1274. Developing a Biomarker-Driven Algorithm to Improve Antibiotic Use in the Pediatric Intensive Care Unit: the Optimizing Antibiotic Strategies in Sepsis (OASIS) Study
Session: Oral Abstract Session: Pediatric Antimicrobial Stewardship
Friday, October 9, 2015: 2:15 PM
Room: 5--AB

Background: Biomarkers that identify children at low risk for bacterial infection may reduce unnecessary antibiotic use. We sought to derive an algorithm to classify critically ill children with SIRS at low likelihood for bacterial infection.

Methods: Prospective cohort study of children with new-onset SIRS admitted to a pediatric ICU from 1/2012 to 3/2014. We enrolled patients upon initiation/expansion of antibiotics (time 0) and measured a panel of 8 serum biomarkers daily over 72h. Patients with immune compromise, neutropenia, or without blood cultures were excluded. Bacterial and viral testing were performed at the discretion of the clinical team.  Microbiology, imaging, and clinical data were reviewed for 48h after enrollment and CDC definitions were used to classify infections as proven/probable bacterial, laboratory-confirmed viral, or no infection; Kruskal-Wallis tests compared biomarkers among the 3 groups. Excess antibiotics (abx) were calculated as use >48h without bacterial infection. We tested various cut points of C-reactive protein (CRP) and procalcitonin (PCT) combinations to maximize the negative predictive value (NPV) for bacterial infection.

Results: Of 87 patients, 24 (28%) had bacterial (11 pneumonia, 4 BSI, 3 sinusitis, and 6 other), 24 had viral, and 39 had no infection; 38 were female and median age was 4.2y (IQR 1.0-12.4). At 0 and 24h, tissue plasminogen activator (TPA), PCT and CRP were higher in children with bacterial v. viral or no infection (p<.05, Figure). Serum amyloid A was higher in patients with bacterial v. viral or no infections at 0h (p=.01, Figure). Patients without bacterial infection received 153 days of excess abx (55% of all abx, mean: 2.4 excess days/patient). Of all biomarker and time point combinations, CRP <4 mg/dL plus PCT <1 ng/mL at time 0 had a NPV of 0.89 (95% CI 0.77-1.01) for identifying patients with bacterial infections.  All patients misclassified as “low-risk” had identifiable bacterial infections on standard diagnostic evaluation at 48h.

Conclusion: Early measurement of CRP and PCT can identify critically ill children with SIRS at low risk of bacterial infection and, in combination with standard diagnostic tests, may augment a clinically useful algorithm to decrease unnecessary antibiotic use.

Kevin Downes, MD1, Scott Weiss, MD, MSCE2, Sarah B. Klieger, MPH1, Julie Fitzgerald, MD, PhD2, Fran Balamuth, MD, PhD, MSCE3, Sherri Kubis, RN2, Pam Tolomeo, MPH4, Warren Bilker, PhD4, Xiaoyan Han, MS4, Irving Nachamkin, DrPH, MPH, FIDSA5, Charles Garrigan, MB5, Jennifer Han, MD, MSCE6, Ebbing Lautenbach, MD, MPH, MSCE, FIDSA, FSHEA6, Jeffrey S. Gerber, MD, PhD1, Susan Coffin, MD, MPH, FPIDS1 and The CDC Prevention Epicenters Program, (1)Department of Pediatrics, Division of Infectious Diseases, The Children's Hospital of Philadelphia, Philadelphia, PA, (2)Anesthesiology and Critical Care Medicine, The Children's Hospital of Philadelphia, Philadelphia, PA, (3)Division of Emergency Medicine, The Children's Hospital of Philadelphia, Philadelphia, PA, (4)Center for Clinical Epidemiology and Biostatistics, University of Pennsylvania, Perelman School of Medicine, Philadelphia, PA, (5)Department of Pathology and Laboratory Medicine, University of Pennsylvania, Perelman School of Medicine, Philadelphia, PA, (6)Division of Infectious Diseases, University of Pennsylvania, Perelman School of Medicine, Philadelphia, PA

Disclosures:

K. Downes, None

S. Weiss, None

S. B. Klieger, None

J. Fitzgerald, None

F. Balamuth, None

S. Kubis, None

P. Tolomeo, None

W. Bilker, None

X. Han, None

I. Nachamkin, None

C. Garrigan, None

J. Han, None

E. Lautenbach, None

J. S. Gerber, None

S. Coffin, None

Findings in the abstracts are embargoed until 12:01 a.m. PDT, Wednesday Oct. 7th with the exception of research findings presented at the IDWeek press conferences.