Quantifying glenoid bone loss in the setting of primary reverse shoulder arthroplasty.

Background: Presently, there is no quantifiable and objective criteria for defining severe glenoid bone loss (GBL) for primary reverse total shoulder arthroplasty (rTSA). In this study, we sought to establish quantifiable thresholds of severe GBL using scapular statistical shape models and intraoperative assessment. In addition, we compared outcomes of severe GBL against moderate and mild cases of GBL. Methods: A retrospective review of 57 primary rTSA cases that were determined intraoperatively to be severe GBL were identified. These were uploaded to pre-operative planning software for three-dimensional (3D) reconstruction. Cases were propensity matched at a 1:1 ratio to moderate and mild bone loss cases based on age, sex, and time to final follow-up (mean 53.5 months). Glenoid version, inclination, maximum erosion depth, and glenoid vault loss (GVL) were compared between groups. Receiver operating characteristic (ROC) curves were used to determine severe bone loss thresholds. Blinded raters were asked to determine bone loss severity based on radiographs, 2D computed tomographies, or 3D reconstructions. Two-year and final follow-up American Shoulder Elbow Surgeons and visual analog scale (VAS) pain and function were compared between groups. Results: There was a higher proportion of rotator cuff disease in the mild and moderate bone loss groups when compared to severe bone loss. ROC curves resulted in minimum thresholds for severe GBL of 5.5 mm maximum erosion depth, 12.5% GVL, and 17.5° of version (area under curve = 0.944, 0.924, 0.812, respectively). Discrimination was highest when combining 2 of the 3 measurements (Youden's J = 0.82). Interobserver reliability was highest when using 3D reconstructions (ƙ = 0.726). The mild bone loss group performed worse on American Shoulder Elbow Surgeons, VAS pain, and VAS function scores at 2 years and final follow-up when compared to the severe bone loss group. Conclusion: In this study, we found that severe GBL in primary rTSA is defined as medialization of at least 5.5 mm, GVL of greater than 12.5%, and retroversion greater than 17.5°. Discrimination of severe bone loss from less severe bone loss is best done with at least 2 of the 3 measurements. In addition, these measurements have substantial agreement when blindly assessed and are strongest when using 3D reconstructions. Severe GBL outcomes are similar to less severe bone loss at a mean of 53.5 months, though this is in part driven by less rotator cuff disease in patients with less severe bone loss.