Predicting good clinical outcomes after revision rotator cuff repair: the revision cuff repair prediction score.

Abstract

Background: The aim of this study was to develop a scoring system based on preoperative characteristics to predict clinical outcomes 2 years after revision rotator cuff repair.

Methods: This retrospective study included 195 patients (mean age, 55.1 years) who underwent revision rotator cuff repair between July 2001 and December 2020 in 12 treatment centers in France. All patients were evaluated preoperatively and at least 2 years after surgery in terms of Constant score, subjective shoulder value (SSV) and patient satisfaction. Variables associated with these clinical outcomes in a multivariable logistic regression model were included in the scoring system and weighted according to their association with 1, 2 or 3 of the considered clinical outcomes (Constant score, SSV, patient satisfaction). The ability of the scoring system to predict Constant scores > 70 points and SSVs > 70% 2 years after surgery was optimized using receiver operating characteristic (ROC) curves.

Results: In multivariate analysis, an isolated supraspinatus tear was associated with higher Constant scores (p = 0.019), SSVs (p < 0.001) and patient satisfaction (p = 0.004); work-related tendon tears were associated with lower Constant scores (p = 0.019) and SSVs (p = 0.005); female sex was associated with lower Constant scores (p = 0.002); manual work was associated with lower SSVs (p = 0.046); and smoking and a delay between primary and revision repair > 6 months were associated with lower patient satisfaction (p = 0.049 and p = 0.020). These variables were weighted in the scoring system from 0 to 9 points as follows: 3 points for an isolated supraspinatus tear, 2 points for a non-work-related tendon tear, and 1 point each for non-manual work, non-smoking status, male sex, and a delay to revision < 6 months. The AUCs of the scoring system to predict a Constant score > 70 points and an SSV > 70% two years after revision repair in these patients were 72% and 73%, respectively. The positive predictive value of a score ≥ 6/9 points was 80% for a Constant score > 70 points, 91% for an SSV > 70% and 91% for patient satisfaction, two years after revision repair.

Conclusion: A simple scoring system for patients with recurrent rotator cuff tear based on six preoperative characteristics was optimized to predict good clinical outcomes 2 years after revision rotator cuff repair. This revision rotator cuff repair prediction (RCRP) score could be used in patient counseling and to identify those most likely to benefit from revision repair.