Epidemiology of Bloodstream Infections in Hospitalized Children in the United States, 2009-2016

Abstract

Background: Bloodstream infections (BSIs) cause significant morbidity and mortality in children. Recent pediatric epidemiological data may inform prevention strategies and empiric antimicrobial therapy selection.

Methods: We conducted a retrospective cohort study from 2009 through 2016 utilizing demographic and microbiologic data on inpatients aged <19 years using the Premier Healthcare Database. BSIs were positive blood cultures without known contaminants. Hospitalization rate was the number of BSI-positive encounters per 1000 admissions. Community-acquired infections (CAIs) were cultures positive ≤2 days of admission among nonneonates. BSI patients were compared to documented positive BSI patients (non-BSI); differences were analyzed using χ2 test, t test, and Cochran-Armitage test for time trends.

Results: Among 1 809 751 encounters from 162 US hospitals, 5340 (0.30%) were BSI positive; CAIs were most common (50%). BSI patients were more often aged 1-5 years and had complex chronic conditions or central lines compared to non-BSI patients. The BSI hospitalization rate declined nonsignificantly over time (3.13 in 2009 to 2.98 in 2016, P = .08). Among pathogens, Escherichia coli (0.80 to 1.26), methicillin-sensitive Staphylococcus aureus (0.83 to 1.98), and group A Streptococcus (0.16 to 0.37) significantly increased for nonneonates, while Streptococcus pneumoniae (1.07 to 0.26) and Enterococcus spp. (0.60 to 0.17) declined. Regional differences were greatest for E. coli and highest in the New England and South Atlantic regions.

Conclusions: Trends in pediatric BSI hospitalization rates varied by pathogen and regionally. Overall the BSI hospitalization rate did not significantly decline, indicating a continued need to improve pediatric BSI assessment and prevention.

Keywords: epidemiology; pathogen distribution; pediatric; bloodstream infection.