We conducted a randomized trial comparing bed rest with tocolysis to determine the therapeutic efficacy, safety, and cost-effectiveness of tocolysis for the treatment of preterm labor after membrane rupture. One hundred nine women participated over a 26-month interval. Treatment groups did not differ significantly in terms of gestational age at membrane rupture, gestational age at delivery, birth weight, maternal or fetal infectious morbidity, respiratory distress syndrome, necrotizing enterocolitis, or perinatal mortality. Prolongation of intrauterine time after the onset of uterine contractions was seen in women receiving tocolysis (105.2 +/- 157 hours versus 62.1 +/- 77 hours, p = 0.06). This prolongation was not associated with a significant reduction in the total cost per surviving infant (tocolysis, $38,593 +/- $40,887 versus bed rest, $43,158 +/- $37,116; p = 0.445). The cost difference was artifactual. The number of very premature infants born (less than 28 weeks' gestation) was unequal in the two groups (12 in the bed rest group and 5 in the tocolysis group) and skewed the results. Before 28 weeks' gestation tocolysis was associated with a significant increase in intrauterine time after the onset of regular contractions (p = 0.05). However, there was no identifiable perinatal benefit garnered from the additional 5 days. After 28 weeks there were no significant differences between treatment groups in terms of intrauterine time after the onset of regular contractions and total cost per surviving infant. Because tocolysis does not improve perinatal outcome and can itself be associated with major maternal morbidity, it should be avoided after 28 weeks' gestation. Before 28 weeks' gestation tocolysis may greatly increase intrauterine time, but the benefit of this prolongation is not clear.