Update on Sepsis Biomarkers

William T. McGee, M.D. MHA, FCCM, FCCP Critical Care Medicine

Associate Professor of Medicine and Surgery University of Massachusetts

759 Chestnut Street, Springfield, MA 01199

Tel: 413-794-5439 | Fax: 413-794-3987 william.mcgee@baystatehealth.org

Community Acquired Pneumonia (CAP)

♦ 5.16-6.11 cases/1000 persons/year • > 900,000 episodes/yr occur in adults > 65 yrs of age

• ~20% require hospitalization

♦ > 60,000 deaths in United States in 2005 • 8th most common cause of death

♦ 30-day mortality for hospitalized pts with CAP: 23%

Marrie TJ et al. Can Respir J 2005 File TM et al. Postgrad Med 2010

Kung HC et al. Natl Vital Stat Rep 2008

Community Acquired Pneumonia

♦ Estimated annual direct medical costs of care: $8.5 billion

♦ Given magnitude of antibiotic (Abx) use associated with CAP, ensuring correct diagnosis and appropriate empiric as well as definitive therapy for optimal course to achieve clinical cure and reduce risk of ADEs is critical

♦ Adherence to Abx guidelines improves survival in pts with CAP

• adherence to Abx treatment guidelines is inconsistent and the erroneous diagnosis of CAP and misuse of Abx is prevalent

♦ ↑’ed LOS, risk of adverse events, and emergence of resistance are negative outcomes associated with unnecessary Abx

Bonafede MM et al. Am J Manag Care 2012 Fakih MG et al. Clin Infect Dis 2003

Schouten JA et al. J Antimicrob Chemother 2005

Interventions to Reduce Abx Exposure

♦ Short-course therapy for CAP: clinically as effective as long-course therapy and associated with fewer adverse events

♦ IV to PO conversion: safe at 48-72 hr, even in pts with severe CAP, who meet criteria for clinical stability

♦ Biomarkers: use to guide the need for Abx at the start of and during therapy leads to reduced Abx exposure without evidence of pt harm

Nussenblatt V et al. Infect Dis Clin N Am 2013

Update on SepsisBiomarkers:

focus on Procalcitonin

6

Objectives

♦ Describe pathophysiology of Procalcitonin (PCT)

♦ Describe use of PCT in diagnosis and prognosis of

severe sepsis and septic shock

♦ Application of critical thinking skills in the context of

sepsis and appropriate use of biomarkers

Biomarkers

♦ Characteristic that is objectively measured and

evaluated as an indicator of normal biological process,

pathogenic process, or pharmacologic response to a

therapeutic intervention

♦ Usefulness is evaluated by:

– Capacity to provide timely information beyond what is readily

available from routine physiologic and clinical data (Speed +

Accuracy)

– Sensitivity and specificity

Marshall JC et al. Crit Care Med. 2009;37:2290-8.

Potential Role(s) for Biomarkers

♦ Identify patient with ↑ probability of disease, adverse outcome, or benefit from intervention

♦ Identify presence or absence of pathologic state or process

♦ Aid in risk stratification/prognosis

♦ Monitor response to an intervention or treatment

♦ Serve as surrogate endpoint

Marshall JC et al. Crit Care Med. 2009;37:2290-8.

Biomarkers in Sepsis Diagnosis

Sepsis is currently diagnosed using

clinical definitions combined with

culture results.

Blander & Sander. Beyond pattern recognition: five immune checkpoints for

scaling the microbial threat. Nature Reviews Immunology 12, 215-225 (March 2012)

Pathogen-Associated Molecular Patterns

(PAMPs) Danger-Associated Molecular Patterns

(DAMPs)

Procalcitonin (PCT)

♦ Precursor peptide of mature hormone calcitonin

♦ Released in multiple tissues in response to bacterial infections via direct stimulation of cytokines

♦ Cytokines such as interleukin (IL)-6 and tumor necrosis factor (TNF) show fast initial spike upon infection

• levels return to normal within a few hrs

• high variability of markers: major challenge for clinical utility

♦ C-reactive protein (CRP): ↑’es slowly with peak ~ 48-72 hr

• considered biomarker for inflammation, rather than infection

Muller B et al. J Clin Endocrinol Metab 2001 Meisner M. J Lab Med 1999

Procalcitonin

♦ Increases promptly within 4-6 hrs upon stimulation and ↓’es by ~ 50% daily if bacterial infection is controlled by immune system and supported by effective Abx

Haeruptle J et al. Eur J Clin Microbiol Infect Dis 2009 Meisner M. J Lab Med 1999

Time course PCT successful treatment

0

2

4

6

8

10

12

14

16

18

20

1 2 3 4 5 6 7 8

PCT Level

PCT Level

Days

14

Simon L. et al. Clin Infect Dis. 2004; 39:206-217.

Adding PCT results to clinical assessment improves the accuracy of the early clinical diagnosis of sepsis

• PCT levels accurately differentiate sepsis from noninfectious inflammation*

• PCT has been demonstrated to be the best marker for differentiating patients with sepsis from those with systemic inflammatory reaction not related to infectious cause

Sensitivity: 89% Specificity: 94%

NPV: 90% PPV:94%

15

Harbarth S et.al. AM J Resp Crit Care Med. 2001; 164:396-402

• When PCT is used as a reference, the sensitivity and specificity of sepsis diagnosis can be significantly increased compared with conventional clinical parameters.

Sensitivity: 94% Specificity: 77%

Adding PCT results to clinical assessment improves the accuracy of the early clinical diagnosis of sepsis

PCT Regulation on Cellular Level ♦ Induced in response to microbial toxins and bacterial-induced

cytokines (IL-1, IL-6, and TNF-α)

• released into bloodstream where it can be measured

♦ Conversely, production attenuated by cytokines released in response to viral infection (interferon (INF)-γ)

Muller B et al. J Clin Endocrinol Metab 2001 Christ-Crain M et al. Eur Respir J 2007 Linscheid P et al. Endocrinology 2005

17

♦ How can we use this cellular signal of infection in the

management of both septic and non

septic patients

♦ Goals

– Provide antibiotic therapy to pts who need it as soon as possible

– Avoid antibiotic prescription to those without infection

– Do both with a strong likelihood of being correct, at least as good

as other markers such as WBC, bands, fever, CRP

Useful Diagnostic Biomarker

♦ Helps distinguish bacterial infections from other inflammatory reactions or viral infections

♦ Strong correlation between concentration of PCT and extent and severity of bacterial infections has been observed in CAP

Gogos CA et al. J Infect Dis 2000

Utility in Clinical Settings

♦ 102 critically ill pts with systemic infections in ICU found similar PCT levels but lower CRP and IL-6 levels in pts treated with systemic corticosteriods

♦ Study of 32 healthy volunteers treated with prednisolone 2 hr before SIRS induced by injection of E. coli lipopolysaccharide (LPS)

• PCT: no inhibition within study period

• Other biomarkers: significantly inhibited in dose-dependant way

♦ PCT production does not rely on WBCs

• Dynamics are expected to be comparable in neutropenic pts

Muller B et al. J Leukoc Biol 2002 de Kruif MD et al. Intensive Care Med 2008

Massaro KS et al. BMC Infect Dis 2007

Commercial Detection Assays

♦ Currently, different assays available for PCT with individual performance characteristics

♦ Reasonably low detection limit especially important when antimicrobial stewardship decisions intended

♦ Kryptor assay: lower level of detection of 0.06 µg/mL

• Based on sheep polyclonal anticalcitonin Ab

• Assay used in most intervention studies

♦ VIDAS system from bioMerieux: equally sensitive assay

Steinback G et al. Clin Chem Lab Med 2004 Schuetz P et al. Clin Biochem 2010

Diagnosis & Prognosis

♦ Inherent complexity evaluating diagnostic capability of a novel biomarker without presence of diagnostic reference standard

• Optimally, morphological verification such as growth of typical pathogens or proving histopathology can be obtained to establish “correct” diagnosis

♦ Causative pathogen cannot be detected in 80% of pts with suspected bloodstream infections

♦ > 70% of pts with radiographically confirmed PNA • Causative microbe never identified/isolated

• Usually a syndromic definition of lower respiratory tract infections (LRTIs)

♦ No “gold standard” to discriminate bacterial from viral etiology Pizzo PA. Eur J Cancer Clin Oncol 1989

Muller B et al. Crit Care Med 2000 Marshall JC. J Leuko Biol 2008

Diagnosis Dilemma

♦ Real-life concept focuses primarily on outcomes of pts with out without Abx therapy, while discarding alleged gold standards

♦ Potential dilemma: ambiguity in deciding whether a pt who received Abx “just to be sure” and subsequently recovers had a good outcome because of OR despite Abx therapy

• For example, self-limiting viral LRTIs will also recover despite Abx therapy

♦ In sepsis and pneumonia, performance of PCT as a tool for Abx guidance is best measured in randomized intervention trials

• Assuming there was no outcome-relevant bacterial infection if the pt recovers without Abx therapy

Haubitz S et al. Expert Rev Respir Med 2013

PCT Cutoff Ranges

♦ Similar treatment recommendation algorithms based on PCT cutoff ranges were all used in all published stewardship studies

♦ 4 classes of Abx treatment recommendations

• Strongly discouraged

• Discouraged

• Recommended

• Strongly recommended

♦ Cutoff ranges derived from multilevel likelihood calculations obtained in observational studies

• Reflect likelihood of bacterial infection in a distinct entity of infection

Muller B et al. Crit Care Med 2000 Haubitz S et al. Expert Rev Respir Med 2013

PCT Algorithm in LRTIs

♦ Initiation or continuation of Abx was more or less encouraged when levels > 0.5 µg/mL or > 0.25 µg/mL or discouraged when levels were < 0.1 µg/mL or < 0.25 µg/mL

♦ When Abx therapy was withheld, clinical re-evaluation and a 2nd measurement of PCT performed after 6-24 hr if clinical condition did not improve spontaneously

♦ Specific “overruling” criteria were defined, were the algorithm could be bypassed and Abx initiated at physician’s discretion

• Namely in life-threatening disease or immediate need for ICU admission because of respiratory or hemodynamic instability

Haubitz S et al. Expert Rev Respir Med 2013

ProRESP Trial

♦ 1st intervention study testing hypothesis that PCT can be used to guide initiation of Abx in pts with LRTI in the ED

♦ 243 pts with LRTI • Outcomes of both groups were not different

• PCT-guided group: less Abx Rx (44 vs 83%)

• Strongest effect seen in pts with acute bronchitis and exacerbated COPD

Christ-Crain M et al. Lancet 2004

ProCAP and ProCOLD

♦ 2 subsequent prospective RCTs evaluating effect of PCT guidance for Abx discontinuation in CAP and to assess pt safety over 6-mo follow-up period in pts with exacerbated COPD

♦ ProCAP: 302 pts with mostly severe CAP (>60% with PSI IV & V) • PCT-guided therapy: Abx courses 65% shorter than in standard regimens

♦ ProCOLD: 208 pts with exacerbated moderate-to-severe COPD • PCT-guided therapy: Abx Rx’s reduced from 72% to 40% initially

• Lasting effect in follow-up period of 6-mo

• Safety and Outcomes: Re-exacerbation rates and FEV-1 improvement over 6-mo were same in intervention and control groups

Christ-Crain M et al. Am J Respir Crit Care Med 2006 Stolz D et al. Chest 2007

♦ Large multicenter trial

– 6 hospitals in Switzerland with > 1300 pts

– Predefined web-based guidelines had to be followed for every pt in control group in order to ensure optimal adherence to guidelines

♦ Abx exposure vs. treatment according to defined guidelines: • ↓ by 32.4% in CAP; 50.4% in exacerbated COPD; 65% in acute bronchitis

♦ Overall reduction in Abx use translated into reduction in associated side effects of ~ 30% in pts treated according to PCT

Schuetz P et al. JAMA 2008

Additional a priori Results

• (B) Predictive value of baseline PCT to determine + culture (blood, urine, respiratory) – Positive vs. Negative culture

• 9.8ng/mL [1.7-41.3] vs. 3.3ng/mL[0.6-15.8] p<0.001

• 61% of cultures were positive

• (C) Predictive value of baseline PCT to determine sepsis severity – Septic shock vs. Sepsis

• 13.6ng/mL [2.7-55.2] vs. 3.6[0.5-15.6], p<0.001

Adapted from Shehabi Y et al. Procalcitonin algorithm in critically ill adults with undifferentiated infection or sepsis. Amer J Resp Crit Care Med 2014

Nov 15;190(10):1102-10

Additional a priori Results

• Baseline PCT was similar in survivors and non-survivors however there was a

significantly faster decline overtime in the serial PCT levels in survivors

• Baseline cut off of ≤ 3ng/mL excluded positive blood culture with a sensitivity of

90% (95% CI, 82-89) and a NPV of 96% (95% CI, 93-99)

• Baseline cut off of ≤ 0.1ng/mL excluded positive culture in the first 72h with a

sensitivity of 100% and NPV of 100%

Adapted from Shehabi Y et al. Procalcitonin algorithm in critically ill adults with undifferentiated infection or sepsis. Amer J Resp Crit Care Med 2014

Nov 15;190(10):1102-10

♦ In 2012, all these data were pooled in large meta-analysis using individual data of 4221 patients in 14 trials

♦ Markedly reduced Abx exposure overall • 8d vs. 4d and adjusted difference of -3.47d (95% CI: -3.78 to -3.17)

• No increase in mortality or treatment failure

Schuetz P et al. Clin Infect Dis 2012

Meta-analysis Results

♦ Primary care setting: mainly due to ↓ Rx rates (23% vs. 63%) • Pts with upper ARI: 15% vs. 48%, adjusted OR: 0.14 [95% CI: 0.09-0.22]

• Pts with bronchitis: 24% vs. 66%, adjusted OR: 0.15 [95% CI: 0.10-0.23]

♦ In ED and ICU pts, shorter courses of Abx treatment were found • ED: 7d vs. 10d, difference - 3.7d [95% CI: - 4.09 to - 3.31]

• ICU: 8d vs. 12d, difference - 3.17d [95% CI: - 4.28 to – 2.06]

♦ In CAP, Abx exposure markedly shorter: • Difference – 3.34d [95% CI: - 3.79 to – 2.88]

• Associated with lower risk of treatment failure (19.1% vs. 21.9% in standard treatment arm) adjusted OR: 0.82; 95% CI: 0.71 – 0.97

Schuetz P et al. Clin Infect Dis 2012

Schuetz P et al. JAMA 2013

Results: All Subjects

Schuetz P et al. Clin Infect Dis 2012

Results: Exacerbated COPD

Schuetz P et al. Clin Infect Dis 2012

Assay Characteristics

♦ Timely decision making: great importance in bacterial infections

♦ 2 retrospective studies: improved mortality in pts with CAP who received 1st dose of Abx within 4-8 hr

♦ In all intervention trials, PCT measured using rapid sensitive assay with assay time of 20 min

• Results readily available around the clock and within 1 hr

Mandell LA et al. Clin Infect Dis 2007 Schuetz P et al. JAMA 2009

Haubitz S et al. Expert Rev Respir Med 2013

Implications for Antimicrobial Stewardship

♦ Key in real-life setting: strong adherence rate to PCT algorithm to have same effect of reduced Abx exposure in clinical practice as found in controlled studies

♦ Unlike controlled studies were adherence is monitored

• Results frequently inadequately implemented in daily practice

♦ Aujesky et al: 37.4% of pts with low PSI scores hospitalized

• Mainly due to comorbidities

• Reflects difficulty of implementing guidelines into clinical practice

Aujesky D et al. Clin Infect Dis 2009

PCT Registry Study

♦ Investigated feasibility/effectiveness of PCT-guided stewardship in real-life setting of medical clinic in Switzerland

♦ 302 pts with LRTIs of all severities (64% PSI IV and V)

♦ 73% of pts treated according to PCT-guided algorithm

• Abx therapy applied in 71% of pts

• Median duration of treatment: 6d

• Significant improvement when compared to control group from previous trial at same hospital (ProHOSP)

Schuetz P et al. Eur J Clin Microbiol Infect Dis 2010

ProREAL

♦ Study of 1759 pts in US, France, Switzerland

♦ Overall adherence to PCT-guided stewardship in 68% pts • Adherence: acute bronchitis (81%); exacerbated COPD (70%); CAP (64%)

• Outpatient setting (86%) vs. Inpatient setting (66%)

♦ After multivariate adjustment, significantly shorter Abx exposure demonstrated with PCT algorithm

• 5.9d vs. 7.4d; difference – 1.51d [95% CI: - 2.04 to – 0.98]

♦ No increase in risk of combined adverse outcome end point within 30d of follow-up in respect to withholding Abx and early cessation

Albrich WC et al. Arch Intern Med 2012

Associations of Procalcitonin Testing with Clinical Outcomes and Antibiotic Use

Procalcitonin Group (n = 3336)

Control Group (n = 3372) Between-Group Difference (95% CI)

Adjusted OR (95% CI)a

P Value

Clinical Outcomes

30-d mortality, No. (%) 286 (8.6) 336 (10.0) 0.83 (0.70 to 0.99) .04

Treatment failure, No. (%)b 768 (23.0) 841 (24.9) 0.90 (0.80 to 1.01) .07

Length of ICU stay, median (IQR), d 8.0 (4.0 to 17.0) 8.0 (4.0 to 17.0) 0.39 (−0.81 to 1.58) .52

Length of hospital stay, median (IQR), d 8.0 (2.0 to 17.0) 8.0 (2.0 to 17.0) −0.19 (−0.96 to 0.58) .63

Antibiotic-related adverse effects, No./total (%) 247/1513 (16.3) 336/1521 (22.1) 0.68 (0.57 to 0.82) .001

Antibiotic Exposure

Rates for initiation of antibiotics, No./total (%) 2351/3288 (71.5) 2894/3353 (86.3) 0.27 (0.24 to 0.32) .001

Duration of antibiotics, median (IQR), d 6.0 (4.0 to 10.0) 8.0 (6.0 to 12.0) −1.83 (−2.15 to −1.50) .001

Total exposure of antibiotics, median (IQR), d 5.0 (0 to 8.0) 7.0 (3.0 to 11.0) −2.43 (−2.71 to −2.15) .001

Procalcitonin Conclusions

♦ Evidence supports PCT as accurate surrogate biomarker for likelihood and severity of bacterial infections

♦ In CAP and other respiratory infections, PCT-guided algorithms resulted in ↓’ed Abx exposure

• Maintaining similar or better level of safety compared with standard care

♦ Reductions in Abx use translate to decreased: • Costs, risk of side effects, and bacterial resistance

• Especially important as acute respiratory infections represent most frequent reason for Abx Rx worldwide

Conclusions

♦ Short-course therapy for CAP: clinically as effective as

long-course therapy and associated with fewer adverse

events

♦ IV to PO conversion: safe at 48-72 hr, even in pts with

severe CAP, who meet criteria for clinical stability

♦ Procalcitonin: use to guide the need for Abx at the start

of therapy leads to reduced Abx exposure without

evidence of pt harm

Questions