Abstract
IMPORTANCE: Systemic postnatal corticosteroids have been shown to reduce rates of bronchopulmonary dysplasia (BPD) in infants born preterm, but both corticosteroids and BPD are associated with cerebral palsy. OBJECTIVE: To describe how the association between systemic postnatal corticosteroids and survival free of cerebral palsy varies with the risk of BPD in infants born preterm, and if the association differs between dexamethasone and hydrocortisone, or with age at starting treatment. DESIGN, SETTING, AND PARTICIPANTS: This comparative effectiveness research used weighted meta-regression analysis of eligible randomized clinical trials (RCTs) of systemic postnatal corticosteroids reported from June 1989 through March 2022 that included rates of all of BPD, mortality, and cerebral palsy in neonatal intensive care units in 10 countries. Infants born preterm at risk of BPD were included. Data were analyzed from April and July 2024. INTERVENTIONS: Systemic dexamethasone or hydrocortisone. MAIN OUTCOMES AND MEASURES: Type and timing of corticosteroid, control group rate of BPD, and risk difference in survival free of cerebral palsy between corticosteroid and control arms. RESULTS: Twenty-six RCTs with data on 3700 randomized infants were eligible; 18 (69%) investigated dexamethasone and 8 (31%) hydrocortisone; 12 (46%) started treatment in the first week after birth. There was evidence for a differential association of the type of corticosteroid with the effect of systemic dexamethasone on survival free of cerebral palsy and the risk of BPD in control groups (interaction coefficient, 0.54; 95% CI, 0.25-0.82; P = .001). For dexamethasone, for every 10-percentage point increase in the risk of BPD, the risk difference for survival free of cerebral palsy increased by 3.74% (95% CI, 1.54 to 5.93; P = .002). Dexamethasone was associated with improved survival free of cerebral palsy at a risk of BPD greater than 70%. Conversely, dexamethasone was associated with harm at a risk of BPD less than 30%. There was some evidence for a negative association with hydrocortisone, with possible benefit with risk of BPD less than 30%. There was no strong evidence for a differential effect of timing among those treated with dexamethasone (interaction coefficient, 0.13; 95% CI, -0.04 to 0.30; P = .14). CONCLUSIONS AND RELEVANCE: The findings suggest that dexamethasone (compared with control) was associated with improved rates of survival free of cerebral palsy in infants at high risk of BPD but should be avoided in those at low risk. A role for hydrocortisone is uncertain.