To assess age-, sex-, and subtype-specific incidence rates of dementia and to calculate the risk of dementia, the authors performed a large, community-based, prospective cohort study on dementia as part of the Rotterdam Study. Participants were recruited among residents of a suburb of Rotterdam, aged 55 yearsand older. Baseline examinations took place between 1990 and 1993. The average follow-up was 2.1 years. Screening for dementia followed a three-stage protocol. Medical records of subjects who had died or couldnot be examined in person were evaluated. Of 7,046 subjects who were nondemented at baseline, 162developed dementia during 15,135 person-years of follow-up, resulting in an overall incidence rate of 10.7 per1,000 person-years. From the youngest to the oldest 5-year age category, the incidence rate increased from 0.6 to 97.2 per 1,000 person-years. Only in men did the increase level off after age 85. Overall, the incidencerate per 1,000 person-years was 7.7 for Alzheimer's disease and 1.5 for vascular dementia. Dementia incidence rates and dementia-free Kaplan-Meier survival tables were used to calculate age- and sex-specific cumulative risks of dementia. Although the incidence rates of men and women up to age 85 were similar, the lifetime risk of dementia for 55-year-old women was twice as high as for men (0.33 vs. 0.16), reflecting both the higher life expectancy of women and the higher dementia risk at very old age.