VKORC1, CYP2C9 AND CYP4F2 GENETIC BASED ALGORITHM FOR WARFARIN DOSING: PRELIMINARY RESULTS OF A PROSPECTIVE ITALIAN STUDY

Background: We have recently developed and published a pharmacogenetic algorithm, based on VKORC1, CYP2C9, CYP4F2 polymorphisms, age and body surface area (BSA). This algorithm could explain 55% of warfarin dose variability in Italian patients. Aim: To verify in a prospective randomized study whether our algorithm (PGX arm) has any clinical advantage over standard warfarin dosing (STD arm). Methods: We enrolled 103 patients with age >18 years, presenting atrial fibrillation as indication for ex-novo warfarin treatment initiation and with target INR of 2,5 (therapeutic INR range 2-3). At enrolment a peripheral blood sample was collected from each patient for genomic DNA extraction. VKORC1 (-1639G>A SNP), CYP2C9 (*1,*2,*3 alleles) (INFINITITM Analyzer,Medical System) and CYP4F2 (*1,*3 alleles) (Taqman chemistry) polymorphisms were analysed and INR was monitored on days 0,5,7,9,12,15 and 19 of treatment. Patients were randomized to receive initial warfarin doses either established according to the standard care (STD arm, n=46) or determined by the pharmacogenetic algorithm (PGX arm, n=46). Eleven patients didn't complete the observational period and resulted drop out. Results: patients randomized to the PGX arm had a lower number of INR above 4 (F=6.381, p<0.05) and a lower time spent at INR>4 (0.2% vs 3.1%) (F=21.9; p<0.0001) than patients in the STD arm, implying a reduced Relative Risk of over-anticoagulation (RR=0.074; 95% C.I. 0.017-0.310). We then subdivided patients in three classes: those expected to require low (<20mg/week), intermediate (20-40mg/week) and high (>40mg/week) warfarin maintenance doses. The lower risk of over-anticoagulation observed for patients overall was confirmed in patients requiring low doses of warfarin (RR=0.077; 95% C.I. 0.010-0.570). Moreover among these patients comparing the PGX arm to the STD arm a lower percentage of INR above 4 (0/30 vs 6/42) (F=4.675, p<0.05) and a reduced time spent at INR>4 (0% vs 13.2%) (F=13.88; p<0.001) were recorded. Among patients expected to require high warfarin doses a lower risk of under-anticoagulation and a higher anticoagulation control were recorded in the PGX arm: the fraction of the study time spent at INR<1.5 in the PGX arm was lower than in the STD arm (20.3% vs 28.2%) (F=4.06; p<0.05) and the fraction of the study time (days) spent within the therapeutic range was higher than in the STD arm (54.1% vs 39.2%) (F=10.42; p=0.0016). No significant difference were recorded considering time to stable anticoagulation in patients overall or subdivided in the three classes of warfarin dosing and when patients requiring intermediate warfarin doses were considered. Conclusion: these preliminary findings suggest that the application of the PGX algorithm might significantly reduce the risk of bleeding events and increase the fraction of time spent at therapeutic INR values especially in patients requiring very low and high warfarin doses respectively.