With new data linking genetics and patient response, who will step up to the plate with a diagnostic capable of guiding dosages for the commonly prescribed — and occasionally dangerous — blood thinner warfarin?
While variations in the CYP450 2C9 enzyme have long been correlated with warfarin response, 2C9-based tests for determining the proper dosage of the anticoagulant have not caught on. But new data linking another gene, VKORC1, to warfarin response has caught the interest of several diagnostic developers, as well as the FDA.
"Initially we could describe anywhere from 20 to 30 percent of the variability with clinical factors and [CYP450] 2C9 variants, and now you add about another 20 to 25 percent of the variability [that] we can explain," with information about a patient's genotype at the VKORC1 locus, Mark Rieder, lead author of a new study on the genetics of warfarin response, told Pharmacogenomics Reporter.
The retrospective research, which genotyped patients on a stable warfarin dose for the VKORC1 gene, appears in the June 2 issue of The New England Journal of Medicine. The study's sample population is the same as that used for a previous study in the May issue of Clinical Pharmacology and Therapeutics headed by Alan Rettie, a University of Washington in Seattle professor of medicinal chemistry, that studied the genetic contribution of CYP450 2C9 to response to the drug.
Although warfarin has many of the right qualities to be a trailblazing pharmacogenomic example — it's a high-selling drug with potentially dangerous adverse events, but genetics govern much of the response — the field of diagnostics related to the drug has been anemic so far. Clinicians appear to need something quick and close at hand to replace the existing method — a prothrombin coagulation test. Also, genetic diagnostics based on 2C9 may not have information for clinicians to feel comfortable using them to aid dosing. It's possible that the new data pointing to VKORC1 could lead to a more effective test.
"We started offering this testing back in 2000," said Christine Ashcraft, a spokesperson for the genetic-testing company Genelex. "We did originally think that that would be a good market to go after; we went after anticoagulation clinics, and there really wasn't very much interest in the 2C9 for warfarin or Coumadin [the Bristol-Myers Squibb trade name for the drug]," Ashcraft said.
The anticoagulation clinics are already comfortable with the common prothrombin blood test for coagulation, said Ashcraft. The test evaluates a patient blood sample for coagulation after an initial dose has been administered. It is a simple test that works quickly, but "[poor responders] take about three months to reach a stable dose, and they have a much worse chance of adverse reactions," under the current regime, said Howard Coleman, Genelex CEO. "Clinical application and utility [are] not as obvious as if it were a psychiatric drug, for example, because when they bring in a patient needing anticoagulation therapy, they [can already] have an answer within an hour," said Coleman.
A genetic test "wouldn't overtake [conventional methods], it would just be a complement to what people already do," said David Veenstra, a research associate professor in the Department of Pharmacy at the University of Washington in Seattle, and a co-author of the recent NEJM paper. "If you talk to a lot of clinicians who care for warfarin patients, they'd say, 'If I get this information, I might modify the starting dose; I'm not sure what else I would do, but I'm going to need that information in hours.' And so, lack of a point-of-care assay has been a stumbling point," he said.
Research-oriented medical centers that have a quick turnaround time are working on such devices, Veenstra said. Unfortunately, a center Veenstra mentioned, Intermountain Health Care in Salt Lake City, Utah, could not be reached for comment before press time.
But point-of-care diagnostics may not be absolutely required for warfarin-related tests. Genetic testing is still useful during the trial-and-error process of warfarin dosing, even if it must be sent to a laboratory through overnight mail, said Richard Stein, a spokesperson for the American Heart Association and director of preventive cardiology at Beth Israel Hospital in New York.
At least two companies are interested enough in the new research on VKORC1 and 2C9 to discuss the possibility of creating diagnostic products, Brian Gage, an associate professor of medicine at the Washington University School of Medicine in St. Louis [and a co-author of the NEJM study], told Pharmacogenomics Reporter. Gage said he was not able to name the companies, but added that they had expressed interested in patents related to the research and assigned to the University of Washington in Seattle. It was not clear whether the companies were working on point-of-care devices.
FDA Gets Involved
There is yet another stumbling block to getting genetic tests between patients and warfarin. "No one has ever done a study showing that using this information improves patient outcomes," Veenstra said.
The US Food and Drug Administration is ponying up some funds to change that situation, however. Brian Gage of Washington University, along with a handful of other researchers have responded to an FDA proposal to fund research into 2C9 and VKORC1 "in order to prospectively determine if knowledge … improves the safety and efficacy of warfarin therapy and warrants genotyping of CYP2C9 when initiating anticoagulant therapy," according to an FDA notice regarding the program. The study will prospectively compare standard of care and pharmacogenomics-guided warfarin treatment in 400 subjects, the notice said. The amount funding offered by the FDA was not available.
Applicant reviews should have occurred by May 31, but this has not been verified. VKORC1 will be studied retrospectively, according to the notice.
What Will VKORC1 Change?
Knowing a patient's VKORC1 status has not yet been shown to help prevent adverse events, although there is still good reason to perform the tests. Larger patient populations will be needed to determine that sort of thing, said Veenstra.
What the recent NEJM study found is that the, the VKORC1 enzyme, which is related to vitamin K, could explain about 25 percent of the variability in warfarin response through its own gene variants, said Reider. Previous factors all added together account for about 30 percent of the variability, he said. "So, essentially, we kind of doubled the amount of variability that we could explain just by looking at this one new gene," said Reider. The amount of response now explained by variability is greater than half with the new information.
Despite any proven ability to prevent adverse events, the new research on VKORC1 "makes a more compelling case for using genetic testing," Coleman said, adding that Genelex is only "contemplating" creating a test for the gene. Anticoagulation drugs are "the best-made case for the impact of polymorphisms on dosage and such," he said. "It's very likely that if these patients knew about their genotype, you could tell them to be more careful about diet and vitamin K consumption," he said. "It's going to take something like a clinical trial to convince the anticoagulation people."
In addition to the disadvantages stemming from the lack of a point-of-care 2C9 test, data about polymorphisms in that gene and clinical factors together haven't yet been useful enough for clinicians to integrate into practice, said Stein, the Beth Israel cardiologist. Instead, a tedious empirical process is used, beginning with a standard low dose and prothrombin test, followed by increasing doses and further tests. "What I would need is to triage patients into a little, medium, and a lot," he said.
Asked whether the addition of VKORC1 to the factors used to estimate dosing would encourage genetic testing, Stein said, "Yes, and I suspect that we'll end up with genetic markers of the other variability" in response to the drug. "If it's a high-risk situation, which about half the time it is, I have to keep them in the hospital on heparin," said Stein. "It's hospital time and it's many more blood tests, because we do them on a seven-day cycle," which could be shortened dramatically for patients likely to tolerate a high warfarin dose, he said.
It appears that there is no warfarin-related test on the market that includes VKORC1. All of the warfarin-related tests offered to consumers and doctors by shops such as Genelex and Aruplab center solely around 2C9 genotyping. The test favored by Arup Laboratories, for example, is a PCR-based 2C9 diagnostic made by Roche Molecular Diagnostics.
Other manufacturers of 2C9 tests include: GE Healthcare, which incorporates this genetic test into a bundle with eight other genes on its CodeLink CYP450 microarray; Jurilab, whose DrugMet genotyping test also includes seven other metabolism-related enzymes; Tm Bioscience, which sells a 2C9 Tag-It mutation detection kit that it provides to LabOne; Genaissance, which offers 2C9 GLP-lab testing to clients, who are often drug makers; Esoterix, which recently became a subsidiary of LabCorp; and Third Wave Technologies, which sells a 2C9 Invader assay.
More than 20 million prescriptions for warfarin were written in 2003, according to NDC Health, a healthcare-industry technology and services provider.
Worldwide net sales of Bristol-Myers Squibb's Coumadin, the company's trade name for warfarin, were $76 million in 2004, with $69 million in the United States, according to documents the company filed with the US Securities and Exchange Commission. During the first quarter of 2005, BMS posted $49 million in total US sales for Coumadin, a rise of 32 percent above first-quarter 2004 sales. Market exclusivity for Coumadin expired in the United States in 1997, the company said.
Bristol-Myers Squibb was not able to respond to questions about warfarin-related diagnostics before this issue went to press. Also not responding were Tm Bioscience, Third Wave Technologies, Roche Molecular Diagnostics, Abbott Diagnostics, Quest Diagnostics, and Genaissance. A GE Healthcare spokesperson could not provide information about any interest on the company's part in VKORC1.
"If you combine serious and life-threatening events, the rate [of adverse events for warfarin] is somewhere between five and 10 events for every 100 patient-years," said Veenstra. Most of the estimates for warfarin events come from academic medical clinics and anticoagulation clinics, "where the monitoring is pretty good," Veenstra said. The situation with general practitioners and others might be more erratic, he said.
— Chris Womack ([email protected])