Japan's age-adjusted rate for mortality from stroke increased after the Second World War until 1965 and then showed a significant decline until 1990; however, the age-adjusted rate for mortality from all heart disease and coronary heart disease (CHD) increased until 1970 and then declined slowly. A puzzling question is why the rate of mortality from CHD declined in spite of an increase in serum total cholesterol level following an increase in fat consumption. It was confirmed that CHD incidence was far lower in several Japanese populations compared to Western countries in the " Monitoring Trends and Determinants in Cardiovascular Disease " (MONICA) project; therefore, the lower CHD mortality in Japan stems from the lower CHD incidence. CHD risk factors based on epidemiologic cohort studies in Japan were no different from those of other industrialized countries: hypertension, hypercholesterolemia, smoking and diabetes mellitus (DM). So, how can we explain this phenomenon?There are three possible explanations. One is the decline in population blood pressure level and the prevalence of hypertension during the years 1965-1990; the second is the decline in smoking rate in men and women; the third is that the serum total cholesterol level for middle-aged and elderly populations remains 5-15 mg/dL lower than that of the US elderly counterpart, although men aged 40-49 in Japan and the US had similar serum total cholesterol levels. It was also noted that elderly people in Japan, as observed in the Seven Countries Study, had far lower serum total cholesterol levels in midlife, i.e., around 160 mg/dL in the 1960s. This was not the case for elderly in the US where a higher serum total cholesterol level was observed in midlife. In conclusion, the lower serum cholesterol level in the past of Japanese middle-aged and elderly people compared to Western counterparts helps to maintain the low CHD incidence and mortality supported by the declining trend in blood pressure level and smoking rate for both men and women.