A study led by researchers from Brigham and Women's Hospital on the effect of CYP2C19 testing on physicians' prescribing patterns has found that when doctors received patient test results for genetic markers associated with clopidogrel response through a pharmacy-benefit manager-initiated program, they only altered treatment decisions for poor metabolizers 20 percent of the time.
The study, published online this month in Circulation: Cardiovascular Quality and Outcomes, tracked the actions of doctors offered CYP2C19 testing through CVS Caremark's personalized medicine program for patients who had recent acute coronary syndrome or percutaneous coronary interventions and who were being treated with clopidogrel.
Among 1,291 patients who the PBM identified as having been hospitalized for ACS or having undergone a PCI and were eligible for testing, physicians for only 945 patients consented to participate. Of the 945 patients, 678 were then successfully contacted and offered the testing, and 499 actually completed testing.
Among those in this cohort of nearly 500, who turned out to be carriers of CYP2C19 reduced-function alleles and thus relatively poor metabolizers of the drug, only one in five had their treatment altered by their doctors based on test results, the study found.
"Clearly [clinicans] are confused. There is great enthusiasm for personalized medicine and the science of pharmacogenomics, but there is still a lot we need to do to help providers make decisions based on this information," Niteesh Choudhry, the study's senior author and a physician in the Division of Pharmacoepidemiology at Brigham and Women's told PGx Reporter.
In the study, Choudhry and his colleagues analyzed physicians' prescribing patterns within the context of a genetic testing program administered by Generation Health, CVS Caremark's genetic testing benefits management arm.
In June of 2012 CVS Caremark fully acquired Generation Health and told PGx Reporter at the time that it planned to continue its PGx programs tracking prescription claims for a subset of drugs in CVS Caremark's system, assessing which patients might benefit from a genetic test, and contacting the doctors to inform them of the availability of a genetic test.
Under the CYP2C19 program, doctors were offered the option of having their patients receive genotyping. Patients who agreed were sent a self-sampling kit, and the results of the testing were sent back to the doctors directly, along with information about the FDA boxed warning for clopidogrel, which advises healthcare professionals to consider the use of alternative antiplatelet medications or dosing strategies for those with particular CYP2C19 genotypes.
The researchers analyzed changes in physicians' prescribing for up to 120 days after reporting test results using prescription claims data from the program and looking for intensification of clopidogrel therapy or replacement of the drug with prasugrel.
The study yielded several important insights into the reception of CYP2C19 testing among physicians and patients. According to the results, among more than 6,000 patients who were potentially eligible for testing during the program, only about 2,600 of their physicians went so far as to provide clinical eligibility information about them to Generation Health to see if they could be genotyped.
Then, of 1,291 patients who met eligibility criteria for testing, only about half had physicians who agreed to test them. Among patients whose doctors were on board with CYP2C19 testing, 678 were successfully contacted and offered testing, 623 consented, and 499 completed testing. Less than 10 percent of the patients offered testing declined.
Reflecting the results of previous studies, about 30 percent of the patients tested were carriers of reduced function CYP2C19 alleles. According to the study authors, 3 percent were ultrarapid metabolizers, 65 percent were extensive metabolizers, 26 percent were intermediate metabolizers, 3 percent were poor metabolizers, and a little more than one percent were classified as unknown due to the presence of both a reduced function and an ultrarapid allele.
According to the data collected on physicians' prescribing patterns, patients who underwent testing were significantly more likely to have their therapy changed than those who did not undergo testing. And, those with reduced function alleles were significantly more likely to have their therapy changed than non carriers.
But, among those patients who were classified as poor metabolizers — potentially those at highest risk of experiencing adverse events on standard clopidogrel therapy — only 20 percent saw a change in their prescribed therapy: either an escalation of clopidogrel dose or a switch to prasugrel.
Physicians' reluctance to change their prescribing based on genetic testing results in the study may be due to either a lack of understanding or skepticism of the clinical utility of CYP2C19 genotype, which Choudhry said is legitimate in light of a lack of clinical utility data.
"We are not saying here that what clinicans were doing was irrational," Choudhry explained. "There may be some [irrationality] with people having drugs switched when they shouldn’t have potentially, but more than that … there is limited [science] about how to use genotype to guide therapy. So, there is still a huge open question about what to do.
"So it's not surprising that very few clinicians changed their action on the basis of this information," he said.
According to the study authors, the 20 percent rate may underestimate physicians' response to CYP2C19 testing when it's used in a more acute context, since patients in the study had already been on clopidogrel, and ostensibly clinically stable, for an average of 100 days, while the risk of adverse events for ACS or PCI patients is highest in the first 30 days.
Overall, Choudhry said, the results make clear that there are many steps along the "cascade" of integrating pharmacogenomic or other genetic testing information into clinical practice, and each offers a stumbling block.
"We need to communicate information, get it to providers, and then they need to know what to do with it. For the majority of this program, clinicians were just provided with this information: 'Here is the test result and here's what it means in terms of metabolism and here's what the FDA label says and… good luck!'," Choudhry said.
"If we are really going to think about this very important space in medicine, and what will be the future of medicine to some extent … we need very specific behavior changes and communication strategies to wrap around the evolving science itself."
Personalized medicine programs launched by PBMs like CVS Caremark have intended to do just that. However, data from this study suggests that such programs may not spur PGx testing adoption in the face of unclear clinical utility evidence. And adoption outside of the context of a PBM program might be even lower.
Before being acquired by Express Scripts, Medco reported data from a study, which found that PGx testing for warfarin and tamoxifen were 45 times and seven times higher, respectively, when doctors were approached by a PBM program compared to test adoption outside such programs. In addition to its programs to educate doctors about PGx testing, Medco was also conducting numerous studies to evaluate whether PBM-initiated testing programs would improve outcomes in patients.
"In order to maximize the benefits of pharmacogenomic testing, when [benefits] do exist, we clearly need to get information in front of providers in a meaningful way,” Choudhry said.
He and his team hope to continue to study the adoption and use of pharmacogenomic testing in a similar vein with other drugs and associated genetic polymorphisms.
"I think the clopidogrel story is not one that's going to evolve a whole lot further, especially with the newer antiplatelet agents now available," he said. "But there are other tests that continue to be of interest and that may be [more clearly] therapeutically relevant."
For example, he said, warfarin PGx testing is a great target, as well as SLOCO1B1 testing for statin drugs.
"As a non-geneticist, but a practicing doctor, [my perspective is that] we have lots of promising potential targets but not a whole lot of this clinical utility data," Choudhry said. "But the science will evolve, and in parallel with that we have to develop almost this consumer science about how we engage patients, providers, and communicate that information."