Stanford Researchers Use Gene Studies to Help Predict Ideal Warfarin Dose


In a study published February 18, 2009 in the New England Journal of Medicine, Dr. Russ Altman, MD, PhD, describes how a patient's genetic profile can be used to custom tailor warfarin drug dosing.

The coumadin derivative warfarin (Coumadin, Jantoven, Marevan, Waran) is a widely prescribed oral anticoagulant used long-term for the prevention of blood clots. However, warfarin has a narrow therapeutic range, and the medication dose leading to this range can vary as much as ten-fold. Altman, a professor of bioengineering, genetics and medicine at Stanford University School of Medicine and lead investigator of this study, reports that "it appears that up to 46 percent of people will require a warfarin dose that is significantly higher or lower than average." Consequently, for many people it can take several months and many blood tests to determine an optimal or ideal warfarin dose. A dose that's too high can lead to uncontrolled bleeding whereas a dose too low can lead to blood clots. An ideal warfarin dose is determined by the results of the International Normalized Ratio (INR) blood test, a measure of how long it takes blood to clot. Problems in selecting an ideal warfarin dose occur because warfarin metabolism is influenced by genetic and environmental factors, including other medications and diet.


The genes that control warfarin pharmacokinetics include the cytochrome p450 gene CYP2C9 (cytochrome P450 2C9), which influences drug metabolism in the liver, and the VKORC1 promoter, a protein involved in the body's synthesis of vitamin K, which is the site of action for warfarin. Individuals with polymorphisms of the CYP2C9 gene are warfarin "sensitive" and require lower doses of warfarin, whereas individuals with VKORC1 missense mutations, particularly the D36Y mutation, are warfarin "resistant" and require higher doses of warfarin. Other inherited differences in the VKORC1 gene increase or decrease the amount of warfarin needed to inhibit the formation of clotting factors. Studies show that the prevalence of these genetic mutations varies among different ethnic groups. About 35 percent of Caucasians have a slow acting form of the CYP2C9 enzyme.

Although other institutions have studied the effects of these genes on warfarin dosing in specific ethnic groups, Stanford's study is the largest and most inclusive study to examine warfarin's genetic influences. The study includes data on more than 5,000 patients from many different ethnic groups. Using previously established warfarin dosing data along with other patient demographics such as weight and ethnicity for 4,000 patients, Stanford researchers have developed a computerized dose-prediction algorithm for warfarin. As part of the study, researchers used this algorithm, both with and without genetic study results, on the remaining 1,000 patients in the study. The researchers found that using the algorithm alone yielded initial warfarin doses that were closer to the optimal dose than using a fixed-dose regimen. When they included the genetic study results of the patients, they were able to arrive at an even more accurate initial warfarin dose. The algorithim and the data used for its derivation are available on Stanford's Pharmacogenetics and Pharmacogenomics Knowledge base, The database was created by Altman, who is also the principal investigator of the PharmGKB database, along with Teri Klein, a senior research scientist in genetics, who is the director. The PharmGKB database is a curated international repository for data and knowledge intended to aid researchers in understanding how genetic variation among individuals contributes to the differences in their reactions to drugs.

The results of Stanford's warfarin study specifically showed that when the pharmacogenetic algorithim (demographs and genetic studies) was used, it yielded dosage predictions that averaged within about 8.5 mg/week of the patient's ideal dose. When the demographic data was used in conjunction with clinical data, it predicted warfarin doses that were within 10 mg/week of the ideal dose. Using the standard fixed dose of 35 mg/week yielded drug doses that differed about 13mg/week from the ideal dose. Researchers are fine-tuning the algorithm in a similar study of about 100 patients at the Stanford Anticoagulation Clinic.

What is the study's greatest value? Currently, physicians can't predict if a patient is best off starting on a low or high or average dose of warfarin. In up to 30 percent of patients it can take as long as 12 weeks to arrive at an ideal dose. The algorithm, if used, flags the 10 to 30 percent of people whose dosing needs deviate from the standard. In the absence of genetic studies, physicians can start patients on a standard dose of warfarin on the first office visit and draw a blood sample for genetic testing. By the second visit, doctors can use these results to fine-tune the dose. This study also paves the way for introducting genetic applications for dosing other drugs, particularly the other 20-30 commonly used drugs that are affected by CYP2C9 variants. Altman and Klein are also working with the International Tamoxifen Pharmacogenomics Consortium to evaluate the use of genetic studies in women with breast cancer using tamoxifen.