Influence of the VKORC1 3730 G > A polymorphism on warfarin dose

Abstract

To the Editor,We have read with great interest the recently publishedarticle “A new warfarin dosing algorithm includingVKORC1 3730 G > A polymorphism: comparison withresults obtained by other published algorithms” by Ciniet al. [1]. The pharmacogenetic dosage of vitamin K-antagonists (VKA) has been a much investigated anddebated topic in the last decade. Despite the abundanceof pharmacogenetic dosage algorithms, including thosederived from large and diverse cohorts [2], currentguidelines still recommend against routine pharmacoge-netic testing of anticoagulated patients [3]. This recom-mendation is based on the lack of evidence fromrandomized trials regarding beneficial effects on end-points such as major bleeding or TIR [time in rangeof International Normalized Ratio (INR)] [4, 5]. Never-theless, in a recent randomized trial, higher TIR andfewer very low INR values were observed during thefirst 3 months of therapy in the pharmacogenetic dosagegroup [6]. All in all, pharmacogenetic dosage couldvery well play a role in the anticoagulant therapy ofthe future.Most studies of VKA pharmacogenetics have shown thatabout half of the large inter-individual dose variation can beexplained by three polymorphisms: one in the VKORC1gene, rs9934438 [7], and two in the CYP2C9 gene,rs1799853, and rs1057910 [8]. All three variant allelescorrelate with lower dose requirements.In their study, Cini et al. investigated another poly-morphism in VKORC1, rs7294 [1]. The authors foundthat warfarin dosage varied substantially among geno-type groups of this polymorphism, with higher require-ments for patients possessing variant alleles (Fig. 1 in[1]). The polymorphism correlated significantly withdose in a univariate analysis. In the following multiplelinear regression model, rs7294 was not a significantpredictor of warfarin dose, having a regression coeffi-cient of 0.025 with a 95 % confidence interval (CI)ranging between −0.201and0.251(Table3in[1]). Onthe other hand, another polymorphism inVKORC1,rs9934438, was an important and significant determinantof warfarin dose with a regression coefficient of−0.571(95 % CI−0.782 to−0.361).Similar findings have been reported previously fromlarger cohorts [9, 10]. Notably, rs7294 was not anindependent predictor of warfarin dose in either study.Table 1 shows the number of patients with variantalleles of rs7294 and rs9934438 based on our data from[10]. The two polymorphisms are not in complete link-age disequilibrium, but there is a statistically significantcorrelation (χ