What are we to make of a new Swedish trial in which women were allocated by chance to either routine induction at 41 weeks or expectant management until 42 weeks (Wennerholm 2019)? The trial was stopped midway through because it had already shown a statistically significant increase, meaning unlikely to be due to chance, in perinatal mortality in the expectant management arm of the trial, and routine induction hadn’t increased the cesarean rate.  You may be thinking, “It has long been established that 41-week induction is the better strategy for just these reasons. The Swedish trial just adds to the pile of studies finding in its favor.”

I would argue that matters aren’t nearly as clear cut as they seem. I’d like to walk you through the evidence, both pro and con. In particular, I’ll compare and contrast two systematic reviews that came to differing conclusions regarding the effect of routine induction on perinatal mortality (Middleton 2018; Rydahl 2019a) and do the same with a Dutch trial that likewise disagreed with the Swedish trial (Keulen 2019). I’d also like to examine the evidence for and against an increase in the cesarean rate. As a bonus, I think our exploration will provide some nuggets of useful information that could help pregnant people determine which course of action is right for them.

Does Routine Induction Reduce Perinatal Mortality?

Let’s start with the systematic reviews. A systematic review is an aggregate of studies on a particular topic.  It is often, although not necessarily, confined to randomized controlled trials.  Randomized controlled trials are studies in which people are allocated by chance to one form of treatment or another (or to no treatment). The great strength of systematic reviews is that they allow researchers to look at multiple studies of the same issue, and if the design of the studies is similar enough, they allow their data to be pooled into a meta-analysis, thereby strengthening conclusions.

Perhaps the most respected collection of systematic reviews is the Cochrane Library, which, I should note, confines included studies to randomized controlled trials. The most recent iteration of the Cochrane review of routine induction, “Induction of labour for improving birth outcomes for women at or beyond term” (Middleton 2018), concluded that routine induction decreased perinatal mortality by 2-3 per 1000 pregnancies compared with expectant management (2/4988 vs. 16/4972 or 0.4 per 1000 vs. 3 per 1000) without increasing the cesarean rate. Reducing perinatal mortality would seem definitive in favor of routine induction at term, especially in the absence of a rise in cesareans . . . but is the case truly closed?

 The authors of “Effects of induction of labor prior to post-term in low-risk pregnancies: a systematic review” say otherwise (Rydahl 2019a). They argue that the Cochrane review didn’t address the pertinent research question, which is: “Does routine induction in the 41^st week in women and babies at low risk for complications cared for in a modern medical system reduce preventable perinatal deaths?” The Cochrane reviewers include deaths that occurred in a trial published in 1969, a trial conducted in Tunisia, and a trial that wasn’t restricted to low-risk women. They also count perinatal deaths that couldn’t have been prevented by induction in the 41^st week because they occurred before 41 weeks or after 42 weeks, and they include deaths that would have occurred under any circumstances because they occurred in babies with congenital malformations or because of a complication (a true knot in the cord) unrelated to the timing of labor.   

In contrast, Rydahl and colleagues restrict their review to studies published within the last 20 years that compare routine induction at 41 weeks + 0-6 days vs. routine induction at 42 weeks in uncomplicated pregnancies, and they exclude deaths unrelated to whether labor was induced or expectantly managed. Their review includes both observational studies and randomized controlled trials. Reporting outcomes solely from the trials, their analysis finds a reduction of 1-2 per 1000 in perinatal mortality (1/2582 vs. 5/2517 or 0.4 per 1000 vs. 2 per 1000), a smaller difference than the 2-3 per 1000 reduction found in the Cochrane review, and one that fails to achieve statistical significance.

Similar to the disagreement between the two systematic reviews, we have a Dutch trial of similar design and published in the same year whose results contradict the Swedish trial. The Swedish trial reports 0 perinatal deaths in 1381 women allocated to routine 41-week induction versus 6 deaths in 1379 women allocated to expectant management (0 vs. 4 per 1000) (Wennerholm 2019).  One of the deaths, however, was in a baby with a heart anomaly and another in a baby who was small for gestational age. If we exclude these deaths on the grounds that the former death was unrelated to the timing of labor and the latter was not a low-risk pregnancy, we now have 4 deaths in 1379 expectantly managed women or a perinatal mortality rate of 0 versus 3 per 1000. The Dutch trial reports one perinatal death in 900 women in the induction group versus two in 901 women allocated to expectant management (1 per 1000 vs. 2 per 1000). This is a difference of only 1 per 1000, and it does not achieve statistical significance (Keulen 2019). As with the Swedish trial, however, one of the deaths in the expectant management group was in a baby who was small for gestational age, again, arguably not a low-risk pregnancy. Excluding this death makes the rates in the Dutch trial identical at 1 per 1000 in both the routine induction and expectant management groups.

These two trials would have fit the inclusion criteria for the Rydahl (2019a) systematic review. If we pool their data with the data from the trials in the Rydahl systematic review, we have 2 deaths in 4863 babies in the routine 41-week induction group compared with 10 in 4797 babies in the expectantly managed group or 0.4 per 1000 versus 2 per 1000 for a difference of 1-2 per 1000, the same difference as reported in the Rydahl review alone.

The reviews and the trials are not the only studies addressing the perinatal mortality controversy. Rydahl and colleagues (2019) took a second look at an analysis of the stillbirth rate in Denmark before and after a national change in policy to induce labor routinely at 41 weeks + 3-5 days instead of 42 weeks. The analysis had found that the stillbirth rate declined after introduction of the new policy. Rydahl and colleagues (one of whom is our beloved Gene Declercq) point out that the analysis failed to take into account the declining stillbirth rate in Denmark. Rydahl plotted the declining rate before the introduction of the new policy, projected it forward, and found that while the induction rate leapt from 41% to 65% at the time of the policy change, the stillbirth and perinatal mortality rates after the policy change exactly matched the projected tren­­d. In other words, the change in induction policy had had no effect whatsoever on the perinatal mortality rate. Finnish investigators undertook a propensity study to determine whether induction prior to 42 weeks improved outcomes (Pyykonen 2018). A propensity study uses large numbers of participants, in this case nearly 213,000 births of a single, head-down baby between 40 weeks 0 days and 42 weeks 0 days, to produce matched pairs according to so many factors that each pair differs only in that one member of the pair received the treatment under study and the other didn’t.[SM1] [SM2]  Outcome data is then analyzed according to the pairs. Investigators divided the time period into five 3-day periods and compared induction within each time period with ongoing pregnancy beyond that period. Neither intrapartum stillbirth nor neonatal death rates were improved by induction.     

All that being said, it could still be argued that you can’t have an antenatal death between week 41 and 42 if the baby is delivered at the beginning of week 41, so it comes down to whether routine induction has potential harms, most notably, whether it increases the likelihood of cesarean delivery. Let’s look at that next.    

Does Routine Induction Affect Cesarean Deliveries?

The Rydahl (2019a) systematic review reports that 71% of women allocated to expectant management began labor spontaneously in the 41^st week. (The rest were either induced for medical reasons or induced routinely at 42 weeks.) If induction increases cesarean deliveries compared with spontaneous labor onset, the potential effect of a 41-week induction policy could be considerable.

As with perinatal mortality, the studies already discussed above disagree on the impact of routine induction on cesarean delivery. The Cochrane review finds a statistically significant, but clinically meaningless, reduction with routine induction (17% vs. 18%) (Middleton 2018). Rydahl (2019a) reports cesarean outcomes separately for trials and observational studies, finding a small excess with induction in the pooled data from four trials (17% vs. 15%) and in a large U.S. observational study of 1^st-time mothers (36% vs. 32%), although it’s hard to interpret these last numbers, considering how extraordinarily high these rates are for a population of healthy women who, regardless of being 1^st-time mothers should have been at much lower risk for cesarean. In the Swedish and Dutch trials, rates are virtually identical (10-11%) (Keulen 2019; Wennerholm 2019). The re-analysis of Danish data that plotted trends before and after a change in induction policy reports that the change had no effect on cesarean rate (Rydahl 2019).

Based on these studies, it appears that the effect of routine induction on cesarean rate is unclear but modest at best. Furthermore, the extreme variation in cesarean rates depending on where the study took place suggests that other factors outweigh the effect of induction in determining likelihood of cesarean whether in general or in 1^st-time mothers. Rates are much lower regardless of whether labor is induced or begins on its own in countries with lower national cesarean rates, such as Sweden (17%), the Netherlands (16%), or Denmark (21%), and higher in a country with a high cesarean rate, such as the U.S. (32%) (OECD 2017).

We have other studies, though, that find routine induction at 41 weeks not so benign. Two studies looked at the effect of an institutional change in policy. A Spanish study evaluated a change to inducing in week 41 rather than at 42 weeks (Burgos 2012). The cesarean rate increased from 11% before the change to 14% afterward, or 3 more cesareans per 100 women. It could be argued that like the before-and-after study critiqued earlier, the increase might just reflect an ongoing trend, except that in this case, the cesarean rates with spontaneous labor onset (8-9%) and with induction (19%) remained the same. The cesarean rate overall went up because the percentage of women who were induced rose from 29% to 50%. The other study, a U.K. study restricted to 1^st-time mothers, looked at a policy change from inducing at 41 weeks 3 days to inducing at 42 weeks (Kassab 2011). The shift to inducing at 42 weeks decreased the cesarean rate from 33% to 24%, or 9 fewer 1^st-time mothers per 100 having a cesarean. The “ongoing trend” argument wouldn’t apply here either because cesarean rates in the U.K. weren’t declining (OECD 2017).  

Studies of various designs have looked at the effect of routinely inducing labor in week 41. One found that the cesarean rate in 1^st-time mothers was 12% when labor started on its own versus 22% when labor was induced, or 10 more 1^st-time mothers per 100 having a cesarean (Mahomed 2016). Another found that even shifting the timing of planned induction from early in week 41 to late in the week had a major effect on the percentage of women being induced (92% vs. 37%) and the cesarean rate (31% vs. 20%) with 11 fewer women per 100 having a cesarean (Oros 2012). A third study, also of 1^st-time mothers, was conducted at a hospital where doctors noted whether their patients were scheduled for routine induction at 41 weeks or not until 42 weeks (Pavicic 2009). Investigators created an expectant-management proxy group by combining all women with planned expectant management until 42 weeks with women for whom induction was planned at 41 weeks but who began labor on their own before the induction was carried out. They then compared the cesarean rate with that of women induced in week 41 as planned. The cesarean rate in the proxy planned expectant management group was 17% vs. 25% in women routinely induced at 41 weeks, or 8 more 1^st-time mothers per 100.

All in all, then, it looks like routine induction at 41 weeks may provoke a meaningful increase in cesarean rates, especially in 1^st-time mothers. An increased risk of cesarean surgery and its attendant harms should be weighed against any benefits of routine induction.

The Take-Away

As we have seen, the choice isn’t as clear as it has been represented. Routine induction at 41 weeks and expectant management until 42 weeks are both reasonable choices. We can best serve the families we work with by helping them determine whether they are good candidates for expectant management, if that is their preference, and advising them on optimal care practices and policies if they prefer to be induced.

For those who prefer to avoid routine induction before 42 weeks, here are a couple of ideas that may help people with apparently uncomplicated pregnancies determine with their care providers whether they’re good candidates for expectant management:

• Late-pregnancy ultrasound scan to confirm uncomplicated pregnancy: Despite congenital anomalies being an exclusion factor in both the Swedish & Dutch trials, one intrapartum death occurred in a baby with an undiagnosed heart defect in the Swedish trial (Wennerholm 2019), and in the Dutch trial (Keulen 2019), 6 of the 11 children admitted to neonatal intensive care had a severe congenital anomaly. Two of the deaths, one in each trial, occurred in babies who were small for gestational age. We don’t know, of course, whether induction would have made a difference in outcome in these cases, but it’s something to factor into a decision.
• Prenatal testing to evaluate fetal wellbeing: In point of fact, systematic reviews of the various prenatal fetal surveillance tests have failed to find benefit (Grivell 2015; Lalor 2008; Morris 2014; Norman 2018), but considered thoughtfully and as part of an overall picture, they may still prove of use. Using the biophysical profile, which measures more than one parameter, or repeating a test or doing a different test if the result suggests there’s a problem reduces the probability of a false-positive result. A single finding that slightly deviates from normal may not be concerning but finding an increasing trend in the wrong direction with serial testing or finding it in combination with other suboptimal maternal or fetal factors would be more worrisome.

To maximize the probability of a complication-free induction and a vaginal birth:

• In the absence of an acute medical problem, wait for a favorable cervix. Studies consistently find that inducing 1^st-time mothers with an unfavorable cervix substantially increases the likelihood of cesarean despite use of cervical ripening agents whereas inducing with a ripe cervix does not (Dunne 2009; Gibson 2014; Le Ray 2007; Laughon 2012; Marroquin 2013; Vahratian 2005).
• If the cervix is favorable, go straight to oxytocin. Cervical ripening agents aren’t necessary and omitting them avoids the possibility of experiencing their adverse effects.
• The oxytocin protocol should “start low & go slow.” The Pitocin package insert includes a recommended regimen.
• Refrain from rupturing membranes before active labor. If membranes are intact and the induction isn’t working, it can be stopped and tried another day. Once membranes are ruptured, that’s no longer an option.
• Turn off the oxytocin drip once labor reaches active phase. In many cases, when the drip is turned off, labor will continue under its own steam. If it doesn’t, the drip can always be turned back on (Saccone 2017). Discontinuing oxytocin decreases likelihood of abnormal contractions and abnormal fetal heart rate and increases likelihood of vaginal birth (Bor 2016; Saccone 2017)
• Have patience. Induced labors tend to take longer than labors that start on their own (Blackwell 2008; Harper 2012; Ostborg 2017; Rinehart 2000; Simon 2005; Vahratian 2005), especially in 1^st-time mothers, but longer labors don’t result in worse maternal or newborn outcomes (Blackwell 2008; Simon 2005).

References

Blackwell, S. C., Refuerzo, J., Chadha, R., & Samson, J. (2008). Duration of labor induction in nulliparous women at term: how long is long enough? American Journal of Perinatology, 25(4), 205-209.

Bor, P., Ledertoug, S., Boie, S., Knoblauch, N. O., & Stornes, I. (2016). Continuation versus discontinuation of oxytocin infusion during the active phase of labour: a randomised controlled trial. BJOG, 123(1), 129-135.

Burgos, J., Rodriguez, L., Otero, B., Cobos, P., Osuna, C., Centeno Mdel, M., . . . Martinez-Astorquiza, T. (2012). Induction at 41 weeks increases the risk of caesarean section in a hospital with a low rate of caesarean sections. J Matern Fetal Neonatal Med, 25(9), 1716-1718.

Dunne, C., Da Silva, O., Schmidt, G., & Natale, R. (2009). Outcomes of elective labour induction and elective caesarean section in low-risk pregnancies between 37 and 41 weeks' gestation. J Obstet Gynaecol Can, 31(12), 1124-1130.

Gibson, K. S., Waters, T. P., & Bailit, J. L. (2014). Maternal and neonatal outcomes in electively induced low-risk term pregnancies. American Journal of Obstetrics and Gynecology, 211(3), 249 e241-249 e216.

Grivell, R. M., Alfirevic, Z., Gyte, G. M., & Devane, D. (2015). Antenatal cardiotocography for fetal assessment. Cochrane Database Syst Rev(9), CD007863.

Harper, L. M., Caughey, A. B., Odibo, A. O., Roehl, K. A., Zhao, Q., & Cahill, A. G. (2012). Normal progress of induced labor. Obstetrics and Gynecology, 119(6), 1113-1118.

Kassab, A., Tucker, A., El-Bialy, G., Mustafa, M., Macaulay, J., & Fox, R. (2011). Comparison of two policies for induction of labour postdates. Journal of Obstetrics and Gynaecology, 31(1), 32-36.

Keulen, J. K., Bruinsma, A., Kortekaas, J. C., van Dillen, J., Bossuyt, P. M., Oudijk, M. A., . . . de Miranda, E. (2019). Induction of labour at 41 weeks versus expectant management until 42 weeks (INDEX): multicentre, randomised non-inferiority trial. BMJ, 364, l344. https://www.bmj.com/content/364/bmj.l344.long

Lalor, J. G., Fawole, B., Alfirevic, Z., & Devane, D. (2008). Biophysical profile for fetal assessment in high risk pregnancies. Cochrane Database Syst Rev(1), CD000038.

Laughon, S. K., Zhang, J., Grewal, J., Sundaram, R., Beaver, J., & Reddy, U. M. (2012). Induction of labor in a contemporary obstetric cohort. American Journal of Obstetrics and Gynecology, 206(6), 486 e481-489.

Le Ray, C., Carayol, M., Breart, G., & Goffinet, F. (2007). Elective induction of labor: failure to follow guidelines and risk of cesarean delivery. Acta Obstetricia et Gynecologica Scandinavica, 86(6), 657-665.

Mahomed, K., Pungsornruk, K., & Gibbons, K. (2016). Induction of labour for postdates in nulliparous women with uncomplicated pregnancy - is the caesarean section rate really lower? Journal of Obstetrics and Gynaecology, 36(7), 916-920.

Marroquin, G. A., Tudorica, N., Salafia, C. M., Hecht, R., & Mikhail, M. (2013). Induction of labor at 41 weeks of pregnancy among primiparas with an unfavorable Bishop score. Archives of Gynecology and Obstetrics, 288(5), 989-993.

Middleton, P., Shepherd, E., & Crowther, C. A. (2018). Induction of labour for improving birth outcomes for women at or beyond term. Cochrane Database Syst Rev, 5, CD004945. https://www.cochranelibrary.com/cdsr/doi/10.1002/14651858.CD004945.pub4/full?highlightAbstract=improving%7Cinduction%7Coutcom%7Coutcomes%7Cwithdrawn%7Cbirth%7Cimprov%7Cinduct

Morris, R. K., Meller, C. H., Tamblyn, J., Malin, G. M., Riley, R. D., Kilby, M. D., . . . Khan, K. S. (2014). Association and prediction of amniotic fluid measurements for adverse pregnancy outcome: systematic review and meta-analysis. BJOG, 121(6), 686-699.

OECD. (2017). “Caesarean sections” in Health at a Glance 2017: OECD Indicators. Paris: OECD Publishing. Retrieved from https://www.oecd-ilibrary.org/social-issues-migration-health/health-at-a-glance-2017_health_glance-2017-en.

Oros, D., Bejarano, M. P., Cardiel, M. R., Oros-Espinosa, D., Gonzalez de Aguero, R., & Fabre, E. (2012). Low-risk pregnancy at 41 weeks: when should we induce labor? J Matern Fetal Neonatal Med, 25(6), 728-731.

Ostborg, T. B., Romundstad, P. R., & Eggebo, T. M. (2017). Duration of the active phase of labor in spontaneous and induced labors. Acta Obstetricia et Gynecologica Scandinavica, 96(1), 120-127.

Pavicic, H., Hamelin, K., & Menticoglou, S. M. (2009). Does routine induction of labour at 41 weeks really reduce the rate of caesarean section compared with expectant management? J Obstet Gynaecol Can, 31(7), 621-626.

Pyykonen, A., Tapper, A. M., Gissler, M., Haukka, J., Petaja, J., & Lehtonen, L. (2018). Propensity score method for analyzing the effect of labor induction in prolonged pregnancy. Acta Obstetricia et Gynecologica Scandinavica, 97(4), 445-453.

Rinehart, B. K., Terrone, D. A., Hudson, C., Isler, C. M., Larmon, J. E., & Perry, K. G., Jr. (2000). Lack of utility of standard labor curves in the prediction of progression during labor induction. American Journal of Obstetrics and Gynecology, 182(6), 1520-1526.

Rydahl, E., Declercq, E., Juhl, M., & Maimburg, R. D. (2019). Routine induction in late-term pregnancies: follow-up of a Danish induction of labour paradigm. BMJ Open, 9(12), e032815. https://bmjopen.bmj.com/content/9/12/e032815.long

Rydahl, E., Eriksen, L., & Juhl, M. (2019a). Effects of induction of labor prior to post-term in low-risk pregnancies: a systematic review. JBI Database System Rev Implement Rep, 17(2), 170-208. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6382053/

Saccone, G., Ciardulli, A., Baxter, J. K., Quinones, J. N., Diven, L. C., Pinar, B., . . . Berghella, V. (2017). Discontinuing Oxytocin Infusion in the Active Phase of Labor: A Systematic Review and Meta-analysis. Obstetrics and Gynecology, 130(5), 1090-1096.

Simon, C. E., & Grobman, W. A. (2005). When has an induction failed? Obstetrics and Gynecology, 105(4), 705-709.

Vahratian, A., Zhang, J., Troendle, J. F., Sciscione, A. C., & Hoffman, M. K. (2005). Labor progression and risk of cesarean delivery in electively induced nulliparas. Obstetrics and Gynecology, 105(4), 698-704.

Wennerholm, U. B., Saltvedt, S., Wessberg, A., Alkmark, M., Bergh, C., Wendel, S. B., . . . Hagberg, H. (2019). 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, 367, l6131. https://www.bmj.com/content/367/bmj.l6131.long

About Henci Goer

Henci Goer, award-winning medical writer, and internationally known speaker is an acknowledged expert on evidence-based maternity care. Her first book, Obstetric Myths Versus Research Realities, was a valued resource for childbirth professionals. Its successor, Optimal Care in Childbirth: The Case for a Physiologic Approach, won the American College of Nurse-Midwives “Best Book of the Year” award. Goer has also written The Thinking Woman's Guide to a Better Birth, which gives pregnant women access to the research evidence, as well as consumer education pamphlets and articles for trade, consumer, and academic periodicals; and she posts regularly on Lamaze International’s Connecting the Dots. Goer is founder and director of Childbirth U, a website offering narrated slide lectures to help pregnant women make informed decisions and obtain optimal care for themselves and their babies.

Tags