Development of a Multivariable Model Based on Individual Risk Factors for Recurrent Lateral Patellar Dislocation

Background Nonoperative treatment after first-time patellar dislocation is the standard of care. There is evidence that certain patients may be at high risk for recurrent instability. The aim of this study was to develop a multivariable model to guide management of patients based on their individual risk of recurrent dislocation. Methods A multivariable model was developed using 291 patients from 4 institutions to identify which patients were at higher risk for recurrent patellar dislocation within 2 years. This model was informed by a univariable logistic regression model developed to test factors based on the patient's history, physical examination, and imaging. The discriminatory ability of the model to classify who will or will not have a recurrent dislocation was measured using the area under the receiver operating characteristic curve (AUC). Results Age, a history of a contralateral patellar dislocation, skeletal immaturity, lateral patellar tilt, tibial tubercle-trochlear groove (TT-TG) distance, Insall-Salvati ratio, and trochlear dysplasia were the most important factors for recurrent patellar dislocation. Sex, mechanism of injury, Caton-Deschamps ratio, sulcus angle, inclination angle, and facet ratio were not factors for recurrent dislocation. The overall AUC for the multivariable model was 71% (95% confidence interval [CI]: 64.7% to 76.6%). Conclusions Optimizing the management of lateral patellar dislocation will improve short-term disability from the dislocation and reduce the long-term risk of patellofemoral arthritis from repeated chondral injury. This multivariable model can identify patients who are at high risk for recurrent dislocation and would be good candidates for early operative treatment. Further validation of this model in a prospective cohort of patients will inform whether it can be used to determine the optimal treatment plan for patients presenting with an initial patellar dislocation. Until validation of the model is done with new patients, it should not be used in clinical practice. Level of evidence Prognostic Level III. See Instructions for Authors for a complete description of levels of evidence.