A new algorithm to predict warfarin dose from polymorphisms of CYP4F2, CYP2C9 and VKORC1 and clinical variables: Derivation in Han Chinese patients with non valvular atrial fibrillation

Few pharmacogenomic dosing regimens of warfarin have been developed for Chinese patients with non valvular atrial fibrillation (NVAF). The objective of this study was to develop a new algorithm by polymorphisms of CYP2C9, VKORC1 and CYP4F2 to predict the daily stable dose of warfarin in Chinese patients with NVAF. A total of 325 Chinese NVAF patients on stable dose of warfarin with a target international normalised ratio of 1.5 to 3.0 were recruited and divided randomly into two cohorts. CYP2C9*3, VKORC1-1639, VKORC1 1173 and CYP4F2 were detected by ligase detection reaction method. The new algorithm was developed with multivariate linear regression in cohort 1 (260 patients) and assessed with Pearson Correlation Analysis (PCA) in cohort 2 (65 patients). From 260 enrolled patients, the model (R2 = 51.7%) was developed as: Dose = 3.47 - 0.022 (AGE) + 0.017 (WT) + 0.189 (PTE) - 0.283 (β-blocker) - 0.471 (AMIO) - 0.586 (CYP2C9 *1/*3) - 0.296 (VKORC1 CT) - 0.648 (VKORC1 TT) + 0.219 (CYP4F2 TT). PCA displayed that the algorithm was good (r = 0.658). The residual plots revealed that the predicted doses by the algorithm tend to be overestimated when lower doses were administered to patients and to be underestimated in higher doses. The algorithm developed by us might predict warfarin dose used by Chinese NVAF patients.