BACKGROUND

Version correction via eccentric reaming reduces clinically important retroversion in Walch type B2 glenoids (those with substantial glenoid retroversion and a second, sclerotic neoglenoid cavity) before total shoulder arthroplasty (TSA). Clinically, an increased risk of glenoid component loosening in B2 glenoids was hypothesized to be the result of compromised glenoid bone quality attributable to eccentric reaming. However, no established guidelines exist regarding how much version correction can be applied without compromising the quality of glenoid bone.

QUESTIONS/PURPOSES: (1) How does version correction correlate to the reaming depth and the volume of resected bone during eccentric reaming of B2 glenoids? (2) How does version correction affect the density of the remaining glenoid bone? (3) How does version correction affect the spatial distribution of high-quality bone in the remaining glenoid?

METHODS

CT scans of 25 patients identified with Walch type B2 glenoids (age, 68 ± 9 years; 14 males, 11 females) were selected from a cohort of 111 patients (age, 69 ± 10 years; 50 males, 61 females) with primary shoulder osteoarthritis who underwent TSA. Virtual TSA with version corrections of 0°, 5°, 10°, and 15° was performed on 25 CT-reconstructed three-dimensional models of B2 scapulae. After simulated eccentric reaming at each version correction angle, bone density (Hounsfield units [HUs]) was analyzed in five adjacent 1-mm layers under the reamed glenoid surface. Remaining high-quality bone (> 650 HUs) distribution in each 1-mm layer at different version corrections was observed on spatial distribution maps.

RESULTS

Larger version corrections required more bone resection, especially from the anterior glenoid. Mean bone densities in the first 1-mm bone bed under the reamed surface were lower with 10° (523.3 ± 79.9 HUs) and 15° (479.5 ± 81.0 HUs) version corrections relative to 0° (0°, 609.0 ± 103.9 HUs; mean difference between 0° and 15°, 129.5 HUs [95% CI, 46.3-212.8 HUs], p < 0.001; mean difference between 0° and 10°, 85.7 HUs [95% CI, 8.6-162.9 HUs], p = 0.021) version correction. Similar results were observed for the second 1-mm bone bed. Spatial distribution maps qualitatively showed a decreased frequency of high-quality bone in the anterior glenoid as version correction increased.

CONCLUSIONS

A version correction as low as 10° was shown to reduce the density of the glenoid bone bed for TSA glenoid fixation in our computational study that simulated reaming on CT-reconstructed B2 glenoid models. Increased version correction resulted in gradual depletion of high-quality bone from the anterior region of B2 glenoids.

CLINICAL RELEVANCE

This computational study of eccentric reaming of the glenoid before TSA quantitatively showed glenoid bone quality is sensitive to version correction via simulated eccentric reaming. The bone density results of our study may benefit surgeons to better plan TSA on B2 glenoids needing durable bone support, and help to clarify goals for development of precision surgical tools.