Objectives To describe real-world patterns of postnatal corticosteroid use and evaluate associations with respiratory outcomes in preterm infants 32 weeks’ gestation. Methods This retrospective cohort study included infants 32 weeks admitted to a tertiary neonatal intensive care unit at King Abdulaziz Medical City, Riyadh, Saudi Arabia, between January 2017 and December 2025. Postnatal steroid exposure was categorized as hydrocortisone, dexamethasone, both, or none. Primary outcomes were bronchopulmonary dysplasia (BPD) severity at 36 weeks’ postmenstrual age, death before 36 weeks, and a composite of BPD or death. Multivariable regression models adjusted for baseline clinical factors, and propensity score weighting was used to address confounding by indication. Results Among 1,228 infants, 326 (26.5%) received postnatal steroids (hydrocortisone 9.5%, dexamethasone 15.2%, both 1.8%). Steroid use increased with decreasing gestational age (54.4% at 26 weeks vs. 7.0% at 30–31 weeks, p 0.001) and was more frequent in infants requiring invasive respiratory support at admission. In adjusted analyses, steroid exposure was associated with higher BPD severity adjusted odds ratio (aOR) 2.71, 95% CI 1.92–3.81, increased home oxygen (aOR 6.36, 95% CI 1.47–27.64), and longer duration of invasive ventilation incidence rate ratio (IRR) 2.65, 95% CI 2.25–3.11. Steroid exposure was associated with lower odds of death before 36 weeks (aOR 0.25, 95% CI 0.15–0.41), although this likely reflects treatment selection and survivor bias. Formulation-specific analyses showed higher BPD severity with combined exposure (aOR 5.67, 95% CI 2.40–13.39). Associations with the composite outcome were consistent across models and confirmed in propensity-weighted analyses (aOR 1.90, 95% CI 1.53–2.34). Conclusions Postnatal corticosteroids were primarily used in infants with severe respiratory disease and were associated with higher observed respiratory morbidity rates in this higher-risk population. These findings reflect real-world corticosteroid use in higher-risk infants and should be interpreted as associations within a selected population rather than causal treatment effects.
Ali et al. (Tue,) studied this question.