How to measure humeral medialization and upward migration after shoulder hemiarthroplasty

The 2 main complications after hemiarthroplasty (HA) are glenoid erosion, resulting in humeral medialization, and rotator cuff insufficiency, resulting in upward migration of the humerus. Currently, qualitative assessments of glenoid erosion, like the radiographic Sperling classification, are primarily used. Hence, we set out to present and validate a novel radiographic method for quantitatively assessing humeral medialization and upward migration following HA implantation.

The immediate postoperative and last follow-up anteroposterior (AP) radiographs from 64 consecutive patients who underwent a pyrocarbon HA for symptomatic osteoarthritis were used to validate the new method. The new method is based on an orthonormal frame drawn upon AP radiographs in neutral rotation. The X-axis corresponds to the supraspinatus fossa line, and the Y-axis is the orthogonal line tangent to the most lateral border of the acromion. Humeral medialization and upward migration were measured and normalized with the humerus shaft diameter to avoid scale errors. Four independent observers applied the new measurements and determined the Sperling classification for glenoid erosion on all radiographs. Measurement reliability was evaluated using intraclass correlation (ICC) and Fleiss' Kappa coefficients.

Interobserver reliability was good to excellent: ICC for humeral medialization = 0.96 (confidence interval: 0.93-0.97; P < .00) and ICC for upward migration = 0.88 (confidence interval: 0.83-0.93; P < .001). A moderate positive correlation (r = 0.54, P < .0001) was observed between humeral medialization and the progression of glenoid erosion as classified by Sperling. After a mean follow-up of 31 months (range: 24-50 months), humeral medialization averaged 18% (±17) of the shaft diameters, while humeral upward migration averaged 13% (±14) of the shaft diameters. Normalized per month, medialization was 0.57% (±0.55) and upward migration was 0.44% (±0.51) of the shaft diameter. Finally, 39% of patients showed almost no glenoid erosion, and 44% showed almost no upward migration.

The new radiologic measurement method, based upon sequential postoperative AP radiographs, can be reliably used to measure humeral medialization and upward migration of HA overtime. Despite a low elastic modulus, pyrocarbon HA still leads to significant glenoid erosion in about 60% of the patients. Here proposed measurement method can help precisely quantify the speed of medialization, thereby enabling the identification of risk factors in patients with pronounced medialization in future studies.