Development of prediction models for successful external cephalic version and delivery outcome.

Abstract

To develop prediction models for the chance of successful external cephalic version (ECV) and delivery outcome. This is a single-center retrospective study including 350 pregnant women with a singleton non-cephalic pregnancy at or after 36weeks of gestational age. We selected 22 factors for ECV prediction and 21 for delivery outcome after successful ECV prediction as candidate predictors. Multivariable logistic regression with a stepwise backward selection procedure was used to construct a prediction model for the chance of successful ECV and the other for the delivery outcome. The discrimination and calibration of the models were assessed and internal validation was done with bootstrapping. ECV was successfully performed in 232 cases (66.3%) among 343 women. Eight predictive factors were identified to be associated with a successful ECV: Gestational week at ECV < 39weeks, multiparous, BMI before pregnancy < 22kg/m3, palpable fetal head, breech engagement, larger AFI, larger BPD and posterior placenta. This model showed good calibration and good discrimination (c-statistic = 0.82, 95% CI 0.76-0.88). Six predictive factors were identified to be associated with vaginal delivery after successful ECV: age < 35, multiparous, BMI before pregnancy < 22kg/m3, anterior placenta, lateral placenta and none-front fetal spine position. This model showed fair discrimination (c-statistic = 0.79, 95% CI 0.72-0.85). However, its calibration was not so satisfactory especially when the predicted probability was low. We validated a prediction model for ECV and delivery outcome, showing that the model's overall performance is good. This can be used in clinical practice after external validation.