Summary

The principal enzyme involved in coumarin metabolism isCYP2C9. Allelic variants of CYP2C9, CYP2C9*2 andCYP2C9*3, code for enzymes with reduced activity. Despiteincreasing evidence that patients with these genetic variantsrequire lower maintenance doses of anticoagulant therapy,there is lack of agreement among studies on the risk ofbleeding and CYP2C9 polymorphisms. It was, therefore, ourobjective to study the effect of the CYP2C9 polymorphismson bleeding complications during initiation and maintenancephases of coumarin anticoagulant therapy.The design of the studywas a population-based cohort in a sample of the RotterdamStudy, a study in 7,983 subjects. All patients who started treatmentwith acenocoumarol or phenprocoumon in the studyperiod from January 1, 1991 through December 31, 1998 andfor whom INR data were available were included. Patients werefollowed until a bleeding complication, the end of their treatment,death or end of the study period. Proportional hazardsregression analysis was used to estimate the risk of a bleedingcomplication in relation to CYP2C9 genotype after adjustmentfor several potentially confounding factors such as age, gender,target INR level, INR, time between INR measurements, andaspirin use.The effect of variant genotype on bleeding risk was separately examined during the initiation phase of 90 days afterstarting therapy with coumarins. The 996 patients withanalysable data had a mean follow-up time of 481 days(1.3 years); 311 (31.2%) had at least 1 variant CYP2C9 alleleand 685 (68.8%) had the wild type genotype. For patients withthe wild type genotype, the rate of minor bleeding, major bleedingand fatal bleeding was 15.9, 3.4 and 0.2 per 100 treatmentyears,respectively. For patients with a variant genotype, the rateof minor, major and fatal bleeding was 14.6, 5.4 and 0.5 per100 treatment-years. Patients with a variant genotype onacenocoumarol had a significantly increased risk for a majorbleeding event (HR 1.83, 95% CI: 1.01-3.32). During the initiationphase of therapy we found no effect of variant genotype onbleeding risk. In this study among outpatients of an anticoagulationclinic using acenocoumarol or phenprocoumon, having avariant allele of CYP2C9 was associated with an increased riskof major bleeding events in patients on acenocoumarol, but notin patients on phenprocoumon. Although one might considerthe assessment of the CYP2C9 genotype of a patient for doseadjustment before starting treatment with acenocoumarol, aprospective randomised trial should demonstrate whether thisreduces the increased risk of major bleeding events.