In a randomized prospective trial, we studied the effect of early high-dose phenobarbital treatment on the early (intraventricular hemorrhage) and late (neurodevelopmental abnormalities) manifestations of hypoxic-ischemic encephalopathy in preterm infants weighing 1500 g or less at birth. The first intravenous dose of 15 mg/kg was given at a mean age of 110 minutes, followed by 15 mg/kg after 4 hours and then by 5 mg/kg at 24-hour intervals for 5 days. The overall incidence of intraventricular hemorrhage was 32% in treated and 46% in control infants, a nonsignificant difference. An ultrasound brain scan at 9 months old revealed no significant difference in the incidence of ventricular dilatation between treated (19%) and control (29%) infants. At 27 months, a similar incidence of mild (10%) and severe (10%) neurodevelopmental handicaps was found in both treated and control groups. Since beneficial effects could not be documented by any of the criteria used, we conclude that routine administration of phenobarbital to low birth weight infants is not justified.