A randomized prospective trial was performed to compare the efficacy and safety of ritodrine and indomethacin in the long-term treatment of preterm labor. Forty patients with intact membranes in preterm labor at 23 to 34 weeks' gestation were randomized to receive either intravenous ritodrine or oral indomethacin as the first-line tocolytic agent. Successful intravenous ritodrine therapy was followed by oral terbutaline therapy, and indomethacin-treated patients continued to receive oral indomethacin. Treatment failures were defined as progressive preterm labor or patient intolerance, and these patients were treated with intravenous magnesium sulfate. Ritodrine and indomethacin were equally successful in delaying preterm birth as defined by interval to delivery, gestational age at delivery, delivery delayed greater than 7 days, attainment of 35 weeks of gestation, percentage of patients who required magnesium sulfate therapy, percentage of patients who were readmitted with premature rupture of membranes, absence of recurrent preterm labor, and infant birth weight. More than 80% of mothers who received ritodrine voiced complaints of beta-sympathomimetic side effects, and one patient discontinued treatment as the result of intolerance. There were minimal patient complaints with indomethacin use. No statistically significant differences were noted in neonatal outcome as defined by Apgar scores, umbilical cord pH, intensive care days, ventilator days, or neonatal deaths. However, three cases of primary pulmonary hypertension were observed in the indomethacin group. We had not previously observed this problem with short-term (24 to 48 hours) indomethacin therapy.