Reducing Empiric Vancomycin Use in the Pediatric Intensive Care Unit: a Multicenter Study

Abstract Background Vancomycin is frequently administered empirically for suspected serious bacterial infections (SBI) in the pediatric intensive care unit (PICU) despite low prevalence of methicillin-resistant Staphylococcus aureus (MRSA). Methods We performed a multicenter interrupted time series (ITS) study including five PICUs in three U.S. children’s hospitals evaluating the impact of a multifaceted stewardship intervention including consensus guidelines, education, and group level audit and feedback on reducing empiric vancomycin use. The primary outcomes were (1) vancomycin administration within 12 hours of suspected SBI onset and (2) empiric vancomycin days of therapy (DOT) during the first 3 days after suspected SBI onset. Secondary outcomes included overall vancomycin DOT/1000 patient days; organ dysfunction at days 3 and 7; 14-day mortality; and invasive MRSA infection not covered empirically. Empiric vancomycin use was modeled using mixed-effects logistic regression and vancomycin DOT was modeled using mixed effects Poisson regression. Results Among 4,549 episodes targeted by the intervention, there was an immediate reduction in empiric vancomycin use after the intervention (OR 0.71, 95% CI 0.54–0.95, P=0.02), with no significant change in post-intervention slope. Empiric vancomycin days after suspected SBI onset decreased by 24% immediately after the intervention (IRR 0.76, 95% CI 0.65–0.89, P<0.01), with no significant change in postintervention slope. There were no increases in mortality, organ dysfunction, or missed MRSA infections postintervention. Conclusion A multifaceted stewardship intervention reduced empiric vancomycin without evidence of harm, supporting the safety and feasibility of targeting empiric antibiotic prescribing in the PICU setting.