Hospital-acquired infections in preterm infants with gestational age <32 weeks: a retrospective study of clinical characteristics, pathogen distribution, and associated factors.

Preterm infants born at <32 weeks' gestation are often highly susceptible to nosocomial infection owing to immune immaturity, prolonged hospitalisation, and frequent invasive procedures. Data on the epidemiology of infection and modifiable risk factors specific to this population in Chinese neonatal intensive care units (NICUs) remain limited. We conducted a retrospective cohort study of 170 preterm infants (<32 weeks' gestation) admitted to the NICU of the People's Hospital of Guangxi Zhuang Autonomous Region between January 2022 and December 2024. The incidence of nosocomial infections, site distribution, and pathogen profiles were characterised. Risk and protective factors were evaluated using univariate and multivariate binary logistic regression. Model discrimination was assessed by receiver operating characteristic (ROC) curve analysis. The incidence of bronchopulmonary dysplasia (BPD), retinopathy of prematurity (ROP), preterm brain injury, and in-hospital mortality was compared between infected and non-infected infants. Fifty-three infants (31.18%) developed nosocomial infections, yielding 61 infection episodes (episode rate 35.88%). Respiratory tract infections predominated (pneumonia 13.53%; ventilator-associated pneumonia 1.18%). Gram-negative bacteria accounted for 68.0% of isolates, with Klebsiella pneumoniae the most frequent species. In the multivariate model, peripherally inserted central catheter (PICC) catheterisation (adjusted OR = 3.172, 95% CI: 1.330-7.564) and 25-hydroxyvitamin D deficiency (adjusted OR = 2.867, 95% CI: 1.232-6.670) were independent risk factors, while delayed cord clamping (DCC) was independently protective (adjusted OR = 0.265, 95% CI: 0.113-0.619). The model AUC was 0.764 (95% CI: 0.693-0.836). BPD incidence was higher in the infection group than in the non-infection group (47.17% vs. 29.91%, P = 0.029); no between-group differences were observed for ROP, brain injury, or mortality. Our analysis shows that nosocomial infection affected approximately one-third of preterm infants <32 weeks' gestation in this cohort. Three modifiable factors, i.e., PICC catheterisation, early vitamin D status, and DCC, were independently associated with infection risk. These findings support standardisation of PICC care bundles, early vitamin D supplementation, and routine DCC implementation as possible targets for infection prevention in this population.