Association of Preoperative Shoulder Osteoarthritis Severity Score With Change in American Shoulder and Elbow Surgeons Score at 2 Years After Rotator Cuff Repair

Abstract

Background: The impact of early glenohumeral osteoarthritis (GHOA) on clinical outcomes after rotator cuff repair (RCR) remains unclear. The magnetic resonance imaging (MRI)–based Shoulder Osteoarthritis Severity (SOAS) score is a comprehensive approach to quantifying glenohumeral degeneration. Purpose: To investigate the association between SOAS scores and changes in American Shoulder and Elbow Surgeons (ASES) scores in patients who underwent RCR. Study Design: Cohort study; Level of evidence, 3. Methods: Two reviewers independently analyzed the preoperative MRI scans of 116 shoulders and assigned SOAS scores. Spearman correlation was used to calculate the association of mean SOAS scores with patient demographic characteristics and change in ASES scores over the 2-year follow-up period (ΔASES). Multivariate regression analysis was performed between the independent variables of patient age, sex, body mass index, and significant SOAS score components as determined by univariate analysis, with the dependent variable being ΔASES. Significance was defined as P < .05 for univariate analysis and P < .0125 after application of the Bonferroni correction for multivariate analysis. Results: The mean ASES scores were 55.8 ± 18.6 preoperatively and 92.1 ± 12.1 at 2 years postoperatively. The mean preoperative SOAS score was 15.2 ± 7.1. On univariate analysis, the total SOAS score was positively correlated with patient age ( rS = 0.41; P < .001), whereas ΔASES was negatively correlated with patient age ( rS = −0.27; P = .0032). Increasing SOAS subscores for supraspinatus/infraspinatus tear size ( rS = −0.28; P = .024), tendon retraction ( rS = −0.23; P = .015), muscle atrophy ( rS = −0.20; P = .034), paralabral ganglia ( rS = −0.23; P = .015), and cartilage degeneration ( rS = −0.21; P = .024) were negatively correlated with ΔASES. A negative correlation was found between increasing total SOAS score and ΔASES ( rS = −0.22; P = .016). On multivariate analysis, increasing supraspinatus/infraspinatus tear size was significantly and negatively correlated with ΔASES (β = −3.3; P = .010). Conclusion: Increasing the total SOAS score was predictive of less improvement in ASES scores at 2 years postoperatively. On univariate analysis, SOAS subscores with the strongest negative correlations with ΔASES scores included tear size, muscle atrophy, tendon retraction, paralabral ganglia, and cartilage wear. On multivariate analysis, only tear size was significantly associated with a lower change in the ASES score.