Dementia-associated mortality at thirteen years in the NEDICES Cohort Study.

To evaluate the mortality, thirteen years after the baseline wave (1994), of participants suffering dementia in the Neurological Disorders in Central Spain (NEDICES) Cohort Study, we conducted a population-based cohort study in the elderly (65 years and more) with 5,278 screened participants at baseline. Mortality has been evaluated by means of the National Death Registry of Spain at 1-5-2007, 13 years after enrolment. Cox's proportional hazards regression models were used to evaluate the hazard of death according to dementia severity and type, adjusting for potential covariates (gender, age, level of education, and co-morbidity). Survival was estimated using Kaplan-Meier method. Of the 5,278 participants screened at baseline, 306 had dementia. Mortality at 13 years was: 275 deaths (89.9%) in dementia subjects; and 2,426 (49.0%) in subjects without dementia. Mortality was higher and statistically significant in dementia subjects. The degree of dementia (DSM-III-R) correlated with the risk of mortality, from mild (HR = 2.23; CI: 1.77-2.82) to moderate (HR =3.10; CI: 2.47-3.89) and severe dementia (HR = 4.98; CI: 3.85-6.44). Survival was similar in Alzheimer's disease and vascular dementia. Factors associated with higher mortality in Cox proportional hazard models were older age, male gender, and comorbidity. Using Population Attributable risk (PAR%), dementia was related to 11.3% of all deaths. Dementia intensity increases the mortality risk at ten years in the NEDICES Study as in other cohort studies. Age, gender, and co-morbidity are associated with higher mortality in dementia patients. Almost one third of deaths in persons over 85 years-old could be attributable to dementia.