August 29th, 2010

Genetic Substudies of Large Trials Question Value of Clopidogrel Genotyping

Genetic substudies across a broad range of large clinical trials that used clopidogrel raise questions about the clinical utility of clopidogrel genotyping. The substudies come from large and important trials like PLATO, TRITON-TIMI 38, CURE, and ACTIVE A.

A genetic substudy of PLATO finds that ticagrelor is superior to clopidogrel irrespective of genetic subtype. Therefore, according to Lars Wallentin and colleagues in their presentation at the ESC in Stockholm and in a simultaneous publication in the Lancet, “use of ticagrelor instead of clopidogrel eliminates the need for presently recommended genetic testing before dual antiplatelet treatment.” Wallentin et al. analyzed the role of CYP2C19 and ABCB1 polymorphisms in more than 10,000 patients enrolled in PLATO. Among patients with any CYP2C19 variation, the primary endpoint occurred less often in those treated with ticagrelor than in those treated with clopidogrel (8.6% vs. 11.2%). A similar pattern occurred with ABCB1.

In a Lancet press release, Wallentin said that the “findings emphasise that ticagrelor will be a simple and reliable treatment to further improve survival and reduce the risk of recurrences in almost all patients with acute coronary syndrome without the need for any specific tests of its activity before or during routine treatment.”

In a second study appearing in the Lancet, Jessica Mega and colleagues genotyped ABCB1 in nearly 3000 patients enrolled in the TRITON-TIMI 38 trial comparing prasugrel with clopidogrel. They also examined the combined effect of ABCB1 variants and CYP2C19 variants and found that a majority of people have a genetic variation that may reduce the efficacy of clopidogrel. In sharp contrast to the PLATO investigators, they conclude that “as clinicians, professional societies, and patients integrate information about genetic factors affecting the response to thienopyridines, the roles of both ABCB1 and CYP2C19 should be considered.”

In an accompanying comment in the Lancet, Betti Giusti and Rosanna Abbate observe that both studies “confirmed the independent role of the CYP2C19 loss-of-function alleles as a determinant of major adverse cardiovascular events in these patients on clopidogrel… whereas they did not show any effects of CYP2C19 genetic variants in patients on prasugrel or ticagrelor.”

In a third paper, published online in the New England Journal of Medicine, Guillaume Paré and colleagues genotyped patients who had been enrolled in the  CURE and ACTIVE A trials and found no reduction in the effect of clopidogrel among patients with a loss-of-function CYP2C19 genotype, although patients with a gain-of-function genotype received greater benefit from clopidogrel.

One Response to “Genetic Substudies of Large Trials Question Value of Clopidogrel Genotyping”

  1. In the 2 placebo-controlled studies (CURE and ACTIVE A), CYP2C19 loss-of-function variants did not modify the efficacy and safety of clopidogrel. Similarly, in the active control study (PLATO), CYP2C19 loss-of-function variants or ABCB1 TT variant did not modify the efficacy and safety of ticagrelor compared with clopidogrel. The TRITON analysis showed that variants in the ABCB1 TT and CYP2C19 loss-of-function allele are associated with adverse cardiovascular outcomes in patients treated with clopidogrel, but not prasugrel. Although prasugrel vs. clopidogrel results are not presented by Mega et al (at least not in the primary report), I find it interesting that amongst the CT/CC carriers for the ABCB1 variant, the MACE event rate was numerically higher in prasugrel-treated (8.7%) compared with clopidogrel-treated patients (7.8%). The data for the TT homozygotes appear to be going in the right direction (12.9% clopidogrel vs. 11% prasugrel).

    When one estimates the predictive accuracies of the genetic testing for major adverse CV outcomes (MACE), the positive predictive values (PPV) for CYP2C19 loss-of-function variants amongst clopidogrel-treated patients range from 8% in CURE, 10% in PLATO, 11% in TRITON and 22% in ACTIVE A. The corresponding PPV for ABCB1 TT variant ranges from 10% in PLATO to 13% in TRITON. In other words, the predictive accuracy of these genetic polymorphisms is low for MACE and even lower for bleeding and stent thrombosis (presumably because of low prevalence of these events). These data, therefore provide support for the recent ACC/AHA clinical alert on clopidogrel that concluded “the evidence base is insufficient to recommend either routine genetic or platelet function testing at the present time”. The overwhelming majority of events cannot be predictable solely on the basis of genetic testing. Thus, improvement in prediction of future cardiovascular events in patients receiving antiplatelet therapy will likely benefit from development of a global risk assessment score based on traditional demographic, clinical, and procedural risk factors, genetics, and biological information rather than any single test result.