The National Heart, Lung, and Blood Institute's 1,200-patient prospective, randomized-controlled trial investigating whether pharmacogenetic-based warfarin dosing is clinically useful, might just be the study that critics and supporters have been waiting for.
NHLBI plans to launch next month the much anticipated Clarification of Optimal Anticoagulation through Genetics trial, which will compare clinical outcomes of one group of patients given warfarin based on just clinical factors and another cohort administered the anticoagulant based on clinical factors and genotype.
The prospective, randomized-controlled trial comes on the heels of a large retrospective study conducted by the International Warfarin Pharmacogenetics Consortium, "Warfarin Dosing Using Clinical and Pharmacogenetic Data," the results of which are published this week in the New England Journal of Medicine.
In this retrospective study, researchers found that when genetic information is factored in, the initial warfarin dose was much closer to the required stable therapeutic warfarin dose than when using just clinical factors or giving a fixed dose. Furthermore, the PGx-based warfarin dosing tended to benefit the 46.2 percent of the "outlier" population, requiring the smallest and largest warfarin doses.
The NEJM study was accompanied by an editorial, "Pharmacogenetics - Tailoring Treatment for the Outliers," written by the Janet Woodcock, the director of the Center for Drug Evaluation and Research, and Lawrence Lesko, the director of the Office of Clinical Pharmacology at the US Food and Drug Administration.
In the editorial the two FDA officials noted that "it is not surprising" that 54 percent of "middle range" patients in the NEJM study derived no benefit from PGx-guided warfarin dosing compared to fixed doses of the drug. "It was the outliers, those persons who required doses higher or lower than the mean, for whom determination of the dose directed by pharmacogenetics was more predictive," Woodcock and Lesko pointed out.
In the editorial, they also noted that randomized-controlled trials may not be necessary to determine the clinical utility of all PGx tests. In the case of warfarin, based on the results of the IWPC trial, future studies should be powered to test the utility of PGx-based warfarin dosing on the "outlier populations."
"In some cases, randomized, controlled trials will be needed to determine whether pharmacogenetic testing is worthwhile; in others, less rigorous approaches will suffice," Woodcock and Lesko wrote in their editorial. "Given the expected volume of genetic information and the relative paucity of randomized, controlled trials involving marketed drugs, we need clear thinking about what is required for the adoption of pharmacogenetic testing."
However, in the case of warfarin, fueling the need for a prospectively-designed, randomized-controlled trial are payors. Most large insurers do not cover PGx-based warfarin dosing, and the Centers for Medicare and Medicaid Services is currently in the process of crafting a national coverage decision on the technology [see PGx Reporter 10-29-2008].
"The NIH research is precisely what is needed to advance the promise of personalized medicine, ensuring that patients receive the safest and most effective drug dose," Frank Torti, acting commissioner of Food and Drugs, said in a statement.
[ pagebreak ]
In August 2007, the FDA updated the label for warfarin to note that people with variations of the genes CYP2C9 and VKORC1 may respond adversely to the drug. The agency did not require physicians to genetically test their patients, however, noting that additional outcomes studies would be necessary [see PGx Reporter 09-05-2007].
In the COAG trial, half of the participants will be randomly selected to have their warfarin doses determined by clinical information alone; the doses for the other study group will be determined by clinical factors, as well as whether they carry variants of the CYP2C9 and VKORC1 genes.
The study participants will be followed for six months. Researchers will assess how long participants in each group maintain the desired level of blood thinning at two and four weeks after starting therapy, as well as at three and six months. Researchers will also review bleeding problems and other complications, quality of life and cost of therapy.
"We think this will be a proof-of-concept trial," Yves Rosenberg, the project officer for the COAG clinical trial, told Pharmacogenomics Reporter this week. "We plan to show that if there is a difference [in outcomes] then it is due to the genetic information and nothing else."
The COAG trial is supported by NHLBI, with additional funding for genome analyses from the National Human Genome Research Institute. Warfarin will be donated by Bristol-Myers Squibb, the maker of Coumadin. The COAG trial is also working with the Critical Path Institute, which will conduct and collect the genotyping data for the study.
With just under $10 million in total funding, the COAG trial represents a significant investment in pharmacogenetics from the NIH. "These efforts showcase NIH's firm commitment to building a future of personalized medicine -- a future in which doctors will be able to prescribe the optimal dosage of medicine for each patient right from the start." Raynard Kington, acting NIH director, said in a statement.
The retrospective study, conducted by members of the IWPC, collected clinical and genetic data from 4,043 patients to dose one set of patients based only on clinical variables and another set of patients on genetic information and clinical variables. The researchers validated their findings in 1,009 patients whose predicted dose of warfarin was within 20 percent of the actual stable therapeutic dose.
The IWPC consortium members had access to anonymized information 5,700 people on stable dosages of warfarin. The samples were drawn from an ethnically diverse cohort, from Taiwan, Japan, Korea, Singapore, Sweden, Israel, Brazil, Britain and the US.
In the study warfarin doses were calculated three ways -- based on standard clinical data; based on clinical data and genetic variations; and fixed daily doses. In developing dosing algorithms for each patient, researcher considered demographic information like age, gender and race; CYP2C9 and VKORC1 variants; and initial, as well as optimized, warfarin dosages. Then, the scientists compared how closely their computational predictions matched the actual, clinically derived stable warfarin dosage for each patient.
"The use of a pharmacogenetic algorithm for estimating the appropriate initial dose of warfarin produces recommendations that are significantly closer to the required stable therapeutic dose than those derived from a clinical algorithm or a fixed-dose approach," the study authors concluded. "The greatest benefits were observed in the 46.2 percent of the population that required 21 mg or less of warfarin per week or 49 mg or more per week for therapeutic anticoagulation."
This means that for more than half of the study cohort, or the "average" patient, the gene-based algorithms did not yield better outcomes than when they received fixed-dose warfarin.
This study received support from NHLBI, the National Institute of General Medical Sciences, the National Institute of Neurological Disorders and Stroke, and the National Center for Research Resources.
RCT: Not Always Gold
According to Jeremy Berg, director of the NIGMS, the large retrospective trial indicated "the possible outcomes" of using pharmacogenetic strategies to dose warfarin. In the prospective COAG trial, researchers are looking at whether this type of PGx-guided dosing "can be implemented in the clinical trial, will it help people get to a faster stabilization of the dose, and will it reduced adverse events?" Berg noted.
[ pagebreak ]
"And that's something that has to be done in a randomized trial going forward," Berg told Pharmacogenomics Reporter this week.
Randomized-controlled clinical trials are the FDA's gold standard for drug development.
However, FDA's Lesko has previously said that demanding randomized-controlled trials to validate the utility of PGx-guided dosing for older drugs, such as warfarin, might be a prohibitively high bar [see PGx Reporter 02-28-2007].
In the NEJM editorial, Woodcock and Lesko point out that "the evidence base for pharmacogenetic testing should be informed by the pharmacologic characteristics of the drug and the characteristics of the outliers.
With regard to warfarin, hypotheses that could be investigated with a randomized-controlled trial include whether dose prediction directed by pharmacogenetics hastens achievement of the target INR; whether genetic testing reduces the consequences of out-of-range INRs; or whether genetic testing reduces visits to doctors' offices.
However, in the case of warfarin, as shown by the trial published in NEJM, only "outlier" patient populations, requiring the highest and lowest doses, will likely benefit from genetic testing. Thus, "trials that investigate these hypotheses only in outliers would have improved power to detect any benefit to them," Lesko and Woodcock write. "Of course in order to conduct such a trial, there needs to be a way to
identify the outliers."
A smaller randomized-controlled trial published in the Nov. 7 issue of Circulation, which enrolled more than 200 patients initiated on warfarin randomized to either PGx-guided or standard dosing, failed to reach its primary endpoint of a reduction in bleeding outcomes, as measured by out-of-range international normalized ratio. However, the study did show that "an algorithm guided by pharmacogenetic and clinical factors improved the accuracy and efficiency of warfarin dose initiation" [see PGx Reporter 12-07-2007].
Based on the results of this study, several large insurers that did not cover PGx-based warfarin dosing, maintained their position that personalized medicine is not ready for prime time [see PGx Reporter 03-19-2008].
However, at the time, Lesko urged the healthcare community to not focus on the trial's negative endpoint. In an editorial in Personalized Medicine last March discussing the trial, Lesko pointed out that the randomized-controlled trial was powered "to detect an ambitious difference in out-of-range [international normalized ratios] between pharmacogenetic (20 percent) and clinical arms (40 percent).
"Thus, future studies should be powered to detect smaller differences," he wrote at the time.