ABSTRACT
OBJECTIVE: To identify any clinical and radiologic findings of rotator cuff injury that predict whether patients will undergo shoulder surgery. METHODS: We retrospectively studied all shoulder sonograms obtained at a single institution over 12 months. Possible predictors of surgical treatment were documented, including patient age and sex, duration and types of symptoms, and the location and severity of tendon damage on sonography. One hundred twenty-eight patients underwent shoulder sonography; 34 patients eventually underwent shoulder surgery. Multivariate logistic regression was performed to identify clinical and sonographic factors associated with the use of surgical therapy. RESULTS: The only statistically significant predictor of surgical intervention was the finding of full-thickness tears (with or without tendon retraction) on sonography (P = .03). Patients with full-thickness tears were 4.3 times more likely to undergo surgery than those with no tears (odds ratio, 4.3). CONCLUSIONS: The sonographic diagnosis of full-thickness rotator cuff tears is the only finding statistically associated with the use of surgical treatment. No single clinical variable was consistently associated with subsequent surgery. Partial-thickness tears on sonography also do not show any statistical association with the eventual use of surgery for rotator cuff symptoms.
