Induction of labour for improving birth outcomes for women at or beyond term

Philippa Middleton; Emily Shepherd; Caroline A Crowther

doi:10.1002/14651858.CD004945.pub4

. 2018 May 9;2018(5):CD004945. doi: 10.1002/14651858.CD004945.pub4

Induction of labour for improving birth outcomes for women at or beyond term

Philippa Middleton ^1,^✉, Emily Shepherd ², Caroline A Crowther ³

Editor: Cochrane Pregnancy and Childbirth Group

PMCID: PMC6494436 PMID: 29741208

Abstract

Background

Beyond term, the risks of stillbirth or neonatal death increase. It is unclear whether a policy of labour induction can reduce these risks. This Cochrane review is an update of a review that was originally published in 2006 and subsequently updated in 2012

Objectives

To assess the effects of a policy of labour induction at or beyond term compared with a policy of awaiting spontaneous labour or until an indication for birth induction of labour is identified) on pregnancy outcomes for infant and mother.

Search methods

We searched Cochrane Pregnancy and Childbirth’s Trials Register, ClinicalTrials.gov and the WHO International Clinical Trials Registry Platform (ICTRP) (9 October 2017), and reference lists of retrieved studies.

Selection criteria

Randomised controlled trials (RCTs) conducted in pregnant women at or beyond term, comparing a policy of labour induction with a policy of awaiting spontaneous onset of labour (expectant management). We also included trials published in abstract form only. Cluster‐RCTs, quasi‐RCTs and trials using a cross‐over design are not eligible for inclusion in this review.

Data collection and analysis

Two reviewers independently assessed trials for inclusion, assessed risk of bias and extracted data. Data were checked for accuracy. We assessed the quality of evidence using the GRADE approach.

Main results

In this updated review, we included 30 RCTs (reporting on 12,479 women). The trials took place in Norway, China, Thailand, the USA, Austria, Turkey, Canada, UK, India, Tunisia, Finland, Spain, Sweden and the Netherlands. They were generally at a moderate risk of bias.

Compared with a policy of expectant management, a policy of labour induction was associated with fewer (all‐cause) perinatal deaths (risk ratio (RR) 0.33, 95% confidence interval (CI) 0.14 to 0.78; 20 trials, 9960 infants; moderate‐quality evidence). There were two perinatal deaths in the labour induction policy group compared with 16 perinatal deaths in the expectant management group. The number needed to treat to for an additional beneficial outcome (NNTB) with induction of labour in order to prevent one perinatal death was 426 (95% CI 338 to 1337). There were fewer stillbirths in the induction group (RR 0.33, 95% CI 0.11 to 0.96; 20 trials, 9960 infants; moderate‐quality evidence); there was one stillbirth in the induction policy arm and 10 in the expectant management group.

For women in the policy of induction arms of trials, there were fewer caesarean sections compared with expectant management (RR 0.92, 95% CI 0.85 to 0.99; 27 trials, 11,738 women; moderate‐quality evidence); and a corresponding marginal increase in operative vaginal births with induction (RR 1.07, 95% CI 0.99 to 1.16; 18 trials, 9281 women; moderate‐quality evidence). There was no evidence of a difference between groups for perineal trauma (RR 1.09, 95% CI 0.65 to 1.83; 4 trials; 3028 women; low‐quality evidence), postpartum haemorrhage (RR 1.09 95% CI 0.92 to 1.30, 5 trials; 3315 women; low‐quality evidence), or length of maternal hospital stay (average mean difference (MD) ‐0.34 days, 95% CI ‐1.00 to 0.33; 5 trials; 1146 women; Tau² = 0.49; I² 95%; very low‐quality evidence).

Rates of neonatal intensive care unit (NICU) admission were lower (RR 0.88, 95% CI 0.77 to 1.01; 13 trials, 8531 infants; moderate‐quality evidence) and fewer babies had Apgar scores less than seven at five minutes in the induction groups compared with expectant management (RR 0.70, 95% CI 0.50 to 0.98; 16 trials, 9047 infants; moderate‐quality evidence).

There was no evidence of a difference for neonatal trauma (RR 1.18, 95% CI 0.68 to 2.05; 3 trials, 4255 infants; low‐quality evidence), for induction compared with expectant management.

Neonatal encephalopathy, neurodevelopment at childhood follow‐up, breastfeeding at discharge and postnatal depression were not reported by any trials.

In subgroup analyses, no clear differences between timing of induction (< 41 weeks versus ≥ 41 weeks' gestation) or by state of cervix were seen for perinatal death, stillbirth, NICU admission, caesarean section, or perineal trauma. However, operative vaginal birth was more common in the inductions at < 41 weeks' gestation subgroup compared with inductions at later gestational ages. The majority of trials (about 75% of participants) adopted a policy of induction at ≥ 41 weeks (> 287 days) gestation for the intervention arm.

Authors' conclusions

A policy of labour induction at or beyond term compared with expectant management is associated with fewer perinatal deaths and fewer caesarean sections; but more operative vaginal births. NICU admissions were lower and fewer babies had low Apgar scores with induction. No important differences were seen for most of the other maternal and infant outcomes.

Most of the important outcomes assessed using GRADE had a rating of moderate or low‐quality evidence ‐ with downgrading decisions generally due to study limitations such as lack of blinding (a condition inherent in comparisons between a policy of acting and of waiting), or imprecise effect estimates. One outcome (length of maternal stay) was downgraded further to very low‐quality evidence due to inconsistency.

Although the absolute risk of perinatal death is small, it may be helpful to offer women appropriate counselling to help choose between scheduled induction for a post‐term pregnancy or monitoring without (or later) induction).

The optimal timing of offering induction of labour to women at or beyond term warrants further investigation, as does further exploration of risk profiles of women and their values and preferences. Individual participant meta‐analysis is likely to help elucidate the role of factors, such as parity, in influencing outcomes of induction compared with expectant management.

Plain language summary

Induction of labour in women with normal pregnancies at or beyond term

What is the issue?

A normal pregnancy lasts about 40 weeks from the start of the woman's last menstrual period, but anything from 37 to 42 weeks is considered as being at term (within the normal range). If a pregnancy goes too long, a woman and her clinician may wish to intervene to bring the birth on, for example, by induction.

Why is this important?

Births after 42 weeks' gestation may slightly increase risks for babies, including a greater risk of death (before or shortly after birth). However induction of labour may also have risks for mothers and their babies, especially if women are not ready to labour. No tests can predict if babies would be better to stay inside their mother or if labour should be induced to make the birth happen sooner. Many hospitals therefore have policies for how long pregnancies should continue. This update (originally published in 2006 and subsequently updated in 2012) looks to see if inducing labour at a set time at or beyond term, could reduce risks for the babies.

What evidence did we find?

We searched for evidence up 9 October 2017 and identified 30 trials with over 12,000 women. The trials took place in Norway, China, Thailand, the USA, Austria, Turkey, Canada, UK, India, Tunisia, Finland, Spain, Sweden and the Netherlands. The evidence was mostly of moderate quality. The trials compared a policy to induce labour at or later than term (usually after 41 completed weeks of gestation (> 287 days)) with waiting for labour to start and/or waiting for a period before inducing labour.

We found that there were fewer deaths of babies in hospitals with a policy to induce when a pregnancy was continuing beyond term (moderate‐quality evidence). Fewer caesarean births were required with induction compared with waiting, but more assisted vaginal births were required with induction. There were fewer admissions to the intensive care nursery and fewer low Apgar scores at five minutes after birth (a simple test to test babies' health) in the induction groups compared with waiting (moderate‐quality evidence). We found that there were no clear differences between a policy to induce at or later than term or waiting in the risks of mothers having trauma to their perineum or bleeding after birth (both low‐quality evidence), in the length of their hospital stay (very‐low quality evidence), or in their babies having trauma (low‐quality evidence), None of the trials provided information on breastfeeding at discharge from hospital, postnatal depression, or whether the babies had encephalopathy (early abnormal neurological function), or child development.

What does this mean?

A policy of labour induction compared with expectant management is associated with fewer deaths of babies and fewer caesarean sections; but more assisted vaginal births. Although the chances of babies dying are small, it may help to offer women appropriate counselling to make an informed choice between induction of labour for pregnancies at, or later than, term ‐ or waiting for labour to start and/or waiting before inducing labour.

The best time to offer induction of labour to women at or beyond term is not yet clear and warrants further investigation. The risk profiles of women as well as their values and preferences could also be considered.

Summary of findings

Background

Description of the condition

A pregnant women is 'at term' when her pregnancy duration reaches 37 weeks. Up to 10% of pregnancies continue beyond 294 days (42⁰ weeks) and are described as being 'post‐term' or 'postdate' (Olesen 2003; Roos 2010; Zeitlin 2007). In 2015 in the USA, 6.5% of pregnancies progressed to 41 weeks and 0.4% continued to 42 weeks or later (Martin 2017).

While the aetiology of post‐term birth is not well elucidated (Mandruzzato 2010), risk factors such as obesity, nulliparity and maternal age greater than 30 years have been associated with an increased risk of post‐term birth (Arrowsmith 2011; Caughey 2009; Heslehurst 2017; Roos 2010). Placental senescence may play a role in the pathophysiology of post‐term birth (Mandruzzato 2010), and genetic/epigenetic factors have also been implicated (Schierding 2014).

Both the mother and the infant are at increased risk of adverse events when the pregnancy continues beyond term (Hilder 1998). In a study from the Norwegian Birth Registry (Heimstad 2008), the perinatal death rate was 0.018% at day 287 (41 weeks) and 0.51% at day 302⁺ (> 43 weeks). These findings are important in that, even in a setting where early booking allows accurate assessment of gestational age and antenatal services are accessible for most women, post‐term pregnancy constitutes a high‐risk situation, especially for the baby. In another Norwegian study of nearly two million births from 1967 to 2006, the risk of post‐term infant death was strongly associated with growth restriction (Morken 2014).

The obstetric problems associated with post‐term pregnancy include induction of labour with an unfavourable cervix, caesarean section, prolonged labour, postpartum haemorrhage and traumatic birth. It is likely that some of these unwanted outcomes result from intervening when the uterus and cervix are not ready for labour (Caughey 2004).

Description of the intervention

Induction of labour is widely practised to try and prevent problems or outcomes such as caesarean section, prolonged labour, postpartum haemorrhage and traumatic birth (Caughey 2004), and to improve health outcomes for women and their infants. In the USA, nearly one in four births is induced (23.8% in 2015 ‐ Martin 2017). For post‐term pregnancies, this may be one in every two births (e.g. 52% induction rate for gestations ≥ 41 weeks ‐ Wolff 2016).

Variation in rates of post‐term births suggests that different policies and practices for managing post‐term pregnancies (especially timing of inductions) are used in Europe (Zeitlin 2007), and elsewhere. There is concern about the high and increasing induction rate in many countries, and increasing caesarean rates despite increasing induction rates (Keirse 2010).

Earlier versions of this review included interventions involving monitoring, such as early pregnancy ultrasound, that may have an effect on the outcomes of pregnancies for women at or beyond term. This topic is addressed in the Cochrane review 'Ultrasound for fetal assessment in early pregnancy' (Whitworth 2015). In this update, we evaluate the effects of timing of labour induction at or beyond term compared with expectant management (which may include various intensities and forms of monitoring).

How the intervention might work

When the cervix is favourable (usually a Bishop score of six or more), induction is often carried out by oxytocin and artificial rupture of amniotic membranes. If the cervix is not favourable then usually a prostaglandin gel or tablet is placed in the vagina or cervix to ripen the cervix and initiate the uterine contractions and labour. Many protocols are used with varying repeat intervals and transition to oxytocin and amniotomy depending on the onset of uterine contractions and progress of cervical dilatation.

Why it is important to do this review

Determining the threshold for induction of post‐term pregnancies has been described as 'the 41 week to 42 week dilemma' (Kortekaas 2014), with many hospitals now adopting a policy of induction at 41 weeks rather than a policy of waiting to induce at 42 weeks if spontaneous labour has not occurred. This 41‐week policy may substantially increase numbers of inductions ‐ in the Netherlands this policy would mean that 18% of all pregnant women would be induced compared with 1.5% if a 42‐week policy was adopted (Kortekaas 2014). It is important to assess whether improved outcomes such as reduced perinatal death and fewer caesarean sections can be achieved with earlier inductions and to determine the optimal gestational threshold for induction.

Objectives

To assess the effects of a policy of labour induction at or beyond term compared with a policy of awaiting spontaneous labour indefinitely (until a later gestational age or until a maternal or fetal indication for induction of labour is identified) on pregnancy outcomes for the infant and the mother.

Methods

Criteria for considering studies for this review

Types of studies

We included randomised controlled trials. We included trials presented only as abstracts as well as trials published in full‐text manuscript format.

Cluster‐randomised trials, quasi‐randomised trials and cross‐over trials are not eligible for inclusion in this review.

Types of participants

Types of interventions

The intervention evaluated in this review is a policy of labour induction at a predetermined gestational age at or beyond term. This policy is compared with 'expectant management' until an indication for birth arises. The trial protocols differ according to:

gestational age used in the policy;
actual method of labour induction (prostaglandins, misoprostol, +/‐ oxytocin), protocol used (dosage of any drugs, timing, frequency of use and mode of administration);
expectant management protocols (intensity of fetal well‐being assessment and fetal monitoring techniques used).

Types of outcome measures

Primary outcomes

The primary outcome of this review was perinatal death, defined as intrauterine deaths plus neonatal deaths in the first week of life.

Secondary outcomes

For the infant/child

Stillbirth
Neonatal death within first week
Birth asphyxia (as defined by trialists)
Admission to neonatal intensive care unit
Neonatal convulsions
Neonatal encephalopathy
Use of anticonvulsants
Meconium aspiration syndrome
Pneumonia
Apgar score less than seven at five minutes
Birthweight
Birthweight > 4000 g
Neonatal trauma
Neurodevelopment at childhood follow‐up

For the mother

Mode of birth (caesarean section)
Operative vaginal birth (forceps or ventouse)
Analgesia used
Perineal trauma
Prolonged labour (cut‐off used by the trialists was used)
Postpartum haemorrhage (cut‐off used by the trialists was used)
Anxiety before birth
Other measures of satisfaction with the approach
Breastfeeding at discharge
Postnatal depression

Health services use

Length of maternal postnatal stay
Length of neonatal postnatal stay
Length of labour

Cost‐related analyses are described in the Discussion.

Search methods for identification of studies

The following methods section of this review is based on a standard template used by Cochrane Pregnancy and Childbirth.

Electronic searches

We searched Cochrane Pregnancy and Childbirth’s Trials Register by contacting their Information Specialist (9 October 2017).

The Register is a database containing over 24,000 reports of controlled trials in the field of pregnancy and childbirth. For full search methods used to populate Pregnancy and Childbirth’s Trials Register including the detailed search strategies for CENTRAL, MEDLINE, Embase and CINAHL; the list of handsearched journals and conference proceedings, and the list of journals reviewed via the current awareness service, please follow this link to the editorial information about the Cochrane Pregnancy and Childbirth in the Cochrane Library and select the ‘Specialized Register ’ section from the options on the left side of the screen.

Briefly, Cochrane Pregnancy and Childbirth’s Trials Register is maintained by their Information Specialist and contains trials identified from:

monthly searches of the Cochrane Central Register of Controlled Trials (CENTRAL);
weekly searches of MEDLINE (Ovid);
weekly searches of Embase (Ovid);
monthly searches of CINAHL (EBSCO);
handsearches of 30 journals and the proceedings of major conferences;
weekly current awareness alerts for a further 44 journals plus monthly BioMed Central email alerts.

Search results are screened by two people and the full text of all relevant trial reports identified through the searching activities described above is reviewed. Based on the intervention described, each trial report is assigned a number that corresponds to a specific Pregnancy and Childbirth review topic (or topics), and is then added to the Register. The Information Specialist searches the Register for each review using this topic number rather than keywords. This results in a more specific search set that has been fully accounted for in the relevant review sections (Included studies; Excluded studies; Studies awaiting classification; Ongoing studies).

In addition, we searched ClinicalTrials.gov and the WHO International Clinical Trials Registry Platform (ICTRP) for unpublished, planned and ongoing trial reports (9 October 2017) (see: Appendix 1 for search methods used).

Searching other resources

We searched the reference lists of retrieved studies. We did not apply any language or date restrictions.

Data collection and analysis

For methods used in the previous version of this review, seeGülmezoglu 2012.

For this update, the following methods were used for assessing the 15 reports that were identified as a result of the updated search. Where required, information pertaining to the previously included studies was updated according to methods outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011).

Selection of studies

Two review authors independently assessed for inclusion all the potential studies identified as a result of the search strategy. We resolved any disagreement through discussion or, if required, we consulted the third review author.

Data extraction and management

We designed a form to extract data. For eligible studies, two review authors extracted the data using the agreed form. We resolved discrepancies through discussion or, if required, we consulted the third review author. Data were entered into Review Manager software (RevMan 2014), and checked for accuracy.

When information regarding any of the above was unclear, we planned to contact authors of the original reports to provide further details.

Assessment of risk of bias in included studies

Two review authors independently assessed risk of bias for each study using the criteria outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). Any disagreement was resolved by discussion or by involving the third review author.

(1) Allocation (checking for possible selection bias)

We described for each included study the method used to generate the allocation sequence in sufficient detail to allow an assessment of whether it should produce comparable groups.

We assessed the method as:

low risk of bias (any truly random process, e.g. random number table; computer random number generator);
high risk of bias (any non‐random process, e.g. odd or even date of birth; hospital or clinic record number);
unclear risk of bias.

(2) Allocation concealment (checking for possible selection bias)

We described for each included study the method used to conceal allocation to interventions prior to assignment and assessed whether intervention allocation could have been foreseen in advance of, or during recruitment, or changed after assignment.

We assessed the methods as:

low risk of bias (e.g. telephone or central randomisation; consecutively numbered sealed opaque envelopes);
high risk of bias (open random allocation; unsealed or non‐opaque envelopes, alternation; date of birth);
unclear risk of bias.

(3.1) Blinding of participants and personnel (checking for possible performance bias)

We described for each included study the methods used, if any, to blind study participants and personnel from knowledge of which intervention a participant received. We considered that studies were at low risk of bias if they were blinded, or if we judged that the lack of blinding unlikely to affect results. We assessed blinding separately for different outcomes or classes of outcomes.

We assessed the methods as:

low, high or unclear risk of bias for participants;
low, high or unclear risk of bias for personnel.

(3.2) Blinding of outcome assessment (checking for possible detection bias)

We described for each included study the methods used, if any, to blinded outcome assessors from knowledge of which intervention a participant received. We assessed blinding separately for different outcomes or classes of outcomes.

We assessed methods used to blinded outcome assessment as:

low, high or unclear risk of bias.

(4) Incomplete outcome data (checking for possible attrition bias due to the amount, nature and handling of incomplete outcome data)

We described for each included study, and for each outcome or class of outcomes, the completeness of data including attrition and exclusions from the analysis. We stated whether attrition and exclusions were reported and the numbers included in the analysis at each stage (compared with the total randomised participants), reasons for attrition or exclusion where reported, and whether missing data were balanced across groups or were related to outcomes. Where sufficient information was reported, or could be supplied by the trial authors, we planned to re‐include missing data in the analyses which we undertook.

We assessed methods as:

low risk of bias (e.g. no missing outcome data; missing outcome data balanced across groups);
high risk of bias (e.g. numbers or reasons for missing data imbalanced across groups; ‘as treated’ analysis done with substantial departure of intervention received from that assigned at randomisation);
unclear risk of bias.

(5) Selective reporting (checking for reporting bias)

We described for each included study how we investigated the possibility of selective outcome reporting bias and what we found.

We assessed the methods as:

low risk of bias (where it is clear that all of the study’s pre‐specified outcomes and all expected outcomes of interest to the review have been reported);
high risk of bias (where not all the study’s pre‐specified outcomes have been reported; one or more reported primary outcomes were not pre‐specified; outcomes of interest are reported incompletely and so cannot be used; study fails to include results of a key outcome that would have been expected to have been reported);
unclear risk of bias.

(6) Other bias (checking for bias due to problems not covered by (1) to (5) above)

We described for each included study any important concerns we had about other possible sources of bias.

(7) Overall risk of bias

We made explicit judgements about whether studies were at high risk of bias, according to the criteria given in the Handbook (Higgins 2011). With reference to (1) to (6) above, we assessed the likely magnitude and direction of the bias and whether we considered it was likely to have impacted on the findings. We explored the impact of the level of bias through undertaking sensitivity analyses ‐ seeSensitivity analysis.

Assessment of the quality of the evidence using the GRADE approach

For this update we used the GRADE approach as outlined in the GRADE handbook, in order to assess the quality of the body of evidence relating to the following outcomes.

For the infant/child

Perinatal death, defined as intrauterine deaths plus neonatal deaths in the first week of life
Stillbirth
Admission to neonatal intensive care unit
Neonatal encephalopathy
Apgar score less than seven at five minutes
Neonatal trauma
Neurodevelopment at childhood follow‐up

For the mother

Mode of birth (caesarean section)
Operative vaginal birth (forceps or ventouse)
Perineal trauma
Postpartum haemorrhage (cut‐off reported by the trialists was used)
Breastfeeding at discharge
Postnatal depression
Length of maternal postnatal stay

We used the GRADEpro Guideline Development Tool to import data from Review Manager 5.3 (RevMan 2014), in order to create ’Summary of findings’ tables, comparing a policy of labour induction versus expectant management. A summary of the intervention effect and a measure of quality for each of the above outcomes was produced using the GRADE approach. The GRADE approach uses five considerations (study limitations, consistency of effect, imprecision, indirectness and publication bias) to assess the quality of the body of evidence for each outcome. The evidence can be downgraded from 'high quality' by one level for serious (or by two levels for very serious) limitations, depending on assessments for risk of bias, indirectness of evidence, serious inconsistency, imprecision of effect estimates or potential publication bias.

Measures of treatment effect

Dichotomous data

For dichotomous data, we presented results as summary risk ratio with 95% confidence intervals.

Continuous data

For continuous data, we used the mean difference. We planned to use the standardised mean difference to combine trials that measured the same outcome, but used different methods.

Unit of analysis issues

Cluster‐randomised trials

Cluster‐randomised trials were not eligible for inclusion.

Cross‐over trials

Cross‐over trials were not eligible for inclusion.

Multiple pregnancies

We did not identify any eligible studies that reported multiple pregnancies separately. If studies with multiple pregnancies are reported separately in trials included in future updates of this review, we will adjust for clustering in the analyses wherever possible, and use the inverse variance method for adjusted analyses, as described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011), and in Yelland 2011.

Multi‐armed trials

Where we included studies with multiple arms, we created single pair‐wise comparisons, by including only the groups relevant to this review, or by combining groups. In Gelisen 2005, we combined the three induction arms for the relevant analyses.

Dealing with missing data

For included studies, we noted levels of attrition. In future updates, if more eligible studies are included, we will explore the impact of including studies with high levels of missing data in the overall assessment of treatment effect by using sensitivity analyses.

For all outcomes, we carried out analyses, as far as possible, on an intention‐to‐treat basis, i.e. we attempted to include all participants randomised to each group in the analyses. The denominator for each outcome in each trial was the number randomised minus any participants whose outcomes were known to be missing.

Assessment of heterogeneity

We assessed statistical heterogeneity in each meta‐analysis using the Tau², I² and Chi² statistics. We regarded heterogeneity as substantial if an I² was greater than 30% and either the Tau² was greater than zero, or there was a low P value (less than 0.10) in the Chi² test for heterogeneity. Where we identified substantial heterogeneity (above 30%), we aimed to explore it using pre‐specified subgroup analyses.

Assessment of reporting biases

Where there were 10 or more studies in the meta‐analyses, we investigated reporting biases (such as publication bias) using funnel plots. We assessed funnel plot asymmetry visually. If asymmetry was suggested by a visual assessment, we planned to perform exploratory analyses to investigate it.

Data synthesis

We carried out statistical analysis using the Review Manager software (RevMan 2014). We used fixed‐effect meta‐analysis for combining data where it was reasonable to assume that studies were estimating the same underlying treatment effect: i.e. where studies were examining the same intervention, and the studies' populations and methods were judged sufficiently similar.

Where there was clinical heterogeneity sufficient to expect that the underlying treatment effects differed between trials, or where substantial statistical heterogeneity was detected, we used random‐effects meta‐analysis to produce an overall summary if an average treatment effect across trials was considered clinically meaningful. The random‐effects summary has been treated as the average of the range of possible treatment effects and we have discussed the clinical implications of treatment effects differing between trials. If the average treatment effect was not clinically meaningful, we decided that we would not combine trials. Where we used random‐effects analyses, the results have been presented as the average treatment effect with 95% confidence intervals, and the estimates of Tau² and I².

Subgroup analysis and investigation of heterogeneity

Where we identified substantial heterogeneity, we planned to investigate it using subgroup and sensitivity analyses. We planned to consider whether an overall summary was meaningful, and if it was, we used random‐effects analysis to produce it.

We carried out the following subgroup analyses.

Gestational age by week of gestation when induction was intended in the intervention arm.

In this update we have presented the main groups as close to this as study reporting would allow ‐ gestational ages ≤ 41 weeks, and > 41 completed weeks (> 287 days). In Brane 2014, the gestational age at induction in the intervention spanned 37 to 42 weeks.

State of the cervix (favourable versus unfavourable).

We were unable to conduct subgroup analyses by method of induction, due to wide variation in reporting of dosage, timing, frequency and mode of administration.

Where possible, we used the following outcomes in subgroup analyses.

For the infant/child

Perinatal death, defined as intrauterine deaths plus neonatal deaths in the first week of life
Stillbirth
Admission to neonatal intensive care unit

For the mother

Mode of birth (caesarean section)
Operative vaginal birth (forceps or ventouse)
Perineal trauma

We assessed subgroup differences by interaction tests available within RevMan (RevMan 2014). We reported the results of subgroup analyses quoting the Chi² statistic and P value, and the interaction test I² value.

Sensitivity analysis

We carried out sensitivity analyses to explore the effect of trial quality assessed by concealment of allocation, high attrition rates, or both, with poor quality studies being excluded from the analyses in order to assess whether this makes any difference to the overall result. Poor quality was defined as high risk of bias. We used the following outcomes in our sensitivity analyses.

For the infant/child

Perinatal death, defined as intrauterine deaths plus neonatal deaths in the first week of life
Stillbirth
Admission to neonatal intensive care unit

For the mother

Mode of birth (caesarean section)
Operative vaginal birth (forceps or ventouse)
Perineal trauma

Results

Description of studies

Results of the search

See Figure 1.

For this update, we assessed 26 new trial reports along with four studies (seven reports) which were previously excluded and one previously ongoing study. We have included eight additional trials (21 reports) (Brane 2014; Cohn 1992; Kortekaas 2014; Martin 1978; Miller 2015; Sande 1983; Tylleskar 1979; Walker 2016), excluded four studies (six reports) (Frass 2011; Gregson 2015; Neri 2014; Rijnders 2011), and added three ongoing studies (four reports) (Elden 2016; Othman 2017; Reddy 2013). Two studies are awaiting classification, one pending a translation (Benito Reyes 2010), and another awaiting further detail (Harrington 2003). We also added an additional earlier report to an already included study (Heimstad 2007a).

This updated review is now comprised of 30 included trials which randomised 12,479 women and their babies (seeCharacteristics of included studies), 64 excluded studies (seeCharacteristics of excluded studies) and three ongoing studies (seeCharacteristics of ongoing studies).

Included studies

Settings

Of the 30 included trials:

six were conducted in the USA (Dyson 1987; Martin 1989; Miller 2015; NICHHD 1994; Nielsen 2005; Witter 1987);
four in the UK (Scotland; England; Ireland) (Cole 1975; Henry 1969; Martin 1978; Walker 2016);
three in China (Bergsjo 1989; Cohn 1992; Roach 1997);
three in Norway (Augensen 1987; Heimstad 2007a; Sande 1983);
two in India (Chakravarti 2000; James 2001);
two in Thailand (Chanrachkul 2003; Herabutya 1992);
two in Sweden (Brane 2014; Tylleskar 1979);
one in Tunisia (Sahraoui 2005);
one in Turkey (Gelisen 2005);
one in Canada (Hannah 1992);
one in France (Breart 1982);
one in Austria (Egarter 1989);
one in Spain (Ocon 1997);
one in the Netherlands (Kortekaas 2014); and
one in Finland (Suikkari 1983).

Cervix status

Fifteen trials did not mention or specify cervix status as an inclusion criterion (Augensen 1987; Bergsjo 1989; Brane 2014; Breart 1982; Chakravarti 2000; Cohn 1992; Cole 1975; Heimstad 2007a; Henry 1969; James 2001; Martin 1978; Roach 1997; Suikkari 1983; Walker 2016; Witter 1987). Nine trials included women with unfavourable cervix (Dyson 1987; Gelisen 2005; Hannah 1992; Herabutya 1992; Martin 1989; Miller 2015; NICHHD 1994; Ocon 1997; Sahraoui 2005), and six with favourable cervical status (Chanrachkul 2003; Egarter 1989; Kortekaas 2014; Nielsen 2005; Sande 1983; Tylleskar 1979).

Interventions

All trials were conducted in hospitals with various intensities of fetal monitoring both in the induction and expectant management groups (seeCharacteristics of included studies).

Timing of induction ‐ induction group

The information on timing of induction in each trial's induction arm is summarised below.

37 to 39 weeks: one trial (Breart 1982), induced women at 37 to 39 weeks' gestation.
39 weeks: one trial (Martin 1978), induced women at 39 weeks' gestation.
39 to 40 weeks: four trials (Cole 1975; Miller 2015; Nielsen 2005; Walker 2016), induced women at 39 to 40 weeks' gestation.
40 weeks: two trials (Egarter 1989; Tylleskar 1979), induced women at their expected due date.
37 to 42 weeks: one trial (Brane 2014), induced women at 37⁰ to 41⁶ weeks' gestation.
40 to 41 weeks: one trial (Sande 1983), induced women between 40 and 41 weeks' gestation.
< 41 weeks: one trial (Chakravarti 2000), induced women at less than 41 weeks' gestation.
41 completed weeks: five trials reported that they induced women at 41 completed weeks (41⁰ or 287 days (Gelisen 2005; James 2001; Kortekaas 2014; Martin 1989); or 41³ or 290 days (Chanrachkul 2003)).
> 41 weeks: in the remaining 15 trials (Augensen 1987; Bergsjo 1989; Cohn 1992; Dyson 1987; Hannah 1992; Heimstad 2007a; Henry 1969; Herabutya 1992; Kortekaas 2014; NICHHD 1994; Ocon 1997; Roach 1997; Sahraoui 2005; Suikkari 1983; Witter 1987), women were generally induced after 41 completed weeks (after 287 days) up to 42 completed weeks (294 days), with Kortekaas 2014 spanning induction across 41^‐2+2 weeks and the NICHHD 1994 trial extending from 41 to 43 completed weeks (43⁰ or 301 days).

In this update of the review, we have collapsed these categories into:

induced at < 41 weeks: 10 trials (Breart 1982; Chakravarti 2000; Cole 1975; Egarter 1989; Martin 1978; Miller 2015; Nielsen 2005; Sande 1983; Tylleskar 1979; Walker 2016);
induced at ≥ 41 weeks: 19 trials (Augensen 1987; Bergsjo 1989; Chanrachkul 2003; Cohn 1992; Dyson 1987; Gelisen 2005; Hannah 1992; Heimstad 2007a; Henry 1969; Herabutya 1992; James 2001; Kortekaas 2014; Martin 1989; NICHHD 1994; Ocon 1997; Roach 1997; Sahraoui 2005; Suikkari 1983; Witter 1987);
inductions spanning 37 to 42 weeks: one trial (Brane 2014).

(In the previous version of this review, we grouped studies into the following three categories: 39 to 40 weeks; 41 weeks; and > 41 weeks.)

In some trials, the actual gestational age at induction in the induction groups may have been slightly later than the gestational threshold specified at trial entry (e.g. Hannah 1992).

See Characteristics of included studies table for further details.

Method of induction ‐ induction group

Labour induction was by oxytocin with or without artificial rupture of membranes in most trials. In trials recruiting women with an unfavourable cervix, priming with prostaglandins or laminaria was often undertaken before induction.

Of the 30 included trials:

two trials did not report the method used (Chakravarti 2000; Cohn 1992);
23 trials used oxytocin infusion in some or all women in their intervention group (Augensen 1987; Bergsjo 1989; Brane 2014; Breart 1982; Chanrachkul 2003; Cole 1975; Dyson 1987; Gelisen 2005; Hannah 1992; Heimstad 2007a; Henry 1969; Herabutya 1992; James 2001; Kortekaas 2014; Martin 1989; Miller 2015; NICHHD 1994; Nielsen 2005; Sande 1983; Suikkari 1983; Tylleskar 1979; Walker 2016; Witter 1987). Of those trials, only one used oxytocin as the sole method of induction (Augensen 1987). Seventeen trials used artificial rupture of membranes (AROM), as well as oxytocin infusion (when possible) (Bergsjo 1989; Brane 2014; Breart 1982; Chanrachkul 2003; Cole 1975; Heimstad 2007a; Henry 1969; Herabutya 1992; James 2001; Kortekaas 2014; Miller 2015; Nielsen 2005; Sande 1983; Suikkari 1983; Tylleskar 1979; Walker 2016; Witter 1987);
none of the included trials used AROM as the sole method of induction;
10 trials used intravaginal prostaglandin E2 for some or all women in the intervention group (in either gel or pessary form) (Brane 2014; Dyson 1987; Egarter 1989; Hannah 1992; Herabutya 1992; NICHHD 1994; Ocon 1997; Roach 1997; Sahraoui 2005; Walker 2016). Four trials used prostaglandin E2 as the sole method of induction (Egarter 1989; Ocon 1997; Roach 1997; Sahraoui 2005) and six trials used a combination of prostaglandin and oxytocin +/‐ AROM (Brane 2014; Dyson 1987; Hannah 1992; Herabutya 1992; NICHHD 1994; Walker 2016);
three trials used vaginal misoprostol in some or all women in the intervention group (Gelisen 2005; Heimstad 2007a; Miller 2015);
two trials had more than one intervention group (Gelisen 2005; NICHHD 1994), although the placebo priming and oxytocin arm in NICHHD 1994 was not included in this review. The Gelisen 2005 trial had three labour induction arms with misoprostol, oxytocin and Foley catheter.

Expectant management group protocols

For the majority of trials, expectant management protocols included various combinations of fetal heart rate monitoring, ultrasound for amniotic fluid measurements and, in earlier studies, biochemical tests.

No gestational age limit for induction was imposed or reported in nine of the trials (Brane 2014; Cohn 1992; Dyson 1987; Henry 1969; James 2001; Ocon 1997; Roach 1997; Suikkari 1983; Witter 1987). In the remaining 21 trials, women were induced at the following times (unless they went into spontaneous labour earlier) in the expectant management groups.

41 weeks (Cole 1975).
41 to 42 weeks (Walker 2016).
42 weeks (Breart 1982; Chakravarti 2000; Egarter 1989; Gelisen 2005; Kortekaas 2014; Martin 1978; Miller 2015; Nielsen 2005; Sahraoui 2005; Sande 1983; Tylleskar 1979).
42 to 43 weeks (Augensen 1987; Heimstad 2007a).
43 weeks (Bergsjo 1989; Martin 1989).
44 weeks (Chanrachkul 2003; Hannah 1992; Herabutya 1992; NICHHD 1994).

See Characteristics of included studies for further details.

Outcomes

The primary outcome of perinatal death was reported in 20 of the 30 included trials. Caesarean section was reported in 27 trials. Other outcomes, such as many of the adverse pregnancy and neonatal outcomes, were reported in less than half of the included trials. Only two trials reported on maternal satisfaction and no trials have yet reported on maternal anxiety or depression, or breastfeeding.

Funding

Nine of the 30 included trials reported their funding sources as follows: Karolinska Institute Foundations and Funds (Brane 2014); Ramathibodi Hospital Research Grants (Chanrachkul 2003; Herabutya 1992); Community Service Program of Kaiser Foundation Hospitals (Dyson 1987); Medical Research Council of Canada (Hannah 1992); and Upjohn Company of Canada supplied the prostaglandin gel for this study; ZonMW (The Netherlands Organisation for Health Research and Development) (Kortekaas 2014); Vicksburg Hospital Medical Foundation (Martin 1989); National Institute of Child Health and Human Development, NIH, USA (NICHHD 1994); and one grant from the Research for Patient Benefit Programme of the National Institute for Health Research (Walker 2016).

Twenty‐one of the 30 included trials did not report their funding sources.

Declarations of interest

Walker 2016 declared relevant interests on the part of one author, Dr Smith: "Dr. Smith reports receiving fees for serving on an advisory board from Roche Diagnostics, consulting fees from GlaxoSmithKline, equipment loans from Roche Diagnostics and General Electric, travel support from Roche Diagnostics and Chiesi, and grant support from GlaxoSmithKline and Action Medical Research, and being named as an inventor on a pending patent (PCT/EP2014/062602) filed by GlaxoSmithKline related to retosiban as a preventive treatment for preterm labor in women with increased uterine stretch". No other potential conflict of interest was reported by the authors of this study.

Kortekaas 2014 declared that they had no conflicts of interest, and Miller 2015 declared that they had no financial conflicts of interest.

Of the remaining 27 studies, none included any declarations of interest.

Excluded studies

Most of the excluded trials were comparisons of different labour induction (17 trials: Ascher‐Walsh 2000; de Aquino 2003; Evans 1983; Kipikasa 2005; Lee 1997; Lemancewicz 1999; Magann 1999; Mancuso 1998; Meydanli 2003; Misra 1994; Müller 1995; Papageorgiou 1992; Rijnders 2011; Satin 1991; Stenlund 1999; Su 1996; Surbek 1997) or cervical ripening protocols (28 trials: Bell 1993; Berghella 1996; Boulvain 1998; Buttino 1990; Damania 1992; Dare 2002; Doany 1997; Elliott 1984; El‐Torkey 1992; Giacalone 1998; Hage 1993; Ingemarsson 1987; Jenssen 1977; Kadar 1990; Klopper 1969; Lien 1998; Lyons 2001; Magann 1998; Newman 1997; Rayburn 1988; Rayburn 1999; Roberts 1986; Sawai 1991; Sawai 1994; Williams 1990; Wing 2000; Wong 2002; Ziaei 2003). Nine studies were not randomised trials (Amano 1999; Cardozo 1986; Garry 2000; Heden 1991; Hernandez‐Castro 2008; Iqbal 2004; Katz 1983; Knox 1979; Ohel 1996) and there were variety of reasons for excluding the remaining nine studies (Alcalay 1996; Conway 2000; Dunn 1989; Frass 2011; Gregson 2015; Imsuwan 1999; Neri 2014; Nicholson 2008; Paul 1988). More details are provided in the Characteristics of excluded studies tables.

Risk of bias in included studies

Three trials (Chakravarti 2000; Cohn 1992; Suikkari 1983), were available only as abstracts and despite extensive searches we could not locate full publications of the studies, which limited our assessment of their risk of bias. Another trial (Kortekaas 2014), is available only in abstract form, with full publication planned.

We judged the majority of included trials to be at moderate risk of bias (Figure 2; Figure 3), largely due to a lack of reporting.

'Risk of bias' graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

'Risk of bias' summary: review authors' judgements about each risk of bias item for each included study.

Allocation

We judged 14 trials (Augensen 1987; Bergsjo 1989; Chanrachkul 2003; Dyson 1987; Heimstad 2007a; James 2001; Kortekaas 2014; Martin 1978; Miller 2015; NICHHD 1994; Nielsen 2005; Sahraoui 2005; Walker 2016; Witter 1987), to be at low risk of selection bias, reporting some form of adequate random sequencing such as a computer‐generated sequence or a list of random numbers. We judged the remaining 16 trials to be at unclear risk of selection bias, as they did not report how a random sequence was generated (Brane 2014; Breart 1982; Chakravarti 2000; Cohn 1992; Cole 1975; Egarter 1989; Gelisen 2005; Hannah 1992; Henry 1969; Herabutya 1992; Martin 1989; Ocon 1997; Roach 1997; Sande 1983; Suikkari 1983; Tylleskar 1979).

Of the 29 included trials, only seven reported a method of allocation concealment likely to have a low risk of bias ‐ either central randomisation or sequentially numbered sealed opaque envelopes (Hannah 1992; Heimstad 2007a; Kortekaas 2014; Miller 2015; NICHHD 1994; Nielsen 2005; Walker 2016). Eight trials reported that they used an envelope system with an unclear risk of bias (Brane 2014; Breart 1982; Dyson 1987; Gelisen 2005; James 2001; Martin 1989; Roach 1997; Witter 1987), one trial reported a partial third party system also with unclear risk of bias (Augensen 1987), and 14 trials did not report a method for concealing allocation and were rated as being at unclear risk of bias (Bergsjo 1989; Chakravarti 2000; Chanrachkul 2003; Cohn 1992; Cole 1975; Egarter 1989; Henry 1969; Herabutya 1992; Martin 1978; Ocon 1997; Sahraoui 2005; Sande 1983; Suikkari 1983; Tylleskar 1979).

Blinding

Performance bias

Given the nature of the intervention (induction of labour) and comparison (expectant management), it was not possible for women or clinicians to be blinded to the treatment group in any of the 29 trials, and thus risk of performance bias was judged to be high. For the more objective outcomes such as perinatal death, this lack of blinding is unlikely to be a major source of bias.

Detection bias

It would have been possible for outcome assessment to have been undertaken by someone blinded to allocation groups. However, only two studies reported whether or not outcome assessment was blinded. One study indicated partial blinding of outcome assessment (Hannah 1992), with an adjudication of abnormal neonatal outcomes undertaken by a neonatologist who was unaware of the mothers' group assignments (rated unclear risk of bias). A further trial (Martin 1978) reported blinded outcome assessment (rated as low risk of bias). The remaining 27 trials did not detail whether outcome assessment was to be blinded, and thus we judged risk of detection bias to be unclear.

Measurement of outcomes such as perinatal death should not be biased by lack of blinding.

Incomplete outcome data

We considered the majority of trials (21) (Augensen 1987; Bergsjo 1989; Breart 1982; Chanrachkul 2003; Dyson 1987; Gelisen 2005; Hannah 1992; Heimstad 2007a; Henry 1969; Herabutya 1992; James 2001; Kortekaas 2014; Martin 1989; Miller 2015; NICHHD 1994; Nielsen 2005; Ocon 1997; Roach 1997; Sahraoui 2005; Walker 2016; Witter 1987) to be at low risk of attrition bias, with minimal/no losses to follow‐up or exclusions. We judged six trials to be at unclear risk of attrition bias (Brane 2014; Chakravarti 2000; Cohn 1992; Cole 1975; Egarter 1989; Suikkari 1983), commonly due to some post‐randomisation exclusions and/or missing data, or due to insufficient information to determine losses or exclusions (due to publication in abstract form only).

We judged three trials to be at high risk of attrition bias. In both Martin 1978 and Tylleskar 1979, between 25% and 30% of the women randomised were excluded post‐randomisation due to going into labour prior to their planned date of induction (for women in the induction group), due to obstetric abnormalities or failure to go into spontaneous labour before 42 weeks (women in the expectant management group of Martin 1978), going into labour prior to their expected delivery date (women in the expectant management group of Tylleskar 1979). In Sande 1983, a per protocol analysis was performed, whereby women were not analysed in the group to which they were randomised, rather according to whether they had their labour induced, or delivered spontaneously.

Selective reporting

Only one trial (Walker 2016) was judged to be at low risk of reporting bias, with outcomes reported as pre‐specified in the published protocol. We judged 23 trials (Brane 2014; Breart 1982; Chakravarti 2000; Chanrachkul 2003; Cohn 1992; Cole 1975; Dyson 1987; Gelisen 2005; Hannah 1992; Heimstad 2007a; Henry 1969; Herabutya 1992; James 2001; Kortekaas 2014; Martin 1978; Miller 2015; NICHHD 1994; Nielsen 2005; Ocon 1997; Roach 1997; Sahraoui 2005; Suikkari 1983; Witter 1987) to be at unclear risk of reporting bias, largely due to insufficient information to assess selective reporting (i.e. no access to trial protocols and limited detail reported in manuscript methods). We considered six trials (Augensen 1987; Bergsjo 1989; Egarter 1989; Martin 1989; Sande 1983; Tylleskar 1979), to be at high risk of reporting bias, predominately due to the incomplete reporting of outcomes data (such as in text or figures only, with statements such as "no significant difference between groups" made) such that outcome data could not be included in review meta‐analyses.

Other potential sources of bias

Most of the trials (21/30) (Augensen 1987; Bergsjo 1989; Brane 2014; Breart 1982; Chanrachkul 2003; Cole 1975; Dyson 1987; Gelisen 2005; Hannah 1992; Heimstad 2007a; Henry 1969; Herabutya 1992; James 2001; Martin 1989; Miller 2015; NICHHD 1994; Ocon 1997; Roach 1997; Sahraoui 2005; Walker 2016; Witter 1987), appeared to be free of other potential sources of bias. We judged the other nine trials to be at unclear risk of bias, six trials due to limited reporting (abstract only or limited methodological detail provided) (Chakravarti 2000; Cohn 1992; Kortekaas 2014; Martin 1978; Suikkari 1983; Tylleskar 1979); and one trial each for imbalance in the numbers of women randomised to each group (Egarter 1989), baseline characteristic imbalance (Nielsen 2005), or lack of reporting of baseline characteristics (Sande 1983).

Effects of interventions

See: Table 1; Table 2

Summary of findings for the main comparison. Labour induction versus expectant management (infant/child outcomes).

Induction of labour for improving birth outcomes for women at or beyond term
Population: pregnant women at or beyond term  Setting: Norway, China, Thailand, the USA, Austria, Turkey, Canada, UK, India, Tunisia, Finland, Spain, Sweden and the Netherlands  Intervention: labour induction  Comparison: expectant management
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect  (95% CI)	№ of participants  (studies)	Quality of the evidence  (GRADE)	Comments
	Risk with expectant management	Risk with labour induction
Perinatal death	Study population		RR 0.33  (0.14 to 0.78)	9960  (20 RCTs)	⊕⊕⊕⊝  MODERATE¹
	3 per 1000	1 per 1000  (0 to 3)
Stillbirth	Study population		RR 0.33  (0.11 to 0.96)	9960  (20 RCTs)	⊕⊕⊕⊝  MODERATE¹
	2 per 1000	1 per 1000  (0 to 2)
Admission to neonatal intensive care unit	Study population		RR 0.88  (0.77 to 1.01)	8531  (13 RCTs)	⊕⊕⊕⊝  MODERATE¹
	85 per 1000	75 per 1000  (60 to 86)
Neonatal encephalopathy	Study population		‐	(0 RCTs)	‐	No RCTs reported data for this outcome.
	see comment	see comment
Apgar score less than 7 at 5 minutes	Study population		RR 0.70  (0.50 to 0.98)	9047  (16 RCTs)	⊕⊕⊕⊝  MODERATE¹
	17 per 1000	12 per 1000  (7 to 17)
Neonatal trauma	Study population		RR 1.18  (0.68 to 2.05)	4255  (3 RCTs)	⊕⊕⊝⊝  LOW^1,2
	10 per 1000	12 per 1000  (7 to 21)
Neurodevelopment at childhood follow‐up	Study population		‐	(0 RCTs)	‐	No RCTs reported data for this outcome.
	see comment	see comment
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).  CI: confidence interval; RCT: randomised controlled trial; RR: risk ratio
GRADE Working Group grades of evidence  High quality: we are very confident that the true effect lies close to that of the estimate of the effect  Moderate quality: we are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different  Low quality: our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect  Very low quality: we have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect

Induction of labour for improving birth outcomes for women at or beyond term
Population: women at or beyond term  Setting: Norway, China, Thailand, the USA, Austria, Turkey, Canada, UK, India, Tunisia, Finland, Spain, Sweden, France and the Netherlands  Intervention: labour induction  Comparison: expectant management
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect  (95% CI)	№ of participants  (studies)	Quality of the evidence  (GRADE)	Comments
	Risk with expectant management	Risk with labour induction
Caesarean section	Study population		RR 0.92  (0.85 to 0.99)	11,738  (27 RCTs)	⊕⊕⊕⊝  MODERATE¹
	184 per 1000	169 per 1000  (157 to 182)
Operative vaginal birth (forceps or ventouse)	Study population		RR 1.07  (0.99 to 1.16)	9281  (18 RCTs)	⊕⊕⊕⊝  MODERATE¹
	193 per 1000	206 per 1000  (191 to 223)
Perineal trauma	Study population		RR 1.09  (0.65 to 1.83)	3028  (4 RCTs)	⊕⊕⊝⊝  LOW^1,2
	17 per 1000	18 per 1000  (11 to 31)
Postpartum haemorrhage	Study population		RR 1.09  (0.92 to 1.30)	3315  (5 RCTs)	⊕⊕⊝⊝  LOW^1,2
	122 per 1000	133 per 1000  (112 to 159
Breastfeeding at discharge	Study population		‐	(0 RCTs)	‐	No RCTs reported data for this outcome.
	see comment	see comment
Postnatal depression	Study population		‐	(0 RCTs)	‐	No RCTs reported data for this outcome.
	see comment	see comment
Length of maternal hospital stay (days)	‐	‐	Average MD 0.34 days shorter for women who were induced  (1 day shorter to 0.33 days longer)	1146  (5 RCTs)	⊕⊝⊝⊝  VERY LOW^1,2,3
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).  CI: confidence interval; MD: mean difference; RCT: randomised controlled trial; RR: risk ratio
GRADE Working Group grades of evidence  High quality: we are very confident that the true effect lies close to that of the estimate of the effect  Moderate quality: we are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different  Low quality: our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect  Very low quality: we have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect

Study	Cause of death
Study	Intervention Group	Control Group
Augensen 1987	No deaths	No deaths
Bergsjo 1989	1. Severe malformations (Livebirth) GA at birth and timing of death after birth not reported	1. Malformation (Livebirth) GA at birth and timing of death after birth not reported 2. Pneumonia (Livebirth) GA at birth and timing of death after birth not reported
Chanrachkul 2003	No deaths	No deaths
Cole 1975	No deaths	1. Congenital heart condition (Stillbirth) GA at detection of death not reported
Dyson 1987	No deaths	1. Meconium aspiration and persistent fetal circulation (Livebirth) GA at birth was 43 + 4 and the timing of death after birth was not reported
Egarter 1989	No deaths	1. Cord complication (Stillbirth) GA at detection of fetal death was 40 + 3 weeks
Gelisen 2005	No deaths	1. Intrauterine fetal death (Stillbirth) GA at death 41 + 5 weeks
Hannah 1992	No deaths	1. Hypoxic ischaemic encephalopathy (Stillbirth) GA at detection of death not reported 2. Massive aspiration of meconium (Stillbirth) GA at detection of death not reported
Heimstad 2007a	No deaths	1. Birth asphyxia secondary to a true knot in the umbilical cord (Livebirth) Birth at 294 days GA; death at 2 days of age
Henry 1969	No deaths	1. Stillbirth in a patient with an abnormal glucose tolerance test (Stillbirth) GA at detection of death not reported 2. Neonatal death from meconium inhalation in a woman with a positive amnioscopy who refused surgical induction of labour (Livebirth) GA at detection of death not reported
Herabutya 1992	No deaths	1. Congenital abnormality (Livebirth) Birth at 43 weeks; death at 3 days of age
James 2001	No deaths	No deaths
Kortekaas 2014	One fetal death (no further details reported)	2 fetal deaths (no further details reported)
Martin 1978	No deaths reported	1. Stillbirth (Stillbirth) Stillbirth after induction of labour at 42 weeks for postmaturity and meconium
Martin 1989	No deaths	No deaths
NICHHD 1994	No deaths	No deaths
Sahraoui 2005	No deaths	1. Intrauterine fetal death (Stillbirth) Death detected at 42 weeks' GA
Sande 1983	No deaths	No deaths
Suikkari 1983	No deaths	No deaths
Walker 2016	No deaths	No deaths

Date	Event	Description
28 December 2017	New search has been performed	Search updated and eight additional trials included (Brane 2014; Cohn 1992; Kortekaas 2014; Martin 1978; Miller 2015; Sande 1983; Tylleskar 1979; Walker 2016). We have updated the methods in line with the standard methods used by Cochrane Pregnancy and Childbirth and we now use GRADE to assess the quality of the body of evidence. For this update, the overall conclusions have not changed. However, there is moderate certainty evidence to suggest that induction was associated with fewer stillbirths and fewer babies with low Apgar scores.
9 October 2017	New citation required but conclusions have not changed	Conclusions not changed.

Date	Event	Description
31 March 2012	New search has been performed	Search updated ‐ no new trials identified. Trial reports that were previously awaiting classification have now been incorporated into the review. We have added three new included trials (Heimstad 2007a; Nielsen 2005; Sahraoui 2005), three new excluded trials (Hernandez‐Castro 2008; Imsuwan 1999; Nicholson 2008) and one ongoing trial (Rijnders 2007). This updated review is now comprised of 22 included studies (reporting on 9383 women); 64 excluded studies and one ongoing study. Results are now presented as 37‐39 weeks; 39‐40 weeks; < 41 weeks, 41 weeks and > 41 weeks. A new author joined the team to help prepare this update.
31 March 2012	New citation required and conclusions have changed	Whilst the overall conclusions have not changed, there is now evidence to show that induction of labour at or beyond term is associated with a lower rate of caesarean section.
6 July 2011	Feedback has been incorporated	Feedback from Amy Marowitz added.
6 July 2011	Amended	Error corrected in response to feedback from Amy Marowitz (Feedback).
14 July 2009	Amended	Search updated. Eight reports of five trials added to Studies awaiting classification (Heimstad 2007a; Hernandez‐Castro 2008a; Imsuwan 1999a; Nicholson 2008a; Rijnders 2007a).
3 September 2008	Amended	Converted to new review format.
28 February 2007	Amended	The Implications for research section has been amended to include the uncertainty about timing of labour induction beyond term, which was unintentionally left out during the revision process.
21 August 2006	New citation required but conclusions have not changed	This version has been re‐written, including a new protocol which now limits the scope to labour induction
30 June 2006	New search has been performed	The previous version of this review included studies up to 1997 and included 21 labour induction trials (Gülmezoglu 2006). This version has been re‐written, including a new protocol which now limits the scope to labour induction, and includes 19 trials. Thirteen of the 21 trials included in the previous version are included in this version. The remaining eight trials were excluded because of alternate allocation (Cardozo 1986; Heden 1991; Katz 1983), a high proportion of postrandomization exclusion (greater than 30% in Martin 1978a and greater than 24% in Tylleskar 1979a), cervical ripening with breast stimulation (Elliott 1984; Kadar 1990), and analysis by intervention received (i.e. groups switched, Sande 1983a). Six trials published since the publication of the previous version have been included in this update (Chakravarti 2000; Chanrachkul 2003; Gelisen 2005; James 2001; Ocon 1997; Roach 1997).

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Perinatal death	20	9960	Risk Ratio (M‐H, Fixed, 95% CI)	0.33 [0.14, 0.78]
2 Stillbirth	20	9960	Risk Ratio (M‐H, Fixed, 95% CI)	0.33 [0.11, 0.96]
3 Neonatal death	19	9776	Risk Ratio (M‐H, Fixed, 95% CI)	0.37 [0.10, 1.38]
4 Birth asphyxia	4	1456	Risk Ratio (M‐H, Fixed, 95% CI)	1.66 [0.61, 4.55]
5 Admission to neonatal intensive care unit	13	8531	Risk Ratio (M‐H, Fixed, 95% CI)	0.88 [0.77, 1.01]
6 Neonatal convulsions	3	4365	Risk Ratio (M‐H, Fixed, 95% CI)	0.54 [0.15, 1.97]
7 Use of anticonvulsants	1	349	Risk Ratio (M‐H, Fixed, 95% CI)	0.34 [0.01, 8.17]
8 Meconium aspiration syndrome	11	7781	Risk Ratio (M‐H, Fixed, 95% CI)	0.77 [0.62, 0.96]
9 Apgar score less than 7 at 5 minutes	16	9047	Risk Ratio (M‐H, Fixed, 95% CI)	0.70 [0.50, 0.98]
10 Birthweight (g)	14	3799	Mean Difference (IV, Fixed, 95% CI)	‐69.43 [‐96.83, ‐42.02]
11 Birthweight > 4000 g	8	5593	Risk Ratio (M‐H, Random, 95% CI)	0.72 [0.54, 0.96]
12 Neonatal trauma	3	4255	Risk Ratio (M‐H, Fixed, 95% CI)	1.18 [0.68, 2.05]
13 Caesarean section	27	11738	Risk Ratio (M‐H, Fixed, 95% CI)	0.92 [0.85, 0.99]
14 Operative vaginal birth (forceps or ventouse)	18	9281	Risk Ratio (M‐H, Fixed, 95% CI)	1.07 [0.99, 1.16]
15 Analgesia used	9		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
16 Perineal trauma	4	3028	Risk Ratio (M‐H, Fixed, 95% CI)	1.09 [0.65, 1.83]
17 Prolonged labour	3		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
17.1 First stage	1	508	Risk Ratio (M‐H, Fixed, 95% CI)	0.55 [0.20, 1.45]
17.2 Second stage	1	508	Risk Ratio (M‐H, Fixed, 95% CI)	0.67 [0.36, 1.22]
17.3 Third stage	1	249	Risk Ratio (M‐H, Fixed, 95% CI)	3.02 [0.12, 73.52]
17.4 No definition	1	112	Risk Ratio (M‐H, Fixed, 95% CI)	0.35 [0.01, 8.30]
18 Postpartum haemorrhage	5	3315	Risk Ratio (M‐H, Fixed, 95% CI)	1.09 [0.92, 1.30]
19 Maternal satisfaction	2		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
19.1 Hoping to be randomised to the same trial arm as they had been in this study	1	496	Risk Ratio (M‐H, Fixed, 95% CI)	1.93 [1.62, 2.30]
19.2 Preferred their allocation	1	184	Risk Ratio (M‐H, Fixed, 95% CI)	0.90 [0.72, 1.13]
20 Length of maternal hospital stay (days)	5	1146	Mean Difference (IV, Random, 95% CI)	‐0.34 [‐1.00, 0.33]
21 Length of neonatal hospital stay (days)	1	302	Mean Difference (IV, Fixed, 95% CI)	‐0.30 [‐0.61, 0.01]
22 Length of labour (hours)	9	1980	Mean Difference (IV, Random, 95% CI)	‐1.01 [‐1.72, ‐0.31]

Methods	RCT
Participants	Number of women randomised: 409  Setting: Bergen, Norway Inclusion criteria Healthy women Normal pregnancy Singleton Cephalic presentation Duration of pregnancy 290‐297 days from the first day of the LMP Reliable dates GA for intervention: 41+ weeks (290‐297 days) Exclusion criteria Use of contraceptive pills during the 2 months before the LMP State of cervix: mixed (about 35% in each group had unripe cervix)
Interventions	Induction group (n = 214): immediate induction with oxytocin (5 IU increased in a stepwise manner). GA at intervention 41+ weeks (290‐297 days) versus Expectant management group (n = 195): NST every 3‐4 days, IOL after 7 days
Outcomes	Mother: caesarean section; assisted vaginal birth; length of labour; length of hospital stay Baby: perinatal death; birthweight; neonatal jaundice; meconium‐stained amniotic fluid; NICU admission
Notes	Funding: not reported Declarations of interest: not reported
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	List of random numbers
Allocation concealment (selection bias)	Unclear risk	Allocation concealment was unclear given that it was not undertaken by a staff member or team clearly uninvolved in the trial. It was reported that the midwife undertook allocation using a random number list, and this list was inaccessible to the participating physicians.
Blinding of participants and personnel (performance bias)   All outcomes	High risk	Blinding was not feasible.
Blinding of outcome assessment (detection bias)   All outcomes	Unclear risk	Blinded outcome assessment was not mentioned.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	4/214 in the IOL group went into labour before IOL but data for these women have been included in the IOL group for analyses. No apparent losses to follow‐up or exclusions.
Selective reporting (reporting bias)	High risk	No outcomes were pre‐specified in the methods; some outcomes reported incompletely in text, e.g. "There was no significant difference between the groups in the use of analgesia, sedatives, and epidural anaesthesia."
Other bias	Low risk	Appears to be free of other bias

Methods	RCT
Participants	Number of women randomised: 188 Setting: Wuhan, Hubei province, China Inclusion criteria All parities Not in labour Intact membranes Normal pregnancy No significant risk factors GA for intervention: 42 completed weeks (294 days) Exclusion criteria No additional criteria State of cervix: not mentioned
Interventions	Induction group (n = 94): stripping of membranes followed by oxytocin infusion and AROM if cervix sufficiently dilated. GA for intervention: 42 completed weeks (294 days) versus  Expectant management group (n = 94): no intervention for 1 week, IOL at 43 weeks.
Outcomes	Mother:operative vaginal birth; duration of labour; caesarean section; breastfeeding (timing of recording of this outcome in relation to birth or discharge time was not specified) Baby: perinatal death; meconium aspiration syndrome
Notes	Funding: not reported Declarations of interest: not reported
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	List of random numbers
Allocation concealment (selection bias)	Unclear risk	The method of allocation concealment was not reported.
Blinding of participants and personnel (performance bias)   All outcomes	High risk	Appears that blinding was not feasible.
Blinding of outcome assessment (detection bias)   All outcomes	Unclear risk	Blinded outcome assessment was not mentioned.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	8/94 in IOL group went into labour before IOL but were kept in the allocated group. No apparent losses to follow‐up or exclusions.
Selective reporting (reporting bias)	High risk	Most pre‐specified outcomes were reported; however, limited information was provided for some outcomes (e.g. combined maternal complications) and neonatal outcomes, e.g. "Maternal complications, including protracted labor, cervical edema, cervical laceration, post‐partum hemorrhage and unspecified post‐partum morbidity accounted to about 15% in both groups, with no significant differences;" and "About 90 mothers in each group were breastfeeding."
Other bias	Low risk	Appears to be free of other bias.

Methods	RCT
Participants	Number of women randomised: 138  Setting: Stockholm, Sweden Inclusion criteria Nulliparous low‐risk women Normal pregnancies With contractions Singleton fetus with cephalic presentation GA: between 37⁺⁰ and 41⁺⁶ confirmed by ultrasound < 20 weeks Cervical dilation < 4 cm and intact membranes Able to speak, read and understand Swedish Exclusion criteria Not stated State of cervix: Bishop score at presentation ranged from 1‐8 in induction group and 2‐9 in expectant management group.
Interventions	Induction group (n = 71): 5 hours after medication to promote 'therapeutic rest' IOL performed, with method dependent on state of cervix: intravaginal PGE2 or transcervical catheter +/‐ AROM if cervical dilation permitted; followed by IV oxytocin (augmented every 20‐30 min) if no progress after AROM. versus Expectant management group (n = 67): spontaneous labour awaited as long as possible; IOL if women wanted, or if the obstetrician/midwife considered suitable.
Outcomes	Mother: mode of birth; experience of birth; duration of labour; labour analgesia; oxytocin for augmentation; birth presentation; postpartum haemorrhage; sphincter tears. Baby: Apgar score < 7 at 5 min; cord artery metabolic acidosis; birthweight; head circumference; admission to NICU.
Notes	Women in both groups given medication to promote 'therapeutic rest' (1 g paracetamol, 10 mg zolpidem, 10 mg morphine). During active phase of labour (cervical dilation ≥ 4 cm or ROM) women monitored according to local protocol; slow progress (arrest of dilation for 203 hours) was treated with AROM or oxytocin. Funding: Karolinska Institute Foundations and Funds Declaration of interests: the authors report no conflicts of interest
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	"The randomization was performed in blocks of 5–10 in each group." Method of sequence generation not reported.
Allocation concealment (selection bias)	Unclear risk	"A sealed envelope containing coded protocols for the respective groups… was opened by the midwife." No further detail provided.
Blinding of participants and personnel (performance bias)   All outcomes	High risk	Blinding was not feasible.
Blinding of outcome assessment (detection bias)   All outcomes	Unclear risk	Blinded outcome assessment was not mentioned.
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	Some attrition and missing data particularly for the women’s views questionnaire. Most of the sample were included in the analyses for the primary outcome (65/71, and 64/67 in main analysis for mode of birth). The reasons for missing data for some clinical outcomes were not reported.
Selective reporting (reporting bias)	Unclear risk	No access to trial protocol to confidently assess selective reporting. Perinatal death not reported.
Other bias	Low risk	Appears to be free of other bias.

Methods	RCT (1:2 randomisation)
Participants	Number of women randomised: 716  Setting: Paris, France Inclusion criteria: GA: 37‐39 weeks Exclusion criteria: high risk, contraindication for IOL  State of cervix: not mentioned
Interventions	Induction group (n = 235): oxytocin and AROM at GA 37‐39 weeks. versus  Expectant management group (n = 481): FHR checking and amnioscopy every 2‐3 days.
Outcomes	Mother: duration of labour; mode of birth  Baby: morbidity (Apgar scores, resuscitation)
Notes	Funding: not reported Declaration of interests: not reported
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	The method of sequence generation was not reported.
Allocation concealment (selection bias)	Unclear risk	It was reported that a closed envelope system was used for allocation concealment, although no further detail was available.
Blinding of participants and personnel (performance bias)   All outcomes	High risk	Blinding was not feasible.
Blinding of outcome assessment (detection bias)   All outcomes	Unclear risk	Blinded outcome assessment was not mentioned.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	202/235 in the induction group and 173/481 in the expectant group followed the trial protocol; trial results were reported for all 716 women and their babies.
Selective reporting (reporting bias)	Unclear risk	No access to trial protocol to confidently assess selective reporting. Perinatal death was not reported.
Other bias	Low risk	Appears to be free of other bias.

Methods	RCT
Participants	Number of women randomised: 231  Setting: Calcutta, India Inclusion criteria Primips Low risk Uncomplicated pregnancy Confirmed dates GA: reported as "before 41 completed weeks" State of cervix: not mentioned
Interventions	Induction group (n = 117): IOL, no details of the method are available.  versus Expectant management group (n = 114 randomised): daily fetal movement counts, biophysical profile and ultrasound; IOL after 1 week.
Outcomes	Only caesarean section rates were adequately reported in the abstract.
Notes	Reported as conference abstract. Only data for caesarean included in meta‐analysis. Funding: not reported Declaration of interests: not reported
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	The method of sequence generation was not reported.
Allocation concealment (selection bias)	Unclear risk	The method of allocation concealment was not reported.
Blinding of participants and personnel (performance bias)   All outcomes	High risk	Appears that blinding was not feasible.
Blinding of outcome assessment (detection bias)   All outcomes	Unclear risk	Blinded outcome assessment was not mentioned.
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	Insufficient information to determine.
Selective reporting (reporting bias)	Unclear risk	No outcomes were pre‐specified in the methods (conference abstract). Insufficient information to determine.
Other bias	Unclear risk	Unable to identify other bias based on the abstract.

Methods	RCT
Participants	Number of women randomised: 249  Setting: Bangkok, Thailand Inclusion criteria Low‐risk No obstetric or medical complication GA: 41³ weeks (290 days) Exclusion criteria No additional criteria State of cervix: favourable (Bishop score 6 or more)
Interventions	Induction group (n = 124): AROM + oxytocin (if uterine contractions inadequate after 2 hours); versus  Expectant management group (n = 125): spontaneous labour awaited unless 1) non‐reactive NST or 2) amniotic fluid index < 5 cm or 3) medical or obstetric indication for birth or 4) reaching 44 completed weeks.
Outcomes	Mother: prolonged labour; modes of birth and their indications; death; postpartum haemorrhage.  Baby: perinatal death, birthweight; birth asphyxia, NICU admission, birthweight > 4000 g; Apgar < 7 at 5 mins.
Notes	Funding: Ramathibodi Hospital Research Grant 2/2542 Declarations of interest: not reported
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation was carried out using computer‐generated numbers.
Allocation concealment (selection bias)	Unclear risk	The method of allocation concealment was not reported.
Blinding of participants and personnel (performance bias)   All outcomes	High risk	Blinding was not feasible.
Blinding of outcome assessment (detection bias)   All outcomes	Unclear risk	Blinded outcome assessment was not mentioned.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	1 woman (in IOL group) excluded after randomisation because of misclassification (breech presentation). No apparent losses to follow‐up.
Selective reporting (reporting bias)	Unclear risk	While pre‐specified outcomes (in the methods) were reported, no access to trial protocol to further assess selective reporting.
Other bias	Low risk	Appears to be free of other bias.

Methods	RCT
Participants	Number of women randomised: 248  Setting: Hong Kong, China Inclusion criteria GA: "on reaching 42 weeks' gestation" No indication for immediate delivery Exclusion criteria Not described State of cervix: not reported
Interventions	Induction group (n = not reported): induction versus Expectant management group (n = not reported): women were seen twice weekly.
Outcomes	Mother: length of stay in hospital; analgesia; caesarean birth; meconium‐stained liquor Baby: neonatal admission for meconium aspiration; birthweight; Apgar scores; umbilical cord vein pH
Notes	Abstract only. No data included in meta‐analyses Funding: not reported Declaration of interests: not reported
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	" A prospective randomised study". The method of sequence generation was not reported.
Allocation concealment (selection bias)	Unclear risk	The method of allocation concealment was not reported.
Blinding of participants and personnel (performance bias)   All outcomes	High risk	Blinding was not feasible.
Blinding of outcome assessment (detection bias)   All outcomes	Unclear risk	Blinded outcome assessment was not mentioned.
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	No information on losses to follow‐up or exclusions.
Selective reporting (reporting bias)	Unclear risk	Abstract only.
Other bias	Unclear risk	Abstract only. Baseline characteristics were not described.

Methods	Randomly allocated, no further details available.
Participants	Number of women randomised: 228  Setting: Glasgow, Scotland Inclusion criteria Primigravidae aged 18‐30 years 1‐3 parity aged 18‐35 years who had previous pregnancies without any obstetric abnormality Certain date of LMP Regular menstrual cycle Early examination which had shown the uterine size to be consistent with the period of amenorrhoea GA: 39‐40 weeks State of cervix: not reported
Interventions	Induction group (n = 111): IOL with AROM + oxytocin  versus Expectant management group (n = 117): no intervention until 41 weeks, thereafter IOL.
Outcomes	Mother: length of labour; mode of birth (including operative versus non operative); analgesia requirements; postpartum blood loss Baby: perinatal deaths; meconium staining; Apgar scores; birthweight; neonatal jaundice
Notes	Funding: not reported Declarations of interest: not reported
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	The method of sequence generation was not reported.
Allocation concealment (selection bias)	Unclear risk	The method of allocation concealment was not reported.
Blinding of participants and personnel (performance bias)   All outcomes	High risk	Blinding was not feasible.
Blinding of outcome assessment (detection bias)   All outcomes	Unclear risk	Blinded outcome assessment was not mentioned.
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	7/118 and 2/119 in the intervention and control groups excluded after randomisation because of misclassification as low risk.
Selective reporting (reporting bias)	Unclear risk	No outcomes were pre‐specified in the methods; no access to trial protocol to further assess selective reporting.
Other bias	Low risk	Appears to be free of other bias.

Methods	RCT
Participants	Number of women randomised: 302  Setting: Kaiser Permanente Medical Care Hospital in California, USA Inclusion criteria Well‐established GA of at least 287 days GA at intervention: at least 41 completed weeks (287 days) Exclusion criteria Non‐reactive NST result Variable decelerations on NST Oligohydramnios Any risk factors known to increase perinatal death and morbidity rates (such as chronic hypertension, pre‐eclampsia, diabetes mellitus, growth retardation and previous stillbirth) Any risk factors known to increase the risk of induction, such as multiple gestation and polyhydramnios Any risk factors known to markedly increase the caesarean section rate, such as breech presentation and previous caesarean section Cervical score of > 6 State of cervix: unfavourable (Bishop score < 6)
Interventions	Induction group (n = 152): PE2 gel (initially 3 mg but later reduced to 0.5 mg). If no labour in 24 hours, repeat PE2 and oxytocin if needed versus  Expectant management group (n = 150): NST twice weekly, pelvic examination and amniotic fluid determination weekly between 41‐42 weeks and twice weekly afterwards.
Outcomes	Mother: length of hospital stay; caesarean section; length of labour Baby: perinatal death; 1 min Apgar score < 7; 5 min Apgar score < 7; meconium‐stained amniotic fluid; meconium aspiration syndrome; post‐maturity syndrome; fetal distress; birthweight; birthweight > 4000 g; infant hospital stay length
Notes	Funding: Community Service Program of Kaiser Foundation Hospitals Declarations of interest: not reported
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	A table of random numbers was used.
Allocation concealment (selection bias)	Unclear risk	The authors reported " using a series of consecutively numbered, sealed envelopes..." for allocation concealment, but no mention was made of envelope opaqueness.
Blinding of participants and personnel (performance bias)   All outcomes	High risk	Blinding was not feasible.
Blinding of outcome assessment (detection bias)   All outcomes	Unclear risk	Blinded outcome assessment was not mentioned.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	No apparent losses to follow‐up or exclusions.
Selective reporting (reporting bias)	Unclear risk	No outcomes were pre‐specified in the methods; no access to trial protocol to further assess selective reporting.
Other bias	Low risk	Appears to be free of other bias.

Methods	RCT
Participants	345 women randomised  Setting: Vienna, Austria Inclusion criteria Length of pregnancy established by early ultrasound Membranes intact Cervix favourable for induction (modified Bishop score of > 4) GA at intervention: 40 completed weeks ("at due date") Exclusion criteria Any fetal or maternal risk factors based on history, gynaecological/obstetrical investigation, CTG and routine lab results State of cervix: favourable (Modified Bishop score > 4)
Interventions	Induction group (n = 180): vaginal PE2 (3 mg) tablets repeated 6 and 24 hours later if no active labour versus  Expectant management group (n = 165): spontaneous labour awaited until 42 weeks. NST monitoring every 2‐3 days.
Outcomes	Mother: birth interval (onset of contractions to birth in hours); rate and indication for operative birth; length of labour; analgesia requirements; caesarean section Baby: birthweight; length of baby at birth; incidence of meconium‐stained amniotic fluid; Apgar scores; results of umbilical cord pH determination; perinatal death
Notes	Funding: not reported Declarations of interest: not reported
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	The method of sequence generation was not reported.
Allocation concealment (selection bias)	Unclear risk	The method of allocation concealment was not reported.
Blinding of participants and personnel (performance bias)   All outcomes	High risk	Blinding was not feasible.
Blinding of outcome assessment (detection bias)   All outcomes	Unclear risk	Blinded outcome assessment was not mentioned.
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	8/180 women in the induction group refused to be induced; and 3/165 women in the expectant group requested induction; and these 11 women were excluded from analysis post‐randomisation.
Selective reporting (reporting bias)	High risk	No outcomes were pre‐specified in the methods, limited information was provided for some outcomes, e.g. "The incidence of prolonged labor was not different in both groups... both groups required analgetic treatment in 35%... Birthweight and length, the incidence of meconium‐stained amniotic fluid, of low Apgar scores, and the results of pH determination were not different between the two groups."
Other bias	Unclear risk	Some imbalance in the numbers randomised to each group (180 versus 165).

Methods	RCT
Participants	Number of women randomised: 600  Setting: teaching hospital in Ankara, Turkey; recruitment dates not reported Inclusion criteria Singleton pregnancy Vertex presentation Intact membranes Bishop score of < 5 Absence of spontaneous uterine contractions (< 4 per hour) Estimated fetal body weight < 4500 g Reactive NST Amniotic fluid index ≥ 5 cm GA at intervention: 41 completed weeks (287 days +/‐ 1 day) Exclusion criteria Allergic to prostaglandins Previous caesarean section Non‐cephalic presentation Body mass index 30 or more before conception Parity 5 or more Low‐lying placenta Previous labour induction attempt State of cervix: unfavourable ‐ Bishop score < 5 Just under half the women were nulliparous.
Interventions	Induction group: labour induction (3 methods) (1) vaginal administration of 50 mg misoprostol (n = 100) (2) oxytocin induction (n = 100), and (3) transcervical insertion of a Foley balloon (n = 100) versus Expectant management group: spontaneous follow‐up with twice‐weekly nonstress testing and amniotic fluid measurement and once‐weekly biophysical scoring (n = 300); 24% of women were induced after 42 completed weeks. the 3 induction arms were combined for analyses
Outcomes	Mother: oligohydramnios; pre‐eclampsia; tachysystole; hyperstimulation; vaginal birth; caesarean (emergent abdominal birth for worrying FHR); failed IOL  Baby: perinatal death; shoulder dystocia; meconium stained amniotic fluid; meconium aspiration syndrome; fetal anomaly; low Apgar scores (< 7 at 5 mins); umbilical artery pH < 7.16; NICU admission; fetal macrosomia; birthweight; birthweight > 4000 g; length of hospital stay
Notes	Funding: not reported Declarations of interest: not reported
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	The method of sequence generation was not reported.
Allocation concealment (selection bias)	Unclear risk	Allocation concealment was by sealed, opaque envelopes but there is no mention of numbering and sequential opening of the envelopes.
Blinding of participants and personnel (performance bias)   All outcomes	High risk	No blinding: "Staff members in charge of labor were not blinded to the type of medication used for induction".
Blinding of outcome assessment (detection bias)   All outcomes	Unclear risk	Blinded outcome assessment not mentioned.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	No apparent losses to follow‐up or exclusions.
Selective reporting (reporting bias)	Unclear risk	While pre‐specified outcomes (in the methods) were reported, no access to trial protocol to further assess selective reporting.
Other bias	Low risk	Appears to be free of other bias.

Methods	RCT
Participants	Number of women randomised: 3418 enrolled (data available for 3407 women only)  Setting: 22 hospitals across Canada Inclusion criteria GA at intervention: 41 completed weeks or more Singleton pregnancy Exclusion criteria Cervix dilated ≥ 3 cm GA ≥ 44 weeks Non‐cephalic presentation Evidence of a lethal congenital anomaly Maternal diabetes mellitus Pre‐eclampsia Intrauterine growth retardation Prelabour rupture of the membranes Need for urgent birth (e.g. fetal distress or antepartum bleeding) Vaginal birth contraindicated (e.g. placenta praevia) Previous caesarean section Addiction to drugs or alcohol State of cervix: unfavourable at trial entry (first ripening and then IOL in the intervention group)
Interventions	Induction group (n = 1701): up to 3 x 0.5 mg doses of PGE2 gel administered intracervically (if NST was normal and cervix unfavourable at time of induction = 77% of women), followed by either AROM or IV oxytocin infusion, or both. versus Expectant management group (n = 1706): daily fetal movement counting, NST and amniotic fluid measurement 2‐3 times per week. If either the NST or amniotic fluid volume assessment was abnormal, or other complications developed, labour was induced (28% of women induced in the expectant group received some form of PE2 (not gel)).
Outcomes	Mother: caesarean section; operative vaginal birth  Baby: perinatal death (stillbirth or neonatal death before discharge excluding deaths caused by lethal congenital abnormalities); birthweight > 4000 g; Apgar score < 7 at 5 min; asphyxial encephalopathy (seizures, alterations in levels of consciousness or tone, or a need for tube feeding during the first 48 hours of life), respiratory distress (oxygen requirement > 40% and respiratory rate > 60 breaths/min, both within 12 hours after birth and persisting for more than 24 hours, or assisted ventilation for more than 24 hours): meconium aspiration syndrome; neonatal trauma; NICU admission
Notes	Most women (89%) were enrolled at 41⁰ to 41⁶ weeks' gestation (3% before 41 weeks and 8% at or beyond 42 weeks), of whom 86.2% in the induced group and 63.6% in the expectant group gave birth before 42 weeks' gestation. In the induction group, 31% of women were not induced and in the expectant management group, 34% of women were induced.  Funding: Medical Research Council of Canada: MA‐8472; Upjohn Company of Canada supplied the prostaglandin gel. Declarations of interest: not reported
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	The method of sequence generation was not reported.
Allocation concealment (selection bias)	Low risk	Randomisation was carried out at a site separate from the trial ("centrally controlled at McMaster University").
Blinding of participants and personnel (performance bias)   All outcomes	High risk	Blinding was not feasible.
Blinding of outcome assessment (detection bias)   All outcomes	Unclear risk	The study was partially blinded; an adjudication of abnormal neonatal outcomes was undertaken by a neonatologist who was unaware of the mothers' group assignments.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	3418 women enrolled (data available for 3407 women); 7 women whose babies had lethal congenital anomalies were excluded after randomisation from the analysis of perinatal and neonatal outcomes ‐ induction group (1 woman) and expectant management group (6 women).
Selective reporting (reporting bias)	Unclear risk	Pre‐specified outcomes (from the methods) were reported; some results reported incompletely in text, e.g. "The frequency of postpartum maternal morbidity (hemorrhage, sepsis, endometritis) did not differ between the two groups (data not shown)." No access to trial protocol to further assess selective reporting.
Other bias	Low risk	Appears to be free of other bias; although methods of induction differed between the induction group and the women requiring induction in the expectant management group.

Methods	RCT
Participants	Number of women randomised: 508  Setting: St. Olavs University Hospital, Trondheim, Norway Inclusion criteria Singleton pregnancies GA: 41+ weeks (at intervention GA = 40⁶ and beyond) Cephalic presentation No PROM State of cervix: all stages included
Interventions	Induction group (n = 254): if cervix favourable (Bishop score ≥ 6) AROM + oxytocin, if not (Bishop score < 6) 50 µg misoprostol vaginally versus Expectant management group (n = 254): twice‐weekly ultrasound and CTG, labour induction after 300 days of pregnancy.
Outcomes	Mother: prolonged labour; mode of birth; perineal trauma; maternal satisfaction; postpartum haemorrhage  Baby: perinatal death; neonatal morbidity, for which a score was tallied (by evaluating the degree of deviation from the potential of a perfect outcome for each newborn as defined by the authors); neonatal trauma; birthweight; birthweight > 4000 g, NICU admission, birth asphyxia, meconium aspiration syndrome, Apgar < 7 at 5 mins
Notes	Funding: not reported Declarations of interest: not reported
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computerised randomisation using blocks of 16 with no stratification.
Allocation concealment (selection bias)	Low risk	Central allocation ‐ clinical trials office.
Blinding of participants and personnel (performance bias)   All outcomes	High risk	Blinding was not feasible.
Blinding of outcome assessment (detection bias)   All outcomes	Unclear risk	Blinded outcome assessment was not mentioned.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	No women were lost to follow‐up for clinical outcomes; 8 women did not complete the inclusion questionnaire; for the post‐birth telephone survey, 12 women were lost to follow‐up (4 in induction group and 8 in expectant management group).
Selective reporting (reporting bias)	Unclear risk	While all pre‐specified outcomes (in the methods) were reported, with no access to trial protocol it is not possible to further assess selective reporting.
Other bias	Low risk	Appears to be free of other bias.

Methods	RCT with inadequately reported randomisation methods.
Participants	Number of women randomised: 112  Setting: Birmingham, UK Inclusion criteria (not well specified) GA: 41⁺ weeks Certain of dates Exclusion criteria Not specified State of cervix: not mentioned as a criterion
Interventions	Induction group (n = 55): AROM and oxytocin ("surgical" group)  versus Expectant management group (n = 57): weekly amnioscopy.
Outcomes	Mother: number of days past term; prolonged labour; mode of birth Baby: perinatal death; birthweight
Notes	4 women in expectant group and 1 in induction group were randomised before 41 weeks Funding: not reported Declaration of interests: not reported
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	The method of sequence generation was not reported.
Allocation concealment (selection bias)	Unclear risk	The method of allocation concealment was not reported.
Blinding of participants and personnel (performance bias)   All outcomes	High risk	Blinding was not feasible.
Blinding of outcome assessment (detection bias)   All outcomes	Unclear risk	Blinded outcome assessment was not mentioned.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	No apparent losses to follow‐up or exclusions.
Selective reporting (reporting bias)	Unclear risk	No outcomes were pre‐specified in the methods; no access to trial protocol to further assess selective reporting.
Other bias	Low risk	Appears to be free of other bias.

Methods	RCT
Participants	Number of women randomised: 108  Setting: Bangkok, Thailand Inclusion criteria Certain dates Low risk GA at intervention: 42 completed weeks (immediately after) Exclusion criteria Women with Bishop scores of > 6 were judged to have a favourable cervix and were excluded from the study State of cervix: unfavourable cervix (Bishop score 6 or less)
Interventions	Induction group (n = 57): PGE2 intracervical, repeated after 6 hours, AROM and oxytocin on day 2 according to contractions versus  Expectant management group (n = 51): a) NST between 42 and 43 completed weeks. 2) NST between 43 and 44 completed weeks; women underwent IOL if there were abnormalities in antepartum fetal testing as non‐reactive NST, or variable decelerations on NST or if Bishop score > 6 on reaching 44 completed weeks' gestation.
Outcomes	Mother: length of first stage of labour; mode of birth; cephalopelvic disproportion; fetal distress Baby: birthweight; meconium staining; Apgar score < 7 at 1 min; Apgar score < 7 at 5 min; intubation required; admission to special care baby unit; perinatal death
Notes	Funding: Ramathibodi Hospital Research Fund Grant 1988 Declaration of interests: not reported
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	The method of sequence generation was not reported.
Allocation concealment (selection bias)	Unclear risk	The method of allocation concealment was not reported.
Blinding of participants and personnel (performance bias)   All outcomes	High risk	Blinding was not feasible.
Blinding of outcome assessment (detection bias)   All outcomes	Unclear risk	Blinded outcome assessment was not mentioned.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	No apparent losses to follow‐up or exclusions.
Selective reporting (reporting bias)	Unclear risk	No outcomes were pre‐specified in the methods; no access to trial protocol to further assess selective reporting.
Other bias	Low risk	Appears to be free of other bias.

PERMALINK

Induction of labour for improving birth outcomes for women at or beyond term

Philippa Middleton

Emily Shepherd

Caroline A Crowther

Abstract

Background

Objectives

Search methods

Selection criteria

Data collection and analysis

Main results

Authors' conclusions

Plain language summary

Summary of findings

Background

Description of the condition

Description of the intervention

How the intervention might work

Why it is important to do this review

Objectives

Methods

Criteria for considering studies for this review

Types of studies

Types of participants

Types of interventions

Types of outcome measures

Primary outcomes

Secondary outcomes

For the infant/child

For the mother

Health services use

Search methods for identification of studies

Electronic searches

Searching other resources

Data collection and analysis

Selection of studies

Data extraction and management

Assessment of risk of bias in included studies

(1) Allocation (checking for possible selection bias)

(2) Allocation concealment (checking for possible selection bias)

(3.1) Blinding of participants and personnel (checking for possible performance bias)

(3.2) Blinding of outcome assessment (checking for possible detection bias)

(4) Incomplete outcome data (checking for possible attrition bias due to the amount, nature and handling of incomplete outcome data)

(5) Selective reporting (checking for reporting bias)

(6) Other bias (checking for bias due to problems not covered by (1) to (5) above)

(7) Overall risk of bias

Assessment of the quality of the evidence using the GRADE approach

For the infant/child

For the mother

Measures of treatment effect

Dichotomous data

Continuous data

Unit of analysis issues

Cluster‐randomised trials

Cross‐over trials

Multiple pregnancies

Multi‐armed trials

Dealing with missing data

Assessment of heterogeneity

Assessment of reporting biases

Data synthesis

Subgroup analysis and investigation of heterogeneity

For the infant/child

For the mother

Sensitivity analysis

For the infant/child

For the mother

Results

Description of studies

Results of the search

1.

Included studies

Settings

Cervix status

Interventions

Timing of induction ‐ induction group

Method of induction ‐ induction group

Expectant management group protocols

Outcomes

Methods	RCT
Participants	Number of women randomised: 74 Setting: Vellore, India Inclusion criteria Low‐risk women Singleton pregnancy Cephalic presentation GA: 41 completed weeks (287 days) Exclusion criteria Presence of risk factors known to increase perinatal death and morbidity such as chronic hypertension, pre‐eclampsia, maternal diabetes mellitus, fetal growth retardation, multiple gestation, hydramnios, PROM, antepartum haemorrhage and previous caesarean section State of cervix: not mentioned as a criterion
Interventions	Induction group (n = 37): Bishop < 5: cervical ripening with extra‐amniotically placed 16F Foley catheter with 20 mL of saline Bishop > 5: stripping of membranes Then, 12 hours later, IOL by AROM and oxytocin infusion versus Expectant management group (n = 37): daily fetal movement counts; biophysical profile every second day
Outcomes	Mother: mode of birth and indications; duration of labour; mean hospital stay Baby: meconium staining of amniotic fluid; meconium aspiration; Apgar scores < 7 (at 1 and 5 min); need for neonatal intubation; birthweight; birthweight > 4000 g; signs of post maturity; perinatal deaths; abnormal electronic fetal trace monitoring
Notes	Funding: not reported Declarations of interest: not reported
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	A table of random numbers was used.
Allocation concealment (selection bias)	Unclear risk	Allocation concealment was unclear since "... a series of consecutively numbered, sealed envelopes..." was used but no mention was made of opaqueness of the envelopes.
Blinding of participants and personnel (performance bias)   All outcomes	High risk	Blinding was not feasible.
Blinding of outcome assessment (detection bias)   All outcomes	Unclear risk	Blinded outcome assessment was not mentioned.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	No loss to follow‐up or post‐randomisation exclusion.
Selective reporting (reporting bias)	Unclear risk	All of the outcomes mentioned in the methods section were reported on in the results section; mean duration of labour was reported with no measure of variance. No access to trial protocol to further assess selective reporting.
Other bias	Low risk	Appears to be free of other bias.

Methods	Multi‐centre RCT (non‐inferiority trial)
Participants	Number of women randomised: 1811 (August 2012 to March 2016) Setting: Dutch Obstetric Consortium in cooperation with the Midwifery Research Network of the Netherlands including 200 centres (university hospitals, teaching hospitals, non‐teaching hospitals and midwifery practices). Inclusion criteria: low‐risk women > 18 years with a singleton pregnancy in stable cephalic position and a certain gestational age of 41^‐2/+2 weeks. Exclusion criteria: age < 18 years, uncertain gestational age, obstetrical indications for secondary care (e.g. hypertension, proteinuria, pre‐existing maternal heart or kidney diseases, gestational diabetes, previous caesarean section, multiple pregnancy, intrauterine growth restriction, non‐reassuring fetal status (no fetal movements, abnormal fetal heart rate, known fetal abnormalities which could influence perinatal outcome (including abnormal karyotype), ruptured membranes at time of randomisation and a non‐reassuring fetal status at time of randomisation).
Interventions	*Induction of labour at 41^0‐2 weeks (n = 902) versus Expectant management until 42 weeks (n = 909)** *Women with a cervix that is judged to be ripe at vaginal examination (Bishop Score of 6 or more), will have labour induced with amniotomy followed by intravenous oxytocin according to local protocol. In case rupturing of membranes is not possible, cervical ripening will be accomplished in accordance with national guidelines.
Outcomes	Primary outcome (of study): composite of perinatal death and neonatal morbidity (adverse perinatal outcomes are defined as a composite of perinatal death, a 5‐minute Apgar‐score < 7 and/or an arterial pH < 7.05, meconium aspiration syndrome, plexus brachialis injury, intracranial haemorrhage and/or NICU admission). Secondary outcomes (of study): individual components of the composite; maternal outcomes (instrumental vaginal birth, Caesarean section), analgesia (epidural, spinal, opiates), postpartum haemorrhage ≥ 1000 mL, severe perineal injury (third‐ or fourth‐degree perineal tear).
Notes	Funding: ZonMW grant number 17120200 Declarations of interest: the authors declare that they have no conflicts of interest
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated code
Allocation concealment (selection bias)	Low risk	Computer‐based randomisation
Blinding of participants and personnel (performance bias)   All outcomes	High risk	Women and staff were aware of assignments
Blinding of outcome assessment (detection bias)   All outcomes	High risk	Staff making treatment decisions and recording outcomes were aware of assignments
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Results available from 96% of 1811 women (but abstract does not report the 109 losses by group)
Selective reporting (reporting bias)	Unclear risk	Trial protocol published, results not fully published yet
Other bias	Unclear risk	Abstract only published to date

Methods	RCT
Participants	Number of women randomised: 264 "admitted to this trial"  Setting: Royal Maternity Hospital, Northern Ireland, UK Trial recruitment: timing not reported Inclusion criteria GA: women in their 38th week of pregnancy All past/present pregnancies obstetrically normal Booked before 18th week Menstrual cycle not exceeding 35 days No contraceptive pill taken, nor any pregnancy for at least 3 months before the LMP Size of uterus at booking corresponded to the period of amenorrhoea Exclusion criteria Not described State of cervix: not clearly reported; assumed to be mixed (Bishop score recorded).
Interventions	Induction group (n = 131 admitted; 92 analysed): IOL at 39 weeks' GA. Women were admitted for fasting at 8:30 am; their forewaters were punctured soon after and Bishop score and cervical dilatation recorded. IV oxytocin commenced at 2.5 mU/min and doubled every 30 min until satisfactory uterine response achieved; dose varied to maintain adequate contractions. All women continuously monitored with internal tocography and fetal scalp electrode versus Expectant management group (n = 133 admitted; 92 analysed): await spontaneous labour until 42 weeks, unless IOL required earlier for medical reasons. Women had, if necessary, augmentation of labour by puncture of the forewaters or IV oxytocin; when possible, they were also monitored.
Outcomes	Mother: mode of birth (assisted birth; caesarean birth); Induction to birth interval; Unexplained postpartum pyrexia; Analgesia demand; Type 1 and 2 diabetes; Meconium staining of the amniotic fluid; duration of gestation; attitudes towards management Baby: Apgar scores at 1 and 5 min; Dubowitz scores < 45; stillbirth; hyperbilirubinaemia
Notes	Funding: not reported Declarations of interest: not reported
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"A prospective randomized controlled trial"; "allocated using random number tables."
Allocation concealment (selection bias)	Unclear risk	The method of allocation concealment was not reported.
Blinding of participants and personnel (performance bias)   All outcomes	High risk	Blinding was not feasible.
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	“All monitor records were examined blind at completion of the trial.”
Incomplete outcome data (attrition bias)   All outcomes	High risk	184/264 women followed up (30% loss to follow‐up). 264 women eligible, 34 excluded: induction group: obstetric abnormality (8 women); refusal (10 women); defaulter (6 women); spontaneous labour before attending clinic (1 woman). Expectant management group: obstetric reasons (9); leaving 230 women in the trial (106 in planned birth, 124 in expectant management group) Of the 106 women in the induction group, 13 went into spontaneous labour before the date of admission and 1 was excluded due to medical reasons – therefore 92 women were induced and analysed. Of the 124 women in the expectant management group, a further 32 were excluded due to obstetric abnormalities or failure to go into spontaneous labour before 42 weeks – therefore only 92 were analysed
Selective reporting (reporting bias)	Unclear risk	Some results reported incompletely, e.g.: “There was no difference with respect to the distribution of Apgar scores…at five minutes.”
Other bias	Unclear risk	Groups appeared comparable at baseline however analyses excluded 30% of women admitted to the trial. Limited methodological detail provided to further assess other bias.

Methods	RCT
Participants	Number of women randomised: 22  Setting: Jackson, USA Inclusion criteria GA: 41 completed weeks Reliable dates Exclusion criteria Oligohydramnios with < 1 cm pocket of amniotic fluid in any dimension A non‐reactive NST Positive concentration stress test Bishop score > 5 State of cervix: unripe (Bishop score 5 or less) included
Interventions	Induction group (n =12): laminaria tents followed by oxytocin versus  Expectant management group (n = 10): weekly ultrasound for amniotic fluid assessment and NST.
Outcomes	Mother: mode of birth; length of labour; type of analgesia; length of hospital stay; labour‐associated morbidity Baby: birthweight; Apgar score; perinatal deaths; neonatal course; meconium staining
Notes	Funding: Vicksburg Hospital Medical Foundation Declarations of interest: not reported
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	The method of sequence generation was not reported.
Allocation concealment (selection bias)	Unclear risk	Allocation in sealed envelopes but no mention of opaqueness, numbering and sequential opening of envelopes.
Blinding of participants and personnel (performance bias)   All outcomes	High risk	Blinding was not feasible.
Blinding of outcome assessment (detection bias)   All outcomes	Unclear risk	Blinded outcome assessment was not mentioned.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	No apparent losses to follow‐up or exclusions.
Selective reporting (reporting bias)	High risk	No outcomes were pre‐specified in the methods; a number of outcomes reported without measures of variance (e.g. birthweight, length of labour, hospital stay), and thus these outcomes could not be used in the meta‐analyses. No access to protocol to further assess selective reporting.
Other bias	Low risk	Appears to be free of other bias.

Methods	RCT
Participants	Number of women randomised: 162  Setting: Military tertiary care medical centre, USA Inclusion criteria GA: between 38⁰ and 38⁶ weeks 18 years or older Nulliparous Singleton pregnancies, cephalic presentation Modified Bishop score ≤ 5 Exclusion criteria Contraindications to vaginal birth Unclear GA Requiring IOL for medical or obstetric indications ≤ 39 weeks' GA Non‐English speaking State of cervix: modified Bishop score ≤ 5
Interventions	Induction group (n = 82): IOL ≤ 1 weeks of randomisation; not before 39 weeks' GA. Where possible IOL by Foley catheter (single balloon, 60 mL water), taped in place until expulsion or 12 hours. After Foley catheter, oxytocin (2 per min, increasing every 20 min to 36 mIU per min max) until adequate contractions. If Foley catheter placement not possible, IOL by misoprostol 25 µg vaginally, repeated every 4 hours; AROM after 3 cm dilation; oxytocin administered after last if adequate contractions not observed with misoprostol versus Expectant management group (n = 80): scheduled for routine appointments and birthed for obstetric indications no later than 42 weeks' GA.
Outcomes	Mother: caesarean birth; mode of delivery; number of visits after randomisation; unscheduled clinic or triage visits; number of antepartum fetal testing appointments; GA at admission; Bishop score at admission; admission diagnosis; indication for operative birth; use of regional anaesthesia; chorioamnionitis; estimated blood loss; blood transfusion; endomyometritis; labour and birth length of stay; postpartum stay Baby: meconium‐stained amniotic fluid; NICU admission; birthweight; SGA; LGA; Apgar score < 5 at 5 min
Notes	Funding: not reported Declarations of interest: the authors declared that they had no financial conflicts of interest
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated list of random numbers in permuted blocks of 4.
Allocation concealment (selection bias)	Low risk	Allocation was concealed by the use of opaque, sealed, sequentially numbered envelopes.
Blinding of participants and personnel (performance bias)   All outcomes	High risk	No blinding: "blinding of health care providers to the indication for delivery was deemed impractical."
Blinding of outcome assessment (detection bias)   All outcomes	Unclear risk	Blinded outcome assessment was not mentioned.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Little loss to follow‐up following randomisation (82/82 and 79/80 women included in the analyses).
Selective reporting (reporting bias)	Unclear risk	No access to trial protocol to assess selective reporting. Perinatal death not reported.
Other bias	Low risk	Appears to be free of other bias.

Methods	RCT
Participants	Number of women randomised: 440  Setting: University hospitals in the USA Inclusion criteria GA at trial entry: at least 287 days GA at intervention: 41 to 43 completed weeks (at least 287 days to < 301 days) Exclusion criteria Any medical or obstetric complications requiring IOL, caesarean section or frequent monitoring of maternal or fetal condition State of cervix: unfavourable (Bishop score 6 or less)
Interventions	Induction group (n = 174): cervical priming with PGE2 gel followed 12 hours later with oxytocin versus  Expectant management group (n = 175): weekly cervix assessments, twice weekly NST and amniotic fluid volume assessment. A total of 265 women were randomised to the intervention arm; however, 91 of these women were randomised to placebo gel with oxytocin 12 hours later and these women have not been included in this review.
Outcomes	Mother: time to birth from randomisation; maternal infection; need for transfusion; uterine hyperactivity; mode of birth; maternal death  Baby: mechanical ventilation; nerve injury; seizures; babies with ≽1 adverse outcome; perinatal death; birthweight; Apgar score < 4 at 5 min; late decelerations in labour; meconium in amniotic fluid; meconium in aspiration pneumonia
Notes	The initial sample size intended was 2800. However, after 18 months and 440 participants, the study was stopped, since the incidence of adverse outcome was only 1.1% and therefore a sample size of 5600 would be required to adequately test the hypothesis proposed. Funding: National Institute of Child Health and Human Development, NIH, USA Declarations of interest: not reported
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	The randomisation sequence generation was performed using a computer‐generated randomisation scheme stratified by site and GA.
Allocation concealment (selection bias)	Low risk	Allocation was concealed by using central allocation by a data co‐ordinating centre.
Blinding of participants and personnel (performance bias)   All outcomes	High risk	Blinding was not feasible.
Blinding of outcome assessment (detection bias)   All outcomes	Unclear risk	Blinded outcome assessment was not mentioned.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	No apparent losses to follow‐up or exclusions.
Selective reporting (reporting bias)	Unclear risk	All pre‐specified outcomes (in methods) were reported; no access to trial protocol to further assess selective reporting.
Other bias	Low risk	Appears to be free of other bias.

Methods	RCT
Participants	Number of women randomised: 226  Setting: Army Medical Center, Tacoma, Washington, USA Inclusion criteria GA at intervention: 39‐40 weeks Cephalic presentation Singleton gestation Maternal age of greater than 17 years Candidate for vaginal birth Semi‐favourable cervical Bishop score defined as a score of 5 or greater in nulliparous or 4 or greater in multiparous women Exclusion criteria No additional criteria reported State of cervix: favourable (≥ 5 for nulliparous and ≥ 4 for multiparous women)
Interventions	Induction group (n = 116): AROM, oxytocin or both versus Expectant management group (n = 110): weekly follow‐up until 42 weeks. Labour induced after 42 weeks. Weekly monitoring with CTG and ultrasound, increased to twice a week after 41 weeks.
Outcomes	Mother: randomisation to birth interval; admission to birth interval; Indication for admission; epidural analgesia; mode of birth; EBL; length of labour; chorioamnionitis; postpartum days Baby: birthweight; admission to NICU; Apgar score < 7 at 5 mins
Notes	The study was discontinued after recruitment of 226 women (target of 600) due to slow recruitment and no observed difference in the 2 groups Funding: not reported Declarations of interest: not reported
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	The randomisation sequence was generated using a computer‐generated list.
Allocation concealment (selection bias)	Low risk	Allocation concealment was achieved using sequentially numbered, opaque, sealed envelopes.
Blinding of participants and personnel (performance bias)   All outcomes	High risk	Blinding was not feasible.
Blinding of outcome assessment (detection bias)   All outcomes	Unclear risk	Blinded outcome assessment was not mentioned.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	No apparent losses to follow‐ up or exclusions. 23/116 (19.8%) in induction group went into spontaneous labour, 10/110 (9.1%) in the expectant management group required labour induction and results for these women were analysed according to which group they were randomised.
Selective reporting (reporting bias)	Unclear risk	No access to trial protocol for confidently assess selective reporting. Perinatal death not reported.
Other bias	Unclear risk	Baseline imbalance for Bishop score "The only significant difference noted was EM patients had a more favorable Bishop score on admission than IND patients (7.2+2.1 versus 8.6+2.0, p<0.0001)."

Methods	RCT (partially translated).
Participants	Number of women randomised: 113  Setting: Gran Canaria, Spain Inclusion criteria Unknown due to not being fully translated GA at intervention: 42 completed weeks Exclusion criteria Unknown, not in translation State of cervix: unfavourable (Bishop score < 5)
Interventions	Induction group (n = 57): intracervical PGE2 gel (0.5 mg); unclear whether further intervention occurred (full translation not available) versus  Expectant management group (n = 56): monitoring by NST, biophysical profile and amnioscopy.
Outcomes	Mother: time to birth; mode of birth  Baby: meconium staining; NICU admission; birthweight > 4000 g; Apgar score < 7 at 5 mins (other outcomes may have been present, but were not reported in the translation)
Notes	Funding: not reported Declarations of interest: not reported
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	The method of sequence generation was not reported according to the translation.
Allocation concealment (selection bias)	Unclear risk	The method of allocation concealment was not reported according to the translation.
Blinding of participants and personnel (performance bias)   All outcomes	High risk	Blinding was not feasible.
Blinding of outcome assessment (detection bias)   All outcomes	Unclear risk	Blinded outcome assessment was not mentioned.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	No apparent losses to follow‐up or exclusions.
Selective reporting (reporting bias)	Unclear risk	No access to trial protocol for confidently assess selective reporting. Perinatal deaths appear not to have been reported according to the translation, although this has not been verified by a second translation.
Other bias	Low risk	Appears to be free of other bias.

Methods	RCT
Participants	Number of women randomised: 201  Setting: Hong Kong, China Inclusion criteria GA at intervention: 42 completed weeks Exclusion criteria Pre‐eclampsia Gestational diabetes Contraindication to vaginal birth (e.g. placenta praevia, non‐cephalic presentation) Evidence of fetal or maternal compromise State of cervix: not mentioned as a criterion
Interventions	Induction group (n = 96): PGE2 pessaries 6‐hourly if necessary versus  Expectant management group (n = 105): serial monitoring with NST (x2) and amniotic fluid index measurements (x1) weekly.
Outcomes	Mother: spontaneous labour; caesarean section; fetal distress in labour Baby: birthweight; Apgar score < 7 (1 min/5 min); cord blood pH; admission to NICU; meconium below the vocal cords
Notes	Funding: not reported Declarations of interest: not reported
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	The method of sequence generation was not reported.
Allocation concealment (selection bias)	Unclear risk	Allocation in a series of identical envelopes but no mention of sealed envelopes, opaqueness and sequential numbered envelopes.
Blinding of participants and personnel (performance bias)   All outcomes	High risk	Blinding was not feasible.
Blinding of outcome assessment (detection bias)   All outcomes	Unclear risk	Blinded outcome assessment was not mentioned.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	No apparent losses to follow‐up or exclusions. 17/96 (18%) in the induction group went into spontaneous labour and 12/105 (11%) in the expectant management group were induced and the results for these women were included in the analyses.
Selective reporting (reporting bias)	Unclear risk	Very few outcomes reported; "We did not address perinatal mortality in this study." No access to trial protocol to further assess selective reporting.
Other bias	Low risk	Appears to be free of other bias.

Methods	RCT
Participants	Number of women randomised: 150  Setting: Sousse, Tunisie (Tunisia) Inclusion criteria 41⁰ to 41⁶ weeks Dates concur with ultrasound before 20 weeks Regular menstrual cycle length 28‐30 days Not on contraception for 3 months prior to conception Singleton pregnancy Morphologically normal ultrasound Intact membranes Bishop score < 4 at initial exam No medical or obstetric complications? Exclusion criteria Presence of risk factors for complication (hypertension, pre‐eclampsia, diabetes, placenta praevia) Fetal‐pelvic disproportion More than 5 previous pregnancies Previous caesarean section Previous IUFD Medical contraindication to the use of prostaglandins (asthma, glaucoma, heart disease, allergy to prostaglandins) State of cervix: cervix unripe (Bishop score < 4)
Interventions	Induction group (n = 75): PGE2 gel intracervically (daily cervical ripening by PGE2 gel, maximum 3 gels) versus Expectant management group (n = 75): CTG every second day until 42 completed weeks. After that, PGE2 gel if no spontaneous labour.
Outcomes	Mother: duration of labour; mode of birth; GA at birth; duration of mother’s hospital stay (hours); need for augmentation of labour using synthetic oxytocin (Recours aux ocytociques); effect of Bishop score on admission on duration of labour (Effet du score de Bishop à l’admission sur la durée (duration) du travail (labour); progress in labour; time between final dose of PE2 gel and birth Baby: duration of infant’s hospital stay (hours); total cost of care; admission to neonatal unit; stained amniotic fluid; Apgar score at 1 min; perinatal death; stillbirth; neonatal death; macrosomia; signs of post‐maturity; need for resuscitation at birth; number of doses of gel administered
Notes	This article is in French. Funding: not reported Declaration of interests: not reported
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation by computer.
Allocation concealment (selection bias)	Unclear risk	Article in French. Appears not to have been reported.
Blinding of participants and personnel (performance bias)   All outcomes	High risk	Blinding was not feasible.
Blinding of outcome assessment (detection bias)   All outcomes	Unclear risk	Blinded outcome assessment was not mentioned.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	No apparent losses to follow‐up or exclusions.
Selective reporting (reporting bias)	Unclear risk	No access to trial protocol to confidently assess selective reporting.
Other bias	Low risk	Appears to be free of other bias.