Prevalence and risk factors for neonatal sepsis among very preterm infants in China: a systematic review and meta-analysis

Background Neonatal sepsis poses a significant risk to very preterm infants (VPIs, gestational age < 32 weeks) in China, with limited nationwide data on its incidence and risk factors. This study aimed to address this gap through a systematic review and meta-analysis to inform prevention strategies. Methods We searched PubMed, Embase, Web of Science, Scopus, CNKI, CBM, CSTJ, and WanFang up to March 31, 2025. Study quality was evaluated using the Newcastle-Ottawa Scale (NOS). Data extraction was performed with Microsoft Excel, and Stata 18.0 was used for meta-analysis. Heterogeneity was assessed using Cochran's Q and I 2 statistics, while publication bias was evaluated with funnel plots and Egger's test. Subgroup analyses were conducted to identify sources of heterogeneity. Results This systematic review encompassed 43 studies involving a total of 138,613 VPIs. The incidence of unclassified sepsis was determined to be 16.41% (95% CI: 11.80%–21.62%), while the incidence rates for early-onset sepsis (EOS) and late-onset sepsis (LOS) were 16.11% (95% CI: 9.52%–24.01%) and 15.10% (95% CI: 12.07%–18.40%), respectively. The study identified six significant risk factors for EOS among VPIs: lower gestational age [Hedges’ g = −0.82, 95% CI: (−1.12, −0.52)], lower birth weight [Hedges’ g = −0.46, 95% CI: (−0.64, −0.27)], chorioamnionitis (OR 3.10, 95% CI: 2.72–3.54), premature rupture of membranes (PROM) (OR 1.51, 95% CI: 1.12–2.04), administration of antenatal antibiotics (OR 1.38, 95% CI: 1.22–1.56), and the requirement for endotracheal intubation (OR 5.87, 95% CI: 3.84–8.97). Conclusions Sepsis imposes a substantial burden on Chinese VPIs, characterized by a highly heterogeneous incidence. The principal risk factors for EOS include lower gestational age, lower birth weight, chorioamnionitis, PROM, antenatal antibiotic administration, and endotracheal intubation. These findings furnish essential evidence for the development of risk prediction models and stratified management strategies for EOS in VPIs. Systematic trial registration The systematic review and meta-analysis were registered with PROSPERO (ID: CRD420251018993).