Does surgical indication influence implant survival in reverse shoulder arthroplasty? A comparative cohort study of fracture and glenohumeral osteoarthritis.

Abstract

Reverse shoulder arthroplasty (RSA) is widely used for the treatment of advanced glenohumeral osteoarthritis and complex proximal humeral fractures. However, comparative data evaluating implant survival and outcomes between fracture and elective indications remain limited. This study aimed to assess mid- to long-term implant survival, complications, radiographic findings, and functional outcomes following RSA, with particular attention to the influence of surgical indication. A retrospective cohort study included 126 patients who underwent primary RSA using a single implant system. Patients were treated for either advanced glenohumeral osteoarthritis or complex proximal humeral fractures. Implant survival was analyzed using Kaplan-Meier estimates, and potential predictors were evaluated with Cox regression analysis. Secondary outcomes included complications, radiographic changes, Oxford Shoulder Score (OSS), and range of motion (ROM). The mean follow-up was 57.4 ± 51.1 months. The overall one-year implant survival rate was 96.8%, declining to 93.4% at eight years and remaining stable thereafter. Survival differed numerically by diagnosis, reaching 98.3% for osteoarthritis and 86.4% for fractures at eight years. No statistically significant predictors of failure were identified; however, uncemented stems showed a numerically higher risk of events (HR 3.01, p = 0.088), and fracture indication demonstrated a numerically higher risk compared with osteoarthritis (HR 2.10, p = 0.220). The overall complication rate was 7.14%. The mean OSS was 46.34 ± 4.72. Postoperative ROM values were within expected ranges, with no significant differences between groups. Reverse shoulder arthroplasty demonstrated satisfactory mid- to long-term implant survival and favorable postoperative outcomes in both fracture and elective indications. Although fracture cases showed numerically lower survivorship, no statistically significant independent effect of surgical indication on implant failure was demonstrated in this cohort. The findings should be interpreted with caution given the retrospective design, limited number of revision events, and absence of preoperative functional data.