In a randomized, controlled study, human surfactant derived from amniotic fluid was administered within 12 hours of birth to infants with severe respiratory distress syndrome who were born at 24 to 32 weeks of gestation weighing less than or equal to 1500 gm. A second dose of surfactant was given to patients in the treatment group if they met ventilator requirements indicating relapse or lack of response to the initial dose. No significant improvement was observed in mortality rate (9/28 vs 15/31) or incidence of bronchopulmonary dysplasia (5/28 vs 3/31) when surfactant-treated infants were compared with control subjects, although there was a significant reduction in initial respirator and inspired oxygen requirements and the arterial/alveolar oxygen ratio improved. In addition, there was a significant reduction in pulmonary air leak in treated infants (10/28 vs 20/31; p less than 0.05). Retreatment was associated with an attenuated ventilatory response and with a higher mortality rate (7/14) than that of infants who did not require a second dose (2/14; p = 0.05), indicating a more severe form of disease. Multiple discriminant analysis, including eight independent variables, revealed that increasing birth weight, earlier age at surfactant treatment, and female gender were significantly associated with survival. These data suggest that early surfactant treatment may reduce mortality rates in very low birth weight infants with severe respiratory distress syndrome, as well as reduce ventilator requirements and the incidence of pulmonary air leaks.