Thursday, August 27, 2009

Plavix Pharmacogenomics: To Test or Not to Test?

By now you've noticed that we are fascinated by the potential for a pharmacogenomic marker--variability in the CYP2C19 genotype--to play a pivotal role in reshaping a blockbuster market: the antiplatelet class currently dominated by Sanofi and Bristol-Myers Squibb's Plavix, but where Lilly and Daichii Sankyo hope make an impact with the newly approved Effient.

We've written about this extensively in "The Pink Sheet" (start here) and debated it during a podcast (available here, or via iTunes).

Two new developments this week have us thinking about it some more. First is the publication of arguably the most compelling data yet about the importance of the biomarker, a study by University of Maryland researcher Alan Shuldiner and colleagues that is the lead article in the Journal of the American Medical Association this week.

To recap, it seems that people with one of the variants of the CYP2C19 polymorphism are poor metabolizers of Plavix; clopidogrel is inactive as ingested, and so poor metabolizers may in effect be taking a placebo rather than antiplatelet therapy. Prasugrel (the active ingredient in Effient) is not metabolized by that pathway, and so should be effective in all patients regardless of genotype.

Given that the faulty genotype is present in something like 30% of the population (though the percentage varies by race)--and especially given that Plavix will soon be generic, while Effient is a key product for Lilly's future--the commercial implications are huge.

Now, the study itself isn't new per se. Indeed, it was Shuldiner's presentation of the data during an Institute of Medicine workshop in March (coupled with the reaction of Food & Drug Administration officials and other IoM panelists to the study) that first convinced us this could be a milestone in the evolution of personalized medicine.

As we point out in "The Pink Sheet" DAILY, publication of the study in JAMA (which is making the full article available for free) certainly ensures greater attention from the medical community. But will it change medical practice? That is the question raised in an accompanying editorial.

Which brings us to the second new development: the public release of FDA's approval letter for a labeling change for Plavix made back in May. That label change (done with no fanfare) added information about the biomarker to labeling for Plavix first time.

Two things in the letter are news to us:

(1) The label change was initiated at the request of FDA, using its new authorities under the FDA Amendments Act.

(2) FDA's initial request included a statement that PKG screening is recommended for all Plavix patients, but that was dropped after discussions with Sanofi.

FDA declined to comment on what prompted the change of heart on an explicit recommendation for the screening--but since the new FDAAA authorities give FDA for the first time the ability to mandate exactly the labeling it wants, Sanofi must have made some strong arguments along the way.

We suspect they pointed out some of the things included in the JAMA editorial, like the perceived lack of ready access to the test and (probably more importantly) the lack of evidence to guide any specific treatment decision following screening. That is, while the hypothesis that prasugrel will work better in patients who don't metabolize clopidogrel is appealing, that hasn't been tested in prospective clinical trials.

And then there are the inevitable questions about whether FDA can or should use labeling to drive change in practice, or instead make sure labeling best supports practice as it evolves.

Use of pharmacogenomic testing in the real world, LabCorp SVP Marcia Eisenberg points out, is really about "the comfortableness of physicians using molecular testing to change the pattern of their behavior.”

So, while infectious disease specialists have made HIV viral typing a routine part of therapeutic diseases, and oncologists are increasingly comfortable with testing for markers like HER2 and now KRAS, those are still very much the outliers. LabCorp alone offers dozens of potentially useful pharmacogenomic screening tests, but--with those few exceptions--most are not routinely ordered.

LabCorp (and other diagnostic firms) have made 2C19 screening available since 2005, and (in their view at least) there is no barrier to routine use. Awareness of the test is low, Eisenberg acknowledges; indeed, Shuldiner himself said during the IoM meeting in March that the test was not readily available outside of academic centers.

FDA, in Eisenberg's view, is helping to bring the overall science of personalized medicine to the forefront, but its regulatory actions inevitably lag behind actual practice. That was certainly the case with KRAS, where oncologists embraced the utility of the biomarker well before FDA approved new labeling for Amgen's Vectibix outlining the data in support of the test.

Will cardiologists do the same with Plavix? It doesn't seem likely right now. After all, the one group with a proven track record of changing behavior in cardiology is the pharmaceutical industry and its marketing prowess. But, as far as we can tell, none of the sponsors involved will be pushing for routine testing. Bristol and Sanofi made clear that they do not want to recommend the testing (indeed, they successfully prevented FDA from directing it in labeling), while Lilly and Daichii would rather not limit prasugrel to the roughly one-third of the population who don't properly metabolize clopidogrel.

That leaves the payors--but even there, the cost-benefit calculation will be complex, since routine screening would limit the population who would be eligible for generic clopidogrel, and may not be sufficient to stop prescribing of prasugrel by physicians who view it as the superior therapy.

The one thing we do know: this won't be the end of the discussion. There is plenty more data to come on the utility of the 2C19 marker in the class, and the looming patent expiration of Plavix means there are billions of dollars at stake. Anyone interested in the future of personalized medicine and blockbuster markets will want to keep paying attention.

1 comment:

Howard Coleman said...

Our sales force found a lot of initial interest among cardiologists in CYP2C19 testing last spring and it seemed like the testing was going to take off. When it came down to ordering though, they just weren't ready. Seems easy enough to figure out if you've got a 2C19 IM you'd up the dose and monitor for platelet function to protect against over medicating. For a PM you'd go to a different treatment.

At Genelex we've been doing this testing since 2000 and have interaction software that makes the results useable for physicians. It's become clear to us that prescribers are not the drivers for PGX adoption. In the case of plavix, the PBMs like Medco, may drive it when the drug goes generic because of the dollars at stake. A pointed relabeling or Medicare payment decision would help, but are roadblocked by the widespread insistence on unrealistic and unreasonable levels of proof. The real cynics think a lawsuit or celebrity death that clearly could have been prevented by the use of PGX would bring the technology into the mainstream.

We believe the evidence shows that variation in the three main cytochromes, 2D6, 2C9, and 2C19 underly the majority of patient variation to medicines seen in everyday medical practice and the widespread adoption of DNA testing for just these three enzymes will save tens of thousands of lives and billions of dollars every year.

Howard Coleman, CEO
Genelex Corporation
Seattle, WA 98121