The reviews included 70 trials using vaginal misoprostol for labour induction and 13 trials using oral misoprostol.

Both vaginal and oral misoprostol were more effective than placebo in cervical ripening and shortening the insertion to delivery interval. The numbers studied were too small to assess the impact on obstetric management and maternal and neonatal complications. There were 5 small trials using vaginal and one using oral misoprostol.

Vaginal misoprostol was more effective than oxytocin for labour induction with marginal statistical significance (five trials, RR of failure to achieve vaginal delivery within 24 hours: 0.66, 95% CI 0.44 to 1.00). However, uterine hyperstimulation without associated fetal heart rate changes were more common with misoprostol (9 trials, RR: 2.22, 95% CI 1.77 to 2.79 respectively). There was a trend to reduced epidural analgesia with misoprostol (3 trials, RR: 0.82, 95% CI 0.67 to 1.00). The trial results with respect to caesarean sections were inconsistent. There were no differences in perinatal or maternal adverse outcomes.

Two small trials evaluated oral misoprostol against oxytocin. There were no clinically or statistically significant differences in the prespecified outcomes.

Vaginal misoprostol versus vaginal prostaglandins: 25 trials were included. Vaginal misoprostol was more effective than other prostaglandins for the induction of labour. Overall, oxytocin augmentation (25 trials, RR: 0.65, 95% CI 0.57 to 0.73) and failure to achieve vaginal delivery within 24 hours (13 trials, RR: 0.80, 95% CI 0.73 to 0.87) were reduced with vaginal misoprostol. Uterine hyperstimulation, with or without fetal heart rate changes and meconium stained amniotic fluid were more common with misoprostol. Caesarean section rates were variable between trials, with no significant differences overall.

Results were similar for the sub-groups of women with unfavourable cervices and those with intact membranes. For sub-groups of primiparous or multiparous women, the numbers were small and no differences in any outcomes were shown.

Oral misoprostol versus vaginal prostaglandins: Only two studies were included. In both trials induction to delivery interval was somewhat shorter in the vaginal prostaglandin group (weighted mean difference 2.7 hours, 95% CI 0.72 to 4.84). Pooled hyperstimulation rate with oral misoprostol was 4.3% compared with 4.9 % in the vaginal prostaglandin. The cesarean section rate was 18.3% and 20.3% respectively. None of these differences were statistically significant. There were no reported cases of serious maternal morbidity, no differences in neonatal mortality and morbidity.

Vaginal misoprostol versus intracervical prostaglandin: 17 trials were included. Failure to achieve delivery within 24 hours (5 trials, RR: 0.68, 95% CI 0.59 to 0.78) and oxytocin augmentation were reduced with misoprostol (13 trials, RR: 0.56, 95% CI 0.51 to 0.61). Uterine hyperstimulation with or without fetal heart rate changes was more common with misoprostol. There were no consistent patterns for vaginal instrumental delivery and cesarean sections.

Meconium stained amniotic fluid was increased with misoprostol (9 trials, RR: 1.27, 95% CI 1.00 to 1.61). There were no statistically significant differences in perinatal or maternal outcomes.

Oral misoprostol versus intracervical prostaglandin: In a single trial including 200 women assigned to a single oral dose of misoprostol 200 mcg or intracervical dinoprostone (0.5 mg four hourly) oral misoprostol was more effective in achieving vaginal delivery within 24 hours than intracervical prostaglandins (76% versus 50%), though the difference was not statistically significant.

The lower dose regimens ranged from 12.5 mg 6 hourly to 50 mg 6 hourly in 13 trials. The lower dose regimens did not show significantly more failures to achieve vaginal delivery within 24 hours. There was significantly more use of oxytocin in the low dose group (12 trials, RR: 1.23, 95% CI 1.08 to 1.40) but no differences in mode of delivery, meconium stained liquor or maternal side effects. There were fewer cases of uterine hyperstimulation with or without FHR changes.

In seven trials oral misoprostol appeared to be less effective than vaginal misoprostol. More women in the oral misoprostol group did not achieve vaginal delivery within 24 hours of randomisation (50.0 %) compared with 39.7 % in the vaginal misoprostol group (RR 1.27, 95% CI 1.09 1.47).

Caesarean section rate was 16.7% in the oral misoprostol group compared with 21.7 % in the vaginal misoprostol (RR 0.77, 95% CI 0.61 0.97). There was no difference in uterine hyperstimulation with fetal heart changes (8.5 % versus 7.4%; RR 1.11, 95% CI 0.78 1.59). There were no reported cases of severe neonatal and maternal morbidity with both.

Subgroup analysis was done according to cervical status (favourable, unfavourable), parity (primiparae, multiparae) and amniotic membrane status (intact, ruptured). In the subgroup with unfavourable cervix, there were no significant differences in reported outcomes. However, women in the oral misoprostol group (primiparous, multiparous, with intact membranes) were less likely to achieve vaginal delivery within 24 hours than subgroup received vaginal misoprostol.

In the subgroup analysis according to dose of misoprostol, four trials compared 50 mcg oral misoprostol with vaginal administration. This oral regimen was less effective than vaginal administration in achieving vaginal delivery within 24 hours.

Planned subgroup analyses could not explain the statistically significant heterogeneity for the outcomes such as failure to achieve vaginal delivery after 24 hours, uterine hyperstimulation, uterine tachysystole and randomization to delivery interval, found in most comparisons with at least three studies.

In one trial including 467 participants, uterine hyperstimulation with fetal heart rate changes was less with the gel preparation. The use of oxytocin and epidural analgesia were increased however.

The reviewers concluded that misoprostol shows promise as a highly effective, inexpensive and convenient agent for labour induction. However, the apparent increase in uterine hyperstimulation and the rare but serious complications like rupture uterus are of concern. It cannot be recommended for routine use at this stage. There is an urgent need for trials to establish its safety. Only large pragmatic trials with adequate sample size will be able to address the risks of uterine hyperstimulation, uterine rupture and serious neonatal and maternal morbidity.

The review included all adequately controlled trials, which could be identified and analysed them appropriately.

This document should be cited as: Hany Abdel-Aleem. Misoprostol for induction of labour: RHL commentary (last revised: 16 September 2004). The WHO Reproductive Health Library, No 9, Update Software Ltd, Oxford, 2006. www.rhlibrary.com