The purpose of this study is to determine if cruciferous vegetables and coffee, two dietary inducers of glutatione-S-transferases, interact with GSTM-1 genotype to alter risk of colon cancer. Data were available on 1579 incident cases of adenocarcinoma of the colon and 1898 population-based controls. Intake of cruciferous vegetables, specific types of cruciferous vegetable, and coffee were not associated with colon cancer; GSTM-1 genotype did not modify these associations. However, age at diagnosis and cigarette smoking appeared to be important effect modifiers of the associations between GSTM-1, cruciferous vegetables and colon cancer. Among GSTM-1 null individuals, <55 years at diagnosis, we observed an inverse association between colon cancer and high levels of cruciferous vegetable intake relative to people who did not eat cruciferous vegetables (ORs 0.23 95% CI 0.10-0.54); broccoli was the cruciferous vegetable associated with the strongest inverse association (OR 0.30 95% CI 0.13-0.70). Among younger individuals who were GSTM-1 present (relative to those with GSTM-1 null), we observed an inverse association with colon cancer regardless of level of cruciferous vegetable intake (OR 0.74 95% CI 0.30-1.79 for no intake; OR 0.44 95% CI 0.21-0.92 for <4 servings/week; and OR 0. 44 95% CI 0.19-0.99 for >/=4 servings/week). These associations were further modified by cigarette smoking. People <65 years of age who smoked had a greater reduction in risk of colon cancer from consumption of cruciferous vegetables than non-smokers at the same age. In summary, although cruciferous vegetables do not appear to modify colon cancer risk in the total population, there are subgroups of the population for whom these vegetables may be important. These subgroups are defined mostly by age and smoking status.