Late-life dementia predicts mortality beyond established midlife risk factors

Objectives: To compare the mortality rates of elderly demented and nondemented subjects and the differential association of midlife risk factors with mortality according to dementia status. DESIGN:: Prospective historical study. SETTINGS:: Community based. Participants: From the 10,059 male Jewish civil servants who participated in the Israel Ischemic Heart Disease study in the 1960s, the 1,713 who were evaluated for dementia in 1999/2000 and who were unequivocally classified as demented or nondemented. Measurements: Midlife sociodemographic and cardiovascular risk factors, late-life dementia, and mortality. Results: Over a period of 6 years, 718 (42%) subjects died. Of the 307 demented subjects, 71.8% died and of the 1,407 nondemented subjects, 35.4% died. Multivariate survival analyses showed that compared with subjects without dementia, demented subjects had a hazard ratio [HR] for mortality of 2.27 (95% confidence interval [CI] 1.92-2.68). Other risk factors associated with mortality were socioeconomic status (HR 0.94 [0.88-1.00]), higher systolic (HR 1.16 per 20 mm Hg [1.06-1.28 mm Hg]) and diastolic blood pressure (HR 1.15 per 10 mm Hg [1.06-1.25 mm Hg]), and ever smoking (HR 1.38 [1.18-1.61]). Midlife total cholesterol was not associated with mortality (1.01 per 40 mg/dL [0.93-1.10 mg/dL]). None of the interactions of the risk factors with dementia was significant. Conclusions: Dementia was associated with more than double the risk of mortality, but this increased risk did not reflect exacerbation by midlife sociodemographic and cardiovascular risk factors. Our findings suggest that the dementing process itself or its consequences may go beyond well-established midlife risk factors for mortality.