A randomized trial comparing outcome of babies treated with a natural surfactant (Curosurf) for moderately severe respiratory distress syndrome (RDS) with corresponding data from babies treated at a more advanced stage of the disease is reported. A total of 182 newborn babies (mean gestational age 29.8 weeks) requiring mechanical ventilation and a fraction of inspired oxygen (FiO2) in the range of 0.40-0.59 for RDS were randomized to immediate ("early") treatment (No = 86) with surfactant (200 mg/kg), or to a control group (No = 96). According to the protocol 49 controls (51%) qualified for a "late" surfactant treatment at an FiO2 requirement of > or = 0.60. In both groups of treated patients administration of surfactant led to a rapid improvement of oxygenation, but the peak value for PaO2 and the variability of the response tended to be lower in babies given immediate treatment. In comparison with the total control group, babies treated immediately had lower incidence of grade III-IV intraventricular hemorrhage (7% vs 18%; p < 0.05), lower mortality (9% vs 23%; p < 0.05), and lower incidence of unfavourable outcome--defined as death or bronchopulmonary dysplasia--(18% vs 34%; p < 0.05) at 28 days. Also significant reductions of time in oxygen > 21% and time on mechanical ventilation were observed. Our data suggest that treatment with surfactant when RDS is moderately severe prevents or reverses the natural progression of the disease in at least 50% of the cases and lowers the risk of serious complications.