The incidence and severity of retinopathy of prematurity (ROP) as affected by vitamin E prophylaxis at pharmacologic serum levels (5 mg/dl) were evaluated in a double-masked clinical trial of infants with a birth weight less than or equal to 2000 gm or a gestational age less than or equal to 36 weeks. The infants were enrolled by age 5 days and randomly assigned to receive parenterally administered, and later orally administered, free alpha-tocopherol (vitamin E) or its placebo. Study medication was continued until retinal vascularization was complete or active ROP had subsided, except in infants with a diagnosis of severe disease, in whom vitamin E was substituted for study medication. Acute ROP data were collected on 755 infants. Logistic regression analysis, with control for immaturity, oxygen exposure, and other illness risk factors, showed a decrease in incidence of ROP in vitamin E-treated infants (p = 0.003, all infants; p = 0.035, infants weighing less than or equal to 1500 gm at birth). Among the 424 infants weighing less than or equal to 1500 gm at birth, the age at enrollment influenced treatment effect (age day 0 to 1, p = 0.006 (n = 288) vs age day 2 to 5, p greater than 0.1 (n = 136]. Overall, 77.6% of infants with ROP had mild disease. Moderate to severe ROP was confined to infants weighing greater than or equal to 1500 gm at birth (25 given placebo, 25 given vitamin E), with progression to severe disease in nine placebo-treated versus three vitamin E-treated infants (p = 0.048). The incidence of severe ROP per se was not significantly decreased (all birth weights, p = 0.086; less than or equal to 1500 gm birth weight, p = 0.080); the sample size was too small, however, to assess this end point adequately. An increased incidence of sepsis and late-onset necrotizing enterocolitis was found among vitamin E-treated infants weighing less than or equal to 1500 gm at birth who received study medication for greater than or equal to 8 days (p = 0.006). Because most ROP is mild in degree and regresses completely, the risk/benefit ratio of pharmacologic prophylaxis for ROP is unfavorable. Treatment of moderate and severe ROP with vitamin E above physiologic serum levels (greater than 3 mg/dl) appears promising and should be further investigated. The interpretation of cicatricial outcome was confounded by the small number of patients involved and by subsequent treatment of severe ROP in placebo-treated infants with vitamin E.