%; P<0.001), which represents an average 18% reduction among the cases. The variant Lys751Gln and Asp312Asn allele frequencies were 0.36 and 0.29, respectively, for the cases and 0.33 and 0.27, respectively, for the controls. For subjects homozygous for the variant genotype at either locus, the adjusted odds ratio [95% confidence interval (CI)] was 1.84 (1.11-3.04; P 5-0.018, for trend). Both cases and controls with the wild-type genotypes exhibited the most proficient DRC. The risk (95% CI) for suboptimal DRC (defined as less than the median DRC value among the controls) was 1.57 (0.743.35) for those with the Gln/Gln751 genotype. For cases with the Asn/Asn312 genotype, the risk (95% CI) was 3.50 (1.06–11.59). For cases who were homozygous at either locus, the risk was 2.29 (1.03-5.12; P 5-0.048, for trend). The pattern was less evident among the controls, although there was a nonsignificant 41% increase in the risk of suboptimal DRC for controls who were homozygous at either locus. These results suggest that the two XPD polymorphisms have a modulating effect on DRC, especially in the cases.