Background Misoprostol is a commonly used prostaglandin to induce labour. A potential risk of induction, however, is caesarean delivery, especially in women with an unfavourable cervix.Objectives To evaluate the use of misoprostol, compared with prostaglandin E 2 (PgE 2 ), for labour induction in women at term with an unfavourable cervix and intact membranes.Search strategy PubMed, Medline, EMBASE and the Cochrane Library were searched for articles published in any language from January 1987 to December 2005, using the keywords 'misoprostol', 'labour/labor' and 'induction'.Selection criteria We identified randomised trials of women at term ( ‡37 weeks of gestation) with intact membranes and unfavourable cervix, undergoing labour induction with misoprostol, orally, vaginally, sublingually or buccally, compared with PgE 2 vaginally or intracervically.Data collection and analysis Caesarean delivery was the primary outcome, with tachysystole and hyperstimulation as secondary outcomes. The primary analysis compared any misoprostol with any PgE 2 for all women, with a subgroup analysis for nulliparous women. Secondary analyses compared different routes and doses of misoprostol (oral or vaginal and 25 microgram or >25 microgram) and PgE 2 (intracervical or vaginal). Relative risks (RR) and 95% confidence intervals (CI) were calculated using random effects models.Main results Fourteen of 611 articles identified met the criteria for systematic review, with three providing information for nulliparous women. There was no difference in the risk of caesarean delivery between misoprostol and PgE 2 groups (RR = 0.99, 95% CI = 0.83-1.17). Any misoprostol was associated with higher risks of tachysystole and hyperstimulation compared with any PgE 2 (RR = 1.86, 95% CI = 1.01-3.43 and RR = 3.71, 95% CI = 2.00-6.88, respectively). There was a higher rate of vaginal delivery within 24 hours among all vaginal deliveries with any misoprostol compared with any PgE 2 (RR = 1.14, 95% CI = 1.00-1.31), and among all deliveries, a lower rate of oxytocin use (RR = 0.71, 95% CI = 0.60-0.85) but a trend towards increased meconium staining was observed (RR = 1.22, 95% CI = 0.96-1.55). The use of misoprostol at starting dosages >25 microgram had similar findings to the primary analysis. Studies of lower misoprostol dosing (starting dose of 25 microgram) did not show any differences in the outcomes of interest, but the sample size of this secondary analysis was small (304 women, 155 receiving misoprostol).Author's conclusions Although misoprostol in women at term with an unfavourable cervix and intact membranes was more effective than PgE 2 in achieving vaginal delivery within 24 hours, misoprostol does not reduce the rate of caesarean delivery either in all women or in the subgroup of nulliparous women, and it increases the rates of tachysystole and hyperstimulation. Further studies of misoprostol using a starting dose of 25 microgram may be warranted.