Effects of CYP2C9 and VKORC1 polymorphisms and drug interactions on coumarin anticoagulation control

Abstract

Oral anticoagulants of the coumarin type are effective drugs for the prevention and treatment of thromboembolic diseases. However, therapy with these drugs is complicated by their small therapeutic range and their substantial variability in anticoagulant response. In this thesis the influence of several drug interactions and two genetic factors on the variability in coumarin response has been examined. Coumarins are metabolised in the liver via the cytochrome P450 isoenzyme CYP2C9, encoded by the homonymous gene CYP2C9. Several polymorphisms of this gene result in the production of an enzyme with decreased metabolic activity. In prospective follow-up studies we found that for acenocoumarol (231 patients) the possession of at least one CYP2C9*3 allele and for phenprocoumon (284 patients) the possession of at least one CYP2C9*2 or *3 allele was associated with 20-30 % lower dose requirements, with an increased risk of severe overanticoagulation and with a decreased chance to achieve a stable anticoagulation during a follow up time of maximally 6 months compared with patients with the normal CYP2C9*1/*1 genotype. The biological target of coumarins is the protein unit VKORC1, on the enzyme vitamin K epoxide reductase (VKOR), which is encoded by the homonymous gene VKORC1. In the above mentioned studies we also studied the effects of the frequently occurring C1173T polymorphism (> 40 % of Caucasian populations). In users of acenocoumarol the possession of a CYP2C9*2 or *3 allele together with the possession of a VKORC1 polymorphism was associated with severe overanticoagulation (3.8 times increased risk compared to patients with no or only one polymorphism). In users of phenprocoumon we found that the difference in dose requirements between carriers of a CYP2C9*2 or *3 variant allele and patients with the normal CYP2C9*1/*1 genotype was only significant among patients with the normal VKORC1 CC genotype, whereas these differences were smaller and largely not significant among carriers of a VKORC1 polymorphism. In patients using either acenocoumarol or phenprocoumon, the VKORC1 genotype explained a greater part of the variability in coumarin dose requirements than the CYP2C9 genotype. In a pharmacoeconomic study on the cost-effectiveness of CYP2C9 genotyping we found several scenarios in which genotyping preceding acenocoumarol therapy could be cost-effective compared with no genotyping. It is well known that coumarin anticoagulants are sensitive to coumarin interactions. In a prospective follow-up study with patients we found that management of the pharmacokinetic interaction of coumarins with the antibiotic co-trimoxazole by preventive or reactive dose reductions resulted in a significantly increased period of undertreatment during the first six weeks after the antibiotic course compared with several other frequently occurring antibiotics. In a case control study selective serotonin reuptake inhibitors (antidepressants) appeared to increase the risk of major non-gastrointestinal bleeding, but not of gastrointestinal bleedings, among users of coumarins. Finally, we found in another case control study that the less extensively explored antiplatelet drugs clopidogrel and dipyridamole, next to the more extensively explored antiplatelet drug acetylsalicylic acid, increase the risk of major bleeding among users of coumarins.