Skip to main content
Premium Trial:

Request an Annual Quote

FDA Defends Actions Against PGx Testing Labs As Necessary to Protect Public Health


WASHINGTON, DC – Health regulators from the US Food and Drug Administration yesterday defended their decision to crack down on bad actors in the lab community making pharmacogenomic tests claims without sufficient evidence.

At the same time, regulators speaking at the American Clinical Laboratory Association's annual meeting on Wednesday, suggested that since the FDA is still practicing enforcement discretion when it comes to lab-developed tests, the agency would allow labs to continue to sell PGx tests without its approval as long as they were making evidence-based claims. While it remains unclear to many industry players how regulators are weighing the evidence underlying PGx tests, FDA officials assured that they will work with the lab community to be more transparent.

At the meeting, FDA Commissioner Stephen Hahn discussed two topics of interest to the lab industry: the agency's efforts to expand coronavirus testing capacity during the current public health emergency and the oversight of laboratory-developed tests (see related story).  While discussing the latter topic, Hahn said the agency was committed to working with Congress and industry on advancing an oversight framework for diagnostics, but he also defended the actions FDA took over the past year-and-a-half to rein in labs that were making inappropriate claims in PGx testing.

"Certainly, not all pharmacogenomic tests are wrong," Hahn acknowledged.  "But some were and some of these mistakes were not minor ones. This both can endanger patients and erode confidence." As an example, he said some labs (without naming which ones) were making claims that certain PGx tests could tell patients they needed higher doses of opioids, "when there was no evidence to support it."

Hahn's comments at the meeting underscore a position the agency had taken on PGx testing well before he became commissioner in December 2019. The agency issued a safety alert in October 2018, in which it called out unapproved PGx tests and software used to make clinical recommendations as a public health risk. 

At the meeting, Elizabeth Hillebrenner, chief of staff and associate director for scientific and regulatory programs at FDA's Center for Device and Radiological Health, noted that the agency became concerned about PGx testing because members of the healthcare community told regulators that some labs were making worrisome claims and asked the agency to look into them. When officials from FDA's device and therapy divisions investigated, they found that some labs were making claims that couldn't be supported by evidence to the point that it was "impacting the safe use of the drugs."

"Seeing the safety issue, we felt it was our responsibility to communicate that with patients and providers," Hillebrenner said.  

The agency followed up this safety alert with a warning letter last April 2019 to Inova Health System's genomics lab for conducting PGx tests without its approval. Then, over the summer, the FDA began calling up labs and asking them to remove any mentions of drugs associated with PGx variants in reports. 

Some in industry pushed back, calling out the FDA's actions as lacking transparency and consistency. A group of stakeholders formed a coalition to ask the agency to stop attempting to shut down PGx labs and to make any new, substantive changes in regulations through notice-and-comment rulemaking, as required by law. 

Hahn acknowledged at the meeting that the agency's actions in the PGx space have been controversial, but maintained that the FDA needed to act. "Even some of the best developers sometimes can get it wrong. That’s precisely why we need an independent check," he said, adding that the agency wants to work with the lab community to strike the right regulatory balance.

Hillebrenner noted that subsequent interactions with the lab community have already resulted in changes. "We actually had a couple of developers say once we called them up, 'Oh, you're right, we actually don't have the evidence behind that,'" she said. "Some [labs] stopped offering some tests, or made some adjustments so their offerings were in line with the evidence behind it." 

During these discussions, the lab community also asked the FDA to be more transparent about how it was evaluating the evidence underlying PGx tests. Toward this end, the FDA recently released a list of PGx variants and grouped them into three actionability categories based on information from FDA-approved drug labeling, and to a lesser extent, evidence from the published literature. In putting together this list, the FDA recognized that drug labels do not always keep up with the science, Hillebrenner said, and that the agency needed to go further.

The three categories of variants are: PGx associations with sufficient evidence to allow their use in guiding therapy management; associations with data to suggest a potential impact on drug safety or response; and associations with data demonstrating only an impact on pharmacokinetic properties. The third category, according to the FDA, lists variants for which the association between the genotype and drug safety or response has not been established.

The FDA has emphasized the list is a first draft and is seeking public input on it. In response to stakeholders' calls for the agency to detail the evidence it relied on from the literature to put the list together, Hillebrenner assured that in the second draft of the list, the FDA will provide literature citations.

One of the consistent criticisms against the FDA from the time it issued its 2018 safety alert has been that the agency's threshold for evidence-based PGx claims differ from expert bodies, such as the Clinical Pharmacogenetics Implementation Consortium. CPIC to date has issued 24 guidelines on 20 PGx genes associated with 61 drugs, and labs widely use these recommendations to design their PGx test offerings.

In a statement, CPIC noted that while this list now shows more overlap between what CPIC and the FDA consider clinically actionable PGx information, there are also key differences. For example, CPIC recommends avoiding phenytoin for phenytoin-naïve patients with HLA-B*15:02, but FDA doesn't list phenytoin it its tables. There are also drug/gene associations where FDA's table directs practitioners to look to drug labeling for dosing recommendations, when the label doesn't contain such information but CPIC guidelines do.

Kelly Caudle from St. Jude Children's Research Hospital, who directs CPIC, described at the meeting how the expert group evaluates the evidence underlying PGx variants and categorizes them into four categories: strong, moderate, optional, or no recommendation. The last category, "no recommendation," denotes that the variant has insufficient evidence to guide clinical practice. CPIC initiated this category a few years ago, Caudle said, in response to some labs returning results for genes that the consortium didn't think were actionable. 

Caudle said the group is now comparing its list against FDA's list and will likely make some changes to "make sure we're not missing any on our list."

In reviewing commercial PGx tests, Hillebrenner said the FDA found labs that were selling tests for hundreds of gene-drug pairs and making claims even beyond CPIC guidelines. It's currently unclear whether the FDA expects the lab community to use this PGx variant list to guide test claims. In releasing the list, the agency said the table doesn't provide comprehensive information on gene-drug interactions, and it doesn't affect current regulatory policies or requirements. 

Hillebrenner said at the meeting that the agency plans to continue to practice enforcement discretion over LDTs. "It's not that we're saying all laboratory-developed pharmacogenetic tests have to come in [for regulatory review.] We're saying we have concerns if you're making claims beyond where the science is and we're going to call you out if you are," she said. "But we're not saying you have to bring all these tests in through the agency. We'd rather you make claims supported by sound science so people can use them."

Some labs have expressed interest in taking their PGx tests through FDA, and CDRH and CDER hope to advance guidance to help sponsors through the process. One software company in the PGx space is already in the process of submitting a 510(K) with the agency.

"We need to keep talking about this and we want to get it right," Hillebrenner said. "We want to make sure our patients have access to good care and providers have access to this information."