Background and Purpose— Identification of risk factors for dementia after stroke is best performed in comparison with stroke-free controls, because older subjects at high risk for stroke also have a substantial risk of dementia in the absence of stroke. Previous case-control studies were hospital-based. We used a nested case-control design to prospectively evaluate these risk factors in the community-based Framingham Study cohort. Methods— We compared 212 subjects who were free of dementia in January 1982 and sustained a first stroke after this date, with 1060 age- and sex-matched, stroke- and dementia-free controls. We calculated 10-year risks of dementia (by Diagnostic and Statistical Manual of Mental Disorders , Volume IV criteria) developing in cases and controls and also estimated the hazard ratios within subgroups defined by exposure to various demographic factors (age, gender, education), stroke-related features (right or left hemisphere, stroke type, second stroke), stroke risk factors (hypertension, diabetes, atrial fibrillation, smoking) and apolipoprotein E genotype. Results— Dementia developed in 19.3% of cases and 11.0% of controls. Baseline stroke doubled the risk of dementia (hazard ratio [HR]: 2.0; 95% confidence interval [CI]: 1.5 to 3.1) and adjustment for age, sex, education, and exposure to individual stroke risk factors did not diminish the risk (HR: 2.4; 95% CI: 1.6 to 3.7). The HR was higher in younger subjects (age younger than 80 years [HR: 2.6; 95% CI: 1.5 to 4.5]), apolipoprotein E 3/3 homozygotes (HR: 3.4; 95% CI: 2.0 to 5.8), and high school graduates (HR: 2.4; 95% CI: 1.5 to 3.9). Conclusion— Stroke increases a subject’s risk of dementia as compared with age- and sex-matched controls. Primary and secondary prevention of stroke should significantly decrease the risk of all dementia.