Pharmacogenetic-guided algorithms to estimate personalized dose or individual responses to anti-thrombotic drugs

Abstract

Purpose

Genotyping seems to be regarded as less useful than expected for predicting the inter-individual variation in drug response. We aim to improve the predictive accuracy of genotyping by developing models that also incorporate certain non-genetic factors.

Study selection and results

The anti-coagulant warfarin is widely used to prevent venous thromboembolic events. Although frequent monitoring of the prothrombin time international normalized ratio (PT-INR) allows an appropriate maintenance dose to be obtained for most individuals, there are some individuals for whom it is difficult to achieve the target PT-INR even when warfarin dose is increased. The anti-platelet drug clopidogrel is typically used with aspirin to prevent cardiovascular events following percutaneous coronary intervention. However, the existence of clopidogrel resistance is a major concern in Asian populations owing to the high prevalence of deficient allele of the CYP2C19 gene, which encodes a major enzyme that produces the active metabolite. Individual response to these anti-thrombotic drugs cannot be accurately predicted based on genetic factors alone. We have constructed two algorithms, one that predicts the maintenance dose of warfarin and one that estimates individual responses to clopidogrel in outpatients without a device-based platelet function test. We applied Akaike's Information Criterion to evaluate the validity of these algorithms.

Conclusions

In addition to genotyping data, inter-individual variation in non-genetic factors, such as clinical laboratory data, should be considered to predict drug response more accurately in each individual.