Disease surveillance and population surveys of risk characteristics in a northeast rural community of Japan (1965 census population, 7,030) are combined in an attempt to relate morbidity and risk factor trends for coronary heart disease and stroke during the last 2 decades. Between 1964 and 1983, the incidence of coronary heart disease (i.e., combined myocardial infarction, angina pectoris, and sudden death) did not change significantly among men and women ages 40-69, and was lower than that for stroke. The incidence of all stroke declined about 60% for both men and women, ages 40-69, with a significant decrease in cerebral hemorrhage for both sexes and in cerebral infarction for men. Between 1963Between -1966Between and 1980Between -1983 upward shifts occurred in the means and distributions of serum total cholesterol and serum total protein in every age and sex group, primarily during the 1st decade. Age-adjusted mean cholesterol levels rose 22 mg/dl to the 1980-1983 mean of 179 mg/dl in men ages 40-69. In women ages 40-69, the mean rose 29 mg/dl to 192 mg/dl. Among nutrients, animal fat intake doubled in men ages 40-59 from 4.5% of daily calories in 1969 to 9.6% in 1980-1983. Animal protein intake also increased, from 5.8% to 7.1%. Most of this increase occurred between 1969 and 1972-1975 and may be attributable to an increased intake of meat, eggs and dairy products. From 1963From -1966From to 1980From -1983 relative weight index rose significantly for all age-sex groups except men ages 50-69. Mean systolic and diastolic blood pressure levels declined for every age-sex group, with a 15-mm Hg age-adjusted decrease in systolic, 4-mm Hg decrease in diastolic pressure among men ages 40-69, and a 11-mm Hg systolic and 4-mm Hg diastolic decrease for women. Two cohorts of men and women ages 40-69 at baseline were followed for disease incidence: an early cohort (2,257 persons) followed from 1963-1966 to 1973 and a later cohort (2,711 persons) followed from 1972-1975 to 1983. In these cohorts, significant risk prediction for cerebral hemorrhage and infarction was obtained with blood pressure level and end organ effects in the electrocardiogram and fundus photographs. Serum cholesterol was inversely associated with cerebral hemorrhage in the early cohort but not in the later. The reduced association of serum cholesterol in the later cohort may be due in part to the marked upward shift in means and distributions of serum cholesterol between these two periods and the lower power of the later analysis. (Circulation 1989;79:503-515) The frequency of coronary heart disease (CHD) and stroke in populations is strongly related to the prevalence of hypertension and atherosclerosis, which are greatly influenced, in turn, by risk characteristics, related behaviors, and associated environmental factors.' Japan has recently experienced rapid change in living and eating patterns.2 This provides a unique opportunity to observe a natural experiment in which population risk may change.In an attempt to relate morbidity t...