May 2009
Feature Story

Karen Lusky

Personalized medicine thus far involves a small but growing list of life-saving drugs. Now there appears to be a new candidate in the wings: the antiplatelet drug clopidogrel (Plavix).

Recent research shows that clopidogrel takers with com­mon variations that reduce the function of CYP2C19 are more likely to suffer negative cardiovascular outcomes than those with a genotype conferring normal function.

CYP2C19 plays a key role in converting clopidogrel into the active compound that prevents adenosine diphosphate (ADP) from binding to platelet P2Y12 receptors. That action, in turn, inhibits the platelet aggregation that clogs arteries or stents, with potentially disastrous results.

The odds of having a genetic variation that impairs CYP2C19’s job performance aren’t exactly remote. The polymorphisms are seen in about 30 percent of whites, 40 percent of blacks, and 55 percent of East Asians, according to a study reported in the Jan. 22 New England Journal of Medicine (Mega JL, et al. 2009;360:354–362).

In the study, researchers at Brigham and Women’s Hospital and drugmaker Eli Lilly found that 162 healthy subjects with at least one CYP2C19 reduced-function allele (about 30 percent of the study population) had a 32.4 percent relative reduction in plasma exposure to the active metabolite of clopidogrel compared with noncarriers. “Carriers also had an absolute reduction in maximal platelet aggregation in response to clopidogrel that was 9 percentage points less than that seen in noncarriers,” the article says.

Pathologist John T. Brandt, MD, one of the study authors and medical fellow II at Eli Lilly, says that in the study, heterozygotes (intermediate metabolizers) for CYP2C19 produced less of the active metabolite—and homozygotes (poor metabolizers) even more so. “It’s like a genedose effect,” he says.

In examining a separate cohort of more than 1,400 subjects with acute coronary syndromes treated with clopidogrel, the Brigham and Lilly resear­chers found that carriers of a reduced-function allele for CYP2C19 had a 53 percent relative higher risk of death over 15 months due to cardiovascular causes, MI, or stroke, compared with their noncarrier counterparts (12.1 percent versus eight percent). And they had a threefold increased risk of stent thrombosis (2.6 percent versus 0.8 percent). The subjects represented a subgroup of a phase three clinical trial testing Eli Lilly’s antiplatelet contender, prasugrel, which is awaiting FDA approval.

Ninety-five percent of those categorized as carrying a reduced-function CYP2C19 allele in the outcomes study had a *2 allele, according to the report. But the researchers also looked at *3, *4, *5A, *6, *7, *8, *9, *10, *12, *13, *14, and *17, says geneticist Sandra Close, PhD, one of the study authors and a research advisor at Lilly. “Some of these variants are very rare and some are population-specific,” Dr. Close explains. “Some alleles, for example, are present in one percent of people of African descent but have not been in other populations.”

The *17 allele makes someone a rapid metabolizer of CYP2C19, resulting in about 30 percent upregulation of the gene, Dr. Close says. “If someone has *17, he or she does appear to have a slightly higher exposure [to the active metabolite] of the drug,” she says. But the results and differences for *17 were not statistically significant in the Brigham/Lilly study, though “there appears to be a trend.”

A study by German researchers published this year in the European Heart Journal found that clopidogrel-treated patients who carried the mutant *2 allele (*1/*1 or *2/*2) had a 1.5 percent cumulative incidence of stent thrombosis within 30 days of percutaneous coronary intervention; only 0.4 percent of those with homozygous wild-type alleles (that is, normal) had that complication (Sibbing D, et al. 2009;30:916–922). Patients with the CYP2C19 *2/*2 genotype showed the highest risk of stent thrombosis (2.1 percent).

But another study published in the same issue of the New England Journal as the Brigham/Lilly research found that only poor metabolizers (those with two CYP2C19 loss-of-function alleles *2, *3, *4, or *5), not intermediate metabolizers, had the higher cardiovascular event rate (21.5 percent versus 13.3 percent for patients who did not have these alleles) over a one-year followup period. Those with two loss-of-function alleles who underwent percutaneous coronary intervention had a 3.58 times higher rate of cardiovascular events (Simon T, et al. 2009;360:363–375).

The prospectively designed study included 2,208 consecutively enrolled patients in the French Registry of Acute ST-Elevation and non-ST Elevation Myocardial Infarction. One explanation for why the study didn’t find intermediate metabolizers to be at risk could be that it included a “heterogeneous population as seen in real life”—all consecutive patients with MI included in different types of institutions (university hospital, general hospital, private clinics) with different management and profile, including age and other risk factors, “which may have diluted some of the genetic effects,” says lead study author Tabassome ­Simon, MD, PhD, associate professor of pharmacology at the Assistance Publique–Hopitaux de Paris, Hospital Saint-Antoine, and Université Pierre et Marie Curie Paris 06.

In the study, Dr. Simon and her colleagues also looked at the impact of ABCB1, a gene associated with drug absorption, on clopidogrel response. And they found that patients with two CYP2C19 loss­of-function alleles and either one or two variant ­alleles for ABCB1 had the highest risk of cardiovascular events compared with normal metabolizers (homozygous wild-type ABCB1 and CYP2C19).

Of the study’s finding for ABCB1, Lilly’s Dr. Close says it’s “a very complicated gene.

“We don’t understand it as well as CYP450 genes. So the jury is still out, although the [French] researchers found a very interesting observation that needs further investigation.”

Dr. Close views 2C19 as “the majority of the genetic story” for clopidogrel, but notes there are likely to be other factors that have different levels of effect on response to the drug. But “2C19 has a strong genetic effect no matter what study or outcome you look at and no matter how you group or ungroup the patients.”

With evidence on CYP2C19 accumulating, many are beginning to consider the potential role of CYP2C19 genotype testing for patients who are candidates for clopidogrel.

Roche Diagnostics has an FDA-cleared test for CYP2C19: the AmpliChip CYP450 test, which includes CYP2D6 and CYP2C19. For the latter, the chip can identify *2 and *3, says Delena Carite, Roche’s marketing manager for genetics/oncology.

David Flockhart, MD, PhD, chief of the Division of Clinical Pharmacology at Indiana University School of Medicine, Indianapolis, says CYP2C19 testing for *2 and *3 covers about 99 percent of the U.S. population. Thus, the testing will “deal with nearly all the Caucasians, Hispanics, and As­ians. It probably only covers 95 percent of Africans but that’s still a good number.”

Dr. Close, however, points to about six or seven CYP2C19 alleles that have “well-established decreased function.” She thus believes comprehensive testing should be available to identify those alleles, even though they may be less common. Otherwise, you can call a person normal (*1, *1) by default when he or she may not be, she cautions.

AutoGenomics is hoping to offer an FDA-cleared comprehensive test panel. The company is collecting clinical data to submit to the FDA for a CYP450 2C19 assay on its automated multiplexing microarray platform, which will include genetic variants from *2 through *10, and *17, a variant associated with ultra-rapid enzyme activity, says Anand Vairavan, the company’s marketing manager for personalized medicine products.

ARUP Laboratories tests for seven CYP2C19 alleles (*2 through *8), says Gwen McMillin, PhD, ARUP co-medical director of pharmacogenomics. (ARUP may add *17 but not for clopidogrel testing.) Clinicians are starting to ask about tests for clopidogrel, though “most of the pharmacogenomics uptake has been really slow,” she says.

“It may be that people are looking for a test called Plavix Risk of Resistance... or something that’s more specific to the drug application. If we named it that way, maybe it would get more attention,” Dr. McMillin says.

Cardiologist Paul A. Gurbel, MD, and his research team first described clopidogrel resistance and response variability in 2003 (Circulation. 2003;107:2908–2913). Dr. Gurbel, who talked about optimizing antiplatelet therapy in a recent AACC audioconference, says the one thing genotyping can do before someone “puts a pill in his or her mouth” is provide a fair idea of whether the drug may be the right choice for the patient. Dr. Gurbel and colleagues presented the results of a study at the American College of Cardiology 2009 Scientific Sessions showing that patients who had the loss-of-function allele CYP2C19*2 had higher platelet aggregation while on clopidogrel and had an approximate two­fold increased risk for adverse post-PCI ischemic events.

But the study also showed that identifying CYP2C19 genotype did not identify all the patients at risk for poor response to clopidogrel, warns Dr. Gurbel, director of the Sinai Center for Thrombosis Research at Sinai Hospital, Baltimore, Md. “That’s because ... the drug is also metabolically activated by other CYP450 enzymes ... and the metabolite formation is also affected by absorption.”

Genotype testing also cannot identify patients with a normal CYP2C19 genotype who are taking medications that inhibit the enzyme’s function or affect drug absorption. The list of potential medications includes proton pump inhibitors, which, says Lilly’s Dr. Brandt, not only inhibit CYP2C19 but may also inhibit absorption of clopidogrel. (The FDA announced in January that it is working with Plavix makers Sanofi-aventis and Bristol-Myers Squibb to study the impact of genetic factors and other drugs, especially the PPIs, on clopidogrel effectiveness.)

Other drugs, Dr. Flockhart says, are known to inhibit CYP2C19. “The other two most commonly potent inhibitors prescribed are fluvoxamine, an antidepressant, and cimetidine, an H2 blocker.” And there are potentially many others.

Platelet testing can identify whether a person is responding as expected to clopidogrel. But the notion of just skipping the genotyping and doing platelet testing isn’t a sound one, in Dr. Close’s view. That’s because, she says, nearly half of clinical events in the first 15 months occur within the first 24 hours after percutaneous coronary intervention.

Dr. Brandt points to three potential ways to determine if a patient is achieving adequate levels of platelet inhibition. One of the options is traditional light transmission aggregation, which few institutions offer because it’s fairly specialized testing, he says.

As an alternative, a new Accumetrics Verify Now test uses a disposable cartridge that can perform ADP-induced platelet aggregation, he says. The cartridge “is designed to fairly specifically check for the effects of thienopyridines like clopidogrel, and Lilly’s [not yet approved] prasugrel,” he adds. “Studies show that [the VerifyNow test] correlates with light transmission testing quite well.” Two clinical studies have shown that patients with a low level of platelet inhibition measured by the device have an increased risk of cardiovascular events, he says (Price MJ, et al. Eur Heart J. 2008;29:992–1000; Patti G, et al. J Am Coll Cardiol. 2008;52:1128–1133).

The third method is called the VASP assay—or vasodilator stimulated phosphoprotein. “It measures the function of P2Y12, the receptor inhibited by thienopyridines [such as clopidogrel and prasugrel], more directly than aggregation studies.”

Dr. Close says “there are still disagreements in the platelet world in terms of the best platelet assay to use for testing people on Plavix.” In her view, more research is needed before definitive guidelines are issued on what assay to use and the level to look for related to the platelet response.

Even if such guidelines were in place, there’s another issue: Test findings showing someone isn’t responding well to clopidogrel leave clinicians without clear direction on how to proceed.

“It’s the same problem,” says hematologist Henny Billett, MD, who moderated the AACC audioconference, “that has affected aspirin for a long time. If you do platelet function testing and see a decreased response, you may be forced to act when you aren’t certain what the action should be.” Dr. Billett is professor of clinical medicine, professor of pathology, and associate chief of hematology at Albert Einstein College of Medicine, Bronx, NY.

One option is to stop clopidogrel to save money and avoid its side effects. But, Dr. Billett cautions, platelet aggregation response to clopidogrel is not an all-or-none phenomenon. “And we don’t really know if taking the drug in these less responsive patients is better than not taking the drug at all.”

Another option would be to increase the dose of the drug. Clopidogrel’s action is related to the level of the active metabolite, Dr. Billett notes. “But there is no study that says if we double the clopidogrel dose in a group, that group has much better outcomes without a significant increase in hemorrhage,” Dr. Billett says.

An international multicenter prospective clinical trial, GRAVITAS, could shed light on that. The trial is examining whether adjusting clopidogrel dosage using Verify­Now as a point-of-care assay will improve outcomes of patients after percutaneous coronary intervention with placement of a drug-eluting stent (Price MJ, et al. Am Heart J. 2009;157:818–824).

What about adding another drug? If the person were not already on aspirin, that might be an option. But, says Dr. Brandt, “virtually everyone on Plavix is on aspirin—it’s usually given as a dual antiplatelet therapy.”

Dr. Gurbel says some experts suggest managing patients who aren’t getting a good response to clopidogrel by adding cilostazol, another plate­let inhibitor. Korean researchers have shown that adding this drug to Plavix has resulted in fewer thrombotic complications after stenting, he says. But the drug is expensive and has flushing, gastrointestinal, and other side effects.

Other antiplatelet drug options may become available soon, which will truly create a new playing field for personalizing antiplatelet therapy. Lilly is awaiting the FDA’s decision on prasugrel, which Dr. Brandt says the FDA’s Cardiovascular Advisory Committee in February recommended for approval.

“Prasugrel is also a prodrug,” he notes, but “it’s converted to an intermediate metabolite by esterases. And that intermediate metabolite is converted to the active metabolite by any one of four different CYP enzymes.” That makes the drug less susceptible to variations in the function in any one of the enzymes, including 2C19.

AstraZeneca also has an antiplatelet contender, ticagrelor, in phase three clinical trials. That drug, like clopidogrel and prasugrel, targets P2Y12, Dr. Brandt says.

“Ticagrelor is an active drug,” explains Astra­Zeneca’s Jay Horrow, MD, MS, executive director for clinical development. Furthermore, “the first metabolic step for the drug produces a compound that’s equally potent to the drug.” And the drug’s metabolic pathways do not involve CYP2C19.

Ticagrelor is also reversible, he notes, in that as the drug is eliminated from the body, the platelets become functional again. (Clopidogrel and prasugrel inhibit platelet function for the life of the platelet.)

The company’s “target” is to get its application for the new drug to the FDA in the fourth quarter of this year, Dr. Horrow says.

Sinai Hospital’s Dr. Gurbel, who thinks the evidence is not yet definitive enough to recommend genotyping or platelet testing for those on clopidogrel, foresees “one-size-fits-all antiplatelet therapy coming rapidly to an end.” But at this stage, he says, none of the data are related to primary prevention.

The question, he says, is “what about people at high risk for coronary artery thrombosis, such as those with diabetes or a family history of aggressive coronary artery disease?” Dr. Gurbel predicts that perhaps testing will eventually be used to target these people with more potent antiplatelet therapy than aspirin before they have an acute MI or stroke or a stent placed. Dr. Gurbel can see this happening within five years if “we come along as quickly as we have been with platelet function measurements.”

Of course, beyond the “could you, should you” questions surrounding pharmacogenomics and platelet testing for clopidogrel is this one: Would clinicians widely adopt it as a routine strategy to guide prescribing?

After all, genotype testing for warfarin has been slow to catch on. But Dr. Close notes there are definitive health outcome and cost studies in the works for warfarin. “And if payers realize it saves them money to do the testing, the uptake may change,” she says.

Robert Epstein, MS, MD, chief medical officer at pharmaceutical benefits manager Medco Health Solutions in Franklin Lakes, NJ, which also does pharmacogenomic testing, predicts that as the understanding improves of how to use testing to identify who will benefit from which drug, “then the market will be shaped by the testing, as well as the economics.

“So, for example, if 70 percent of people are extensive [normal] metabolizers for CYP2C19 and evidence comes out showing that either clopidogrel or prasugrel works equally well—and one is in generic form, then testing can identify who can take the generic drug.”

Meanwhile, drugmakers like those discovering possible substitutes for clopidogrel are looking for ways to design around the need for pharmacogenomics. But even if they are successful, Dr. Flockhart says, new drugs in the pipeline are a “drop in the ocean—a tiny drop.” To personalize medicine, he says, there’s so much work to be done on drugs already on the market.