Alcohol and mortality: a cohort study of male Japanese physicians.

The relationship between drinking habit surveyed in 1965 and cause-specific mortality over 19 years was investigated in 5135 male Japanese physicians taking into account smoking habit and separating ex-drinker from non-drinker. As compared with non-drinkers, daily drinkers with high consumption had a significantly increased mortality from all causes. Drinking was significantly related to the so-called alcohol-related causes of death; upper aerodigestive cancer, liver cancer and liver cirrhosis. Mortality from acute myocardial infarction was inversely related to drinking, whereas other coronary heart disease showed a somewhat higher mortality among men consuming a large amount of alcohol than among non-drinkers. There was a weak, but significant, association between stroke and drinking, and the relation did not differ between haemorrhagic stroke and other stroke. No obvious relationship with drinking was observed for cancers of the stomach, large bowel, pancreas and lung.