RESUMO
BACKGROUND: The aim of this study is to assess whether a simple comprehensive geriatric assessment (CGA) score predicts hospital mortality among very elderly patients admitted with heart failure (HF). METHODS: This is a prospective follow-up of 581 individuals aged ≥75 years admitted for decompensated HF to an acute geriatric unit from October 2006 to September 2009. A CGA score (range, 0-10) was constructed using baseline individual data on 5 domains: dependence in activities of daily living (Katz index), mobility (qualitative mobility scale), cognition (Mini-Mental State Examination), comorbidity (Charlson index), and number of prescribed medications. RESULTS: Mean age of patients was 85.8 ± 5.8 years, 67% were women, and 75% had preserved ventricular function (ejection fraction >45%). Fifty percent of patients required assistance in ≥1 activities of daily living, 66% had mobility problems, 45% had cognitive impairment, the mean Charlson index was 3.97 ± 3.01, and 36% had >7 medications prescribed. As a result, the mean CGA score was 4.8 ± 2.2. Hospital mortality was 8.2%. In multivariate analysis, variables associated with hospital mortality included New York Heart Association functional class III (odds ratio [OR] 4.1, 95% CI 1.5-10.8), class IV (OR 19.6, 95% CI 6.3-61), pulmonary edema on chest radiography (OR 3.0, 95% CI 1.3-6.6), renal failure (OR 2.8, 95% 1.2-6.2), and the CGA score (OR 1.2, 95% CI 1.02-1.4 for each point of increment). The area under the receiver operating characteristic curve was 0.856 (95% CI 0.790-0.921), and the model classified 93.4% of cases correctly. CONCLUSIONS: In our cohort of very old patients with HF, a simple CGA score predicts hospital mortality.
