Background: Clinical trials of genotype-guided dosing of warfarin have yielded mixed results, which may in part reflect ethnic differences among study participants. However, no previous study has compared genotype-guided versus clinically guided or standard-of-care dosing in a Chinese population, whereas those involving African-Americans were underpowered to detect significant differences. We present a preclinical strategy that integrates pharmacogenetics (PG) and pharmacometrics to predict the outcome or guide the design of dosing strategies for drugs that show large interindividual variability. We use the example of warfarin and focus on two underrepresented groups in warfarin research.
Materials and methods: We identified the parameters required to simulate a patient population and the outcome of dosing strategies. PG and pharmacogenetic plus loading (PG+L) algorithms that take into account a patient's VKORC1 and CYP2C9 genotype status were considered and compared against a clinical (CA) algorithm for a simulated Chinese population using a predictive Monte Carlo and pharmacokinetic-pharmacodynamic framework. We also examined a simulated population of African-American ancestry to assess the robustness of the model in relation to real-world clinical trial data.
Results and conclusion: The simulations replicated similar trends observed with clinical data in African-Americans. They further predict that the PG+L regimen is superior to both the CA and the PG regimen in maximizing percentage time in therapeutic range in a Chinese cohort, whereas the CA regimen poses the highest risk of overanticoagulation during warfarin initiation. The findings supplement the literature with an unbiased comparison of warfarin dosing algorithms and highlights interethnic differences in anticoagulation control.