Radiologic midterm results of cemented and uncemented glenoid components in primary osteoarthritis of the shoulder: a matched pair analysis.

INTRODUCTION: Cemented all-polyethylene glenoid components are considered the gold standard in anatomic shoulder arthroplasty. New designs of cementless metal backed glenoid components showed promising early and midterm results. The aim of this matched-pair analysis was to compare the radiologic results of two cemented glenoid components and a cementless glenoid component in patients with primary osteoarthritis (OA).

METHODS: Sixty shoulders were clinically and radiologically evaluated after a mean follow-up of 59 months. Mean patient age was 70.4 years at surgery. Based on the design of the glenoid component (keel, peg, MB), 3 groups with, respectively, 20 shoulders were formed according to the matching criteria time of follow-up, patient age and gender. RLL and osteolysis in anteroposterior and axillary X-ray images were quantified and combined in a radiologic score (R-Score). Higher scores expressed worse radiologic outcomes. Further radiological parameters such as lateral glenohumeral offset (LGHO) and subluxation index were measured according to Walch. The functional results were documented using the age and gender normalized Constant-Murley score.

RESULTS: Postoperative R-Score was highest in pegged components (peg: 5.7, keel: 2.4, MB: 1.6; p < 0.001) when combining both radiographs and after separate analysis of anteroposterior radiographs. MB glenoids had the lowest R-score in axillary radiographs (peg: 2.2, keel: 1.4, MB: 0.6; MB vs. keel: p = 0.004, MB vs peg: p < 0.001). RLL were more common (p = 0.004) and severe (p = 0.005) in pegged glenoids (RLL incidence: 77.8%, RLL-score: 2.5) than in MB glenoids (RLL incidence: 30%, RLL-score 0.7) and tended (p = 0.084) to have a higher RLL-score than keeled glenoids (RLL incidence: 63.2%, RLL-score:1.4). Both the osteolysis score (keel vs. peg: p < 0.001, MB vs. peg p < 0.001) and the incidence of osteolysis (keel vs. peg: p = 0.008, MB vs peg: p = 0.003) were significant higher in pegged glenoids (peg: osteolysis score: 3.2, osteolysis incidence: 100%; keel: osteolysis score: 1.0, osteolysis incidence: 63.2%, MB: osteolysis score: 0.9, osteolysis incidence: 60%), while the osteolysis score in axillary images was lowest for MB glenoids (peg: 1.2, keel: 0.9, MB: 0.4; peg vs. MB: p = 0.009, keel vs. MB: p = 0.047). Osteolysis in the central axillary zone was least common in MB glenoids (peg: 50%, keel: 47.4%, MB: 15%; peg vs. MB p = 0.035, keel vs. MB p = 0.041). LGHO was highest in MB glenoids (peg: 54.1, keel: 54.5, MB: 57.8; p < 0.001) but did not increase radiographic loosening (r = 0.007; p = 0.958). Preoperative posterior humeral head subluxation seemed to affect incidence of RLL negatively (pre-op posterior decentered 64.3%, pre-op centered 31.9%; p = 0.201) but did not reach statistical significance.

CONCLUSION: Pegged glenoid components had a concerning rate of RLL and osteolysis. MB glenoid components had a better outcome in axillary radiographs concerning RLL and osteolysis. Increased LGHO did not increase radiographic loosening.

LEVEL OF EVIDENCE: Retrospective comperative treatment study Level III.