Severity of Cognitive Impairment and Mortality After Hip Fracture: Implications for Risk Stratification and Care...

Background/Objectives: Proximal femur fracture (PFF) is associated with high mortality in older adults. Cognitive impairment is common in this population and linked to adverse outcomes, but differences in mortality between mild cognitive impairment (MCI) and dementia and the performance of prediction models in these subgroups remain unclear. Methods: A retrospective cohort study included 1579 patients aged ≥ 60 years undergoing surgery for PFF between 2017 and 2019. Patients were stratified into no, MCI and dementia. The primary outcome was all-cause mortality at 2 years, with additional analyses at predefined timepoints. Associations were assessed using Kaplan-Meier analysis, Cox regression, and timepoint-specific logistic regression. The performance of three prediction models (NHFS, AHFS, ACS-NSQIP) was evaluated using discrimination (AUC) and calibration metrics. Results: Cognitive impairment was present in 469 patients (29.3%), including 307 (19.2%) with mild and 162 (10.1%) with dementia. Mortality increased substantially with cognitive impairment, reaching 24.0% in patients without impairment, 43.7% with mild impairment, and 62.4% with dementia at 2 years. Both mild and severe impairment were independently associated with increased mortality (adjusted hazard ratios 1.48 and 1.95, respectively). Dementia was associated with early mortality, while MCI showed a delayed effect becoming significant from 6 to 12 months onward. Survival differed significantly between patients with and without impairment, but not between MCI and dementia groups. All three risk scores showed moderate and comparable discrimination (AUC ~0.71-0.75). ACS-NSQIP demonstrated the best calibration, while NHFS and AHFS tended to overestimate risk. Conclusions: Cognitive impairment is a strong independent predictor of mortality after PFF. Cognitive impairment was associated with increased mortality after proximal femur fracture. Associations differed across follow-up timepoints, although survival differences between patients with mild cognitive impairment and dementia were not statistically significant in unadjusted Kaplan-Meier analyses. These findings support the clinical relevance of cognitive impairment as a marker of vulnerability in this population.