BACKGROUND The study objective was to determine the association between reported alcohol consumption and total mortality, mortality from selected causes, and incident nonfatal chronic disease events in middle-aged (51 to 64 years old) and elderly (65 to 75 years old) men during an approximate 15-year follow-up period. RESULTS We conducted a prospective epidemiological study of Japanese-American men who were participating in the Honolulu Heart Program and were free from coronary heart disease, cerebrovascular disease, and cancer at baseline examination and at subsequent reexamination 6 years later. Self-reported alcohol consumption was determined twice: at the baseline examination in 1965 through 1968 and at reexamination approximately 6 years later (1971 through 1974). Four primary alcohol consumption groups who reported similar alcohol intake at the time of these two clinical examinations were considered: abstainers and light (1 to 14 mL of alcohol per day), moderate (15 to 39 mL of alcohol per day), and heavy (> or = 40 mL of alcohol per day) drinkers. Study end points were also determined in very light (1 to 4.9 mL of alcohol per day) drinkers and in men who reported a change in their alcohol intake between examinations. Longitudinal follow-up was carried out through the end of 1988 with determination of selected fatal and nonfatal events according to alcohol intake. After controlling for several potentially confounding factors, total mortality exhibited a J-shaped pattern in relation to alcohol consumption in middle-aged and elderly men. There was a trend for lower rates of occurrence of combined fatal and nonfatal coronary heart disease events with increasing alcohol consumption in both middle-aged and elderly men. Increasing alcohol consumption was related to an increased risk of fatal and nonfatal strokes in middle-aged men, whereas elderly light and moderate drinkers were at increased risk for fatal and nonfatal strokes. Heavy drinkers were at increased risk for fatal and nonfatal malignant neoplasms in the two age groups examined. CONCLUSIONS The results of this long-term prospective study provide a balanced perspective of the health effects of alcohol consumption in middle-aged and elderly men. High levels of alcohol consumption were shown to be related to an increasing risk of diseases of considerable public health importance. These findings suggest that caution be taken in formulating population-wide recommendations for increases in the population levels of alcohol consumed given the associated significant social and biological problems of high consumption levels.