Anteroposterior tear size, age, hospital, and case number are important predictors of repair integrity: an analysis of 1962 consecutive arthroscopic single-row rotator cuff repairs

• Published in: Journal of shoulder and elbow surgery 2021-08, Vol.30 (8), p.1907-1914
• Main Authors: Duong, Julian K.H., Lam, Patrick H., Murrell, George A.C.
• Format: Article
• Language: English
• Subjects: determinants; interaction effects; predictors; repair; retear; Rotator cuff
• ISSN: 1058-2746; 1532-6500
• DOI: 10.1016/j.jse.2020.09.038

Retear or nonhealing of a surgically repaired rotator cuff is common, but the determinants of retear are poorly understood. We aimed to determine the relationship between preoperative and intraoperative factors and retear rate following rotator cuff repair and to formulate a predictive model based on this relationship, including any interaction effects between tear size, patient age, and surgical experience in contributing to the retear rate. We performed a post hoc analysis of prospectively collected data from 1962 consecutive patients who underwent a primary arthroscopic single-row rotator cuff repair performed by the senior author from 2007 through 2018 and postoperatively returned for 6-month follow-up ultrasonography. Multiple logistic regression analysis revealed anteroposterior tear length to be the most important independent predictor of retear (Wald statistic, 90; P < .001). Other independent predictors included case number (Wald statistic, 59; P < .001), patient age at surgery (Wald statistic, 30; P < .001), and type of hospital (public vs. private) (Wald statistic, 17; P < .001). The data indicated that following rotator cuff repair, there was a 4-fold increase in the retear rate as the anteroposterior tear size increased from 1 cm to 3 cm; a 8-fold decrease when comparing case number 1000 with case number 3000; a 2-fold increase as patient age increased from 50 years to 70 years; and a 3-fold increase when comparing surgery performed in a public hospital vs. a private hospital. The area under the receiver operating characteristic curve was 0.81 (95% confidence interval, 0.78-0.84), indicating that our equation could predict 81% of retears. Anteroposterior tear length was the most important independent predictor of retear or nonhealing. Other predictive factors included case number, patient age at surgery, and hospital type. The predictive effect of anteroposterior tear length, patient age, and case number in contributing to retear was additive. Our regression equation may be used to calculate patient rotator cuff retear risk at 6 months after repair.