Clopidogrel: To Test or Not to Test? That Is the Question—Still

“Personalized medicine” holds great promise to help target therapies to achieve the best clinical benefit for patients. One area where cardiovascular disease and personalized medicine intersect is in the use of genetic and platelet-function testing for evaluating the antiplatelet effects of the antiplatelet agent clopidogrel. While some have begun early adoption of this approach and the US Food and Drug Administration (FDA)2 has recently called attention to this issue via a “black box” warning (1), caution has recently been raised with a study by Paré et al. (2), who found the value of platelet genetic testing to be less clear. An emerging consensus is that additional larger clinical-outcome studies are needed to specifically evaluate any potential strategy of genetic and platelet testing with antiplatelet therapy adjustment before this approach can be recommended for clinical use. The focus on clopidogrel began with a series of studies that showed marked variation in the level of platelet inhibition achieved with the standard fixed dose of this drug. Because clopidogrel is a prodrug, its metabolism in the liver to the active metabolite became a focus of research to explain this variation in platelet response. In particular, several studies showed that clopidogrel-treated patients who had a reduced-function allele of CYP2C19 (cytochrome P450, family 2, subfamily C, polypeptide 19) (most commonly the *2 or *3 allele, roughly 30% of patients) had a significantly higher risk of cardiovascular events than those who had a normally functioning CYP2C19 enzyme (3). Patients with reduced-function alleles were also documented to have lower concentrations of the active clopidogrel metabolite and lower levels of platelet inhibition (3). Thus, data appeared to support the full chain of events: A genetic polymorphism of the reduced-function enzyme can lead to reduced active metabolite, less platelet inhibition, and an increased risk of cardiovascular events. Several cautionary notes are worth discussing, however. Most of the studies that have linked genetic polymorphisms to an increased risk of cardiovascular events were observational studies that compared different groups of patients who were defined by their CYP2C19 status but who also differed with respect to many other factors. Therefore, some of the other factors might explain the observed difference in event rates between 2 groups (a classic issue in observational studies). Although multivariate analysis tries to account for these differences, we know that it often cannot sufficiently correct for these variables; thus, apparent differences in event rates can be “confounded” by other differences in patient characteristics. The other major caution is that the observed difference in event rates was far greater (roughly 50% higher event rates, and even higher hazard ratios) than the clopidogrel benefit seen in the randomized trials (20% reduction in events). Thus, one might ask how patients who have a reduced-function CYP2C19 allele, are taking clopidogrel, and therefore have a reduced level of inhibition, could have an increase in events greater than that if they were not taking clopidogrel at all (i.e., no platelet inhibition). This increase points to the likely confounding effects of other variables that explain the observed differences. A recent metaanalysis of the initial observational studies found that most of the studies had been conducted with patients who were treated by percutaneous coronary intervention, with a relatively high percentage using drug-eluting stents. In this setting, the clinical benefit of the thienopyridines was approximately 75% in the placebo-controlled trials with the older thienopyridine ticlopidine. Therefore, the observed 50%-higher event rate in patients with a reduced-function CYP2C19 allele seems more plausible. The FDA asked the drug manufacturer to do additional pharmacokinetic and pharmacodynamic studies. A study of 40 healthy individuals found reduced concentrations of the active clopidogrel metabolite and a reduced degree of platelet inhibition in the group they characterized as “poor metabolizers,” i.e., those who were homozygous for the CYP2C19*2 or *3 allele (4). Those who were “intermediate metabolizers” (i.e., heterozygous for a reduced-function allele) were not significantly different in the pharmacokinetic or pharmacodynamic assays. Consequently, the FDA issued a black box warning for clopidogrel [(1); full prescribing information available at (4)] that focused only on “poor metabolizers” (2%–4% of the population but up to 14% of Asians), which stated: Warn about reduced effectiveness in patients who are poor metabolizers of Plavix. Poor metabolizers do not effectively convert Plavix to its active form in the body. Inform healthcare professionals that tests are available to identify genetic differences in CYP2C19 function. Advise healthcare professionals to consider use of other anti-platelet medications or alternative dosing strategies for Plavix in patients identified as poor metabolizers. Of note, however, is that this warning and the recommendations were based only on the pharmacokinetic and pharmacodynamic data, not on clinical outcomes. Further complicating a clinician's response to this warning is that the FDA warning recommends using alternate doses of clopidogrel and other therapies, most of which would have been untested in large clinical trials and therefore would be “off label” for most indications. The above-mentioned observational studies, however, do not address the real clinical issue: Does a patient who possesses a genetic polymorphism for a reduced-function CYP2C19 enzyme exhibit reduced clinical efficacy for clopidogrel? Data that address this key question were recently published by Paré et al., who carried out a genetics substudy from the Clopidogrel in Unstable Angina to Prevent Recurrent Events (CURE) randomized trial of clopidogrel vs placebo in >5000 patients with acute coronary syndrome (2). The report also featured an additional analysis of approximately 1000 patients from another placebo-controlled trial of atrial fibrillation. Surprisingly (given all the prior studies), clopidogrel was equally effective in reducing cardiovascular death, myocardial infarction, and stroke in patients with a reduced-function CYP2C19 allele, compared with those with a full-function CYP2C19 allele (2). This finding also applied to the small number of homozygotes for CYP2C19 reduced-function alleles. Note that this study was conducted a decade ago, and approximately 15% of the patients underwent a percutaneous coronary intervention (none with a drug-eluting stent); consequently, we have no data on what would potentially be the highest-risk situation for evaluating the efficacy of platelet-inhibitory drugs. Therefore, we have randomized-trial data showing the clopidogrel benefit to be similar regardless of the CYP2C19 allele. On the other hand, a wealth of data demonstrates that individuals with a CYP2C19 reduced-function allele are at higher risk for cardiovascular events. How do we reconcile these discrepant findings? My current thought is that CYP2C19 status is certainly a factor, but not the only factor, in determining the clopidogrel antiplatelet response and that the many other factors will tend to “drown out” the effect observed in a large randomized trial. This idea comes from an interesting analysis of a study that correlated different clinical and genetic factors with the variation in the platelet-inhibition response (5). This analysis revealed that CYP2C19 status explained only 6% of the observed variation. Other factors that explained variation were obesity, diabetes, and hypertension. Therefore, one would want to incorporate both genetic and clinical factors in assessing whether an adequate response could be expected. What about just testing the platelets and seeing what level of inhibition was achieved? That would integrate the clinical and genetic factors and assess the phenotype: Was there adequate or inadequate platelet inhibition? The next (and final) step would then be to alter therapy for patients with an “inadequate” response. Pilot studies of platelet testing and increasing the clopidogrel dose have been promising, and a larger (but still modestly sized) study has recently been published (6). The GRAVITAS (Gauging Responsiveness with a VerifyNow Assay—Impact on Thrombosis and Safety) study used the VerifyNow assay (Accumetrics) to identify poor responders and randomized them to a standard 75-mg clopidogrel dose vs 150 mg daily for 6 months. A modest difference in platelet inhibition was seen with the higher dose: The percentage of nonresponders was 62% at 30 days after a percutaneous intervention with standard clopidogrel dosing and 40% in those who received the 150-mg dose. The 2 groups showed no difference in the rates of cardiovascular events, although the number of events was small (just 50), thus limiting the power of the study. Other trials with stronger interventions (use of more powerful agents, such as prasugrel) for the nonresponders are ongoing, so the final word on the strategy of testing and dose alteration is not yet in. So, what should clinicians and their hospitals do? At present, I believe selective use of genetic and platelet testing could be considered for some patients, with input from cardiologists who have an interest in this area of research. An example would be a patient who has experienced stent thrombosis; one would want to make sure to prevent a second episode in such cases. A reasonable approach would be to test whether platelet inhibition is low. If it is, one could use a more potent agent, such as prasugrel (or ticagrelor, if available). A second example would be patients at the extremes of body weight. In a very obese patient, is standard-dose clopidogrel adequate? Or, in a patient with a very low body weight, is standard-dose prasugrel too strong, and should the FDA-suggested 5-mg prasugrel dose be used instead? Platelet-function testing could determine whether the platelet response was adequate. Genetic testing could help determine whether genetics was a contributing factor to a low platelet response, if identified. As illustrated by the complexity of these few examples, however, widespread use cannot be recommended at present because clear algorithms for testing and altering treatment have not been fully validated. Thus, at our institution we have testing available on a limited basis and use it selectively in current practice, but we are carrying out more studies, including ELEVATE-TIMI56 (http://www.clinicaltrials.gov/ct2/show/NCT01235351). Fortunately, we will have data from additional randomized trials within the next year, so the full answer may soon be coming. US Food and Drug Administration Clopidogrel in Unstable Angina to Prevent Recurrent Events (trial) Gauging Responsiveness with a VerifyNow Assay—Impact on Thrombosis and Safety.