Tailoring Antiplatelet Therapy Based on Pharmacogenomics: Title and subTitle BreakHow Well Do the Data Fit?

Antiplatelet therapy has a prominent role in the treatment of a broad range of cardiovascular diseases, and data supporting the role of aspirin in secondary prevention are robust.¹ As monotherapy, the thienopyridine clopidogrel has also been shown to have modestly superior efficacy compared with aspirin in secondary prevention for patients with recent myocardial infarction or stroke or with established peripheral arterial disease.² When given in addition to aspirin, clopidogrel has been demonstrated to have an incremental benefit in patients with acute coronary syndromes, in those undergoing percutaneous coronary intervention (PCI), and most recently, in those who have atrial fibrillation but are not candidates for warfarin.^{3 - 4}

Variability in response to aspirin and clopidogrel were identified and offered as potential explanations for why some patients had recurrent ischemic events despite taking aspirin and clopidogrel.^{5 - 6} In the case of aspirin, much of the apparent “aspirin resistance” was attributable to factors such as noncompliance, and true biochemical resistance was present in no more than 1% of patients.⁷ With clopidogrel, however, some patients did have minimal ex vivo antiplatelet effect with conventional clopidogrel dosing. More recently, genetic polymorphisms have been described that may be associated with impaired pharmacokinetic and pharmacodynamic response to clopidogrel, with possible clinical consequences.^{8 - 12} These genomic analyses have used a candidate gene approach, examining single-nucleotide polymorphisms (SNPs) involved with metabolic conversion of the prodrug clopidogrel to its active metabolite that binds to the adenosine diphosphate (ADP) receptor.

In this issue of JAMA, Shuldiner and colleagues¹³ report the first careful genome-wide association study of clopidogrel response. The authors first studied a genetically homogenous population (the Old Order Amish), which minimizes the potential for confounding. They found that the loss-of-function cytochrome (CYP) P450 2C19*2 variant was associated with diminished effect of clopidogrel on inhibition of ADP-induced platelet aggregation. They then examined a separate population of patients undergoing PCI. Importantly, patients with this variant were twice as likely as those without the variant to sustain an ischemic event in the year following PCI. This elevated rate of ischemic events was largely confined to patients taking clopidogrel and who had the CYP2C19*2 polymorphism (vs those who did not have this polymorphism); the excess risk was not evident in patients not taking clopidogrel.

The authors demonstrated that the CYP2C19 polymorphisms were not associated with baseline platelet aggregation or response to aspirin but only affected aggregation in response to clopidogrel. Furthermore, there was a gene-dose effect, because persons heterozygous for the CYP2C19*2 allele had intermediate platelet inhibition after clopidogrel compared with CYP2C19*2 homozygous persons and noncarriers. These observations provide supportive evidence that the association between CYP2C19 and clopidogrel response may be real. Also, because the results of prior candidate gene studies and the current genome-wide association study approach are concordant, the likelihood of the association being true is high.¹³

Nevertheless, it remains possible that some proportion of the excess risk associated with this polymorphism is independent of clopidogrel and may have to do with another mechanism, such as impaired metabolism of other common cardiovascular drugs. More data from completed or future trials of clopidogrel in which genetic samples are obtained are needed to answer that question.¹⁴ Additionally, response to medication, much like cardiovascular disease, is a complex interplay between nature and nurture. Shuldiner et al¹³ found that a percentage of platelet response at baseline and an even larger percentage of platelet response to clopidogrel was heritable—but not the entire response. Diabetes, obesity, and acute ischemic syndromes all increase platelet aggregability.¹⁵

Thus, a genomic profile may identify some proportion of risk, but this proportion may change over time in a given individual, depending on the specific clinical scenario. This observation implies that even a future era of pharmacogenomic profiling could include a complementary role for point-of-care testing of platelet function.¹⁶ This appealing concept needs to be evaluated in prospective studies. Also, as Shuldiner et al¹³ rightly point out, because the presence of the potentially relevant polymorphism ranges from approximately 25% in whites to approximately 50% in East Asians, large studies in ethnically diverse populations are really necessary. Therefore, additional work is needed before routine testing for this CYP2C19 polymorphism can be recommended.

Beyond merely identifying risk, the major reason to perform pharmacogenomic testing would be to identify patients in whom an alternative antiplatelet approach would decrease ischemic events. Potentially, higher doses of clopidogrel could overcome impaired antiplatelet responsiveness. Novel ADP receptor antagonists, such as prasugrel (a more potent thienopyridine than clopidogrel but similarly an irreversible ADP receptor antagonist), ticagrelor (a reversible ADP receptor antagonist), or elinogrel (a reversible ADP receptor antagonist) could be used instead of clopidogrel.¹⁷ Alternative antithrombotic agents undergoing evaluation, such as direct thrombin inhibitors or factor Xa inhibitors, might supplant the need for dual antiplatelet therapy in some patients. Regardless of those developments, the allure of higher dosing of clopidogrel or selective use based on knowledge of genotype lies in the imminent generic availability of the drug.

If a certain genotypic profile precludes clopidogrel effect at any dose, however, this would be critical information for the patient and physician. One case series of highly selected patients with prior myocardial infarction who subsequently developed stent thrombosis has suggested that a strategy of increasing clopidogrel dose in carriers of CYP2C19*2 is time-consuming and largely ineffective at providing adequate platelet inhibition (although prasugrel was able to suppress platelet aggregation successfully).¹⁸ However, this does not necessarily mean that all CYP2C19*2 carriers who are doing well clinically while taking clopidogrel would benefit from an alternative therapeutic approach—only a prospective randomized trial could answer that tantalizing question.

The effect of polymorphisms on bleeding also remains to be elucidated. Lack of a relationship between bleeding and CYP2C19 polymorphisms has been explained as resulting from lack of statistical power in this and other studies. Although that is a possible explanation, an alternative hypothesis would be that the relationship between antiplatelet effect and clinical events is nonlinear and somewhat different for thrombosis and bleeding.¹⁹ Development of genotyping strategies to decrease risk of bleeding would be a major advance, and it remains possible that larger studies will elucidate such a relationship for antiplatelet agents.

Polymorphisms that predict anticoagulation response to warfarin have already been identified, but testing is not widely used because of a lack of large randomized outcome trials demonstrating that this approach actually reduces bleeding.²⁰ Similarly, the lack of prospectively defined algorithms to react to the presence of clopidogrel-associated polymorphisms should limit such testing at present. Although such testing currently is expensive, the cost will decrease and hopefully will coincide with supportive data. Furthermore, if such testing allowed use of a less expensive generic antiplatelet drug, the test might essentially pay for itself.

The study by Shuldiner et al¹³ moves closer to fulfilling the promise of pharmacogenomic testing in tailoring antiplatelet therapy to the individual patient. Strides toward individualized therapy have already been made in cancer care and human immunodeficiency virus treatment. Antiplatelet therapy seems well suited to a tailored approach, and future investigations should pursue this promising area.