Abstract
The relationship between alcohol and site-specific cancers was investigated in a follow-up study of 5,139 male Japanese physicians. Information on drinking habits was obtained by mail questionnaire in 1965, and cancer deaths over 12.7 years were analyzed with drinking habits classified into five categories; nondrinker, ex-drinker, occasional drinker, and daily drinker whose intake of alcohol was equivalent to less than 2 or 2 and more go of sake (1 go sake congruent to 27 ml alcohol). Both age and smoking habits were taken into account in the calculation of death rates based on man-years at risk. Logistic regression analysis was also performed on cummulative mortality data. Upper aerodigestive cancer was strongly associated with alcohol consumption, giving some confidence in the validity of the present study. Excluding ex-drinkers, the risk of stomach cancer and liver cancer was gradually increased from nondrinkers to daily drinkers with lower intake of alcohol, but no further increase was noted for daily drinkers with larger consumption. Logistic regression did not show any significant associations between drinking habits and these two cancers, but the number of deaths from liver cancer was still small. Not particular patterns were observed for cancers of the large bowel and lung.