The Elbow UCL Injury Prognosis Score

Objectives: No model exists to predict which patients with elbow ulnar collateral ligament (UCL) injuries will successfully return to play (RTP) with non-operative treatment. The reported rates for successful RTP after non-operative management of UCL injuries are limited and vary widely. Furthermore, patient and UCL injury characteristics that influence failure of non-operative treatment have not been established. We sought to identify which factors were most significant in predicting which patients would have successful RTP after non-operative management of UCL injury. In doing so, we created the Elbow UCL Injury Prognosis Score to identify and weigh factors that predict failure of non-operative treatment. The parameters include: age, position, competition level, moving valgus stress test (MVST), 5th finger sensation, UCL tenderness, and MRI findings including tear type, location, and chronicity features. This score will predict which patients will succeed non-operative management, avoiding unnecessary surgery while simultaneously identify patients for whom non-operative management would be delaying the inevitable need for surgical intervention. Methods: 205 patients were evaluated by the senior author for a UCL injury. Physical exam findings, imaging characteristics and patient demographics were retrospectively collected and entered into a HIPAA-compliant REDCap database after IRB approval of the study design. Of these patients, 138 underwent a trial of non-operative treatment and were either cleared for RTP (n= 55) or eventually underwent surgical intervention (n= 83) after a minimum one-month trial of non-operative management. Sixty-seven patients decided to have surgery immediately without trialing non-operative management and were excluded from the scoring system. The Elbow UCL Injury Prognosis Score was developed and validated using stepwise feature selection by Akaike information criterion (AIC). Three-fold cross validation was performed using these ten variables as predictors. The mean and standard deviation of area under the receiver operating characteristic (ROC) curve (AUC), sensitivity, specificity, positive predictive value, and negative predictive value were evaluated. Results: The area under curve for non-operative failure predicting non-operative management failure was 0.79 (0.054), with sensitivity=0.85 (0.088), specificity=0.72 (0.032), positive predictive value=0.83 (0.023), negative predictive value=0.77 (0.091) [mean (standard deviation (SD))]. Patients were predicted to succeed non-operative management if they scored less than 10, while for a score greater than or equal to 10 patients would likely fail non-operative management and require surgical intervention for their UCL injury (Table 1). Conclusion: The Elbow UCL Injury Prognosis Score is a statistically rigorous, powerful tool for predicting which patients will most likely RTP with non-operative treatment. We hope this score can provide surgeons with a useful tool to enhance and enable shared decision making with patients when faced with UCL injuries at all competitive levels. [Table: see text]