We examined the role of smoking and diet in the cross-cultural variation in lung-cancer mortality, using aggregated data of the Seven Countries Study, a follow-up study comprising 12,763 middle-aged men in 16 cohorts in Europe, the United States and Japan, which started around 1960. Smoking habits were assessed with a standardised questionnaire. Dietary intake was collected in random sub-samples of each cohort by the dietary record method. Cohort-specific 25-year lung-cancer mortality among all men and among categories of smoking behaviour was related to smoking prevalence and population average dietary intake, respectively, using Poisson regression. Smoking prevalence was positively associated with lung-cancer mortality [risk ratio 1.47, 95% confidence interval (CI) 1.05-2.07, for an increase of 10 percentage points]. Lung-cancer mortality among smokers, which varied significantly among cultures, was positively associated with average fat intake, especially saturated fat intake (rate ratio 1.10, 95% CI 1.04-1.17, for an increase of 4.6 g) but not with unsaturated fat intake. Average fruit and vegetable intake were not related to lung-cancer mortality. Among never-smokers, the power to detect associations was low. In conclusion, both smoking prevalence and average fat intake, especially saturated fat, may play a role in the cross-cultural variation in lung-cancer mortality, either independently or by effect modification.
Copyright 2000 Wiley-Liss, Inc.