A randomised trial was conducted in 82 centres using the porcine surfactant extract, Curosurf, to compare two regimens of multiple doses to treat infants with respiratory distress syndrome and arterial to alveolar oxygen tension ratio < 0.22. Infants were randomly allocated to a low dosage group (100 mg/kg initially, with two further doses at 12 and 24 hours to a maximum cumulative total of 300 mg/kg; n = 1069) or a high dosage group (200 mg/kg initially with up to four further doses of 100 mg/kg to a maximum cumulative total of 600 mg/kg; n = 1099). There was no difference between those allocated low and high dosage in the rates of death or oxygen dependency at 28 days (51.1% v 50.8%; difference -0.3%, 95% confidence interval (CI) -4.6% to 3.9%), death before discharge (25.0% v 23.5%; difference -1.5%, 95% CI -5.1% to 2.2%), and death or oxygen dependency at the expected date of delivery (32.2% v 31.0%; difference -1.2%, 95% CI -5.2% to 2.7%). For 14 predefined secondary measures of clinical outcome there were no significant differences between the groups but the comparison of duration of supplemental oxygen > 40% did attain significance; 48.4% of babies in the low dose group needed > 40% oxygen after three days compared with 42.6% of those in the high dose group. The total amount of surfactant administered in the low dose regimen (mean 242 mg phospholipid/kg) was probably enough to replace the entire pulmonary surfactant pool. Adopting the low dose regimen would lead to considerable cost savings, with no clinically significant loss in efficacy.