Genetic Testing Influences Post-PCI Antiplatelet Prescriptions, but Other Factors Also Play a Role

ORLANDO, FL—While pharmacogenomic information does influence antiplatelet prescribing patterns—especially for patients with loss of function alleles—clinicians appear to also rely heavily on clinical factors, according to new randomized data.

With evidence lacking that genetic testing influences outcomes, it’s appropriate today for clinicians to “incorporate the genetic testing with your clinical gestalt,” senior author Jay Giri, MD (Hospital of the University of Pennsylvania, Philadelphia), told TCTMD after the presentation of the ADAPT study yesterday at the American College of Cardiology 2018 Scientific Session. “The clinical practice of precision medicine is going to be about more than just precise genetics,” he said.

Giri and colleagues, including Sony Tuteja, PharmD, MS (Perelman School of Medicine at the University of Pennsylvania, Philadelphia), who presented the findings here, randomized 504 patients undergoing PCI and stenting to CYP2C19 genotyping in the cath lab or to usual care. Genotyping was used to identify patients who carried clopidogrel loss-of-function alleles. The ultimate choice of antiplatelet agent remained at the discretion of the treating interventional cardiologist, however, and patients were preloaded with an antiplatelet agent before the genetic information was available (average turnaround time was 1.4 hours).

Overall, prasugrel (Effient; Eli Lilly) or ticagrelor (Brilinta; AstraZeneca) were used more often in patients who underwent genotyping than in the standard care arm (30% vs 21%; P = 0.03). Specifically, the more potent agents were used more often in loss-of-function allele carriers who were genotyped than in “normal” clopidogrel responders (53% vs 22%; P < 0.001).

Interestingly, the rate of agreement between clinicians and what the genetic information recommended was 71%. Reasons for nonagreement among patients who were intermediate or poor clopidogrel metabolizers typically related to presence of stable CAD, cost, contraindications, and physician preference. In patients who were considered normal or rapid metabolizers, the disagreement typically arose because of disease characteristics, the patient already being on the therapy, ACS diagnosis, or recurrent events.

Patients who were on clopidogrel before hospital admission were two times more likely to stay on the drug regardless of genotype (OR 2.04; 95% CI 1.22-3.45). However, patients who were on ticagrelor or prasugrel prior to genotyping were 99 times more likely to stick with them (OR 99.3; 95% CI 13.2-744).

Post hoc analysis showed no difference in clinical outcomes, although the study was not powered to determine this, both Giri and Tuteja acknowledged.

What Really Guides Prescribing Habits?

“There’s more to making these decisions than just the genetics,” Tuteja said in a discussion following her presentation. “The genetics only explain 12% of the platelet aggregation results for clopidogrel nonresponse, so I think there’s a lot of other factors that go into deciding whether a patient will have an adverse event after a stent. That’s why we gave the physicians choice whether to switch or not.”

With societal guidelines recommending prasugrel and ticagrelor in patients with ACS, this trial will only add to the “controversy,” she adds. “We need more studies to show what to do in those normal metabolizers, especially in the context of ACS.”

Likewise, Giri added, when physicians in this study saw complex disease yet were told by the genetic information the patient was a good responder to clopidogrel, “they still went ahead in many cases . . . to use prasugrel or ticagrelor because it made them feel better.” Whether this makes the patient feel better, though, remains uncertain, he said.

Tuteja explained that they designed the trial in a randomized fashion because of the skepticism that surrounded pharmacogenomic testing when they began. “We didn’t want to just produce a study that said, 'Yep, the physicians followed the recommendations and all is good.' We wanted to do it in a very thoughtful manner that might provide some insight on what goes into antiplatelet prescribing behaviors,” she told TCTMD.

Panelist Kristin Patton, MD (University of Washington Medical Center, Seattle), commented that the ADAPT findings raise the “question of whether the doctors were continuing to prescribe out of habit versus because of real clinical decision-making. I think that would be an interesting area to look at in the future.”

The study was conducted at two hospitals, which allowed researchers to identify differences in agreement patterns between the two, Tuteja observed. “A lot of education” needs to accompany the continued expansion of genotyping across institutions, she suggested. Examples of that could be “live education or clinical decision support tools that come up in electronic health records to alert physicians not to prescribe certain drugs based on genotype,” she suggested.

‘Tip of the Iceberg’

Is a large randomized trial needed to determine whether genotyping has an effect on outcomes? Panelist Michelle O’Donoghue, MD (Brigham and Women’s Hospital, Boston, MA), told TCTMD she’s not sure.

“So many have been done, and yet they all have different pieces and approaches as to the recommendations that they're offering,” she said. “I would like to see what some of the newer studies show. There are a bunch that are percolating, but once this round of clinical trials wraps up, I think that it really comes down to a bunch of folks sitting down and offering guideline recommendations based on the observations so far.”

Along with cost, the biggest barrier to use now for clinicians is that they don’t seem to know what to do with the genetic information provided after genotyping, O’Donoghue said. “Certainly, it would need to be very clear in order for it to make sense for clinicians to use clinically.”

Paul Gurbel, MD (Inova Heart and Vascular Institute, Falls Church, VA), presented a separate poster at ACC 2018 today, also on CYP2C19 genotyping, that found a similar degree of agreement with regard to what the test recommended and what was ultimately prescribed. In his opinion, larger outcomes studies guided by genetic information would be useful. “This personalized therapy that can be delivered at the bedside promises to reduce bleeding and thrombosis as compared to a blanket one-size-fits-all mentality that is pervasive in our current practice,” he told TCTMD in an email.

Giri noted that with the research expected in the coming years, “there are going to be a lot of drug-gene pairs for a lot of drugs, both in cardiovascular medicine and outside of cardiovascular medicine.” To come up with outcomes data for all of these pairings would be “completely unfeasible,” he continued. The two things that clinicians need to figure out, Giri added, is if they will be able to make decisions based on a “trust in the mechanistic data that's represented by drug-gene pairing” and how to “adjudicate this concept of having pharmacogenomic data that tells you one thing [but might be in contrast] to what's been seen in a large, population-based outcomes trial.”

The whole concept of precision medicine is to target therapies to patients who will benefit, Giri stated. “But does that necessarily mean that physicians are going to feel comfortable following those targeted recommendations when we've all been trained to follow the conclusion statements of these large randomized trials?” he asked. “We're just at the tip of the iceberg by running an implementation study here, but it is one that I think is way, way bigger picture than just the CYP2C19 [genotyping].”