OBJECTIVE To compare the efficacy of labour induction using sublingual misoprostol versus combined artificial rupture of membranes and oxytocin infusion for nulliparous women with a favourable cervix at term. METHODS Open randomised controlled trial. METHODS Regional hospital, Hong Kong. METHODS Fifty nulliparous women with a favourable cervix (Bishop score 6 or more) at term and indications for labour induction. METHODS With their informed consent, 100 eligible women were to be randomised to receive either sublingual misoprostol 50 micrograms every 4 hours for up to five doses or oxytocin infusion after artificial rupture of membranes. Interim analysis was planned at a sample size of 50. METHODS Vaginal delivery within 24 hours of induction. RESULTS The study was terminated when interim analysis of the first 50 recruits showed that a significantly smaller proportion of misoprostol-treated women delivered vaginally within 24 hours of induction than in the conventional treatment group (68% vs 100%; relative risk, 0.68; 95% confidence interval, 0.51-0.91; P=0.009), although comparable numbers of women eventually delivered vaginally. The mean induction to vaginal delivery interval was 4.5 hours longer in the misoprostol group (P=0.027). After misoprostol treatment, all women went into labour. Forty percent of them delivered without oxytocin. There was no significant difference in uterine hyperstimulation rate, operative delivery rate, and neonatal outcomes. Maternal satisfaction was higher in the misoprostol group (92% vs 60%; relative risk, 1.53; 95% confidence interval, 1.09-2.16; P=0.008). CONCLUSIONS Despite being well accepted by women, labour induction using this regimen of sublingual misoprostol is less effective in achieving vaginal delivery within 24 hours.