by Henci Goer

Commentary on: Crowley P. Interventions for preventing or improving the outcome of delivery at or beyond term. (Cochrane Review). In: The Cochrane Library. Issue 2. Oxford: Update Software, 2002. [Abstract]

Study design and results: systematic review with accompanying meta-analysis of 19 randomized controlled trials (RCTs) of routine induction versus expectant management. It concludes that routine induction of labor at 41 weeks reduces the perinatal mortality rate without increasing the cesarean rate.

• perinatal mortality: 0.2 per 1000 (1/4122) routine induction vs. 2.4 per 1000 (9/3803) expectant management; relative risk 0.20, 95% confidence interval 0.06-0.90
• cesarean section rates after 41 weeks: 19.6% routine induction vs. 21.7% expectant management; relative risk 0.87, 95% confidence interval 0.77-0.99

Problems include but are not limited to the following:

• The individual studies that the authors review have multiple, major problems: If the constituent trials are flawed, the review's conclusions will be equally flawed, a principle known as “garbage in, garbage out.”
  
  
• crossover: Many women in the expectant management arms of trials were induced while many in the labor induction groups began labor spontaneously. Investigators in randomized controlled trials (RCTs) keep participants in their assigned groups when they analyze data regardless of which treatment they received. This is referred to as “intent to treat” analysis. In conventional RCTs—drug A versus B or procedure A versus B—crossover is low, and analysis by “intent to treat” maintains the advantages of random assignment. But high crossover rates flatten out the differences between groups.
  
  
• high cesarean rates in both arms of the trial: This creates the illusion that inducing labor doesn't make a difference. High cesarean rates result from obstetric management, but many obstetricians assume them to be intrinsic to birth.
  
  
• combining data from first-time mothers and women with previous vaginal births: Because inducing labor has little if any effect on cesarean rates in multiparous women, not reporting cesarean rates separately will mask the major effect on cesarean rate in primiparous women.
  
  
• expectant management not really expectant management: In most studies, labor was induced when the participants met some different set of criteria from the routine induction group. In some studies women in the so-called “expectant management” group were induced when they attained a ripe cervix. In others, they were induced at 42 weeks while the “routine induction” women were induced at 41 weeks.  Inducing labor carries with it a high potential for complications.
  
  
• women expectantly managed had fetal surveillance testing: These tests have high false positive rates, that is, the test indicates a problem, but the baby is actually fine. Nonetheless, a woman with a non-reassuring test would then be induced under circumstances where the obstetrician believed the baby to be compromised. The heightened concern would increase the likelihood of cesarean. In some cases, the obstetrician might opt to proceed directly to a cesarean rather than subject a supposedly compromised baby to the known extra stresses of an induced labor.
  
  
• The gap in perinatal mortality is narrower than the systematic review makes it appear: Two of the 9 deaths in the await-labor arm of the meta-analysis occurred before 41 weeks' gestation. A policy of routine induction at 41 weeks would not have prevented them. Two deaths occurred in a 1960s study of only 112 women, of whom 57 were assigned to expectant management. That calculates to a perinatal mortality rate of 35.1 per 1000 in the expectant management group, a 15-fold difference from the 2.4 per 1000 aggregate rate in the systematic review. So enormous a discrepancy indicates that this study and those deaths should not have been included. Either the population studied, the obstetrics of that era, or both clearly cannot be extrapolated to contemporary women receiving modern-day care. One death occurred at 43 weeks + 3 days. This may make an argument for routine induction at 42 or 43 weeks, but it is irrelevant to an argument for induction at 41 weeks. That leaves 4 perinatal deaths in 3746 women versus 1 in 4067 women, or a perinatal mortality rate of 1.1 per 1000 in the expectant management group versus 0.2 per 1000 in the routine induction group, a difference of less than 1 per 1000 between groups. A difference this small may be due to chance.

Comment: A major conceptual problem with routine induction at 41 weeks is that the median length of pregnancy in healthy first-time mothers is 41 weeks 1 day. The conventional 40 weeks is just that: a convention. It is based on nothing more than a German obstetrician's fiat two centuries ago that since women cycle according to the moon, pregnancy lasts 10 moon months, that is, 10 months of 4 weeks each. Practitioners may argue over how great a deviation from normal warrants intervention, but in the case of routine induction at 41 weeks, they are arguing for intervening when there is no deviation from normal. The same study that reported a 41 week 1 day median pregnancy length in primiparous women found a 40 week 3 day average pregnancy length in women who had had babies before. First-time mothers are notoriously more likely to have problem labors and cesarean sections than multiparous women. This means that the increasing complication rates and cesarean rates seen with advancing gestational length may well be nothing more than an artifact created by having a higher and higher proportion of primiparous women in the mix as the days roll by after 40 weeks.

Practice philosophy aside, a policy of routine induction at 41 weeks produces more than a conceptual problem. Primiparous women have roughly double the risk of having an induced labor end in a c-section. A policy of routine induction at 41 weeks exposes large numbers of a vulnerable population to a greatly heightened risk of surgical delivery with all of the attendant problems of a major operation as well as all the future reproductive consequences of having a uterine scar. In addition, crowding the labor ward with women undergoing an unnecessary intervention means there may be no room for a woman who really needs care. In their paper criticizing routine 41-week induction, Menticoglou and Hall (2002) cite a case where admission was delayed for a pregnant woman requiring IV antihypertensive drugs for severe hypertension because no beds were available. Several were filled with women undergoing routine 41-week inductions. The woman died of a stroke before she could be admitted. To quote Menticoglou and Hall's conclusion: “Routine induction at 41 weeks is ritual induction at term, unsupported by rational evidence of benefit. It is unacceptable, illogical and unsupportable interference with a normal physiologic situation.”

Bibliography:

Childbirth Connection. What should I know about cesarean section? 2004. Available at: http://www.childbirthconnection.org/article.asp?ClickedLink=274&ck=10168&area=27

Hannah ME et al. Induction of labor as compared with serial antenatal monitoring in
post-term pregnancy: a randomized controlled trial. N Engl J Med 1992;326(24):1587-1592.

Hannah ME et al. Postterm pregnancy: putting the merits of a policy of induction of labor into perspective. Birth 1996;23(1):13-9.

Menticoglou SM and Hall PF. Routine induction of labour at 41 weeks gestation: nonsensus consensus. BJOG 2002;109:485-91.

Mittendorf R et al. The length of uncomplicated human gestation. Obstet Gynecol 1990;75(6):929-932.