Abstract 4136716: Frailty and Malnutrition Predict Mortality in Older Adults with Atherosclerotic Cardiovascular Disease

Background: Frailty and malnutrition are related geriatric syndromes but it remains unknown if both provide independent prognostic information in older adults with atherosclerotic cardiovascular disease (ASCVD). Aims: We compared the prognostic value of combined frailty and malnutrition assessment with either parameter alone in predicting all-cause mortality. Methods: We analyzed 719 older adults (>70 years old) with prevalent ASCVD from visit-6 of the Atherosclerosis Risk in Communities Study. ASCVD was defined as prevalent coronary heart disease, stroke, or history of MI at visit-6. Malnutrition was assessed using the geriatric nutritional risk index (score <98), and physical frailty was determined using the Fried frailty phenotype. Participants were categorized as high-risk if both frail and malnourished, intermediate-risk if frail or malnourished, and low-risk if neither. Cox regression and model performance metrics were used to assess frailty, malnutrition, and their combination, both alone and in conjunction with a base model. Results: The mean age of the cohort was 80.5±4.5 years and 39.4% were women. A total of 67.2% had frailty, 13.4% had malnutrition and 11.1% had both. Median follow-up was 580 days with 61 mortality events. Compared to low-risk adults, intermediate-risk (2.3 [1.04-5.0], p=0.039) and high-risk (OR 3.1 [95% CI 1.1-8.5], p=0.027) adults had significantly higher adjusted risk of mortality (Figure). In contrast, frailty or malnutrition alone were not statistically significant predictors after multivariable adjustment. Addition of the combined risk to the base model led to a modest increase in the c-index compared to the malnutrition model (0.668 vs. 0.647) but not the frailty model (0.668 vs 0.667). There was no significant improvement in reclassification or integrated discrimination index, but the combined model had superior goodness-of-fit than the frailty and malnutrition models, as demonstrated by a difference >|2.0| in the Akaike information criterion (-3.4&-13.5 respectively). Conclusion: Frailty and malnutrition are highly overlapping syndromes. Incorporation of readily available nutritional risk measures to traditional frailty assessment may improve risk prediction in older adults with ASCVD.