Skip to main content
Premium Trial:

Request an Annual Quote

PGx Utility, Cost Effectiveness Touted at AMP, Countering FDA Message on Evidence


NEW YORK (GenomeWeb) – Presentations at last week's Association for Molecular Pathology annual meeting suggested that despite newer areas like liquid biopsy and immunotherapy taking the limelight, there is still attention being paid to more legacy areas like pharmacogenomics, where test makers face different and potentially greater hurdles.

PGx has been a feature of the molecular pathology space for years but has struggled to gain a wide foothold in clinical practice due to challenges in establishing clinical utility or winning reimbursement.

During a corporate workshop day ahead of the AMP meeting last Wednesday, Thermo Fisher Scientific highlighted talks from two PGx providers that use its products in clinical practice, taking the opportunity to remind attendees of the suite of tools it provides to the space, including microarrays, quantitative PCR, and next-generation sequencing. A company representative said at the meeting that customers have run over 7 million samples to date using its qPCR solution in routine PGx testing, for example.

The Thermo-sponsored presentations — by Coriell Life Sciences President and CEO Scott McGill and by Betsy Bove, assistant lab director and vice president of laboratory regulation compliance at Genomind — took place with interesting, and perhaps unexpected timing in light of the US Food and Drug Administration's warning one day later regarding the use of unapproved pharmacogenomics tests. FDA had issued the note after authorizing the direct-to-consumer PGx analyses sold by 23andMe.

Beyond just cautioning consumers away from unproven PGx tests, the FDA safety alert specifically called out assays focused on psychiatric medications as lacking evidence. Genomind's Genecept, as well as other assays like the GeneSight test marketed by Myriad Genetics, are specifically focused on this class of drugs, and others, like Coriell's services, have also included this area.

In a statement in response to the FDA note, Coriell's McGill wrote that the company supports the FDA's position and believes the agency "acted in the best interest of the public."

"Genetic testing for medication response is a relatively new science, and while the evidence is extremely strong for some genes and correlating drugs, many other correlations are still emerging," he added. "It is important that consumers are made aware of these facts, and that healthcare providers exercise diligence in choosing a partner for important new pharmacogenomics information."

In its authorization of 23andMe, the FDA specified that the company's testing should not be used to guide medical decision-making, and results should be followed up by a clinical PGx test. In its cautionary follow up, it then advised the community not to use unapproved, unproven PGx information.

Neither Coriell's services, nor the Genecept assay from Genomind are FDA approved, nor are a host of clinical assays, raising the question of what, if any, tests the agency views as potential confirmation tools for a 23andMe result.

In the safety alert, FDA did not go so far as to recommend that the laboratory-developed tests in the space now seek its review. Instead, it advised genetic test manufacturers to assure that their test claims and reports are in line with information in FDA-approved drug labels about genotype/drug associations.

In this, FDA and PGx testmakers seem to be at odds. During the Thermo workshop at AMP, Coriell's McGill mentioned a widely used citation that the FDA includes in pharmacogenomic warnings on the labels of more than 200 medications. But he and Bove also described how their firms are hoping to make the case for a more holistic utility view of tests — seeking recognition and reimbursement of their assays as whole, rather than the piecemeal payment for specific drug-gene pairs that is currently in place.

"The way that PGx is reimbursed today is very anachronistic," McGill said referring to a gene-by-gene schema as opposed to reimbursement for the larger "purpose of the test."

"We have to create a voice that pushes back," he added, "that this is an intervention that when done appropriately has a significant outcome for both the system and the patient.

"We need to be advocates," Bove said.  

In his talk, McGill highlighted some of Coriell's takeaways from data it has collected in its business model of facilitating PGx services for large healthcare systems, pension plans, and employers — something he called "enterprise pharmacogenomics."

The firm's approach marries genetic information about drug metabolism with analyses of the variety of clinical factors that can also impact drug response. Coriell sends out saliva sampling kits, and contracts with a group of labs — most of whom use Thermo Fisher platforms — for genetic testing, which a group of company pharmacists then curate into a report to an individual's physician.

The company also spearheads education, promoting the availability and usefulness of getting the testing to individuals within whatever health system that has become its customer.

At AMP, McGill presented a case study from its work with a pension fund of about 34,000 retired teachers. On average, members were on 15 medications each, with expected prevalence of diseases like heart disease, high blood pressure, and diabetes.

"It’s a relatively medicated population, but not atypical by any measure," he said.

Based just on variant prevalence measures from population data, Coriell calculated that about 83 percent of the group would be expected to be taking at least one drug with "clear, credible" evidence for a genetic proscription.

The company has so far tested about 2,100 individuals from the cohort since January. In this subset, the company is seeing a reduction of both medical issues and healthcare costs versus a non-tested control group.

About 64 percent tested had an immediate medication change recommendation issued to their doctor, he added, and 94 percent of those recommendations were accepted by physicians, based on an analysis of claims data.

Following McGill, Bove led a presentation discussing the findings of recent studies by Genomind demonstrating the ability of its Genecept test to reduce utilization of healthcare services and related costs among patients with mental illness.

The latest iteration of Genecept is a an assay that covers 60 targets in 18 genes, using a combination of Thermo's OpenArray platform with the QuantStudio 12KFlex instrument, and a handful of individual qPCR assays run on the 7KFlex.

In a study published earlier this year in Depression and Anxiety, researchers working with the company used health claims data from Aetna to compare health services utilization by about 800 Genecept-tested psychiatry patients compared to a control group. The group reported that the tested patients experienced 40 percent fewer emergency room visits and 58 percent fewer hospitalizations compared to the control group.

Although Bove spoke before the FDA announcement last Thursday, she also touched on the question of relying on specific gene-drug approvals for assessing the utility of PGx testing.

"We can't afford clinical trials for every gene-drug pair," she said, adding that in certain cases, it may even be unethical to do so.

Based on industry comments this week, the question of whether PGx validity and utility should be hemmed by what is published in FDA-approved drug labels, or should be viewed through the lens of holism that Bove and McGill described at AMP, will remain up for debate.

In an email to GenomeWeb, FDA Spokesperson Deborah Katz said that physicians seeking to confirm or deny a variant from a 23 & Me PGx report should use an "independent pharmacogenetic test" which is "not indicated as only for preliminary testing," and "should use professional judgment in the interpretation of results."

But this leaves open the question of how the community should or shouldn't employ the host of PGx LDTs employing platforms from Thermo Fisher and other providers in the space, and offering information beyond just FDA-approved drug-gene associations. Genomind's Genecept and Coriell's testing are neither "indicated" nor "not indicated" because, as LDTs, they have not been reviewed by FDA.