Trends in stroke incidence and mortality are available from few sites worldwide. The unique nature of the medical care in patients from Rochester, Minn, has allowed the ascertainment of virtually all cases of stroke for community residents for the period dating from 1955. This study updates incidence rates, stroke mortality, radiological imaging availability, and prevalence for stroke and subtypes of stroke though 1989 and evaluates trends since 1955. Medical records of all residents of Rochester with potential diagnosis of stroke during the 5-year period through December 31, 1989 were screened to determine whether the case met the criteria for stroke. The type of stroke was determined with the use of all available data. Average annual age- and sex-adjusted incidence rates for stroke and various types of stroke were calculated for 5-year periods from 1955 to 1989. Prevalence of stroke was also determined. Poisson regression was used to better clarify the dependence of stroke incidence on multiple variables. Cox proportional hazards modeling evaluated trends in short-term and long-term stroke survival. We found that 496 incidence cases of stroke occurred among Rochester residents during 1985 to 1989. The annual age- and sex-adjusted stroke incidence rates was 145 per 100 000 population, which was virtually unchanged from 1980 to 1984 and 13% higher than the rate determined in 1975 to 1979. Increasing incidence rates compared with the 1970s were noted in all groups aged older than 54 years and in both sexes. Survival after cerebral infarction was dependent on age and calendar year when trends dating from 1955 were evaluated. The annual incidence rate of intraparenchymal hemorrhage was twice that of subarachnoid hemorrhage. Incidence rates for stroke have remained at a high rate than those determined in the 1970s. Although the occurrence of untreated hypertension has been reported to be stable or decreasing, the incidence rate of stroke is higher than those noted during the decline in stroke incidence during the 1960s and 1970s. In addition to the impact of radiological imaging, increased contribution of patients with ischemic heart disease, and the possible detection of milder cases of stroke, other factors, which are undefined, may be contributing to the increased stroke incidence rates detected over the last 10 years.