OBJECTIVE Our purpose was to compare the efficacy of 2 different prostaglandin E2 delivery methods for preinduction cervical ripening. METHODS Ninety patients admitted for labor induction with a Bishop score <8 were randomized to either 0.5 mg prostaglandin E2 intracervical gel (Prepidil) every 6 hours for 2 doses or 10 mg prostaglandin E2 slow release vaginal pessary (Cervidil). Oxytocin induction was begun after 12 hours. It was estimated that enrollment of 90 women would be required to identify a 30% difference in the percent delivered in <24 hours (1 - beta = .80, alpha = .05). Data were analyzed with use of chi2 analysis or the Student t test. RESULTS There were 45 subjects in each treatment arm. The percent delivered by 24 hours was 53% with intracervical gel and 63% with vaginal pessary (P = .28). Mean change in Bishop score was 1.8 +/- 1.9 for the intracervical gel versus 3.2 +/- 3.1 for the vaginal pessary (P = .01). No difference was demonstrated in mean time to delivery, 28.3 versus 24.0 hours (P = .19) or percent requiring cesarean section. CONCLUSIONS Preinduction cervical ripening with a slow release prostaglandin E2 vaginal pessary resulted in greater change in Bishop score than with intracervical prostaglandin E2. There was a trend toward shorter time to delivery with the pessary. There was no statistically significant difference in percent delivered in <24 hours.