Dose response to vitamin K antagonists (VKA) exhibits large within- and between-individual variation in both adults and children, presenting challenges to clinicians attempting to maximize the safety and efficacy of these agents. In part to address limitations with VKAs, alternative oral anticoagulant agents (eg, direct Xa and thrombin inhibitors) are being developed and evaluated in clinical trials. However, for the foreseeable future, VKAs will continue to be used given their familiarity, cost, and ease of access. Thus, anticoagulant research must continue to focus on optimizing use of these agents.
Warfarin produces its anticoagulant effect through inhibition of vitamin K epoxide reductase, which results in diminished recycling of vitamin K and, thereby, impaired gamma-carboxylation of the vitamin K–dependent procoagulants (factors II, VII, IX, and X) and intrinsic anticoagulants (proteins C and S). The net effect of this process is measured by the prothrombin time, and calculated as an international normalized ratio (INR) of the patient's plasma relative to international reference plasma. The INR is affected by a variety of factors, including vitamin K status, warfarin target enzyme (ie, VKOR) activity, and warfarin metabolism. Warfarin metabolism, in turn, is impacted by age and cytochrome P450 (CYP) activity, which itself is modulated by genetic (eg, CYP2C9 *2 and *3 polymorphisms) and environmental (eg, diet, concomitant medication) influences. Given all these variables, the development and validation of predictive models for VKA dose response that inform dosing algorithms have been emphasized in recent years.
In this issue of Blood, Nowak-Göttl and colleagues present the findings of a single-institution prospective cohort study evaluating the influence of VKOR complex 1 (VKORC1) and CYP2C9 gene polymorphisms and clinical factors on therapeutic dosing of VKA in children with venous thromboembolism or thromboembolic ischemic arterial stroke.1 This report is among the first such studies in children. It identifies age as a far more important factor than VKORC1 and CYP2C9 genotypes in determining therapeutic VKA dose (mg/kg) and demonstrates that these genotypes had no appreciable influence on time to achievement of therapeutic INR. Age accounted for 28% of the variation in therapeutic VKA dose (compared with 17% in adults2 ). VKORC1 polymorphisms explained only 3% of variation in warfarin dose requirements, compared with 11% for phenprocoumon; CYP2C9 polymorphisms explained less than 1% of variation in dose requirements for either agent.
While a relatively small study (n = 59 with complete data), it is among the largest of its kind in children, and (by our calculations) its sample size provides 80% power to detect a variance in VKA dose requirements explained by genotype of at least 12%. Given the low incidence of bleeding in children receiving warfarin (0.5% per patient-year in one large prospective study3 ), the findings from this cohort suggest that routine testing of VKORC1 and CYP2C9 genotypes contributes little to the management of children receiving VKA. However, given the size and homogeneity of this pediatric study population and its restriction to warfarin and phenprocoumon, the results should be extrapolated with caution, particularly to non-white children, those with other or mixed races and ethnicities, and children receiving acenocoumarol (the third VKA in clinical use internationally).
The importance of age in determining pediatric VKA therapeutic dosing may be explained by the fact that weight-adjusted clearance of warfarin (like that of other conventional anticoagulant such as low-molecular-weight heparins) is greater in children than adults. This clearance decreases with age toward young adult values.4 Importantly, nearly two-thirds of the variability in therapeutic VKA dose in this pediatric cohort (and ∼ 40% in adults2 ) remains unexplained; it is likely that vitamin K intake, other dietary factors, and concomitant medications are key contributors. From a broader perspective, because the magnitude of association between INR and bleeding risk in children receiving VKA has not been established, future efforts should seek to determine predictors of bleeding and/or need for warfarin reversal in this setting, as a key clinical outcome of interest.
Contribution: N.A.G. and M.A.C. conceived of and wrote the paper and approved of its submission for publication.
Conflict-of-interest disclosure: The authors declare no competing financial interests. ■
