# Quality improvement interventions to prevent late-onset sepsis in premature infants: a systematic review and meta-analysis

## TL;DR

Quality improvement bundles significantly reduce late-onset sepsis in preterm infants, especially those with very low or extremely low birth weight.

## Contribution

This study provides the first meta-analysis showing the effectiveness of QI bundles in reducing late-onset sepsis in preterm infants.

## Key findings

- QI bundles reduced late-onset sepsis in very low birth weight infants by 53%.
- Core components like hand hygiene and central line management were consistently included in effective bundles.
- Late-onset sepsis reduction varied based on how sepsis was defined in studies.

## Abstract

Late-onset sepsis (LOS) is a life-threatening complication in preterm infants, with reported incidence rates of 1%–30% that vary by clinical and geographical factors. Quality improvement (QI) bundles integrating infection control, nutrition, and device management show promise in reducing LOS, but evidence remains fragmented due to heterogeneous definitions and mixed study populations.

A systematic review and meta-analysis of QI studies was conducted across PubMed, Embase, Cochrane Library, and Web of Science (inception to March 19th, 2025). Studies were included if they reported pre-post QI outcomes for LOS in preterm infants (gestational age < 37 weeks), with effect sizes synthesized as odds ratios (ORs) and 95% confidence intervals (CIs). Heterogeneity was evaluated using I2, with random-effects models for I2 ≥ 50%. Subgroup analyses explored LOS definition impacts (time windows: ≥48 h, ≥72 h, other) and meta-regression tested covariates (study year, sample size, quality).

Of 9,705 identified studies, 29 (21 for meta-analysis, n = 29,120 infants) met criteria. QI bundles significantly reduced LOS in very low birth weight (VLBW) infants (pooled OR = 0.47, 95% CI [0.38–0.58], I2 = 88.7%) and extremely low birth weight (ELBW) infants (OR = 0.49, 95% CI [0.29–0.83], I2 = 80.6%). Core components included multidisciplinary teams (25/29), hand hygiene (21/29), and central line management (22/29). Subgroup analysis showed varying effects by LOS definition: OR = 0.35 (95% CI [0.19–0.64]) for ≥ 48 h, OR = 0.50 (95% CI [0.39–0.64]) for ≥ 72 h, and OR = 0.71 (95% CI [0.61–0.82]) for ill-defined thresholds. Meta-regression identified no significant modifiers (all p > 0.0), but publication bias was detected in VLBW analyses (Egger’s test, p < 0.0).

QI bundles significantly reduce LOS in preterm infants, including VLBW and ELBW subgroups, through core components like multidisciplinary teams, hand hygiene, and central line management. Given outcome variations by diagnostic criteria, standardizing LOS definitions is crucial. These bundles should be integrated into routine care globally. Future efforts should prioritize standardized reporting, antibiotic stewardship, and equitable implementation, especially in low-resource settings, by building on existing evidence and broader healthcare principles.