JAMA Letters Continue Debate Over Genotype, Clopidogrel Response

New correspondence in the Journal of the American Medical Association refutes a meta-analysis published late last year that concluded the CYP2C19 genotype does not predict clinical outcomes in patients taking clopidogrel. The letters, which appear in the April 11, 2012, issue of JAMA, caution that the findings are based on faulty calculations and therefore should not be used to make the case against genotyping.

However, study authors Michael V. Holmes, MBBS, MSc, Aroon D. Hingorani, FRCP, PhD, and Juan P. Casas, MD, PhD, all of University College London (London, United Kingdom), stand by their argument, as does Steven E. Nissen, MD, of the Cleveland Clinic Foundation (Cleveland, OH), who wrote an editorial accompanying the original paper in the December 28, 2011, issue of JAMA.

At the time, Dr. Nissen said the study countered the “irrational exuberance” for genotyping prompted by the US Food and Drug Administration’s black box warning on clopidogrel.

It All Adds up to No Effect

For the meta-analysis, Dr. Holmes and colleagues identified 32 studies involving 42,016 patients that focused on the link between clopidogrel therapy and platelet and/or clinical response to CYP2C19 genotype, which is involved in converting the antiplatelet drug to its active form.

Understandably, results changed according to the type of studies included in various calculations—21 looked at ACS and 8 stable CHD, with 3 not reporting CHD status—but the researchers ultimately honed in on a pooled analysis of data from 22 studies using clopidogrel alone plus the clopidogrel-treatment arms of 4 randomized trials (26,251 patients and 2,465 events). It showed that patients with any loss-of-function allele had a higher risk of CV events than those with wild-type or gain-of-function alleles (RR 1.18; 95% CI 1.09-1.28). However, when the studies were stratified by number of events, there was a trend toward no effect in larger studies, consistent with small-study bias. In fact, in the 4 largest studies reporting at least 200 events, the point estimate was attenuated (RR 0.97; 95% CI 0.86-1.09).

But Are We Looking in the Wrong Place?

Alan R. Shuldiner, MD, Mark R. Vesely, MD, and Adam Fisch, BS, all of the University of Maryland, Baltimore (Baltimore, MD), point out in their letter to JAMA that clopidogrel is most protective in patients receiving stents. The inclusion of patients with other indications may have masked the effect of genotype, they say, noting, “This effect might have been even more apparent if studies in which stent thrombosis was the primary endpoint were not excluded.”

In addition, Dr. Shuldiner and colleagues express doubt that small-study bias could be responsible because “sample size is not a determinant of effect size, only the variance around the estimate.” Many of these small studies happened to have higher proportions of stented patients, they add.

Dr. Holmes and colleagues respond with more data. Excluding studies involving patients who received clopidogrel for indications other than PCI yields a relative risk of 1.22 (95% CI 1.11-1.33), while adding 3 studies that assessed stent thrombosis also resulted in a “modest” effect estimate of 1.21 (95% CI 1.12-1.30), they report. Moreover, stratifying the studies by PCI or stent insertion revealed no heterogeneity.

In another letter, Jessica L. Mega, MD, MPH, and Marc S. Sabatine, MD, MPH, both of Brigham and Women’s Hospital (Boston, MA), along with Eric J. Topol, MD, of Scripps Translational Science Institute (La Jolla, CA), allege that the meta-analysis is weakened by its inclusion of studies that measured outcomes in patients no longer taking clopidogrel.

They note: “Testing the association of CYP2C19 genotype with responsiveness to clopidogrel in patients not taking clopidogrel is problematic.”

Depends on How the Evidence is Weighed

Beyond multiple methodological issues, Dr. Mega and colleagues say, the “data presented by Holmes et al do not support their conclusion that ‘there was no significant association of genotype with cardiovascular events.’” Namely, the paper’s treatment-only analysis found a 37% increased risk of MI and 75% increased risk of stent thrombosis, both significant.

The study authors counter that their conclusions “were based on the totality of evidence, and we caution against focusing on a selected subgroup of outcomes, particularly when there is evidence of selective outcome reporting and small study bias.” Again, they stress that the 4 randomized trials evaluating the genotype by treatment interaction failed to find an association.

CYP2C19 Does Not Act Alone

But CYP2C19 is only part of the story, assert Gerasimos Siasos, MD, MSc, PhD, Dimitris Tousoulis, MD, PhD, and Christodoulos Stefanadis, MD, PhD, all of the University of Athens Medical School (Athens, Greece), in their letter.

“Although these findings are interesting, other factors involved in clopidogrel response such as other genetic polymorphisms or cellular and clinical factors have not been completely evaluated,” they write. Different enzymes in the metabolic pathway might matter, for example, as well as the speed of platelet turnover at a cellular level. Whether a patient is compliant, has higher body mass index, is taking other drugs, or has diabetes might also play roles, they add.

Based on these complexities, the “management of clopidogrel remains a controversial issue,” Dr. Siasos and colleagues note.

Dr. Holmes and colleagues concede: “Multiple genetic and nongenetic influences on clopidogrel response exist, although these groups should be distributed evenly among CYP2C19 genotype groups.”

Genotyping Debate Carries On

Drs. Shuldiner and colleagues refer to Dr. Nissen’s editorial in their letter, criticizing his earlier call “for a prospective randomized trial before clinicians adopt CYP2C19 genetics into everyday practice. We contend that until the results of such studies are available, clinicians must evaluate the full body of evidence and use their best clinical judgment to decide which patients may benefit from testing.

“Other directives may be denying patients access to potentially life-saving alternative therapies,” they argue.

In a rebuttal, Dr. Nissen says, “I could not more strongly disagree.”

He cites the example of hormone therapy in postmenopausal women, once thought helpful based on observational evidence but then shown harmful by the randomized Women’s Health Initiative trial, as well as the use of surrogate endpoints that gave false hope for drugs meant to increase HDL cholesterol. “Platelet reactivity testing represents another unvalidated surrogate endpoint and cannot be a substitute for studies that measure morbidity and mortality, which are the outcomes of importance to patients and physicians,” Dr. Nissen comments.

“Advocates of pharmacogenomics want the standards for evidence-based medicine to be lowered, substituting belief for well-controlled randomized trials,” he states, contending that a misstep could be expensive. “With a cost ranging up to $500 for each test, treating patients based on genetic profiling could cost the healthcare system as much as $1 billion annually.”

Even so, Dr. Nissen concludes, “I continue to believe that pharmacogenomics represents a promising approach in cardiovascular medicine but must adhere to the same rigorous standards applied to other medical advances.”

 

---

Sources:
1. Shuldiner AR, Vesely MR, Fisch A. CYP2C19 genotype and cardiovascular events [letter to the editor]. JAMA. 2012;307:1482.

2. Mega JL, Topol EJ, Sabatine MS. CYP2C19 genotype and cardiovascular events [letter to the editor]. JAMA. 2012;307:1482-1483.

3. Siasos G, Tousoulis D, Stefanadis C. CYP2C19 genotype and cardiovascular events [letter to the editor]. JAMA. 2012;307:1483-1484.

4. Holmes MV, Hingorani AD, Casas JP. CYP2C19 genotype and cardiovascular events [author reply]. JAMA. 2012;307:1484-1485.

5. Nissen SE. CYP2C19 genotype and cardiovascular events [editorialist reply]. JAMA. 2012;307:1485.

6. Holmes MV, Perel P, Shah T, et al. CYP2C19 genotype, clopidogrel metabolism, platelet function, and cardiovascular events: A systematic review and meta-analysis. JAMA. 2011;306:2704-2714.

7. Nissen SE. Pharmacogenomics and clopidogrel: Irrational exuberance? JAMA. 2011;306:2727-2728.

Related Stories:

• Meta-analysis: Data Do Not Support Genotyping for Clopidogrel Response
• Genetic Variants Affect Clopidogrel Response Even When Doses Are Increased
• Cases Made For, Against Genotyping to Guide Antiplatelet Therapy

 

Click here for a listing of companies that provide support to the Cardiovascular Research Foundation, owner and operator of TCTMD.