CYP2C9 polymorphism and warfarin dose requirements

We were interested to read Dr Goldstein's review on the clinical relevance of genetic polymorphisms in the human CYP2C subfamily [1]. While we believe this to be a generally wide-ranging and useful contribution, we were surprised to read that a recent paper by us reporting that CYP2C9 genotype was a good predictor of warfarin dose requirement [2] was said to involve the use of a PCR assay with primers which were not specific for CYP2C9 but also amplified CYP2C19. The review therefore suggests that our study might have underrepresented the incidence of the ‘PM allele’ in the low warfarin dose requirement group. Though we agree that the study of Yasar and colleagues [3] indicated that the primers we used for CYP2C9 genotyping, which were originally designed by Wang et al. [4], might amplify other related sequences at a low level, we were aware of this confounder prior to the commencement of our warfarin study. We found that the problem could be avoided by using the maximum annealing temperatures (60 ° C for CYP2C9*2 and 61 ° C for CYP2C9*3 for 2 min on a Techne PHC-3 heating block), compatible with successful PCR amplification. Restriction digests of samples homozygous for CYP2C9*2 and CYP2C9*3 showed that no CYP2C sequences other than CYP2C9 were amplified.

We believe that the accuracy of the assay in our hands is also confirmed by the fact that in our control group, we obtained similar genotype distributions to those obtained by others for Northern European populations (see Table 1). In addition, our findings with respect to warfarin dose requirement and CYP2C9 genotype have been replicated in a recent large study [5] which used different PCR primers for CYP2C9 amplification. While the finding that CYP2C9*2 heterozygotes have a lower warfarin dose requirement to homozygous wild-type subjects might seem unexpected in the light of early studies using a yeast expression system which suggested that the Arg144Cys substitution did not affect S-warfarin 7-hydroxylation [6], others have observed a highly significant decrease in V_max for S-warfarin 7-hydroxylation by this substitution in a human cell line expression system [7]. This discrepancy between expression systems may be due to differences in cytochrome P450-cytochrome P450 oxidoreductase ratios between human and yeast cell lines [8]. S-warfarin does not appear to be the only CYP2C9 substrate where the CYP2C9*2 allele gene product shows functionally significant differences in catalytic activity from the wild-type since individuals heterozygous for CYP2C9*2 have also been demonstrated to show slower metabolism of phenytoin compared with homozygous wild-type individuals [9].

Table 1.

CYP2C9 genotype frequencies in Northern European controls.

	UK [2]n = 100	UK [10]n = 100	Sweden [3]n = 430
*1/*1	60 (0.60)	62 (0.62)	287 (0.67)
*1/*2	20 (0.20)	19 (0.19)	80 (0.19)
*1/*3	17 (0.17)	15 (0.15)	50 (0.12)
*2/*3	2 (0.02)	0	8 (0.02)
*2/*2	0	3 (0.03)	2 (0.005)
*3/*3	1 (0.01)	1 (0.01)	3 (0.01)

The findings that individuals heterozygous for either CYP2C9*2 or CYP2C9*3 (approximately 40% of Northern Europeans) as well as those with two variant CYP2C9 alleles (approximately 3% of Northern Europeans) have on average a lower warfarin dose requirement than homozygous wild-type individuals are, we believe, of particular clinical relevance in view of the large number of patients who fall into these categories.