Dosing predictions for the anticoagulant Warfarin
Michael Wagner - UNC School of Pharmacy
Alison Motsinger-Reif - North Carolina State University
Key Words: pharmacogenomics, dosing algorithm, human genetics
Warfarin is the most widely used oral anticoagulant worldwide, withover 30 million users in the United States
along, and has been used for over half a century. Despite its widespread use, it ranks among the top ten drugs
in the United States associated with serious adverse events over the last 20 years, as over-dosing leads to serious
bleeding side effects. These side effects are largely caused by the extreme variation in the dose of the drug required
for therapeutic levels that cannot be predicted based on clinical variables alone. Many studies have reliably and
consistently found that variations in two genes, cytochrome P4502C9 (CYP2C9) and Vitamin K epoxide reductase
complex 1 (VKORC1), are significantly correlated with warfarin dose, and the U.S. Food and Drug Administration
(FDA) includes this finding on the drug label.
Recently, the International Warfarin Pharmacogenetics Consortium (IWPC), composed of 21 research groups from
around the world pooled their data to derive a global dosing algorithm based on genetic and clinical factors. Based
on data from over 5000 patients worldwide, the pharmacogenomic algorithm (with both genetic and clinical predictors)
predicts over 40% of the variation in therapeutic dosing, and greatly outperforms traditional fixed dose approaches,
or a dosing algorithm with only clinical predictors. While this algorithm represents an important clinical dosing tool,
there are still several important questions on how to best model these known predictors, and model the relationship
between genotype and phenotype. Additionally, the predictive performance and overall fitting criteria of the dosing
algorithms across different human subpopulations needs to be carefully evaluated.
The International Warfarin Pharmacogenetics Consortium: Estimation of the Warfarin Dose with Clinical and
Pharmacogenetic Data (paper).
Web applet of predictive algorithm Background Information on Genetic Association studies (paper)