Comparison of early and late aseptic baseplate failure in primary reverse shoulder arthroplasty with and without structural glenoid autograft.

Abstract

Background: Early reverse total shoulder arthroplasty (RTSA) designs demonstrated high glenoid baseplate complication and revision rates. Although contemporary designs have reduced the incidence of glenoid baseplate failures, there are reports of elevated failure risks in RTSA with glenoid bone grafting within the first 2 years. This study aims to evaluate the incidence and etiology of aseptic glenoid baseplate failure with a contemporary central screw baseplate. The null hypothesis is that majority of the baseplate failure occurs within the first 2 years and that use of glenoid bone grafting does not lead to a higher risk of baseplate failure.

Methods: In 2014-2019, a total of 753 consecutive patients who underwent primary RSA using the same inlay press-fit humeral stem and monoblock central screw baseplate were retrospectively reviewed. Fracture and septic arthropathy cases were excluded. All patients underwent preoperative radiographic and computed tomographic evaluation. If there was significant glenoid erosion (Walch A2, B2, B3, C1, C2, E2, E3, and/or E4 variants), patient-specific structural glenoid bone grafting was performed. All patients underwent standardized radiographic follow-up, and failure was strictly defined as any hardware breakage and/or shift in glenoid baseplate position. Failures were defined as "early" if occurring within 2 years and "late" if occurring >2 years after surgery. Comparative analysis was performed to evaluate demographics, glenoid graft use, and graft union rates between the cohorts.

Results: There were 23 patients with baseplate failures (23 of 753, 3.0%) at a mean of 23 months. Twenty-two failures (96%) occurred in patients who received structural glenoid bone grafting. Only 1 failure (0.2%) occurred when bone grafting was not indicated (P < .001). The most common failure pattern was associated with the B2 glenoid (16 of 23, 70%). There were 5 early failures (22%) and 18 late failures (78%). There were no differences in any patient demographic characteristics between cohorts. All 5 early failures had graft nonunion, and 4 of them occurred without trauma. In the 18 late failures, 9 (50%) occurred without trauma (P = .135). Seventeen of these patients had glenoid grafting, among which 9 (53%) had graft nonunion.

Conclusions: Contemporary RTSA glenoid baseplate designs have an acceptably low incidence of failure. However, the addition of structural bone graft to correct glenoid wear leads to higher aseptic baseplate failure rate. The majority of these patients suffer failure after the 2-year postoperative mark, highlighting the necessity of longer follow-up. Further analysis is necessary to quantify glenoid characteristics (severity of glenoid erosion, critical size of graft) associated with failure.