Glenoid morphology after reaming in computer-simulated total shoulder arthroplasty

Background: The relationships between reaming parameters for glenoid-implant surface area and bone loss in total shoulder arthroplasty have not been well established. The hypotheses of this study are: (1) for large version corrections, a large reaming depth of 5 mm is not sufficient to obtain complete glenoid implant contact; (2) glenoid bone is removed in a linear proportion with reaming depth; and (3) initial reamer placement has no effect on glenoid bone removal. Methods: Ten computer models from computed tomography scans of patients with advanced osteoarthritis were created for computer-simulated reaming as performed during total shoulder arthroplasty. Reaming variables studied included reaming depth, reamer placement, and version correction. The resulting reamed glenoid surface area available for implantation and bone volume removed were calculated for each permutation. Results: Reamed surface area significantly increased with larger depths of reaming (P < .0001) and smaller version corrections (P < .0001). Bone volume removed and reaming depth had a strong quadratic relationship (r² = 0.999). With off-center reamer placement, volume removed when deviating in the posterior direction was significantly greater than when deviating in the anterior, superior, or inferior direction (P < .05). Conclusion: Performing smaller version corrections allows for greater attainable implant-bone surface contact because increasing reaming depth results in small increases in conforming surface area but large losses in glenoid bone stock. Bone volume removed was most sensitive to off-center position errors in the posterior direction.