Abstract
OBJECTIVE: Cervical insufficiency is a significant cause of mid-trimester miscarriage and preterm birth, for which cervical cerclage is the standard treatment. However, some patients experience extreme preterm delivery even after undergoing a repeat cerclage following initial cerclage failure. This study aimed to investigate the optimal timing for repeat cerclage and to identify independent predictors of its failure. METHODS: The sociodemographic characteristics and clinical data of 47 singleton pregnancies who underwent a repeat cerclage at Fujian Maternity and Child Health Hospital from October 2018 to May 2025 were retrospectively analyzed. Participants were divided into a failure group (delivery < 28 weeks, n = 17) and a success group (delivery ≥ 28 weeks, n = 30). Univariate and multivariate logistic regression analyses were used to identify independent risk factors. Receiver operating characteristic (ROC) curves were employed to determine optimal cut-off values for key continuous variables. Moreover, the BP neural network was used to perform a secondary validation of the independent risk factors. RESULTS: Multivariate analysis using Firth’s penalized logistic regression identified that cervical dilation before repeat cerclage (aOR = 3.098, 95% CI: 1.492–6.434, P = 0.002), elevated neutrophil-to-lymphocyte ratio (NLR) (aOR = 2.015, 95% CI: 1.055–3.849, P = 0.034), and positive cervical culture (aOR = 10.554, 95% CI: 2.506–44.442, P = 0.001) were independent risk factors for repeat cerclage failure. Exploratory ROC curve analysis suggested that preoperative cervical dilation ≥ 1.75 cm (AUC = 0.76) and NLR ≥ 4.9 (AUC = 0.83) may have predictive value for failure. A Back-Propagation (BP) neural network validation using Leave-One-Out Cross-Validation (LOOCV) achieved an accuracy of 74.5%, further confirming the predictive robustness of cervical dilation, NLR, and cervical culture. CONCLUSION: Preoperative cervical dilation, systemic inflammation (NLR), and local cervical infection are key predictors of repeat cerclage failure. Our findings suggest that repeat cerclage may be most effective when cervical dilation is less than 2 cm and inflammatory markers are not elevated, though this requires prospective validation.