Skip to main content
Premium Trial:

Request an Annual Quote

Clinical Utility of PGx-Guided Warfarin Dosing Not Proven in Mayo-Medco Study, Reviewer Concludes at Cardiology Meeting


By Turna Ray

ATLANTA – The final results of an observational pharmacogenomic-warfarin dosing study led by Medco and the Mayo Clinic do not support the clinical utility of genotyping for heart patients, according to a physician who reviewed the data presented this week at the American College of Cardiology's annual meeting.

In the study, presented at the meeting by Medco's chief medical officer Robert Epstein, researchers found that hospitalizations due to any cause and due to blood clots or excessive bleeding for heart patients taking warfarin dropped by approximately 30 percent when genetic information was available to doctors.

However, according to a review by Mandeep Mehra, chief of cardiology at the University of Maryland Medical Center, the design of the Medco/Mayo study did not sufficiently account for the fact that physicians, aware that their patients were being genotyped as part of a clinical trial, may have followed those patients more closely, leading to better outcomes.

Researchers in the Mayo/Medco study genotyped 896 patients for alterations in their CYP2C9 and VKORC1 genes, which account for the nearly 33 percent of variation seen with warfarin administration, and used this genetic data clinical information to dose the anticoagulant. The outcomes seen by the genetically tested group were then compared to the experience of a control group of 2,688 patients whose warfarin doses were managed using standard methods.

In order to conduct the prospective, observational cohort study, the pharmacy-benefit manager identified patients for genotyping within its medical and claims database, garnered physician approval and patient consent to participate in the trial, and sent representatives to patients' homes to collect a biological sample for analysis by Mayo.

The patients picked to participate in the study, between July 2007 and February 2009, came from 49 states and were insured by 29 plan sponsors within Medco's system. The participants were followed for six months to observe the effect of PGx-guided dosing on hospitalizations and thromboembolism/bleeding.

Of the physicians contacted about the study, approximately 75 percent agreed to allow their patients to participate in the trial, even though most doctors were not familiar with genetic testing to dose warfarin. After the first 21 days of the study, researchers noted that many of the genotyped patients deemed at 'moderate' or 'high risk' of being sensitive to the drug refilled their prescriptions in the dose recommended after factoring in their genetic data. Researchers also noted that the closer the genotyping was done to when patients got their first warfarin dose, the better their outcomes were.

However, Mehra, who ACC selected to review the data, pointed out some holes that may have confounded the study findings. "We did not see data with target INRs achieved, and proportion of target INRs achieved between the [genotyped and the non-genotyped groups]," Mehra noted, further questioning whether the 30 percent reduction in hospitalizations in the study was "the effect of genotyping [or] the effect of more closer attention to the patient?"

Ultimately, Mehra concluded that what Medco and Mayo demonstrated is "more likely the effect of [doctors paying] closer attention to their patients by sending them a piece of paper that asked them to change their dosing schedule, or consider changing their dosing schedule.

"This trial was not adequate to answer the question posed and most definitely the use of genotyping for warfarin was not established in this study, in the real world," Mehra said.

"In order to push the personalized medicine portfolio forward we need to ensure and demand well-constructed clinical trials," he added. "You're effort was clearly laudable, but for every one step we take forward, we shouldn't take two backwards"

[ pagebreak ]

Epstein acknowledged in his presentation that the study may be limited by several factors, including non-randomization, the use of administrative claims data, and the length of time it sometimes took to secure patient samples. Additionally, he admitted that since physicians knew their patients were being genotyped, the study could have been impacted by the so-called Hawthorne effect, the scenario in which study subjects improve behavior simply in response to being studied and not as a result of the intervention.

However, the adoption of genetic testing does require physicians to change their behavior to a certain extent. As such, the fact that physicians in the Mayo/Medco study may have followed their patients more closely because of genetic testing shouldn't be a mark against the clinical utility of the intervention, in Epstein's view.

"If, in fact, receiving genotyping information helps the physician pay closer attention to the patient, like make them do more INRs because it points out the person has a rare genotype and needs to be tracked closer, I don't think that's a bad thing. I think that's a good thing," Epstein said in response to Mehra's conclusions. "The incremental information you receive can only help patients achieve a closer stable dose."

By that logic, Mehra countered, genotyping may not have been necessary. "You could have as easily done a structured follow-up in titration … rather than actually genotyping the patient," he noted.

Medco launched the PGx-guided warfarin dosing study nearly four years ago, hoping the study results would encourage its payor customers to begin reimbursing for the test by showing it was clinically useful and cost effective [see PGx Reporter 12-06-2006].

Most private payors and the Centers for Medicare and Medicaid services have refused to cover PGx-guided warfarin dosing, demanding more robust clinical utility data in randomized-controlled studies. CMS will only cover genetic testing to dose warfarin if it is done for Medicare beneficiaries who are part of a prospectively designed, randomized-controlled trial that shows pharmacogenomics-guided dosing strategies improves health outcomes over standard dosing methods [see PGx Reporter 05-06-2009].

Announcing the results of the study, Epstien said in a statement that Medco's "health plan sponsors recognize the importance of genetic testing in improving patient outcomes and avoiding medical costs due to adverse drug events. If it costs a few hundred dollars for the genetic test but avoids the $13,500 hospital bill, it very quickly pays for itself." However, it is unclear if after Mehra's review of the Mayo-Medco study payors will be particularly swayed to cover genetic testing in this setting.

A previously reported meta-analysis from the University of Cincinnati concluded that it may not be cost-effective to perform genetic tests to guide initial warfarin dosing in "typical" patients with atrial fibrillation, but it may be cost-effective to test individuals at high risk for hemorrhage. The study was published in the Annals of Internal Medicine in January [see PGx Reporter 01-21-2008].

The US Food and Drug Administration has updated the labeling for warfarin twice, to provide physicians additional guidelines and encourage them to use PGx-guided warfarin dosing strategies when appropriate. In August 2007, the FDA updated the label for warfarin to note that people with variations of the genes CYP2C9 and VKORC1 may respond adversely to the drug.

Then, in January, the FDA updated the "Dosage and Administration" section of the label, which advises doctors that knowing "the patient's CYP2C9 and VKORC1 genotype information, when available, can assist in selection of the starting dose." The label also instructs healthcare providers to refer to a table containing stable maintenance doses observed in multiple patients having different combinations of CYP2C9 and VKORC1 variants. "Consider these ranges in choosing the initial dose," the label recommends [see PGx Reporter 02-03-2010].

According to Epstein, researchers from Mayo and Medco have submitted their data to the Journal of the American College of Cardiology for publication. The manuscript has been accepted by the JACC and is awaiting publication.