The objectives of the study were to clarify the relationship between blood pressure and mortality from stroke, heart disease, cardiovascular diseases and all causes of death among representative population of Japanese and to estimate category-specific excess mortality from stroke due to blood pressure (BP) level. The study design comprised a retrospective cohort study using the 1980 National Survey on Cardiovascular Diseases and identification of underlying causes of death using national vital statistics data. In 1994, a 14-year follow-up cohort study was conducted among participants of the National Survey on Cardiovascular Diseases in 1980, randomly selected from the Japanese population. With a collaboration of 300 public health centres, which had conducted the original survey in 1980, 91.4% of the participants of the original survey could be followed up. Total observed person-years were 53948 for men and 70932 for women. During follow-up, 1327 deaths were observed. BP levels were significantly related to mortality from strokes, cardiovascular diseases and all causes of death for both sexes (P<0.001). Heart disease mortality was significantly related to BP levels among men (P<0.05) while not among women. Estimated excess mortality was 130% for men and 42% for women and chiefly observed among moderate hypertensives (48% for men and 16% for women). In conclusion, high blood pressure was a risk factor for mortality from all causes as well as those from cardiovascular diseases, stroke and heart disease among Japanese. Since the major part of excess mortality was due to mild hypertension, a population strategy to reduce blood pressure should be encouraged.