Survival in a population-based cohort of dementia patients: predictors and causes of mortality

OBJECTIVES: To examine predictors of survival time and causes of mortality in a population-based cohort of demented subjects. DESIGN: Longitudinal naturalistic follow-up study. SETTING: A rural area in The Netherlands. PARTICIPANTS: The study population consisted of 102 demented subjects derived from a population-based, two-stage prevalence study of dementia among subjects aged 65 and over. MEASUREMENTS AND ANALYSIS: Survival analysis was performed using the Realized Probability of Dying as a measure of survival time. Adjustment for duration and severity at study entry was employed in all analyses. RESULTS: A substantial excess mortality was present among demented subjects as compared with their non-demented birth cohort. Disorders related to dementia, such as cachexia, dehydration and pneumonia, were major causes of death. Variables related to severity of functional impairment, aphasia and a shorter duration at study entry predicted a shorter survival in the study population.