One hundred six patients in preterm labor with intact amniotic membranes and gestational age less than or equal to 32 weeks were randomized to receive either ritodrine hydrochloride or a 48-hour course of indomethacin for tocolysis. The relative efficacy, maternal and neonatal safety, and costs were evaluated to determine which may be the more appropriate first-line pharmacologic agent used to manage preterm labor. Fifty-four patients and 52 patients were randomized to receive ritodrine hydrochloride or indomethacin, respectively. Ritodrine hydrochloride and indomethacin were equally effective in inhibiting uterine contractions and delaying delivery. Delivery was delayed for at least 48 hours in 83 and 94%, and for at least 7 days in 70 and 75% of patients receiving ritodrine or indomethacin, respectively. Tocolysis with indomethacin was associated with no maternal side effects, whereas tocolysis with ritodrine hydrochloride was associated with a 24% incidence of serious cardiovascular and metabolic adverse effects prompting discontinuation of the drug. There were no differences in outcome between the infants exposed to indomethacin versus ritodrine hydrochloride when delivered either remote from therapy or during therapy, except for a statistically higher serum glucose in the infants exposed to ritodrine hydrochloride when delivered during tocolytic therapy. There were no cases of premature closure of the ductus arteriosus or pulmonary hypertension. Tocolysis with indomethacin was 17 times less costly than tocolysis with ritodrine hydrochloride. For gestations less than or equal to 32 weeks complicated by preterm labor, indomethacin may be an appropriate alternative as a first-line tocolytic agent.