To measure the incidence and outcome of stroke in Estonia a community-based study of first-ever stroke was carried out in Tartu from 1991 through 1993. A total of 829 first-ever stroke cases were registered during the 3-year period. Most stroke patients were prospectively identified from the Department of Neurology and Neurosurgery of the University Hospital, other sources where death certificates, autopsy protocols, outpatient and hospital records. The hemorrhagic stroke subtype constituted 15% of all cases. Although pathological verification was available in 32% of cases (autopsy or CT/MRI), the incidence and 30 days'' case fatality rates for the hemorrhagic subtype were similar to those of other studies, while the rates for the ischemic subtype were higher than those reported by many authors. The total incidence rate (2.50 per 1,000), the age-specific rates for the patients < 65 and the overall case fatality rate at 30 days were higher than those found by most study centers. Possibly, a study of stroke risk factors would explain the high burden of stroke in Estonia.