Shorter or biomarker-guided antibiotic durations for common serious neonatal infections: a collection of non-inferiority meta-analyses.

Background: Unnecessary antibiotic prolongation promotes antimicrobial resistance. Evidence-based guidance on antibiotic durations for treating neonatal infections is lacking. Noninferiority meta-analyses evaluating shorter durations or biomarker-guided durations have not been conducted. Methods: We conducted systematic reviews and meta-analyses using a unified search strategy (MEDLINE, EMBASE, Cochrane Library, January 1990-December 2024) to determine whether short-course antibiotics are noninferior to "standard" courses for culture-positive sepsis, culture-negative sepsis, UTIs, uncomplicated/complicated meningitis, and fungal sepsis. Likewise, we determined whether biomarker-guided antibiotic durations are noninferior to "standard" courses for any sepsis. We included randomized controlled trials (RCTs) without language restriction comparing short-courses or biomarker-guided courses versus standard courses for treating the above neonatal infections. Minimum clinically important differences (MCIDs) for critical outcomes were pre-specified [mortality (28-day, in-hospital or 12-month corrected age) = +1%, culture-positive relapse = +3%, culture-negative relapse = +5%]. We evaluated risk of bias (Cochrane RoB 2.0) and certainty of evidence (CoE) using GRADE. Registration: PROSPERO CRD42023311895. Findings: We reviewed 146 full-text articles and included 26 in the review. For culture-positive sepsis (7 randomized controlled trials [RCTs]), 7-10-day courses were noninferior to longer courses for 28-day mortality [upper bound of 95% CI for risk difference +0.85%, below 1% MCID (low CoE)], in-hospital mortality [+0.51% (moderate CoE)], culture-positive relapse [+0.75%, below 3% MCID (low CoE)], culture-negative relapse [+4.7%, below 5% MCID (very low CoE)]. For culture-negative and -positive sepsis, biomarker-guided courses (5 RCTs) were noninferior to standard courses for mortality and relapses (all very low CoE). For culture-negative sepsis (6 RCTs), shorter courses (3-4 days) were noninferior for culture-positive relapses (very low CoE) but Interpretation: Compared to standard durations, 7-10-day antibiotic courses and biomarker-guided courses may be noninferior for treating culture-positive neonatal sepsis and any sepsis, respectively. Three-to-four-day courses may Funding: Not funded.