Sorting recent data
To obtain more precise estimates of morbidity and mortality risks, Eileen Hutton and I performed a meta-analysis of the literature published between 1989 and 1999.9 We included studies in which women undergoing a TOL and those choosing ERC had both been candidates for vaginal birth. That is, we tried to evaluate studies in which those choosing either of the 2 treatments were as comparable as possible at the outset. Unfortunately, in the absence of randomized treatment assignments, it is impossible to exclude all intrinsic differences.
Overall, a TOL results in a more favorable maternal risk profile than ERC.
Fifteen studies involving a total of 47,682 women met our inclusion criteria. We evaluated 8 outcomes of interest: the rates of vaginal birth, uterine rupture, perinatal death, maternal death, maternal febrile morbidity, maternal blood transfusion, hysterectomy, and 5-minute Apgar scores less than 7.
Of the 28,813 women undergoing a TOL, 72.3% achieved vaginal birth. Although the risk of uterine rupture among those choosing TOL was about twice that of those choosing ERC, the absolute risk of this complication was quite small for both groups (0.4% and 0.2%, respectively). The risk of perinatal death was significantly increased in the TOL group, compared with the ERC group; however, the unadjusted risk of death was quite low (0.6% for TOL compared with 0.3% for ERC). When deaths attributable to extreme prematurity or lethal anomalies and intrauterine deaths before labor were excluded, the risk of death decreased to 0.2% in the TOL group and 0.1% in the ERC group. Thus, we calculated the risk of perinatal death attributable to trial of labor to be about 1 in 1,000.
There were about twice as many babies with 5-minute Apgar scores below 7 in the TOL group as in the ERC group. However, we were unable to identify studies comparing long-term infant or childhood morbidity by intended delivery method. The assumption that ERC will result in significantly fewer cases of long-term neurologic impairment is unproven at this time.
For all measures of maternal morbidity, a TOL produced more favorable outcomes than ERC. We found that women electing TOL were about half as likely to require a blood transfusion as those choosing ERC. Women choosing TOL also were much less likely to undergo hysterectomy than those selecting ERC (odds ratio [OR] 0.39; 95% confidence interval [CI], 0.27 to 0.57). Maternal mortality did not differ between the groups.
Overall, our systematic review of a large body of literature supported the contention that a TOL results in a more favorable maternal risk profile than ERC. Although we found that uterine ruptures and fetal risks may be slightly increased with a TOL, our estimates of these complications were reassuringly small. In addition, it should be noted that rates of placenta previa and placenta accreta—potentially life-threatening complications—tend to increase with the number of sections, further supporting a TOL.
Additional research is needed to investigate any possible pelvic-floor trauma associated with VBAC, as well as potential long-term morbidities—such as bowel obstructions, chronic pelvic pain, and dyspareunia—from cesarean section.
Induction of labor
Although TOL became increasingly popular in the 1980s and ’90s, many investigators have been reluctant to recommend induction of labor in the VBAC setting, fearing an increased risk of uterine rupture when oxytocic agents are used. There is less evidence of the safety and efficacy of induction of labor than of spontaneous labor in VBAC attempts. Most reports of induction of labor in this setting come from small series, which may lack the statistical power to identify differences in relatively rare outcomes such as uterine rupture.
A number of authors have published series of labor inductions involving oxytocin. In 1987, Molloy et al reported on 418 women with prior cesarean section whose labors were induced with oxytocin.10 Of these women, 374 (89.5%) delivered vaginally, and 3 (0.7%) experienced uterine ruptures. In contrast, Zelop and colleagues reported a relatively high risk of uterine rupture in a similar population.11 In their series of 458 women attempting VBAC who had labor induced with oxytocin, 9 (2%) experienced uterine rupture. In 2 Canadian series, the risks reported for induction of labor with oxytocin were lower. Ravasia et al12 reported a uterine rupture rate of 2 in 248 (0.8%), and Bebbington and Waterman13 reported a rate of 2 in 460 (0.4%).
There also have been several large series reporting on the use of prostaglandin E2 (PGE2) for induction of labor in VBAC attempts. In 1991, MacKenzie published a series describing 482 women attempting VBAC who had labor induced with PGE2.14 Seventy-five percent delivered vaginally, and 1 (0.2%) experienced uterine rupture. In a 1994 German series, 161 women received PGE2 for induction of labor.15 No uterine ruptures occurred in this cohort.