Reverse total shoulder arthroplasty versus open reduction internal fixation for proximal humerus fractures in elderly patients: an age-stratified propensity score-matched analysis.

BACKGROUND

The optimal surgical management of displaced proximal humerus fractures in elderly patients remains controversial. Reverse total shoulder arthroplasty (rTSA) has gained increasing acceptance for complex fracture patterns, yet comparative data against open reduction and internal fixation (ORIF) across age strata are limited. This study compared reoperation rates, emergency department (ED) utilization, and dislocation following rTSA versus ORIF in propensity score-matched cohorts stratified by age.

METHODS

Using the TriNetX multi-institutional research network, patients aged 65-74 years and 75-84 years who underwent rTSA or ORIF for proximal humerus fractures were identified. Propensity score matching was performed using nearest-neighbor matching with a 0.150 standard deviation caliper. The primary outcome was coded shoulder-related return to surgery at 1 and 2 years. Secondary outcomes included all-cause ED utilization and coded dislocation when available.

RESULTS

After matching, 748 patients (374 per group) were analyzed in the 65 to 74-year cohort and 882 patients (441 per group) in the 75 to 84-year cohort. Patients undergoing rTSA had significantly more 4-part fractures (34.2% vs. 6.4% in the 65 to 74-year cohort, P < .001). At 2-year follow-up, reoperation rates did not differ significantly in either age group (65-74 years: rTSA 5.62% vs. ORIF 3.48%, risk ratio 1.62, P = .16; 75-84: rTSA 4.54% vs. ORIF 5.67%, risk ratio 0.80, P = .444). Kaplan-Meier survival analysis confirmed no significant differences in freedom from surgical failure. ED utilization did not differ significantly between groups, and dislocation was comparable in the 75 to 84-year cohort for which dislocation data were available.

CONCLUSION

In propensity score-matched cohorts, rTSA and ORIF demonstrated comparable coded reoperation rates and all-cause ED utilization at 1- and 2-year follow-up. These findings persisted despite rTSA patients having more complex fracture patterns; however, because fracture morphology and patient-reported outcomes were not directly matched or captured, the results should be interpreted as hypothesis-generating real-world associations rather than evidence of treatment superiority.