Determining the effect of dietary factors on cancer is a crucial issue when accounting for the effect of other major risks, such as smoking and drinking. A total of 15,563 adults from the Korean National Cancer Center Community Cohort were analyzed to determine and to compare the effect of dietary factors on stomach and colorectal cancer in overall and in the subgroup of non-smokers (or urinary cotinine concentrations <5 ng/mg) and non-drinkers with Cox proportional-hazard models. During the mean follow-up (13.7 years), 469 and 299 cases of stomach and colorectal cancer were identified, respectively. The preventive effect of vegetable, fish, and soybean/tofu intake on colorectal cancer was found in women after adjustment for smoking, drinking, BMI, and sociodemographic factors. In the subgroup analysis of non-smokers and non-drinkers, the effect on colorectal cancer was increased in women (≥1 time/week vs. almost never, vegetables: hazard ratio (HR) 0.30, 95% confidence interval (CI) 0.13-0.69; fish: HR 0.46, 95% CI 0.26-0.83), and the fresh fish intake effect on stomach cancer was newly identified in men (HR 0.36, 95% CI 0.15-0.86). These effects were more pronounced and additionally shown in other dietary factors such as soybean or tofu in women and vegetables and fish in men, when subjects with <5 ng/mg urinary cotinine concentrations applied. The protective effect of healthy eating on the risk of stomach and colorectal cancer were different by smoking and drinking status. Rigorous control of smoking and drinking effects is necessary when measuring the effect of dietary factors on cancer, properly.
Read full abstract