Antibiotic Choice for Community-Acquired Pneumonia in Children With Medical Complexity

BACKGROUND There are no guideline recommendations for antibiotic choice for community-acquired pneumonia (CAP) in children with medical complexity (CMC). We compared broad- vs narrow-spectrum antibiotics in CMC hospitalized with nonsevere CAP. METHODS This cohort study was conducted using target trial emulation within an academic health system. We included hospitalizations from January 2016 to April 2024 with International Classification of Diseases, 10th Revision (ICD-10) codes for CAP and complex chronic conditions among children aged 3 months to 21 years who received 3 or more days of antibiotics. We excluded children with severe/complicated CAP, chronic pulmonary conditions, sickle cell disease, tracheostomies, immunosuppression, other bacterial infections, or recent CAP admission. The exposure was initial broad-spectrum antibiotics (CAP antibiotics other than aminopenicillins) within 12 hours of presentation. The primary outcome was time to discharge, censored at day 10. Secondary outcomes were progression to severe CAP and 30-day CAP-related revisits. To mitigate confounding, we applied inverse probability of treatment weighting in outcome models, including a Royston-Parmar flexible survival model for time to discharge, and logistic regression for binary outcomes. RESULTS A total of 847 encounters (749 patients) were included. After weighting and trimming, 742 encounters remained with adequate covariate balance. Compared with the narrow-spectrum group, the broad-spectrum group demonstrated no significant differences in time to discharge (hazard ratio, 0.84; 95% CI, 0.68–1.04), progression to severe CAP (odds ratio [OR], 1.29; 95% CI, 0.7–2.4), or 30-day CAP-related revisits (OR, 0.39; 95% CI, 0.1–1.2). CONCLUSIONS In CMC hospitalized with nonsevere CAP, initial broad-spectrum antibiotics did not improve clinical outcomes compared with narrow-spectrum antibiotics, supporting the use of aminopenicillins for nonsevere CAP in many CMC.