Establishment and validation of a prediction model for successful trial of labor after cesarean in scarred uteruses.

OBJECTIVE

To develop and validate a prediction model for successful trial of labor after cesarean (TOLAC) in women with a scarred uterus using clinical data from a single center with an external temporal validation set.

METHODS

This retrospective study developed a prediction model using data from a single center (Huai'an Maternal and Child Health Hospital) with scarred uterus planning TOLAC, collected from January 2016 to December 2023. Assessed variables included maternal characteristics, obstetric history, and prenatal examination findings. Univariate and multivariate logistic regression were used to identify predictors and develop the model. The model was temporally validated using an independent external dataset from two other tertiary hospitals. Its performance was evaluated by the area under the ROC curve (AUC), sensitivity, and specificity, and was compared with the Grobman model.

RESULTS

Among 2,386 eligible women, 1,721 (72.1%) had a successful TOLAC. Univariate analysis identified 10 significant variables (P<0.05). Multivariate analysis retained six independent predictors: age, body mass index, history of vaginal delivery, cervical score, estimated fetal weight, and gestational age. The model expression was: Logit(P) = -3.82 + 0.04 × age - 0.12 × BMI + 1.56 × vaginal delivery history + 0.37 × cervical score - 0.02 × fetal weight + 0.18 × gestational age. Internal validation showed an AUC of 0.85 (95% CI: 0.82-0.88), sensitivity of 82.3%, and specificity of 78.5%. Temporal Validation (n=524) yielded an AUC of 0.83 (95% CI: 0.79-0.87), which was significantly higher than the Grobman model (AUC=0.79, P<0.05).

CONCLUSIONS

The developed prediction model demonstrates good performance and generalizability for predicting TOLAC success, potentially aiding clinical decision-making and improving maternal and neonatal outcomes.