In order to assess the efficacy of a combination of systemic and nebulized corticosteroids in reducing the incidence and severity of chronic lung disease (CLD) in very low birthweight (VLBW) infants, 60 ventilator-dependent infants < or = 1500 g were randomly assigned to receive either steroids or placebo as of 7 d. The steroid group (n = 30, GA = 25.8 +/- 1.6 weeks, BW = 731 +/- 147 g) received systemic dexamethasone for 3 d, followed by nebulized budesonide for 18 d. Control infants (n = 30, GA = 25.9 +/- 1.8 weeks, BW = 796 +/- 199 g) received systemic and inhaled saline. Steroid-treated infants required less ventilatory support between 9 and 17 d (p < 0.01), and had greater lung compliance at 10 d (p = 0.01), but not subsequently. CLD incidence at 36 weeks was 45.5% vs 56.0% in controls, and fewer steroid-treated infants required dexamethasone rescue (23.3% vs 56.7%, p = 0.017). Survival to discharge was similar (73.3% vs 83.3%), as were the durations of mechanical ventilation, supplemental oxygen use, and hospitalization. Tracheal effluent elastase/albumin ratios and serum cortisol values did not differ between groups, and no adverse effects were noted. We conclude that early dexamethasone administration was associated with improved pulmonary function, which was not sustained with nebulized budesonide. However, the steroid regimen studied reduced the need for dexamethasone rescue in infants with CLD.