# The association of early antibiotic exposure with subsequent development of late-onset sepsis in preterm infants: a systematic review and meta-analysis studies

## TL;DR

This study finds that prolonged early antibiotic use in preterm infants does not significantly increase the risk of late-onset sepsis.

## Contribution

The study provides a meta-analysis clarifying the relationship between early antibiotic exposure and late-onset sepsis in preterm infants.

## Key findings

- Prolonged antibiotic exposure was not significantly associated with late-onset sepsis (aOR = 1.2, 95% CI 0.99–1.46).
- Shorter antibiotic exposure might be associated with a slightly lower risk of late-onset sepsis (aOR = 0.87, 95% CI 0.77–0.98).

## Abstract

Early antibiotic exposure in preterm infants may disrupt gut microbiome development, affecting health. However, its link to late-onset sepsis (LOS) remains unclear. This meta-analysis aims to clarify the association while addressing confounding bias.

This systematic review and meta-analysis, conducted per PRISMA guidelines, utilized PubMed, Scopus, Google Scholar, and Web of Science for comprehensive literature retrieval. Studies comparing preterm infants with sterile blood cultures who received early antibiotics (short or prolonged) to those without, using LOS as the primary outcome, were included. Comparisons between short- and prolonged-course antibiotics were also considered. Only studies with adjusted analyses for confounders were considered. Adjusted odds ratios (aOR) were meta-analyzed, and the prediction interval (PI) was calculated using R software.

Ten studies met the eligibility criteria, comprising a total sample size of 55,089 preterm infants. Among these, nine studies included 33,549 preterm infants and compared prolonged antibiotic exposure to short exposure. Prolonged exposure was not significantly associated with LOS (pooled aOR = 1.2, 95% CI 0.99–1.46, P = 0.066, PI = 0.66 to 2.19, I² = 67%). Limiting the analysis to five studies with sample sizes over 1,000 reduced heterogeneity (I² = 30%) and provided a more precise confidence interval (pooled aOR = 1.03, 95% CI 0.91–1.15). Four studies, involving 41,938 preterm infants, examined preterm infants exposed to prolonged antibiotics versus those not exposed and found no significant association (aOR = 0.91, 95% CI 0.82–1.02, P = 0.1, PI = 0.72 to 1.16, I² = 0). All four studies had sample sizes exceeding 1,000. Additionally, these studies compared preterm infants with short antibiotic exposure to non-exposure, revealing a slightly lower risk of LOS (aOR = 0.87, 95% CI 0.77–0.98, P = 0.024, I² = 0) and a PI of 0.76 to 1.14.

Our findings indicate that prolonged early antibiotic exposure in preterm infants with sterile cultures does not significantly increase the risk of LOS compared to no antibiotic exposure. Interestingly, a shorter duration of antibiotic exposure might be associated with a slightly lower risk of LOS.

The online version contains supplementary material available at 10.1186/s12245-025-00869-5.