| Summary: | It has been estimated that, at any one time, more than one million people in the U.K. are taking the anticoagulant drug warfarin, for the treatment or prevention of venous thromboembolism. The incidence of life threatening haemorrhage, due to overdose, is approximately two per one hundred patient years. It is well known that there is great inter-individual variability in reaching and maintaining a therapeutic level of oral anticoagulants, in part, due to the combined effect of gender, age, body size and drug interactions. In recent years, single nucleotide polymorphisms (SNPs) have been identified, which significantly reduce the amount of warfarin required for an individual to reach a therapeutic level. Consequently, the optimum dose of warfarin can be predicted in a higher percentage of patients using an algorithm, which includes pharmacogenomic information rather than one with clinical and demographic data alone. The aim of this study was to create two rigorous, composite algorithms, one clinical and one pharmacogenetic, which combined as many influencing factors as possible, in an effort to improve the predictability of warfarin dosing beyond that of other published studies to date. The study was carried out using three groups of subjects, after obtaining ethical committee approval and informed consent. Group 1 subjects (n=12) were healthy, non-warfarin treated laboratory staff, whose DNA was used to optimise the DNA extraction procedure. Group 2 subjects (n=207) consisted of warfarin patients, who had had a stable therapeutic International Normalised Ratio (INR) for at least two months. A comprehensive list of clinical and demographic data was obtained from each patient, as well as DNA samples for SNP analysis of the VKORC1, CYP2C9 and CYP4F2 genes. Group 3 subjects (n=20) comprised of pre-warfarinised patients who provided the same data as in group 2. In addition, venous blood samples were obtained for the measurement of the baseline levels of the vitamin K dependent coagulation factors and albumin. The stable warfarin dose for each of the group 3 patients was obtained retrospectively, after several weeks of warfarin therapy. The two algorithms were then constructed using the data from a random selection of group 2 patients (n=160). These were then used to predict the warfarin dose of the remaining patients in the group (n=47). By plotting the predicted dose against the actual stable dose, the percentage predictability of the new algorithms was calculated. In addition, the predictability of eleven previously published algorithms, eight pharmacogenetic and three clinical, was calculated using the same 47 patients. The clinical algorithm from this study showed the lowest predictability (R2=0.188) when compared to the three published algorithms (R2=0.203-0.268). However, the pharmacogenetic algorithm was able to account for a higher proportion of the warfarin dose (R2=0.553) than any of the other eight published algorithms (R2=0.383-0.525).In the group 3 patients, no relationship was demonstrated between the warfarin dose and either the albumin levels or the baseline levels of the vitamin K-dependent coagulation factors, with the exception of factor IX, which showed a negative correlation.
|
|---|
