Successful vaginal trial of labor: Can it be predicted?

November 1, 2010

Some women who had an earlier cesarean delivery are not candidates for a trial of labor.

Key Points

The cesarean delivery rate in the United States continues to rise, and as a result more women than ever are entering pregnancy with a prior cesarean scar. Some women who had an earlier cesarean delivery are not candidates for a trial of labor (TOL) because of a previous classical uterine incision or obstetric conditions such as placenta previa or breech presentation; however, others deliver at hospitals where TOL is not allowed and are not given the choice of undergoing a TOL, although they may be candidates for it.

After an in-depth counseling session with the clinician detailing the risks, benefits, and alternatives to TOL, a woman's decision to pursue a TOL often hinges on her perception of the likelihood of achieving a vaginal birth after cesarean (VBAC) and the risks particular to her clinical situation. Ideally, the clinician should be able to present information on the likelihood of success specific to that patient.

Factors affecting TOL outcome

Certain factors influence the likelihood of successful TOL to a greater extent than others. For example, a history of any prior vaginal delivery, particularly VBAC, is much more highly associated with successful TOL than maternal age.1 However, many studies that have identified factors associated with TOL success did not address the independence of and potential interactions among these factors. In addition, although knowledge of factors associated with success or failure of TOL can provide some guidance, they are insufficient to predict TOL outcomes.

Estimating probability of successful TOL

Several strategies have been used to counsel women about the probability of achieving a VBAC. One approach has been to inform the patient that the overall likelihood of successful TOL is approximately 60% to 82%, based on studies of most populations.2 This approach, however, fails to individualize the risk profile, potentially placing women of vastly different risk statuses into a single group with a uniform range.

A second, more individualized approach recognizes several patient characteristics (eg, BMI, maternal age, prior indication for cesarean delivery, prior vaginal delivery) to qualitatively place a woman at the higher or lower end of that generalized 60% to 82% range. Although more individualized than a single unvarying estimate, this approach does not accurately or quantitatively combine multiple patient factors into a specific, accurate assessment of the likelihood of successful TOL.

Various predictive models have been proposed in the literature.3 These models form the basis of another counseling strategy: using individual patient factors to arrive at a specific risk estimate. In each of these prediction models, patient factors associated with successful TOL are assigned a certain number of points, the total of which indicates a woman's chance of successful TOL. However, several of these models have limited accuracy and clinical usefulness. For example, a model was based on a univariate analysis that identified 4 variables associated with successful TOL. In the scoring system, women were awarded 1 point for each variable: previous dysfunctional labor, no previous vaginal delivery, nonreassuring fetal tracing on admission, and induction of labor in the current pregnancy. Notably, 2 of these variables cannot be known until presentation to labor and delivery, thereby precluding use of the model in the antepartum period. Also, a variable such as a nonreassuring fetal heart tracing can directly cause the outcome of interest, limiting its predictive value. Moreover, in a predictive model such as this, the range of points is limited (ie, 0 points if none of the 4 variables is present, giving a 91.5% chance of VBAC, to 4 points if all variables are present, yielding a 46.1% chance of VBAC).4 This, in effect, means that patients with potentially very different risk estimates are lumped into groups with equal likelihood of TOL success. Other models have had similar limitations.5-7