The Patellar Instability Probability Calculator: A Multivariate-Based Model to Predict the Individual Risk of Recurrent Lateral Patellar Dislocation

Abstract

Background: Predicting the risk of recurrence is of great interest when counseling patients after primary lateral patellar dislocation (LPD). Purpose: To investigate a multivariate model to predict the individual risk of recurrent LPD. Study Design: Case-control study; Level of evidence, 3. Methods: The study population included patients with primary LPD, knee imaging, and a minimum 2-year follow-up after nonoperative treatment. Data including patient characteristics and anatomic patellar instability risk factors were collected retrospectively from 7 national study centers. Bivariate and multivariate regression analyses were carried out to identify risk factors for recurrent LPD and to generate an accuracy-optimized model for out-of-sample prediction. Results: In total, 115 of 201 patients (57%) experienced recurrent LPD within 2 years after primary LPD. Age ≤16 years at primary LPD (odds ratio [OR], 5.0), history of contralateral instability (OR, 2.4), and trochlear dysplasia (Dejour type B-D: OR, 2.5; lateral trochlear inclination ≤12°: OR, 2.7) were significant risk factors for recurrent LPD (P < .05). The prediction accuracy including these 3 risk factors was 79%. Patella alta, an increased tibial tubercle to trochlear groove distance, and patellar tilt had neither an association with increased recurrence rates nor an influence on prediction accuracy of recurrent LPD. Conclusion: Young age and trochlear dysplasia are major risk factors for early recurrent LPD. A multivariate model including age at primary LPD, lateral trochlear inclination, and history of contralateral LPD achieved the highest prediction accuracy. Based on these findings, the patellar instability probability calculator is proposed to estimate the individual risk of early recurrence when counseling patients after primary LPD.