Mid-trimester cervical length and prediction of vaginal birth after cesarean delivery in Chinese parturients: A retrospective study

Guangpu Liu,Chaofan Zhou,Shengpu Wang,Huixin Zhang

DOI: https://doi.org/10.1016/j.jogoh.2023.102647

Abstract:Background: A successful trial of labor after cesarean (TOLAC) is linked with the best maternal/neonatal outcomes and is more cost-effective than elective repeat cesarean section (ERCS). Predictive models of vaginal birth after cesarean (VBAC) have been established worldwide to improve the success rate of TOLAC. Objective: To validate a VBAC prediction model (the updated Grobman's predictive model without ethnicity) and identify whether mid-trimester cervical lengths (MCL) improve the prediction of VBAC among Chinese women undergoing a TOLAC. Methods: In this retrospective cohort study, the inclusion criteria were a previous history of cesarean delivery (CD) as well as a singleton gestation in the vertex position with routine CL measurements between 20 and 24 weeks and the experience of a TOLAC. MCL as well as identifiable characteristics in early prenatal care that have been used in updated Grobman's predictive model (maternal age, height, pre-pregnancy weight, vaginal delivery history, VBAC history, arrest disorder in previous CD, and treated chronic hypertension) were obtained from the medical records. Associations of maternal characteristics and MCL with VBAC were evaluated using multivariate logistic regression. Two multivariable regression models with and without MCL as one of the risk factors were established and their predictive accuracy for VBAC was critically compared based on receiver-operating characteristic (ROC) curves. Results: This study involved 409 women, among which, 347 (84.8%) achieved a VBAC. The mean MCL was significantly shorter in women who had a successful VBAC than in those who required an intrapartum CD (4.16±0.49 cm vs. 4.35±0.46 cm, P=0.007). Multivariable logistic regression revealed that a longer MCL (cm) was significantly related to a lower success rate of TOLAC [adjusted odds ratio (aOR), 0.48; 95% confidence interval (CI), 0.26-0.88]. The areas under the ROCs of Grobman's model with and without MCL as one of the risk factors were 0.785 (95% CI, 0.725-0.844) and 0.774 (95% CI, 0.710-0.837), respectively, but not significantly different (Z = -0.968, P = 0.333). Conclusions: We first evaluated the efficiency of the updated Grobman's model (without race and ethnicity) in the Chinese population. The area under the curve is relatively high, indicating that the model can be used efficiently in China. The shorter MCL was associated with a greater chance of VBAC and MCL was the independent factor from the factors of Grobman's model. However, the predictive capacity of the modified model by adding MCL as one of the risk factors did not improve significantly.