Three-dimensional analysis of biplanar glenoid deformities: What are they and can they be virtually reconstructed with anatomic total shoulder arthroplasty implants?

BACKGROUND: Descriptions of glenoid deformities in glenohumeral osteoarthritis (GHOA) have focused on the axial plane. Less is known regarding arthritic glenoids with higher amounts of superior inclination and little evidence exists to guide management of inclination or combined version-inclination deformity when performing anatomic total shoulder arthroplasty (aTSA). We hypothesized that biplanar deformities (BD) would be present in a higher proportion of GHOA patients than previously appreciated, and these deformities would be difficult to adequately reconstruct with contemporary aTSA implants.

METHODS: A retrospective query was performed of GHOA patients indicated for TSA 2012-2017 with a computed tomography (CT) scan within three months of surgery. Images were uploaded to three-dimensional software for automated measurements. Glenoids with superior inclination ≥10˚, retroversion ≥20˚ were considered to have BD. Walch classification was determined, C-type glenoids were excluded. Rotator-cuff muscle cross-sectional area (CSA) was measured and fatty infiltration graded. Glenoids with BD were virtually planned for aTSA with correction to neutral inclination and version, then 5˚ superior inclination and 10˚ retroversion.

RESULTS: Two-hundred sixty-eight shoulders in 250 patients were included; average age 65 years, 67% male. There were no differences in inclination between Walch types (p=0.25). Twenty-nine shoulders with BD were identified (11%). These deformities were not associated with age (p=0.47) or gender (p=0.50) but were skewed towards Walch B-type, specifically B2 (p=0.03). Acromial index and posterior humeral head subluxation were higher in BD patients (p=0.04, p<0.001, respectively). Biplanar deformities had similar cuff CSA compared to those without but were less frequently associated with fatty infiltration of the subscapularis (p=0.05). When correcting to neutral version and inclination, 41% BD could not be reconstructed. Of those that could, 94% required augmented implants. When correcting to 5˚ superior inclination and 10˚ retroversion, 10% could not be reconstructed. Of those that could, 58% required augmented implants. With partial correction, augment use was predicted by retroversion >26˚ (p=0.009). Inclination did not predict augment use (p=0.90). Final implant position commonly involved unseating in the posterosuperior quadrant and cancellous exposure in the anteroinferior quadrant.

CONCLUSIONS: This retrospective CT-based study of 268 shoulders with GHOA found a 11% prevalence of BD. These deformities were commonly associated with Walch B2 wear patterns. Virtual aTSA planning showed a high failure rate (41%) when correcting to neutral version and inclination. Posteriorly augmented implants were frequently required, and often still involved unseating in the posterosuperior quadrant, increased cancellous exposure in the anteroinferior quadrant, and vault perforation.