The trusted source for
healthcare information and
A Best-Evidence Review of Indications for Induction of Labor
Abstract & Commentary
By John C. Hobbins, MD, Professor and Chief of Obstetrics, University of Colorado Health Sciences Center, Denver, is Associate Editor for OB/GYN Clinical Alert.
Dr. Hobbins reports no financial relationship to this field of study.
Synopsis: To review the evidence supporting indications for inductions of labor.
Source: Mozurkewich E, et al. Indications for induction of labour: A best-evidence review. BJOG 2009;116:626-636; doi: 10.1111/j.1471-0528.2008.02065.x.
Fortunately, at a time when there is an emphasis on evidence-based reasons for medical decisions, more reviews are appearing in the literature looking at various common interventions and the evidence available to support these practices. In the February issue of the British Journal of Obstetrics and Gynaecology a "best-evidence review" focused on the indications for induction of labor. In the small space available here, I will limit the discussion to only a few of the indications that have the greatest impact on the way we practice.
The authors sifted through MEDLINE and the Cochrane databases to pull articles from 1980 through 2008 that dealt with the most common reasons for induction. They started with 1387 abstracts and 418 articles, and, after subjecting each one to scientific scrutiny, they coned down to 34 submissions — some reviews, some randomized and observational trials, and meta-analyses. The following represents a very condensed version of the authors' findings and conclusions regarding induction of labor for some of the most common indications.
Most of the studies lumped together pregnancies after 41 weeks with those over 42 weeks, but there did appear to be less perinatal morbidity and meconium aspiration with induction over expectant management (with fetal heart rate monitoring). Induction did not increase the cesarean section rate. Interestingly, in one study, 74% of those having inductions said they would choose this again under the same circumstances, while only 32% in the "watchful waiting" group wanted to do this next time.
Rupture of Membranes at Term
This can be handled quickly. Studies comparing expectant management to immediate induction evaluated waiting periods from 24 hours to 4 days for spontaneous labor to ensue. At term, "expedited" induction resulted in less endometritis, less amnionitis, and fewer admissions to the NICU (12% vs 17%).
Preterm Premature Rupture of Membranes
Even though it is a commonly accepted practice to induce labor at 34 weeks when premature rupture occurs, surprisingly, there are very few (4) randomized trials dealing with the efficacy of inducing labor, rather than awaiting spontaneous labor, in the face of ruptured membranes between 30 and 36 weeks. Two studies required documentation of pulmonic maturity prior to induction, and the remaining two studies did not.1,2 Also, antibiotics and steroids were not given in any of the above studies.
Other than a decrease in maternal hospital days and a decreased incidence of the amnionitis, there were no real differences in neonatal sepsis, respiratory distress syndrome (RDS), intraventricular hemorrhage, or days in the NICU.
The authors concluded that there was "moderate" evidence for the induction of labor in this group of patients, but the authors' recommendation for its use in this clinical backdrop was "weak."
Induction for suspected macrosomia has been more common over the last few years, and it is interesting that the authors found a paucity of studies with solid design to even address this practice. The largest chunk of data came from a meta-analysis by Sanchez-Ramos et al,2 which incorporated two randomized clinical trials and nine observational studies. It involved a total of 3751 patients. There were no differences in the incidence of shoulder dystocia between those patients undergoing induction and those having expectant management, but in the observational studies there was an increase in cesarean deliveries in the induction group.
It is becoming quite common to induce patients with twin pregnancies around 37 weeks. This is based on the assumption that the average time when spontaneous labor occurs in twins is 37 weeks, thus any patient undelivered by this date should be considered "post term." Yet, the authors found that the concept really is no more than a hunch, as there is nothing in the literature at this point with enough statistical power to show benefit of empiric induction at 37 weeks.
The only pure study addressing induction for isolated oligohydramnios at 41 weeks (compared with waiting until 42 weeks) showed no difference in any outcome studied.3 Unfortunately, the study was grossly underpowered to say much at all. There is little else in the literature dealing with isolated oligohydramnios at other gestational ages. However, although this does not address the question directly, a study showing a tendency to over-diagnose oligohydramnios with an amniotic fluid index (AFI) vs a single vertical pocket suggested that acting on the AFI will increase the cesarean section rate without affecting in a positive way any other outcome variable.
One randomized trial of 200 women found that inducing labor after 37 weeks decreased the incidence of macrosomia (> 4000 g) by a relative risk of 0.54 (confidence interval, 0.32-0.98). Yet, there were no statistically significant differences in maternal or neonatal morbidities (including shoulder dystocia) between those who were induced and those who were managed expectantly.
The latest available statistics show that 24% of patients in the United States between 37 and 41 weeks of gestation are induced. When adding in an unknown number of preterm inductions, it could mean that 1 of 3 women are undergoing inductions of labor. Yet, in only 3 of the most commonly used indications for induction was there "high" or even "moderate" quality of evidence of its benefit.
It is possible that future studies will demonstrate the efficacy of a blanket policy of induction in some of the above categories — even if we develop, in parallel, better ways to assess fetal condition. However, until then, there is at least as much justification for taking the alternative, "first, do no harm" approach of sitting on our hands in some of the above clinical situations when there is nothing to indicate that either the mother or her fetus is in danger.