The Minnesota Heart Survey is a population-based study designed to monitor and explain trends in cardiovascular mortality, morbidity, and risk factors in the Minneapolis-St. Paul (Twin Cities) metropolitan area. Trends in stroke mortality were examined from 1960 to 1991; stroke morbidity trends were examined in 50% samples of patients discharged with an acute stroke diagnosis in 1970, 1980, and 1985. Stroke mortality in Twin Cities residents aged 30 to 74 years declined by more than 70% from 1960 to 1991 in both men and women. The decline occurred at a rate of 2.5% per year until 1972, accelerated between 1972 and 1984 (7.9% per year), and slowed considerably thereafter (1.8% per year). Hospitalized acute-stroke discharge rates among those aged 30 to 74 years declined substantially between 1970 and 1985 in both sexes (P < 0.01), but there was no change in definite stroke rates defined by standardized clinical criteria. Both short-term (28 days) and long-term (5 years) survival of definite-stroke patients improved significantly between 1970 and 1985. These improvements, however, were not found in the entire samples of unverified acute-stroke discharges. The proportion of hospitalized acute-stroke patients who had computed tomography performed increased from 0% in 1970 to 75% in 1985. There were also improvements in hospital records pertaining to the documentation of stroke symptoms and signs. These data indicate that the impressive declines in stroke mortality observed in the 1970s and early 1980s have slowed dramatically in the latter half of the 1980s. The decline in stroke mortality likely reflects both a decline in attack rate and improved survival after stroke.(ABSTRACT TRUNCATED AT 250 WORDS)