Introduction.
Genetic differences in warfarin metabolism (via CYP2C9) and vitamin K recycling (VKORC1) have been found to significantly influence warfarin maintenance dose requirements. A Bayesian forecasting method for warfarin therapy should predict dosing requirements without prior genetic data.
Aims.
The aim of this study was to determine the influence of CYP2C9 and VKORC1 genotype on the predictive performance of a Bayesian warfarin dosing method.
Methods.
Patients who were initiating warfarin therapy were genotyped for CYP2C9 *1, *2 [rs1799853] and *3 [rs1057910] alleles and for the VKORC1 promoter variant -1639G>A [rs9923231]. The warfarin dosing history for each patient was retrospectively recorded up to the steady-state INR (INRss). TCIWorks, a Bayesian dose individualisation tool, was used for predicting maintenance doses using all INR measurements up to the last dose change. Observed and predicted maintenance doses were compared using measures of bias (mean prediction error [MPE]) and imprecision (root mean square error [RMSE]).
Results.
A total of 102 patients completed the study. Maintenance dose predictions were positively biased for patients with VKORC1 G/G genotype (MPE +0.6 mg/day [95% CI 0.89, 0.29]) and CYP2C9 *1*1 genotype (MPE +0.39 [95%CI 0.69, 0.09]. The 95% CI of the MPE included zero for patients with CYP2C9 *1*2 and *1*3 genotypes and for VKORC1 A/G and A/A genotypes indicating unbiased predictions.
Discussion.
Warfarin maintenance dose predictions were over-predicted by +0.60mg/day on average in patients with VKORC1 G/G genotype and were less precise than other VKORC1 genotypes, likewise predictions for CYP2C9*1*1 genotype was over-predicted by +0.39mg/day on average and less precise than other CYP2C9 genotypes. Further research to explore the positive bias seen in the these genotypes is warranted.
Wadelius M et al (2009) Blood 113:784-792
Wright DFW & Duffull SB (2013) Clin Pharmacokinet 52:59-68