The review concluded that elective induction of labour for post-term pregnancies appeared to be effective in reducing perinatal morbidity and mortality compared with expectant management and should be offered to women with post-term pregnancies after discussing the benefits and risks. Given the wide confidence intervals in some outcomes, including stillbirth, the authors' conclusion should be interpreted with caution.

To determine the efficacy of elective induction of labour for post-term pregnancies (beyond 41 weeks gestation) compared with expectant management on stillbirths.

PubMed, The Cochrane Library and WHO regional databases were searched without language restriction; search terms but not search dates were reported. Bibliographies of relevant reviews were handsearched. Experts in the field were also contacted to identify any relevant unpublished trials. Only articles published in English or the abstracts of non-English studies with relevant data were included.

Eligible randomised controlled trials (RCTs), quasi-randomised trials and observational studies compared elective induction of labour for low-risk pregnancies at or beyond 41 weeks gestation with expectant management that reported stillbirth or perinatal mortality. Different methods of induction and/or techniques for monitoring expectantly managed pregnancies were accepted. Studies of induction for reasons other than prolonged gestation (such as macrosomia) were excluded. Morbidity outcomes associated with post-term pregnancies were also considered (such as meconium aspiration syndrome, birth asphyxia and macrosomia).

Methods of induction and monitoring of the controlled group varied across studies. Interventions were delivered through routine health system in all studies. In the RCTs, the most frequently reported method of induction was oxytocin infusion and amniotomy; others included membrane stripping, prostaglandin gel cervical ripening and extraamniotic saline infusion. RCT settings included hospitals and clinics in an urban environments, where reported, and definition of post-term pregnancy ranged from 287 to 294 days.

The authors did not state how many reviewers performed the study selection

Two reviewers assessed the quality of included studies. Data on allocation concealment, blinding assessment, rates of withdrawal/loss to follow-up and intention-to-treat analysis were extracted.

Two reviewers independently extracted data to enable the calculation of risk ratios (RRs) along with 95% confidence intervals (CIs). A standardised data extraction form was used. The definitions of stillbirth and perinatal mortality were taken as defined by the study author(s).

Summary risk ratios and 95% confidence intervals were estimated using a fixed-effect model, or a random-effects model where significant statistical heterogeneity was found. Statistical heterogeneity was assessed using Χ² and Ι² and by observing the overlap of the confidence intervals in the included studies. Subgroup analyses were performed based on gestational weeks completed (41 or 42). Number-needed-to-treat was also calculated (method used not reported). Summary estimates were only calculated for RCTs; a narrative synthesis was presented for other types of study.

Twenty-five studies were included in the review; 16 RCTs (7,001 patients, range 22 to 3,407), three quasi-randomised trials (658 patients, range 100 to 402), and six observational studies (at least 5,574 patients, range 254 to 3,914 where reported). Seven RCTs reported allocation concealment (reported as adequate in three), one trial reported blinding and six trials indicated intention-to-treat analyses was performed.

No significant between-group difference was found for stillbirth (RR 0.29, 95% CI 0.06 to 1.38; Ι²=0%, 14 RCTs); 10 RCTs had no reported events. Subgroup analyses for induction of labour at 41 weeks remained statistically insignificant for incidence of stillbirth (RR 0.29, 95% CI 0.06 to 1.38; Ι²=0%, 12 RCTs). No estimate was possible for induction of labour at 42 weeks. Elective induction of labour for post-term pregnancies beyond 41 weeks gestation significantly reduced (all-cause) perinatal mortality compared with expectant management (RR 0.31, 95% CI 0.11 to 0.88; Ι²=0%, 14 RCTs). Six of the included RCT had no reported events. Results remained statistically significant for induction of labour at 41 weeks (RR 0.27, 95% CI 0.08 to 0.98; Ι²=0%, 12 RCTs), but no significant between-group difference was found for induction of labour at 42 weeks (RR 0.41, 95% CI 0.06 to 2.73; Ι²=0%, two RCTs). The total number of reported events for the primary outcomes was low.

Elective induction of labour significantly reduced the incidence of meconium aspiration syndrome (RR 0.43, 95% CI 0.23 to 0.79; Ι²=0%, seven RCTs) and macrosomia (RR 0.72, 95% CI 0.54 to 0.98; Ι²=71%, seven RCTs), but no significant difference was found for birth asphyxia (two RCTs).

Results from individual quasi-randomised trials and observational studies were also reported; evidence from these studies was mixed, approximately half the studies demonstrated a significant difference in the rate of stillbirths and/or perinatal mortality for active compared with expectant management.

Elective induction of labour appeared to be effective in reducing perinatal morbidity and mortality associated with post-term pregnancies compared with expectant management, and should be offered to women with post-term pregnancies after discussing the benefits and risks involved.

The review question was supported by defined inclusion criteria. Several electronic databases were searched and some attempt was made to locate unpublished trials. Whilst the search was not restricted by language, non-English articles were not translated and if the abstract did not provide sufficient data were excluded so some studies may have been missed. No formal assessment of publication bias was reported. Appropriate steps were taken to minimise the likelihood of reviewer error or bias for data extraction and quality assessment but it was not clear whether a similar procedure was undertaken for study selection.

Quality assessment of individual trials, although somewhat buried in their approach to grading the quality of overall evidence, included some relevant criteria, but results were only available from a supplemental table. Standard techniques were used in the meta-analyses and assessment of heterogeneity was performed, but many of the most commonly used meta-analytical methods were biased when events were rare (in particular inverse variance methods should be avoided). The authors acknowledged that the reported number of events was low, and appeared to have taken this into account when interpreting the results. Overall, the authors' cautious conclusion might still be considered a little strong given the wide confidence intervals for some outcomes, in particular stillbirth.

Practice: The authors stated that routine induction of labour at or beyond 41 weeks should be offered to women with post-term pregnancies after discussing the benefits and risks of induction of labour.

Research: Using CHERG (Child Health Epidemiology Reference Group) rules, the authors recommend a reduction in perinatal mortality of 69% as a surrogate for reduction in stillbirths with induction of labour for pregnancies beyond 41 weeks gestation (based on the association of induction of labour beyond 41 weeks with a reduction in perinatal deaths, and assuming these were mostly due to stillbirths).

USA Fund for UNICEF; Bill and Melinda Gates Foundation.

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.