Influence of glenoid wear pattern on glenoid component placement accuracy in shoulder arthroplasty

Accurate glenoid component placement in shoulder arthroplasty is often difficult even with the use of preoperative planning. Computer navigation and patient-specific guides increase component placement accuracy, but which patients benefit most is unknown. Our purpose was to assess surgeons' accuracy in placing a glenoid component in vivo using 3-dimensional preoperative planning and standard instruments among various glenoid wear patterns.

We conducted a retrospective review of 170 primary anatomic total shoulder arthroplasty (aTSA) and reverse total shoulder arthroplasty (rTSA) performed at a single institution. Commercially available preoperative planning software was used in all arthroplasties with multiplanar 2-dimensional computed tomography and a 3-dimensional implant overlay. After registration of intraoperative bony landmarks to the navigation system, participating surgeons with knowledge of the preoperative plan were blinded to the computer screen and attempted to implement their preoperative plan by simulating placement of a central-axis glenoid guide pin. Two hundred thirty-three screenshots of surgeon's simulated guide pin placement were included. Glenoid displacement, error in version and inclination, and overall malposition from the preoperatively planned target point were stratified by posterior wear status (with [Walch B2 or B3] or without [A1, A2, or B1]) and Walch classification (A1, A2, B1, B2, or B3). The glenoid component was considered malpositioned when version or inclination errors exceeded 10° or the starting point displacement exceeded 4 mm.

For rTSA, errors in version were greater for glenoids with posterior wear compared with those without (8.1° ± 5.6° vs. 4.7° ± 4.0°; P < .001). On post hoc analysis, B2 glenoids had greater version error than A1, A2, and B1 glenoids. A greater proportion of glenoids undergoing rTSA that possessed posterior wear had an error in version >10° compared with those without (31% vs. 8%; P < .001). Consequently, glenoids undergoing rTSA with posterior wear were malpositioned at a greater rate compared with those without (73% vs. 53%). In contrast, glenoids undergoing aTSA with and without posterior wear did not differ based on displacement error, version error, inclination error, or malposition occurrence.

Posterior glenoid bone loss more commonly resulted in glenoid version errors exceeding 10 degrees and component malposition in rTSA, but not for aTSA. Malposition was still relatively high in patients without significant posterior wear for both aTSA (36%) and rTSA (53%). Surgeons should consider alternate techniques beyond preoperative planning and standard instrumentation when performing shoulder arthroplasty in patients with posteriorly worn glenoids.