Comparative performance of warfarin pharmacogenetic algorithms in Chinese patients

Introduction

Multiple warfarin pharmacogenetic algorithms have been confirmed to predict warfarin dose more accurately than clinical algorithm or the fixed-dose approach. However, their performance has never been objectively evaluated in patients under low intensity warfarin anticoagulation, which is optimal for prevention of thromboembolism in Asian patients.

Material and methods

We sought to compare the performances of 8 eligible pharmacogenetic algorithms in a cohort of Chinese patients (n = 282) under low intensity warfarin anticoagulation with target international normalized ratio (INR) ranged from 1.6 to 2.5. The performance of each algorithm was evaluated by calculating the percentage of patients whose predicted dose fell within 20% of their actual therapeutic dose (percentage within 20%), and the mean absolute error (MAE) between each predicted dose and actual stable dose.

Results

In the entire cohort, the pharmacogenetic algorithms could predict warfarin dose with the average MAE of 0.87 ± 0.17 mg/day (0.73-1.17 mg/day), and the average percentage within 20% of 43.8% ± 8.1% (29.1% - 52.1%). By pairwise comparison, warfarin dose prediction was significantly more accurate with the algorithms derived from Asian patients (48.6% - 50.0%) than those from Caucasian patients (29.1% - 39.7%; odds ratio OR]: 1.61-3.36, p ≤ 0.02). Algorithms with additional covariates of INR values or CYP4F2*3 performed better than those without the covariates (adding INR: OR: 1.71 (1.08-2.72), p = 0.029; adding CYP4F2*3: OR: 2.67(1.41-5.05), p = 0.004). When the patients were stratified according to the dose range, the algorithms from Caucasian and racially mixed populations tended to perform better in higher dose group (≥ 4.5 mg/day), and algorithms from Asian populations performed better in intermediate dose group (1.5-4.5 mg/day). None of the algorithms performed well in lower dose group (≤ 1.5 mg/day).

Conclusions

No eligible pharmacogenetic algorithm could perform the best for all dosing range in the Chinese patients under low intensity warfarin anticoagulation. Construction of a refinement pharmacogenetic algorithm integrating 3 genotypes (CYP2C9, VKORC1 and CYP4F2) and INR data should be warranted to improve the warfarin dose prediction in such patients.