Mid-term to long-term results of open posterior bone block grafting in recurrent posterior shoulder instability: a clinical and CT-based analysis.

Background: There is little consensus on the best treatment after failed conservative management of recurrent posterior shoulder instability. The purpose of this study was to analyze our clinical and radiological mid-term to long-term results of an open, posterior bone block procedure for the treatment of recurrent posterior shoulder instability.

Methods: From 1999 to 2015, 14 patients were included in the study and available for clinical and radiographic follow-up (FU). FU included a standardized physical examination, assessment of the Constant-Murley-Score, subjective shoulder value, American Shoulder and Elbow Surgeons score, and Western Ontario Shoulder Instability Index. Conventional radiographs and a computed tomography (CT)-scan were performed preoperatively and at latest FU. Glenohumeral arthropathy was classified as per Samilson and Prieto. The CT scans were used to evaluate the structure of the graft (resorption, union), graft positioning, glenoid version, centering of the humeral head, and glenoid erosion and morphology.

Results: The median age at the time of surgery was 26 years (range 16-41 years) and the median FU period was 9 years (range 4-20 years). The rate of reported dynamic postoperative subluxation and instability was 46% (n = 6) and the rate of dynamic posterior instability during clinical testing at FU was 31% (n = 4). The tested instability rate in the traumatic group was 14% (n = 1) compared to the atraumatic group with 50% (n = 3) during clinical FU. The mean Constant-Murley-Score increased from preoperatively (77 ± 11 points) to postoperatively (83 ± 14 points,  = .158). The last FU showed an American Shoulder and Elbow Surgeons score of 85 ± 12; the Western Ontario Shoulder Instability Index score was 715 ± 475 points. The mean subjective shoulder value increased from 58% ± 19 preoperatively to 73% ± 17 at final FU ( = .005). Degenerative changes increased by at least one grade in 67% of the patients. Mean preoperative glenoid retroversion (CT) was 7.5° ± 6°. The position of the graft was optimal in 86% (n = 12). In 62% of the cases, a major resorption of the graft (Zhu grade II) was observed.

Conclusion: The rate of tested recurrent instability at last FU was as high as 31% (n = 4, atraumatic [n = 3] vs. traumatic [n = 1]) after a median FU of 9 years. Given the moderate improvement of clinical outcome scores, shoulder stability and the increase of degenerative joint changes by at least one grade (Samilson/Prieto) in 67% of patients, a posterior bone block procedure is not a uniformly satisfying treatment option for recurrent posterior shoulder subluxation, especially in cases of atraumatic posterior instability.