Genetic Tests Not Useful for Managing Warfarin Dosing, According to New Studies

At an August 2007 press conference, with much fanfare, the Food and Drug Administration (FDA) announced that the drug label for warfarin (COUMADIN, JANTOVEN) — one of the oldest and most widely prescribed blood thinners (anticoagulants) — was being updated. The new label would indicate that some patients needing blood thinners were likely to respond differently to warfarin if they carried certain variations of two genes related to the breakdown and action of the drug.[1],[2]

Although stopping short of explicitly recommending that patients who need warfarin undergo genetic testing to screen for these gene variations, FDA officials strongly implied that such testing could be useful in determining the initial warfarin dose. Indeed, touting genetic testing to guide warfarin dosing as a great advancement in the realm of personalized medicine, a senior FDA official noted:[3]

Today is a significant event for those who foresee a day when medical care will be tailored to fit the unique genetic makeup of every single one of us. ... This means that personalized medicine is no longer an abstract concept but has moved into the mainstream where it is recognized as a factor in a product used by millions of Americans every day.

However, the results of three large clinical trials appearing in The New England Journal of Medicine (NEJM) on Dec. 12, 2013, demonstrate that the FDA’s enthusiasm about the promise of genetic testing in management of warfarin dosing was premature and overstated. The collective data from these trials reveal that such genetic testing offers little meaningful clinical benefit to patients being started on warfarin and is not worth the additional cost.

How does warfarin work?

Warfarin is a drug that reduces the blood’s ability to clot (coagulate) and prevents blood clots from forming in veins, arteries and the heart. It does so by blocking the liver’s production of certain proteins called clotting factors, which form the building blocks of a clot.

The drug is commonly prescribed to people who have developed abnormal blood clots (for example, a clot in the large veins of the legs, called deep venous thrombosis, or the lung, called a pulmonary embolus) or who are at high risk of forming abnormal clots (for example, people with atrial fibrillation, a common irregular heart rhythm).

The dose of warfarin must be regularly monitored and adjusted in order to achieve the appropriate balance between the drug’s lifesaving anti-clotting benefits and its risk of serious, potentially life-threatening bleeding. This monitoring requires periodic blood tests to measure how quickly a patient’s blood clots.

Multiple factors influence how much warfarin an individual patient needs to achieve the desired anti-clotting effect. Those factors include age, weight, diet, health status, use of other medicines and genetic makeup. One of the most important factors is dietary intake of vitamin K, which plays a key role in the liver’s ability to make certain clotting factors. Vitamin K is found in high levels in green leafy vegetables, broccoli, Brussels sprouts, asparagus and many other foods. After starting warfarin, patients need to maintain a stable intake of vitamin K to avoid large fluctuations in the level of the drug’s anti-clotting effect.

Many other drugs and dietary supplements — including vitamin products containing vitamin K — interact with warfarin and can increase or decrease the drug’s effects. Patients with significant liver disease also generally require lower warfarin doses to achieve the desired effect because their diseased livers make less of the clotting factors.

Genetics and warfarin

Patients with variations in two genes respond very differently to warfarin. One gene, known as CYP2C9, is used by liver cells to make an enzyme that breaks down warfarin, eliminating the drug from the body.[4] Some people have a variant or mutation of this gene that causes the enzyme to break down warfarin more slowly than the usual version of the enzyme. As a result, patients with this gene variant need less warfarin to achieve the same anti-clotting effect.

Variants of the second gene, known as VKORC1, cause patients to have increased sensitivity to warfarin, which likewise results in a need for lower warfarin dosing.[5],[6]

Most major laboratories across the country, as well as many smaller laboratories at major medical centers, offer genetic testing for the two genes together that affect warfarin dosing.[7] According to the FDA, in 2007, prices for such testing ranged from $125 to $500.[8] In September 2007, the FDA cleared the first commercially marketed genetic tests for assessing warfarin responsiveness. Following the FDA’s lead, the manufacturer of the new test, Nanosphere Inc., stated that information on an individual’s ability to metabolize warfarin is “critical to determining safe and appropriate dosing” of the drug.[9]

When the FDA announced the label change for warfarin in 2007, it was not the first time that the agency had approved a drug label with information on the use of genetic testing to guide drug dosing.[10] (The first time involved the cancer drug irinotecan.[11]) However, it did mark the first time that such genetic information was included in the label of a widely used drug.[12]

In approving the inclusion of genetic testing information in the drug label for warfarin, the agency apparently assumed that knowing whether a patient had one or more of the noted gene variants of the CYP2C9 or VKORC1 genes that predispose a patient to needing less warfarin would help physicians select an initial dose when starting patients on the drug. However, no large, well-designed clinical trials had been conducted to test this assumption prior to the completion of the recent NEJM studies. Moreover, such genetic testing would not alter the need for routine periodic blood testing to monitor and adjust the warfarin dose.

In August 2009, the Centers for Medicare and Medicaid Services (CMS) concluded the available evidence failed to demonstrate that genetic testing for warfarin responsiveness improved health outcomes in Medicare beneficiaries, and therefore, the tests were not reasonable or necessary.[13] Thus, CMS decided to cover such genetic testing only in Medicare patients who were being initiated on warfarin and were enrolled in a randomized, controlled trial (RCT) assessing such testing, such as those discussed below.

Three NEJM studies

The first — and largest — of the three trials published in the NEJM in December 2013 was an RCT funded by the National Institutes of Health and conducted at 18 medical centers in the U.S.[14] The study was titled the Clarification of Optimal Anticoagulation through Genetics (COAG) trial.

Between September 2009 and April 2013, researchers randomly assigned 1,015 patients needing to start warfarin to one of two groups. In one group (the genetic-testing group, 514 patients), dosing of warfarin during the first five days of treatment was determined based on baseline patient characteristics and certain clinical criteria (including age, race, weight, height, current smoking status, whether the patient was taking the heart drug amiodarone [CORDARONE, NEXTERONE, PACERONE], the desired level of anticoagulation and the reason for starting anticoagulation treatment) plus results of genetic testing for variants of the two genes that affect warfarin responsiveness. The second group (the control group, 501 patients) had initial warfarin dosing based only on baseline patient characteristics.

The COAG trial researchers found no difference between the two groups in the primary outcome of interest: the average amount of time patients spent during the first four weeks within the desired anticoagulation range based on usual blood tests (the result was 45 percent for both groups). There also were no significant differences between the two groups in the number of patients whose blood tests showed an overdosing of warfarin, major bleeding or formation of abnormal blood clots.

The second NEJM study was a smaller RCT conducted at three medical centers in the U.K. and two in Sweden: the European Pharmacogenetics of Anticoagulation Therapy group (EU-PACT) warfarin trial.[15] Between January 2011 and April 2013, researchers randomly assigned 455 patients needing to start warfarin to one of two groups. In one group (the genetic-testing group, 227 patients), dosing of warfarin during the first five days of treatment was determined based on baseline patient characteristics (including age, weight, height, amiodarone use and the desired level of anticoagulation, but, unlike the COAG trial, not race, smoking status or the reason for starting anticoagulation treatment) plus results of genetic testing for warfarin responsiveness. The control group (228 patients), in contrast to that of the COAG trial, had initial warfarin dosing based only on patient age.

Unlike in the COAG trial, the EU-PACT warfarin trial researchers did find a small statistically significant difference in their primary outcome: The average amount of time patients spent during the first 12 weeks within the desired anticoagulation range was greater in the genetic-testing group than in the control group (67 percent versus 60 percent, respectively). Also, fewer patients in the genetic-testing group had at least one blood test showing an overdosing of warfarin compared with control group patients (27 percent versus 37 percent).

However, most importantly, there were no differences between the two groups in major bleeding, bleeding episodes or formation of abnormal blood clots, which are the primary undesirable clinical outcomes of incorrect dosing.

One plausible reason for the difference in results between this trial and the COAG trial was the difference in the number of factors used to determine initial warfarin dosing in the control groups.

The third NEJM study combined data from two additional EU-PACT RCTs conducted at the same time at multiple medical centers in the Netherlands and Greece.[16] The two RCTs together enrolled 548 patients being started on blood thinners that are similar to warfarin but not available in the U.S.

The design of these EU-PACT trials was nearly identical to that of the U.S. COAG trial. In the genetic-testing group (273 patients), dosing of the blood thinner during the first five to seven days of treatment was determined based on baseline patient characteristics (including age, sex, weight, height and amiodarone use, but, unlike the COAG trial, not race, smoking status or the reason for starting anticoagulation treatment). In the control group (275 patients), initial blood thinner dosing was based only on the baseline patient characteristics.

The data from the third NEJM study showed no statistically significant differences between the two groups in the average amount of time patients spent within the desired anticoagulation range during the first 12 weeks, the study’s primary outcome (the results were 62 percent for the genetic-testing group and 60 percent for the control group). There also were no significant differences between the two groups in the number of patients who had blood tests showing an overdosing of warfarin or in the rates of adverse events, bleeding or formation of abnormal blood clots.

Conclusions

The new data from the three similar RCTs discussed here provide strong evidence that genetic testing warfarin responsiveness does not provide any meaningful clinical benefit to patients being started on warfarin. Initial dosing of warfarin, therefore, should be based on consideration of baseline patient characteristics alone, such as age, weight, height, race and smoking status, among others.

What You Can Do

If you develop a condition that requires initiation of the blood thinner warfarin and you are offered genetic testing to assess your responsiveness to the drug, you should decline the testing.

References

[1] Food and Drug Administration. Transcript of FDA press conference on warfarin. August 16, 2007. http://www.fda.gov/downloads/newsevents/newsroom/mediatranscripts/ucm123583.pdf. Accessed May 8, 2014.

[2] Brown D. For first time, FDA recommends gene testing. The Washington Post. August 17, 2007.

[3] Food and Drug Administration. Transcript of FDA press conference on warfarin. August 16, 2007. http://www.fda.gov/downloads/newsevents/newsroom/mediatranscripts/ucm123583.pdf. Accessed May 8, 2014.

[4] Bristol-Myers Squibb. Drug label for COUMADIN. August 2007. http://www.accessdata.fda.gov/drugsatfda_docs/label/2007/009218s105lblv2.pdf. Accessed May 8, 2014.

[5] Ibid.

[6] Wadelius M, Chen LY, Downes K, et al. Common VKORC1 and GGCX polymorphisms associated with warfarin dose. Pharmacogenomics J. 2005;5:262-270.

[7] Food and Drug Administration. Transcript of FDA press conference on warfarin. August 16, 2007. http://www.fda.gov/downloads/newsevents/newsroom/mediatranscripts/ucm123583.pdf. Accessed May 8, 2014.

[8] Ibid.

[9] Nanosphere. Press release: Nanosphere announces first FDA cleared genetic test for warfarin sensitivity and nanotechnology-based molecular diagnostics platform. September 18, 2007. http://ir.nanosphere.us/phoenix.zhtml?c=214748&p=irol-newsArticle&ID=1075936&highlight. Accessed May 8, 2014.

[10] Food and Drug Administration. Transcript of FDA press conference on warfarin. August 16, 2007. http://www.fda.gov/downloads/newsevents/newsroom/mediatranscripts/ucm123583.pdf. Accessed May 8, 2014.

[11] Ibid.

[12] Ibid.

[13] Centers for Medicare & Medicaid Services. National Coverage Determination (NCD) for pharmacogenomics testing for warfarin responsiveness. Effective August 3, 2009. http://www.cms.gov/medicare-coverage-database/details/ncd-details.aspx?NCDId=333&bc=AgAAQAAAAAAA&ncdver=1. Accessed May 10, 2014.

[14] Kimmel SE, French B, Kasner SE, et al. A pharmacogenetic versus a clinical algorithm for warfarin dosing. N Engl J Med. 2013;369(24):2283-2293.

[15] Pirmohamed M, Burnside G, Eriksson N, et al. A randomized trial of genotype-guided dosing of warfarin. N Engl J Med. 2013;369(24):2294-2303.

[16] Verhoef TI, Ragia G, de Boer A, et al. ArRandomized trial of genotype-guided dosing of acenocoumarol and phenprocoumon. N Engl J Med. 2013;369(24):2304-2312.