Prediction of vaginal birth after cesarean delivery

Objective To examine factors associated with successful vaginal birth after cesarean (VBAC) and to validate a previously established prediction model. Methods In a retrospective study, data were obtained for women with one prior low-transverse cesarean procedure who underwent a trial of labor with a cephalic singleton pregnancy at term at one UK hospital between January 2000 and August 2013. Univariate analysis and logistic regression analysis were used to identify maternal demographic characteristics significantly associated with successful VBAC and factors independently associated with this outcome, respectively. A prediction model was built, and predicted probabilities were compared with observed frequencies. For validation, probabilities were also calculated by a previous prediction model. Results Overall, 1463 women formed the cohort. Successful vaginal delivery was achieved in 1050 (71.8%) women. The only factors significantly associated with unsuccessful VBAC were Asian (odds ratio [OR] 1.59, 95% confidence interval [CI] 1.14-2.23) or African (OR 1.80, 95% CI 1.23-2.64) ethnic origin, and previous cesarean for failure to progress (OR 6.39, 95% CI 4.81-8.49). The predicted and observed probability of successful VBAC were well correlated (Spearman ρ, 0.905; P = 0.002). The established prediction model was less accurate. Conclusion Previous cesarean performed for failure to progress and Asian/African ethnic origin were associated with unsuccessful VBAC. The performance of a previous prediction model was inferior.