Radiographic bone resorption after reverse total shoulder arthroplasty using a single onlay platform: risk factors...

Bone resorption around the humeral stem is a frequent radiographic finding after reverse total shoulder arthroplasty (rTSA). Short stems have been introduced to reduce stress shielding and preserve proximal bone, but comparative data in Asian populations are limited. We evaluated the incidence, distribution, and factors associated with humeral bone resorption among Japanese patients who underwent cementless rTSA using a single onlay humeral platform and compared standard and short stems.

This multicenter retrospective study included shoulders treated with cementless rTSA in Japan. After excluding 6 shoulders with incomplete data, 128 shoulders were eligible. Propensity score matching was performed based on age, sex, body mass index, diagnosis, and pre-operative range of motion. Bone resorption in the lateral metaphysis, lateral diaphysis (LD), medial metaphysis, and medial diaphysis zones was graded using the Inoue classification. Canal filling ratios were calculated at metaphyseal and diaphyseal levels, and stem alignment was categorized as valgus, neutral, or varus. Multivariable logistic regression identified factors associated with grade ≥3 bone resorption.

After propensity score matching, baseline characteristics were balanced between the standard stem (n = 35) and short stem (n = 35) groups. At 2 years post-operatively, grade ≥3 bone resorption was observed more often with standard stems (23 of 35 shoulders, 65.7%) than with short stems (13 of 35 shoulders, 37.1%; unadjusted P = .03). The LD zone most frequently exhibited resorption and differed significantly between groups after correction for multiple secondary comparisons (standard: 12 of 35 shoulders, 34.3%; short: 2 of 35 shoulders, 5.7%; unadjusted P < .01), whereas the lateral metaphysis, medial metaphysis, and medial diaphysis zones did not differ significantly. Compared with the nonresorption group, the resorption group included more female patients (52.8% vs. 20.6%; unadjusted P < .01). Age, body mass index, canal filling ratios, and alignment did not differ between these groups. Female sex (odds ratio, 3.64; 95% confidence interval, 1.07-12.37; P = .04) and standard stem use (odds ratio, 5.54; 95% confidence interval, 1.19-25.87; P = .03) were independently associated with resorption. Range of motion at 2 years did not differ according to resorption status.

In this Japanese cohort treated with a single onlay humeral platform, grade ≥3 bone resorption occurred more frequently with standard stems than with short stems, particularly in the LD zone. Female sex and standard stem use were independently associated with resorption.