Estimation of Warfarin Maintenance Dose Based on VKORC1 ( 1639 G A) and CYP2C9 Genotypes

Transcription

1 Clinical Chemistry 53: (2007) Molecular Diagnostics and Genetics Estimation of Warfarin Maintenance Dose Based on VKORC1 ( 1639 G A) and CYP2C9 Genotypes Yusheng Zhu, 1* Michael Shennan, 2 Kristen K. Reynolds, 1,3 Nancy A. Johnson, 3 Matthew R. Herrnberger, 3 Roland Valdes, Jr., 1,3 and Mark W. Linder 1,3* 1 Department of Pathology and Laboratory Medicine, University of Louisville School of Medicine, Louisville, KY. 2 Luminex Molecular Diagnostics (formerly Tm Bioscience), Toronto, ON, Canada. 3 PGXL Laboratories, Louisville KY. Current affiliation: Department of Pathology and Laboratory Medicine, Medical University of South Carolina, Charleston, SC. * Address correspondence to these authors at: Department of Pathology and Laboratory Medicine, University of Louisville School of Medicine, Louisville, KY Fax ; mwlind01@louisville.edu or Department of Pathology and Laboratory Medicine, Medical University of South Carolina, Charleston, SC Fax ; zhuyu@musc.edu. Received August 11, 2006; accepted April 12, Previously published online at DOI: /clinchem Background: CYP2C9 polymorphisms are associated with decreased S-warfarin clearance and lower maintenance dosage. Decreased expression of VKORC1 resulting from the 1639G>A substitution has also been implicated in lower warfarin dose requirements. We investigated the additional contribution of this polymorphism to the variance in warfarin dose. Methods: Sixty-five patients with stable anticoagulation were genotyped for CYP2C9 and VKORC1 with Tag-It allele-specific primer extension technology. Plasma S- warfarin concentrations and warfarin maintenance dose were compared among patients on the basis of the VKORC1 1639G>A genotype. Results: Eighty percent of CYP2C9*1/*1 patients stabilized on <4.0 mg/day warfarin had at least 1 VKORC1 1639A allele. Mean warfarin doses (SD) were 6.7 (3.3), 4.3 (2.2), and 2.7 (1.2) mg/day for patients with the VKORC1 1639GG, GA, and AA genotypes, respectively. Steady-state plasma concentrations of S-warfarin were lowest in patients with the VKORC1 1639AA genotype and demonstrated a positive association with the VKORC1 1639G allele copy number (trend P 0.012). A model including VKORC1 and CYP2C9 genotypes, age, sex, and body weight accounted for 61% of the variance in warfarin daily maintenance dose. Conclusions: The VKORC1 1639A allele accounts for low dosage requirements of most patients without a CYP2C9 variant. Higher plasma S-warfarin concentrations corresponding to increased warfarin maintenance dosages support a hypothesis for increased expression of the VKORC1 1639G allele. VKORC1 and CYP2C9 genotypes, age, sex, and body weight account for the majority of variance in warfarin dose among our study population American Association for Clinical Chemistry The management of warfarin therapy is challenging because of the large variability in dose requirements ( mg/day) (1, 2) and potential for severe bleeding (3). The CYP2C9 enzyme accounts for 80% 85% of the metabolism of the pharmacologically more potent S-warfarin enantiomer. Warfarin inhibits vitamin K epoxide reductase (VKOR), 5 an enzyme complex responsible for recycling reduced vitamin K, through interaction with subunit protein 1 (VKORC1). Genetic variants of both CYP2C9 6 and VKORC1 influence warfarin maintenance dose requirements (4). Five principal alleles of CYP2C9, including CYP2C9*2 (430C T), *3 (1075A C), *4 (1076T C), *5 (1080C G), and *6 (818delA) (5) ( ki.se), are associated with decreased metabolic efficiency of the CYP2C9 enzyme (6, 7) and increased risk of bleeding. Several common single nucleotide polymorphisms of VKORC1 represent haplotypes associated with clinically significant differences in warfarin maintenance dose re- 5 Nonstandard abbreviations: VKOR, vitamin K epoxide reductase; VKORC1, VKOR subunit protein 1; MFI, median fluorescence intensity. 6 Human genes: CYP2C9, cytochrome P450, family 2, subfamily C, polypeptide 9; VKORC1, vitamin K epoxide reductase complex, subunit

2 1200 Zhu et al.: VKORC1 and CYP2C9 and Warfarin Maintenance Dose quirements (8). Of the nucleotide substitutions that differentiate interindividual differences in warfarin dose requirements, the 1639G A nucleotide substitution has been shown to be useful for diagnosis (9, 10). In addition, several other nucleotide sequence variants of VKORC1 have been identified in patients who require extremely high warfarin doses (range, 15 to 30 mg/day) (8, 11 13). The frequencies of these VKORC1 variants in general populations and among patients who demonstrate warfarin resistance are not known. Recently, warfarin dosing algorithms based on age, height or weight, CYP2C9*2, CYP2C9*3, and VKORC1 ( 1639G A) genotypes have been developed to estimate the warfarin dosage for patients during maintenance therapy (9, 14). A multiplex assay is commercially available for the simultaneous determination of both CYP2C9 and VKORC1 genotypes. This assay includes what can be considered 3 categories of genetic variants based on the level of information available for each. Category 1 would include the CYP2C9*2 and CYP2C9*3 alleles in addition to the VKORC1 1639G A substitution. For these genetic variants, there is sufficient evidence for their influence on warfarin dose to permit their inclusion into quantitative dosing models. Category 2 includes the CYP2C9*5 and CYP2C9*6 alleles. For these alleles, there is strong evidence to support their utility for identification of individuals with decreased S-warfarin metabolism; however, owing to the rarity of these genetic variants, insufficient information is available to permit their inclusion into quantitative dosing models. Finally, category 3 includes the rare variants of VKORC1 for which there is the least information available and CYP2C9*4, which to date has been identified in only a few individuals. This assay provides for a diagnostic tool based on the validated CYP2C9 and VKORC1 genetic variants and will permit future studies directed toward quantifiable influences of CYP2C9*5 and CYP2C9*6 on dose requirement. Furthermore, this tool will facilitate defining the associations between the coding region variants of VKORC1 and warfarin resistance. We evaluated this multiplex genotyping assay, which simultaneously detects 5 CYP2C9 and 7 VKORC1 singlenucleotide variations, to further investigate the relationship of VKORC1 variants, warfarin maintenance dose, and steady-state S-warfarin plasma concentrations in stabilized patients and to determine the most significant factors within our study population that contribute to variance in warfarin maintenance dosages. Materials and Methods vkorc1 and cyp2c9 genotyping We simultaneously detected CYP2C9*2 (rs ), *3 (rs ), *4, *5 (rs ), and *6 (rs ) alleles and VKORC1 1639G A (rs ), 85G T, 121G T, 134T C, 172A G, 1331G A, and 3487T G variants by use of the 2C9 VKORC1 Mutation Detection Kit as indicated in the manufacturer s product label (formerly Tm Bioscience, now Luminex Molecular Diagnostics) (15, 16). Detection of the CYP2C9*1 allele is based on the absence of all CYP2C9 nucleotide changes detected by the assay. Briefly, an initial PCR produced 7 amplicons encompassing the locations of 12 single-nucleotide variations, which we treated with shrimp alkaline phosphatase and exonuclease I. We subjected the treated PCR product to allele-specific primer extension in the presence of tag-labeled allele-specific primers and biotinylated dctp. We then incubated products of this reaction in the presence of fluorescent microspheres with covalently attached anti-tag sequences for hybridization capture and sorting of the extended tag-labeled primers. After hybridization and washing, we incubated extended allele-specific primers containing biotinylated cytosine in the presence of streptavidin-r-phycoerythrin (Molecular Probes). Finally, we sorted beads on the basis of fluorescence and measured the median fluorescence intensity (MFI) of R-phycoerythrin by use of a Luminex 100 TM System (16). samples The patient samples were used with institutional review board approval and informed consent and have been described in detail (17). Data for warfarin maintenance doses and steady-state S-warfarin concentrations were obtained from 65 white patients [35 men, mean age 66.8 years (41 89); 30 women, mean age 70.6 years (48 92)] undergoing warfarin maintenance anticoagulation therapy (17). All patients were treated with warfarin for the primary indication of atrial fibrillation. All patients were receiving multiple therapies, with no differences in polypharmacy between groups used for comparison (P 0.6). Patients were enrolled when consistent dosing yielded international normalized ratios between 2.0 and 3.0 for 3 consecutive months. We used DNA samples stored at 70 C (17) for the combined CYP2C9 and VKORC1 genotyping. A subset of these samples was also used to evaluate the imprecision and reproducibility of the multiplex assay for the specific alleles identified in this population. assay evaluation Imprecision of allelic ratios and reproducibility of genotypes. We determined imprecision of allelic ratios and reproducibility of genotype calls for selected patients (n 11 total) with CYP2C9*1/*1, *1/*2, *2/*2, *1/*3, and *3/*3 and VKORC1 ( 1639G A) GG, GA, and AA genotypes on the basis of replicate runs of each sample on 8 separate days. Accuracy of the assay. By use of the ABI PRISM BigDye Terminator v3.1 Cycle Sequencing Kit (Perkin-Elmer), we confirmed the genotype determination for 30 patient samples and 7 synthetic DNA constructs containing the rare variant nucleotides of interest, CYP2C9*4 and VKORC1 85G T, 121G T, 134T C, 172A G, 1331G A, and 3487T G, by bidirectional sequencing across all 7 PCR amplicons generated in the assay.

3 Clinical Chemistry 53, No. 7, measurement of steady-state concentrations of s-warfarin We measured plasma S-warfarin by HPLC with ultraviolet detection according to the method of Henne et al. (18). statistical analysis We used one-way ANOVA followed by post hoc test with Bonferroni correction to compare allelic ratios, warfarin doses, and plasma S-warfarin concentrations between groups with GraphPad Prism version 3.0 (GraphPad). We compared variation frequencies of VKORC1 ( 1639G A) in patients with CYP2C9*1/*1 genotype vs patients whose CYP2C9 genotypes include 1 or more variant alleles and estimated departure from Hardy Weinberg equilibrium by use of 2 test with 1 degree of freedom. We calculated 95% CIs according to the method of Newcombe and Altman (19) and performed multiple linear-regression analysis by use of Stata/SE 9.1 (StataCorp). Results imprecision of allelic ratios and reproducibility of genotype calls We determined genotype on the basis of the ratio of MFI for each single nucleotide variation tested according to the following formula: M/(M W), where W is the MFI of the wild-type allele signal and M is the MFI of the variant allele signal. Imprecision results (samples measured on 8 separate days) showed that genotype calls for each of the samples were 100% concordant in each of the separate runs. The means and SDs of the allelic ratios are shown in Table 1. accuracy We tested 30 genomic DNA samples and 7 synthetic DNA plasmids each containing 1 of the rare (CYP2C9*4 and VKORC1 85G T, 121G T, 134T C, 172A G, 1331G A, and 3487T G) nucleotide variants with the allele-specific primer extension assay and confirmed the results by direct sequencing. Table 1. Allelic ratios of patients with different genotypes. Genotype Allelic ratio, mean (SD) n a CYP2C9*1/*1 (430CC) (0.005) 63 CYP2C9*1/*2 (430CT) (0.009) 25 CYP2C9*2/*2 (430TT) (0.007) 9 CYP2C9*1/*1 (1075AA) (0.012) 70 CYP2C9*1/*3 (1075AC) (0.012) 18 CYP2C9*3/*3 (1075CC) (0.005) 9 VKORC1 1639GG (0.006) 54 VKORC1 1639GA (0.010) 34 VKORC1 1639AA (0.009) 9 a Number of replicates of genotyping. frequency of vkorc1 ( 1639g a) among patients prescribed warfarin Using the multiplex assay, we genotyped 65 white patients stabilized on warfarin. Genotype results for CYP2C9 are in complete agreement with our prior results. As previously reported, the frequencies of the CYP2C9*2 and CYP2C9*3 alleles in this population are and 0.065, respectively (18). In this small sample, CYP2C9*4, *5, and *6 were not detected, nor were any of the rare VKORC1 coding-region variants. With regard to the VKORC G A substitution, we identified 34 (52.3%; GG), 23 (35.4%) heterozygous variant (GA), and 8 (12.3%) homozygous variant (AA) individuals. The allele frequencies for 1639G and 1639A alleles were 0.70 (95% CI ) and 0.30 (95% CI ), respectively. The frequencies of the G allele in this population are slightly higher than previously observed in 2 prior reports involving predominantly white populations, in which reported frequencies for 1639G allele were 0.57 (9) and (20). The observed VKORC genotype distribution showed no deviation from Hardy Weinberg equilibrium ( , P 0.05). warfarin dose requirement and vkorc1 ( 1639g a) genotype The mean (SD) daily doses for the individuals with VKORC1 GG, GA, and AA genotypes were 6.7 (3.3; n 34), 4.3 (2.2; n 23), and 2.7 (1.2) mg (n 8), respectively. Differences in warfarin daily maintenance dose between the GG and GA groups and the GG and AA groups were statistically significant (P 0.01). We previously reported differences in warfarin dose requirement based on CYP2C9 genotype alone for this population of individuals (17). As illustrated in Fig. 1, although the majority (60%) of patients requiring dosages of 4 mg/day were found to be carriers of either the CYP2C9*2 or CYP2C9*3 alleles (15 of 25, designated V for variant), 10 patients in this dose requirement interval were identified as CYP2C9*1/*1. Reanalysis of this population based on the more comprehensive multiplex assay ruled out any contribution of the more rare CYP2C9*5 and *6 variants and revealed 80% of these individuals as VKORC GA (4 patients) or VKORC1 1639AA (4 patients). Overall, 23 of 25 patients (92%) with dose requirements of 4.0 mg/day were found to possess variant alleles of either CYP2C9 (4 patients), VKORC1 (8 patients), or CYP2C9 and VKORC1 combined (11 patients). s-warfarin concentrations of patients with various vkorc1 ( 1639g a) genotypes Previous work has demonstrated that there are no differences in S-warfarin clearance among patients with different VKORC genotypes (9). Therefore, to account for the differences in warfarin dose requirements among the VKORC genotypes, we compared the mean plasma concentrations of S-warfarin among these groups. The mean (SD) plasma S-warfarin concentrations of the

4 1202 Zhu et al.: VKORC1 and CYP2C9 and Warfarin Maintenance Dose not include other CYP2C9 alleles. The full model accounts for 61% of the variance in log warfarin daily maintenance dose (R , P ), with an SE of the dose estimate of 0.39 mg (Fig. 2). Fig. 1. Histogram representing the CYP2C9 and VKORC1 genotypes and warfarin maintenance dose requirements of the study population. Patients were divided into 4 genotype categories and data were plotted in the form of a histogram. CYP2C9*1/*1: VKORC1GG (open squares); CYP2C9*1/*1: VKORC1GA or AA (filled squares); CYP2C9 with V (V indicates all CYP2C9 genotypes that include 1 or 2 variant alleles, CYP2C9*1/*2, *1/*3, *2*2, *2*3, and *3*3); VKORC1GG (open triangles) and CYP2C9 with V: VKORC1GA or AA (filled triangles). Note that 8 of 10 patients with maintenance dosages of 4 mg/day previously determined to have the CYP2C9*1/*1 genotype were found to possess 1 or 2 VKORC1 1639A alleles. patients in the GG, GA, and AA groups were 0.73 (0.33; n 29), 0.54 (0.21; n 21), and 0.48 (0.04) mg/l (n 8), respectively (P 0.035). There was a positive association between S-warfarin plasma concentrations at steady state and the 1639G allele copy number (trend P 0.012). multivariate analysis of the relationship between maintenance dose and patient characteristics We conducted a natural logarithm transformation of daily dose to create a gaussian distribution for further analysis (P ). Multiple linear regression of the logarithm of daily dose was modeled for each of the variables listed in Table 2 to evaluate individual contributions to the variance in warfarin maintenance dose. Because only CYP2C9*2 and *3 were found in this study, our model did Discussion Recent studies clearly demonstrate an association between warfarin dose requirements and genetic variations of VKORC1 and CYP2C9 (9, 14, 21, 22). In collaboration with Tm Bioscience (now Luminex Molecular Diagnostics), we developed a robust multiplex genotyping assay for the simultaneous determination of 5 CYP2C9 and 7 VKORC1 variants, by use of Tag-It TM allele-specific primer extension technology. The rationale for developing the assay on this platform is that several clinical laboratories are now providing pharmacogenetic testing services for cytochrome P450 genes and other clinically relevant molecular diagnostic services that use this technology. This technology differentiates genotypes on the basis of the ratio of the signals produced for each of the alternative nucleotides for a given position in the gene. Thus 1 critical operational aspect is for there to be clear differentiation between the allelic ratios obtained for each of the potential genotype determinations. As is true for other assays that we have developed with this technology (16), this assay demonstrates a high degree of precision in the allelic ratio for each nucleotide position tested by the method, thus yielding a high degree of certainty with regard to differentiation of the genotype. Furthermore, through the analysis of a variety of synthetic and genomic DNA samples, providing a diverse array of possible genotype structures, we have demonstrated the high degree of accuracy that can be achieved with this technology. This assay was purposefully designed to serve both clinical and research laboratories. Inclusion of the promoter region nucleotide variant of VKORC1 ( 1639 G A) in combination with CYP2C9 variants is useful for accounting for variability of warfarin dose requirements in the interval of 1.5 to 12 mg/day. Additional evi- Table 2. Multiple regression analysis for modeling warfarin daily dose requirements based on age, sex, weight, VKORC1 ( 1639G>A), and CYP 2C9 genotypes. a Predictor Regression equation Model P value R 2 Age ln (D) Sex ln (D) Weight ln (D) VKORC1 ( 1639G A) ln (D) (VKORC1-AA) (VKORC1-GG) CYP2C9 ln (D) (2C9*2) (2C9*3) Full model (all variables) ln (D) (age) (sex) (weight) (VKORC1-AA) (VKORC1-GG) (2C9*2) (2C9*3) a Age, input age in years; sex, input 0 for female and 1 for male; weight, input weight in lb; VKORC1 ( 1639AA), input 0 for GG, 0 for GA, and 1 for AA; VKORC1 ( 1639GG), input 1 for GG, 0 for GA, and 0 for AA; CYP2C9, input 0, 1, or 2 for the number of CYP*2 and *3 alleles.

5 Clinical Chemistry 53, No. 7, Fig. 2. Correlation between actual and calculated daily warfarin maintenance dosages. Actual daily maintenance doses for each of the 65 individuals were plotted against the dose requirements estimated with the full computational model shown in Table 2. dence is needed to advance the knowledge pertaining to the association between rare coding region variants of VKORC1 and warfarin dose requirements that exceed mg/day. Warfarin resistance that is unexplained by excessive vitamin K ingestion, poor drug absorption, or drug interaction has been recognized previously as a heritable trait (2, 23 25); however, only recently have genetic alterations in VKORC1 coding regions been implicated as contributing to this trait (8, 11 13). For example, Rost et al. (11) described 4 cases of warfarin resistance associated with 4 different gene mutations of VKORC1: Val29Leu (85 G T), Val45Ala (134 T C), Arg58Gly (172 A G), and Leu 128Arg (3487 T G). Rieder et al. (8) found a mutation of VKORC1 (Ala41Ser, 121G T) in a patient requiring 15.5 mg/day, and Harrington et al. (12) identified another mutation of VKORC1 (Val66Met, 1331 G A) in a patient who required more than 25 mg warfarin per day. Bodin et al. (13) also reported a Leu128Arg (3487 T G) mutation of VKORC1 in a patient resistant to all vitamin K antagonists, including warfarin up to 45 mg/day. However, the frequencies of the above mutations of the VKORC1 gene in different populations are not well documented. The application of this assay will help address these questions and potentially help identify patients with extremely high warfarin dose requirements. By use of this assay, we determined the VKORC1 genotypes for 65 white patients stabilized on warfarin therapy who had been previously genotyped for CYP2C9. In the original study we were able to demonstrate significant differences in warfarin dose requirements based solely on CYP2C9 genotype; however, 10 patients who were identified as having the CYP2C9*1/*1 genotype were found to have daily warfarin dose requirements of 3.5 mg/day. Reanalysis of this population including the VKORC1 1639G A substitution revealed that 8 (80%) of these patients have 1 or 2 VKORC1 1639A alleles, a finding that likely accounts for their low dose requirement. Within this population, we demonstrated significant differences in the mean daily warfarin dose requirement among all 3 possible VKORC genotypes. The mechanism to account for decreased warfarin dose requirements as a consequence of the 1639A-containing haplotype is believed to be associated with decreased transcription of this allele (8, 20). Our data support this hypothesis by demonstrating a positive trend between the VKORC1 1639G allele copy number and higher steadystate plasma concentrations of S-warfarin in stabilized patients. In late 2005, Sconce et al. (9) reported a significant advancement in the approach to using genomic information for estimation of warfarin dose requirements. A multiple linear regression analysis was developed on the basis of patient age, height, VKORC1 1639G A, and CYP2C9 genotype, including unique contributions of both the CYP2C9*2 and CYP2C9*3 alleles for estimation of warfarin dose requirement. Since that publication, to our knowledge, 3 other groups in addition to our current report have published similar approaches to evaluating the contribution of multiple factors to the variation in warfarin dose requirement. Important similarities and differences exist between these studies and ours. We and 2 other groups (9, 14) report that warfarin dose requirements of the study populations do not show a gaussian distribution. Sconce et al. (9) handled this using a square root transformation of the data, whereas we and Tham et al. (14) reported logarithmic transformations. We compared both transformations and found that the square root transformation nearly failed the Shapiro Wilk test for gaussian distribution (P 0.06), whereas logarithmic transformation yielded a highly significant test for gaussian distribution of the transformed data (P 0.69). Kimura et al. (21) found that the variables of age, sex, and weight were significantly associated with daily warfarin dose. Tham et al. (14) also noted significant influences of age and weight; however, in their study population sex was not found to be a significant predictor of warfarin dose requirement. The inclusion of height as a variable is unique to the Sconce et al. (9) report. Weight-based dosing of warfarin is a more conventional approach to warfarin dose adjustments and is included in our model. Another important aspect of our model is the ability to include unique coefficients for each of the CYP2C9*2 and CYP2C9*3 alleles. Although 2 other reports also include a weight-based adjustment to dosing, owing to the low frequency of the CYP2C9*2 allele among the populations studied, neither Tham et al. (14) nor Kimura et al. (21) were able to include the significance of the CYP2C9*2 allele in their final models. Overall, the model that we developed combines many of the most favorable aspects of prior approaches, specifically including logarithmic transformation of the dose, inclusion of terms for sex and weight, and the ability to individually represent the influence of the CYP2C9*2 vs CYP2C9*3 alleles. This

6 1204 Zhu et al.: VKORC1 and CYP2C9 and Warfarin Maintenance Dose model will require validation by comparison of estimated vs actual dosages of an independent population sample. Regardless of the model applied, 40% of the variance in warfarin dose requirement remains unaccounted for on the basis of known factors (26 28). Although at this time there is insufficient evidence to support inclusion of additional known genetic variables beyond CYP2C9 and VKORC1 in algorithms for estimation of warfarin maintenance dose, several investigators have begun to identify additional genetic characteristics that may eventually help estimate dose requirements. For example, Wadelius et al. (27) added gamma-glutamyl carboxylase genotypes to the model; the explanatory value increased only marginally. Shikata et al. (28) found warfarin sensitivity to be independently associated with the 402G A nucleotide substitution of the factor VII gene, the (CAA repeat)n of the gamma-glutamyl carboxylase gene, CYP2C9*3, and the 165 Thr Met variant of the factor II gene. The combined influence of these attributes accounted for 50% of variance in warfarin dose. It is possible that this profile could be further improved by the addition of VKORC1. Additionally, variation in apolipoprotein E (29), multidrug resistance 1 (MDR1) (27), calumenin (30), and possibly genes encoding other components of the VKOR complex (11) may also play important roles. Our data are the first to support the hypothesis that the abundance of VKORC1 protein may be greater based on the 1639G-containing haplotype (8), a hypothesis originally based on VKORC1 mrna measurements. Furthermore, we report a multiple linear regression model for estimation of warfarin maintenance dose. The physical characteristics included in our model are a unique combination that includes the more conventional measure of weight vs height. Although our study involves 70 individuals, we were able to demonstrate statistically significant and independent effects of both CYP2C9*2 and CYP2C9*3 in addition to the VKORC G A. Genotyping both VKORC1 and CYP2C9 in conjunction with the patient s physical characteristics will help physicians and pharmacists estimate warfarin dose more precisely and thus improve the efficiency of the dosage titration process. Grant/funding support: This work was supported in part by K23 AA (to M.W.L.), and Y.Z. received partial support for postdoctoral training through a grant from Tm Bioscience. Financial disclosures: M.W.L. has received honoraria from Tm Bioscience as an invited speaker. Acknowledgments: We acknowledge the helpful discussions with Jim Gordon (Tm Bioscience) concerning assay design. References 1. James AH, Britt RP, Raskino CL, Thompson SG. Factors affecting the maintenance dose of warfarin. J Clin Pathol 1992;45: Hallak HO, Wedlund PJ, Modi MW, Patel IH, Lewis GL, Woodruff B, et al. High clearance of (S)-warfarin in a warfarin-resistant subject. Br J Clin Pharmacol 1993;35: Gullov AL, Koefoed BG, Petersen P. Bleeding complications to long-term oral anticoagulant therapy. J Thromb Thrombolysis 1994;1: Reynolds KK, Valdes R Jr, Hartung BR, Linder MW. Individualizing warfarin therapy. Pers Med 2007;4: Lee CR, Goldstein JA, Pieper JA. Cytochrome P450 2C9 polymorphisms: a comprehensive review of the in-vitro and human data. Pharmacogenetics 2002;12: Takahashi H, Kashima T, Nomoto S, Iwade K, Tainaka H, Shimizu T, et al. Comparisons between in-vitro and in-vivo metabolism of (S)-warfarin: catalytic activities of cdna-expressed CYP2C9, its Leu359 variant and their mixture versus unbound clearance in patients with the corresponding CYP2C9 genotypes. Pharmacogenetics 1998;8: Rettie AE, Wienkers LC, Gonzalez FJ, Trager WF, Korzekwa KR. Impaired (S)-warfarin metabolism catalysed by the R144C allelic variant of CYP2C9. Pharmacogenetics 1994;4: Rieder MJ, Reiner AP, Gage BF, Nickerson DA, Eby CS, McLeod HL, et al. Effect of VKORC1 haplotypes on transcriptional regulation and warfarin dose. N Engl J Med 2005;352: Sconce EA, Khan TI, Wynne HA, Avery P, Monkhouse L, King BP, et al. The impact of CYP2C9 and VKORC1 genetic polymorphism and patient characteristics upon warfarin dose requirements: proposal for a new dosing regimen. Blood 2005;106: Bodin L, Verstuyft C, Tregouet DA, Robert A, Dubert L, Funck- Brentano C, et al. Cytochrome P450 2C9 (CYP2C9) and vitamin K epoxide reductase (VKORC1) genotypes as determinants of acenocoumarol sensitivity. Blood 2005;106: Rost S, Fregin A, Ivaskevicius V, Conzelmann E, Hortnagel K, Pelz HJ, et al. Mutations in VKORC1 cause warfarin resistance and multiple coagulation factor deficiency type 2. Nature 2004;427: Harrington DJ, Underwood S, Morse C, Shearer MJ, Tuddenham EG, Mumford AD. Pharmacodynamic resistance to warfarin associated with a Val66Met substitution in vitamin K epoxide reductase complex subunit 1. Thromb Haemost 2005;93: Bodin L, Horellou MH, Flaujac C, Loriot MA, Samama MM. A vitamin K epoxide reductase complex subunit-1 (VKORC1) mutation in a patient with vitamin K antagonist resistance. J Thromb Haemost 2005;3: Tham LS, Goh BC, Nafziger A, Guo JY, Wang LZ, Soong R, et al. A warfarin-dosing model in Asians that uses single-nucleotide polymorphisms in vitamin K epoxide reductase complex and cytochrome P450 2C9. Clin Pharmacol Ther 2006;80: Bortolin S, Black M, Modi H, Boszko I, Kobler D, Fieldhouse D, et al. Analytical validation of the Tag-It high-throughput microspherebased universal array genotyping platform: application to the multiplex detection of a panel of thrombophilia-associated singlenucleotide polymorphisms. Clin Chem 2004;50: Zhu Y, Hein DW, Doll MA, Reynolds KK, Abudu N, Valdes R Jr, et al. Simultaneous determination of 7 N-acetyltransferase-2 singlenucleotide variations by allele-specific primer extension assay. Clin Chem 2006;52: Linder MW, Looney S, Adams JE, 3rd, Johnson N, Antonino-Green D, Lacefield N, et al. Warfarin dose adjustments based on CYP2C9 genetic polymorphisms. J Thromb Thrombolysis 2002; 14: Henne KR, Gaedigk A, Gupta G, Leeder JS, Rettie AE. Chiral phase analysis of warfarin enantiomers in patient plasma in relation to CYP2C9 genotype. J Chromatogr B Biomed Sci Appl 1998;710:

7 Clinical Chemistry 53, No. 7, Newcombe RG, Altman DG. Statistics with Confidence. London: British Medical Journal, 1989: Yuan HY, Chen JJ, Lee MT, Wung JC, Chen YF, Charng MJ, et al. A novel functional VKORC1 promoter polymorphism is associated with inter-individual and inter-ethnic differences in warfarin sensitivity. Hum Mol Genet 2005;14: Kimura R, Miyashita K, Kokubo Y, Akaiwa Y, Otsubo R, Nagatsuka K, et al. Genotypes of vitamin K epoxide reductase, gammaglutamyl carboxylase, and cytochrome P450 2C9 as determinants of daily warfarin dose in Japanese patients. Thromb Res 2006 [Epub ahead of print]. 22. Wadelius M, Chen LY, Eriksson N, Bumpstead S, Ghori J, Wadelius C, et al. Association of warfarin dose with genes involved in its action and metabolism. Hum Genet 2007;121: Diab F, Feffer S. Hereditary warfarin resistance. South Med J 1994;87: Alving BM, Strickler MP, Knight RD, Barr CF, Berenberg JL, Peck CC. Hereditary warfarin resistance: investigation of a rare phenomenon. Arch Intern Med 1985;145: Warrier I, Brennan CA, Lusher JM. Familial warfarin resistance in a black child. Am J Pediatr Hematol Oncol 1986;8: Takahashi H, Wilkinson GR, Nutescu EA, Morita T, Ritchie MD, Scordo MG, et al. Different contributions of polymorphisms in VKORC1 and CYP2C9 to intra- and inter-population differences in maintenance dose of warfarin in Japanese, Caucasians and African-Americans. Pharmacogenet Genomics 2006;16: Wadelius M, Chen LY, Downes K, Ghori J, Hunt S, Eriksson N, et al. Common VKORC1 and GGCX polymorphisms associated with warfarin dose. Pharmacogenomics J 2005;5: Shikata E, Ieiri I, Ishiguro S, Aono H, Inoue K, Koide T, et al. Association of pharmacokinetic (CYP2C9) and pharmacodynamic (factors II, VII, IX, and X; proteins S and C; and gamma-glutamyl carboxylase) gene variants with warfarin sensitivity. Blood 2004; 103: Visser LE, Trienekens PH, De Smet PA, Vulto AG, Hofman A, van Duijn CM, et al. Patients with an ApoE epsilon4 allele require lower doses of coumarin anticoagulants. Pharmacogenet Genomics 2005;15: Vecsler M, Loebstein R, Almog S, Kurnik D, Goldman B, Halkin H, et al. Combined genetic profiles of components and regulators of the vitamin K-dependent gamma-carboxylation system affect individual sensitivity to warfarin. Thromb Haemost 2006;95: