We investigated CYP1A1*2A, CYP1A1*2C, CYP1A2*1C, CYP1A2*1F, GSTM1 and NAT2 gene polymorphisms, involving enzymes which metabolize many carcinogens, with reference to colorectal cancer risk. The distribution of these genotypes was not associated with risk overall. However, the CYP1A1*2A T/C genotype showed a significant association with colorectal cancer risk in never-smokers (odds ratio [OR], 3.06; 95% confidence interval [95% CI], 1.11-8.40; p = 0.030). The risk of the NAT2 rapid genotype in never-smokers was also statistically significantly increased (OR, 5.38; 95%CI, 1.80-16.1; p = 0.003). Furthermore, the joint effects of NAT2 rapid plus other genotypes were associated with colorectal cancer overall (OR, 3.12; 95%CI, 1.15-8.51; p = 0.026, for NAT2 rapid plus combined CYP1A1*2C Ile/Val and Val/Val, OR, 3.25; 95%CI, 1.09-9.74; p = 0.035, for NAT2 rapid plus CYP1A2*1C G/G, and OR, 4.20; 95%CI, 1.09-16.1; p = 0.037, for NAT2 rapid plus GSTM1 null, respectively). In never-smokers, the joint effects of NAT2 rapid plus other genotypes were remarkable (OR, 15.9; 95%CI, 1.87-135.8; p = 0.011, for NAT2 rapid plus combined CYP1A1*2A T/C and C/C, OR, 5.71; 95%CI, 1.49-21.9; p = 0.011, for NAT2 rapid plus combined CYP1A1*2C Ile/Val and Val/Val, and OR, 9.14; 95%CI, 2.05-40.7; p = 0.004, for NAT2 rapid plus CYP1A2*1F A/A, respectively). The joint effect of CYP1A2*1F A/A plus CYP1A2*1C G/G genotypes was also increased in never-smokers (OR, 6.16; 95%CI, 1.26-30.1; p = 0.025). Our findings suggest that the CYP1A1*2A T/C and NAT2 rapid genotypes is associated with colorectal cancer susceptibility without smoking exposure. These results also indicate that the NAT2 in combination with CYP1A1*2C, CYP1A2*1C, or GSTM1 genotypes may strongly confer susceptibility to colorectal cancer. In particular, the combination of NAT2 plus CYP1A1*2A, CYP1A1*2C, or CYP1A2*1F genotypes, and that of CYP1A2*1F plus CYP1A2*1C genotype may define a group of persons who are genetically susceptible to colorectal cancer in never smokers.