Eight RCTs (n=880 infants, range 34 to 360) were included in the review. All studies were described as methodologically sound. Treatment allocation was concealed from the study investigators. Investigators were blinded to the intervention. There was blinding of outcome measures. There was complete reporting of in-hospital outcome measures and follow-up rates of at least 85% after discharge.
There was no significant difference between hydrocortisone and placebo for any of the specified outcome measures, (death, bronchopulmonary dysplasia, death or bronchopulmonary dysplasia, cerebral palsy, death or cerebral palsy). These results did not change when the analyses were restricted to studies with the primary aim of preventing bronchopulmonary dysplasia.
Analyses of early complications revealed that use of hydrocortisone was associated with a decreased risk of patent ductus arteriosus (RR 0.85, 95% CI 0,73 to 0.99; six RCTs, significant between-study heterogeneity) and with an increased risk of gastrointestinal perforation (RR 2.02, 95% CI 1.13 to 3.59; six RCTs). Results were similar when analyses were restricted to studies with the primary aim of preventing bronchopulmonary dysplasia.
There were no significant differences between hydrocortisone and placebo for the outcomes of discharged on oxygen and failure to extubate.