Ie

Ja

b

c

d

e

f

a

ARA

KGLEP

1

aEoifcm[i

U

0d

International Journal of Antimicrobial Agents 39 (2012) 69– 72

Contents lists available at SciVerse ScienceDirect

International Journal of Antimicrobial Agents

jou rn al h om epa ge: h t tp : / /www.e lsev ier .com/ locate / i jant imicag

mproving vancomycin prescription in critical illness through a drug usevaluation process: a weight-based dosing intervention study

anice Lia,b,∗, Andrew A. Udyc,d, Carl M.J. Kirkpatricka,e, Jeffrey Lipmanc,d, Jason A. Robertsc,d,f

School of Pharmacy, The University of Queensland, Brisbane, AustraliaDepartment of Pharmacy, National University Hospital, SingaporeBurns, Trauma and Critical Care Research Centre, The University of Queensland, Brisbane, AustraliaDepartment of Intensive Care Medicine, Royal Brisbane and Women’s Hospital, Brisbane, AustraliaCentre for Medicine Use and Safety, Faculty of Pharmacy and Pharmaceutical Sciences, Monash University, Melbourne, AustraliaPharmacy Department, Royal Brisbane and Women’s Hospital, Brisbane, Australia

r t i c l e i n f o

rticle history:eceived 12 August 2011ccepted 26 August 2011

eywords:lycopeptidesoading doseducational interventionharmacokinetics

a b s t r a c t

Vancomycin is currently recommended as first-line therapy for many meticillin-resistant Staphylococcusaureus (MRSA) infections. Recent guidelines have advocated loading doses (25–30 mg/kg) in critically illpatients in order to achieve therapeutic concentrations rapidly. However, weight-based loading dosesare still not widely practised. A drug use evaluation was performed to improve the appropriateness ofvancomycin initial doses in a population of critically ill adults. An educational intervention incorpo-rating a vancomycin dosing protocol was carried out. Data were collected pre and post intervention.Vancomycin exposure [area under the concentration–time curve (AUC)] in the first 24 h was determinedusing serum concentrations and the Bayesian software TCIWorks. Initial vancomycin doses and expo-sures were compared between the pre- and post-intervention groups using �2 and Mann–Whitney tests.A total of 111 vancomycin courses were analysed in the pre-intervention (n = 80) and post-intervention(n = 31) groups. Patients in the post-intervention group had significantly higher median weight-basedinitial doses (20.0 mg/kg vs. 12.5 mg/kg; P < 0.001) compared with the pre-intervention group. This cor-responded to significantly higher median vancomycin exposures (366.0 mg h/L vs. 262.5 mg h/L; P < 0.01)

in the post-intervention group. Despite higher weight-based initial doses, only 32.3% of patients in thepost-intervention group had achieved optimal vancomycin exposures (AUC/minimum inhibitory con-centration ratio ≥400) in the first 24 h of therapy. A vancomycin dosing protocol improved the initialdosing of vancomycin and the proportion of patients who rapidly achieved optimal vancomycin expo-sures. However, subtherapeutic exposures were still prevalent and may warrant more vigilant promotionof the dosing protocol to ensure that recommended vancomycin doses are used in this population.

lsevie

© 2011 E

. Introduction

Infection in the Intensive Care Unit (ICU) continues to represent significant cause of mortality in critically ill patients. The recentxtended Prevalence of Infection in Intensive Care (EPIC II) studybserved that Staphylococcus aureus was the most common organ-sm isolated, with meticillin-resistant S. aureus (MRSA) accountingor two-thirds of such isolates in some areas [1]. Of significant con-ern, the high frequency of isolation of MRSA is matched by high

ortality rates, which have been reported to be as high as 50%

2,3]. The challenge of effectively treating Gram-positive infectionsn the present clinical environment of escalating resistance has

∗ Corresponding author. Present address: Department of Pharmacy, Nationalniversity Hospital, Singapore. Tel.: +65 6772 5239; fax: +65 6772 4705.

E-mail address: janice li@nuhs.edu.sg (J. Li).

924-8579/$ – see front matter © 2011 Elsevier B.V. and the International Society of Chemoi:10.1016/j.ijantimicag.2011.08.017

r B.V. and the International Society of Chemotherapy. All rights reserved.

necessitated re-evaluation of dosing approaches for older antibi-otics such as vancomycin. Although regarded as the mainstay oftherapy for such infections, controversy has reigned as to the opti-mal way to dose vancomycin in critically ill patients.

Emerging data are now addressing this knowledge gap,highlighting the importance of rapidly achieving therapeutic con-centrations [4–7]. Rapid attainment of therapeutic vancomycintrough concentrations has been suggested to improve efficacywhilst decreasing the emergence of resistance. Furthermore, recentclinical practice guidelines endorsed by the Infectious DiseasesSociety of America (IDSA) have recommended targeting highervancomycin trough concentrations (15–20 mg/L), particularly incomplicated infections [8,9]. Such trough concentrations are pro-

posed to increase the likelihood of achieving an area under theconcentration–time curve to minimum inhibitory concentration(AUC/MIC) ratio ≥400, which correlates with the clinical efficacyof vancomycin [10]. It follows that more consistent achievement of

otherapy. All rights reserved.

7 f Antimicrobial Agents 39 (2012) 69– 72

tt

ioa[pm1acaafi[

ecwiosoc

2

2

oeasobi2

ocftsca

2

odpaw

2

caftv

0102030405060708090100

% o

f pat

ient

s

Pre-intervention

Post-intervention

0 J. Li et al. / International Journal o

herapeutic vancomycin exposures early in therapy may increasehe likelihood of treatment success [11,12].

Recent consensus recommendations from major governing bod-es have advocated loading doses in the range of 25–30 mg/kg basedn total body weight in order to hasten the achievement of ther-peutic vancomycin trough concentrations in critically ill patients10]. These recommendations have been supported by populationharmacokinetic modelling data [13]. Despite this, it remains com-on practice to administer a standard vancomycin dose of 1 g every

2 h for all patients regardless of body weight. Although dosesre subsequently adjusted based on therapeutic drug monitoringoncentration data, this practice results in a significant delay inchieving therapeutic concentrations in critically ill patients. Thevailable literature, however, clearly points to a survival bene-t of early appropriate antibiotic therapy in critically ill patients14–16].

Thus, loading doses that rapidly result in optimal vancomycinxposures have been proposed as a strategy to improve clinical out-omes. In light of these data, we conducted a study to determinehether introduction of a vancomycin dosing protocol emphasis-

ng the need for loading doses would lead to an increased numberf appropriate initial vancomycin doses (25–30 mg/kg). We alsoought to determine whether this would lead to achievement ofptimal initial vancomycin exposures, defined as higher troughoncentrations and AUCs.

. Methods

.1. Study design and patient population

A full drug use evaluation cycle was performed, which consistedf retrospective data collection, intervention and post-interventionvaluation phases. The study was performed in a 33-bed ICU of

major tertiary referral hospital. The only patients not repre-ented include paediatric, post-operative cardiothoracic and solidrgan transplant patients. The retrospective study was conductedetween January 2009 and January 2010 (pre intervention). Post-

ntervention data collection was undertaken from April 2010 to July010.

All critically ill, adult (age >18 years) patients who were initiatedn vancomycin in the ICU with at least one serum vancomycin con-entration measured during the course of therapy were screenedor eligibility. Exclusion criteria included vancomycin adminis-ration by continuous infusion and renal dysfunction (defined aserum creatinine concentration >170 �mol/L on initiation of van-omycin or those requiring renal replacement therapy). Ethicalpproval was granted by the local institutional review board.

.2. Intervention

A written vancomycin dosing protocol advocating loading dosesf 25–30 mg/kg based on total body weight was developed andistributed to all medical staff in the ICU. The protocol was cham-ioned by senior medical and pharmacy staff, including writtennd electronic reminders. A series of continuing education lecturesere also provided on a regular basis.

.3. Data collection

Demographic and clinical data were obtained from the medi-

al notes. These included age, weight, height, sex, serum creatininend albumin concentrations upon initiation of vancomycin, needor mechanical ventilation, vasopressor requirement, and informa-ion relating to drug prescription and administration at the time ofancomycin dosing.

Vancomycin initial dose

Fig. 1. Prescription patterns of vancomycin initial doses pre- and post intervention.

2.4. Prediction of vancomycin trough concentrations and areasunder the concentration–time curve (AUCs)

As this study was essentially observational in nature, we did notspecify the timing of vancomycin concentrations, which was deter-mined solely by the treating clinician. As such, inconsistent localpractice meant that many patients did not have vancomycin troughconcentrations taken within the first 24 h of treatment. There-fore, predicted initial vancomycin concentrations and AUCs weredetermined using the Bayesian dose individualisation software TCI-Works (http://www.tciworks.info) with a pharmacokinetic modeldeveloped in a population including 102 critically ill patients [5].Relevant patient covariate data were entered into TCIWorks toobtain the predicted trough concentrations and AUCs at 24–36 h.

2.5. Data analysis

Vancomycin initial doses, trough concentrations and AUCs werecompared in the pre- and post-intervention groups. The AUC/MICratios were calculated based on the assumption that MICs were1 mg/L for all cases, given local susceptibility patterns. Target expo-sures were defined as AUC/MIC ≥400 [10].

Data were analysed statistically using PASW Statistics GradPackv.17.0 (SPSS Inc., Chicago, IL). Kolmogorov–Smirnov tests were usedto determine the normality of the different variables. Pre- and post-intervention groups were compared using a �2 test/Fisher’s exacttest where analysis assumptions were met, an independent Stu-dent’s T-test or Mann–Whitney test as appropriate. A P-value of<0.05 was considered statistically significant for all tests.

3. Results

A total of 111 vancomycin treatment courses were analysed inthe pre-intervention (n = 80) and post-intervention (n = 31) phases.Demographic and clinical characteristics of the included patientswere similar in both groups (Table 1).

The post-intervention group had a significantly smaller pro-portion of patients who were prescribed a standard 1 g initialdose compared with the pre-intervention group (32.3% vs. 86.3%;P < 0.001), indicating improved individualisation of dosing (Fig. 1).Compared with none in the pre-intervention group, 8/31 patients(26%) in the post-intervention group had an initial vancomycinweight-based dose ≥25 mg/kg. The proportion of patients who hadachieved the target AUC/MIC ≥400 was significantly increased in

the post-intervention group (10.0% vs. 32.3%; P = 0.008).

There was a significant increase in the median weight-basedinitial vancomycin dose from 12.5 mg/kg in the pre-interventiongroup to 20.0 mg/kg in the post-intervention group (P < 0.001)

J. Li et al. / International Journal of Antimicrobial Agents 39 (2012) 69– 72 71

Table 1Characteristics of patients included in the study.

Characteristic Pre-intervention phasea Post-intervention phase P-value

Total no. of courses 80 31Type of patient [n (%)]

Burns 10 (12.5) 1 (3.2)Trauma 14 (17.5) 2 (6.5)Surgical 17 (21.3) 8 (25.8)Neurosurgical 15 (18.8) 7 (22.6)Medical 24 (30.0) 13 (41.9)

Female sex [n (%)] 28 (35.0) 14 (45.2) 0.32b

Age (years) (mean ± S.D.) 50.2 ± 15.8 54.9 ± 17.8 0.17c

Weight (kg) [median (range)] 80 (65.0–90.8) 78.0 (62.0–90.0) 0.58d

Height (cm) [median (range)] 172.5 (165.0–179.5) (n = 68) 170.0 (160.0–178.0) 0.69d

BMI (kg/m2) [median (range)] 26.1 (22.9–29.4) (n = 68) 26.2 (21.8–29.7) 0.73d

Baseline serum creatinine concentration (�mol/L) [median (range)] 62.0 (49.0–86.0) 64.0 (47.0–89.0) 0.83d

Baseline serum albumin concentration (g/L) [median (range)] 24.0 (21.0–28.0) (n = 77) 24.0 (21.0–28.0) 0.89d

Need for mechanical ventilation [n (%)] 60 (75.0) 24 (77.4) 0.79b

Need for vasopressor [n (%)] 31 (38.8) 8 (25.8) 0.20b

Length of stay in ICU before initiation (days) [median (range)] 6.0 (1.0–10.0) 6.0 (1.0–13.0) 0.78d

S.D., standard deviation; BMI, body mass index; ICU, Intensive Care Unit.a n = 80 unless otherwise stated.

(p(

4

ipvt>dtosawdmidhp

Fclr

b �2 test.c Student’s T-test.d Mann–Whitney test.

Fig. 2). This corresponded with a significant increase in the medianredicted trough concentration at 24 h from 6.8 mg/L to 10.1 mg/LP = 0.013) and AUC/MICs from 262.5 to 365.0 (P < 0.001).

. Discussion

This study has described a drug use evaluation process lead-ng to improved initial prescribing of vancomycin in critically illatients. The results demonstrate that a higher initial dose ofancomycin is more likely to result in therapeutic trough concen-rations and AUCs in the first 24 h of therapy (trough concentrations15 mg/L and AUCs >400 mg h/L). A dosing approach using loadingoses of 25–30 mg/kg will give therapeutic vancomycin concen-rations rapidly, which in turn may facilitate improved clinicalutcomes [10], although this is yet to be demonstrated conclu-ively. However, what has been shown is the importance of earlynd adequate antibiotic cover in ensuring optimal outcomes. Thisas recently highlighted by Lodise and McKinnon [17] wherebyelayed treatment was found to be an independent predictor ofortality in patients with MRSA bacteraemia. Giving a standard

nitial dose of 1 g and waiting for vancomycin serum concentrationata before optimising therapy might mean missing the ‘goldenour’ for optimal antibiotic therapy, especially in critically ill septicatients [18]; such a delay may consequently result in unfavourable

ig. 2. Box plots comparing weight-normalised initial doses and predicted van-omycin trough concentrations in the pre- and post-intervention phases. Thick blackines represent median values and boxes represent the 25th to 75th percentile. Tailsepresent the range.

outcomes. Furthermore, the altered pharmacokinetics of van-comycin in critically ill patients means that standard dosingapproaches are even more likely to result in subtherapeutic antibi-otic exposure [4,19,20], which predisposes these patients to a riskof subsequent treatment failure. Our dosing approach concurs withmuch of the current research on optimising vancomycin dosing,whereby an emphasis is placed on achieving therapeutic concen-trations early in therapy [5,6,21]. Current research supports timelyadministration of appropriate antibiotics, thus it would appear ben-eficial to achieve optimal vancomycin exposures early in therapyto reduce the likelihood of suboptimal outcomes.

The current results also demonstrate the positive impact of aneducational intervention on vancomycin prescribing practices ata tertiary referral ICU. Not surprisingly, the majority of standardvancomycin 1 g doses pre intervention were consistent with trendsreported elsewhere [22–25]. Implementation of a vancomycin dos-ing protocol helped to improve the initial dosing of vancomycin.The greater variation in vancomycin initial doses prescribed inthe post-intervention group demonstrates the increased propen-sity towards individualised dosing, with a tendency to prescribehigher initial doses in this group.

However, although there was an increase in weight-based ini-tial vancomycin doses in the post-intervention group, the majorityof initial doses were still not in the target range of 25–30 mg/kg.The median initial dose achieved was 20 mg/kg, which may not besufficient to achieve target vancomycin exposures rapidly. Conse-quently, the initial trough concentrations and AUC/MICs achievedin the post-intervention group were still suboptimal at 10 mg/L and365, respectively, with only ca. 30% of patients having achievedoptimal AUCs. It is likely that higher loading doses given inthe recommended range of 25–30 mg/kg may have enabled theachievement of target trough concentrations (15–20 mg/L) andoptimal AUC/MICs within the first 24 h of therapy.

We recognise that one potential barrier in the uptake of this dos-ing protocol could be the presence of significant obesity. Although itis assumed that dosing vancomycin according to total body weightis a relatively uncomplicated process, it must not be overlookedthat morbidly obese patients would thus require a larger than com-monly accepted dose of vancomycin. The need for higher doses

that the prescriber does not have experience with, and may havereduced confidence in, could represent a potential barrier in pre-scribing an appropriate initial dose. Thus, we recognise that work isstill needed to improve further the prescribing of appropriate initial

7 f Antim

visiapb

pivtpcampmlisata

bitatrsd

bf

R

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

2 J. Li et al. / International Journal o

ancomycin doses in this population. Ongoing education and famil-arisation with the protocol is still required to improve uptake andustained application of such a dosing approach. Visual remindersn the form of posters as well as verbal reminders by nursing stafft the time of administration could help facilitate improved com-liance with the protocol. Regular audits and feedback have alsoeen shown to help improve compliance with protocols.

There are some limitations to this study. The retrospectivere-intervention data collection may have resulted in inherent

naccuracies, although every attempt has been made to ensure dataalidity. Second, the educational intervention may not have cap-ured all physicians, which may have affected compliance with therotocol and subsequent results. Third, the use of predicted troughoncentrations as opposed to true trough concentrations may haveffected the results, although this inaccuracy is thought to be mini-al as the population pharmacokinetic model used in the validated

rogramme TCIWorks had been previously evaluated for perfor-ance in a similar population. Lastly, the small sample size may

imit the generalisability of these results; however, these results aren agreement with previous studies of vancomycin dosage optimi-ation in critically ill patients [5,6], which suggest that such dosingpproaches may be applicable in the general ICU population. Cau-ion should be exercised in the population with renal impairments these were not evaluated in this study.

In conclusion, this study has demonstrated that use of a weight-ased dosing protocol can improve initial vancomycin prescription

n a population of critically ill patients. However, subtherapeu-ic exposures within the first 24 h of therapy were still prevalentnd may warrant more vigilant promotion of the dosing protocolo encourage greater compliance with recommended doses. Weecommend that further prospective pharmacokinetic and clinicaltudies be carried out to confirm the utility of vancomycin loadingoses and its association with outcomes in critically ill patients.

Funding: No funding sources.Competing interests: None declared.Ethical approval: The Institutional Review Board of Royal Bris-

ane and Women’s Hospital (Brisbane, Australia) granted approvalor this research under Low and Negligible Risk Research.

eferences

[1] Vincent JL, Rello J, Marshall J, Silva E, Anzueto A, Martin CD, et al. Internationalstudy of the prevalence and outcomes of infection in intensive care units. JAMA2009;302:2323–9.

[2] Subbe CP, Rao GG, Sedgwick P, Van Heerden N, Groba CB. MRSA: predictor ofoutcome in critically ill patients. Br J Anaesth 2000;84:662.

[3] Ho KM, Robinson JO. Risk factors and outcomes of methicillin-resistant Staphy-lococcus aureus bacteraemia in critically ill patients: a case control study.Anaesth Intensive Care 2009;37:457–63.

[4] Revilla N, Martín-Suárez A, Pérez MP, González FM, Fernández de Gatta Mdel M.Vancomycin dosing assessment in intensive care unit patients based on a pop-

ulation pharmacokinetic/pharmacodynamic simulation. Br J Clin Pharmacol2010;70:201–12.

[5] Thomson AH, Staatz CE, Tobin CM, Gall M, Lovering AM. Development andevaluation of vancomycin dosage guidelines designed to achieve new targetconcentrations. J Antimicrob Chemother 2009;63:1050–7.

[

icrobial Agents 39 (2012) 69– 72

[6] Pea F, Furlanut M, Negri C, Pavan F, Crapis M, Cristini F, et al. Prospec-tively validated dosing nomograms for maximizing the pharmacodynamicsof vancomycin administered by continuous infusion in critically ill patients.Antimicrob Agents Chemother 2009;53:1863–7.

[7] Patanwala AE, Norris CJ, Nix DE, Kopp BJ, Erstad BL. Vancomycin dosing forpneumonia in critically ill trauma patients. J Trauma 2009;67:802–4.

[8] Liu C, Bayer A, Cosgrove SE, Daum RS, Fridkin SK, Gorwitz RJ, et al. Clinical prac-tice guidelines by the Infectious Diseases Society of America for the treatmentof methicillin-resistant Staphylococcus aureus infections in adults and children.Clin Infect Dis 2011;52:11–8.

[9] Tunkel AR, Hartman BJ, Kaplan SL, Kaufman BA, Roos KL, Scheld WM, et al.Practice guidelines for the management of bacterial meningitis. Clin Infect Dis2004;39:1267–84.

10] Rybak M, Lomaestro B, Rotschafer JC, Moellering Jr R, Craig W, Billeter M, et al.Therapeutic monitoring of vancomycin in adult patients: a consensus reviewof the American Society of Health-System Pharmacists, the Infectious DiseasesSociety of America, and the Society of Infectious Diseases Pharmacists. Am JHealth Syst Pharm 2009;66:82–98.

11] Hidayat LK, Hsu DI, Quist R, Shriner KA, Wong-Beringer A. High-dose van-comycin therapy for methicillin-resistant Staphylococcus aureus infections:efficacy and toxicity. Arch Intern Med 2006;166:2138–44.

12] Mohr JF, Murray BE. Point: Vancomycin is not obsolete for the treatment ofinfection caused by methicillin-resistant Staphylococcus aureus. Clin Infect Dis2007;44:1536–42.

13] Roberts JA, Taccone FS, Udy AA, Vincent J-L, Jacobs F, Lipman J. Vancomycin dos-ing in critically ill patients: robust methods for improved continuous-infusionregimens. Antimicrob Agents Chemother 2011;55:2704–9.

14] Gaieski DF, Mikkelsen ME, Band RA, Pines JM, Massone R, Furia FF, et al. Impactof time to antibiotics on survival in patients with severe sepsis or septic shock inwhom early goal-directed therapy was initiated in the emergency department.Crit Care Med 2010;38:1045–53.

15] Kumar A, Roberts D, Wood KE, Light B, Parrillo JE, Sharma S, et al. Dura-tion of hypotension before initiation of effective antimicrobial therapy isthe critical determinant of survival in human septic shock. Crit Care Med2006;34:1589–96.

16] Garnacho-Montero J, Aldabo-Pallas T, Garnacho-Montero C, Cayuela A, JimenezR, Barroso S, et al. Timing of adequate antibiotic therapy is a greater determinantof outcome than are TNF and IL-10 polymorphisms in patients with sepsis. CritCare 2006;10:R111.

17] Lodise TP, McKinnon PS. Clinical and economic impact of methicillin resistancein patients with Staphylococcus aureus bacteremia. Diagn Microbiol Infect Dis2005;52:113–22.

18] Nathwani D, Tillotson GS. Vancomycin for Staphylococcus aureus therapyof respiratory tract infections: the end of an era? Int J Antimicrob Agents2003;21:521–4.

19] del Mar Fernandez de Gatta Garcia M, Revilla N, Calvo MV, Dominguez-GilA, Sanchez Navarro A. Pharmacokinetic/pharmacodynamic analysis of van-comycin in ICU patients. Intensive Care Med 2007;33:279–85.

20] Llopis-Salvia P, Jimenez-Torres NV. Population pharmacokinetic parame-ters of vancomycin in critically ill patients. J Clin Pharm Ther 2006;31:447–54.

21] Nunn MO, Corallo CE, Aubron C, Poole S, Dooley MJ, Cheng AC. Vancomycindosing: assessment of time to therapeutic concentration and predictive accu-racy of pharmacokinetic modeling software. Ann Pharmacother 2011;45:757–63.

22] Kitzis MD, Goldstein FW. Monitoring of vancomycin serum levels for the treat-ment of staphylococcal infections. Clin Microbiol Infect 2006;12:92–5.

23] Hall 2nd RG, Payne KD, Bain AM, Rahman AP, Nguyen ST, Eaton SA, et al. Mul-ticenter evaluation of vancomycin dosing: emphasis on obesity. Am J Med2008;121:515–18.

24] Pea F, Viale P. Should the currently recommended twice-daily dosingstill be considered the most appropriate regimen for treating MRSAventilator-associated pneumonia with vancomycin? Clin Pharmacokinet2008;47:147–52.

25] Tobin CM, Darville JM, Thomson AH, Sweeney G, Wilson JF, MacGowanAP, et al. Vancomycin therapeutic drug monitoring: is there a consensusview? The results of a UK National External Quality Assessment Scheme (UKNEQAS) for Antibiotic Assays questionnaire. J Antimicrob Chemother 2002;50:713–18.