Development and comparison of warfarin dosing algorithms in stroke patients

Abstract

The genes for cytochrome P450 2C9 (CYP2C9) and vitamin k epoxide reductase complex subunit 1 (VKORC1) have been identified and studied as important genetic determinants of warfarin dosing. We developed warfarin algorithm for Korean patients with stroke and compared the accuracy of warfarin dose prediction algorithms based on the pharmacogenetics. A total of 95 patients on stable maintenance dose of warfarin were enrolled. Warfarin dosing algorithm was developed using multiple linear regression analysis. The performance of all the algorithms was characterized with coefficient of determination, determined by linear regression, and the mean of percent deviation predicted doses from the actual dose. In addition, we compared the performance of the algorithms using percentage of predicted dose falling within ±20% of clinically observed dose and dividing the patients into a low-dose group (≤3mg/day), an intermediate-dose group (3-7mg/day), and a high-dose group (≥7mg/day). A newly developed algorithm included the variables of age, body weight, and CYP2C9 and VKORC1 genotype. Our algorithm accounted for 46.5% of variation in stable warfarin doses. The predicted doses using algorithms derived from Anderson and this study showed the best correlation with the actual maintenance doses. Our algorithm performed best in predicting dose within 20% of actual dose and intermediate-dose. Our warfarin dosing algorithm may be useful for Korean patients with stroke.