Genotype-guided dosing of warfarin through modeling and simulation
Current genotype-guided algorithms for warfarin dosing fail to deliver optimal performance in two aspects: 1) these algorithms are not able to achieve the same level of benefits in non-white populations, since they were developed based on multivariate regression analysis with mostly European/White data and did not include genetic variants found frequently in non-white populations; 2) these algorithms do not account for the dynamic dose/response relationship and were limited in their usefulness to guide dosing during the initiation phase, as the possession of variant VKORC1 and/or CYP2C9 polymorphisms has been associated with a more rapid attainment of target international normalized ratio (INR) and higher risk of over-anticoagulation even in genotype-guided patients. To address these shortcomings, we report on the novel use of a previously published kinetic/pharmacodynamic (K/PD) model to develop a warfarin dosing nomogram to be used across genotypes and ethnicities. Our approach leverages data from both ethnically diverse and European patients, while accounting for the differential dose/response behaviors due to VKORC1 and CYP2C9 genotypes. According to simulations, the utilization of our dosing nomogram could enable effective attainment of therapeutic INR within one week in both ethnically diverse and European populations, while maintaining uniform INR response profiles across genotypes. Furthermore, in silico clinical trial simulations using the K/PD model could be a feasible approach to help to further refine our dosing nomogram to be more applicable in the clinical setting and explore possible outcomes even before prospective clinical trials are initiated.