cant decreases in peak airway pressure and delta pressure in both PAV and PSV groups Adjusting for baseline measure-ments as a covariate the PAV group received 7 cm H2

O∙s less ventilator assistance (PaoArea) per liter of minute ventilation (p lt 00001) and increased their VT by 115 mLcm H

2O∙s

relative to the PSV group (p lt 0001)PAV significantly reduced AI postrandomization (Fig 2A)

In total nine patients (18) had a high AI (gt 10 of breaths) on PSV or on AC mode at baseline No patients had a high AI on PAV In the PSV group six patients (26) had a high AI

at baseline four (17) immediately postrandomization and one patient (7) after further adjustment of ventilator settings according to the PSV protocol (Fig 2B) Adjusting for baseline differences patients in the PAV group tended to have a lower AI than patients in the PSV group immediately postrandom-ization but this did not reach statistical significance (adjusted difference ndash45 [95 CI ndash95 to 04] p = 007)

Clinical OutcomesTable 3 shows the results for the main clinical endpoints Weaning duration and time to successful SBT did not dif-fer significantly among survivors between the groups Among ICU survivors median time to ICU discharge was shorter in the group randomized to PAV 73 days (IQR 52ndash114 d) ver-sus 124 days (IQR 75ndash308 d) in the PSV group (p = 003) Kaplan-Meier plots of the proportion of at-risk patients achiev-ing successful extubation and live ICU discharge are shown in Figure 3 A and B No difference was found for hospital length of stay ICU mortality or hospital mortality Four patients (15) in the PAV group and three patients (13) in the PSV group died in ICU (respiratory rate [RR] of PAV 114 95 CI 028 to 46 p = 086) Overall in-hospital mortality was nine patients (33) in the PAV group and six patients (26) in the PSV group (RR of PAV 128 95 CI 054 to 31 p = 058)

Cointerventions Sedating MedicationsTable 4 compares the use of medications that may prolong weaning (by causing sedation or neuromuscular weakness) in the two groups pre- and postrandomization Type and dose of sedative medication were highly variable among individual patients and therefore we analyzed pre- and postrandomiza-tion changes to determine the impact of the ventilation pro-tocol on sedative use within each patient In the PAV group benzodiazepine narcotic and propofol use measured as both cumulative dose and average daily dose all decreased signifi-cantly postrandomization as did neuromuscular blocker use In the PSV group average daily dose of lorazepam equiva-lent and average daily dose of propofol decreased signifi-cantly postrandomization whereas the proportion of patients

Figure 2 A Asynchrony index pre- and postrandomization on proportional assist ventilation B Asynchrony index pre- and postrandomization on pressure support ventilationPAV群 PSV群

Asynchrony Index () = 同調イベントの回数全呼吸回数 times 100

IntensiveCareMed(2006)321515ndash1522

AI>10は3名いたがPAVに設定変更後0名となった

AI>10は6名いたがPSVに設定変更後4名となった

Clinical Outcomes

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

成功したSBTまでの数抜管成功までの数に変わりなしICU退室までの数はPAV群で有意に短い

Kaplan-Meier curves

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

抜管成功までの時間 ICU退室までの時間

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 9

TABLE 4 Comparison of Medications That May Alter Weaning Duration Pre- and Postrandomization

Variables

PAV Group (n = 27)

PSV Group (n = 23) Postrandomization

Pre randomization Pos trandomization Pre randomization Pos trandomizationAdjusted Difference PSV ndash PAV (95 CI)

Lorazepam equivalents

Cumulative dose mg

308 (207) 48 (257)a 137 (440) 54 (258) 237 (ndash580 to 1055)

Average daily dose mg

44 (255) 09 (43)a 23 (85) 06 (40)a 24 (ndash62 to 110)

Single-day dose mgb

08 (72) 12 (50) 16 (70) 10 (64) 32 (ndash13 to 78)

Morphine equivalents

Cumulative dose mg

5003 (6188) 1140 (3140)b 1710 (13310) 1610 (4860) 591 (ndash2515 to 3696)

Average daily dose mg

858 (1638) 170 (578)c 428 (1724) 136 (320)d 108 (ndash316 to 532)

Single-day dose mga

180 (1240) 187 (77) 100 (1920) 50 (630)d 334 (ndash026 to 670)d

Propofol

Cumulative dose mg

7400 (55500) 0 (700)a 29450 (85800) 0 (27660) 226 (ndash42156 to 42607)

Average daily dose mg

529 (11690) 0 (53)a 4879 (15980) 0 (2738)a 651 (ndash3211 to 4514)

Single-day dose mga

0 (0) 0 (0) 0 (1000) 0 (0) 344 (ndash3781 to 4469)

Neuroleptics

Days receiving neuroleptics n

0 (2) 0 (4) 0 (0) 1 (4)a ndash16 (ndash34 to 03)

Proportion of patients who received a neuroleptic

407 407 174 565b

NMBs

Days receiving NMBs n

1 (1) 0 (0)a 0 (1) 0 (0) ndash01 (ndash05 to 03)

Proportion of patients who received an NMB

518 148a 348 174

Corticosteroids

Days receiving corticosteroids n

0 (0) 0 (0) 0 (2) 0 (2) 02 (ndash06 to 10)

Proportion of patients who received corticosteroids

185 148 348 261

PAV = proportional assist ventilation PSV = pressure support ventilation NMB = neuromuscular blockera p lt 005b Single-day dose = total dose given day prior to randomization or total dose give day after randomizationc p lt 0005d p lt 00005e p le 007All data points except percentage of patients receiving neuroleptics NMBs or corticosteroids are presented as median (interquartile range)

鎮静剤鎮痛剤についてはPAV群とPSV群で変わりなし

その他bull 不整脈気胸プロトコル施中の死亡

は両群とも起こらなかった

筆者らの考察bull PSVはweaningに最も使われているモードだ

がそれと較しても遜ないweaningができることがわかった

bull アシストが少ない傾向であったがPAV群のが呼吸窮迫となる回数が少なかった

rarr同調性が良くなった結果の可能性

ICU退室までの期間はPAV群で短いという結果であったがpowerが少ない

筆者らの考察bull 単施設での結果であり他の施設にこの

新しいプロトコルそのものが受けれられるかは不明である

bull SBTに失敗した患者を主に対象としておりICU患者すべてに適応はできない

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる

Page 22: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

Enrolment Procedure

Supplemental Digital Content Figure 1 Enrolment Procedure

Daily Screening of all ICU patients gt18 yrs of age Intubated gt36 hours

Exclude if Exclusion Criteria

Successful SBT awaiting extubation Withdrawal of life support considered High spinal cord injury Neurosurgical patient Neuromuscular disease

Follow using daily eligibility checklist until meet all Clinical Stability Criteria to trial mode of partial ventilatory support (PAV or PSV)

partialcomplete reversal of cause of respiratory failure

body temperature between 36ordmC and 39ordmC metabolic disorders corrected pH gt732 hemoglobin gt70 gL no ongoing bleeding stable hemodynamic status Able to trigger PaO2 ge 60 mmHg or SaO2 ge 90 on FiO2le

060 and PEEP le 15 cm H2O ACVC Pplatle30cmH2O with Vt 6-8 mlkg

OR ACPC PC+PEEPle30cmH2O

Exclude if Exclusion Criteria Has met all Criteria for Switching to

Spontaneous Mode of Ventilation gt24 hours

Pre-inclusion Test on PSV 15 cmH2O for 30 minutes

Patient fails Pre-inclusion test Return to AC ventilation and reassess within 24 hour

Patient passes Pre-inclusion test Assess for Criteria to Initiate Weaning

PaO2FiO2gt200 on FiO2 le050 and PEEP le8 cmH2O No inotropesvasopressors except low dose dopamine

RSBI lt105 Patient passes Pre-SBT readiness assessment Proceed to SBT Place patient on CPAP 5 cmH2O and FiO2 040 for 30 minutes

Patient does not meet Criteria to Initiate Weaning

Patient meets Criteria to Initiate Weaning Proceed to Pre-SBT readiness assessment Place patient on CPAP for 1-2 min measure RSBI

RSBIge 105 Patient fails Pre-SBT readiness assessment

Patient passes SBT Patient fails SBT

Exclude if Exclusion Criteria

Passed SBT and tolerating PSVle7cmH2O and PEEPle5 cmH2O while awaiting extubation

Obtain informed consent from Patient or SDM as soon as available (deferred consent permitted)

Randomize

Exclude if Exclusion Criteria

Patient SDM refuse consent

呼吸窮迫がなければプレテストクリアbull FiO2≦05 PEEP≦8cmH2OでPF≧200bull 強薬昇圧剤の使なしであればSBTへ

まず1-2分CPAP 5cmH2OにしてRSBIを測定

RSBI<105であればSBTへPSV 5cmH2Oにして30-120分

酸素化の条件を満たさないRSBI≧105SBT失敗であれば割り付けへ

Ventilation Protocols bull 呼吸器はPuritan Bennett 840 を使bull 後述するPAVプロトコールPSVプロトコー

ルに従って呼吸療法(RRT)が設定を変更(respiratory distress 呼吸窮迫がないようにできるだけサポートを下げていく)

bull 設定したmaxのサポートが必要になった場合はACに戻す

bull PAV群はPAV+とACPSV群はPSVとACしか使しない

Ventilation Protocols bull RCTの開始前に15名の患者をいて観察

研究をいRRTにプロトコルの確認をってもらった

bull プロトコルはすぐにられるようにラミネートしてベッドサイドに置いた

bull RRTは毎SBTをえるかどうか後述するチェックリストをいて確認除外項を満たさない場合は毎RSBIのアセスメントとSBTをう

Respiratory Distress bull 呼吸数>35回minbull 脈拍>140min もしくはベースラインより20以上の

上昇bull 収縮期圧>180mmHg もしくは<90mmHg ベース

ラインより30以上の変化bull 不安bull 発汗bull 副呼吸筋の使bull 奇異呼吸bull 呼吸困難の訴え

PAV protocolSupplemental Digital Content Figure 2a Proportional Assist Ventilation Protocol

bull 開始の設定はサポート率70

bull 2-3時間おきにRRTがアセスメントし問題なければ10-20ずつPAVサポートを下げる

サポート率90まで上げてもダメな場合はACに戻す

PSV protocolSupplemental Digital Content Figure 2b Pressure Support Ventilation Protocol

bull 開始の設定はPS 15cmH2O

bull まずRRlt35 VTgt5mLkgで耐えられるPSまで下げる

bull 2-3時間おきにRRTがアセスメントしてPS2-3cmH2Pずつ下げる

Data collection form for RRTsSupplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

SBTの開始基準bull PF≧200bull FiO2≦05bull pH≧732bull トリガーできるbull 24時間以内に筋虚イベントなしbull 強薬昇圧剤の使なし

Data collection form for RRTs

Supplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

CPAPにしてRSBIを測定RSBI<105であればSBT施

PAV群はサポート率20PSV群はPS 5cmH2OでSBTを施

PEEP≦8cmH2OでSaO2≧90呼吸窮迫なしであればSBTクリア

抜管へ

Endpoints 呼吸器離脱までの時間ICU病院滞在期間

呼吸器離脱成功の定義抜管後48時間再挿管なし侵襲的呼吸(NIV)の使もなし気管切開使時は呼吸器から最終的に外れた時点1のうち12時間以上NIVを使していない

Statistical Analysis bull Pilot studyのためサンプルサイズの計

算はしていないbull P値<005を統計学的有意としたbull Intention to treat approachをった

RESULTS

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial

エンロール期間は42か80名が基準を満たしが最終的に割付されたのは54名

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial最終的に解析がわれたのはPAV群27名PSV群23名

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

ベースラインは両群間で変わりなし重症度はAPACHEⅡ 27点程度患者は肺炎が3割もともとCOPDを持っている患者は2割程度

Feasibility Outcomes bull 呼吸窮迫でRRTが呼吸器サポートを上昇

させた回数はPAV群で76回232daysPSV群で124回295daysと有意にPSV群で多かった(p=0002)

bull ACに戻した回数はPAV群で66回PSV群で90回と変わりなし(p=061)

bull 全くACに戻さなかった患者はPAV群で627名PSV群で623名(p=10)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

ベースラインは両群間で変わりなしPAVにすることでΔP圧Peak圧が低く吸気時間呼吸時間がくなるPSV群と較するとPAV群のがPaoAreaPaoAreaVE(呼吸器のサポートの強さをす)が低くVTPaoAreaがくsecond post-randomizationの解析においてΔP圧Peak圧が低い

Asynchrony index

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 7

B635) documenting compliance with the daily screening for SBT noncompletion of either the CPAP screening test or SBT occurred on 5 of 292 of patient study days (17) and this violation involved two patients in each study arm The second audit examining for nine different potential deviations found complete adherence to study protocol on 254 of 292 of patient study days (900) There were no adverse events (pneumo-thorax arrhythmia or death) resulting from any of the study interventions

Physiologic Response to Ventilation ProtocolsRecordings of flow and airway pressure were available for analysis for all 50 patients prerandomization 49 patients immediately (median 1 interquartile range [IQR] 0ndash15 hr) postrandomization and for 26 patients during subsequent days (median 27 IQR 20ndash70-hr postrandomization) Approximately 15000 breaths in 125 recordings were analyzed Table 2 compares the respiratory parameters in both groups pre- and postrandomization There were no statistically sig-nificant differences between the PAV and PSV groups at base-line prerandomization (p values not shown) In the PAV group there was a significant decrease in peak airway pressure delta pressure (measured as peak airway opening pressure minus the PEEP) and area under the inspiratory pressure curve immediately postrandomization and an increase in the per-cent of total cycle time spent in inspiration (TiTtot) whereas there were no significant differences immediately postran-domization in the PSV group The second measurement of postrandomization respiratory parameters revealed signifi-cant decreases in peak airway pressure and delta pressure in both PAV and PSV groups Adjusting for baseline measure-ments as a covariate the PAV group received 7 cm H2

O∙s less ventilator assistance (PaoArea) per liter of minute ventilation (p lt 00001) and increased their VT by 115 mLcm H

2O∙s

relative to the PSV group (p lt 0001)PAV significantly reduced AI postrandomization (Fig 2A)

In total nine patients (18) had a high AI (gt 10 of breaths) on PSV or on AC mode at baseline No patients had a high AI on PAV In the PSV group six patients (26) had a high AI

at baseline four (17) immediately postrandomization and one patient (7) after further adjustment of ventilator settings according to the PSV protocol (Fig 2B) Adjusting for baseline differences patients in the PAV group tended to have a lower AI than patients in the PSV group immediately postrandom-ization but this did not reach statistical significance (adjusted difference ndash45 [95 CI ndash95 to 04] p = 007)

Clinical OutcomesTable 3 shows the results for the main clinical endpoints Weaning duration and time to successful SBT did not dif-fer significantly among survivors between the groups Among ICU survivors median time to ICU discharge was shorter in the group randomized to PAV 73 days (IQR 52ndash114 d) ver-sus 124 days (IQR 75ndash308 d) in the PSV group (p = 003) Kaplan-Meier plots of the proportion of at-risk patients achiev-ing successful extubation and live ICU discharge are shown in Figure 3 A and B No difference was found for hospital length of stay ICU mortality or hospital mortality Four patients (15) in the PAV group and three patients (13) in the PSV group died in ICU (respiratory rate [RR] of PAV 114 95 CI 028 to 46 p = 086) Overall in-hospital mortality was nine patients (33) in the PAV group and six patients (26) in the PSV group (RR of PAV 128 95 CI 054 to 31 p = 058)

Cointerventions Sedating MedicationsTable 4 compares the use of medications that may prolong weaning (by causing sedation or neuromuscular weakness) in the two groups pre- and postrandomization Type and dose of sedative medication were highly variable among individual patients and therefore we analyzed pre- and postrandomiza-tion changes to determine the impact of the ventilation pro-tocol on sedative use within each patient In the PAV group benzodiazepine narcotic and propofol use measured as both cumulative dose and average daily dose all decreased signifi-cantly postrandomization as did neuromuscular blocker use In the PSV group average daily dose of lorazepam equiva-lent and average daily dose of propofol decreased signifi-cantly postrandomization whereas the proportion of patients

Figure 2 A Asynchrony index pre- and postrandomization on proportional assist ventilation B Asynchrony index pre- and postrandomization on pressure support ventilationPAV群 PSV群

Asynchrony Index () = 同調イベントの回数全呼吸回数 times 100

IntensiveCareMed(2006)321515ndash1522

AI>10は3名いたがPAVに設定変更後0名となった

AI>10は6名いたがPSVに設定変更後4名となった

Clinical Outcomes

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

成功したSBTまでの数抜管成功までの数に変わりなしICU退室までの数はPAV群で有意に短い

Kaplan-Meier curves

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

抜管成功までの時間 ICU退室までの時間

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 9

TABLE 4 Comparison of Medications That May Alter Weaning Duration Pre- and Postrandomization

Variables

PAV Group (n = 27)

PSV Group (n = 23) Postrandomization

Pre randomization Pos trandomization Pre randomization Pos trandomizationAdjusted Difference PSV ndash PAV (95 CI)

Lorazepam equivalents

Cumulative dose mg

308 (207) 48 (257)a 137 (440) 54 (258) 237 (ndash580 to 1055)

Average daily dose mg

44 (255) 09 (43)a 23 (85) 06 (40)a 24 (ndash62 to 110)

Single-day dose mgb

08 (72) 12 (50) 16 (70) 10 (64) 32 (ndash13 to 78)

Morphine equivalents

Cumulative dose mg

5003 (6188) 1140 (3140)b 1710 (13310) 1610 (4860) 591 (ndash2515 to 3696)

Average daily dose mg

858 (1638) 170 (578)c 428 (1724) 136 (320)d 108 (ndash316 to 532)

Single-day dose mga

180 (1240) 187 (77) 100 (1920) 50 (630)d 334 (ndash026 to 670)d

Propofol

Cumulative dose mg

7400 (55500) 0 (700)a 29450 (85800) 0 (27660) 226 (ndash42156 to 42607)

Average daily dose mg

529 (11690) 0 (53)a 4879 (15980) 0 (2738)a 651 (ndash3211 to 4514)

Single-day dose mga

0 (0) 0 (0) 0 (1000) 0 (0) 344 (ndash3781 to 4469)

Neuroleptics

Days receiving neuroleptics n

0 (2) 0 (4) 0 (0) 1 (4)a ndash16 (ndash34 to 03)

Proportion of patients who received a neuroleptic

407 407 174 565b

NMBs

Days receiving NMBs n

1 (1) 0 (0)a 0 (1) 0 (0) ndash01 (ndash05 to 03)

Proportion of patients who received an NMB

518 148a 348 174

Corticosteroids

Days receiving corticosteroids n

0 (0) 0 (0) 0 (2) 0 (2) 02 (ndash06 to 10)

Proportion of patients who received corticosteroids

185 148 348 261

PAV = proportional assist ventilation PSV = pressure support ventilation NMB = neuromuscular blockera p lt 005b Single-day dose = total dose given day prior to randomization or total dose give day after randomizationc p lt 0005d p lt 00005e p le 007All data points except percentage of patients receiving neuroleptics NMBs or corticosteroids are presented as median (interquartile range)

鎮静剤鎮痛剤についてはPAV群とPSV群で変わりなし

その他bull 不整脈気胸プロトコル施中の死亡

は両群とも起こらなかった

筆者らの考察bull PSVはweaningに最も使われているモードだ

がそれと較しても遜ないweaningができることがわかった

bull アシストが少ない傾向であったがPAV群のが呼吸窮迫となる回数が少なかった

rarr同調性が良くなった結果の可能性

ICU退室までの期間はPAV群で短いという結果であったがpowerが少ない

筆者らの考察bull 単施設での結果であり他の施設にこの

新しいプロトコルそのものが受けれられるかは不明である

bull SBTに失敗した患者を主に対象としておりICU患者すべてに適応はできない

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる

Page 23: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

Ventilation Protocols bull 呼吸器はPuritan Bennett 840 を使bull 後述するPAVプロトコールPSVプロトコー

ルに従って呼吸療法(RRT)が設定を変更(respiratory distress 呼吸窮迫がないようにできるだけサポートを下げていく)

bull 設定したmaxのサポートが必要になった場合はACに戻す

bull PAV群はPAV+とACPSV群はPSVとACしか使しない

Ventilation Protocols bull RCTの開始前に15名の患者をいて観察

研究をいRRTにプロトコルの確認をってもらった

bull プロトコルはすぐにられるようにラミネートしてベッドサイドに置いた

bull RRTは毎SBTをえるかどうか後述するチェックリストをいて確認除外項を満たさない場合は毎RSBIのアセスメントとSBTをう

Respiratory Distress bull 呼吸数>35回minbull 脈拍>140min もしくはベースラインより20以上の

上昇bull 収縮期圧>180mmHg もしくは<90mmHg ベース

ラインより30以上の変化bull 不安bull 発汗bull 副呼吸筋の使bull 奇異呼吸bull 呼吸困難の訴え

PAV protocolSupplemental Digital Content Figure 2a Proportional Assist Ventilation Protocol

bull 開始の設定はサポート率70

bull 2-3時間おきにRRTがアセスメントし問題なければ10-20ずつPAVサポートを下げる

サポート率90まで上げてもダメな場合はACに戻す

PSV protocolSupplemental Digital Content Figure 2b Pressure Support Ventilation Protocol

bull 開始の設定はPS 15cmH2O

bull まずRRlt35 VTgt5mLkgで耐えられるPSまで下げる

bull 2-3時間おきにRRTがアセスメントしてPS2-3cmH2Pずつ下げる

Data collection form for RRTsSupplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

SBTの開始基準bull PF≧200bull FiO2≦05bull pH≧732bull トリガーできるbull 24時間以内に筋虚イベントなしbull 強薬昇圧剤の使なし

Data collection form for RRTs

Supplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

CPAPにしてRSBIを測定RSBI<105であればSBT施

PAV群はサポート率20PSV群はPS 5cmH2OでSBTを施

PEEP≦8cmH2OでSaO2≧90呼吸窮迫なしであればSBTクリア

抜管へ

Endpoints 呼吸器離脱までの時間ICU病院滞在期間

呼吸器離脱成功の定義抜管後48時間再挿管なし侵襲的呼吸(NIV)の使もなし気管切開使時は呼吸器から最終的に外れた時点1のうち12時間以上NIVを使していない

Statistical Analysis bull Pilot studyのためサンプルサイズの計

算はしていないbull P値<005を統計学的有意としたbull Intention to treat approachをった

RESULTS

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial

エンロール期間は42か80名が基準を満たしが最終的に割付されたのは54名

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial最終的に解析がわれたのはPAV群27名PSV群23名

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

ベースラインは両群間で変わりなし重症度はAPACHEⅡ 27点程度患者は肺炎が3割もともとCOPDを持っている患者は2割程度

Feasibility Outcomes bull 呼吸窮迫でRRTが呼吸器サポートを上昇

させた回数はPAV群で76回232daysPSV群で124回295daysと有意にPSV群で多かった(p=0002)

bull ACに戻した回数はPAV群で66回PSV群で90回と変わりなし(p=061)

bull 全くACに戻さなかった患者はPAV群で627名PSV群で623名(p=10)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

ベースラインは両群間で変わりなしPAVにすることでΔP圧Peak圧が低く吸気時間呼吸時間がくなるPSV群と較するとPAV群のがPaoAreaPaoAreaVE(呼吸器のサポートの強さをす)が低くVTPaoAreaがくsecond post-randomizationの解析においてΔP圧Peak圧が低い

Asynchrony index

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 7

B635) documenting compliance with the daily screening for SBT noncompletion of either the CPAP screening test or SBT occurred on 5 of 292 of patient study days (17) and this violation involved two patients in each study arm The second audit examining for nine different potential deviations found complete adherence to study protocol on 254 of 292 of patient study days (900) There were no adverse events (pneumo-thorax arrhythmia or death) resulting from any of the study interventions

Physiologic Response to Ventilation ProtocolsRecordings of flow and airway pressure were available for analysis for all 50 patients prerandomization 49 patients immediately (median 1 interquartile range [IQR] 0ndash15 hr) postrandomization and for 26 patients during subsequent days (median 27 IQR 20ndash70-hr postrandomization) Approximately 15000 breaths in 125 recordings were analyzed Table 2 compares the respiratory parameters in both groups pre- and postrandomization There were no statistically sig-nificant differences between the PAV and PSV groups at base-line prerandomization (p values not shown) In the PAV group there was a significant decrease in peak airway pressure delta pressure (measured as peak airway opening pressure minus the PEEP) and area under the inspiratory pressure curve immediately postrandomization and an increase in the per-cent of total cycle time spent in inspiration (TiTtot) whereas there were no significant differences immediately postran-domization in the PSV group The second measurement of postrandomization respiratory parameters revealed signifi-cant decreases in peak airway pressure and delta pressure in both PAV and PSV groups Adjusting for baseline measure-ments as a covariate the PAV group received 7 cm H2

O∙s less ventilator assistance (PaoArea) per liter of minute ventilation (p lt 00001) and increased their VT by 115 mLcm H

2O∙s

relative to the PSV group (p lt 0001)PAV significantly reduced AI postrandomization (Fig 2A)

In total nine patients (18) had a high AI (gt 10 of breaths) on PSV or on AC mode at baseline No patients had a high AI on PAV In the PSV group six patients (26) had a high AI

at baseline four (17) immediately postrandomization and one patient (7) after further adjustment of ventilator settings according to the PSV protocol (Fig 2B) Adjusting for baseline differences patients in the PAV group tended to have a lower AI than patients in the PSV group immediately postrandom-ization but this did not reach statistical significance (adjusted difference ndash45 [95 CI ndash95 to 04] p = 007)

Clinical OutcomesTable 3 shows the results for the main clinical endpoints Weaning duration and time to successful SBT did not dif-fer significantly among survivors between the groups Among ICU survivors median time to ICU discharge was shorter in the group randomized to PAV 73 days (IQR 52ndash114 d) ver-sus 124 days (IQR 75ndash308 d) in the PSV group (p = 003) Kaplan-Meier plots of the proportion of at-risk patients achiev-ing successful extubation and live ICU discharge are shown in Figure 3 A and B No difference was found for hospital length of stay ICU mortality or hospital mortality Four patients (15) in the PAV group and three patients (13) in the PSV group died in ICU (respiratory rate [RR] of PAV 114 95 CI 028 to 46 p = 086) Overall in-hospital mortality was nine patients (33) in the PAV group and six patients (26) in the PSV group (RR of PAV 128 95 CI 054 to 31 p = 058)

Cointerventions Sedating MedicationsTable 4 compares the use of medications that may prolong weaning (by causing sedation or neuromuscular weakness) in the two groups pre- and postrandomization Type and dose of sedative medication were highly variable among individual patients and therefore we analyzed pre- and postrandomiza-tion changes to determine the impact of the ventilation pro-tocol on sedative use within each patient In the PAV group benzodiazepine narcotic and propofol use measured as both cumulative dose and average daily dose all decreased signifi-cantly postrandomization as did neuromuscular blocker use In the PSV group average daily dose of lorazepam equiva-lent and average daily dose of propofol decreased signifi-cantly postrandomization whereas the proportion of patients

Figure 2 A Asynchrony index pre- and postrandomization on proportional assist ventilation B Asynchrony index pre- and postrandomization on pressure support ventilationPAV群 PSV群

Asynchrony Index () = 同調イベントの回数全呼吸回数 times 100

IntensiveCareMed(2006)321515ndash1522

AI>10は3名いたがPAVに設定変更後0名となった

AI>10は6名いたがPSVに設定変更後4名となった

Clinical Outcomes

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

成功したSBTまでの数抜管成功までの数に変わりなしICU退室までの数はPAV群で有意に短い

Kaplan-Meier curves

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

抜管成功までの時間 ICU退室までの時間

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 9

TABLE 4 Comparison of Medications That May Alter Weaning Duration Pre- and Postrandomization

Variables

PAV Group (n = 27)

PSV Group (n = 23) Postrandomization

Pre randomization Pos trandomization Pre randomization Pos trandomizationAdjusted Difference PSV ndash PAV (95 CI)

Lorazepam equivalents

Cumulative dose mg

308 (207) 48 (257)a 137 (440) 54 (258) 237 (ndash580 to 1055)

Average daily dose mg

44 (255) 09 (43)a 23 (85) 06 (40)a 24 (ndash62 to 110)

Single-day dose mgb

08 (72) 12 (50) 16 (70) 10 (64) 32 (ndash13 to 78)

Morphine equivalents

Cumulative dose mg

5003 (6188) 1140 (3140)b 1710 (13310) 1610 (4860) 591 (ndash2515 to 3696)

Average daily dose mg

858 (1638) 170 (578)c 428 (1724) 136 (320)d 108 (ndash316 to 532)

Single-day dose mga

180 (1240) 187 (77) 100 (1920) 50 (630)d 334 (ndash026 to 670)d

Propofol

Cumulative dose mg

7400 (55500) 0 (700)a 29450 (85800) 0 (27660) 226 (ndash42156 to 42607)

Average daily dose mg

529 (11690) 0 (53)a 4879 (15980) 0 (2738)a 651 (ndash3211 to 4514)

Single-day dose mga

0 (0) 0 (0) 0 (1000) 0 (0) 344 (ndash3781 to 4469)

Neuroleptics

Days receiving neuroleptics n

0 (2) 0 (4) 0 (0) 1 (4)a ndash16 (ndash34 to 03)

Proportion of patients who received a neuroleptic

407 407 174 565b

NMBs

Days receiving NMBs n

1 (1) 0 (0)a 0 (1) 0 (0) ndash01 (ndash05 to 03)

Proportion of patients who received an NMB

518 148a 348 174

Corticosteroids

Days receiving corticosteroids n

0 (0) 0 (0) 0 (2) 0 (2) 02 (ndash06 to 10)

Proportion of patients who received corticosteroids

185 148 348 261

PAV = proportional assist ventilation PSV = pressure support ventilation NMB = neuromuscular blockera p lt 005b Single-day dose = total dose given day prior to randomization or total dose give day after randomizationc p lt 0005d p lt 00005e p le 007All data points except percentage of patients receiving neuroleptics NMBs or corticosteroids are presented as median (interquartile range)

鎮静剤鎮痛剤についてはPAV群とPSV群で変わりなし

その他bull 不整脈気胸プロトコル施中の死亡

は両群とも起こらなかった

筆者らの考察bull PSVはweaningに最も使われているモードだ

がそれと較しても遜ないweaningができることがわかった

bull アシストが少ない傾向であったがPAV群のが呼吸窮迫となる回数が少なかった

rarr同調性が良くなった結果の可能性

ICU退室までの期間はPAV群で短いという結果であったがpowerが少ない

筆者らの考察bull 単施設での結果であり他の施設にこの

新しいプロトコルそのものが受けれられるかは不明である

bull SBTに失敗した患者を主に対象としておりICU患者すべてに適応はできない

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる

Page 24: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

Ventilation Protocols bull RCTの開始前に15名の患者をいて観察

研究をいRRTにプロトコルの確認をってもらった

bull プロトコルはすぐにられるようにラミネートしてベッドサイドに置いた

bull RRTは毎SBTをえるかどうか後述するチェックリストをいて確認除外項を満たさない場合は毎RSBIのアセスメントとSBTをう

Respiratory Distress bull 呼吸数>35回minbull 脈拍>140min もしくはベースラインより20以上の

上昇bull 収縮期圧>180mmHg もしくは<90mmHg ベース

ラインより30以上の変化bull 不安bull 発汗bull 副呼吸筋の使bull 奇異呼吸bull 呼吸困難の訴え

PAV protocolSupplemental Digital Content Figure 2a Proportional Assist Ventilation Protocol

bull 開始の設定はサポート率70

bull 2-3時間おきにRRTがアセスメントし問題なければ10-20ずつPAVサポートを下げる

サポート率90まで上げてもダメな場合はACに戻す

PSV protocolSupplemental Digital Content Figure 2b Pressure Support Ventilation Protocol

bull 開始の設定はPS 15cmH2O

bull まずRRlt35 VTgt5mLkgで耐えられるPSまで下げる

bull 2-3時間おきにRRTがアセスメントしてPS2-3cmH2Pずつ下げる

Data collection form for RRTsSupplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

SBTの開始基準bull PF≧200bull FiO2≦05bull pH≧732bull トリガーできるbull 24時間以内に筋虚イベントなしbull 強薬昇圧剤の使なし

Data collection form for RRTs

Supplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

CPAPにしてRSBIを測定RSBI<105であればSBT施

PAV群はサポート率20PSV群はPS 5cmH2OでSBTを施

PEEP≦8cmH2OでSaO2≧90呼吸窮迫なしであればSBTクリア

抜管へ

Endpoints 呼吸器離脱までの時間ICU病院滞在期間

呼吸器離脱成功の定義抜管後48時間再挿管なし侵襲的呼吸(NIV)の使もなし気管切開使時は呼吸器から最終的に外れた時点1のうち12時間以上NIVを使していない

Statistical Analysis bull Pilot studyのためサンプルサイズの計

算はしていないbull P値<005を統計学的有意としたbull Intention to treat approachをった

RESULTS

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial

エンロール期間は42か80名が基準を満たしが最終的に割付されたのは54名

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial最終的に解析がわれたのはPAV群27名PSV群23名

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

ベースラインは両群間で変わりなし重症度はAPACHEⅡ 27点程度患者は肺炎が3割もともとCOPDを持っている患者は2割程度

Feasibility Outcomes bull 呼吸窮迫でRRTが呼吸器サポートを上昇

させた回数はPAV群で76回232daysPSV群で124回295daysと有意にPSV群で多かった(p=0002)

bull ACに戻した回数はPAV群で66回PSV群で90回と変わりなし(p=061)

bull 全くACに戻さなかった患者はPAV群で627名PSV群で623名(p=10)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

ベースラインは両群間で変わりなしPAVにすることでΔP圧Peak圧が低く吸気時間呼吸時間がくなるPSV群と較するとPAV群のがPaoAreaPaoAreaVE(呼吸器のサポートの強さをす)が低くVTPaoAreaがくsecond post-randomizationの解析においてΔP圧Peak圧が低い

Asynchrony index

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 7

B635) documenting compliance with the daily screening for SBT noncompletion of either the CPAP screening test or SBT occurred on 5 of 292 of patient study days (17) and this violation involved two patients in each study arm The second audit examining for nine different potential deviations found complete adherence to study protocol on 254 of 292 of patient study days (900) There were no adverse events (pneumo-thorax arrhythmia or death) resulting from any of the study interventions

Physiologic Response to Ventilation ProtocolsRecordings of flow and airway pressure were available for analysis for all 50 patients prerandomization 49 patients immediately (median 1 interquartile range [IQR] 0ndash15 hr) postrandomization and for 26 patients during subsequent days (median 27 IQR 20ndash70-hr postrandomization) Approximately 15000 breaths in 125 recordings were analyzed Table 2 compares the respiratory parameters in both groups pre- and postrandomization There were no statistically sig-nificant differences between the PAV and PSV groups at base-line prerandomization (p values not shown) In the PAV group there was a significant decrease in peak airway pressure delta pressure (measured as peak airway opening pressure minus the PEEP) and area under the inspiratory pressure curve immediately postrandomization and an increase in the per-cent of total cycle time spent in inspiration (TiTtot) whereas there were no significant differences immediately postran-domization in the PSV group The second measurement of postrandomization respiratory parameters revealed signifi-cant decreases in peak airway pressure and delta pressure in both PAV and PSV groups Adjusting for baseline measure-ments as a covariate the PAV group received 7 cm H2

O∙s less ventilator assistance (PaoArea) per liter of minute ventilation (p lt 00001) and increased their VT by 115 mLcm H

2O∙s

relative to the PSV group (p lt 0001)PAV significantly reduced AI postrandomization (Fig 2A)

In total nine patients (18) had a high AI (gt 10 of breaths) on PSV or on AC mode at baseline No patients had a high AI on PAV In the PSV group six patients (26) had a high AI

at baseline four (17) immediately postrandomization and one patient (7) after further adjustment of ventilator settings according to the PSV protocol (Fig 2B) Adjusting for baseline differences patients in the PAV group tended to have a lower AI than patients in the PSV group immediately postrandom-ization but this did not reach statistical significance (adjusted difference ndash45 [95 CI ndash95 to 04] p = 007)

Clinical OutcomesTable 3 shows the results for the main clinical endpoints Weaning duration and time to successful SBT did not dif-fer significantly among survivors between the groups Among ICU survivors median time to ICU discharge was shorter in the group randomized to PAV 73 days (IQR 52ndash114 d) ver-sus 124 days (IQR 75ndash308 d) in the PSV group (p = 003) Kaplan-Meier plots of the proportion of at-risk patients achiev-ing successful extubation and live ICU discharge are shown in Figure 3 A and B No difference was found for hospital length of stay ICU mortality or hospital mortality Four patients (15) in the PAV group and three patients (13) in the PSV group died in ICU (respiratory rate [RR] of PAV 114 95 CI 028 to 46 p = 086) Overall in-hospital mortality was nine patients (33) in the PAV group and six patients (26) in the PSV group (RR of PAV 128 95 CI 054 to 31 p = 058)

Cointerventions Sedating MedicationsTable 4 compares the use of medications that may prolong weaning (by causing sedation or neuromuscular weakness) in the two groups pre- and postrandomization Type and dose of sedative medication were highly variable among individual patients and therefore we analyzed pre- and postrandomiza-tion changes to determine the impact of the ventilation pro-tocol on sedative use within each patient In the PAV group benzodiazepine narcotic and propofol use measured as both cumulative dose and average daily dose all decreased signifi-cantly postrandomization as did neuromuscular blocker use In the PSV group average daily dose of lorazepam equiva-lent and average daily dose of propofol decreased signifi-cantly postrandomization whereas the proportion of patients

Figure 2 A Asynchrony index pre- and postrandomization on proportional assist ventilation B Asynchrony index pre- and postrandomization on pressure support ventilationPAV群 PSV群

Asynchrony Index () = 同調イベントの回数全呼吸回数 times 100

IntensiveCareMed(2006)321515ndash1522

AI>10は3名いたがPAVに設定変更後0名となった

AI>10は6名いたがPSVに設定変更後4名となった

Clinical Outcomes

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

成功したSBTまでの数抜管成功までの数に変わりなしICU退室までの数はPAV群で有意に短い

Kaplan-Meier curves

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

抜管成功までの時間 ICU退室までの時間

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 9

TABLE 4 Comparison of Medications That May Alter Weaning Duration Pre- and Postrandomization

Variables

PAV Group (n = 27)

PSV Group (n = 23) Postrandomization

Pre randomization Pos trandomization Pre randomization Pos trandomizationAdjusted Difference PSV ndash PAV (95 CI)

Lorazepam equivalents

Cumulative dose mg

308 (207) 48 (257)a 137 (440) 54 (258) 237 (ndash580 to 1055)

Average daily dose mg

44 (255) 09 (43)a 23 (85) 06 (40)a 24 (ndash62 to 110)

Single-day dose mgb

08 (72) 12 (50) 16 (70) 10 (64) 32 (ndash13 to 78)

Morphine equivalents

Cumulative dose mg

5003 (6188) 1140 (3140)b 1710 (13310) 1610 (4860) 591 (ndash2515 to 3696)

Average daily dose mg

858 (1638) 170 (578)c 428 (1724) 136 (320)d 108 (ndash316 to 532)

Single-day dose mga

180 (1240) 187 (77) 100 (1920) 50 (630)d 334 (ndash026 to 670)d

Propofol

Cumulative dose mg

7400 (55500) 0 (700)a 29450 (85800) 0 (27660) 226 (ndash42156 to 42607)

Average daily dose mg

529 (11690) 0 (53)a 4879 (15980) 0 (2738)a 651 (ndash3211 to 4514)

Single-day dose mga

0 (0) 0 (0) 0 (1000) 0 (0) 344 (ndash3781 to 4469)

Neuroleptics

Days receiving neuroleptics n

0 (2) 0 (4) 0 (0) 1 (4)a ndash16 (ndash34 to 03)

Proportion of patients who received a neuroleptic

407 407 174 565b

NMBs

Days receiving NMBs n

1 (1) 0 (0)a 0 (1) 0 (0) ndash01 (ndash05 to 03)

Proportion of patients who received an NMB

518 148a 348 174

Corticosteroids

Days receiving corticosteroids n

0 (0) 0 (0) 0 (2) 0 (2) 02 (ndash06 to 10)

Proportion of patients who received corticosteroids

185 148 348 261

PAV = proportional assist ventilation PSV = pressure support ventilation NMB = neuromuscular blockera p lt 005b Single-day dose = total dose given day prior to randomization or total dose give day after randomizationc p lt 0005d p lt 00005e p le 007All data points except percentage of patients receiving neuroleptics NMBs or corticosteroids are presented as median (interquartile range)

鎮静剤鎮痛剤についてはPAV群とPSV群で変わりなし

その他bull 不整脈気胸プロトコル施中の死亡

は両群とも起こらなかった

筆者らの考察bull PSVはweaningに最も使われているモードだ

がそれと較しても遜ないweaningができることがわかった

bull アシストが少ない傾向であったがPAV群のが呼吸窮迫となる回数が少なかった

rarr同調性が良くなった結果の可能性

ICU退室までの期間はPAV群で短いという結果であったがpowerが少ない

筆者らの考察bull 単施設での結果であり他の施設にこの

新しいプロトコルそのものが受けれられるかは不明である

bull SBTに失敗した患者を主に対象としておりICU患者すべてに適応はできない

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる

Page 25: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

Respiratory Distress bull 呼吸数>35回minbull 脈拍>140min もしくはベースラインより20以上の

上昇bull 収縮期圧>180mmHg もしくは<90mmHg ベース

ラインより30以上の変化bull 不安bull 発汗bull 副呼吸筋の使bull 奇異呼吸bull 呼吸困難の訴え

PAV protocolSupplemental Digital Content Figure 2a Proportional Assist Ventilation Protocol

bull 開始の設定はサポート率70

bull 2-3時間おきにRRTがアセスメントし問題なければ10-20ずつPAVサポートを下げる

サポート率90まで上げてもダメな場合はACに戻す

PSV protocolSupplemental Digital Content Figure 2b Pressure Support Ventilation Protocol

bull 開始の設定はPS 15cmH2O

bull まずRRlt35 VTgt5mLkgで耐えられるPSまで下げる

bull 2-3時間おきにRRTがアセスメントしてPS2-3cmH2Pずつ下げる

Data collection form for RRTsSupplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

SBTの開始基準bull PF≧200bull FiO2≦05bull pH≧732bull トリガーできるbull 24時間以内に筋虚イベントなしbull 強薬昇圧剤の使なし

Data collection form for RRTs

Supplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

CPAPにしてRSBIを測定RSBI<105であればSBT施

PAV群はサポート率20PSV群はPS 5cmH2OでSBTを施

PEEP≦8cmH2OでSaO2≧90呼吸窮迫なしであればSBTクリア

抜管へ

Endpoints 呼吸器離脱までの時間ICU病院滞在期間

呼吸器離脱成功の定義抜管後48時間再挿管なし侵襲的呼吸(NIV)の使もなし気管切開使時は呼吸器から最終的に外れた時点1のうち12時間以上NIVを使していない

Statistical Analysis bull Pilot studyのためサンプルサイズの計

算はしていないbull P値<005を統計学的有意としたbull Intention to treat approachをった

RESULTS

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial

エンロール期間は42か80名が基準を満たしが最終的に割付されたのは54名

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial最終的に解析がわれたのはPAV群27名PSV群23名

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

ベースラインは両群間で変わりなし重症度はAPACHEⅡ 27点程度患者は肺炎が3割もともとCOPDを持っている患者は2割程度

Feasibility Outcomes bull 呼吸窮迫でRRTが呼吸器サポートを上昇

させた回数はPAV群で76回232daysPSV群で124回295daysと有意にPSV群で多かった(p=0002)

bull ACに戻した回数はPAV群で66回PSV群で90回と変わりなし(p=061)

bull 全くACに戻さなかった患者はPAV群で627名PSV群で623名(p=10)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

ベースラインは両群間で変わりなしPAVにすることでΔP圧Peak圧が低く吸気時間呼吸時間がくなるPSV群と較するとPAV群のがPaoAreaPaoAreaVE(呼吸器のサポートの強さをす)が低くVTPaoAreaがくsecond post-randomizationの解析においてΔP圧Peak圧が低い

Asynchrony index

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 7

B635) documenting compliance with the daily screening for SBT noncompletion of either the CPAP screening test or SBT occurred on 5 of 292 of patient study days (17) and this violation involved two patients in each study arm The second audit examining for nine different potential deviations found complete adherence to study protocol on 254 of 292 of patient study days (900) There were no adverse events (pneumo-thorax arrhythmia or death) resulting from any of the study interventions

Physiologic Response to Ventilation ProtocolsRecordings of flow and airway pressure were available for analysis for all 50 patients prerandomization 49 patients immediately (median 1 interquartile range [IQR] 0ndash15 hr) postrandomization and for 26 patients during subsequent days (median 27 IQR 20ndash70-hr postrandomization) Approximately 15000 breaths in 125 recordings were analyzed Table 2 compares the respiratory parameters in both groups pre- and postrandomization There were no statistically sig-nificant differences between the PAV and PSV groups at base-line prerandomization (p values not shown) In the PAV group there was a significant decrease in peak airway pressure delta pressure (measured as peak airway opening pressure minus the PEEP) and area under the inspiratory pressure curve immediately postrandomization and an increase in the per-cent of total cycle time spent in inspiration (TiTtot) whereas there were no significant differences immediately postran-domization in the PSV group The second measurement of postrandomization respiratory parameters revealed signifi-cant decreases in peak airway pressure and delta pressure in both PAV and PSV groups Adjusting for baseline measure-ments as a covariate the PAV group received 7 cm H2

O∙s less ventilator assistance (PaoArea) per liter of minute ventilation (p lt 00001) and increased their VT by 115 mLcm H

2O∙s

relative to the PSV group (p lt 0001)PAV significantly reduced AI postrandomization (Fig 2A)

In total nine patients (18) had a high AI (gt 10 of breaths) on PSV or on AC mode at baseline No patients had a high AI on PAV In the PSV group six patients (26) had a high AI

at baseline four (17) immediately postrandomization and one patient (7) after further adjustment of ventilator settings according to the PSV protocol (Fig 2B) Adjusting for baseline differences patients in the PAV group tended to have a lower AI than patients in the PSV group immediately postrandom-ization but this did not reach statistical significance (adjusted difference ndash45 [95 CI ndash95 to 04] p = 007)

Clinical OutcomesTable 3 shows the results for the main clinical endpoints Weaning duration and time to successful SBT did not dif-fer significantly among survivors between the groups Among ICU survivors median time to ICU discharge was shorter in the group randomized to PAV 73 days (IQR 52ndash114 d) ver-sus 124 days (IQR 75ndash308 d) in the PSV group (p = 003) Kaplan-Meier plots of the proportion of at-risk patients achiev-ing successful extubation and live ICU discharge are shown in Figure 3 A and B No difference was found for hospital length of stay ICU mortality or hospital mortality Four patients (15) in the PAV group and three patients (13) in the PSV group died in ICU (respiratory rate [RR] of PAV 114 95 CI 028 to 46 p = 086) Overall in-hospital mortality was nine patients (33) in the PAV group and six patients (26) in the PSV group (RR of PAV 128 95 CI 054 to 31 p = 058)

Cointerventions Sedating MedicationsTable 4 compares the use of medications that may prolong weaning (by causing sedation or neuromuscular weakness) in the two groups pre- and postrandomization Type and dose of sedative medication were highly variable among individual patients and therefore we analyzed pre- and postrandomiza-tion changes to determine the impact of the ventilation pro-tocol on sedative use within each patient In the PAV group benzodiazepine narcotic and propofol use measured as both cumulative dose and average daily dose all decreased signifi-cantly postrandomization as did neuromuscular blocker use In the PSV group average daily dose of lorazepam equiva-lent and average daily dose of propofol decreased signifi-cantly postrandomization whereas the proportion of patients

Figure 2 A Asynchrony index pre- and postrandomization on proportional assist ventilation B Asynchrony index pre- and postrandomization on pressure support ventilationPAV群 PSV群

Asynchrony Index () = 同調イベントの回数全呼吸回数 times 100

IntensiveCareMed(2006)321515ndash1522

AI>10は3名いたがPAVに設定変更後0名となった

AI>10は6名いたがPSVに設定変更後4名となった

Clinical Outcomes

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

成功したSBTまでの数抜管成功までの数に変わりなしICU退室までの数はPAV群で有意に短い

Kaplan-Meier curves

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

抜管成功までの時間 ICU退室までの時間

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 9

TABLE 4 Comparison of Medications That May Alter Weaning Duration Pre- and Postrandomization

Variables

PAV Group (n = 27)

PSV Group (n = 23) Postrandomization

Pre randomization Pos trandomization Pre randomization Pos trandomizationAdjusted Difference PSV ndash PAV (95 CI)

Lorazepam equivalents

Cumulative dose mg

308 (207) 48 (257)a 137 (440) 54 (258) 237 (ndash580 to 1055)

Average daily dose mg

44 (255) 09 (43)a 23 (85) 06 (40)a 24 (ndash62 to 110)

Single-day dose mgb

08 (72) 12 (50) 16 (70) 10 (64) 32 (ndash13 to 78)

Morphine equivalents

Cumulative dose mg

5003 (6188) 1140 (3140)b 1710 (13310) 1610 (4860) 591 (ndash2515 to 3696)

Average daily dose mg

858 (1638) 170 (578)c 428 (1724) 136 (320)d 108 (ndash316 to 532)

Single-day dose mga

180 (1240) 187 (77) 100 (1920) 50 (630)d 334 (ndash026 to 670)d

Propofol

Cumulative dose mg

7400 (55500) 0 (700)a 29450 (85800) 0 (27660) 226 (ndash42156 to 42607)

Average daily dose mg

529 (11690) 0 (53)a 4879 (15980) 0 (2738)a 651 (ndash3211 to 4514)

Single-day dose mga

0 (0) 0 (0) 0 (1000) 0 (0) 344 (ndash3781 to 4469)

Neuroleptics

Days receiving neuroleptics n

0 (2) 0 (4) 0 (0) 1 (4)a ndash16 (ndash34 to 03)

Proportion of patients who received a neuroleptic

407 407 174 565b

NMBs

Days receiving NMBs n

1 (1) 0 (0)a 0 (1) 0 (0) ndash01 (ndash05 to 03)

Proportion of patients who received an NMB

518 148a 348 174

Corticosteroids

Days receiving corticosteroids n

0 (0) 0 (0) 0 (2) 0 (2) 02 (ndash06 to 10)

Proportion of patients who received corticosteroids

185 148 348 261

PAV = proportional assist ventilation PSV = pressure support ventilation NMB = neuromuscular blockera p lt 005b Single-day dose = total dose given day prior to randomization or total dose give day after randomizationc p lt 0005d p lt 00005e p le 007All data points except percentage of patients receiving neuroleptics NMBs or corticosteroids are presented as median (interquartile range)

鎮静剤鎮痛剤についてはPAV群とPSV群で変わりなし

その他bull 不整脈気胸プロトコル施中の死亡

は両群とも起こらなかった

筆者らの考察bull PSVはweaningに最も使われているモードだ

がそれと較しても遜ないweaningができることがわかった

bull アシストが少ない傾向であったがPAV群のが呼吸窮迫となる回数が少なかった

rarr同調性が良くなった結果の可能性

ICU退室までの期間はPAV群で短いという結果であったがpowerが少ない

筆者らの考察bull 単施設での結果であり他の施設にこの

新しいプロトコルそのものが受けれられるかは不明である

bull SBTに失敗した患者を主に対象としておりICU患者すべてに適応はできない

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる

Page 26: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

PAV protocolSupplemental Digital Content Figure 2a Proportional Assist Ventilation Protocol

bull 開始の設定はサポート率70

bull 2-3時間おきにRRTがアセスメントし問題なければ10-20ずつPAVサポートを下げる

サポート率90まで上げてもダメな場合はACに戻す

PSV protocolSupplemental Digital Content Figure 2b Pressure Support Ventilation Protocol

bull 開始の設定はPS 15cmH2O

bull まずRRlt35 VTgt5mLkgで耐えられるPSまで下げる

bull 2-3時間おきにRRTがアセスメントしてPS2-3cmH2Pずつ下げる

Data collection form for RRTsSupplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

SBTの開始基準bull PF≧200bull FiO2≦05bull pH≧732bull トリガーできるbull 24時間以内に筋虚イベントなしbull 強薬昇圧剤の使なし

Data collection form for RRTs

Supplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

CPAPにしてRSBIを測定RSBI<105であればSBT施

PAV群はサポート率20PSV群はPS 5cmH2OでSBTを施

PEEP≦8cmH2OでSaO2≧90呼吸窮迫なしであればSBTクリア

抜管へ

Endpoints 呼吸器離脱までの時間ICU病院滞在期間

呼吸器離脱成功の定義抜管後48時間再挿管なし侵襲的呼吸(NIV)の使もなし気管切開使時は呼吸器から最終的に外れた時点1のうち12時間以上NIVを使していない

Statistical Analysis bull Pilot studyのためサンプルサイズの計

算はしていないbull P値<005を統計学的有意としたbull Intention to treat approachをった

RESULTS

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial

エンロール期間は42か80名が基準を満たしが最終的に割付されたのは54名

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial最終的に解析がわれたのはPAV群27名PSV群23名

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

ベースラインは両群間で変わりなし重症度はAPACHEⅡ 27点程度患者は肺炎が3割もともとCOPDを持っている患者は2割程度

Feasibility Outcomes bull 呼吸窮迫でRRTが呼吸器サポートを上昇

させた回数はPAV群で76回232daysPSV群で124回295daysと有意にPSV群で多かった(p=0002)

bull ACに戻した回数はPAV群で66回PSV群で90回と変わりなし(p=061)

bull 全くACに戻さなかった患者はPAV群で627名PSV群で623名(p=10)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

ベースラインは両群間で変わりなしPAVにすることでΔP圧Peak圧が低く吸気時間呼吸時間がくなるPSV群と較するとPAV群のがPaoAreaPaoAreaVE(呼吸器のサポートの強さをす)が低くVTPaoAreaがくsecond post-randomizationの解析においてΔP圧Peak圧が低い

Asynchrony index

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 7

B635) documenting compliance with the daily screening for SBT noncompletion of either the CPAP screening test or SBT occurred on 5 of 292 of patient study days (17) and this violation involved two patients in each study arm The second audit examining for nine different potential deviations found complete adherence to study protocol on 254 of 292 of patient study days (900) There were no adverse events (pneumo-thorax arrhythmia or death) resulting from any of the study interventions

Physiologic Response to Ventilation ProtocolsRecordings of flow and airway pressure were available for analysis for all 50 patients prerandomization 49 patients immediately (median 1 interquartile range [IQR] 0ndash15 hr) postrandomization and for 26 patients during subsequent days (median 27 IQR 20ndash70-hr postrandomization) Approximately 15000 breaths in 125 recordings were analyzed Table 2 compares the respiratory parameters in both groups pre- and postrandomization There were no statistically sig-nificant differences between the PAV and PSV groups at base-line prerandomization (p values not shown) In the PAV group there was a significant decrease in peak airway pressure delta pressure (measured as peak airway opening pressure minus the PEEP) and area under the inspiratory pressure curve immediately postrandomization and an increase in the per-cent of total cycle time spent in inspiration (TiTtot) whereas there were no significant differences immediately postran-domization in the PSV group The second measurement of postrandomization respiratory parameters revealed signifi-cant decreases in peak airway pressure and delta pressure in both PAV and PSV groups Adjusting for baseline measure-ments as a covariate the PAV group received 7 cm H2

O∙s less ventilator assistance (PaoArea) per liter of minute ventilation (p lt 00001) and increased their VT by 115 mLcm H

2O∙s

relative to the PSV group (p lt 0001)PAV significantly reduced AI postrandomization (Fig 2A)

In total nine patients (18) had a high AI (gt 10 of breaths) on PSV or on AC mode at baseline No patients had a high AI on PAV In the PSV group six patients (26) had a high AI

at baseline four (17) immediately postrandomization and one patient (7) after further adjustment of ventilator settings according to the PSV protocol (Fig 2B) Adjusting for baseline differences patients in the PAV group tended to have a lower AI than patients in the PSV group immediately postrandom-ization but this did not reach statistical significance (adjusted difference ndash45 [95 CI ndash95 to 04] p = 007)

Clinical OutcomesTable 3 shows the results for the main clinical endpoints Weaning duration and time to successful SBT did not dif-fer significantly among survivors between the groups Among ICU survivors median time to ICU discharge was shorter in the group randomized to PAV 73 days (IQR 52ndash114 d) ver-sus 124 days (IQR 75ndash308 d) in the PSV group (p = 003) Kaplan-Meier plots of the proportion of at-risk patients achiev-ing successful extubation and live ICU discharge are shown in Figure 3 A and B No difference was found for hospital length of stay ICU mortality or hospital mortality Four patients (15) in the PAV group and three patients (13) in the PSV group died in ICU (respiratory rate [RR] of PAV 114 95 CI 028 to 46 p = 086) Overall in-hospital mortality was nine patients (33) in the PAV group and six patients (26) in the PSV group (RR of PAV 128 95 CI 054 to 31 p = 058)

Cointerventions Sedating MedicationsTable 4 compares the use of medications that may prolong weaning (by causing sedation or neuromuscular weakness) in the two groups pre- and postrandomization Type and dose of sedative medication were highly variable among individual patients and therefore we analyzed pre- and postrandomiza-tion changes to determine the impact of the ventilation pro-tocol on sedative use within each patient In the PAV group benzodiazepine narcotic and propofol use measured as both cumulative dose and average daily dose all decreased signifi-cantly postrandomization as did neuromuscular blocker use In the PSV group average daily dose of lorazepam equiva-lent and average daily dose of propofol decreased signifi-cantly postrandomization whereas the proportion of patients

Figure 2 A Asynchrony index pre- and postrandomization on proportional assist ventilation B Asynchrony index pre- and postrandomization on pressure support ventilationPAV群 PSV群

Asynchrony Index () = 同調イベントの回数全呼吸回数 times 100

IntensiveCareMed(2006)321515ndash1522

AI>10は3名いたがPAVに設定変更後0名となった

AI>10は6名いたがPSVに設定変更後4名となった

Clinical Outcomes

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

成功したSBTまでの数抜管成功までの数に変わりなしICU退室までの数はPAV群で有意に短い

Kaplan-Meier curves

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

抜管成功までの時間 ICU退室までの時間

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 9

TABLE 4 Comparison of Medications That May Alter Weaning Duration Pre- and Postrandomization

Variables

PAV Group (n = 27)

PSV Group (n = 23) Postrandomization

Pre randomization Pos trandomization Pre randomization Pos trandomizationAdjusted Difference PSV ndash PAV (95 CI)

Lorazepam equivalents

Cumulative dose mg

308 (207) 48 (257)a 137 (440) 54 (258) 237 (ndash580 to 1055)

Average daily dose mg

44 (255) 09 (43)a 23 (85) 06 (40)a 24 (ndash62 to 110)

Single-day dose mgb

08 (72) 12 (50) 16 (70) 10 (64) 32 (ndash13 to 78)

Morphine equivalents

Cumulative dose mg

5003 (6188) 1140 (3140)b 1710 (13310) 1610 (4860) 591 (ndash2515 to 3696)

Average daily dose mg

858 (1638) 170 (578)c 428 (1724) 136 (320)d 108 (ndash316 to 532)

Single-day dose mga

180 (1240) 187 (77) 100 (1920) 50 (630)d 334 (ndash026 to 670)d

Propofol

Cumulative dose mg

7400 (55500) 0 (700)a 29450 (85800) 0 (27660) 226 (ndash42156 to 42607)

Average daily dose mg

529 (11690) 0 (53)a 4879 (15980) 0 (2738)a 651 (ndash3211 to 4514)

Single-day dose mga

0 (0) 0 (0) 0 (1000) 0 (0) 344 (ndash3781 to 4469)

Neuroleptics

Days receiving neuroleptics n

0 (2) 0 (4) 0 (0) 1 (4)a ndash16 (ndash34 to 03)

Proportion of patients who received a neuroleptic

407 407 174 565b

NMBs

Days receiving NMBs n

1 (1) 0 (0)a 0 (1) 0 (0) ndash01 (ndash05 to 03)

Proportion of patients who received an NMB

518 148a 348 174

Corticosteroids

Days receiving corticosteroids n

0 (0) 0 (0) 0 (2) 0 (2) 02 (ndash06 to 10)

Proportion of patients who received corticosteroids

185 148 348 261

PAV = proportional assist ventilation PSV = pressure support ventilation NMB = neuromuscular blockera p lt 005b Single-day dose = total dose given day prior to randomization or total dose give day after randomizationc p lt 0005d p lt 00005e p le 007All data points except percentage of patients receiving neuroleptics NMBs or corticosteroids are presented as median (interquartile range)

鎮静剤鎮痛剤についてはPAV群とPSV群で変わりなし

その他bull 不整脈気胸プロトコル施中の死亡

は両群とも起こらなかった

筆者らの考察bull PSVはweaningに最も使われているモードだ

がそれと較しても遜ないweaningができることがわかった

bull アシストが少ない傾向であったがPAV群のが呼吸窮迫となる回数が少なかった

rarr同調性が良くなった結果の可能性

ICU退室までの期間はPAV群で短いという結果であったがpowerが少ない

筆者らの考察bull 単施設での結果であり他の施設にこの

新しいプロトコルそのものが受けれられるかは不明である

bull SBTに失敗した患者を主に対象としておりICU患者すべてに適応はできない

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる

Page 27: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

PSV protocolSupplemental Digital Content Figure 2b Pressure Support Ventilation Protocol

bull 開始の設定はPS 15cmH2O

bull まずRRlt35 VTgt5mLkgで耐えられるPSまで下げる

bull 2-3時間おきにRRTがアセスメントしてPS2-3cmH2Pずつ下げる

Data collection form for RRTsSupplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

SBTの開始基準bull PF≧200bull FiO2≦05bull pH≧732bull トリガーできるbull 24時間以内に筋虚イベントなしbull 強薬昇圧剤の使なし

Data collection form for RRTs

Supplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

CPAPにしてRSBIを測定RSBI<105であればSBT施

PAV群はサポート率20PSV群はPS 5cmH2OでSBTを施

PEEP≦8cmH2OでSaO2≧90呼吸窮迫なしであればSBTクリア

抜管へ

Endpoints 呼吸器離脱までの時間ICU病院滞在期間

呼吸器離脱成功の定義抜管後48時間再挿管なし侵襲的呼吸(NIV)の使もなし気管切開使時は呼吸器から最終的に外れた時点1のうち12時間以上NIVを使していない

Statistical Analysis bull Pilot studyのためサンプルサイズの計

算はしていないbull P値<005を統計学的有意としたbull Intention to treat approachをった

RESULTS

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial

エンロール期間は42か80名が基準を満たしが最終的に割付されたのは54名

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial最終的に解析がわれたのはPAV群27名PSV群23名

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

ベースラインは両群間で変わりなし重症度はAPACHEⅡ 27点程度患者は肺炎が3割もともとCOPDを持っている患者は2割程度

Feasibility Outcomes bull 呼吸窮迫でRRTが呼吸器サポートを上昇

させた回数はPAV群で76回232daysPSV群で124回295daysと有意にPSV群で多かった(p=0002)

bull ACに戻した回数はPAV群で66回PSV群で90回と変わりなし(p=061)

bull 全くACに戻さなかった患者はPAV群で627名PSV群で623名(p=10)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

ベースラインは両群間で変わりなしPAVにすることでΔP圧Peak圧が低く吸気時間呼吸時間がくなるPSV群と較するとPAV群のがPaoAreaPaoAreaVE(呼吸器のサポートの強さをす)が低くVTPaoAreaがくsecond post-randomizationの解析においてΔP圧Peak圧が低い

Asynchrony index

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 7

B635) documenting compliance with the daily screening for SBT noncompletion of either the CPAP screening test or SBT occurred on 5 of 292 of patient study days (17) and this violation involved two patients in each study arm The second audit examining for nine different potential deviations found complete adherence to study protocol on 254 of 292 of patient study days (900) There were no adverse events (pneumo-thorax arrhythmia or death) resulting from any of the study interventions

Physiologic Response to Ventilation ProtocolsRecordings of flow and airway pressure were available for analysis for all 50 patients prerandomization 49 patients immediately (median 1 interquartile range [IQR] 0ndash15 hr) postrandomization and for 26 patients during subsequent days (median 27 IQR 20ndash70-hr postrandomization) Approximately 15000 breaths in 125 recordings were analyzed Table 2 compares the respiratory parameters in both groups pre- and postrandomization There were no statistically sig-nificant differences between the PAV and PSV groups at base-line prerandomization (p values not shown) In the PAV group there was a significant decrease in peak airway pressure delta pressure (measured as peak airway opening pressure minus the PEEP) and area under the inspiratory pressure curve immediately postrandomization and an increase in the per-cent of total cycle time spent in inspiration (TiTtot) whereas there were no significant differences immediately postran-domization in the PSV group The second measurement of postrandomization respiratory parameters revealed signifi-cant decreases in peak airway pressure and delta pressure in both PAV and PSV groups Adjusting for baseline measure-ments as a covariate the PAV group received 7 cm H2

O∙s less ventilator assistance (PaoArea) per liter of minute ventilation (p lt 00001) and increased their VT by 115 mLcm H

2O∙s

relative to the PSV group (p lt 0001)PAV significantly reduced AI postrandomization (Fig 2A)

In total nine patients (18) had a high AI (gt 10 of breaths) on PSV or on AC mode at baseline No patients had a high AI on PAV In the PSV group six patients (26) had a high AI

at baseline four (17) immediately postrandomization and one patient (7) after further adjustment of ventilator settings according to the PSV protocol (Fig 2B) Adjusting for baseline differences patients in the PAV group tended to have a lower AI than patients in the PSV group immediately postrandom-ization but this did not reach statistical significance (adjusted difference ndash45 [95 CI ndash95 to 04] p = 007)

Clinical OutcomesTable 3 shows the results for the main clinical endpoints Weaning duration and time to successful SBT did not dif-fer significantly among survivors between the groups Among ICU survivors median time to ICU discharge was shorter in the group randomized to PAV 73 days (IQR 52ndash114 d) ver-sus 124 days (IQR 75ndash308 d) in the PSV group (p = 003) Kaplan-Meier plots of the proportion of at-risk patients achiev-ing successful extubation and live ICU discharge are shown in Figure 3 A and B No difference was found for hospital length of stay ICU mortality or hospital mortality Four patients (15) in the PAV group and three patients (13) in the PSV group died in ICU (respiratory rate [RR] of PAV 114 95 CI 028 to 46 p = 086) Overall in-hospital mortality was nine patients (33) in the PAV group and six patients (26) in the PSV group (RR of PAV 128 95 CI 054 to 31 p = 058)

Cointerventions Sedating MedicationsTable 4 compares the use of medications that may prolong weaning (by causing sedation or neuromuscular weakness) in the two groups pre- and postrandomization Type and dose of sedative medication were highly variable among individual patients and therefore we analyzed pre- and postrandomiza-tion changes to determine the impact of the ventilation pro-tocol on sedative use within each patient In the PAV group benzodiazepine narcotic and propofol use measured as both cumulative dose and average daily dose all decreased signifi-cantly postrandomization as did neuromuscular blocker use In the PSV group average daily dose of lorazepam equiva-lent and average daily dose of propofol decreased signifi-cantly postrandomization whereas the proportion of patients

Figure 2 A Asynchrony index pre- and postrandomization on proportional assist ventilation B Asynchrony index pre- and postrandomization on pressure support ventilationPAV群 PSV群

Asynchrony Index () = 同調イベントの回数全呼吸回数 times 100

IntensiveCareMed(2006)321515ndash1522

AI>10は3名いたがPAVに設定変更後0名となった

AI>10は6名いたがPSVに設定変更後4名となった

Clinical Outcomes

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

成功したSBTまでの数抜管成功までの数に変わりなしICU退室までの数はPAV群で有意に短い

Kaplan-Meier curves

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

抜管成功までの時間 ICU退室までの時間

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 9

TABLE 4 Comparison of Medications That May Alter Weaning Duration Pre- and Postrandomization

Variables

PAV Group (n = 27)

PSV Group (n = 23) Postrandomization

Pre randomization Pos trandomization Pre randomization Pos trandomizationAdjusted Difference PSV ndash PAV (95 CI)

Lorazepam equivalents

Cumulative dose mg

308 (207) 48 (257)a 137 (440) 54 (258) 237 (ndash580 to 1055)

Average daily dose mg

44 (255) 09 (43)a 23 (85) 06 (40)a 24 (ndash62 to 110)

Single-day dose mgb

08 (72) 12 (50) 16 (70) 10 (64) 32 (ndash13 to 78)

Morphine equivalents

Cumulative dose mg

5003 (6188) 1140 (3140)b 1710 (13310) 1610 (4860) 591 (ndash2515 to 3696)

Average daily dose mg

858 (1638) 170 (578)c 428 (1724) 136 (320)d 108 (ndash316 to 532)

Single-day dose mga

180 (1240) 187 (77) 100 (1920) 50 (630)d 334 (ndash026 to 670)d

Propofol

Cumulative dose mg

7400 (55500) 0 (700)a 29450 (85800) 0 (27660) 226 (ndash42156 to 42607)

Average daily dose mg

529 (11690) 0 (53)a 4879 (15980) 0 (2738)a 651 (ndash3211 to 4514)

Single-day dose mga

0 (0) 0 (0) 0 (1000) 0 (0) 344 (ndash3781 to 4469)

Neuroleptics

Days receiving neuroleptics n

0 (2) 0 (4) 0 (0) 1 (4)a ndash16 (ndash34 to 03)

Proportion of patients who received a neuroleptic

407 407 174 565b

NMBs

Days receiving NMBs n

1 (1) 0 (0)a 0 (1) 0 (0) ndash01 (ndash05 to 03)

Proportion of patients who received an NMB

518 148a 348 174

Corticosteroids

Days receiving corticosteroids n

0 (0) 0 (0) 0 (2) 0 (2) 02 (ndash06 to 10)

Proportion of patients who received corticosteroids

185 148 348 261

PAV = proportional assist ventilation PSV = pressure support ventilation NMB = neuromuscular blockera p lt 005b Single-day dose = total dose given day prior to randomization or total dose give day after randomizationc p lt 0005d p lt 00005e p le 007All data points except percentage of patients receiving neuroleptics NMBs or corticosteroids are presented as median (interquartile range)

鎮静剤鎮痛剤についてはPAV群とPSV群で変わりなし

その他bull 不整脈気胸プロトコル施中の死亡

は両群とも起こらなかった

筆者らの考察bull PSVはweaningに最も使われているモードだ

がそれと較しても遜ないweaningができることがわかった

bull アシストが少ない傾向であったがPAV群のが呼吸窮迫となる回数が少なかった

rarr同調性が良くなった結果の可能性

ICU退室までの期間はPAV群で短いという結果であったがpowerが少ない

筆者らの考察bull 単施設での結果であり他の施設にこの

新しいプロトコルそのものが受けれられるかは不明である

bull SBTに失敗した患者を主に対象としておりICU患者すべてに適応はできない

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる

Page 28: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

Data collection form for RRTsSupplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

SBTの開始基準bull PF≧200bull FiO2≦05bull pH≧732bull トリガーできるbull 24時間以内に筋虚イベントなしbull 強薬昇圧剤の使なし

Data collection form for RRTs

Supplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

CPAPにしてRSBIを測定RSBI<105であればSBT施

PAV群はサポート率20PSV群はPS 5cmH2OでSBTを施

PEEP≦8cmH2OでSaO2≧90呼吸窮迫なしであればSBTクリア

抜管へ

Endpoints 呼吸器離脱までの時間ICU病院滞在期間

呼吸器離脱成功の定義抜管後48時間再挿管なし侵襲的呼吸(NIV)の使もなし気管切開使時は呼吸器から最終的に外れた時点1のうち12時間以上NIVを使していない

Statistical Analysis bull Pilot studyのためサンプルサイズの計

算はしていないbull P値<005を統計学的有意としたbull Intention to treat approachをった

RESULTS

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial

エンロール期間は42か80名が基準を満たしが最終的に割付されたのは54名

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial最終的に解析がわれたのはPAV群27名PSV群23名

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

ベースラインは両群間で変わりなし重症度はAPACHEⅡ 27点程度患者は肺炎が3割もともとCOPDを持っている患者は2割程度

Feasibility Outcomes bull 呼吸窮迫でRRTが呼吸器サポートを上昇

させた回数はPAV群で76回232daysPSV群で124回295daysと有意にPSV群で多かった(p=0002)

bull ACに戻した回数はPAV群で66回PSV群で90回と変わりなし(p=061)

bull 全くACに戻さなかった患者はPAV群で627名PSV群で623名(p=10)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

ベースラインは両群間で変わりなしPAVにすることでΔP圧Peak圧が低く吸気時間呼吸時間がくなるPSV群と較するとPAV群のがPaoAreaPaoAreaVE(呼吸器のサポートの強さをす)が低くVTPaoAreaがくsecond post-randomizationの解析においてΔP圧Peak圧が低い

Asynchrony index

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 7

B635) documenting compliance with the daily screening for SBT noncompletion of either the CPAP screening test or SBT occurred on 5 of 292 of patient study days (17) and this violation involved two patients in each study arm The second audit examining for nine different potential deviations found complete adherence to study protocol on 254 of 292 of patient study days (900) There were no adverse events (pneumo-thorax arrhythmia or death) resulting from any of the study interventions

Physiologic Response to Ventilation ProtocolsRecordings of flow and airway pressure were available for analysis for all 50 patients prerandomization 49 patients immediately (median 1 interquartile range [IQR] 0ndash15 hr) postrandomization and for 26 patients during subsequent days (median 27 IQR 20ndash70-hr postrandomization) Approximately 15000 breaths in 125 recordings were analyzed Table 2 compares the respiratory parameters in both groups pre- and postrandomization There were no statistically sig-nificant differences between the PAV and PSV groups at base-line prerandomization (p values not shown) In the PAV group there was a significant decrease in peak airway pressure delta pressure (measured as peak airway opening pressure minus the PEEP) and area under the inspiratory pressure curve immediately postrandomization and an increase in the per-cent of total cycle time spent in inspiration (TiTtot) whereas there were no significant differences immediately postran-domization in the PSV group The second measurement of postrandomization respiratory parameters revealed signifi-cant decreases in peak airway pressure and delta pressure in both PAV and PSV groups Adjusting for baseline measure-ments as a covariate the PAV group received 7 cm H2

O∙s less ventilator assistance (PaoArea) per liter of minute ventilation (p lt 00001) and increased their VT by 115 mLcm H

2O∙s

relative to the PSV group (p lt 0001)PAV significantly reduced AI postrandomization (Fig 2A)

In total nine patients (18) had a high AI (gt 10 of breaths) on PSV or on AC mode at baseline No patients had a high AI on PAV In the PSV group six patients (26) had a high AI

at baseline four (17) immediately postrandomization and one patient (7) after further adjustment of ventilator settings according to the PSV protocol (Fig 2B) Adjusting for baseline differences patients in the PAV group tended to have a lower AI than patients in the PSV group immediately postrandom-ization but this did not reach statistical significance (adjusted difference ndash45 [95 CI ndash95 to 04] p = 007)

Clinical OutcomesTable 3 shows the results for the main clinical endpoints Weaning duration and time to successful SBT did not dif-fer significantly among survivors between the groups Among ICU survivors median time to ICU discharge was shorter in the group randomized to PAV 73 days (IQR 52ndash114 d) ver-sus 124 days (IQR 75ndash308 d) in the PSV group (p = 003) Kaplan-Meier plots of the proportion of at-risk patients achiev-ing successful extubation and live ICU discharge are shown in Figure 3 A and B No difference was found for hospital length of stay ICU mortality or hospital mortality Four patients (15) in the PAV group and three patients (13) in the PSV group died in ICU (respiratory rate [RR] of PAV 114 95 CI 028 to 46 p = 086) Overall in-hospital mortality was nine patients (33) in the PAV group and six patients (26) in the PSV group (RR of PAV 128 95 CI 054 to 31 p = 058)

Cointerventions Sedating MedicationsTable 4 compares the use of medications that may prolong weaning (by causing sedation or neuromuscular weakness) in the two groups pre- and postrandomization Type and dose of sedative medication were highly variable among individual patients and therefore we analyzed pre- and postrandomiza-tion changes to determine the impact of the ventilation pro-tocol on sedative use within each patient In the PAV group benzodiazepine narcotic and propofol use measured as both cumulative dose and average daily dose all decreased signifi-cantly postrandomization as did neuromuscular blocker use In the PSV group average daily dose of lorazepam equiva-lent and average daily dose of propofol decreased signifi-cantly postrandomization whereas the proportion of patients

Figure 2 A Asynchrony index pre- and postrandomization on proportional assist ventilation B Asynchrony index pre- and postrandomization on pressure support ventilationPAV群 PSV群

Asynchrony Index () = 同調イベントの回数全呼吸回数 times 100

IntensiveCareMed(2006)321515ndash1522

AI>10は3名いたがPAVに設定変更後0名となった

AI>10は6名いたがPSVに設定変更後4名となった

Clinical Outcomes

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

成功したSBTまでの数抜管成功までの数に変わりなしICU退室までの数はPAV群で有意に短い

Kaplan-Meier curves

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

抜管成功までの時間 ICU退室までの時間

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 9

TABLE 4 Comparison of Medications That May Alter Weaning Duration Pre- and Postrandomization

Variables

PAV Group (n = 27)

PSV Group (n = 23) Postrandomization

Pre randomization Pos trandomization Pre randomization Pos trandomizationAdjusted Difference PSV ndash PAV (95 CI)

Lorazepam equivalents

Cumulative dose mg

308 (207) 48 (257)a 137 (440) 54 (258) 237 (ndash580 to 1055)

Average daily dose mg

44 (255) 09 (43)a 23 (85) 06 (40)a 24 (ndash62 to 110)

Single-day dose mgb

08 (72) 12 (50) 16 (70) 10 (64) 32 (ndash13 to 78)

Morphine equivalents

Cumulative dose mg

5003 (6188) 1140 (3140)b 1710 (13310) 1610 (4860) 591 (ndash2515 to 3696)

Average daily dose mg

858 (1638) 170 (578)c 428 (1724) 136 (320)d 108 (ndash316 to 532)

Single-day dose mga

180 (1240) 187 (77) 100 (1920) 50 (630)d 334 (ndash026 to 670)d

Propofol

Cumulative dose mg

7400 (55500) 0 (700)a 29450 (85800) 0 (27660) 226 (ndash42156 to 42607)

Average daily dose mg

529 (11690) 0 (53)a 4879 (15980) 0 (2738)a 651 (ndash3211 to 4514)

Single-day dose mga

0 (0) 0 (0) 0 (1000) 0 (0) 344 (ndash3781 to 4469)

Neuroleptics

Days receiving neuroleptics n

0 (2) 0 (4) 0 (0) 1 (4)a ndash16 (ndash34 to 03)

Proportion of patients who received a neuroleptic

407 407 174 565b

NMBs

Days receiving NMBs n

1 (1) 0 (0)a 0 (1) 0 (0) ndash01 (ndash05 to 03)

Proportion of patients who received an NMB

518 148a 348 174

Corticosteroids

Days receiving corticosteroids n

0 (0) 0 (0) 0 (2) 0 (2) 02 (ndash06 to 10)

Proportion of patients who received corticosteroids

185 148 348 261

PAV = proportional assist ventilation PSV = pressure support ventilation NMB = neuromuscular blockera p lt 005b Single-day dose = total dose given day prior to randomization or total dose give day after randomizationc p lt 0005d p lt 00005e p le 007All data points except percentage of patients receiving neuroleptics NMBs or corticosteroids are presented as median (interquartile range)

鎮静剤鎮痛剤についてはPAV群とPSV群で変わりなし

その他bull 不整脈気胸プロトコル施中の死亡

は両群とも起こらなかった

筆者らの考察bull PSVはweaningに最も使われているモードだ

がそれと較しても遜ないweaningができることがわかった

bull アシストが少ない傾向であったがPAV群のが呼吸窮迫となる回数が少なかった

rarr同調性が良くなった結果の可能性

ICU退室までの期間はPAV群で短いという結果であったがpowerが少ない

筆者らの考察bull 単施設での結果であり他の施設にこの

新しいプロトコルそのものが受けれられるかは不明である

bull SBTに失敗した患者を主に対象としておりICU患者すべてに適応はできない

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる

Page 29: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

Data collection form for RRTs

Supplemental Digital Content Figure 3 Daily data collection form for Registered Respiratory Therapists

CPAPにしてRSBIを測定RSBI<105であればSBT施

PAV群はサポート率20PSV群はPS 5cmH2OでSBTを施

PEEP≦8cmH2OでSaO2≧90呼吸窮迫なしであればSBTクリア

抜管へ

Endpoints 呼吸器離脱までの時間ICU病院滞在期間

呼吸器離脱成功の定義抜管後48時間再挿管なし侵襲的呼吸(NIV)の使もなし気管切開使時は呼吸器から最終的に外れた時点1のうち12時間以上NIVを使していない

Statistical Analysis bull Pilot studyのためサンプルサイズの計

算はしていないbull P値<005を統計学的有意としたbull Intention to treat approachをった

RESULTS

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial

エンロール期間は42か80名が基準を満たしが最終的に割付されたのは54名

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial最終的に解析がわれたのはPAV群27名PSV群23名

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

ベースラインは両群間で変わりなし重症度はAPACHEⅡ 27点程度患者は肺炎が3割もともとCOPDを持っている患者は2割程度

Feasibility Outcomes bull 呼吸窮迫でRRTが呼吸器サポートを上昇

させた回数はPAV群で76回232daysPSV群で124回295daysと有意にPSV群で多かった(p=0002)

bull ACに戻した回数はPAV群で66回PSV群で90回と変わりなし(p=061)

bull 全くACに戻さなかった患者はPAV群で627名PSV群で623名(p=10)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

ベースラインは両群間で変わりなしPAVにすることでΔP圧Peak圧が低く吸気時間呼吸時間がくなるPSV群と較するとPAV群のがPaoAreaPaoAreaVE(呼吸器のサポートの強さをす)が低くVTPaoAreaがくsecond post-randomizationの解析においてΔP圧Peak圧が低い

Asynchrony index

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 7

B635) documenting compliance with the daily screening for SBT noncompletion of either the CPAP screening test or SBT occurred on 5 of 292 of patient study days (17) and this violation involved two patients in each study arm The second audit examining for nine different potential deviations found complete adherence to study protocol on 254 of 292 of patient study days (900) There were no adverse events (pneumo-thorax arrhythmia or death) resulting from any of the study interventions

Physiologic Response to Ventilation ProtocolsRecordings of flow and airway pressure were available for analysis for all 50 patients prerandomization 49 patients immediately (median 1 interquartile range [IQR] 0ndash15 hr) postrandomization and for 26 patients during subsequent days (median 27 IQR 20ndash70-hr postrandomization) Approximately 15000 breaths in 125 recordings were analyzed Table 2 compares the respiratory parameters in both groups pre- and postrandomization There were no statistically sig-nificant differences between the PAV and PSV groups at base-line prerandomization (p values not shown) In the PAV group there was a significant decrease in peak airway pressure delta pressure (measured as peak airway opening pressure minus the PEEP) and area under the inspiratory pressure curve immediately postrandomization and an increase in the per-cent of total cycle time spent in inspiration (TiTtot) whereas there were no significant differences immediately postran-domization in the PSV group The second measurement of postrandomization respiratory parameters revealed signifi-cant decreases in peak airway pressure and delta pressure in both PAV and PSV groups Adjusting for baseline measure-ments as a covariate the PAV group received 7 cm H2

O∙s less ventilator assistance (PaoArea) per liter of minute ventilation (p lt 00001) and increased their VT by 115 mLcm H

2O∙s

relative to the PSV group (p lt 0001)PAV significantly reduced AI postrandomization (Fig 2A)

In total nine patients (18) had a high AI (gt 10 of breaths) on PSV or on AC mode at baseline No patients had a high AI on PAV In the PSV group six patients (26) had a high AI

at baseline four (17) immediately postrandomization and one patient (7) after further adjustment of ventilator settings according to the PSV protocol (Fig 2B) Adjusting for baseline differences patients in the PAV group tended to have a lower AI than patients in the PSV group immediately postrandom-ization but this did not reach statistical significance (adjusted difference ndash45 [95 CI ndash95 to 04] p = 007)

Clinical OutcomesTable 3 shows the results for the main clinical endpoints Weaning duration and time to successful SBT did not dif-fer significantly among survivors between the groups Among ICU survivors median time to ICU discharge was shorter in the group randomized to PAV 73 days (IQR 52ndash114 d) ver-sus 124 days (IQR 75ndash308 d) in the PSV group (p = 003) Kaplan-Meier plots of the proportion of at-risk patients achiev-ing successful extubation and live ICU discharge are shown in Figure 3 A and B No difference was found for hospital length of stay ICU mortality or hospital mortality Four patients (15) in the PAV group and three patients (13) in the PSV group died in ICU (respiratory rate [RR] of PAV 114 95 CI 028 to 46 p = 086) Overall in-hospital mortality was nine patients (33) in the PAV group and six patients (26) in the PSV group (RR of PAV 128 95 CI 054 to 31 p = 058)

Cointerventions Sedating MedicationsTable 4 compares the use of medications that may prolong weaning (by causing sedation or neuromuscular weakness) in the two groups pre- and postrandomization Type and dose of sedative medication were highly variable among individual patients and therefore we analyzed pre- and postrandomiza-tion changes to determine the impact of the ventilation pro-tocol on sedative use within each patient In the PAV group benzodiazepine narcotic and propofol use measured as both cumulative dose and average daily dose all decreased signifi-cantly postrandomization as did neuromuscular blocker use In the PSV group average daily dose of lorazepam equiva-lent and average daily dose of propofol decreased signifi-cantly postrandomization whereas the proportion of patients

Figure 2 A Asynchrony index pre- and postrandomization on proportional assist ventilation B Asynchrony index pre- and postrandomization on pressure support ventilationPAV群 PSV群

Asynchrony Index () = 同調イベントの回数全呼吸回数 times 100

IntensiveCareMed(2006)321515ndash1522

AI>10は3名いたがPAVに設定変更後0名となった

AI>10は6名いたがPSVに設定変更後4名となった

Clinical Outcomes

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

成功したSBTまでの数抜管成功までの数に変わりなしICU退室までの数はPAV群で有意に短い

Kaplan-Meier curves

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

抜管成功までの時間 ICU退室までの時間

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 9

TABLE 4 Comparison of Medications That May Alter Weaning Duration Pre- and Postrandomization

Variables

PAV Group (n = 27)

PSV Group (n = 23) Postrandomization

Pre randomization Pos trandomization Pre randomization Pos trandomizationAdjusted Difference PSV ndash PAV (95 CI)

Lorazepam equivalents

Cumulative dose mg

308 (207) 48 (257)a 137 (440) 54 (258) 237 (ndash580 to 1055)

Average daily dose mg

44 (255) 09 (43)a 23 (85) 06 (40)a 24 (ndash62 to 110)

Single-day dose mgb

08 (72) 12 (50) 16 (70) 10 (64) 32 (ndash13 to 78)

Morphine equivalents

Cumulative dose mg

5003 (6188) 1140 (3140)b 1710 (13310) 1610 (4860) 591 (ndash2515 to 3696)

Average daily dose mg

858 (1638) 170 (578)c 428 (1724) 136 (320)d 108 (ndash316 to 532)

Single-day dose mga

180 (1240) 187 (77) 100 (1920) 50 (630)d 334 (ndash026 to 670)d

Propofol

Cumulative dose mg

7400 (55500) 0 (700)a 29450 (85800) 0 (27660) 226 (ndash42156 to 42607)

Average daily dose mg

529 (11690) 0 (53)a 4879 (15980) 0 (2738)a 651 (ndash3211 to 4514)

Single-day dose mga

0 (0) 0 (0) 0 (1000) 0 (0) 344 (ndash3781 to 4469)

Neuroleptics

Days receiving neuroleptics n

0 (2) 0 (4) 0 (0) 1 (4)a ndash16 (ndash34 to 03)

Proportion of patients who received a neuroleptic

407 407 174 565b

NMBs

Days receiving NMBs n

1 (1) 0 (0)a 0 (1) 0 (0) ndash01 (ndash05 to 03)

Proportion of patients who received an NMB

518 148a 348 174

Corticosteroids

Days receiving corticosteroids n

0 (0) 0 (0) 0 (2) 0 (2) 02 (ndash06 to 10)

Proportion of patients who received corticosteroids

185 148 348 261

PAV = proportional assist ventilation PSV = pressure support ventilation NMB = neuromuscular blockera p lt 005b Single-day dose = total dose given day prior to randomization or total dose give day after randomizationc p lt 0005d p lt 00005e p le 007All data points except percentage of patients receiving neuroleptics NMBs or corticosteroids are presented as median (interquartile range)

鎮静剤鎮痛剤についてはPAV群とPSV群で変わりなし

その他bull 不整脈気胸プロトコル施中の死亡

は両群とも起こらなかった

筆者らの考察bull PSVはweaningに最も使われているモードだ

がそれと較しても遜ないweaningができることがわかった

bull アシストが少ない傾向であったがPAV群のが呼吸窮迫となる回数が少なかった

rarr同調性が良くなった結果の可能性

ICU退室までの期間はPAV群で短いという結果であったがpowerが少ない

筆者らの考察bull 単施設での結果であり他の施設にこの

新しいプロトコルそのものが受けれられるかは不明である

bull SBTに失敗した患者を主に対象としておりICU患者すべてに適応はできない

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる

Page 30: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

Endpoints 呼吸器離脱までの時間ICU病院滞在期間

呼吸器離脱成功の定義抜管後48時間再挿管なし侵襲的呼吸(NIV)の使もなし気管切開使時は呼吸器から最終的に外れた時点1のうち12時間以上NIVを使していない

Statistical Analysis bull Pilot studyのためサンプルサイズの計

算はしていないbull P値<005を統計学的有意としたbull Intention to treat approachをった

RESULTS

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial

エンロール期間は42か80名が基準を満たしが最終的に割付されたのは54名

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial最終的に解析がわれたのはPAV群27名PSV群23名

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

ベースラインは両群間で変わりなし重症度はAPACHEⅡ 27点程度患者は肺炎が3割もともとCOPDを持っている患者は2割程度

Feasibility Outcomes bull 呼吸窮迫でRRTが呼吸器サポートを上昇

させた回数はPAV群で76回232daysPSV群で124回295daysと有意にPSV群で多かった(p=0002)

bull ACに戻した回数はPAV群で66回PSV群で90回と変わりなし(p=061)

bull 全くACに戻さなかった患者はPAV群で627名PSV群で623名(p=10)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

ベースラインは両群間で変わりなしPAVにすることでΔP圧Peak圧が低く吸気時間呼吸時間がくなるPSV群と較するとPAV群のがPaoAreaPaoAreaVE(呼吸器のサポートの強さをす)が低くVTPaoAreaがくsecond post-randomizationの解析においてΔP圧Peak圧が低い

Asynchrony index

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 7

B635) documenting compliance with the daily screening for SBT noncompletion of either the CPAP screening test or SBT occurred on 5 of 292 of patient study days (17) and this violation involved two patients in each study arm The second audit examining for nine different potential deviations found complete adherence to study protocol on 254 of 292 of patient study days (900) There were no adverse events (pneumo-thorax arrhythmia or death) resulting from any of the study interventions

Physiologic Response to Ventilation ProtocolsRecordings of flow and airway pressure were available for analysis for all 50 patients prerandomization 49 patients immediately (median 1 interquartile range [IQR] 0ndash15 hr) postrandomization and for 26 patients during subsequent days (median 27 IQR 20ndash70-hr postrandomization) Approximately 15000 breaths in 125 recordings were analyzed Table 2 compares the respiratory parameters in both groups pre- and postrandomization There were no statistically sig-nificant differences between the PAV and PSV groups at base-line prerandomization (p values not shown) In the PAV group there was a significant decrease in peak airway pressure delta pressure (measured as peak airway opening pressure minus the PEEP) and area under the inspiratory pressure curve immediately postrandomization and an increase in the per-cent of total cycle time spent in inspiration (TiTtot) whereas there were no significant differences immediately postran-domization in the PSV group The second measurement of postrandomization respiratory parameters revealed signifi-cant decreases in peak airway pressure and delta pressure in both PAV and PSV groups Adjusting for baseline measure-ments as a covariate the PAV group received 7 cm H2

O∙s less ventilator assistance (PaoArea) per liter of minute ventilation (p lt 00001) and increased their VT by 115 mLcm H

2O∙s

relative to the PSV group (p lt 0001)PAV significantly reduced AI postrandomization (Fig 2A)

In total nine patients (18) had a high AI (gt 10 of breaths) on PSV or on AC mode at baseline No patients had a high AI on PAV In the PSV group six patients (26) had a high AI

at baseline four (17) immediately postrandomization and one patient (7) after further adjustment of ventilator settings according to the PSV protocol (Fig 2B) Adjusting for baseline differences patients in the PAV group tended to have a lower AI than patients in the PSV group immediately postrandom-ization but this did not reach statistical significance (adjusted difference ndash45 [95 CI ndash95 to 04] p = 007)

Clinical OutcomesTable 3 shows the results for the main clinical endpoints Weaning duration and time to successful SBT did not dif-fer significantly among survivors between the groups Among ICU survivors median time to ICU discharge was shorter in the group randomized to PAV 73 days (IQR 52ndash114 d) ver-sus 124 days (IQR 75ndash308 d) in the PSV group (p = 003) Kaplan-Meier plots of the proportion of at-risk patients achiev-ing successful extubation and live ICU discharge are shown in Figure 3 A and B No difference was found for hospital length of stay ICU mortality or hospital mortality Four patients (15) in the PAV group and three patients (13) in the PSV group died in ICU (respiratory rate [RR] of PAV 114 95 CI 028 to 46 p = 086) Overall in-hospital mortality was nine patients (33) in the PAV group and six patients (26) in the PSV group (RR of PAV 128 95 CI 054 to 31 p = 058)

Cointerventions Sedating MedicationsTable 4 compares the use of medications that may prolong weaning (by causing sedation or neuromuscular weakness) in the two groups pre- and postrandomization Type and dose of sedative medication were highly variable among individual patients and therefore we analyzed pre- and postrandomiza-tion changes to determine the impact of the ventilation pro-tocol on sedative use within each patient In the PAV group benzodiazepine narcotic and propofol use measured as both cumulative dose and average daily dose all decreased signifi-cantly postrandomization as did neuromuscular blocker use In the PSV group average daily dose of lorazepam equiva-lent and average daily dose of propofol decreased signifi-cantly postrandomization whereas the proportion of patients

Figure 2 A Asynchrony index pre- and postrandomization on proportional assist ventilation B Asynchrony index pre- and postrandomization on pressure support ventilationPAV群 PSV群

Asynchrony Index () = 同調イベントの回数全呼吸回数 times 100

IntensiveCareMed(2006)321515ndash1522

AI>10は3名いたがPAVに設定変更後0名となった

AI>10は6名いたがPSVに設定変更後4名となった

Clinical Outcomes

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

成功したSBTまでの数抜管成功までの数に変わりなしICU退室までの数はPAV群で有意に短い

Kaplan-Meier curves

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

抜管成功までの時間 ICU退室までの時間

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 9

TABLE 4 Comparison of Medications That May Alter Weaning Duration Pre- and Postrandomization

Variables

PAV Group (n = 27)

PSV Group (n = 23) Postrandomization

Pre randomization Pos trandomization Pre randomization Pos trandomizationAdjusted Difference PSV ndash PAV (95 CI)

Lorazepam equivalents

Cumulative dose mg

308 (207) 48 (257)a 137 (440) 54 (258) 237 (ndash580 to 1055)

Average daily dose mg

44 (255) 09 (43)a 23 (85) 06 (40)a 24 (ndash62 to 110)

Single-day dose mgb

08 (72) 12 (50) 16 (70) 10 (64) 32 (ndash13 to 78)

Morphine equivalents

Cumulative dose mg

5003 (6188) 1140 (3140)b 1710 (13310) 1610 (4860) 591 (ndash2515 to 3696)

Average daily dose mg

858 (1638) 170 (578)c 428 (1724) 136 (320)d 108 (ndash316 to 532)

Single-day dose mga

180 (1240) 187 (77) 100 (1920) 50 (630)d 334 (ndash026 to 670)d

Propofol

Cumulative dose mg

7400 (55500) 0 (700)a 29450 (85800) 0 (27660) 226 (ndash42156 to 42607)

Average daily dose mg

529 (11690) 0 (53)a 4879 (15980) 0 (2738)a 651 (ndash3211 to 4514)

Single-day dose mga

0 (0) 0 (0) 0 (1000) 0 (0) 344 (ndash3781 to 4469)

Neuroleptics

Days receiving neuroleptics n

0 (2) 0 (4) 0 (0) 1 (4)a ndash16 (ndash34 to 03)

Proportion of patients who received a neuroleptic

407 407 174 565b

NMBs

Days receiving NMBs n

1 (1) 0 (0)a 0 (1) 0 (0) ndash01 (ndash05 to 03)

Proportion of patients who received an NMB

518 148a 348 174

Corticosteroids

Days receiving corticosteroids n

0 (0) 0 (0) 0 (2) 0 (2) 02 (ndash06 to 10)

Proportion of patients who received corticosteroids

185 148 348 261

PAV = proportional assist ventilation PSV = pressure support ventilation NMB = neuromuscular blockera p lt 005b Single-day dose = total dose given day prior to randomization or total dose give day after randomizationc p lt 0005d p lt 00005e p le 007All data points except percentage of patients receiving neuroleptics NMBs or corticosteroids are presented as median (interquartile range)

鎮静剤鎮痛剤についてはPAV群とPSV群で変わりなし

その他bull 不整脈気胸プロトコル施中の死亡

は両群とも起こらなかった

筆者らの考察bull PSVはweaningに最も使われているモードだ

がそれと較しても遜ないweaningができることがわかった

bull アシストが少ない傾向であったがPAV群のが呼吸窮迫となる回数が少なかった

rarr同調性が良くなった結果の可能性

ICU退室までの期間はPAV群で短いという結果であったがpowerが少ない

筆者らの考察bull 単施設での結果であり他の施設にこの

新しいプロトコルそのものが受けれられるかは不明である

bull SBTに失敗した患者を主に対象としておりICU患者すべてに適応はできない

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる

Page 31: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

Statistical Analysis bull Pilot studyのためサンプルサイズの計

算はしていないbull P値<005を統計学的有意としたbull Intention to treat approachをった

RESULTS

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial

エンロール期間は42か80名が基準を満たしが最終的に割付されたのは54名

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial最終的に解析がわれたのはPAV群27名PSV群23名

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

ベースラインは両群間で変わりなし重症度はAPACHEⅡ 27点程度患者は肺炎が3割もともとCOPDを持っている患者は2割程度

Feasibility Outcomes bull 呼吸窮迫でRRTが呼吸器サポートを上昇

させた回数はPAV群で76回232daysPSV群で124回295daysと有意にPSV群で多かった(p=0002)

bull ACに戻した回数はPAV群で66回PSV群で90回と変わりなし(p=061)

bull 全くACに戻さなかった患者はPAV群で627名PSV群で623名(p=10)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

ベースラインは両群間で変わりなしPAVにすることでΔP圧Peak圧が低く吸気時間呼吸時間がくなるPSV群と較するとPAV群のがPaoAreaPaoAreaVE(呼吸器のサポートの強さをす)が低くVTPaoAreaがくsecond post-randomizationの解析においてΔP圧Peak圧が低い

Asynchrony index

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 7

B635) documenting compliance with the daily screening for SBT noncompletion of either the CPAP screening test or SBT occurred on 5 of 292 of patient study days (17) and this violation involved two patients in each study arm The second audit examining for nine different potential deviations found complete adherence to study protocol on 254 of 292 of patient study days (900) There were no adverse events (pneumo-thorax arrhythmia or death) resulting from any of the study interventions

Physiologic Response to Ventilation ProtocolsRecordings of flow and airway pressure were available for analysis for all 50 patients prerandomization 49 patients immediately (median 1 interquartile range [IQR] 0ndash15 hr) postrandomization and for 26 patients during subsequent days (median 27 IQR 20ndash70-hr postrandomization) Approximately 15000 breaths in 125 recordings were analyzed Table 2 compares the respiratory parameters in both groups pre- and postrandomization There were no statistically sig-nificant differences between the PAV and PSV groups at base-line prerandomization (p values not shown) In the PAV group there was a significant decrease in peak airway pressure delta pressure (measured as peak airway opening pressure minus the PEEP) and area under the inspiratory pressure curve immediately postrandomization and an increase in the per-cent of total cycle time spent in inspiration (TiTtot) whereas there were no significant differences immediately postran-domization in the PSV group The second measurement of postrandomization respiratory parameters revealed signifi-cant decreases in peak airway pressure and delta pressure in both PAV and PSV groups Adjusting for baseline measure-ments as a covariate the PAV group received 7 cm H2

O∙s less ventilator assistance (PaoArea) per liter of minute ventilation (p lt 00001) and increased their VT by 115 mLcm H

2O∙s

relative to the PSV group (p lt 0001)PAV significantly reduced AI postrandomization (Fig 2A)

In total nine patients (18) had a high AI (gt 10 of breaths) on PSV or on AC mode at baseline No patients had a high AI on PAV In the PSV group six patients (26) had a high AI

at baseline four (17) immediately postrandomization and one patient (7) after further adjustment of ventilator settings according to the PSV protocol (Fig 2B) Adjusting for baseline differences patients in the PAV group tended to have a lower AI than patients in the PSV group immediately postrandom-ization but this did not reach statistical significance (adjusted difference ndash45 [95 CI ndash95 to 04] p = 007)

Clinical OutcomesTable 3 shows the results for the main clinical endpoints Weaning duration and time to successful SBT did not dif-fer significantly among survivors between the groups Among ICU survivors median time to ICU discharge was shorter in the group randomized to PAV 73 days (IQR 52ndash114 d) ver-sus 124 days (IQR 75ndash308 d) in the PSV group (p = 003) Kaplan-Meier plots of the proportion of at-risk patients achiev-ing successful extubation and live ICU discharge are shown in Figure 3 A and B No difference was found for hospital length of stay ICU mortality or hospital mortality Four patients (15) in the PAV group and three patients (13) in the PSV group died in ICU (respiratory rate [RR] of PAV 114 95 CI 028 to 46 p = 086) Overall in-hospital mortality was nine patients (33) in the PAV group and six patients (26) in the PSV group (RR of PAV 128 95 CI 054 to 31 p = 058)

Cointerventions Sedating MedicationsTable 4 compares the use of medications that may prolong weaning (by causing sedation or neuromuscular weakness) in the two groups pre- and postrandomization Type and dose of sedative medication were highly variable among individual patients and therefore we analyzed pre- and postrandomiza-tion changes to determine the impact of the ventilation pro-tocol on sedative use within each patient In the PAV group benzodiazepine narcotic and propofol use measured as both cumulative dose and average daily dose all decreased signifi-cantly postrandomization as did neuromuscular blocker use In the PSV group average daily dose of lorazepam equiva-lent and average daily dose of propofol decreased signifi-cantly postrandomization whereas the proportion of patients

Figure 2 A Asynchrony index pre- and postrandomization on proportional assist ventilation B Asynchrony index pre- and postrandomization on pressure support ventilationPAV群 PSV群

Asynchrony Index () = 同調イベントの回数全呼吸回数 times 100

IntensiveCareMed(2006)321515ndash1522

AI>10は3名いたがPAVに設定変更後0名となった

AI>10は6名いたがPSVに設定変更後4名となった

Clinical Outcomes

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

成功したSBTまでの数抜管成功までの数に変わりなしICU退室までの数はPAV群で有意に短い

Kaplan-Meier curves

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

抜管成功までの時間 ICU退室までの時間

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 9

TABLE 4 Comparison of Medications That May Alter Weaning Duration Pre- and Postrandomization

Variables

PAV Group (n = 27)

PSV Group (n = 23) Postrandomization

Pre randomization Pos trandomization Pre randomization Pos trandomizationAdjusted Difference PSV ndash PAV (95 CI)

Lorazepam equivalents

Cumulative dose mg

308 (207) 48 (257)a 137 (440) 54 (258) 237 (ndash580 to 1055)

Average daily dose mg

44 (255) 09 (43)a 23 (85) 06 (40)a 24 (ndash62 to 110)

Single-day dose mgb

08 (72) 12 (50) 16 (70) 10 (64) 32 (ndash13 to 78)

Morphine equivalents

Cumulative dose mg

5003 (6188) 1140 (3140)b 1710 (13310) 1610 (4860) 591 (ndash2515 to 3696)

Average daily dose mg

858 (1638) 170 (578)c 428 (1724) 136 (320)d 108 (ndash316 to 532)

Single-day dose mga

180 (1240) 187 (77) 100 (1920) 50 (630)d 334 (ndash026 to 670)d

Propofol

Cumulative dose mg

7400 (55500) 0 (700)a 29450 (85800) 0 (27660) 226 (ndash42156 to 42607)

Average daily dose mg

529 (11690) 0 (53)a 4879 (15980) 0 (2738)a 651 (ndash3211 to 4514)

Single-day dose mga

0 (0) 0 (0) 0 (1000) 0 (0) 344 (ndash3781 to 4469)

Neuroleptics

Days receiving neuroleptics n

0 (2) 0 (4) 0 (0) 1 (4)a ndash16 (ndash34 to 03)

Proportion of patients who received a neuroleptic

407 407 174 565b

NMBs

Days receiving NMBs n

1 (1) 0 (0)a 0 (1) 0 (0) ndash01 (ndash05 to 03)

Proportion of patients who received an NMB

518 148a 348 174

Corticosteroids

Days receiving corticosteroids n

0 (0) 0 (0) 0 (2) 0 (2) 02 (ndash06 to 10)

Proportion of patients who received corticosteroids

185 148 348 261

PAV = proportional assist ventilation PSV = pressure support ventilation NMB = neuromuscular blockera p lt 005b Single-day dose = total dose given day prior to randomization or total dose give day after randomizationc p lt 0005d p lt 00005e p le 007All data points except percentage of patients receiving neuroleptics NMBs or corticosteroids are presented as median (interquartile range)

鎮静剤鎮痛剤についてはPAV群とPSV群で変わりなし

その他bull 不整脈気胸プロトコル施中の死亡

は両群とも起こらなかった

筆者らの考察bull PSVはweaningに最も使われているモードだ

がそれと較しても遜ないweaningができることがわかった

bull アシストが少ない傾向であったがPAV群のが呼吸窮迫となる回数が少なかった

rarr同調性が良くなった結果の可能性

ICU退室までの期間はPAV群で短いという結果であったがpowerが少ない

筆者らの考察bull 単施設での結果であり他の施設にこの

新しいプロトコルそのものが受けれられるかは不明である

bull SBTに失敗した患者を主に対象としておりICU患者すべてに適応はできない

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる

Page 32: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

RESULTS

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial

エンロール期間は42か80名が基準を満たしが最終的に割付されたのは54名

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial最終的に解析がわれたのはPAV群27名PSV群23名

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

ベースラインは両群間で変わりなし重症度はAPACHEⅡ 27点程度患者は肺炎が3割もともとCOPDを持っている患者は2割程度

Feasibility Outcomes bull 呼吸窮迫でRRTが呼吸器サポートを上昇

させた回数はPAV群で76回232daysPSV群で124回295daysと有意にPSV群で多かった(p=0002)

bull ACに戻した回数はPAV群で66回PSV群で90回と変わりなし(p=061)

bull 全くACに戻さなかった患者はPAV群で627名PSV群で623名(p=10)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

ベースラインは両群間で変わりなしPAVにすることでΔP圧Peak圧が低く吸気時間呼吸時間がくなるPSV群と較するとPAV群のがPaoAreaPaoAreaVE(呼吸器のサポートの強さをす)が低くVTPaoAreaがくsecond post-randomizationの解析においてΔP圧Peak圧が低い

Asynchrony index

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 7

B635) documenting compliance with the daily screening for SBT noncompletion of either the CPAP screening test or SBT occurred on 5 of 292 of patient study days (17) and this violation involved two patients in each study arm The second audit examining for nine different potential deviations found complete adherence to study protocol on 254 of 292 of patient study days (900) There were no adverse events (pneumo-thorax arrhythmia or death) resulting from any of the study interventions

Physiologic Response to Ventilation ProtocolsRecordings of flow and airway pressure were available for analysis for all 50 patients prerandomization 49 patients immediately (median 1 interquartile range [IQR] 0ndash15 hr) postrandomization and for 26 patients during subsequent days (median 27 IQR 20ndash70-hr postrandomization) Approximately 15000 breaths in 125 recordings were analyzed Table 2 compares the respiratory parameters in both groups pre- and postrandomization There were no statistically sig-nificant differences between the PAV and PSV groups at base-line prerandomization (p values not shown) In the PAV group there was a significant decrease in peak airway pressure delta pressure (measured as peak airway opening pressure minus the PEEP) and area under the inspiratory pressure curve immediately postrandomization and an increase in the per-cent of total cycle time spent in inspiration (TiTtot) whereas there were no significant differences immediately postran-domization in the PSV group The second measurement of postrandomization respiratory parameters revealed signifi-cant decreases in peak airway pressure and delta pressure in both PAV and PSV groups Adjusting for baseline measure-ments as a covariate the PAV group received 7 cm H2

O∙s less ventilator assistance (PaoArea) per liter of minute ventilation (p lt 00001) and increased their VT by 115 mLcm H

2O∙s

relative to the PSV group (p lt 0001)PAV significantly reduced AI postrandomization (Fig 2A)

In total nine patients (18) had a high AI (gt 10 of breaths) on PSV or on AC mode at baseline No patients had a high AI on PAV In the PSV group six patients (26) had a high AI

at baseline four (17) immediately postrandomization and one patient (7) after further adjustment of ventilator settings according to the PSV protocol (Fig 2B) Adjusting for baseline differences patients in the PAV group tended to have a lower AI than patients in the PSV group immediately postrandom-ization but this did not reach statistical significance (adjusted difference ndash45 [95 CI ndash95 to 04] p = 007)

Clinical OutcomesTable 3 shows the results for the main clinical endpoints Weaning duration and time to successful SBT did not dif-fer significantly among survivors between the groups Among ICU survivors median time to ICU discharge was shorter in the group randomized to PAV 73 days (IQR 52ndash114 d) ver-sus 124 days (IQR 75ndash308 d) in the PSV group (p = 003) Kaplan-Meier plots of the proportion of at-risk patients achiev-ing successful extubation and live ICU discharge are shown in Figure 3 A and B No difference was found for hospital length of stay ICU mortality or hospital mortality Four patients (15) in the PAV group and three patients (13) in the PSV group died in ICU (respiratory rate [RR] of PAV 114 95 CI 028 to 46 p = 086) Overall in-hospital mortality was nine patients (33) in the PAV group and six patients (26) in the PSV group (RR of PAV 128 95 CI 054 to 31 p = 058)

Cointerventions Sedating MedicationsTable 4 compares the use of medications that may prolong weaning (by causing sedation or neuromuscular weakness) in the two groups pre- and postrandomization Type and dose of sedative medication were highly variable among individual patients and therefore we analyzed pre- and postrandomiza-tion changes to determine the impact of the ventilation pro-tocol on sedative use within each patient In the PAV group benzodiazepine narcotic and propofol use measured as both cumulative dose and average daily dose all decreased signifi-cantly postrandomization as did neuromuscular blocker use In the PSV group average daily dose of lorazepam equiva-lent and average daily dose of propofol decreased signifi-cantly postrandomization whereas the proportion of patients

Figure 2 A Asynchrony index pre- and postrandomization on proportional assist ventilation B Asynchrony index pre- and postrandomization on pressure support ventilationPAV群 PSV群

Asynchrony Index () = 同調イベントの回数全呼吸回数 times 100

IntensiveCareMed(2006)321515ndash1522

AI>10は3名いたがPAVに設定変更後0名となった

AI>10は6名いたがPSVに設定変更後4名となった

Clinical Outcomes

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

成功したSBTまでの数抜管成功までの数に変わりなしICU退室までの数はPAV群で有意に短い

Kaplan-Meier curves

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

抜管成功までの時間 ICU退室までの時間

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 9

TABLE 4 Comparison of Medications That May Alter Weaning Duration Pre- and Postrandomization

Variables

PAV Group (n = 27)

PSV Group (n = 23) Postrandomization

Pre randomization Pos trandomization Pre randomization Pos trandomizationAdjusted Difference PSV ndash PAV (95 CI)

Lorazepam equivalents

Cumulative dose mg

308 (207) 48 (257)a 137 (440) 54 (258) 237 (ndash580 to 1055)

Average daily dose mg

44 (255) 09 (43)a 23 (85) 06 (40)a 24 (ndash62 to 110)

Single-day dose mgb

08 (72) 12 (50) 16 (70) 10 (64) 32 (ndash13 to 78)

Morphine equivalents

Cumulative dose mg

5003 (6188) 1140 (3140)b 1710 (13310) 1610 (4860) 591 (ndash2515 to 3696)

Average daily dose mg

858 (1638) 170 (578)c 428 (1724) 136 (320)d 108 (ndash316 to 532)

Single-day dose mga

180 (1240) 187 (77) 100 (1920) 50 (630)d 334 (ndash026 to 670)d

Propofol

Cumulative dose mg

7400 (55500) 0 (700)a 29450 (85800) 0 (27660) 226 (ndash42156 to 42607)

Average daily dose mg

529 (11690) 0 (53)a 4879 (15980) 0 (2738)a 651 (ndash3211 to 4514)

Single-day dose mga

0 (0) 0 (0) 0 (1000) 0 (0) 344 (ndash3781 to 4469)

Neuroleptics

Days receiving neuroleptics n

0 (2) 0 (4) 0 (0) 1 (4)a ndash16 (ndash34 to 03)

Proportion of patients who received a neuroleptic

407 407 174 565b

NMBs

Days receiving NMBs n

1 (1) 0 (0)a 0 (1) 0 (0) ndash01 (ndash05 to 03)

Proportion of patients who received an NMB

518 148a 348 174

Corticosteroids

Days receiving corticosteroids n

0 (0) 0 (0) 0 (2) 0 (2) 02 (ndash06 to 10)

Proportion of patients who received corticosteroids

185 148 348 261

PAV = proportional assist ventilation PSV = pressure support ventilation NMB = neuromuscular blockera p lt 005b Single-day dose = total dose given day prior to randomization or total dose give day after randomizationc p lt 0005d p lt 00005e p le 007All data points except percentage of patients receiving neuroleptics NMBs or corticosteroids are presented as median (interquartile range)

鎮静剤鎮痛剤についてはPAV群とPSV群で変わりなし

その他bull 不整脈気胸プロトコル施中の死亡

は両群とも起こらなかった

筆者らの考察bull PSVはweaningに最も使われているモードだ

がそれと較しても遜ないweaningができることがわかった

bull アシストが少ない傾向であったがPAV群のが呼吸窮迫となる回数が少なかった

rarr同調性が良くなった結果の可能性

ICU退室までの期間はPAV群で短いという結果であったがpowerが少ない

筆者らの考察bull 単施設での結果であり他の施設にこの

新しいプロトコルそのものが受けれられるかは不明である

bull SBTに失敗した患者を主に対象としておりICU患者すべてに適応はできない

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる

Page 33: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial

エンロール期間は42か80名が基準を満たしが最終的に割付されたのは54名

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial最終的に解析がわれたのはPAV群27名PSV群23名

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

ベースラインは両群間で変わりなし重症度はAPACHEⅡ 27点程度患者は肺炎が3割もともとCOPDを持っている患者は2割程度

Feasibility Outcomes bull 呼吸窮迫でRRTが呼吸器サポートを上昇

させた回数はPAV群で76回232daysPSV群で124回295daysと有意にPSV群で多かった(p=0002)

bull ACに戻した回数はPAV群で66回PSV群で90回と変わりなし(p=061)

bull 全くACに戻さなかった患者はPAV群で627名PSV群で623名(p=10)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

ベースラインは両群間で変わりなしPAVにすることでΔP圧Peak圧が低く吸気時間呼吸時間がくなるPSV群と較するとPAV群のがPaoAreaPaoAreaVE(呼吸器のサポートの強さをす)が低くVTPaoAreaがくsecond post-randomizationの解析においてΔP圧Peak圧が低い

Asynchrony index

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 7

B635) documenting compliance with the daily screening for SBT noncompletion of either the CPAP screening test or SBT occurred on 5 of 292 of patient study days (17) and this violation involved two patients in each study arm The second audit examining for nine different potential deviations found complete adherence to study protocol on 254 of 292 of patient study days (900) There were no adverse events (pneumo-thorax arrhythmia or death) resulting from any of the study interventions

Physiologic Response to Ventilation ProtocolsRecordings of flow and airway pressure were available for analysis for all 50 patients prerandomization 49 patients immediately (median 1 interquartile range [IQR] 0ndash15 hr) postrandomization and for 26 patients during subsequent days (median 27 IQR 20ndash70-hr postrandomization) Approximately 15000 breaths in 125 recordings were analyzed Table 2 compares the respiratory parameters in both groups pre- and postrandomization There were no statistically sig-nificant differences between the PAV and PSV groups at base-line prerandomization (p values not shown) In the PAV group there was a significant decrease in peak airway pressure delta pressure (measured as peak airway opening pressure minus the PEEP) and area under the inspiratory pressure curve immediately postrandomization and an increase in the per-cent of total cycle time spent in inspiration (TiTtot) whereas there were no significant differences immediately postran-domization in the PSV group The second measurement of postrandomization respiratory parameters revealed signifi-cant decreases in peak airway pressure and delta pressure in both PAV and PSV groups Adjusting for baseline measure-ments as a covariate the PAV group received 7 cm H2

O∙s less ventilator assistance (PaoArea) per liter of minute ventilation (p lt 00001) and increased their VT by 115 mLcm H

2O∙s

relative to the PSV group (p lt 0001)PAV significantly reduced AI postrandomization (Fig 2A)

In total nine patients (18) had a high AI (gt 10 of breaths) on PSV or on AC mode at baseline No patients had a high AI on PAV In the PSV group six patients (26) had a high AI

at baseline four (17) immediately postrandomization and one patient (7) after further adjustment of ventilator settings according to the PSV protocol (Fig 2B) Adjusting for baseline differences patients in the PAV group tended to have a lower AI than patients in the PSV group immediately postrandom-ization but this did not reach statistical significance (adjusted difference ndash45 [95 CI ndash95 to 04] p = 007)

Clinical OutcomesTable 3 shows the results for the main clinical endpoints Weaning duration and time to successful SBT did not dif-fer significantly among survivors between the groups Among ICU survivors median time to ICU discharge was shorter in the group randomized to PAV 73 days (IQR 52ndash114 d) ver-sus 124 days (IQR 75ndash308 d) in the PSV group (p = 003) Kaplan-Meier plots of the proportion of at-risk patients achiev-ing successful extubation and live ICU discharge are shown in Figure 3 A and B No difference was found for hospital length of stay ICU mortality or hospital mortality Four patients (15) in the PAV group and three patients (13) in the PSV group died in ICU (respiratory rate [RR] of PAV 114 95 CI 028 to 46 p = 086) Overall in-hospital mortality was nine patients (33) in the PAV group and six patients (26) in the PSV group (RR of PAV 128 95 CI 054 to 31 p = 058)

Cointerventions Sedating MedicationsTable 4 compares the use of medications that may prolong weaning (by causing sedation or neuromuscular weakness) in the two groups pre- and postrandomization Type and dose of sedative medication were highly variable among individual patients and therefore we analyzed pre- and postrandomiza-tion changes to determine the impact of the ventilation pro-tocol on sedative use within each patient In the PAV group benzodiazepine narcotic and propofol use measured as both cumulative dose and average daily dose all decreased signifi-cantly postrandomization as did neuromuscular blocker use In the PSV group average daily dose of lorazepam equiva-lent and average daily dose of propofol decreased signifi-cantly postrandomization whereas the proportion of patients

Figure 2 A Asynchrony index pre- and postrandomization on proportional assist ventilation B Asynchrony index pre- and postrandomization on pressure support ventilationPAV群 PSV群

Asynchrony Index () = 同調イベントの回数全呼吸回数 times 100

IntensiveCareMed(2006)321515ndash1522

AI>10は3名いたがPAVに設定変更後0名となった

AI>10は6名いたがPSVに設定変更後4名となった

Clinical Outcomes

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

成功したSBTまでの数抜管成功までの数に変わりなしICU退室までの数はPAV群で有意に短い

Kaplan-Meier curves

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

抜管成功までの時間 ICU退室までの時間

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 9

TABLE 4 Comparison of Medications That May Alter Weaning Duration Pre- and Postrandomization

Variables

PAV Group (n = 27)

PSV Group (n = 23) Postrandomization

Pre randomization Pos trandomization Pre randomization Pos trandomizationAdjusted Difference PSV ndash PAV (95 CI)

Lorazepam equivalents

Cumulative dose mg

308 (207) 48 (257)a 137 (440) 54 (258) 237 (ndash580 to 1055)

Average daily dose mg

44 (255) 09 (43)a 23 (85) 06 (40)a 24 (ndash62 to 110)

Single-day dose mgb

08 (72) 12 (50) 16 (70) 10 (64) 32 (ndash13 to 78)

Morphine equivalents

Cumulative dose mg

5003 (6188) 1140 (3140)b 1710 (13310) 1610 (4860) 591 (ndash2515 to 3696)

Average daily dose mg

858 (1638) 170 (578)c 428 (1724) 136 (320)d 108 (ndash316 to 532)

Single-day dose mga

180 (1240) 187 (77) 100 (1920) 50 (630)d 334 (ndash026 to 670)d

Propofol

Cumulative dose mg

7400 (55500) 0 (700)a 29450 (85800) 0 (27660) 226 (ndash42156 to 42607)

Average daily dose mg

529 (11690) 0 (53)a 4879 (15980) 0 (2738)a 651 (ndash3211 to 4514)

Single-day dose mga

0 (0) 0 (0) 0 (1000) 0 (0) 344 (ndash3781 to 4469)

Neuroleptics

Days receiving neuroleptics n

0 (2) 0 (4) 0 (0) 1 (4)a ndash16 (ndash34 to 03)

Proportion of patients who received a neuroleptic

407 407 174 565b

NMBs

Days receiving NMBs n

1 (1) 0 (0)a 0 (1) 0 (0) ndash01 (ndash05 to 03)

Proportion of patients who received an NMB

518 148a 348 174

Corticosteroids

Days receiving corticosteroids n

0 (0) 0 (0) 0 (2) 0 (2) 02 (ndash06 to 10)

Proportion of patients who received corticosteroids

185 148 348 261

PAV = proportional assist ventilation PSV = pressure support ventilation NMB = neuromuscular blockera p lt 005b Single-day dose = total dose given day prior to randomization or total dose give day after randomizationc p lt 0005d p lt 00005e p le 007All data points except percentage of patients receiving neuroleptics NMBs or corticosteroids are presented as median (interquartile range)

鎮静剤鎮痛剤についてはPAV群とPSV群で変わりなし

その他bull 不整脈気胸プロトコル施中の死亡

は両群とも起こらなかった

筆者らの考察bull PSVはweaningに最も使われているモードだ

がそれと較しても遜ないweaningができることがわかった

bull アシストが少ない傾向であったがPAV群のが呼吸窮迫となる回数が少なかった

rarr同調性が良くなった結果の可能性

ICU退室までの期間はPAV群で短いという結果であったがpowerが少ない

筆者らの考察bull 単施設での結果であり他の施設にこの

新しいプロトコルそのものが受けれられるかは不明である

bull SBTに失敗した患者を主に対象としておりICU患者すべてに適応はできない

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる

Page 34: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

割付

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 3

ventilation protocols In both protocols registered respiratory therapists (RRTs) used stepwise reductions in level of sup-port to determine the lowest level tolerated in order to avoid both overassistance and respiratory muscle fatigue If respi-ratory distress developed according to the predefined criteria (Supplemental Table 1 Supplemental Digital Content 1 httplinkslwwcomCCMB636) the level of support was increased If distress or respiratory acidosis was unresolved despite maximum levels of support patients were switch to assistcontrol (AC) mode and reassessed daily to resume PAV or PSV Patients in the PAV group could be ventilated using PAV+ or AC only and patients in the PSV group could be ventilated using PSV or AC only In both arms weaning was

conducted by the RRTs with daily screening and assessment for SBTs (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCMB635) Level of support was not used as an indicator of readiness to wean The RSBI was con-ducted on continuous positive airway pressure (CPAP) set at the prior level of positive end-expiratory pressure (PEEP) SBTs were conducted on PEEP of 5 cm H

2O plus PSV of 5ndash6 cm

H2O or PAV of 10ndash20 gain respectively

Strategies to Enhance Protocol CompliancePrior to the start of this RCT the ventilation protocols were trialed and modified with input and collaboration from RRTs during an observational study of 15 patients

Figure 1 Flow of patients through the study MD = medical doctor PAV = proportional assist ventilation PSV = pressure support ventilation SBT = spontaneous breathing trial最終的に解析がわれたのはPAV群27名PSV群23名

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

ベースラインは両群間で変わりなし重症度はAPACHEⅡ 27点程度患者は肺炎が3割もともとCOPDを持っている患者は2割程度

Feasibility Outcomes bull 呼吸窮迫でRRTが呼吸器サポートを上昇

させた回数はPAV群で76回232daysPSV群で124回295daysと有意にPSV群で多かった(p=0002)

bull ACに戻した回数はPAV群で66回PSV群で90回と変わりなし(p=061)

bull 全くACに戻さなかった患者はPAV群で627名PSV群で623名(p=10)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

ベースラインは両群間で変わりなしPAVにすることでΔP圧Peak圧が低く吸気時間呼吸時間がくなるPSV群と較するとPAV群のがPaoAreaPaoAreaVE(呼吸器のサポートの強さをす)が低くVTPaoAreaがくsecond post-randomizationの解析においてΔP圧Peak圧が低い

Asynchrony index

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 7

B635) documenting compliance with the daily screening for SBT noncompletion of either the CPAP screening test or SBT occurred on 5 of 292 of patient study days (17) and this violation involved two patients in each study arm The second audit examining for nine different potential deviations found complete adherence to study protocol on 254 of 292 of patient study days (900) There were no adverse events (pneumo-thorax arrhythmia or death) resulting from any of the study interventions

Physiologic Response to Ventilation ProtocolsRecordings of flow and airway pressure were available for analysis for all 50 patients prerandomization 49 patients immediately (median 1 interquartile range [IQR] 0ndash15 hr) postrandomization and for 26 patients during subsequent days (median 27 IQR 20ndash70-hr postrandomization) Approximately 15000 breaths in 125 recordings were analyzed Table 2 compares the respiratory parameters in both groups pre- and postrandomization There were no statistically sig-nificant differences between the PAV and PSV groups at base-line prerandomization (p values not shown) In the PAV group there was a significant decrease in peak airway pressure delta pressure (measured as peak airway opening pressure minus the PEEP) and area under the inspiratory pressure curve immediately postrandomization and an increase in the per-cent of total cycle time spent in inspiration (TiTtot) whereas there were no significant differences immediately postran-domization in the PSV group The second measurement of postrandomization respiratory parameters revealed signifi-cant decreases in peak airway pressure and delta pressure in both PAV and PSV groups Adjusting for baseline measure-ments as a covariate the PAV group received 7 cm H2

O∙s less ventilator assistance (PaoArea) per liter of minute ventilation (p lt 00001) and increased their VT by 115 mLcm H

2O∙s

relative to the PSV group (p lt 0001)PAV significantly reduced AI postrandomization (Fig 2A)

In total nine patients (18) had a high AI (gt 10 of breaths) on PSV or on AC mode at baseline No patients had a high AI on PAV In the PSV group six patients (26) had a high AI

at baseline four (17) immediately postrandomization and one patient (7) after further adjustment of ventilator settings according to the PSV protocol (Fig 2B) Adjusting for baseline differences patients in the PAV group tended to have a lower AI than patients in the PSV group immediately postrandom-ization but this did not reach statistical significance (adjusted difference ndash45 [95 CI ndash95 to 04] p = 007)

Clinical OutcomesTable 3 shows the results for the main clinical endpoints Weaning duration and time to successful SBT did not dif-fer significantly among survivors between the groups Among ICU survivors median time to ICU discharge was shorter in the group randomized to PAV 73 days (IQR 52ndash114 d) ver-sus 124 days (IQR 75ndash308 d) in the PSV group (p = 003) Kaplan-Meier plots of the proportion of at-risk patients achiev-ing successful extubation and live ICU discharge are shown in Figure 3 A and B No difference was found for hospital length of stay ICU mortality or hospital mortality Four patients (15) in the PAV group and three patients (13) in the PSV group died in ICU (respiratory rate [RR] of PAV 114 95 CI 028 to 46 p = 086) Overall in-hospital mortality was nine patients (33) in the PAV group and six patients (26) in the PSV group (RR of PAV 128 95 CI 054 to 31 p = 058)

Cointerventions Sedating MedicationsTable 4 compares the use of medications that may prolong weaning (by causing sedation or neuromuscular weakness) in the two groups pre- and postrandomization Type and dose of sedative medication were highly variable among individual patients and therefore we analyzed pre- and postrandomiza-tion changes to determine the impact of the ventilation pro-tocol on sedative use within each patient In the PAV group benzodiazepine narcotic and propofol use measured as both cumulative dose and average daily dose all decreased signifi-cantly postrandomization as did neuromuscular blocker use In the PSV group average daily dose of lorazepam equiva-lent and average daily dose of propofol decreased signifi-cantly postrandomization whereas the proportion of patients

Figure 2 A Asynchrony index pre- and postrandomization on proportional assist ventilation B Asynchrony index pre- and postrandomization on pressure support ventilationPAV群 PSV群

Asynchrony Index () = 同調イベントの回数全呼吸回数 times 100

IntensiveCareMed(2006)321515ndash1522

AI>10は3名いたがPAVに設定変更後0名となった

AI>10は6名いたがPSVに設定変更後4名となった

Clinical Outcomes

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

成功したSBTまでの数抜管成功までの数に変わりなしICU退室までの数はPAV群で有意に短い

Kaplan-Meier curves

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

抜管成功までの時間 ICU退室までの時間

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 9

TABLE 4 Comparison of Medications That May Alter Weaning Duration Pre- and Postrandomization

Variables

PAV Group (n = 27)

PSV Group (n = 23) Postrandomization

Pre randomization Pos trandomization Pre randomization Pos trandomizationAdjusted Difference PSV ndash PAV (95 CI)

Lorazepam equivalents

Cumulative dose mg

308 (207) 48 (257)a 137 (440) 54 (258) 237 (ndash580 to 1055)

Average daily dose mg

44 (255) 09 (43)a 23 (85) 06 (40)a 24 (ndash62 to 110)

Single-day dose mgb

08 (72) 12 (50) 16 (70) 10 (64) 32 (ndash13 to 78)

Morphine equivalents

Cumulative dose mg

5003 (6188) 1140 (3140)b 1710 (13310) 1610 (4860) 591 (ndash2515 to 3696)

Average daily dose mg

858 (1638) 170 (578)c 428 (1724) 136 (320)d 108 (ndash316 to 532)

Single-day dose mga

180 (1240) 187 (77) 100 (1920) 50 (630)d 334 (ndash026 to 670)d

Propofol

Cumulative dose mg

7400 (55500) 0 (700)a 29450 (85800) 0 (27660) 226 (ndash42156 to 42607)

Average daily dose mg

529 (11690) 0 (53)a 4879 (15980) 0 (2738)a 651 (ndash3211 to 4514)

Single-day dose mga

0 (0) 0 (0) 0 (1000) 0 (0) 344 (ndash3781 to 4469)

Neuroleptics

Days receiving neuroleptics n

0 (2) 0 (4) 0 (0) 1 (4)a ndash16 (ndash34 to 03)

Proportion of patients who received a neuroleptic

407 407 174 565b

NMBs

Days receiving NMBs n

1 (1) 0 (0)a 0 (1) 0 (0) ndash01 (ndash05 to 03)

Proportion of patients who received an NMB

518 148a 348 174

Corticosteroids

Days receiving corticosteroids n

0 (0) 0 (0) 0 (2) 0 (2) 02 (ndash06 to 10)

Proportion of patients who received corticosteroids

185 148 348 261

PAV = proportional assist ventilation PSV = pressure support ventilation NMB = neuromuscular blockera p lt 005b Single-day dose = total dose given day prior to randomization or total dose give day after randomizationc p lt 0005d p lt 00005e p le 007All data points except percentage of patients receiving neuroleptics NMBs or corticosteroids are presented as median (interquartile range)

鎮静剤鎮痛剤についてはPAV群とPSV群で変わりなし

その他bull 不整脈気胸プロトコル施中の死亡

は両群とも起こらなかった

筆者らの考察bull PSVはweaningに最も使われているモードだ

がそれと較しても遜ないweaningができることがわかった

bull アシストが少ない傾向であったがPAV群のが呼吸窮迫となる回数が少なかった

rarr同調性が良くなった結果の可能性

ICU退室までの期間はPAV群で短いという結果であったがpowerが少ない

筆者らの考察bull 単施設での結果であり他の施設にこの

新しいプロトコルそのものが受けれられるかは不明である

bull SBTに失敗した患者を主に対象としておりICU患者すべてに適応はできない

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる

Page 35: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

ベースラインは両群間で変わりなし重症度はAPACHEⅡ 27点程度患者は肺炎が3割もともとCOPDを持っている患者は2割程度

Feasibility Outcomes bull 呼吸窮迫でRRTが呼吸器サポートを上昇

させた回数はPAV群で76回232daysPSV群で124回295daysと有意にPSV群で多かった(p=0002)

bull ACに戻した回数はPAV群で66回PSV群で90回と変わりなし(p=061)

bull 全くACに戻さなかった患者はPAV群で627名PSV群で623名(p=10)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

ベースラインは両群間で変わりなしPAVにすることでΔP圧Peak圧が低く吸気時間呼吸時間がくなるPSV群と較するとPAV群のがPaoAreaPaoAreaVE(呼吸器のサポートの強さをす)が低くVTPaoAreaがくsecond post-randomizationの解析においてΔP圧Peak圧が低い

Asynchrony index

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 7

B635) documenting compliance with the daily screening for SBT noncompletion of either the CPAP screening test or SBT occurred on 5 of 292 of patient study days (17) and this violation involved two patients in each study arm The second audit examining for nine different potential deviations found complete adherence to study protocol on 254 of 292 of patient study days (900) There were no adverse events (pneumo-thorax arrhythmia or death) resulting from any of the study interventions

Physiologic Response to Ventilation ProtocolsRecordings of flow and airway pressure were available for analysis for all 50 patients prerandomization 49 patients immediately (median 1 interquartile range [IQR] 0ndash15 hr) postrandomization and for 26 patients during subsequent days (median 27 IQR 20ndash70-hr postrandomization) Approximately 15000 breaths in 125 recordings were analyzed Table 2 compares the respiratory parameters in both groups pre- and postrandomization There were no statistically sig-nificant differences between the PAV and PSV groups at base-line prerandomization (p values not shown) In the PAV group there was a significant decrease in peak airway pressure delta pressure (measured as peak airway opening pressure minus the PEEP) and area under the inspiratory pressure curve immediately postrandomization and an increase in the per-cent of total cycle time spent in inspiration (TiTtot) whereas there were no significant differences immediately postran-domization in the PSV group The second measurement of postrandomization respiratory parameters revealed signifi-cant decreases in peak airway pressure and delta pressure in both PAV and PSV groups Adjusting for baseline measure-ments as a covariate the PAV group received 7 cm H2

O∙s less ventilator assistance (PaoArea) per liter of minute ventilation (p lt 00001) and increased their VT by 115 mLcm H

2O∙s

relative to the PSV group (p lt 0001)PAV significantly reduced AI postrandomization (Fig 2A)

In total nine patients (18) had a high AI (gt 10 of breaths) on PSV or on AC mode at baseline No patients had a high AI on PAV In the PSV group six patients (26) had a high AI

at baseline four (17) immediately postrandomization and one patient (7) after further adjustment of ventilator settings according to the PSV protocol (Fig 2B) Adjusting for baseline differences patients in the PAV group tended to have a lower AI than patients in the PSV group immediately postrandom-ization but this did not reach statistical significance (adjusted difference ndash45 [95 CI ndash95 to 04] p = 007)

Clinical OutcomesTable 3 shows the results for the main clinical endpoints Weaning duration and time to successful SBT did not dif-fer significantly among survivors between the groups Among ICU survivors median time to ICU discharge was shorter in the group randomized to PAV 73 days (IQR 52ndash114 d) ver-sus 124 days (IQR 75ndash308 d) in the PSV group (p = 003) Kaplan-Meier plots of the proportion of at-risk patients achiev-ing successful extubation and live ICU discharge are shown in Figure 3 A and B No difference was found for hospital length of stay ICU mortality or hospital mortality Four patients (15) in the PAV group and three patients (13) in the PSV group died in ICU (respiratory rate [RR] of PAV 114 95 CI 028 to 46 p = 086) Overall in-hospital mortality was nine patients (33) in the PAV group and six patients (26) in the PSV group (RR of PAV 128 95 CI 054 to 31 p = 058)

Cointerventions Sedating MedicationsTable 4 compares the use of medications that may prolong weaning (by causing sedation or neuromuscular weakness) in the two groups pre- and postrandomization Type and dose of sedative medication were highly variable among individual patients and therefore we analyzed pre- and postrandomiza-tion changes to determine the impact of the ventilation pro-tocol on sedative use within each patient In the PAV group benzodiazepine narcotic and propofol use measured as both cumulative dose and average daily dose all decreased signifi-cantly postrandomization as did neuromuscular blocker use In the PSV group average daily dose of lorazepam equiva-lent and average daily dose of propofol decreased signifi-cantly postrandomization whereas the proportion of patients

Figure 2 A Asynchrony index pre- and postrandomization on proportional assist ventilation B Asynchrony index pre- and postrandomization on pressure support ventilationPAV群 PSV群

Asynchrony Index () = 同調イベントの回数全呼吸回数 times 100

IntensiveCareMed(2006)321515ndash1522

AI>10は3名いたがPAVに設定変更後0名となった

AI>10は6名いたがPSVに設定変更後4名となった

Clinical Outcomes

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

成功したSBTまでの数抜管成功までの数に変わりなしICU退室までの数はPAV群で有意に短い

Kaplan-Meier curves

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

抜管成功までの時間 ICU退室までの時間

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 9

TABLE 4 Comparison of Medications That May Alter Weaning Duration Pre- and Postrandomization

Variables

PAV Group (n = 27)

PSV Group (n = 23) Postrandomization

Pre randomization Pos trandomization Pre randomization Pos trandomizationAdjusted Difference PSV ndash PAV (95 CI)

Lorazepam equivalents

Cumulative dose mg

308 (207) 48 (257)a 137 (440) 54 (258) 237 (ndash580 to 1055)

Average daily dose mg

44 (255) 09 (43)a 23 (85) 06 (40)a 24 (ndash62 to 110)

Single-day dose mgb

08 (72) 12 (50) 16 (70) 10 (64) 32 (ndash13 to 78)

Morphine equivalents

Cumulative dose mg

5003 (6188) 1140 (3140)b 1710 (13310) 1610 (4860) 591 (ndash2515 to 3696)

Average daily dose mg

858 (1638) 170 (578)c 428 (1724) 136 (320)d 108 (ndash316 to 532)

Single-day dose mga

180 (1240) 187 (77) 100 (1920) 50 (630)d 334 (ndash026 to 670)d

Propofol

Cumulative dose mg

7400 (55500) 0 (700)a 29450 (85800) 0 (27660) 226 (ndash42156 to 42607)

Average daily dose mg

529 (11690) 0 (53)a 4879 (15980) 0 (2738)a 651 (ndash3211 to 4514)

Single-day dose mga

0 (0) 0 (0) 0 (1000) 0 (0) 344 (ndash3781 to 4469)

Neuroleptics

Days receiving neuroleptics n

0 (2) 0 (4) 0 (0) 1 (4)a ndash16 (ndash34 to 03)

Proportion of patients who received a neuroleptic

407 407 174 565b

NMBs

Days receiving NMBs n

1 (1) 0 (0)a 0 (1) 0 (0) ndash01 (ndash05 to 03)

Proportion of patients who received an NMB

518 148a 348 174

Corticosteroids

Days receiving corticosteroids n

0 (0) 0 (0) 0 (2) 0 (2) 02 (ndash06 to 10)

Proportion of patients who received corticosteroids

185 148 348 261

PAV = proportional assist ventilation PSV = pressure support ventilation NMB = neuromuscular blockera p lt 005b Single-day dose = total dose given day prior to randomization or total dose give day after randomizationc p lt 0005d p lt 00005e p le 007All data points except percentage of patients receiving neuroleptics NMBs or corticosteroids are presented as median (interquartile range)

鎮静剤鎮痛剤についてはPAV群とPSV群で変わりなし

その他bull 不整脈気胸プロトコル施中の死亡

は両群とも起こらなかった

筆者らの考察bull PSVはweaningに最も使われているモードだ

がそれと較しても遜ないweaningができることがわかった

bull アシストが少ない傾向であったがPAV群のが呼吸窮迫となる回数が少なかった

rarr同調性が良くなった結果の可能性

ICU退室までの期間はPAV群で短いという結果であったがpowerが少ない

筆者らの考察bull 単施設での結果であり他の施設にこの

新しいプロトコルそのものが受けれられるかは不明である

bull SBTに失敗した患者を主に対象としておりICU患者すべてに適応はできない

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる

Page 36: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

Baseline Clinical Characteristics

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 5

TABLE 1 Baseline Clinical Characteristics of the Study Patients at Enrollment

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support

Ventilation Group (n = 23)

Age yr mean plusmn SD 63 plusmn 14 67 plusmn 12

Gender femalemale (female) 1512 (56) 1013 (43)

Body mass index mean plusmn SD 270 plusmn 82 305 plusmn 80

APACHE II score at admission 27 plusmn 9 26 plusmn 8

APACHE II score at randomization 19 plusmn 7 20 plusmn 5

Duration of mechanical ventilation at randomization d median (interquartile range)

63 (41ndash79) 52 (39ndash102)

Hospital admission type n ()

Medical 19 (70) 19 (83)

Surgical 8 (30) 4 (17)

ICU admission diagnosis n ()

Pneumonia 8 (30) 6 (26)

Sepsis nonrespiratory 3 (11) 5 (21)

Cardiac arrest 3 (11) 4 (17)

COPD exacerbation 0 (0) 2 (9)

Postoperative 4 (15) 1 (4)

Gastrointestinal bleed 1 (4) 1 (4)

Congestive heart failure 0 (0) 1 (4)

Acute respiratory distress syndrome 2 (7) 0 (0)

Hepatic encephalopathy 2 (7) 0 (0)

Other 4 (15) 3 (13)

Pulmonary comorbidities n ()

COPD 4 (15) 6 (26)

Restrictiveinterstitial lung disease 1 (4) 2 (9)

Asthma 0 (0) 1 (4)

Home oxygen therapy 1 (4) 3 (13)

Nonpulmonary comorbidities n ()

Diabetes mellitus 11 (41) 10 (43)

Congestive heart failure 3 (11) 3 (13)

Ischemic heart disease 2 (7) 3 (13)

Immunosuppression 4 (15) 2 (9)

Stroke 2 (7) 0 (0)

At least one comorbidity 16 (59) 19 (83)

APACHE chronic score gt 0 n () 17 (63) 14 (61)

APACHE = Acute Physiology and Chronic Health Evaluation COPD = chronic obstructive pulmonary disease

common reasons for switching to AC were similar between PAV and PSV groups and were ongoing distress at maxi-mum support (36 vs 33 events) sedated for procedure (10 vs 8 events) apnea (3 vs 8 events) and pH less than 732 (4 vs 2 events) respectively

An initial audit demonstrated that the most common protocol violation was noncompletion of the daily two-step weaning assessment After modification of the daily data collection form to include a checklist (Supplemental Fig 3 Supplemental Digital Content 4 httplinkslwwcomCCM

ベースラインは両群間で変わりなし重症度はAPACHEⅡ 27点程度患者は肺炎が3割もともとCOPDを持っている患者は2割程度

Feasibility Outcomes bull 呼吸窮迫でRRTが呼吸器サポートを上昇

させた回数はPAV群で76回232daysPSV群で124回295daysと有意にPSV群で多かった(p=0002)

bull ACに戻した回数はPAV群で66回PSV群で90回と変わりなし(p=061)

bull 全くACに戻さなかった患者はPAV群で627名PSV群で623名(p=10)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

ベースラインは両群間で変わりなしPAVにすることでΔP圧Peak圧が低く吸気時間呼吸時間がくなるPSV群と較するとPAV群のがPaoAreaPaoAreaVE(呼吸器のサポートの強さをす)が低くVTPaoAreaがくsecond post-randomizationの解析においてΔP圧Peak圧が低い

Asynchrony index

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 7

B635) documenting compliance with the daily screening for SBT noncompletion of either the CPAP screening test or SBT occurred on 5 of 292 of patient study days (17) and this violation involved two patients in each study arm The second audit examining for nine different potential deviations found complete adherence to study protocol on 254 of 292 of patient study days (900) There were no adverse events (pneumo-thorax arrhythmia or death) resulting from any of the study interventions

Physiologic Response to Ventilation ProtocolsRecordings of flow and airway pressure were available for analysis for all 50 patients prerandomization 49 patients immediately (median 1 interquartile range [IQR] 0ndash15 hr) postrandomization and for 26 patients during subsequent days (median 27 IQR 20ndash70-hr postrandomization) Approximately 15000 breaths in 125 recordings were analyzed Table 2 compares the respiratory parameters in both groups pre- and postrandomization There were no statistically sig-nificant differences between the PAV and PSV groups at base-line prerandomization (p values not shown) In the PAV group there was a significant decrease in peak airway pressure delta pressure (measured as peak airway opening pressure minus the PEEP) and area under the inspiratory pressure curve immediately postrandomization and an increase in the per-cent of total cycle time spent in inspiration (TiTtot) whereas there were no significant differences immediately postran-domization in the PSV group The second measurement of postrandomization respiratory parameters revealed signifi-cant decreases in peak airway pressure and delta pressure in both PAV and PSV groups Adjusting for baseline measure-ments as a covariate the PAV group received 7 cm H2

O∙s less ventilator assistance (PaoArea) per liter of minute ventilation (p lt 00001) and increased their VT by 115 mLcm H

2O∙s

relative to the PSV group (p lt 0001)PAV significantly reduced AI postrandomization (Fig 2A)

In total nine patients (18) had a high AI (gt 10 of breaths) on PSV or on AC mode at baseline No patients had a high AI on PAV In the PSV group six patients (26) had a high AI

at baseline four (17) immediately postrandomization and one patient (7) after further adjustment of ventilator settings according to the PSV protocol (Fig 2B) Adjusting for baseline differences patients in the PAV group tended to have a lower AI than patients in the PSV group immediately postrandom-ization but this did not reach statistical significance (adjusted difference ndash45 [95 CI ndash95 to 04] p = 007)

Clinical OutcomesTable 3 shows the results for the main clinical endpoints Weaning duration and time to successful SBT did not dif-fer significantly among survivors between the groups Among ICU survivors median time to ICU discharge was shorter in the group randomized to PAV 73 days (IQR 52ndash114 d) ver-sus 124 days (IQR 75ndash308 d) in the PSV group (p = 003) Kaplan-Meier plots of the proportion of at-risk patients achiev-ing successful extubation and live ICU discharge are shown in Figure 3 A and B No difference was found for hospital length of stay ICU mortality or hospital mortality Four patients (15) in the PAV group and three patients (13) in the PSV group died in ICU (respiratory rate [RR] of PAV 114 95 CI 028 to 46 p = 086) Overall in-hospital mortality was nine patients (33) in the PAV group and six patients (26) in the PSV group (RR of PAV 128 95 CI 054 to 31 p = 058)

Cointerventions Sedating MedicationsTable 4 compares the use of medications that may prolong weaning (by causing sedation or neuromuscular weakness) in the two groups pre- and postrandomization Type and dose of sedative medication were highly variable among individual patients and therefore we analyzed pre- and postrandomiza-tion changes to determine the impact of the ventilation pro-tocol on sedative use within each patient In the PAV group benzodiazepine narcotic and propofol use measured as both cumulative dose and average daily dose all decreased signifi-cantly postrandomization as did neuromuscular blocker use In the PSV group average daily dose of lorazepam equiva-lent and average daily dose of propofol decreased signifi-cantly postrandomization whereas the proportion of patients

Figure 2 A Asynchrony index pre- and postrandomization on proportional assist ventilation B Asynchrony index pre- and postrandomization on pressure support ventilationPAV群 PSV群

Asynchrony Index () = 同調イベントの回数全呼吸回数 times 100

IntensiveCareMed(2006)321515ndash1522

AI>10は3名いたがPAVに設定変更後0名となった

AI>10は6名いたがPSVに設定変更後4名となった

Clinical Outcomes

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

成功したSBTまでの数抜管成功までの数に変わりなしICU退室までの数はPAV群で有意に短い

Kaplan-Meier curves

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

抜管成功までの時間 ICU退室までの時間

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 9

TABLE 4 Comparison of Medications That May Alter Weaning Duration Pre- and Postrandomization

Variables

PAV Group (n = 27)

PSV Group (n = 23) Postrandomization

Pre randomization Pos trandomization Pre randomization Pos trandomizationAdjusted Difference PSV ndash PAV (95 CI)

Lorazepam equivalents

Cumulative dose mg

308 (207) 48 (257)a 137 (440) 54 (258) 237 (ndash580 to 1055)

Average daily dose mg

44 (255) 09 (43)a 23 (85) 06 (40)a 24 (ndash62 to 110)

Single-day dose mgb

08 (72) 12 (50) 16 (70) 10 (64) 32 (ndash13 to 78)

Morphine equivalents

Cumulative dose mg

5003 (6188) 1140 (3140)b 1710 (13310) 1610 (4860) 591 (ndash2515 to 3696)

Average daily dose mg

858 (1638) 170 (578)c 428 (1724) 136 (320)d 108 (ndash316 to 532)

Single-day dose mga

180 (1240) 187 (77) 100 (1920) 50 (630)d 334 (ndash026 to 670)d

Propofol

Cumulative dose mg

7400 (55500) 0 (700)a 29450 (85800) 0 (27660) 226 (ndash42156 to 42607)

Average daily dose mg

529 (11690) 0 (53)a 4879 (15980) 0 (2738)a 651 (ndash3211 to 4514)

Single-day dose mga

0 (0) 0 (0) 0 (1000) 0 (0) 344 (ndash3781 to 4469)

Neuroleptics

Days receiving neuroleptics n

0 (2) 0 (4) 0 (0) 1 (4)a ndash16 (ndash34 to 03)

Proportion of patients who received a neuroleptic

407 407 174 565b

NMBs

Days receiving NMBs n

1 (1) 0 (0)a 0 (1) 0 (0) ndash01 (ndash05 to 03)

Proportion of patients who received an NMB

518 148a 348 174

Corticosteroids

Days receiving corticosteroids n

0 (0) 0 (0) 0 (2) 0 (2) 02 (ndash06 to 10)

Proportion of patients who received corticosteroids

185 148 348 261

PAV = proportional assist ventilation PSV = pressure support ventilation NMB = neuromuscular blockera p lt 005b Single-day dose = total dose given day prior to randomization or total dose give day after randomizationc p lt 0005d p lt 00005e p le 007All data points except percentage of patients receiving neuroleptics NMBs or corticosteroids are presented as median (interquartile range)

鎮静剤鎮痛剤についてはPAV群とPSV群で変わりなし

その他bull 不整脈気胸プロトコル施中の死亡

は両群とも起こらなかった

筆者らの考察bull PSVはweaningに最も使われているモードだ

がそれと較しても遜ないweaningができることがわかった

bull アシストが少ない傾向であったがPAV群のが呼吸窮迫となる回数が少なかった

rarr同調性が良くなった結果の可能性

ICU退室までの期間はPAV群で短いという結果であったがpowerが少ない

筆者らの考察bull 単施設での結果であり他の施設にこの

新しいプロトコルそのものが受けれられるかは不明である

bull SBTに失敗した患者を主に対象としておりICU患者すべてに適応はできない

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる

Page 37: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

Feasibility Outcomes bull 呼吸窮迫でRRTが呼吸器サポートを上昇

させた回数はPAV群で76回232daysPSV群で124回295daysと有意にPSV群で多かった(p=0002)

bull ACに戻した回数はPAV群で66回PSV群で90回と変わりなし(p=061)

bull 全くACに戻さなかった患者はPAV群で627名PSV群で623名(p=10)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

ベースラインは両群間で変わりなしPAVにすることでΔP圧Peak圧が低く吸気時間呼吸時間がくなるPSV群と較するとPAV群のがPaoAreaPaoAreaVE(呼吸器のサポートの強さをす)が低くVTPaoAreaがくsecond post-randomizationの解析においてΔP圧Peak圧が低い

Asynchrony index

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 7

B635) documenting compliance with the daily screening for SBT noncompletion of either the CPAP screening test or SBT occurred on 5 of 292 of patient study days (17) and this violation involved two patients in each study arm The second audit examining for nine different potential deviations found complete adherence to study protocol on 254 of 292 of patient study days (900) There were no adverse events (pneumo-thorax arrhythmia or death) resulting from any of the study interventions

Physiologic Response to Ventilation ProtocolsRecordings of flow and airway pressure were available for analysis for all 50 patients prerandomization 49 patients immediately (median 1 interquartile range [IQR] 0ndash15 hr) postrandomization and for 26 patients during subsequent days (median 27 IQR 20ndash70-hr postrandomization) Approximately 15000 breaths in 125 recordings were analyzed Table 2 compares the respiratory parameters in both groups pre- and postrandomization There were no statistically sig-nificant differences between the PAV and PSV groups at base-line prerandomization (p values not shown) In the PAV group there was a significant decrease in peak airway pressure delta pressure (measured as peak airway opening pressure minus the PEEP) and area under the inspiratory pressure curve immediately postrandomization and an increase in the per-cent of total cycle time spent in inspiration (TiTtot) whereas there were no significant differences immediately postran-domization in the PSV group The second measurement of postrandomization respiratory parameters revealed signifi-cant decreases in peak airway pressure and delta pressure in both PAV and PSV groups Adjusting for baseline measure-ments as a covariate the PAV group received 7 cm H2

O∙s less ventilator assistance (PaoArea) per liter of minute ventilation (p lt 00001) and increased their VT by 115 mLcm H

2O∙s

relative to the PSV group (p lt 0001)PAV significantly reduced AI postrandomization (Fig 2A)

In total nine patients (18) had a high AI (gt 10 of breaths) on PSV or on AC mode at baseline No patients had a high AI on PAV In the PSV group six patients (26) had a high AI

at baseline four (17) immediately postrandomization and one patient (7) after further adjustment of ventilator settings according to the PSV protocol (Fig 2B) Adjusting for baseline differences patients in the PAV group tended to have a lower AI than patients in the PSV group immediately postrandom-ization but this did not reach statistical significance (adjusted difference ndash45 [95 CI ndash95 to 04] p = 007)

Clinical OutcomesTable 3 shows the results for the main clinical endpoints Weaning duration and time to successful SBT did not dif-fer significantly among survivors between the groups Among ICU survivors median time to ICU discharge was shorter in the group randomized to PAV 73 days (IQR 52ndash114 d) ver-sus 124 days (IQR 75ndash308 d) in the PSV group (p = 003) Kaplan-Meier plots of the proportion of at-risk patients achiev-ing successful extubation and live ICU discharge are shown in Figure 3 A and B No difference was found for hospital length of stay ICU mortality or hospital mortality Four patients (15) in the PAV group and three patients (13) in the PSV group died in ICU (respiratory rate [RR] of PAV 114 95 CI 028 to 46 p = 086) Overall in-hospital mortality was nine patients (33) in the PAV group and six patients (26) in the PSV group (RR of PAV 128 95 CI 054 to 31 p = 058)

Cointerventions Sedating MedicationsTable 4 compares the use of medications that may prolong weaning (by causing sedation or neuromuscular weakness) in the two groups pre- and postrandomization Type and dose of sedative medication were highly variable among individual patients and therefore we analyzed pre- and postrandomiza-tion changes to determine the impact of the ventilation pro-tocol on sedative use within each patient In the PAV group benzodiazepine narcotic and propofol use measured as both cumulative dose and average daily dose all decreased signifi-cantly postrandomization as did neuromuscular blocker use In the PSV group average daily dose of lorazepam equiva-lent and average daily dose of propofol decreased signifi-cantly postrandomization whereas the proportion of patients

Figure 2 A Asynchrony index pre- and postrandomization on proportional assist ventilation B Asynchrony index pre- and postrandomization on pressure support ventilationPAV群 PSV群

Asynchrony Index () = 同調イベントの回数全呼吸回数 times 100

IntensiveCareMed(2006)321515ndash1522

AI>10は3名いたがPAVに設定変更後0名となった

AI>10は6名いたがPSVに設定変更後4名となった

Clinical Outcomes

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

成功したSBTまでの数抜管成功までの数に変わりなしICU退室までの数はPAV群で有意に短い

Kaplan-Meier curves

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

抜管成功までの時間 ICU退室までの時間

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 9

TABLE 4 Comparison of Medications That May Alter Weaning Duration Pre- and Postrandomization

Variables

PAV Group (n = 27)

PSV Group (n = 23) Postrandomization

Pre randomization Pos trandomization Pre randomization Pos trandomizationAdjusted Difference PSV ndash PAV (95 CI)

Lorazepam equivalents

Cumulative dose mg

308 (207) 48 (257)a 137 (440) 54 (258) 237 (ndash580 to 1055)

Average daily dose mg

44 (255) 09 (43)a 23 (85) 06 (40)a 24 (ndash62 to 110)

Single-day dose mgb

08 (72) 12 (50) 16 (70) 10 (64) 32 (ndash13 to 78)

Morphine equivalents

Cumulative dose mg

5003 (6188) 1140 (3140)b 1710 (13310) 1610 (4860) 591 (ndash2515 to 3696)

Average daily dose mg

858 (1638) 170 (578)c 428 (1724) 136 (320)d 108 (ndash316 to 532)

Single-day dose mga

180 (1240) 187 (77) 100 (1920) 50 (630)d 334 (ndash026 to 670)d

Propofol

Cumulative dose mg

7400 (55500) 0 (700)a 29450 (85800) 0 (27660) 226 (ndash42156 to 42607)

Average daily dose mg

529 (11690) 0 (53)a 4879 (15980) 0 (2738)a 651 (ndash3211 to 4514)

Single-day dose mga

0 (0) 0 (0) 0 (1000) 0 (0) 344 (ndash3781 to 4469)

Neuroleptics

Days receiving neuroleptics n

0 (2) 0 (4) 0 (0) 1 (4)a ndash16 (ndash34 to 03)

Proportion of patients who received a neuroleptic

407 407 174 565b

NMBs

Days receiving NMBs n

1 (1) 0 (0)a 0 (1) 0 (0) ndash01 (ndash05 to 03)

Proportion of patients who received an NMB

518 148a 348 174

Corticosteroids

Days receiving corticosteroids n

0 (0) 0 (0) 0 (2) 0 (2) 02 (ndash06 to 10)

Proportion of patients who received corticosteroids

185 148 348 261

PAV = proportional assist ventilation PSV = pressure support ventilation NMB = neuromuscular blockera p lt 005b Single-day dose = total dose given day prior to randomization or total dose give day after randomizationc p lt 0005d p lt 00005e p le 007All data points except percentage of patients receiving neuroleptics NMBs or corticosteroids are presented as median (interquartile range)

鎮静剤鎮痛剤についてはPAV群とPSV群で変わりなし

その他bull 不整脈気胸プロトコル施中の死亡

は両群とも起こらなかった

筆者らの考察bull PSVはweaningに最も使われているモードだ

がそれと較しても遜ないweaningができることがわかった

bull アシストが少ない傾向であったがPAV群のが呼吸窮迫となる回数が少なかった

rarr同調性が良くなった結果の可能性

ICU退室までの期間はPAV群で短いという結果であったがpowerが少ない

筆者らの考察bull 単施設での結果であり他の施設にこの

新しいプロトコルそのものが受けれられるかは不明である

bull SBTに失敗した患者を主に対象としておりICU患者すべてに適応はできない

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる

Page 38: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

ベースラインは両群間で変わりなしPAVにすることでΔP圧Peak圧が低く吸気時間呼吸時間がくなるPSV群と較するとPAV群のがPaoAreaPaoAreaVE(呼吸器のサポートの強さをす)が低くVTPaoAreaがくsecond post-randomizationの解析においてΔP圧Peak圧が低い

Asynchrony index

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 7

B635) documenting compliance with the daily screening for SBT noncompletion of either the CPAP screening test or SBT occurred on 5 of 292 of patient study days (17) and this violation involved two patients in each study arm The second audit examining for nine different potential deviations found complete adherence to study protocol on 254 of 292 of patient study days (900) There were no adverse events (pneumo-thorax arrhythmia or death) resulting from any of the study interventions

Physiologic Response to Ventilation ProtocolsRecordings of flow and airway pressure were available for analysis for all 50 patients prerandomization 49 patients immediately (median 1 interquartile range [IQR] 0ndash15 hr) postrandomization and for 26 patients during subsequent days (median 27 IQR 20ndash70-hr postrandomization) Approximately 15000 breaths in 125 recordings were analyzed Table 2 compares the respiratory parameters in both groups pre- and postrandomization There were no statistically sig-nificant differences between the PAV and PSV groups at base-line prerandomization (p values not shown) In the PAV group there was a significant decrease in peak airway pressure delta pressure (measured as peak airway opening pressure minus the PEEP) and area under the inspiratory pressure curve immediately postrandomization and an increase in the per-cent of total cycle time spent in inspiration (TiTtot) whereas there were no significant differences immediately postran-domization in the PSV group The second measurement of postrandomization respiratory parameters revealed signifi-cant decreases in peak airway pressure and delta pressure in both PAV and PSV groups Adjusting for baseline measure-ments as a covariate the PAV group received 7 cm H2

O∙s less ventilator assistance (PaoArea) per liter of minute ventilation (p lt 00001) and increased their VT by 115 mLcm H

2O∙s

relative to the PSV group (p lt 0001)PAV significantly reduced AI postrandomization (Fig 2A)

In total nine patients (18) had a high AI (gt 10 of breaths) on PSV or on AC mode at baseline No patients had a high AI on PAV In the PSV group six patients (26) had a high AI

at baseline four (17) immediately postrandomization and one patient (7) after further adjustment of ventilator settings according to the PSV protocol (Fig 2B) Adjusting for baseline differences patients in the PAV group tended to have a lower AI than patients in the PSV group immediately postrandom-ization but this did not reach statistical significance (adjusted difference ndash45 [95 CI ndash95 to 04] p = 007)

Clinical OutcomesTable 3 shows the results for the main clinical endpoints Weaning duration and time to successful SBT did not dif-fer significantly among survivors between the groups Among ICU survivors median time to ICU discharge was shorter in the group randomized to PAV 73 days (IQR 52ndash114 d) ver-sus 124 days (IQR 75ndash308 d) in the PSV group (p = 003) Kaplan-Meier plots of the proportion of at-risk patients achiev-ing successful extubation and live ICU discharge are shown in Figure 3 A and B No difference was found for hospital length of stay ICU mortality or hospital mortality Four patients (15) in the PAV group and three patients (13) in the PSV group died in ICU (respiratory rate [RR] of PAV 114 95 CI 028 to 46 p = 086) Overall in-hospital mortality was nine patients (33) in the PAV group and six patients (26) in the PSV group (RR of PAV 128 95 CI 054 to 31 p = 058)

Cointerventions Sedating MedicationsTable 4 compares the use of medications that may prolong weaning (by causing sedation or neuromuscular weakness) in the two groups pre- and postrandomization Type and dose of sedative medication were highly variable among individual patients and therefore we analyzed pre- and postrandomiza-tion changes to determine the impact of the ventilation pro-tocol on sedative use within each patient In the PAV group benzodiazepine narcotic and propofol use measured as both cumulative dose and average daily dose all decreased signifi-cantly postrandomization as did neuromuscular blocker use In the PSV group average daily dose of lorazepam equiva-lent and average daily dose of propofol decreased signifi-cantly postrandomization whereas the proportion of patients

Figure 2 A Asynchrony index pre- and postrandomization on proportional assist ventilation B Asynchrony index pre- and postrandomization on pressure support ventilationPAV群 PSV群

Asynchrony Index () = 同調イベントの回数全呼吸回数 times 100

IntensiveCareMed(2006)321515ndash1522

AI>10は3名いたがPAVに設定変更後0名となった

AI>10は6名いたがPSVに設定変更後4名となった

Clinical Outcomes

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

成功したSBTまでの数抜管成功までの数に変わりなしICU退室までの数はPAV群で有意に短い

Kaplan-Meier curves

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

抜管成功までの時間 ICU退室までの時間

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 9

TABLE 4 Comparison of Medications That May Alter Weaning Duration Pre- and Postrandomization

Variables

PAV Group (n = 27)

PSV Group (n = 23) Postrandomization

Pre randomization Pos trandomization Pre randomization Pos trandomizationAdjusted Difference PSV ndash PAV (95 CI)

Lorazepam equivalents

Cumulative dose mg

308 (207) 48 (257)a 137 (440) 54 (258) 237 (ndash580 to 1055)

Average daily dose mg

44 (255) 09 (43)a 23 (85) 06 (40)a 24 (ndash62 to 110)

Single-day dose mgb

08 (72) 12 (50) 16 (70) 10 (64) 32 (ndash13 to 78)

Morphine equivalents

Cumulative dose mg

5003 (6188) 1140 (3140)b 1710 (13310) 1610 (4860) 591 (ndash2515 to 3696)

Average daily dose mg

858 (1638) 170 (578)c 428 (1724) 136 (320)d 108 (ndash316 to 532)

Single-day dose mga

180 (1240) 187 (77) 100 (1920) 50 (630)d 334 (ndash026 to 670)d

Propofol

Cumulative dose mg

7400 (55500) 0 (700)a 29450 (85800) 0 (27660) 226 (ndash42156 to 42607)

Average daily dose mg

529 (11690) 0 (53)a 4879 (15980) 0 (2738)a 651 (ndash3211 to 4514)

Single-day dose mga

0 (0) 0 (0) 0 (1000) 0 (0) 344 (ndash3781 to 4469)

Neuroleptics

Days receiving neuroleptics n

0 (2) 0 (4) 0 (0) 1 (4)a ndash16 (ndash34 to 03)

Proportion of patients who received a neuroleptic

407 407 174 565b

NMBs

Days receiving NMBs n

1 (1) 0 (0)a 0 (1) 0 (0) ndash01 (ndash05 to 03)

Proportion of patients who received an NMB

518 148a 348 174

Corticosteroids

Days receiving corticosteroids n

0 (0) 0 (0) 0 (2) 0 (2) 02 (ndash06 to 10)

Proportion of patients who received corticosteroids

185 148 348 261

PAV = proportional assist ventilation PSV = pressure support ventilation NMB = neuromuscular blockera p lt 005b Single-day dose = total dose given day prior to randomization or total dose give day after randomizationc p lt 0005d p lt 00005e p le 007All data points except percentage of patients receiving neuroleptics NMBs or corticosteroids are presented as median (interquartile range)

鎮静剤鎮痛剤についてはPAV群とPSV群で変わりなし

その他bull 不整脈気胸プロトコル施中の死亡

は両群とも起こらなかった

筆者らの考察bull PSVはweaningに最も使われているモードだ

がそれと較しても遜ないweaningができることがわかった

bull アシストが少ない傾向であったがPAV群のが呼吸窮迫となる回数が少なかった

rarr同調性が良くなった結果の可能性

ICU退室までの期間はPAV群で短いという結果であったがpowerが少ない

筆者らの考察bull 単施設での結果であり他の施設にこの

新しいプロトコルそのものが受けれられるかは不明である

bull SBTに失敗した患者を主に対象としておりICU患者すべてに適応はできない

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる

Page 39: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

Respiratory Parameters

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

6 wwwccmjournalorg

TABLE 2 Comparison of Respiratory Parameters Pre- and Postrandomization on Proportional Assist Ventilation With Load-Adjustable Gain Factors Versus Pressure Support Ventilation

Variables

PAV PSVAdjusted Difference

(PAV ndash PSV) (95 CI)

Pre-randomization

(n = 27)

First Post-randomization

(n = 26)

Second Post-randomization

(n = 12)

Pre-randomization

(n = 23)

First Post-randomization

(n = 23)

Second Post-randomization

(n = 15)

First Post-

randomization

Second Post -

randomization

Set level of support on PSV cm H2O

128 plusmn 31 134 plusmn 24 129 plusmn 28 97 plusmn 39

Set level of support on PAV

583 plusmn 158 313 plusmn 105

∆P cm H2O 132 plusmn 33 108 plusmn 38a 63 plusmn 18b 136 plusmn 26 123 plusmn 54 104 plusmn 37b ndash15 (ndash41 to 11)

ndash41 (ndash65 to ndash16)a

Peak pressure at the airway opening cm H2O

218 plusmn 39 193 plusmn 41b 159 plusmn 37a 229 plusmn 28 229 plusmn 35 192 plusmn 45a ndash30 (ndash48 to ndash12)a

ndash30 (ndash64 to 03)c

Positive end- expiratory pressure cm H2O

86 plusmn 22 84 plusmn 20 95 plusmn 25 93 plusmn 26 95 plusmn 24 87 plusmn 24 ndash03 (ndash10 to 04)

08 (ndash09 to 25)

Inspiratory time total breath time

353 plusmn 79 380 plusmn 65d 383 plusmn 50 380 plusmn 85 388 plusmn 85 374 plusmn 74 10 (ndash22 to 41)

03 (ndash45 to 50)

Peak inspiratory flow Ls

10 plusmn 02 10 plusmn 02 09 plusmn 02 10 plusmn 02 10 plusmn 02 09 plusmn 02d 00 (ndash01 to 01)

ndash00 (ndash02 to 02)

VT mL 564 plusmn 162 560 plusmn 136 589 plusmn 98 628 plusmn 193 634 plusmn 190 563 plusmn 158d ndash00 (ndash01 to 004)

00 (ndash01 to 01)

VTkg ideal body weight mL

96 plusmn 28 96 plusmn 30 89 plusmn 20 100 plusmn 27 100 plusmn 25 87 plusmn 20d ndash00 (ndash12 to 12)

04 (ndash10 to 19)

Respiratory rate breathsmin

257 plusmn 55 272 plusmn 62c 251 plusmn 63 243 plusmn 75 257 plusmn 67 272 plusmn 65 09 (ndash23 to 41)

ndash28 (ndash79 to 24)

VE Lmin 139 plusmn 32 146 plusmn 31 143 plusmn 29 145 plusmn 48 159 plusmn 53 145 plusmn 39 ndash05 (ndash24 to 13)

ndash02 (ndash26 to 22)

PaoArea cm H2Os

67 (43) 38 (26)b 71 (39) 82 (68) ndash41 (ndash54 to ndash28)b

PaoAreaVE cm H2OsL

122 (65) 73 (67)b 145 (112) 148 (124) ndash74 (ndash98 to ndash49)b

VTPaoArea mL cm H2Os

84 (40) 140 (158)b 66 (62) 68 (68) 115 (48 to 181)a

PAV = proportional assist ventilation PSV = pressure support ventilation ∆P (delta pressure) = peak pressure at the airway opening ndash positive end-expiratory pressure VT = tidal volume VE = minute ventilation PaoArea = area under the inspiratory pressure curvea p lt 0005b p lt 00005c p le 007d p lt 005All data points are presented as mean plusmn SD or median (interquartile range)

ベースラインは両群間で変わりなしPAVにすることでΔP圧Peak圧が低く吸気時間呼吸時間がくなるPSV群と較するとPAV群のがPaoAreaPaoAreaVE(呼吸器のサポートの強さをす)が低くVTPaoAreaがくsecond post-randomizationの解析においてΔP圧Peak圧が低い

Asynchrony index

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 7

B635) documenting compliance with the daily screening for SBT noncompletion of either the CPAP screening test or SBT occurred on 5 of 292 of patient study days (17) and this violation involved two patients in each study arm The second audit examining for nine different potential deviations found complete adherence to study protocol on 254 of 292 of patient study days (900) There were no adverse events (pneumo-thorax arrhythmia or death) resulting from any of the study interventions

Physiologic Response to Ventilation ProtocolsRecordings of flow and airway pressure were available for analysis for all 50 patients prerandomization 49 patients immediately (median 1 interquartile range [IQR] 0ndash15 hr) postrandomization and for 26 patients during subsequent days (median 27 IQR 20ndash70-hr postrandomization) Approximately 15000 breaths in 125 recordings were analyzed Table 2 compares the respiratory parameters in both groups pre- and postrandomization There were no statistically sig-nificant differences between the PAV and PSV groups at base-line prerandomization (p values not shown) In the PAV group there was a significant decrease in peak airway pressure delta pressure (measured as peak airway opening pressure minus the PEEP) and area under the inspiratory pressure curve immediately postrandomization and an increase in the per-cent of total cycle time spent in inspiration (TiTtot) whereas there were no significant differences immediately postran-domization in the PSV group The second measurement of postrandomization respiratory parameters revealed signifi-cant decreases in peak airway pressure and delta pressure in both PAV and PSV groups Adjusting for baseline measure-ments as a covariate the PAV group received 7 cm H2

O∙s less ventilator assistance (PaoArea) per liter of minute ventilation (p lt 00001) and increased their VT by 115 mLcm H

2O∙s

relative to the PSV group (p lt 0001)PAV significantly reduced AI postrandomization (Fig 2A)

In total nine patients (18) had a high AI (gt 10 of breaths) on PSV or on AC mode at baseline No patients had a high AI on PAV In the PSV group six patients (26) had a high AI

at baseline four (17) immediately postrandomization and one patient (7) after further adjustment of ventilator settings according to the PSV protocol (Fig 2B) Adjusting for baseline differences patients in the PAV group tended to have a lower AI than patients in the PSV group immediately postrandom-ization but this did not reach statistical significance (adjusted difference ndash45 [95 CI ndash95 to 04] p = 007)

Clinical OutcomesTable 3 shows the results for the main clinical endpoints Weaning duration and time to successful SBT did not dif-fer significantly among survivors between the groups Among ICU survivors median time to ICU discharge was shorter in the group randomized to PAV 73 days (IQR 52ndash114 d) ver-sus 124 days (IQR 75ndash308 d) in the PSV group (p = 003) Kaplan-Meier plots of the proportion of at-risk patients achiev-ing successful extubation and live ICU discharge are shown in Figure 3 A and B No difference was found for hospital length of stay ICU mortality or hospital mortality Four patients (15) in the PAV group and three patients (13) in the PSV group died in ICU (respiratory rate [RR] of PAV 114 95 CI 028 to 46 p = 086) Overall in-hospital mortality was nine patients (33) in the PAV group and six patients (26) in the PSV group (RR of PAV 128 95 CI 054 to 31 p = 058)

Cointerventions Sedating MedicationsTable 4 compares the use of medications that may prolong weaning (by causing sedation or neuromuscular weakness) in the two groups pre- and postrandomization Type and dose of sedative medication were highly variable among individual patients and therefore we analyzed pre- and postrandomiza-tion changes to determine the impact of the ventilation pro-tocol on sedative use within each patient In the PAV group benzodiazepine narcotic and propofol use measured as both cumulative dose and average daily dose all decreased signifi-cantly postrandomization as did neuromuscular blocker use In the PSV group average daily dose of lorazepam equiva-lent and average daily dose of propofol decreased signifi-cantly postrandomization whereas the proportion of patients

Figure 2 A Asynchrony index pre- and postrandomization on proportional assist ventilation B Asynchrony index pre- and postrandomization on pressure support ventilationPAV群 PSV群

Asynchrony Index () = 同調イベントの回数全呼吸回数 times 100

IntensiveCareMed(2006)321515ndash1522

AI>10は3名いたがPAVに設定変更後0名となった

AI>10は6名いたがPSVに設定変更後4名となった

Clinical Outcomes

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

成功したSBTまでの数抜管成功までの数に変わりなしICU退室までの数はPAV群で有意に短い

Kaplan-Meier curves

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

抜管成功までの時間 ICU退室までの時間

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 9

TABLE 4 Comparison of Medications That May Alter Weaning Duration Pre- and Postrandomization

Variables

PAV Group (n = 27)

PSV Group (n = 23) Postrandomization

Pre randomization Pos trandomization Pre randomization Pos trandomizationAdjusted Difference PSV ndash PAV (95 CI)

Lorazepam equivalents

Cumulative dose mg

308 (207) 48 (257)a 137 (440) 54 (258) 237 (ndash580 to 1055)

Average daily dose mg

44 (255) 09 (43)a 23 (85) 06 (40)a 24 (ndash62 to 110)

Single-day dose mgb

08 (72) 12 (50) 16 (70) 10 (64) 32 (ndash13 to 78)

Morphine equivalents

Cumulative dose mg

5003 (6188) 1140 (3140)b 1710 (13310) 1610 (4860) 591 (ndash2515 to 3696)

Average daily dose mg

858 (1638) 170 (578)c 428 (1724) 136 (320)d 108 (ndash316 to 532)

Single-day dose mga

180 (1240) 187 (77) 100 (1920) 50 (630)d 334 (ndash026 to 670)d

Propofol

Cumulative dose mg

7400 (55500) 0 (700)a 29450 (85800) 0 (27660) 226 (ndash42156 to 42607)

Average daily dose mg

529 (11690) 0 (53)a 4879 (15980) 0 (2738)a 651 (ndash3211 to 4514)

Single-day dose mga

0 (0) 0 (0) 0 (1000) 0 (0) 344 (ndash3781 to 4469)

Neuroleptics

Days receiving neuroleptics n

0 (2) 0 (4) 0 (0) 1 (4)a ndash16 (ndash34 to 03)

Proportion of patients who received a neuroleptic

407 407 174 565b

NMBs

Days receiving NMBs n

1 (1) 0 (0)a 0 (1) 0 (0) ndash01 (ndash05 to 03)

Proportion of patients who received an NMB

518 148a 348 174

Corticosteroids

Days receiving corticosteroids n

0 (0) 0 (0) 0 (2) 0 (2) 02 (ndash06 to 10)

Proportion of patients who received corticosteroids

185 148 348 261

PAV = proportional assist ventilation PSV = pressure support ventilation NMB = neuromuscular blockera p lt 005b Single-day dose = total dose given day prior to randomization or total dose give day after randomizationc p lt 0005d p lt 00005e p le 007All data points except percentage of patients receiving neuroleptics NMBs or corticosteroids are presented as median (interquartile range)

鎮静剤鎮痛剤についてはPAV群とPSV群で変わりなし

その他bull 不整脈気胸プロトコル施中の死亡

は両群とも起こらなかった

筆者らの考察bull PSVはweaningに最も使われているモードだ

がそれと較しても遜ないweaningができることがわかった

bull アシストが少ない傾向であったがPAV群のが呼吸窮迫となる回数が少なかった

rarr同調性が良くなった結果の可能性

ICU退室までの期間はPAV群で短いという結果であったがpowerが少ない

筆者らの考察bull 単施設での結果であり他の施設にこの

新しいプロトコルそのものが受けれられるかは不明である

bull SBTに失敗した患者を主に対象としておりICU患者すべてに適応はできない

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる

Page 40: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

Asynchrony index

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 7

B635) documenting compliance with the daily screening for SBT noncompletion of either the CPAP screening test or SBT occurred on 5 of 292 of patient study days (17) and this violation involved two patients in each study arm The second audit examining for nine different potential deviations found complete adherence to study protocol on 254 of 292 of patient study days (900) There were no adverse events (pneumo-thorax arrhythmia or death) resulting from any of the study interventions

Physiologic Response to Ventilation ProtocolsRecordings of flow and airway pressure were available for analysis for all 50 patients prerandomization 49 patients immediately (median 1 interquartile range [IQR] 0ndash15 hr) postrandomization and for 26 patients during subsequent days (median 27 IQR 20ndash70-hr postrandomization) Approximately 15000 breaths in 125 recordings were analyzed Table 2 compares the respiratory parameters in both groups pre- and postrandomization There were no statistically sig-nificant differences between the PAV and PSV groups at base-line prerandomization (p values not shown) In the PAV group there was a significant decrease in peak airway pressure delta pressure (measured as peak airway opening pressure minus the PEEP) and area under the inspiratory pressure curve immediately postrandomization and an increase in the per-cent of total cycle time spent in inspiration (TiTtot) whereas there were no significant differences immediately postran-domization in the PSV group The second measurement of postrandomization respiratory parameters revealed signifi-cant decreases in peak airway pressure and delta pressure in both PAV and PSV groups Adjusting for baseline measure-ments as a covariate the PAV group received 7 cm H2

O∙s less ventilator assistance (PaoArea) per liter of minute ventilation (p lt 00001) and increased their VT by 115 mLcm H

2O∙s

relative to the PSV group (p lt 0001)PAV significantly reduced AI postrandomization (Fig 2A)

In total nine patients (18) had a high AI (gt 10 of breaths) on PSV or on AC mode at baseline No patients had a high AI on PAV In the PSV group six patients (26) had a high AI

at baseline four (17) immediately postrandomization and one patient (7) after further adjustment of ventilator settings according to the PSV protocol (Fig 2B) Adjusting for baseline differences patients in the PAV group tended to have a lower AI than patients in the PSV group immediately postrandom-ization but this did not reach statistical significance (adjusted difference ndash45 [95 CI ndash95 to 04] p = 007)

Clinical OutcomesTable 3 shows the results for the main clinical endpoints Weaning duration and time to successful SBT did not dif-fer significantly among survivors between the groups Among ICU survivors median time to ICU discharge was shorter in the group randomized to PAV 73 days (IQR 52ndash114 d) ver-sus 124 days (IQR 75ndash308 d) in the PSV group (p = 003) Kaplan-Meier plots of the proportion of at-risk patients achiev-ing successful extubation and live ICU discharge are shown in Figure 3 A and B No difference was found for hospital length of stay ICU mortality or hospital mortality Four patients (15) in the PAV group and three patients (13) in the PSV group died in ICU (respiratory rate [RR] of PAV 114 95 CI 028 to 46 p = 086) Overall in-hospital mortality was nine patients (33) in the PAV group and six patients (26) in the PSV group (RR of PAV 128 95 CI 054 to 31 p = 058)

Cointerventions Sedating MedicationsTable 4 compares the use of medications that may prolong weaning (by causing sedation or neuromuscular weakness) in the two groups pre- and postrandomization Type and dose of sedative medication were highly variable among individual patients and therefore we analyzed pre- and postrandomiza-tion changes to determine the impact of the ventilation pro-tocol on sedative use within each patient In the PAV group benzodiazepine narcotic and propofol use measured as both cumulative dose and average daily dose all decreased signifi-cantly postrandomization as did neuromuscular blocker use In the PSV group average daily dose of lorazepam equiva-lent and average daily dose of propofol decreased signifi-cantly postrandomization whereas the proportion of patients

Figure 2 A Asynchrony index pre- and postrandomization on proportional assist ventilation B Asynchrony index pre- and postrandomization on pressure support ventilationPAV群 PSV群

Asynchrony Index () = 同調イベントの回数全呼吸回数 times 100

IntensiveCareMed(2006)321515ndash1522

AI>10は3名いたがPAVに設定変更後0名となった

AI>10は6名いたがPSVに設定変更後4名となった

Clinical Outcomes

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

成功したSBTまでの数抜管成功までの数に変わりなしICU退室までの数はPAV群で有意に短い

Kaplan-Meier curves

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

抜管成功までの時間 ICU退室までの時間

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 9

TABLE 4 Comparison of Medications That May Alter Weaning Duration Pre- and Postrandomization

Variables

PAV Group (n = 27)

PSV Group (n = 23) Postrandomization

Pre randomization Pos trandomization Pre randomization Pos trandomizationAdjusted Difference PSV ndash PAV (95 CI)

Lorazepam equivalents

Cumulative dose mg

308 (207) 48 (257)a 137 (440) 54 (258) 237 (ndash580 to 1055)

Average daily dose mg

44 (255) 09 (43)a 23 (85) 06 (40)a 24 (ndash62 to 110)

Single-day dose mgb

08 (72) 12 (50) 16 (70) 10 (64) 32 (ndash13 to 78)

Morphine equivalents

Cumulative dose mg

5003 (6188) 1140 (3140)b 1710 (13310) 1610 (4860) 591 (ndash2515 to 3696)

Average daily dose mg

858 (1638) 170 (578)c 428 (1724) 136 (320)d 108 (ndash316 to 532)

Single-day dose mga

180 (1240) 187 (77) 100 (1920) 50 (630)d 334 (ndash026 to 670)d

Propofol

Cumulative dose mg

7400 (55500) 0 (700)a 29450 (85800) 0 (27660) 226 (ndash42156 to 42607)

Average daily dose mg

529 (11690) 0 (53)a 4879 (15980) 0 (2738)a 651 (ndash3211 to 4514)

Single-day dose mga

0 (0) 0 (0) 0 (1000) 0 (0) 344 (ndash3781 to 4469)

Neuroleptics

Days receiving neuroleptics n

0 (2) 0 (4) 0 (0) 1 (4)a ndash16 (ndash34 to 03)

Proportion of patients who received a neuroleptic

407 407 174 565b

NMBs

Days receiving NMBs n

1 (1) 0 (0)a 0 (1) 0 (0) ndash01 (ndash05 to 03)

Proportion of patients who received an NMB

518 148a 348 174

Corticosteroids

Days receiving corticosteroids n

0 (0) 0 (0) 0 (2) 0 (2) 02 (ndash06 to 10)

Proportion of patients who received corticosteroids

185 148 348 261

PAV = proportional assist ventilation PSV = pressure support ventilation NMB = neuromuscular blockera p lt 005b Single-day dose = total dose given day prior to randomization or total dose give day after randomizationc p lt 0005d p lt 00005e p le 007All data points except percentage of patients receiving neuroleptics NMBs or corticosteroids are presented as median (interquartile range)

鎮静剤鎮痛剤についてはPAV群とPSV群で変わりなし

その他bull 不整脈気胸プロトコル施中の死亡

は両群とも起こらなかった

筆者らの考察bull PSVはweaningに最も使われているモードだ

がそれと較しても遜ないweaningができることがわかった

bull アシストが少ない傾向であったがPAV群のが呼吸窮迫となる回数が少なかった

rarr同調性が良くなった結果の可能性

ICU退室までの期間はPAV群で短いという結果であったがpowerが少ない

筆者らの考察bull 単施設での結果であり他の施設にこの

新しいプロトコルそのものが受けれられるかは不明である

bull SBTに失敗した患者を主に対象としておりICU患者すべてに適応はできない

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる

Page 41: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

Clinical Outcomes

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

成功したSBTまでの数抜管成功までの数に変わりなしICU退室までの数はPAV群で有意に短い

Kaplan-Meier curves

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

抜管成功までの時間 ICU退室までの時間

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 9

TABLE 4 Comparison of Medications That May Alter Weaning Duration Pre- and Postrandomization

Variables

PAV Group (n = 27)

PSV Group (n = 23) Postrandomization

Pre randomization Pos trandomization Pre randomization Pos trandomizationAdjusted Difference PSV ndash PAV (95 CI)

Lorazepam equivalents

Cumulative dose mg

308 (207) 48 (257)a 137 (440) 54 (258) 237 (ndash580 to 1055)

Average daily dose mg

44 (255) 09 (43)a 23 (85) 06 (40)a 24 (ndash62 to 110)

Single-day dose mgb

08 (72) 12 (50) 16 (70) 10 (64) 32 (ndash13 to 78)

Morphine equivalents

Cumulative dose mg

5003 (6188) 1140 (3140)b 1710 (13310) 1610 (4860) 591 (ndash2515 to 3696)

Average daily dose mg

858 (1638) 170 (578)c 428 (1724) 136 (320)d 108 (ndash316 to 532)

Single-day dose mga

180 (1240) 187 (77) 100 (1920) 50 (630)d 334 (ndash026 to 670)d

Propofol

Cumulative dose mg

7400 (55500) 0 (700)a 29450 (85800) 0 (27660) 226 (ndash42156 to 42607)

Average daily dose mg

529 (11690) 0 (53)a 4879 (15980) 0 (2738)a 651 (ndash3211 to 4514)

Single-day dose mga

0 (0) 0 (0) 0 (1000) 0 (0) 344 (ndash3781 to 4469)

Neuroleptics

Days receiving neuroleptics n

0 (2) 0 (4) 0 (0) 1 (4)a ndash16 (ndash34 to 03)

Proportion of patients who received a neuroleptic

407 407 174 565b

NMBs

Days receiving NMBs n

1 (1) 0 (0)a 0 (1) 0 (0) ndash01 (ndash05 to 03)

Proportion of patients who received an NMB

518 148a 348 174

Corticosteroids

Days receiving corticosteroids n

0 (0) 0 (0) 0 (2) 0 (2) 02 (ndash06 to 10)

Proportion of patients who received corticosteroids

185 148 348 261

PAV = proportional assist ventilation PSV = pressure support ventilation NMB = neuromuscular blockera p lt 005b Single-day dose = total dose given day prior to randomization or total dose give day after randomizationc p lt 0005d p lt 00005e p le 007All data points except percentage of patients receiving neuroleptics NMBs or corticosteroids are presented as median (interquartile range)

鎮静剤鎮痛剤についてはPAV群とPSV群で変わりなし

その他bull 不整脈気胸プロトコル施中の死亡

は両群とも起こらなかった

筆者らの考察bull PSVはweaningに最も使われているモードだ

がそれと較しても遜ないweaningができることがわかった

bull アシストが少ない傾向であったがPAV群のが呼吸窮迫となる回数が少なかった

rarr同調性が良くなった結果の可能性

ICU退室までの期間はPAV群で短いという結果であったがpowerが少ない

筆者らの考察bull 単施設での結果であり他の施設にこの

新しいプロトコルそのものが受けれられるかは不明である

bull SBTに失敗した患者を主に対象としておりICU患者すべてに適応はできない

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる

Page 42: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

Kaplan-Meier curves

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Bosma et al

8 wwwccmjournalorg

receiving a neuroleptic medication and the median number of days patients received a neuroleptic medication increased significantly postrandomization After adjusting for the pres-enceabsence of neuroleptic medication prerandomization the odds of receiving a neuroleptic medication postrandomization on PSV was 35 (95 CI 09ndash136 p = 007) compared with PAV ANCOVA adjusting for baseline values revealed no sta-tistically significant differences between PAV and PSV groups in postrandomization values for any of the medications

DISCUSSIONThis study is the first RCT examining the physiologic and clinical outcomes of a protocol for PAV with load-adjustable gain factors employed from randomization to final extubation

compared with a protocol for PSV the most commonly used weaning mode Patients spent a median 6 days (3ndash12 d) on study protocol and an average of 18 plusmn 5 hoursd without need for ventilation with AC mode demonstrating that both pro-tocols could be used safely and effectively for the majority of the recovery period Patients on PSV met criteria for respira-tory distress prompting an increase in level of support more frequently than patients on PAV but patients met criteria for switching to AC mode with equal frequency between the two groups No patient experienced an adverse event as a physio-logic consequence of excessive work of breathing or instability

The physiologic impact of PAV+ was a more marked reduc-tion in peak airway pressures and area under the inspiratory pressure curve (a measure of ventilator assistance) while achieving equivalent VTs indicating that patients on PAV+

Figure 3 A Time to successful extubation for the two treatment groups (Kaplan-Meier curves) B Time to live ICU discharge for the two treatment groups (Kaplan-Meier curves) PAV = proportional assist ventilation PSV = pressure support ventilation

TABLE 3 Comparison of Clinical Outcomes Between Study Groups

VariablesProportional Assist

Ventilation Group (n = 27)Pressure Support Ventilation

Group (n = 23) p

Time to successful pass of spontaneous breathing trial d median (IQR)

37 (26ndash70) 47 (29ndash259) 034

Time to successful extubation d median (IQR) 39 (28ndash84) 49 (29ndash263) 039

Time to live ICU discharge d median (IQR) 73 (52ndash114) 124 (75ndash308) 003

Time to live hospital discharge d median (IQR) 265 (123ndash744) 250 (159ndash435) 042

ICU mortality n () 4 (15) 3 (13) 086

Hospital mortality n () 9 (33) 6 (26) 058

Reintubation n () 3 (11) 5 (22) 031

Tracheostomy n () 4 (15) 6 (26) 032

Noninvasive ventilation use postfinal extubation 3 (11) 5 (22) 052

Total duration of MV d median (IQR) 121 (78ndash204) 139 (71ndash346) 051

Total duration of MV gt 21 d n () 6 (22) 6 (26) 075

IQR = interquartile range MV = mechanical ventilation

抜管成功までの時間 ICU退室までの時間

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 9

TABLE 4 Comparison of Medications That May Alter Weaning Duration Pre- and Postrandomization

Variables

PAV Group (n = 27)

PSV Group (n = 23) Postrandomization

Pre randomization Pos trandomization Pre randomization Pos trandomizationAdjusted Difference PSV ndash PAV (95 CI)

Lorazepam equivalents

Cumulative dose mg

308 (207) 48 (257)a 137 (440) 54 (258) 237 (ndash580 to 1055)

Average daily dose mg

44 (255) 09 (43)a 23 (85) 06 (40)a 24 (ndash62 to 110)

Single-day dose mgb

08 (72) 12 (50) 16 (70) 10 (64) 32 (ndash13 to 78)

Morphine equivalents

Cumulative dose mg

5003 (6188) 1140 (3140)b 1710 (13310) 1610 (4860) 591 (ndash2515 to 3696)

Average daily dose mg

858 (1638) 170 (578)c 428 (1724) 136 (320)d 108 (ndash316 to 532)

Single-day dose mga

180 (1240) 187 (77) 100 (1920) 50 (630)d 334 (ndash026 to 670)d

Propofol

Cumulative dose mg

7400 (55500) 0 (700)a 29450 (85800) 0 (27660) 226 (ndash42156 to 42607)

Average daily dose mg

529 (11690) 0 (53)a 4879 (15980) 0 (2738)a 651 (ndash3211 to 4514)

Single-day dose mga

0 (0) 0 (0) 0 (1000) 0 (0) 344 (ndash3781 to 4469)

Neuroleptics

Days receiving neuroleptics n

0 (2) 0 (4) 0 (0) 1 (4)a ndash16 (ndash34 to 03)

Proportion of patients who received a neuroleptic

407 407 174 565b

NMBs

Days receiving NMBs n

1 (1) 0 (0)a 0 (1) 0 (0) ndash01 (ndash05 to 03)

Proportion of patients who received an NMB

518 148a 348 174

Corticosteroids

Days receiving corticosteroids n

0 (0) 0 (0) 0 (2) 0 (2) 02 (ndash06 to 10)

Proportion of patients who received corticosteroids

185 148 348 261

PAV = proportional assist ventilation PSV = pressure support ventilation NMB = neuromuscular blockera p lt 005b Single-day dose = total dose given day prior to randomization or total dose give day after randomizationc p lt 0005d p lt 00005e p le 007All data points except percentage of patients receiving neuroleptics NMBs or corticosteroids are presented as median (interquartile range)

鎮静剤鎮痛剤についてはPAV群とPSV群で変わりなし

その他bull 不整脈気胸プロトコル施中の死亡

は両群とも起こらなかった

筆者らの考察bull PSVはweaningに最も使われているモードだ

がそれと較しても遜ないweaningができることがわかった

bull アシストが少ない傾向であったがPAV群のが呼吸窮迫となる回数が少なかった

rarr同調性が良くなった結果の可能性

ICU退室までの期間はPAV群で短いという結果であったがpowerが少ない

筆者らの考察bull 単施設での結果であり他の施設にこの

新しいプロトコルそのものが受けれられるかは不明である

bull SBTに失敗した患者を主に対象としておりICU患者すべてに適応はできない

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる

Page 43: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

Copyright copy 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health Inc All Rights Reserved

Clinical Investigation

Critical Care Medicine wwwccmjournalorg 9

TABLE 4 Comparison of Medications That May Alter Weaning Duration Pre- and Postrandomization

Variables

PAV Group (n = 27)

PSV Group (n = 23) Postrandomization

Pre randomization Pos trandomization Pre randomization Pos trandomizationAdjusted Difference PSV ndash PAV (95 CI)

Lorazepam equivalents

Cumulative dose mg

308 (207) 48 (257)a 137 (440) 54 (258) 237 (ndash580 to 1055)

Average daily dose mg

44 (255) 09 (43)a 23 (85) 06 (40)a 24 (ndash62 to 110)

Single-day dose mgb

08 (72) 12 (50) 16 (70) 10 (64) 32 (ndash13 to 78)

Morphine equivalents

Cumulative dose mg

5003 (6188) 1140 (3140)b 1710 (13310) 1610 (4860) 591 (ndash2515 to 3696)

Average daily dose mg

858 (1638) 170 (578)c 428 (1724) 136 (320)d 108 (ndash316 to 532)

Single-day dose mga

180 (1240) 187 (77) 100 (1920) 50 (630)d 334 (ndash026 to 670)d

Propofol

Cumulative dose mg

7400 (55500) 0 (700)a 29450 (85800) 0 (27660) 226 (ndash42156 to 42607)

Average daily dose mg

529 (11690) 0 (53)a 4879 (15980) 0 (2738)a 651 (ndash3211 to 4514)

Single-day dose mga

0 (0) 0 (0) 0 (1000) 0 (0) 344 (ndash3781 to 4469)

Neuroleptics

Days receiving neuroleptics n

0 (2) 0 (4) 0 (0) 1 (4)a ndash16 (ndash34 to 03)

Proportion of patients who received a neuroleptic

407 407 174 565b

NMBs

Days receiving NMBs n

1 (1) 0 (0)a 0 (1) 0 (0) ndash01 (ndash05 to 03)

Proportion of patients who received an NMB

518 148a 348 174

Corticosteroids

Days receiving corticosteroids n

0 (0) 0 (0) 0 (2) 0 (2) 02 (ndash06 to 10)

Proportion of patients who received corticosteroids

185 148 348 261

PAV = proportional assist ventilation PSV = pressure support ventilation NMB = neuromuscular blockera p lt 005b Single-day dose = total dose given day prior to randomization or total dose give day after randomizationc p lt 0005d p lt 00005e p le 007All data points except percentage of patients receiving neuroleptics NMBs or corticosteroids are presented as median (interquartile range)

鎮静剤鎮痛剤についてはPAV群とPSV群で変わりなし

その他bull 不整脈気胸プロトコル施中の死亡

は両群とも起こらなかった

筆者らの考察bull PSVはweaningに最も使われているモードだ

がそれと較しても遜ないweaningができることがわかった

bull アシストが少ない傾向であったがPAV群のが呼吸窮迫となる回数が少なかった

rarr同調性が良くなった結果の可能性

ICU退室までの期間はPAV群で短いという結果であったがpowerが少ない

筆者らの考察bull 単施設での結果であり他の施設にこの

新しいプロトコルそのものが受けれられるかは不明である

bull SBTに失敗した患者を主に対象としておりICU患者すべてに適応はできない

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる

Page 44: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

その他bull 不整脈気胸プロトコル施中の死亡

は両群とも起こらなかった

筆者らの考察bull PSVはweaningに最も使われているモードだ

がそれと較しても遜ないweaningができることがわかった

bull アシストが少ない傾向であったがPAV群のが呼吸窮迫となる回数が少なかった

rarr同調性が良くなった結果の可能性

ICU退室までの期間はPAV群で短いという結果であったがpowerが少ない

筆者らの考察bull 単施設での結果であり他の施設にこの

新しいプロトコルそのものが受けれられるかは不明である

bull SBTに失敗した患者を主に対象としておりICU患者すべてに適応はできない

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる

Page 45: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

筆者らの考察bull PSVはweaningに最も使われているモードだ

がそれと較しても遜ないweaningができることがわかった

bull アシストが少ない傾向であったがPAV群のが呼吸窮迫となる回数が少なかった

rarr同調性が良くなった結果の可能性

ICU退室までの期間はPAV群で短いという結果であったがpowerが少ない

筆者らの考察bull 単施設での結果であり他の施設にこの

新しいプロトコルそのものが受けれられるかは不明である

bull SBTに失敗した患者を主に対象としておりICU患者すべてに適応はできない

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる

Page 46: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

筆者らの考察bull 単施設での結果であり他の施設にこの

新しいプロトコルそのものが受けれられるかは不明である

bull SBTに失敗した患者を主に対象としておりICU患者すべてに適応はできない

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる

Page 47: HO Journal Club PAV vs PSV - JSEPTICrally adjusted ventilatory assist ventilation modes are activated and controlled by diaphragm excitation. Proportional assist ven-tilation modes

私bull 本試験では実際にPAVをweaning困難な患者に適

応すべきかどうかはわからない(どのような患者に適応すべきかrarr離脱まで時間のかかるprolongedweaningの患者に適しているのでは

bull 患者の呼吸努に応じたサポートをして(同調が少ない可能性)呼吸仕事量をモニタリングしながらweaningすることができるので(適切にサポートをする)weaningに適したモードである可能性はある

bull 今後の臨床アウトカムを据えた規模多施設研究が待たれる