We admitted 48 preterm neonates (600 to 1250 gm birth weight, normal 6-hour echoencephalograms) to a randomized prospective indomethacin or placebo trial for the prevention of neonatal intraventricular hemorrhage. Beginning at 6 postnatal hours, indomethacin or placebo was administered intravenously every 12 hours for a total of five doses. Cardiac ultrasound studies to assess the status of the ductus arteriosus were performed at 6 postnatal hours and on day 5. Urinary output, serum electrolytes, and renal and clotting functions were monitored. No differences in birth weight, gestational age, Apgar scores, or ventilatory needs were noted between the two groups. Six infants given indomethacin had intraventricular hemorrhage, compared to 14 control infants (P = 0.02). The indomethacin-treated group had significant decreases in serum prostaglandin values 30 hours after the initiation of therapy. The overall incidence of patent ductus arteriosus was 82% at 6 postnatal hours; 84% of the indomethacin-treated infants experienced closure of the ductus, compared to 60% of the placebo-treated patients. Closure of the ductus was not related to incidence of intraventricular hemorrhage. We speculate that indomethacin may provide some protection against neonatal intraventricular hemorrhage by acting on the cerebral microvasculature.