Stemless and stemmed total shoulder arthroplasty: a comparison of short-term clinical and radiographic outcomes

Stemless total shoulder arthroplasty could provide benefits over stemmed arthroplasty which has represented the gold standard for decades. Proposed benefits of stemless arthroplasty include better reproduction of anatomy and reduction in stress shielding; however, this does not appear to be confirmed by any study. The hypothesis was there would be no clinical differences between the stemless and the short-stem prosthesis, but the stemless prosthesis would better reproduce coronal radiographic anatomy and have less radiographic evidence of stress shielding. A prospectively collected data of patients undergoing primary, anatomic total shoulder arthroplasty for osteoarthritis were retrospectively reviewed. Patient-determined outcomes including the Western Ontario Osteoarthritis Index, American Shoulder and Elbow Surgeons score, Single Assessment Numeric Evaluation, Simple Shoulder Test, and Shoulder Activity Level were recorded preoperatively, at 1 year, and at 2 years. Preoperative and 1-year postoperative range of motion was recorded. Radiographic parameters to assess restoration of proximal humeral anatomy included humeral head height, humeral neck angle, humeral centering on the glenoid, and postoperative restoration of the anatomic center of rotation. Final postoperative radiographs were assessed for evidence of stress shielding. Forty-eight patients had a stemmed humeral prosthesis, and 109 patients had a stemless prosthesis. Patient-determined outcomes were available from 2 years postoperatively in 99.4%. Both groups had significant improvements in all patient-reported outcomes and range-of-motion metrics, but there were no differences between the stemless and stemmed groups in these outcomes. The prosthetic humeral head of the stemmed components was more likely to extend further superior to the humeral osseous margin than that of the stemless group (2.0 ± 2.4 vs. 0.8 ± 1.4 mm; P = .0004). The stemless group had a smaller postoperative deviation from the anatomic center of rotation than the stemmed group (2.5 ± 1.9 vs. 3.2 ± 2.1 mm; P = .04). The humeral neck angle was comparable between the stemmed group and the stemless group (133 ± 7° vs. 131 ± 8°; P = .06). There was similar mean deviation of humeral head centering on the glenoid prosthesis between the stemmed and stemless groups (1.9 ± 1.8 vs. 1.6 ± 1.7 mm; P = .20). There was evidence of stress shielding in 10 patients (21%) with a stemmed prosthesis and in no patients with a stemless prosthesis at 1-year follow-up (P < .0001). There were no differences in patient-determined outcomes between the groups at 2-year follow-up. Restoration of proximal humeral anatomy was either better or equivalent with the stemless prosthesis compared to the stemmed one. Radiographic evidence of stress shielding was found in the stemmed prosthesis but not in the stemless prosthesis at 1-year follow-up.