Hypothermia and Adverse Outcomes in Very Preterm Infants: A Systematic Review

CONTEXT Hypothermia after very preterm birth, typically defined as a temperature less than 36 °C, is variably linked to neonatal mortality and morbidities. OBJECTIVE To examine the association between admission hypothermia and adverse outcomes in very preterm infants with a gestational age (GA) of less than 32 weeks. DATA SOURCES CENTRAL, MEDLINE, and Embase from inception to February 18, 2024. STUDY SELECTION Observational or randomized designs reporting on the association between admission temperature and adverse outcomes in very preterm infants. DATA EXTRACTION Two reviewers screened abstracts and full texts, extracted the data, and assessed the risk of bias, following Meta-analysis Of Observational Studies in Epidemiology /Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. We planned to perform random-effects meta-analyses, subgroup (GA, birthweight [BW], and income), sensitivity analysis (NOS, study type), and meta-regression (GA, BW). Outcomes included mortality and neonatal morbidities: bronchopulmonary dysplasia (BPD), intraventricular hemorrhage (IVH), necrotizing enterocolitis (NEC), retinopathy (ROP), and sepsis. RESULTS This review included 32 studies with >300 000 infants. The mean hypothermia rate was 42% (range 14%–88%). Hypothermia was associated with increased mortality (crude odds ratio [cOR] [95% CI] 2.02[1.84;2.21]; adjusted OR 1.55[1.29;1.87]). Subgroup and sensitivity analyses upheld these results. Meta-regression analysis showed an inversed relationship between effect size and BW. Hypothermia was associated with higher risks of BPD (cOR 1.13[1.01;1.27]), IVH (cOR 1.37[1.17;1.61]), ROP (cOR 1.55[1.41;1.69]), and sepsis (cOR 1.32[1.16;1.51]). LIMITATIONS Only observational studies were included. CONCLUSIONS Hypothermia is associated with increased mortality and morbidity in very preterm infants. The strength of this association may be influenced by BW, definitions of hypothermia and outcomes, and exclusion criteria. Given the robustness of our results and our sample size, identical cohort studies might not provide different insights.