American Association for Clinical Chemistry
Better health through laboratory medicine
June 2009 Clinical Laboratory News: Warfarin Dosing


June 2009: Volume 35, Number 6

Warfarin Dosing
Should Labs Offer Pharmacogenetic Testing?
By Charles Eby, MD


Pharmacogenetic-based dosing of warfarin is a frequently cited example of personalized medicine, the concept that information about a patient’s genotype could be used to further tailor medical care to an individual’s needs. Warfarin acts to keep blood from clotting, but one of the side effects of the drug is an increased risk for bleeding. Too much can lead to serious bleeding events, while too little may allow dangerous blood clots to form. Therefore, correct dosing is critical to successful patient management. Researchers have identified common single nucleotide polymorphisms (SNPs) in genes that affect warfarin metabolism and activity and could potentially be used to predict an individual’s response to warfarin, thereby avoiding over- or under-dosing. In fact, several algorithms have been developed to improve initial dosing accuracy based on the patient’s genotype.

Following these developments, in August 2007, FDA revised the labeling for warfarin suggesting that clinicians consider genetic testing before administering the drug. With the recent FDA clearance of several methods and instruments for identifying these warfarin-related SNPs, labs now have the opportunity to add this pharmacogenetic-based (PGx) test for warfarin dosing to their menus.

However, organizations representing cardiovascular physicians and clinical geneticists have not endorsed PGx warfarin dosing. Furthermore, clinicians do not appear to be clamoring for labs to provide warfarin genetic testing. Should laboratorians be advocating for personalized warfarin dosing, or should we be patient and wait for stronger support from ongoing clinical investigations? This article describes the current status of PGx warfarin dosing.

The Challenges of Prescribing Warfarin

Physicians prescribe warfarin, a member of the coumarin family of oral anticoagulants, to prevent and treat venous and arterial thromboembolic disorders. By inhibiting reduction of oxidized vitamin K, warfarin prevents post-translational γ-carboxylation of selective glutamic acid residues on coagulation factors X, IX, VII, and II (prothrombin), leading to an acquired coagulopathy (Figure 1).

Figure 1
Warfarin Metabolism and Anticoagulation Activity

After oral absorption, warfarin is transported to the liver where CYP1A1, CYP1A2, and CYP3A4 metabolize the R-enantiomer and CYP2C9 metabolizes the more potent S- enantiomer. Warfarin inhibits vitamin K reductase, which is synthesized by VKORC1. By impairing the regeneration of the reduced form of vitamin K, R- and S-warfarin interfere with the vitamin-K–dependent carboxylation of clotting factors prothrombin (II), VII, IX, and X.

Reprinted with permission of the authors from Thrombosis Journal 2008; 6:7.

Due to warfarin’s narrow therapeutic range and unpredictably wide maintenance dose, patients may require from <1mg/day to >15mg/day. To assess the drug’s effectiveness, physicians monitor patients’ anticoagulant status with the International Normalized Ratio (INR), a value derived from the measured prothrombin time. The current “trial-and-error” approach to initiating warfarin therapy is accompanied by frequent INR measurements, dose adjustments, and bleeding complications (1). Once a patient’s therapeutic dose is determined empirically, less frequent INR monitoring is necessary, and the risk of major bleeding is diminished, although persistent (1).

Genetic Polymorphisms Affecting Therapeutic Warfarin Dose

Clinical and demographic variables including age, body mass, and certain medications, also affect warfarin dosing. However, clinical-based algorithms that incorporate these characteristics only explain ~20% of the variability in maintenance dose (2). Recent biochemical and molecular discoveries have further advanced the understanding of warfarin pharmacology, leading to major advances in predicting therapeutic warfarin dose.

Warfarin consists of equal amounts of S- and R-enantiomers. S-warfarin accounts for ~80% of warfarin’s anticoagulant activity, and it undergoes hydroxylation by cytochrome P450 2C9 (CYP2C9). Two SNPs in CYP2C9 affect warfarin metabolism: *2 allele is 420 C>T in exon 3, substituting cysteine for arginine at position 144, and *3 allele is 1075 A>C in exon 7, substituting leucine for isoleucine at position 359. CYP2C9 *2 and *3 reduce warfarin metabolism moderately and severely, respectively, compared to the common, wild-type allele, CYP2C9*1.

Following the discovery of the gene for vitamin K reductase activity (VKORC1) in 2004, investigators identified SNPs associated with warfarin maintenance dose that define two common haplotypes: A, associated with a lower maintenance dose; and B, associated with a higher dose (3). A SNP in the VKORC1 promoter region substituting alanine (A) for guanine (G) at -1639, is associated with lower mRNA expression, decreased VKORC1 enzyme, and lower maintenance warfarin dose (4). The prevalence of these SNPs varies among different ethic and racial populations (Table 1), partially explaining the hierarchy in average warfarin dose: African American>Caucasian>Asian.

Table 1
Variant Allele Frequencies for VKORC1 and CYP2C9 in Different Populations


European American

African American














Adapted from Marsh S. et al. Thromb Haemost 2006;4:473.

Combining patients’ clinical, demographic, and genetic information for SNPs CYP2C9*2 and *3, and VKORC1 -1639G>A, or other VKORC1 SNPs in high linkage disequilibrium with -1639 such as 1173C>T, accounts for ~45-60% of the variation in therapeutic warfarin doses in predominantly Caucasian and Asian populations (2, 5). At least a dozen pharmacogenetic dosing algorithms have been published, but few have been prospectively validated. In fact, when several models were used to predict warfarin doses in a cohort of African Americans monitored in an anticoagulation clinic, they performed poorly (6), suggesting additional genetic variations affecting warfarin dosing in African Americans await discovery.

At the same time, a growing number of prospective studies have demonstrated the feasibility of using pharmacogenetic algorithms to guide initiation of warfarin therapy (2). Moreover, one study conducted at my institution demonstrated that dosing accuracy is improved by using a dose-revision algorithm that includes INR results obtained following the third or fourth warfarin dose (7). Based on our experience, my colleague, Brian Gage, MD, associate professor of medicine at Washington University in St. Louis, now curates a website that calculates initial and revised predicted warfarin doses for patients using either clinical or pharmacogenetic algorithms (see box, Estimating Warfarin Dose, below).

Estimating Warfarin Dose is a free web-site to help doctors and other clinicians begin warfarin therapy by estimating the therapeutic dose in patients new to warfarin. This site is supported by the Barnes-Jewish Hospital at Washington University Medical Center, the NIH, and donations. Estimates are based on clinical factors and (when available) genotypes of two genes: cytochrome P450 2C9 (CYP2C9) and vitamin K epoxide reductase (VKORC1). 

Recommendations on the website are based on data from over 1,000 patients. Once information is entered, the initial estimate of therapeutic dose explains 53% of the variability in a warfarin dose. If you return to the website and enter an INR value after 3 and/or 4 warfarin doses, the dose refinement is even more accurate. 

Dosing: Clinical Utility Uncertain

However, improving the accuracy of initial warfarin dosing must produce convincing improvement in clinically meaningful outcomes in order to justify the additional cost of genetic testing. Inadequate evidence of clinical efficacy is currently the major impediment to acceptance of warfarin pharmacogenetics. To date, only three small, prospective, randomized trials comparing different PGx-based dosing to different empiric nomogram-based dosing strategies have been published (8–10) (Table 2).

Lab-defined endpoints, including time to therapeutic INR and time within therapeutic range, were significantly better in the pharmacogenetic arm of one trial (9), but major bleeding complications were rare, 11 out of 435 subjects, and were not significantly different between dosing arms in any of the studies. A recent cost-benefit analysis, based on the currently available data, and an estimated charge of $400 to perform CYP2C9 *2/*3 and VKORC1 genotyping, recommended against routine genotyping due to an estimated cost of >$170,000 per quality adjusted life-years (11).

The COAG Trial

In order to address the uncertainty of the clinical efficacy of PGx-based initiation of warfarin therapy, the National Heart Lung and Blood Institute announced in February that it is sponsoring a prospective, multicenter, double-blinded, randomized study (Coag website). The Clarification of Optimal Anticoagulation through Genetics (COAG) Trial will randomize eligible patients to clinical plus genetic-based or clinical only- based algorithms for the initial week of warfarin dosing followed by subsequent protocol-driven dose adjustments based on INRs (Figure 2). The primary outcome that researchers will measure is the time within therapeutic INR range during the first four weeks of anticoagulation therapy. This will be a cost-effective and clinically meaningful target since improved INR control is a valid surrogate for less common major bleeding and thrombotic complications (12).

Eligible patients will be randomized to initial dosing based on either a clinical algorithm or pharmacogenetic+clinical algorithm, and a second algorithm-based dose revision after an INR following the third or fourth dose. An INR nomogram is used to make subsequent adjustment. Patients and clinicians are blinded to warfarin doses for days 1 to 30.

Outside the U.S., investigators are also planning similar prospective, randomized trials. The availability of results from multiple studies will provide opportunities to pool outcomes and better assess the impact of warfarin PGx on major hemostasis complications.

Warfarin Genotyping Methods

Multiple methods are available to perform warfarin PGx testing, including direct sequencing, PCR/RFLP, real time PCR, and several commercial platforms. Studies evaluating different molecular diagnostic methods uniformly confirm excellent accuracy genotyping SNPs CYP2C9*2/*3 and VKORC1 SNPs (13). However, there are differences in turn around time (TAT), technical complexity, instrument size, and extended CYP2C9 and VKORC1 SNP menus, which may influence a lab’s platform selection (13). Additional factors include other diagnostic genetic tests available on the platform, cost of the instrument and consumables, and reimbursement rates. Presently, five manufactures have FDA-cleared medical devices for warfarin testing: Nanosphere (Verigene), Autogenomics (INFINITI), Osmetech (eSENSOR), ParagonDX reagents and primers with the Cephied SmartCycler, and TrimGen reagents and primers with Roche LightCycler.

Discovery of New Polymorphisms

Researchers continue to search for additional genetic polymorphisms to account for the ~50% of warfarin dose variability that is currently unexplained. Using a drug metabolism and transport gene chip, Caldwell and colleagues recently identified a C>T SNP in the CYP4F2 gene (rs2108622) associated with ~1mg/day increase in warfarin dose among homozygous Caucasian subjects, yet it only accounts for ~2% of dosing variability due to the low frequency of the variant (T) allele (14). Italian investigators have confirmed these findings (15), and recent experimental data indicate CYP4F2 V433M C>T delays vitamin K metabolism.

While it is likely that future discoveries of genetic variants affecting warfarin dose will have small, incremental affects, the costs of incorporating additional SNPs into genotypying platforms will be small. Existing algorithms can be easily adjusted to accommodate new information and further improve dosing accuracy.

Warfarin PGx Testing Programs

As mentioned at the beginning of this article, demand for warfarin PGx testing has not been strong, despite the update to the drug’s label. Several ongoing projects are exploring warfarin PGx testing services. For example, ParagonDX has started a project in the Research Triangle region of North Carolina to provide local primary care physicians with warfarin PGx results within 24 hours accompanied by qualitative dosing advice.

Nationally, Medco Life Insurance Company is conducting a study to determine if warfarin PGx testing will reduce hospitalizations for bleeding or thrombotic complications. The insurer is offering warfarin PGx testing to patients when they begin taking warfarin for the first time. Medco plans to monitor patients’ hospitalizations, INR results, warfarin dose changes, and healthcare resource utilization. Outcomes will be compared to historical controls and concurrent patients who do not participate in the study. Although the project is not a randomized, prospective, efficacy trial designed to optimally control for potential biases, it may provide valuable, real-world information about personalized anticoagulation therapy.

News Update
Warfarin PGx Testing Fails to Pass CMS Coverage Hurdle

In a proposed decision memo, the Centers for Medicare and Medicaid Services (CMS) announced in early May that it would not pay for genetic tests to help guide warfarin dosing for Medicare recipients, noting that “available evidence does not demonstrate that pharmacogenomic testing to predict warfarin responsiveness improves health outcomes in Medicare beneficiaries.” The agency did, however, leave open an option for payment under its Coverage with Evidence Development authority as a means of gathering further information on the topic. Reimbursement would be limited to testing provided to Medicare beneficiaries who are candidates for anticoagulation therapy with warfarin and part of a prospective, randomized, controlled clinical study designed to show that pharmacogenomics-guided dosing strategies improve health outcomes over standard dosing methods.

CMS’s action concludes a 9-month national coverage analysis initiated by the agency to seek public and expert opinion on the matter. AACC, along with a number of other professional societies, supported extending coverage for the testing, whereas others stated that it was premature. A copy of the CMS decision, along with a synopsis of public input, is available on the agency’s website.

Ready or Not?

Many stakeholders are participating in the evolution of personalized medicine, and in particular in PGx-based warfarin dosing. FDA’s mandate to ensure the public’s safety motivated a recent revision of Coumadin package insert to include information about CYP2C9 and VKORC1 SNPs that could warrant lower initial doses for patients with those genotypes.

But at this point, clinical geneticists and major medical societies have not endorsed routine genotyping due to concerns over lack of evidence supporting clinical utility of genotyping and the associated costs of testing. Clinical efficacy studies, such as the COAG Trial, are expensive, lengthy, and, despite careful design and execution, may not support definitive conclusions upon completion. If results of novel approaches to assessing clinical utility, such as the Medco project and innovative statistical methods, withstand critical review, they may provide efficient alternative methods for evaluating future molecular-based personalized medicine initiatives.

But incumbent upon a lab director’s decision to offer warfarin PGx testing is the need for a plan to rapidly provide results and to inform physicians how to use this new molecular information to modify a patient’s warfarin dose. For the time being, there is no mandate for community or academic hospitals to offer patients starting warfarin therapy PGx testing with a turn around time of 1–3 days.

Local circumstances will vary, and some local or regional healthcare systems may elect to introduce PGx-based warfarin dosing based on currently available clinical science. However, if PGx-based initiation of warfarin therapy is to be the “poster child” for personalized medicine, then we should wait for more compelling evidence of its clinical utility before recognizing it as the standard of care.


  1. Linkins L, Choi P, Douketis J. Clinical impact of bleeding in patients taking oral anticoagulant therapy for venous thromboembolism: a meta analysis. Ann Intern Med 2003;139:893–900.
  2. Gage BF, Eby C, Johnson JA, Deych E, et al. Use of pharmacogenetic and clinical factors to predict the therapeutic dose of warfarin. Clin Pharmacol and Ther 2008;84:326–31.
  3. Rieder MJ, Reiner AP, Gage BF, Nickerson DA, et al. Effect of VKORC1 haplotypes on transcriptional regulation and warfarin dose. N Engl J Med 2005;352:2285–93.
  4. Yuan HY, Chen JJ, Lee MT, Wung JC, et al. A novel functional VKORC1 promoter polymorphism is associated with inter-individual and inter-ethnic differences in warfarin sensitivity. Hum Mol Genet 2005;14:1745–51.
  5. IWPC. Estimation of the warfarin dose with clinical and pharmacogenetic data. N Engl J Med 2009;360:753–764.
  6. Schelleman H, Chen J, Chen Z, Christie J, et al. Dosing algorithms to predict warfarin maintenance dose in Caucasians and African Americans. Clin Pharmacol and Ther 2008;84:332–9.
  7. Millican E, Lenzini P, Milligan P, Grosso L, et al. Genetic-based dosing in orthopaedic patients beginning warfarin therapy. Blood 2007;110:1511–5.
  8. Hillman MA, Wilke RA, Yale SH, Vidaillet HJ, et al. A prospective, randomized pilot trial of model-based warfarin dose initiation using CYP2C9 genotype and clinical data. Clin Med Res 2005;3:137–45.
  9. Caraco Y, Blotnick S, Muszkat M. CYP2C9 genotype-guided warfarin prescribing enhances the efficacy and safety of anticoagulation: a prospective randomized controlled study. Clin Pharmacol and Ther 2008;83:460–70.
  10. Anderson JL, Horne BD, Stevens SM, Grove AS, et al. Randomized trial of genotype-guided versus standard warfarin dosing in patients initiating oral anticoagulation. Circulation 2007;116:2563–70.
  11. Eckman MH, Rosand J, Greenberg SM, Gage B. Cost-effectiveness of using pharmacogenetic information in warfarin dosing for patients with nonvalvular atrial fibrillation. Ann Intern Med 2008;150:73–83.
  12. Chiquette E, Amato MG, Bussey HI. Comparison of an anticoagulation clinic with usual medical care: anticoagulation control, patient outcomes, and health care cost. Arch Intern Med 1998;158:1641–7.
  13. King C, Porche-Sorbet R, Gage B, Ridker P, et al. Performance of commercial platforms for rapid genotyping of polymorphisms affecting warfarin dose. Am J Clin Path 2008;129:876–883.
  14. Caldwell MD, Awad T, Johnson JA, Gage BF, et al. CYP4F2 genetic variant alters required warfarin dose. Blood 2008;111:4106–12.
  15. Borgiani P, Ciccacci C, Forte V, Sirianni E, et al. CYP4F2 genetic variant (rs218622) significantly contributes to warfarin dosing variability in the Italian population. Pharmacogenomics 2009;10:261–6.

Charles S. Eby, MD, is medical director of the Hematology Laboratory at Barnes-Jewish Hospital, and associate professor of pathology, immunology, and medicine at Washington University School of Medicine in St. Louis, Mo. Dr. Eby also directs the Hemostasis Research Laboratory in the Division of Laboratory and Genomic Medicine.


Disclosure: Dr. Eby receives grant/research support and honoraria/expenses from Osmetech, and honoraria/expenses from Autogenomics.