BackgroundDespite a new trend to systematically use reverse shoulder arthroplasty (RSA) in elderly population regardless of the indication, total anatomical shoulder arthroplasty can get good functional results in this population. The purpose of this study was to evaluate clinical and radiological outcomes of uncemented short-stem anatomic total shoulder arthroplasty (TSA) for primary glenohumeral osteoarthritis in patients older than 70 years and to compare these results to a matched population with an uncemented short-stem RSA.MethodsIn this retrospective monocentric study, clinical outcomes were based on constant score (Cst), subjective shoulder value (SSV) score, and range of motion. The aim of radiographic analysis was to identify glenoid component loosening and humeral bone remodeling around the uncemented short stem.ResultsAt an average follow-up of 44 ± 12.5 months, 32 uncemented short-stem TSA in 31 patients with a minimum of 2 years of follow-up were included and were compared to 32 uncemented RSA. Fifty three percent of the patients had “a forgotten prosthesis”. ROM was significantly improved in all cases. Cst reached 73 ± 9 pts and SSV 90 ± 10.8% (P < .001). In 8 patients with repairable supraspinatus tendon tears, clinical outcomes were not statistically different from patients with an intact rotator cuff: Cst (77 ± 6.2 points vs 72 ± 9.6 points, P = .3) and SSV (88 ± 11.5% vs. 91 ± 10.5%; P = .59). The type of glenoid wear (A vs B) did not influence the constant score: 73 ± 9 points versus 74 ± 11 points respectively; P = .81. Despite a complication rate of 6% (n = 2), no prosthesis revision was performed. At last follow-up, range of motion was better in the TSA group compared to the RSA group for internal (7.8 ± 1.3 vs 6.25 ± 2; P = .001) and external (47 ± 14 vs 24 ± 21; P < .001) rotations. The postoperative SSV score was also better in the TSA group (91.3 ± 10% vs 82.2 ± 13%; P = .002).ConclusionsAt medium-term, uncemented short-stem anatomic TSA in patients older than 70 years provided satisfactory clinical results. Patients have forgotten their prosthesis in over 50% of cases. This prosthetic design is still indicated in this patient population in case of primary osteoarthritis with a functional rotator cuff with an almost normal rotator cuff muscle trophicity.