Efficacy of short-course antibiotics for culture-positive neonatal sepsis: A systematic review and meta-analysis.

Sepsis is a common cause of neonatal mortality and morbidity. Though antibiotics are the mainstay of treatment in culture-positive neonatal sepsis, the dilemma persists for the optimum duration of antimicrobial therapy. The present study aimed to evaluate the efficacy of short-course antibiotics for uncomplicated culture-positive neonatal sepsis. This systematic review and meta-analysis (PROSPERO: CRD42023444899) identified, appraised, and synthesised the available evidence from randomised and quasi-randomised controlled trials related to the efficacy of short-course (7-10 days) versus standard-course (14 days) antibiotics for uncomplicated culture-positive neonatal sepsis on the rate of treatment failure, mortality, duration of hospitalisation, morbidities including antibiotics-related adverse events, long-term neurodevelopmental outcomes and cost analysis. Data were pooled using RevMan 5.4 software. Certainty of evidence (COE) for predefined outcomes was analysed by GRADE. Available evidence showed no significant difference in the rate of treatment failure between 7- to 10-day versus 14-day antibiotics courses [risk ratio (95% confidence interval, CI), 2.45 (0.93-6.47), I = 0%, six studies, n = 573, very low COE]. No incidence of death was reported in either treatment arm in the two included studies. Duration of hospitalisation was significantly shorter with the short-course antibiotics arm compared to standard-course [mean difference (95% CI), -3.88 (-4.22 to -3.54) days, I = 0%, five studies, n = 507, low COE]. Morbidities reported in the three studies were similar. Other outcomes were not reported. To conclude the evidence is very uncertain about the effect of short-course antibiotic regimen, compared to a standard-course, on the treatment failure rate in uncomplicated culture-positive neonatal sepsis. Adequately powered trials with outcomes including death and long-term neurodevelopmental impairment are needed.