Seventy-seven very low birthweight (VLBW) infants (mean birthweight 891 +/- 209 g) with a diagnosis of bronchopulmonary dysplasia (BPD) were treated with a steroid (dexamethasone) in an attempt to wean them from mechanical ventilation. Seventeen of 77 (22%) treated infants died. Death from respiratory failure occurred in 13 infants; sepsis occurred in six infants (7.8%) and contributed to death in one. During steroid therapy systemic hypertension occurred in 18 surviving infants (30%), glucose intolerance occurred in 29 infants (38%), and marked irritability occurred in three infants (3.8%). Elevated blood pressure returned to normal and glucose intolerance resolved in all infants following discontinuation of therapy. Fifty infants were available for follow-up at a mean corrected age of 14.9 +/- 9.8 months. Twenty-two percent required rehospitalization in the first year of life for respiratory illnesses. Results of testing by Bayley Scales of Infant Development were normal in 60% of infants. Fifty percent were considered normal based on both developmental testing and physical examination. Twenty-eight percent had mild to moderate abnormalities, and 22% were severely handicapped. These follow-up results are statistically similar to those recorded in LBW infants with BPD not treated with steroids who were hospitalized during the same period. We conclude that the side effects of steroid therapy for BPD consist primarily of blood pressure elevation, glucose intolerance, and irritability. Causes of death are unchanged by steroids. The incidence of severe infection and the long-term neurologic outcome of high-risk infants with BPD are not appreciably compromised by this therapy. These data suggest that concern for steroid side effects should not prevent additional prospective investigation to determine the role of steroid therapy in the overall management of BPD.