Intended for healthcare professionals


When to induce late term pregnancies

BMJ 2019; 367 doi: (Published 20 November 2019) Cite this as: BMJ 2019;367:l6486

Linked Research

Induction of labour at 41 weeks versus expectant management and induction of labour at 42 weeks

  1. Sara Kenyon, professor of evidence based maternity care1,
  2. Lee Middleton, senior statistician2,
  3. Magdalena Skrybant, patient and public involvement and engagement lead for NIHR Applied Research Collaboration (ARC) West Midlands1,
  4. Tracey Johnston, consultant in fetal maternal medicine3
  1. 1Institute of Applied Research Centre, Public Health, Epidemiology and Biostatistics, University of Birmingham, Birmingham B15 2TT, UK
  2. 2Birmingham Clinical Trials Unit, Institute of Applied Health Research, Birmingham, UK
  3. 3Birmingham Women’s and Children’s NHS Foundation Trust, Birmingham, UK
  1. Correspondence to: S Kenyon s.kenyon{at}

41 weeks looks like the safer option for women and their babies

Giving clear unbiased information to pregnant women enabling them to make their own choices is extremely important. For those with late term pregnancies a new trial reported by Wennerholm and colleagues (doi:10.1136/bmj.l6131) suggests that waiting longer than 41 weeks to induce labour increases the risk of perinatal mortality.1

The SWEdish Post-term Induction Study (SWEPIS) randomised 2760 women with low risk singleton pregnancies to induction of labour at 41 weeks or expectant management until induction at 42 weeks. The primary outcome was a composite of perinatal mortality and morbidity and no difference was found between the groups. The authors observed an average of three days difference in gestation, and the proportions of caesarean sections, instrumental births, and major maternal morbidity were similar. A systematic review of composite primary outcomes in clinical trials has identified their use as problematic, with components often unreasonably combined, inconsistently defined, and inadequately reported.2 While the first could be said of this trial, the definitions and reporting were clear.

The trial was stopped early by the independent Data Safety Monitoring Board, well short of the planned 10 038 sample size, after six babies died in the expectant management group compared with none in the induction at 41 weeks group. Stopping a trial early on the basis of safety outcomes is controversial when numbers of events are low (n=6)—particularly without the reassurance of a convincing statistical signal—but in this instance the decision appears reasonable. Whether the Data Safety Monitoring Board considered any external evidence when making this decision is not clear.

A recent Cochrane systematic review found a policy of induction at or beyond the expected date of birth compared with expectant management to be associated with fewer perinatal deaths, admissions to neonatal intensive care, and caesarean sections but more operative vaginal births.3 However, uncertainty about the optimal timing of induction remains.

Two previous trials have made the same comparisons as SWEPIS; the first, from Turkey, recorded one death, in the expectant management group.4 The second, the INDEX (INDuction of labour at 41 weeks with a policy of EXpectant management until 42 weeks) trial, found a reduced risk of adverse perinatal outcomes in favour of induction, and the study recorded one perinatal death in the induction group compared with two in the expectant management group.5

When mortality data from these trials are combined with the results of SWEPIS a total of 0.4 deaths per 1000 (1/2581) occur with induction at 41 weeks compared with 3.5 events per 1000 (9/2580) with expectant management (Peto odds ratio from meta-analysis 0.20, 95% confidence interval 0.06 to 0.70). The number of events is small and a lot of imprecision exists in the estimate of treatment effect, but this evidence is likely to be compelling enough to question the ethics of starting any further trials.

Choice is important within maternity care, and clear information about available options should be accessible to all pregnant women, enabling them to make fully informed and timely decisions. A substantial number of eligible women did not join SWEPIS (77.9% (9792/12 554) declined), suggesting that many women at 41 gestational weeks prefer either induction or expectant management. Regrettably, women were not involved in the design or conduct of this trial, or in dissemination of the findings. Involving them would add real value to ensuring a clear message that, although the overall risk of an adverse outcome at 42 weeks is low for both mother and baby, induction of labour at 41 weeks is associated with fewer perinatal deaths and no difference in neonatal and maternal morbidity. The combined evidence from SWEPIS and previous trials suggests that induction at 41 weeks is a reasonable choice for women, but maternity units must now find ways to overcome the challenge of implementation.

Summary table of perinatal deaths

View this table: