To evaluate the minimal effective dose regimen of the oxytocin antagonist atosiban in the treatment of acute preterm labor and the effect of a bolus on uterine activity within the first 2 hours compared with no bolus and the same infusion rate.A randomized, double-blind (except the ritodrine group), parallel group, multicenter study compared four different intravenous atosiban regimens (6.5 mg bolus plus 300 micrograms/minute, placebo bolus plus 300 micrograms/minute, 2mg bolus plus 100 micrograms/minute, and 0.6 mg plus 30 micrograms/minute) and intravenous ritodrine with respect to the cessation of uterine contractions for 1 hour or more during infusion, four or fewer contractions per hour in the last hour of therapy, and discontinuation because of adverse experiences. Three hundred two patients were enrolled.The lowest dose of atosiban (0.6 mg plus 30 micrograms/minute) was significantly less effective than ritodrine with respect to cessation of contractions and four or fewer contractions per hour in the last hour of therapy. Other atosiban regimens were comparable to ritodrine, except for the drug discontinuation rate for adverse experiences. Bolus therapy with high-dose atosiban resulted in a significantly greater proportion of patients who stopped contracting within the first 2 hours of treatment (17 of 63) compared with those not receiving a bolus (six of 58, P = .017). Because of adverse experiences, the study drug was discontinued in one of 244 atosiban patients and 15 of 58 ritodrine patients.Atosiban's effect on uterine activity in preterm labor was enhanced by bolus infusion and was similar to the effect of ritodrine, but with fewer side effects.