Abstract
A cohort of 5,477 male Japanese physicians was studied to examine the relationship between smoking habits and mortalities from cancer, coronary heart disease (CHD) and stroke over 12.7 years. The logistic regression analysis based on proportional hazard models was used for statistical assessment. The risks of both lung cancer and CHD were strongly associated with smoking habits in terms of the number of cigarettes smoked per day, inhalation level and age at starting to smoke. These associations were not influenced by the effect of drinking habits. However, the risk increment of lung cancer due to cigarette smoking was fairly small as compared with the data from other studies of male Caucasians. A statistically significant association was observed between upper aerodigestive cancer and cigarette smoking. But this relationship became insignificant after adjustment for drinking habits, and the risk of heavy smokers was drastically reduced. No clear association was noted between smoking and mortalities from gastric cancer and stroke.