Smoking and lung cancer mortality in Japanese men: estimates for dose and duration of cigarette smoking based on the Japan vital statistics data

Abstract

For the purpose of understanding human carcinogenesis and making a quantitative prediction of lung cancer mortality in a general population of Japanese males, we evaluated a statistical model which assumes lung cancer mortality to be proportional to the 4.5th power of the effective duration of cigarette smoking among smokers and to the 4th power of age among nonsmokers, using Japan Vital Statistics data. For the male birth cohorts aged 30-69 in 1965 in the age range of 40-79, studied by quinquennial calendar time intervals from 1955 to 1985, it was found that, (i) for nonsmokers, the estimated lung cancer mortality rate was comparable to the rates reported in the US or Britain, assigning 20 to 25% proportions of nonsmokers, (ii) for smokers, the estimated duration of smoking was shorter than would be expected from the age when smoking was started according to various epidemiological surveys, and (iii) the estimated average numbers of cigarettes smoked per day by smokers were similar to those obtained by epidemiological studies, when these were estimated by incorporating a part of Doll and Peto's dose-response relationship. Also discussed is the possibility of assessing lung cancer mortality risk for Japanese male smokers by means of the statistical model, alpha x (cigarettes smoked per day + beta) x (age - (age started smoking) - gamma)4.5.