Pharmacogenomics & Precision Medicine

March 31, 2016

Study looks at doctors’ response to genetic testing

A new clinical implementation study from Vanderbilt University Medical Center measures how physicians have responded to the introduction of routine genetic testing to predict patient response to the commonly prescribed antiplatelet drug clopidogrel.

A new clinical implementation study from Vanderbilt University Medical Center (VUMC) measures how physicians have responded to the introduction of routine genetic testing to predict patient response to the commonly prescribed antiplatelet drug clopidogrel.

The study, appearing in Clinical Pharmacology & Therapeutics, involves the nuts and bolts of personalized medicine, where newly employed genetic data figure prominently in efforts to guide more precise and efficient therapy.

Clopidogrel (brand name Plavix) is used to prevent heart attack and stroke. Scores of observational studies have validated an association between certain genetic variants and poor or intermediate response to clopidogrel after a coronary stent is placed. When prescribing this drug, there is currently no clinical guideline within the cardiology community to genotype patients for their likely response.

The new study shows that physicians were often responsive to these new and unfamiliar clinical lab data.

Josh Peterson, M.D., MPH

“We found that physicians were far more likely to switch patients away from clopidogrel when genetic data characterized patients as poor metabolizers of the drug, as opposed to intermediate metabolizers. And that very clearly shows sensitivity to degree of genetic risk, and receptivity to clinical pharmacogenetic testing,” said study author Josh Peterson, M.D., MPH, associate professor of Biomedical Informatics and Medicine.

In all, 36 percent of patients who fit the genetic profile for compromised response to clopidogrel were switched by their Vanderbilt doctor to an alternative therapy.

In an analysis comparing various likely reasons that patients tracked in the study might have been switched by their doctors to an alternative therapy, the single factor that loomed greatest, posing an eightfold increased likelihood of a switch, was a genetic test result forecasting poor clopidogrel response.

“I think at Vanderbilt we have certainly pioneered the integration of pharmacogenomics into our clinical systems, and have begun to change the culture of medical practice here. But like any new technology, there are barriers to using it effectively,” Peterson said.

In this case, “We found some physicians were happy to have the genetics, while others thought it was premature.”

A coronary stent is a small tube implanted to prop open a blood vessel. Patients receiving stents are at risk for blood clots, and antiplatelet therapy is a routine measure to decrease this risk. Aspirin is the most commonly used antiplatelet drug. Patients receiving stents are routinely given an additional antiplatelet drug on top of aspirin, and clopidogrel is the current standard treatment.

Two newer alternatives to clopidogrel are available, metabolized through other channels and not affected by known genetic variants.

The patients followed in this study received stents at VUMC between October 2010 and March 2013. These were the first three years of the PREDICT pharmacogenomics program. Under PREDICT, patients receiving or predicted to receive certain drugs are genotyped to determine their likely drug responses. Clopidogrel was the program’s first drug — the first to be associated with a genetic test result in the medical record, and the first to be subject to pharmacy surveillance and automated inpatient clinical decision support.

In the study, 514 patients who received stents and were prescribed standard dose clopidogrel were characterized by the PREDICT genetic test as more likely to benefit from an alternative antiplatelet therapy. This included 64 patients characterized as poor metabolizers of clopidogrel and 450 characterized as intermediate metabolizers.

The PREDICT test results were posted to the medical record. For all but 33 of these patients, pharmacists also notified attending physicians of the test result. Also, in 134 of these patients, when physicians used the inpatient order entry system to initiate a clopidogrel order, they saw an automated alert about the patient’s PREDICT test result.

Over a 12-month period, 58 percent of patients who tested as poor metabolizers of clopidogrel with two copies of defective gene, and 33 percent of those who tested as intermediate metabolizers, with one defective copy, were switched by their Vanderbilt doctor to another antiplatelet therapy.

For 304 of these patients, pharmacists took the added step of suggesting an alternative therapy to the prescribing physician. This was ventured after confirming that the patient wasn’t documented to have contraindications for other antiplatelet drugs. Over 12 months, 130 of these 304 patients, or 43 percent, were switched by their doctor to an alternative therapy.

Among 13 high-volume cardiologists, for patients who tested as poor or intermediate metabolizers of clopidogrel, one doctor switched to an alternative antiplatelet in 23 percent cases, another in 68 percent of cases, and the remaining 11 doctors were in between.

“Not all physicians were ready to adopt this as standard practice, as some are waiting for randomized controlled trial data. However, one of the dilemmas of precision medicine is that that type of evidence is not feasible to generate for every biomarker or combination of biomarkers that might impact care,” Peterson said.

“A second important issue is that many other factors influence physician prescribing. Like cost: if physicians switch to a brand name medication, one without the genetic concern, and the patient’s co-pay doubles, triples or more, then the fear is that the patient would discontinue their antiplatelet drug prematurely.

“Also, the alternative medications come with their own risks, which need to be considered before switching.”

Kevin Johnson, M.D., Cornelius Vanderbilt Professor and chair of Biomedical Informatics, was among the study authors.

“This study reminds us that it is not only important to present the genomic status of the patient; it is also important to educate providers about these test results and their implications for patients,” Johnson said.

Peterson said physician response to new routine genetic testing should be expected to vary for different drugs. At a conference this spring, research will be presented regarding Vanderbilt physician responses to routine genetic testing for response to the anticoagulant drug warfarin. According to Peterson, this warfarin study shows a high rate of acceptance among physicians for using genetic factors to calculate medication dose.

Joining Peterson in the study were Julie Field, Ph.D., Kim Unertl, Ph.D., Jonathan Schildcrout, Ph.D., Daniel Johnson, Pharm.D., Yaping Shi, M.S., Ioana Danciu, M.S., John Cleator, M.D., Ph.D., Jill Pulley, MBA, John McPherson, M.D., Josh Denny, M.D., Michael Laposata, M.D., Ph.D., Dan Roden, M.D., and Kevin Johnson, M.D.

The study was supported by grants from the National Institutes of Health (HL122904, HL105198, HL065962, HG006378, HG007253, TR000445) and the Centers for Disease Control and Prevention.