American Association for Clinical Chemistry
Better health through laboratory medicine
Is Warfarin Pharmacogenomic Testing Ready for Prime Time?

CLN Banner Logo

Is Warfarin Pharmacogenomic Testing Ready for Prime Time?
Today’s Debate to Focus on Implementation Issues
By Deborah Levenson

Cover Image

In recent years, personalized medicine has become the subject of much hype in newspapers and magazines, although it has yet to become a part of routine healthcare. But with an August 2007 update to the warfarin package insert, FDA may have nudged patient care closer to an era in which genetic information helps tailor care for millions. The updated insert makes clinicians aware that patients with certain variants of the CYP2C9 and VKORC1 genes probably need lower initial doses. But despite FDA’s suggestion that warfarin pharmacogenomic assays might aid dosing, use of such tests remains controversial.

While most experts on pharmacogenomic testing say it holds great potential for patient care, they disagree about whether now is the time to implement it widely. “Everyone says personalized medicine is the wave of the future, but the question is really the rate at which it is rolled out,” said Michael Hallworth, FRCPath, Consultant Clinical Scientist at Royal Shrewsbury Hospital in Shropshire, U.K. “Warfarin genomic testing is like the poster child for personalized medicine because it addresses diseases that affect millions. It’s a paradigm for widespread adoption of genomic testing.” Hallworth is moderator of a debate titled “Warfarin Pharmacogenomic Testing: Now Ready for Prime Time?” scheduled for 10:30 a.m. in room 150AB of the Walter E. Washington Convention Center.

The debate format represents a new twist to the standard sypmposia presented at the AACC Annual Meeting. Everyone who attends the debate will get a chance to weigh in on the issue, Hallworth said. The audience will vote both before and after the debate to determine if the speakers were able to sway any opinions.

Proponents Emphasize Potential

Shiew-Mei Huang, PhD, Deputy Director of FDA’s Office of Clinical Pharmacology, will propose a motion in favor of more widespread testing, reflecting FDA’s stance on the issue. FDA’s label language, which indirectly suggests that physicians consider pharmacogenomic testing to determine initial warfarin dose, is based on data showing that doing so keeps warfarin patients’ INRs in an acceptable range of 2–3. In the absence of definitive data from randomized, controlled trials that look at harder clinical outcomes like excessive bleeding events, FDA and virtually all practitioners have used INR as a surrogate outcome, she explained.

“Recent studies have showed poor patient outcomes that we need to improve upon,” Huang added, pointing to a 1,015-patient study by a research team led by Brian Gage, MD, of Washington University in St. Louis. The investigators found that a dosing algorithm based on clinical factors explained 17–22% of dose variability, while another algorithm that incorporated information about CYP2C9 and VKORC1 explained 53–54% of dose variability (Clinical Pharmacology Therapeutics 2008 doi: 10.38/clpt.2008.10).

Huang also plans to discuss data that support expected cost effectiveness of widespread testing. One recent report estimates that formally integrating genetic tests into routine warfarin therapy could allow American warfarin users to avoid between 4,500 and 22,000 serious bleeding events annually, she pointed out (Personalized Medicine 2008; 5: 279–284).

Meanwhile, various studies also show that particular variants of the CYP2C9 and VKORC1 genes affect clearance and dose requirements. “There’s a consistent relationship. Even the worst critics of the test acknowledge this consistent and known relationship,” added Mark Linder, PhD, who is also speaking in favor of more widespread testing. Linder is Associate Director, Chemistry and Toxicology and Associate Director of the Pharmacogenetics Diagnostic Laboratory at University of Louisville Hospital in Louisville, Ky. and Senior Vice President of PGXL Laboratories, a private company that provides warfarin testing services.

Wanted: Harder Outcomes

Noting that the expected benefits—improved patient safety in the form of fewer bleeding and clotting incidents—have been detailed in various reports, opponents to more widespread use of pharmacogenomic warfarin testing are waiting for hard evidence of these benefits. “There’s good evidence that two SNPs on CYP2C9 and several on VKORC1 affect warfarin metabolism. This we know well,” said Amir Jaffer, MD, Associate Professor of Medicine at the University of Miami’s Leonard M. Miller School of Medicine and Service Chief of Medicine at Miami Hospital. “What we don’t know well is if testing translates into evidence of improved clinical outcomes.” Jaffer is speaking in opposition to more widespread testing.

Both Jaffer and Charles Eby, MD, Associate Professor in the Department of Pathology and Immunology at Washington University St. Louis School of Medicine, Mo., want to see data that show warfarin pharmacogenomic testing actually reduces incidence of bleeding and mortality. Eby, who will also speak against widespread testing, considers such evidence the only true measure of the tests’ worth. “Not a single study shows a clinical benefit to testing,” he pointed out.

While Huang maintained that a greater percentage of time within the target therapeutic target range and fewer patients with INRs >4 are indeed clinical benefits, Eby noted that only two trials that have used therapeutic INR as endpoints were randomized controlled studies. Those two studies yielded conflicting results. The first involved 206 patients and did not achieve a reduction in out-of-range INRs (Circulation 2007;116: 2563–2570), but the other, with 201 patients, found incorporating genetic information into a dosing algorithm increased patients’ time in a therapeutic INR range (Clinical Pharmacology & Therapeutics 2008; 83: 460–470). However, these studies’ control arms employed different strategies for using genetic information for warfarin dosing, Eby noted.

Paragon Dx Offers Warfarin Genotyping

Whatever their stance on warfarin pharmacogenomic testing, visitors to the Paragon Dx booth (3729) in the exhibit hall may learn their CYP2C9 and VKORC1 genotypes at no cost. While Paragon Dx recently received FDA clearance for its Rapid Genotyping kit , the company also markets ASR reagents for lab-developed tests and will be using them to genotype meeting attendees onsite. According to Paragon Dx CEO Michael Murphy, the company hopes to show lab directors that it’s possible to run these warfarin pharmacogenomic tests in a space as small as the company’s exhibit booth. While labs can genotype samples with Paragon Dx reagents in 60 minutes, company representatives at the meeting will be batching tests because they expect to do several hundred of them.

Other Considerations

Data aside, Eby raised concerns about how labs should report genetic information and how clinicians would use it. “Genetic information by itself has minimal value unless you put it into a dosing algorithm someone can carry around in their head,” he explained, “So should we provide an interpretation by saying that a particular genotype is associated with a decreased dose? Should we provide an algorithm?”

Whatever labs report to physicians, “it’s more than the usual accurate, precise results,” Eby noted. While most hospital labs have experience providing specific interpretive comments for for certain tests, like D-dimer, “we’ve never had to customize results for each patient,” he pointed out. Eby also worries about labs communicating genetic information to an audience much wider than the oncology subspecialties that now receive it. Lab staff may be talking to clinicians who lack the education and judgment to take into consideration several other necessary factors— like age, body mass, sex, and use of other drugs—into dosing decisions, according to both Eby and Jaffer.

Linder doesn’t share these fears. He thinks opponents of wider testing are applying a double standard to the issue. They are worried about physicians learning to use the test through a process of trial and error, Linder said, but such processes are inherent to the practice of medicine in general and INR—the current standard of care—in particular. “People are applying a new set of rules to this test,” he said.

Jaffer also worries that genetic data would be foisted upon physicians who are unprepared to use it safely, responsibly, and soon enough to yield benefits. One of his concerns is that existing algorithms might be too complicated for most physicians to use in routine practice. “If we make dosing more difficult, physicians might not start therapy as soon. That would have a counter effect on how patients do,” he asserted.

But complex algorithms are nothing new, according to Huang. “Physicians adjust dosage of drugs used in renal failure with algorithms that are far more complicated,” she maintained.

Linder disagrees with Jaffer’s expectation that most physicians who prescribe warfarin will use the pharmacogenomic tests. “The test isn’t for all physicians, but only those who understand it. It’s available and useful now to those with specific knowledge about how to use and interpret it,” he explained.

Will Huang, Eby, Linder, and Jaffer sway any opinions about warfarin pharmacogenomic testing? The best way for meeting attendees to find out is by attending what promises to be an exciting debate.