Development and Validation of a Nomogram Model for Predicting Postoperative Nonunion in Femoral Shaft Fractures

Objective：To identify independent risk factors for nonunion following femoral shaft fracture surgery and develop a clinically applicable nomogram model for personalized risk prediction. Methods：A retrospective cohort study included 804 patients with femoral shaft fractures treated at Xijing Hospital (2014–2020). Patients were divided into development (n=561) and validation (n=243) cohorts. Variables were screened via LASSO regression, and a nomogram was constructed using multivariate logistic regression. Model performance was assessed using ROC curves, calibration plots, Hosmer-Lemeshow tests, and decision curve analysis (DCA). Results：Five independent predictors of nonunion were identified: smoking (OR=3.094, 95% CI:1.790–5.350), high-energy injury (OR=2.454, 95% CI:1.167–5.159), multiple injuries (OR=2.897, 95% CI:1.580–5.312), internal fixation method (OR=3.437, 95% CI:1.519–7.778), and fixation failure (OR=3.437, 95% CI:1.519–7.778). The nomogram demonstrated excellent discrimination (AUC=0.828 in development, 0.835 in validation cohorts) and calibration (Hosmer-Lemeshow P=0.463 and P=0.858, respectively). DCA confirmed clinical utility at threshold probabilities >15%. Conclusion：This nomogram provides a practical tool for predicting nonunion risk in femoral shaft fractures, enabling early intervention for high-risk patients. Clinical trial number：Not applicable.