We admitted 36 preterm neonates (600 to 1250 gm birth weight) with normal 6-hour echoencephalograms to a randomized, placebo-controlled prospective trial to determine whether a low dose of indomethacin would prevent germinal matrix or intraventricular hemorrhage and permit adequate urinary output. Between the sixth and tenth postnatal hours, indomethacin (0.1 mg/kg) or placebo was administered intravenously every 24 hours for a total of three 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, serum indomethacin levels, and renal and clotting functions were monitored. No differences in birth weight, gestational age, or Apgar scores were noted between the two groups of infants. Two indomethacin-treated infants and three infants given placebo had significant urinary output difficulties, requiring that the study medication be withheld. Of 19 infants given indomethacin, two had germinal matrix or intraventricular hemorrhage, in comparison with 8 of 17 infants given saline solution (p = 0.02). Of the infants who had a left-to-right patent ductus arteriosus shunt before treatment, 64% of the indomethacin-treated and 33% of the saline solution-treated infants no longer had a patent ductus arteriosus on day 5. Ductal status appeared unrelated to the development of germinal matrix or intraventricular hemorrhage.