Prognostic model on niche development after a first caesarean section: development and internal validation

Abstract

Objective: To develop and internally validate a prognostic prediction model for development of a niche in the uterine scar after a first caesarean section (CS). Study design: Secondary analyses on data of a randomized controlled trial, performed in 32 hospitals in the Netherlands among women undergoing a first caesarean section. We used multivariable backward logistic regression. Missing data were handled using multiple imputation. Model performance was assessed by calibration and discrimination. Internal validation using bootstrapping techniques took place. The outcome was ‘development of a niche in the uterus’, defined as an indentation of ≥ 2 mm in the myometrium. Results: We developed two models to predict niche development: in the total population and after elective CS. Patient related risk factors were: gestational age, twin pregnancy and smoking, and surgery related risk factors were double-layer closure and less surgical experience. Multiparity and Vicryl suture material were protective factors. The prediction model in women undergoing elective CS revealed similar results. After internal validation, Nagelkerke R2 ranged from 0.01 to 0.05 and was considered low; median area under the curve (AUC) ranged from 0.56 to 0.62, indicating failed to poor discriminative ability. Conclusions: The model cannot be used to accurately predict the development of a niche after a first CS. However, several factors seem to influence scar healing which indicates possibilities for future prevention such as surgical experience and suture material. The search for additional risk factors that play a role in development of a niche should be continued to improve the discriminative ability.