Whither Personalized Medicine? Warfarin Study May Help with the Answer
The National Institutes of Health announced in late February that it will soon begin a clinical trial that will prescribe the anticoagulant drug warfarin based on genetic data collected by an international consortium that the NIH spearheaded. The decision and certainly the outcome could have enormous implications for personalized medicine, a new medical field where the results of genetic tests or other biomarker assessments are used to tailor drugs and treatments to individual patients
NIH’s Pharmacogenomics Research Network has compiled data from 5,700 patients across the globe who are being prescribed the widely-used blood-thinning drug. Every year, 2 million Americans with certain heart conditions start taking warfarin, but their doctors often encounter difficulty with prescribing the drug since the optimal dosage for each individual patient varies widely. This usually results in doctors taking a trial-and-error approach to dosing that takes months to perfect and runs the risk of causing harmful side effects in the meantime. Warfarin is especially risky for patients for whom the optimal dosage is either very high or very low.
The patient data collected by the NIH included both the patients’ demographic and clinical information as well as genetic information on two gene variants, CYP2C9 and VKORC1, both of which are known to influence a patient’s ability to metabolize warfarin. The consortium also collected data on the initial and optimized doses of warfarin that the patients were prescribed.
According to an article published by the consortium in the New England Journal of Medicine, the 5,700 subjects were divided into two groups along an 80/20 split. This means the data for 80 percent of the subjects were used to construct a dosing algorithm and the other 20 percent were used as a validation cohort to test the algorithm. After running some statistical analyses, the researchers discovered that the addition of the genetic data to the demographic and clinical data allowed for a better prediction of the optimal dosage than just the demographic and clinical data alone. The genetic data was most helpful for those patients whose optimal doses are outside the average range—either very high or very low. This led an accompanying editorial in NEJM by Janet Woodcock and Lawrence Lesko of the Center for Drug Evaluation and Research at the FDA to argue that for warfarin, the “evidence base for pharmacogenetic testing should be informed by…the characteristics of the outliers.”
The lead investigator for the consortium, Stanford University’s Russ Altman, reported that his team plans to tweak the dosing algorithm through an ongoing study of 100 patients from the San Francisco Bay area and report the data on the PharmGKB website. Altman and the other authors also admitted in the report that research still needs to be done on whether the improved dosing leads to better clinical outcomes. This is where the NIH’s upcoming clinical trial falls right into place.
NIH will now embark on a prospective double-blind clinical trial where approximately 1,200 subjects will be split into two groups, one being prescribed warfarin according to clinical data and one using both clinical and genetic data. GenomeWeb Daily News reports that a European team at Newcastle University and the University of Liverpool are also working on a similar clinical trial that will take place at 13 research centers in seven countries where they expect to enroll 2,700 subjects.
The question of clinical outcomes is an important one since the true potential of pharmacogenomics for improving our nation’s health lies not just in the scientific advancements but in the clinical effectiveness advancements it generates. Indeed, pharmacogemomics-based treatments, like all treatments, need to pass the practicality test if biomedical innovation is to make a constructive contribution to a larger system-based healthcare infrastructure.
Last month, for instance, ScienceDaily reported on an analysis conducted by the University of Cincinnati which found that even though pharmacogenetic-based dosing of warfarin improved outcomes, it did so at very high cost—$170,000 per quality-adjusted life year gained. The current rule governing the interpretation of most cost-effective analyses is $50,000 per quality-adjusted life year gained.
The main focus of the UC research was to determine whether pharmacogenetic-dosing decreased the risk of major bleeds. The analysis was conducted on the combined data of the only three clinical studies of pharmacogenetic-guided warfarin dosing that had been conducted by that time. The researchers also found that there is only a 10 percent chance that the pharmacogenetic-based would be cost effective. The lead investigator, Dr. Mark Eckman, recommended a number of conditions that could make the dosing more cost-effective.
Specifically, it should be used for patients who have a high risk for hemorrhage, prevent more than 32 percent of major bleeding events, be available within 24 hours, and cost less than $200.
He recommended that the upcoming NIH clinical trial “examine the impact of pharmacogenetic-guided dosing on bleeding risk and monitor outcomes long enough to determine the true duration of benefit,” suggesting that patients with a higher risk of bleeding should not be excluded if it is determined that they need warfarin. Eckman summed it all up by saying, “personalized, predictive medicine offers great promise, but we need to carefully examine the benefits and understand the cost-effectiveness of such strategies before we spend a lot of money on very expensive tests.”
Image: AP/ED ANDRIESKI
Comments on this article