Fixation-specific modeling of age-frailty interactions to predict early mortality after hip fracture surgery.

Background

Early mortality after hip fracture surgery remains high. Existing risk models treat age and frailty as independent predictors and may not reflect how physiologic reserve modifies risk across fixation strategies. We evaluated whether modeling age-frailty interactions improves estimation of 30-day mortality.

Methods

A retrospective cohort study using the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) database (2019-2023) included adults undergoing intramedullary nailing (IMN), hemiarthroplasty, or plate fixation for femoral neck, pertrochanteric, or subtrochanteric fractures. Frailty was measured using a validated 5-item modified frailty index. Generalized additive models with bivariate spline surfaces generated procedure-specific mortality risk estimates. Secondary outcomes were assessed using propensity score-matched cohorts.

Results

Among 115,643 patients, 30-day mortality was 5.25%. Predicted mortality varied across age-frailty combinations, ranging from 1.41% to 28.83% for IMN, 1.26%-32.06% for hemiarthroplasty, and 0.48%-20.37% for plate fixation. IMN showed higher transfusion rates, while hemiarthroplasty demonstrated higher cardiopulmonary complication rates. Length of stay did not differ between procedures.

Conclusions

Modeling age and frailty as interacting variables produces procedure-specific mortality risk estimates after hip fracture surgery. These estimates support patient counseling and perioperative decision-making using routinely available clinical data.