First approved for use in 1954, the “blood thinner” warfarin effectively prevents clots but must be dosed carefully. A too-high starting dose can cause internal bleeding, while an insufficient dose may fail to protect against clots. In a hint at the promise of pharmacogenomics, inexpensive, 59-year-old warfarin may become more valuable each time a newly identified genetic marker makes more accurate algorithms used to predict dose for each patient.
As it stands, most physicians start patients on a standard five-milligram dose, closely monitor their clotting speed and adjust their dose based on it. The effort to reach a safe dose is complicated by the 20-fold variability in dose requirements observed in African Americans. The best available dosing models miss their mark often enough that warfarin causes a third of U.S. hospital visits related to drug side effects in patients aged 65 and older, with nearly two thirds of those due to bleeds.
Seeking to fix the problem, three previous genome-wide association studies (GWAS) sought to identify genetic factors that could improve warfarin dose prediction, but none of them included patients of African ancestry. That makes the current study, a seventeen-institution effort led by the University of Florida, the first GWAS of its kind. The University of Alabama at Birmingham was the largest enrolling center in the study, with local demographics and a tradition of tackling disparities best positioning it to recruit African-American patients.
“Our goal is to help more patients arrive at their optimal dose more quickly,” said Nita Limdi, Pharm.D., Ph.D., associate professor in the Department of Neurology within the UAB School of Medicine and a first author of the study. “As the field continues to identify genetic variations that affect dose, knowing which patients don’t have certain variations will be as valuable as knowing which do.”
GWAS look at differences at many points in the genetic code to see if, across a population, one or more variations are found more often in those with a given trait, be it high risk for a disease or lower needed drug dose. The current study found that a single genetic marker is associated with a 20 percent warfarin dose reduction in people of African ancestry, an effect not present in patients of European or Asian ancestry. When incorporated into dosing algorithms, the new variant enabled the prediction of dose with 21 percent greater accuracy than the standard formula.
“Clinic by clinic, doctors are starting to check genetic codes as a first step in determining each patient’s disease risk, likelihood of benefiting from a given drug or best dose,” said Julie Johnson, director of the Center for Pharmacogenomics at the University of Florida and corresponding author for The Lancet publication. “As this revolution unfolds, we need to ensure that all share in the benefits.”
Where to start?
Clotting prevents small injuries from causing extended bleeds, but it can also be triggered by diseases like atrial fibrillation and cancer, as well as by surgery, aging and inactivity. Once formed, clots can float through the circulatory system to clog blood vessels elsewhere, causing heart attacks, strokes or pulmonary emboli. For these reasons, 25 million people each year get a prescription for warfarin for the treatment and prevention of clots.One reason warfarin is still the most widely prescribed anticoagulant, instead of newer, more easily tolerated drugs like dabigatran, rivaroxaban and apixaban, is its low cost. As a mainstay in the treatment of the underinsured in the United States, and for patients in developing countries, especially in Africa, making warfarin safer remains a priority.
With the wrong dose often leading to dire consequences, the field has for years been building statistical models meant to determine individuals’ perfect warfarin dose based on measurable characteristics. Researchers joined forces in 2007 to form the International Warfarin Pharmacogenetics Consortium (IWPC) , which began adding genetic information to the traditional, consensus dosing formula based factors including weight, race, age, sex and other medicines taken.
IWPC studies in recent years determined that patients of European, Asian and African ancestry with certain genetic variations in two genes, VKORC1 and CYP2C9, require lower doses. Conventional dosing in such patients can lead to a higher risk of bleeding (“over-anticoagulation”). Going into the current study, however, dosing algorithms, including variations in these genes, left unexplained most of the variability seen in warfarin dosing. Furthermore, the known VKORC1 and CYP2C9 variations accounted for 30-35 percent of variability in people of European or Asian origin, but just 7-10 percent in those of African ancestry.
Even the smallest genetic variations, called single nucleotide polymorphisms (SNPS), can have a major impact on a trait by swapping just one of 3.2 billion “letters” making up the human DNA code. The current GWA study found a statistically significant association between a SNP called rs12777823 and reduced warfarin dose requirements in African-Americans.
Data analysis further revealed that a person with rs12777823 requires as many as nine fewer milligrams of warfarin per week, a good portion of the 40 mg average weekly dose for African-Americans. Often overlooked, said Limdi, is that patients who do not possess these variants are under-dosed, placing them at higher risk for clots.
Given that the known VKORC1 and CYP2C9 SNPs account for just seven percent of variability in African-Americans, the authors say the discovery of a SNP that explains another five percent represents “a huge leap.”
Looking forward
Associations found by GWA studies between an SNP and any trait do not prove that one causes the other. The patterns they suggest must be confirmed by randomized clinical trials before they can change clinical practice.| First approved for use in 1954, the “blood thinner” warfarin effectively prevents clots but must be dosed carefully. One reason warfarin is still the most widely prescribed anticoagulant, instead of newer, more easily tolerated drugs, is its low cost. |
Beyond COAG, Limdi’s focus is on creating the infrastructure at UAB needed to incorporate genetic information into better informed clinical decisions in daily practice. She leads a project called PRIME, Pharmacogenomic Resource for Improving Medication Effectiveness, which identifies projects where adjusting treatment based on patient genotypes can dramatically improve a drug’s effectiveness.
The first such effort is focused on antiplatelet drugs given after patients receive an implanted stent to keep open partially blocked arteries. Despite the design of coatings meant to prevent it, stents can attract notice from blood platelets, which sense their foreign surfaces and stick to them to encourage clotting. Many of the 450,000 U.S. stent recipients each year receive anti-platelet drugs like Plavix, but one in five patients has SNPs that prevent them from converting the drug into its active form. Limdi and colleagues are finalizing a genotype-guided system expected to come online this fall that will administer Plavix alternatives (e.g. ticagrelor or prasugrel) when needed.
Next in line for clinical launch, should ongoing trials confirm its value, would be a warfarin genotyping initiative, an effort Limdi expects to mature in 2014. On the horizon, she hopes to apply genomic information to treatment choice and dosing in cancer as well.
Along with Limdi, Minoli Perera, Pharm.D., in the Section of Genetic Medicine in the Department of Medicine at University of Chicago and Larisa Cavallari, Pharm.D., in the Department of Pharmacy Practice at University of Illinois at Chicago, were named first authors for having designed the study, gathered and interpreted data and drafted the study report. All told, 42 authors made vital contributions to the work. Three of them declared consulting relationships with industry or held a related patent, with the details included in The Lancet journal text. The study was funded by institutions including the National Institutes of Health, the American Heart Association, Howard Hughes Medical Institute and the Wellcome Trust.
Along with Limdi, UAB authors of the study were Nianjun Liu, Ph.D., and Jelai Wang in the Section on Statistical Genetics, which is part of the Department of Biostatistics in the UAB School of Public Health.