Abstract Background Timely active empirical therapy (AET) may improve outcomes following bloodstream infection. In high-income settings, low pediatric mortality following gram-negative bloodstream infection (GNBSI) limits its value as a sole endpoint. Composite measures that incorporate multiple clinically relevant outcomes may detect differences more readily. We assessed the effect of AET on clinical outcomes in pediatric GNBSI. Methods We analyzed data from a prospective surveillance study of hospitalized Australian children with GNBSI. A ranked composite endpoint included 30-day all-cause mortality (30D ACM), time to death, intensive care unit (ICU) admission, relapse, and hospitalization duration. Baseline differences were adjusted using inverse probability of treatment weighting. Generalized pairwise comparison (GPC) generated win statistics. Analyses were stratified by comorbidity, acquisition setting, and organism. Results AET occurred in 597 episodes and inactive empirical therapy (IET) in 62. Median treatment duration was similar (AET 10 vs IET 11 days, p = 0.44). GPC analysis showed an adjusted win ratio of 1.04 (95% CI 0.72–1.52, p = 0.83), indicating no significant difference. Contributions from each composite component (AET vs IET) were: 30D ACM (2.2% vs 1.3%), time to death (0.02% vs 0.01%), ICU admission (7.5% vs 9.1%), relapse (2.6% vs 1.4%) and hospitalization duration (37.3% vs 36.0%). Stratified estimates were similar for comorbidity and acquisition. Conclusion In this large multicenter pediatric GNBSI cohort, AET was not associated with improved outcomes. In high-resource settings, the impact of initial antibiotic activity may be smaller than traditionally assumed, though timely effective therapy remains essential for severe illness and low-resource settings.
Wen et al. (Tue,) studied this question.