Cochrane Summaries

Trusted evidence. Informed decisions. Better health.
Language:
English

Mechanical methods for induction of labour

Jozwiak M, Bloemenkamp KWM, Kelly AJ, Mol BWJ, Irion O, Boulvain M
Published Online: 
14 March 2012

Labour induction is a common obstetric procedure that is carried out when the risk of continuing pregnancy outweighs the benefits.

Mechanical methods of induction were developed to promote cervical ripening and the onset of labour by stretching the cervix. They are amongst the oldest methods to initiate labour. More recently pharmacological prostaglandins (vaginal and intracervical prostaglandin E2, vaginal and oral misoprostol) and oxytocin have partly replaced mechanical methods. The goal of this review of 71 randomised controlled trials (9722 women) was to determine the effects of mechanical methods for cervical ripening or induction of labour in comparison with no treatment, prostaglandins and oxytocin for women in the third trimester of pregnancy.

The review includes 71 randomised controlled trials (total of 9722 women), ranging from 39 to 588 women per study. Most studies reported on caesarean section; all other outcomes are based on substantially fewer women. Mechanical methods were as effective as prostaglandins in achieving delivery within 24 hours of the start of the intervention, with fewer episodes of excessive uterine contractions. The risk of caesarean section did not differ. Few studies addressed the issue of infection, which appeared not to be higher when using mechanical methods. Thus mechanical methods can be considered to have less side effects compared with prostaglandins. The one study that reported on maternal discomfort showed more discomfort during ripening with prostaglandins compared with Foley catheter insertion, and with double balloon devices compared with Foley catheters. This outcome may influence the choice of method and is an important issue to be addressed in future studies. Mechanical methods were more effective than induction with oxytocin. Serious neonatal and maternal ill effects were not often reported and did not differ between the interventions.