Introduction

Vitamin K antagonists such as warfarin, acenocoumarol, and phenprocoumon are commonly used for the prevention and treatment of venous and arterial thrombosis. The molecular target of these anticoagulants is the vitamin K epoxide reductase complex of which the precise composition is largely unknown. Recently, one component of this complex, the vitamin K epoxide reductase complex subunit 1 (VKORC1), was identified by positional cloning [1]. This VKORC1 gene is mutated in individuals with combined deficiency of vitamin K-dependent clotting factors type 2 or with warfarin resistance [1,2].

The required dose of antagonist in vitamin K antagonist therapy (VKA) to achieve a target level of anticoagulation is variable, in particular between individuals, but also within a single individual, and depends on, for example, dietary intake and variations in pharmacokinetics. Earlier this year, polymorphisms in the VKORC1 gene have been reported that explain up to 30% of the variability in the pharmacological response to VKAs [3–5].

Bleeding is an important complication of treatment with VKA. In The Netherlands, intensive efforts by anticoagulation clinics are aimed at monitoring treatment with phenprocoumon and acenocoumarol, the two VKAs that are registered for use. These efforts are largely focussed on maintaining the level of anticoagulation, generally expressed as the international normalized ratio (INR), within a specified target range. In these efforts, individual patient characteristics such as those related to inherited determinants of the pharmacological response are only indirectly taken into account. We set out to estimate the contribution of a C1173T polymorphism in the VKORC1 gene to the acenocoumarol and phenprocoumon dose requirement and to the bleeding risk.

Methods

For the present analyses we used DNA from a previously reported case-control study [6]. This study includes 110 patients classified as “severe bleeders” and 220 “non-bleeders” (more than 96% of the participants are Caucasian) who were selected from two anticoagulation clinics. Major bleeding was defined as a bleeding that was clinically overt and met one of the following criteria: associated with a haemoglobin drop of ≥20 g/l, hospitalization, blood transfusion of two or more units, intracranial bleeding, intramuscular bleeding, intra-articular bleeding, or intraocular bleeding. Bleeders and non-bleeders were visited at home, during which a questionnaire was completed and blood was collected for the preparation of plasma and DNA. Genotyping for the C1173T polymorphism in the VKORC1 gene was performed using a simple restriction enzyme digestion of PCR-amplified DNA. Details of the protocol are available from the corresponding author upon request. Statistical analyses were performed for acenocoumarol and phenprocoumon separately and combined. We calculated odds ratios as measures of relative risks from the contingency table as the exposure odds ratios (OR), with 95% confidence intervals (CI) based on the assumption of a Poisson distribution, according to Woolf [7]. The institutional review boards of the Academic Medical Center and the Leiden University Medical Center approved the study protocol, and all patients gave written informed consent.

Results

Table 1 shows the dose requirements for the same target INR categorized for the different genotypes in controls. The data confirm the claims that were made in three earlier papers [3–5]. In comparison to CC genotype carriers, the average dose of acenocoumarol or phenprocoumon was 15%–30% less in CT heterozygotes, and about 40% less in homozygous TT individuals. There does not seem to be a major difference between acenocoumarol and phenprocoumon in this respect. The results are similar in the cases (data not shown).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Relationship between C1173T Genotype and VKA Mean Daily Dose Requirement in Control Patients

https://doi.org/10.1371/journal.pmed.0020312.t001

Next we examined whether the C1173T genotype influenced the bleeding risk (Table 2). Overall the OR for major haemorrhage for carriers of at least one T-allele was 1.7 (95% CI: 1.1–2.5). When analysed separately, phenprocoumon seems to more strongly modify the bleeding risk of the C1173T genotype (OR = 2.6 [95% CI: 1.2–5.7 for T-allele carriers]), whereas in acenocoumarol users the genotype had little effect (OR = 1.2 [95% CI: 0.6–2.3]). When we add all controls together in the calculations of the odds ratios in order to obtain statistically more stable results (which is reasonable because genotype does not influence the prescription), the ORs for bleeding in T-carriers are 1.4 (95 % CI: 0.8–2.5) for acenocoumarol and 2.1 (95% CI: 1.1–4.2) for phenprocoumon.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Numbers of Cases and Controls for the Different VKORC1 Genotypes and Crude ORs for Bleeding under VKA Treatment

https://doi.org/10.1371/journal.pmed.0020312.t002

We examined the quality of anticoagulation in the various groups of genotypes and VKAs. The percentage of time in range was considerably higher in phenprocoumon treated controls (CC genotype 60% in range (95% CI: 45–75, CT/TT genotype 66% in range (95% CI: 62–71) than in individuals treated with acenocoumarol (CC genotype 41% in range (95% CI: 35–47, CT/TT genotype 46% in range (95% CI: 42–51), whereas genotype had no appreciable effect on time in range.

Discussion

The results, although based on a small sample size of individuals with bleeding, support the suggestion that the bleeding risk for T-carriers is higher in phenprocoumon than in acenocoumarol users. If this finding is confirmed in additional studies and extended to the more frequently occurring and clinically relevant cases of nonmajor bleeding, it may imply that CT and TT carriers should be preferentially treated with acenocoumarol despite the fact that the genotype-related excess bleeding in phenprocoumon users is not mediated through unstable anticoagulation that is detected by the INR, because genotypes were associated with bleeding but not with unstable INR values. In fact, stability of treatment is better in phenprocoumon than in acenocoumarol users, which is in agreement with previous publications [8]. This paradox—apparently good INR control by all measures, but higher bleeding risks in phenprocoumon-treated patients—has been reported before, in particular in relationship to minor bleeding [8].

At present we have no clear explanation for risk differences between the two coumarin anticoagulants. It is unlikely that these drugs have differential effects on the carboxylation status of vitamin K–dependent clotting factors. In support of this, levels of factors IX (which do not influence the INR) are similar between the two treatment groups (data not shown). More likely, a difference in bleeding risk may find an explanation in the vastly different pharmokinetics of acenocoumarol (biological half-life = 11 h) and phenprocoumon (half-life = 140 h). Receivers of anticoagulants are often frail people with significant health problems. Perhaps alterations in their health status that affect coumarin sensitivity are not effectively managed with a long half-life product.

The increased bleeding risks give support to the notion that genotyping for this polymorphism in the VKORC1 gene may go beyond an academic interest and may become standard for every individual starting on acenocoumarol and phenprocoumon treatment. It may help in the identification of those individuals who are at the highest bleeding risk and thus should be monitored more intensely. Such VKORC1 testing may need to be complemented with testing for Cyp2C9 polymorphisms, which also influence the pharmacokinetics of VKAs and for which there is also increasing evidence for a relevant relationship between genotype and bleeding [9].

Acknowledgments

We thank Dr. F.J.M. van der Meer from the Thrombosis Service Leiden and Dr. M.G.H. Remkes from the Thrombosis Service Amsterdam for providing access to the patients under their care. The FACTORS study was supported by The Netherlands Heart Foundation (#99.165). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. The corresponding author had access to all data in the study and had final responsibility for the decision to submit for publication.