Using data from a case-control study conducted between 1985 and 1992 in northern Italy on 828 cases of colon cancer, 498 cases of rectal cancer and 2,024 controls in hospital for acute, non-neoplastic, non-digestive tract disorders, we estimated the percent population attributable risk (PAR) for colorectal cancer in relation to β-carotene, vitamin C (as markers of a diet rich in fruit and vegetables), red meat and seasoning fat intake, daily meal frequency and family history of the disease. On the basis of multivariate odds ratios, adjusted for total calorie intake, a low intake of β-carotene accounted for 39% of all the cases and a low intake of vitamin C for 14%. These two micronutrients together explained 43% of all colorectal cancer cases in this population. A high frequency of intake of red meat consumption explained 17% of all cases, and a high score of seasoning fats 4%. A higher daily meal frequency was responsable for 13% of the cases, and these 5 dietary factors together explained 63% of colorectal cancer cases in this population. Family history of colorectal cancer accounted for 4% of all cases. These estimates were similar for colon and rectal cancers separately, in males and females, and in younger and elderly subjects, except for seasoning fats and family history, whose PARs were apparently greater for colon cancer and at younger age. Thus, even though available dietary data were limited in several aspects, and the PAR estimates were based on somewhat arbitrary assumptions regarding the exposure distribution, about two-thirds of all colorectal cancers in this population could be explained in terms of a few risk factors or risk indicators considered. This would correspond to the avoidance of a large proportion of the over 18,000 deaths from colorectal cancer registered per year in the whole of Italy. © 1996 Wiley-Liss, Inc.