Decoding the Risk of Postoperative Elbow Stiffness: A Predictive Model Integrating Fracture Complexity and Surgical Trauma

Background Post-traumatic elbow stiffness (PTES) is a prevalent and functionally debilitating complication following surgical treatment of elbow fractures, with an incidence ranging from 10–30%. It significantly compromises upper limb mobility and quality of life, and may result in permanent disability. Despite the identification of multiple risk factors, the etiology of PTES remains multifactorial, and no widely accepted tool exists for early risk stratification. This study aimed to identify independent predictors of PTES and to develop and validate a clinical nomogram for individualized risk prediction. Methods We conducted a retrospective cohort study involving patients who underwent surgical treatment for elbow fractures between January 2021 and December 2024. Functional outcomes were evaluated using range of motion (ROM) and the Mayo Elbow Performance Score (MEPS), and patients were categorized into a non-stiffness group and a stiffness group. The dataset was randomly divided into training and validation cohorts. Univariate and multivariate logistic regression analyses were performed to identify independent risk factors, which were then used to construct a predictive nomogram. Model performance was evaluated through discrimination (C-index), calibration (Hosmer–Lemeshow test and calibration plots), and clinical utility (decision curve analysis). Results A total of 376 patients were included, with 264 in the training set and 112 in the validation set. Multivariate logistic regression identified six independent predictors of elbow stiffness: fracture type (OR 11.00), joint dislocation (OR 3.95), mechanism of injury (OR 6.91), surgical approach (OR 4.57), surgical duration (OR 1.01 per minute), and postoperative immobilization time (OR 1.04 per day). The model demonstrated excellent discrimination with a C-index of 0.931 in the training set and 0.821 in the validation set. Calibration curves indicated strong agreement between predicted and observed outcomes. Decision curve analysis showed a favorable net clinical benefit across a wide range of threshold probabilities (10–90%). Conclusion We present a clinically validated prediction model that decodes the multifactorial risk landscape of postoperative elbow stiffness. By integrating fracture complexity and surgical trauma variables, the nomogram offers a practical tool for early risk stratification, enabling personalized rehabilitation planning and improved functional outcomes.