For those of you struggling with the news sweeping the internet that a trial has definitively established that women should be routinely induced at 39 weeks, a new study offers push back. (FYI: I deconstructed the ARRIVE trial and its hidden assumptions for Science & Sensibility.) A large, well-conducted study takes a fresh look at the issue and fails to find that routinely inducing labor prior to 42 weeks, let alone earlier, improves outcomes and reduces cesarean rates. Let’s take a look.
The Set-Up
Prompted by concern that one in four Finnish pregnancies continues past the beginning of the 41st week, which means substantial numbers of Finnish women will undergo labor induction with a policy of inducing at 41 weeks, investigators decided to take a new approach that would avoid the fallibilities of both conventional observational studies and randomized controlled trials. A big problem in their eyes was that observational studies may have unrecognized confounding factors. For this reason, they conducted a propensity study, a study design that, ideally, produces matched pairs differing only in that one member of the pair received the treatment under study. Outcome data is then analyzed according to the pairs.
The investigators chose the years between 2006 and 2012 because the usual policy during that time period was to wait until the beginning of the 42nd week before inducing for postdates. The study population consisted of 212,716 births occurring after 40 wk 0 day and before 42 wk 1 day in which there was a single, head down baby who was alive at labor onset. (Planned cesareans were excluded.) The labor induction rate overall was 16%, and the cesarean and instrumental delivery rates were both 10%. Not quite half the population (46%) was 1st-time mothers.
Investigators divided the population into 5 groups, each encompassing a 3-day period:
- Group 1: 40+0 to 40+2
- Group 2: 40+3 to 40+5
- Group 3: 40+6 to 41+1
- Group 4: 41+2 to 41+4
- Group 5: 41+5 to 42+0
They then compared outcomes for induced labors within each 3-day period with labors beginning spontaneously during the same time period plus expectant management beyond the time period, which could mean either spontaneous labor onset or subsequent labor induction.
Women undergoing labor induction were matched with women having expectant management according to 1st or subsequent birth; maternal age; BMI; smoking; previous cesarean; infertility treatment; labor unit; and year, month, and weekday of birth. They were also matched according to 450 ICD-10 diagnostic codes, which were grouped into 12 categories. These were abnormal fetal growth or disproportion; abnormal presentation; eclamptic diseases, hypertonia; diabetic diseases; infection or sexually transmitted disease; suspected or known fetal problems; placental disorders; pregnancy with other or unspecified complications; very low or no risk increase; low risk increase; moderate risk increase; and substantial risk increase.
The Findings
Neither perinatal mortality nor deaths during labor differed between groups. In addition, the 125 antenatal deaths that were excluded from the study population calculated to a rate of 6 per 10,000 in the population overall. This means that the study’s findings disconfirm the primary reason for routine induction before 42 weeks, which is that it will reduce perinatal deaths. Induction had no effect on intrapartum and neonatal deaths compared with expectant management, and the extremely low rate of antenatal deaths means that routine induction could have no meaningful effect on antenatal deaths either. Furthermore, if that is true in a population in which most inductions would have been for medical indication, it would certainly be true when inducing healthy women. Induction did have one benefit: it reduced meconium aspiration syndrome in groups 1-4 by around 60%; however, the excess in expectantly managed women was small (range: 2-3 per 1000), and it didn’t translate into an increased need for respiratory assistance or an increase in prolonged hospitalization (> 7 days). In fact, babies born after induced labors in groups 1 and 2 were at 20-30% increased risk of prolonged hospitalization, although as with meconium aspiration syndrome, the absolute difference was small (1 per 1000). Again, this population included women with health problems and pregnancy complications, so differences would likely be negligible in healthy women.
As for maternal outcomes, while the incidence of anal sphincter injury increased from 8-10 per 1000 in group 1 to 12 per 1000 in group 5, rates in expectantly managed women were similar to rates in induced women, which means that induction at an earlier gestational age provided no benefit. On the other hand, inducing labor increased the cesarean rate by about 20% in groups 3 and 4. Here, too, though, the absolute difference was small (2.5%), although the gap might have been wider had the study been confined to 1st-time mothers. Women with prior vaginal births are much less vulnerable to the effects of factors that inhibit labor progress.
The investigators conclude:
An intervention should be proven beneficial, not only non-harmful in order to be recommended over a wait-and-see policy. . . . The results . . . call into question the increasingly proactive management of prolonged pregnancy. To do no harm, we regard expectant management as a justified treatment option, at least until close to 42 weeks (pp. 448, 451).
Unfortunately, that ship may already have sailed. The idea that induction for impending postdates improves newborn outcomes and prevents antenatal deaths is firmly embedded in the obstetric consciousness, and 39-week induction is being proposed at this year’s American College of Obstetricians & Gynecologists’ annual meeting as a strategy for reducing cesareans.
The Take-Away
It is difficult to determine the applicability of this study to U.S. women. For one thing, the study doesn’t report episiotomy rates, and any episiotomies would have been mediolateral episiotomies, which are much less likely to extend into the anal sphincter. For another, as we saw above, Finnish practitioners have a much lower propensity to induce labor and perform cesareans than those in the U.S. Without a population of similar characteristics with which to perform an apples-to-apples comparison, we have no way to gauge the effect of U.S. obstetricians’ much greater propensity to do both. One thing remains clear, though, this study joins others that throw doubt on the validity of routine induction prior to 42 weeks.
###
