To compare the efficacy and safety of 50 microg of sublingual misoprostol with 25 microg of vaginal misoprostol administered for labour induction at term. Design Double-blinded, randomised controlled trial. Setting University Hospital, Kaunas, Lithuania. Sample A total of 140 women at term with indications for labour induction. Methods Women were randomised to receive either 50 microg of sublingual misoprostol with vaginal placebo (n = 70) or sublingual placebo with 25 microg of vaginal misoprostol (n = 70) every 4 hours (maximum six doses). Main outcome measures The number of women delivering vaginally within 24 hours of labour induction. Results Fifty-eight women (83%) in the sublingual misoprostol group and 53 (76%) in the vaginal misoprostol group delivered vaginally within 24 hours [relative risk (RR) 1.1, 95% confidential interval (CI) 0.9-1.3]. However, the induction to vaginal delivery time was significantly shorter in the sublingual group (15.0 +/- 3.7 hours) compared with the vaginal group (16.7 +/- 4.1 hours, P = 0.03). The incidence of tachysystole was more than three-fold higher in the sublingual than in the vaginal group (14 versus 4.3%; RR 3.3, 95% CI 0.9-11.6), but this was not statistically significant. There were no significant differences in the incidence of hypertonus or hyperstimulation syndrome, mode of delivery, interventions for fetal distress or neonatal outcomes between the two groups. Conclusion A 50 microg of sublingual misoprostol 4 hourly for labour induction at term seems to have similar efficacy as 25 microg of vaginal misoprostol. Further studies on safety with larger numbers of women need to be conducted before routine sublingual misoprostol use in this setting.