A randomized, placebo-controlled trial of human surfactant given intratracheally at birth (prophylactic) versus rescue administration after the onset of severe respiratory distress syndrome (RDS) was conducted among preterm infants born at 24 to 29 weeks of gestation. Singleton fetuses were randomly assigned to receive (1) placebo (air), (2) prophylactic surfactant treatment, or (3) rescue surfactant treatment; infants of multiple births received either (1) prophylactic or (2) rescue treatment. Of 282 potentially eligible fetuses, 246 infants received treatments at birth and 200 infants had RDS. Outcomes are presented both as an intention-to-treat analysis (including infants who met exclusion criteria at or after birth) and as a full treatment protocol analysis for those infants with RDS and likely to benefit from surfactant. Preterm infants (mean 1.0 kg birth weight, 27 to 28 weeks of gestational age) randomly assigned to receive prophylactic treatment received surfactant soon after birth; those assigned to receive rescue surfactant had instillation at a mean age of 220 minutes if the lecithin-sphingomyelin ratio was less than or equal to 2.0 and no phosphatidylglycerol was detected in either amniotic fluid or initial airway aspirate, oxygen requirements were a fraction of inspired oxygen of greater than 0.5, and mean airway pressure was greater than or equal to 7 cm H2O from 2 to 12 hours after birth. Up to four treatment doses (or air) were permitted within 48 hours; approximately 60% of surfactant-treated infants required two or more doses. Surfactant-treated infants had significantly less pulmonary interstitial emphysema than placebo-treated infants (p = 0.02), but there were no other significant differences in mortality rates or morbidity. Indexes of oxygenation and ventilation were improved in surfactant recipients during the first 24 hours. An intention-to-treat analysis found no significant differences between infants given placebo and surfactant-treated infants or between prophylactic- and rescue-treated infants; an improved total mortality rate (p = 0.002) was found among surfactant-treated infants in Helsinki but not in San Diego. Among infants with RDS, the total mortality rate was significantly improved (p = 0.004) with surfactant treatment but not the proportion alive and without bronchopulmonary dysplasia at 28 days (p = 0.052), or the proportion alive and without bronchopulmonary dysplasia at 38 weeks of postconceptional age (p = 0.18) to adjust for differences in prematurity. Deaths caused by RDS or bronchopulmonary dysplasia were significantly reduced among surfactant recipients (p = 0.0001). Neither among singletons nor among multiple-birth infants was there a selective advantage to prophylactic versus rescue treatment.(ABSTRACT TRUNCATED AT 400 WORDS)