Offset reconstruction in stemless and stemmed shoulder hemiarthroplasty: influence on long-term outcomes

Shoulder hemiarthroplasty (HA) is a well-established treatment for various humeral-side pathologies. Overstuffing should be avoided because it is associated with glenoid wear; however, specific radiological reconstruction targets remain unclear. This study assessed the association between radiological joint restoration in stemmed and stemless HA and long-term functional outcomes.

Patients who underwent stemmed or stemless HA between 2001 and 2011 were included in a long-term follow-up. Cases with incomplete or non-calibratable digital radiographs or concomitant glenoid procedures were excluded. Clinical outcomes were evaluated using the Constant-Murley Score (CMS) and satisfaction questionnaires. Radiographic parameters-lateral glenohumeral offset (LGHO), lateral humeral offset (LHO), and center of rotation (COR)-were measured on standardized anteroposterior radiographs pre- and postoperatively by two independent observers. Differences from baseline to final follow-up (ΔLHO, ΔLGHO, ΔCOR) were calculated. Interobserver reliability was determined using the intraclass correlation coefficient (ICC). Associations between radiological measures and clinical outcomes were analyzed using linear regression; intergroup comparisons used unpaired t-tests (p < 0.05).

Forty-eight patients (12 stemmed, 36 stemless HA) were examined after a mean follow-up of 16.6 ± 2.5 years. CMS did not differ between groups (stemmed 50.5 ± 20.8 vs. stemless 54.1 ± 20.4, p = 0.603). Complete radiographic data were available for 25 patients. Interobserver agreement was excellent for LHO (ICC = 0.93) and very good for LGHO (ICC = 0.85). Mean reconstruction values were close to zero for LHO (ΔLHO = - 0.3 ± 5.1 mm), slightly underreconstructed for LGHO (ΔLGHO = - 6.3 ± 12.9 mm), and slightly medialized for COR (ΔCOR = - 3.5 ± 8.5 mm). Linear regression revealed a significant negative association between ΔLHO and CMS (R² = 0.45; coefficient = - 4.9; p = 0.02).

Both stemmed and stemless HA achieve satisfactory long-term outcomes after over 16 years. Accurate reconstruction of the lateral humeral offset correlates with improved functional results.