A review of trials comparing induction of labour with expectant management has concluded that evidence is lacking for the recommendation to induce labour at 41 weeks instead of 42 weeks for the improvement of perinatal outcome (Keulen et al 2018).
This is an important review and reminder, especially in the wake of recent studies looking at even earlier induction of labour. Keulen et al (2018) highlight some important points arising from their systematic review of 22 trials, including the fact that the risk of perinatal mortality between 41 and 42 weeks is low, there is no difference in caesarean section rate between induction at 41 or 42 weeks of pregnancy and is it thus unclear whether it would be better to recommend induction at 41 or 42 weeks. Expectant management, for anyone who doesn’t know, describes what happens where we await spontaneous labour while keeping an eye on things in case earlier intervention is warranted. Expectant management isn’t about saying an absolute ‘no’ to induction of labour; it’s the decision to ‘wait and see’; to tailor care to the needs of the individual woman and baby. No matter what the research says, of course, the decision is the woman’s, and the goal is to help women make the decision that is right for them.
One important issue that the authors of this review discuss is the methodological (or research design) problems that have to be overcome and taken into account when we are looking at the outcomes of studies and reviews on induction of labour and other birth-related interventions. Research studies aren’t carried out in a vacuum; they happen in the real world and are affected by human factors. It is very clear to those who design, carry out, use and appraise research that, sometimes, we see certain results or outcomes not because there is a clear difference between two courses of action in reality, or because one course of action is actually superior, but because the research design or something else that was going on at the same time either unwittingly or intentionally biased the outcomes. A good example of this is where studies show that induction of labour reduces the caesarean section rate. Many people who work on birth suites say that their reality is that induction of labour frequently leads to other interventions, and deeper analysis of reviews on this area also raises many questions, because the decision to do a caesarean section is a human one and not a hard outcome measure. Henci Goer has illustrated some of these issues really well in her analysis of the recent ARRIVE trial, Harvard professor Neal Shah also contributed to the conversation about how such trials can be biased. and ‘Why ARRIVE should not thrive in Australia’ was published in Woman and Birth. More recently, I spotted this article in Midwifery Today on the same topic and this one from Sarah Buckley.
Keulen et al (2018) have provided a really useful discussion of some of the issues that are debatable, and which may be interpreted differently by different researchers and reviewers. Their work further confirms that it isn’t good enough to simply cite a research study as evidence that a particular intervention is associated with a particular outcome. We need to dig deeper, think hard about the absolute and the relative risk of different options and carefully consider what was happening behind the scenes of these studies which might have affected the outcomes. That’s not because researchers are sloppy, uncaring or determined to reach a particular conclusion. It’s because research is carried out in our complex, messy, human world, which contains loads of factors that can affect the outcomes and leads us to need to engage our brains when thinking about whether the results are relevant to us and those we serve.
Background: Postterm pregnancy is associated with increased perinatal risk. The WHO defines postterm pregnancy as a pregnancy at or beyond 42 weeks +0 days, though currently labour is induced at 41 weeks in many settings. Guidelines on timing of labour induction are frequently based on the Cochrane systematic review ‘Induction of labour for improving birth outcomes for women at or beyond term’ in which is concluded that a policy of induction of labour is associated with fewer adverse perinatal outcome and fewer Caesarean sections. However, the included trials differed regarding the timing of induction, ranging from 39 to beyond 42 weeks while the upper limit of expectant management exceeded a gestational age of 42 weeks in most studies.
Objective: to evaluate perinatal mortality, meconium aspiration syndrome and Caesarean section rate of trials comparing a policy of elective induction of labour and expectant management according to timeframes of comparison with a focus on studies within the 41-42 weeks’ timeframe.
Design: Review. Methods: The systematic review of Cochrane was used as a starting point for assessing relevant trials and a search was performed for additional recent trials. We evaluated incidence and causes of perinatal mortality, incidence of meconium aspiration syndrome and Caesarean section according to three time frames of comparison. We pooled estimates and heterogeneity was tested. The quality of the included trials was assessed using the Quality Assessment Tool for Quantative Studies (EPHPP).
Findings: In total 22 trials were included which had all different timeframes of comparison. Only one trial compared induction of labour at 41 weeks +0-2 days with induction at 42 weeks +0 days, three other trials compared induction of labour at 41 weeks +0-6 days with induction at 42 weeks +0-6 days. In 18 trials the comparison was outside the 41-42 weeks’ timeframe: in six trials induction was planned ≤40 weeks and in another 12 trials expectant management was beyond 43 weeks. The incidence of potentially gestational age associated perinatal mortality between 41 and 42 weeks was 0/2.444 [0%] (induction) versus 4/2.452 [0.16%] (expectant management), NNT 613; 95%CI 613 – infinite. Two trials in the timeframe of comparison 41-42 weeks were available for evaluation of meconium aspiration syndrome (6/554 (induction) versus 14/554 (expectant management), RR 0.44; 95%CI 0.17-1.16). Three trials in the timeframe 41-42 weeks could be evaluated for Caesarean section, with different inclusion criteria regarding Bishop score. There was no significant difference in the Caesarean section rate 93/629 (induction) versus 106/629 (expectant management), RR 0.88; 95%CI 0.68-1.13.
Conclusion: Evidence is lacking for the recommendation to induce labour at 41 weeks instead of 42 weeks for the improvement of perinatal outcome. More studies comparing both timeframes with an adequate sample size are needed to establish the optimal timing of induction of labour in late-term pregnancies.