SWEPIS – The Swedish Post-term Induction Study

The evidence on whether there is a benefit in inducing labour if a pregnancy would otherwise last beyond 41 or 42 weeks is far from clear.¹

The SWEPIS study², published in 2019, was designed to look at whether induction of labour at 41 weeks improves outcomes for mothers and babies with a low risk pregnancy compared with waiting till 42 weeks and inducing labour then if it has not already started (referred to as expectant management).

Studies which have looked at data from a large population of women suggest that there may be a small increase in the risk of stillbirth in later pregnancy. A recent review³ of population studies estimated that there was one additional stillbirth for every 604 pregnancies which lasted to 42 weeks compared with 41 weeks. There is limited evidence to show whether induction at an earlier point would help to avoid any of these additional stillbirths. The best evidence is from a meta-analysis⁴ which concluded that if 544 women had an induction before 42 weeks of pregnancy that would avoid one perinatal death. The authors of the review found no difference in any of the outcomes they measured whether the induction took place before or after 41 weeks of pregnancy, except that assisted births were more common with induction before 41 weeks than after.

The SWEPIS study was intended to recruit 10 038 women, which was calculated to be the number required to detect a statistically significant difference in the outcomes for the two groups. Even then, the authors were only expecting to be able to look for a difference in a ‘composite primary outcome’ which meant grouping together the total of stillbirths and neonatal deaths (referred to together as perinatal deaths) , and the number of babies showing signs of being in poor condition at birth or having various birth-associated health problems.

In fact the trial was stopped after recruiting only 2762 women (just over a quarter of the intended sample) because at that point there had been six perinatal deaths in the expectant management group (equalling about four deaths per 1000 pregnancies) and none in the induction group, which was calculated to be a significant difference.

Stopping the trial early means that the results are invalid. It is the equivalent of stopping a football match at half time because one team has already scored several goals, or a board game as someone has thrown several sixes. These extremes of statistical variation are to be expected to occur in some cases.

It is worth noting that the perinatal death rate was much higher than previous studies have found, and much higher than the authors expected to see. Because the trial was stopped early, there is no way of knowing whether this was due to chance with most of the perinatal deaths having happened to occur early on and in one group. It’s possible that had the trial continued the apparently high rate in the expectant management group compared to the induction group would have evened out or at least been reduced to the kind of level seen in other studies – but now we will never know.

It is also important to note that there was no significant difference between the two groups when looking at the ‘composite primary outcome’ as originally intended. Changing the intended measure means that an unwarranted value is attached to what may be a spurious result.

Despite saying “Although these results should be interpreted cautiously” and “It is not clear whether the results are broadly generalisable” the authors concluded that “induction of labour ought to be offered to women no later than at 41 weeks.” However, others may consider that this remains an open question – and of course the important word is “offered”.

References

1. Chippington Derrick D. & Higson N. “Labour Induction at Term – How great is the risk of refusing it?” AIMS Journal 31 (1) 2019
2. Wennerholm U-B et al “Induction of labour at 41 weeks versus expectant management and induction of labour at 42 weeks (Swedish Post-term Induction Study, SWEPIS): multicentre, open label, randomised, superiority trial” BMJ 2019; 367
3. Muglu J. et al “Risks of stillbirth and neonatal death with advancing gestation at term: A systematic review and meta-analysis of cohort studies of 15 million pregnancies.” PLOS Medicine July 2019
4. Middleton P. et al Cochrane review: Induction of labour at or beyond 37 weeks' gestation 2020