Development of A Genomic-based Predictive Model for Warfarin Dosing

Abstract

Genetic diversity among patients is a significant factor in deciding the appropriate warfarin dose when oral anticoagulation is started; nevertheless, useful applications of genetic data have not been tested in a large and diverse population. As an anticoagulant, warfarin can be administered at doses ranging from less than 10 mg/week to more than 100 mg/week to provide the same therapeutic amount of anticoagulation to patients. The appropriate dose must be determined because underdosing puts patients at risk for thromboembolic events and overdoing can produce side effects including bleeding. Genes related to warfarin's pharmacokinetic and pharmacodynamic properties have genetic differences that affect the dosage needed. The two most significant gene variations affecting warfarin dosage are those encoding VKORC1, the target of warfarin, and CYP2C9, a key metabolising enzyme. Dosing algorithms that take into account clinical characteristics and important polymorphisms from these genes have the potential to help prevent the side effects associated with long-term "guess-and-test" dosing by partially predicting stable warfarin doses. The genetics of CYP2C9 and VKORC1 account for about 30% of the variation in warfarin dose required. Even when clinical parameters are added to the genetic data, current dosage algorithms are unable to completely account for the difference in dose. For Warfarin dosage, we created and employed a genomic-based predictive model.