Patterns of glenoid bone loss in revision of anatomic shoulder arthroplasty: the impact of prior glenoid resurfacing

Background: Historically, glenoid resurfacing has been avoided in the young patient due to a perceived risk of early glenoid loosening and destructive bone loss. As a result, many young patients with end-stage arthritis undergo humeral-based hemiarthroplasty with subsequent incomplete pain relief. The aim of this study was to determine the status of glenoid bone stock in patients undergoing revision of a painful hemiarthroplasty versus anatomic total shoulder arthroplasty (aTSA). Methods: Using an institutional shoulder arthroplasty revision database, all patients undergoing revision for aseptic glenoid-based pain after hemiarthroplasty or glenoid loosening of aTSA were identified. Patient demographics, implant longevity prior to revision, and implanted construct at revision were collected for each patient. Based upon preoperative plain films, computed tomography when available, and intraoperative documentation, the status of the glenoid bone stock was described. Results: A total of 39 patients (14 aTSA and 25 hemiarthroplasty) met the inclusion criteria. No statistically significant difference in age at the time of index arthroplasty (56.4 vs. 52.4 years; P = .18), age at time of revision (64.5 vs. 59.0 years; P = .07), proportion of males (64.2% vs. 56.0%; P = .74), or longevity of the revised implant (110.9 vs. 79.2 months; P = .08) were identified between the aTSA and hemiarthroplasty cohorts, respectively. Patients undergoing revision of aTSA were more likely to receive a reverse total shoulder arthroplasty at revision (85.7% vs. 48%; P = .04). Peripheral wear alone was noted in 40% (10/25) of hemiarthroplasty revisions and central defect alone was noted in 79% (11/14) of aTSA revisions. Combined defects (central and peripheral bone loss) were noted in three patients undergoing revision of an aTSA (21.4%) and five undergoing revision of a hemiarthroplasty (20%; P = 1.0). The number of cases with medialization at or beyond the coracoid (Kocsis Grade 2; 42.9% vs. 40%) and at or beyond the spinoglenoid notch (Kocsis Grade 3; 14% vs. 12%) was not different between groups (P = 1.0). Structural bone graft or augmented glenoid components were used at the time of revision in 29% (4/14) and 32% (8/25) of failed aTSA and hemiarthroplasty, respectively (P = 1.0). Conclusion: Despite variability in the cohorts, this study demonstrates no significant improvement in residual glenoid bone stock by deferring glenoid resurfacing at the time of shoulder arthroplasty. At the time of glenoid-based failure, glenoid resurfacing results in central bone loss while hemiarthroplasty results in peripheral/eccentric bone loss.