This article has been updated to clarify that the FDA is a member of the Clinical Pharmacogenetics Implementation Consortium, but only participates as an observer.
NEW YORK (GenomeWeb) – The US Food and Drug Administration's warning letter to Inova Genomics Laboratory last week contains language that may be interpreted as an attempt by the agency to regulate the practice of medicine, according to legal experts and medical professionals.
In the letter, the agency told the lab, part of Northern Virginia's Inova Health System, that it is illegally selling its suite of MediMap pharmacogenetic tests without FDA approval. Specifically, the agency accused the lab of marketing its tests for uses that aren't "established" and described in FDA-approved drug labeling.
"We are unaware of data establishing the relationships between the genotypes assessed by your tests and your assertions regarding drug response for multiple drugs," wrote Donald St. Pierre, deputy director at the Office of In Vitro Diagnostics and Radiological Health in FDA's device division. As examples, the agency cited MediMap pharmacogenetic tests for two depression drugs escitalopram (Lexapro) and sertraline (Zoloft), and told Inova that the relationship between the CYP2C19 genotype and response to these therapies "is not established" and "not described in the FDA-approved labeling for these drugs."
Inova has stopped offering MediMap tests while it responds to the agency. A spokesperson said Inova believes it was offering its tests in line with the FDA's longstanding policy of enforcement discretion for laboratory-developed tests and that its lab and LDTs meet federal lab standards known as the Clinical Laboratory Improvement Amendments (CLIA). But she added that the warning letter provided further clarification regarding the agency's views on the use of LDTs for pharmacogenomics testing. "Inova takes seriously FDA's warning letter and is considering this recent clarification to assess the appropriate path forward to address FDA's concerns," the spokesperson said.
However, FDA's language in the warning letter didn't sit well with experts in pharmacogenetics, medical professionals, and regulatory lawyers, particularly the parts that read like the agency is asserting itself as the sole arbiter of when drug/gene relationships can be adopted into medical practice. They pointed out that while FDA approval and labeling for drugs and tests do inform patient care, doctors use a lot of other information, including best practices, guidelines, published data, and their own experience to discern when it's appropriate to pursue a treatment strategy that's "off-label."
The agency's claim that it isn't aware of data to support the association between CYP2C19 genotypes and the mental health drugs struck Howard McLeod, the director of the DeBartolo Family Personalized Medicine Institute at Moffitt Cancer Center in Florida, as misleading. McLeod and others interviewed for this story pointed out that experts within the Clinical Pharmacogenetics Implementation Consortium have reviewed the literature and found "substantial evidence" that CYP2C19 genotypes impact patients' ability to respond to selective serotonin reuptake inhibitors, such as escitalopram and sertraline. An FDA official is listed as a CPIC member with observer status.
CDRH wants to control laboratory testing … [and] they only have so many different sticks they can throw at people, and this warning letter is one of them. But this was done in a disingenuous way.
"To say there is no data when there is enough data for CPIC to issue guidelines is really disappointing," said McLeod, who is widely recognized for his expertise in pharmacogenomics. "Because either they believe that, which means there is a lack of credible expertise [at the FDA's] Center for Devices and Radiological Health, or it means they're playing games, which is not what you want from a place like CDRH."
It remains to be seen how the lab industry and pathologists groups historically opposed to FDA oversight of LDTs react to this language. The American Clinical Laboratory Association and the Association for Molecular Pathology declined to comment for this article. But in the warning letter, the agency certainly gives industry players a lot to which they can object.
"The FDA seems to be coming down squarely on the idea that it is within its domain to decide" what are "established" drug/gene relationships, observed Jeff Gibbs, a lawyer at Hyman, Phelps & McNamara who represents life sciences firms on regulatory matters. "That is certainly something people can question in an area as fast moving as these drug/gene associations, and given the regulatory process, it can take a while before information makes it into drug labeling, if it ever does."
The FDA has published guidance telling diagnostics firms that tests that are essential for the safe and effective administration of a drug requiring premarket review and has provided drugmakers similar guidance on when pharmacogenomic tests data must be submitted to the agency. However, guidances are not legally binding documents, and besides, industry players have argued that the FDA regulatory process is too slow to keep up with the pace of advancing knowledge about drug/gene relationships, and too expensive for most diagnostics firms to pursue without the support of drugmakers. Unless there is a serious safety issue, pharma sponsors can be unwilling to return to the FDA to add PGx information to labeling for an already approved drug, and there is even less incentive to do so once a drug has gone generic.
Few labs have taken their pharmacogenetic tests through FDA review as companion or complementary diagnostics for specific drugs. In fact, the majority of tests on the market are performed as LDTs without FDA approval, because the agency has for decades practiced enforcement discretion over LDTs, leaving the Centers for Medicare & Medicaid Services to oversee lab activities in accordance with standards outlined in CLIA. Noting a shift in the way labs marketed tests, FDA has attempted to lift its enforcement discretion many times in the past and bring these tests under its oversight, but it has been repeatedly thwarted by interest groups representing labs and pathologists, who maintain that labs provide services that are not devices and therefore outside the agency's legal purview.
The FDA's warning letter to Inova, however, contains very strong statements from the agency describing its policy of enforcement discretion against LDTs, asserting that it can change its mind whenever it sees fit, for example, when it determines public health is at risk.
"I understand that CDRH wants to control laboratory testing … [and] I realize they only have so many different sticks they can throw at people, and this warning letter is one of them," said McLeod. "But this was done in a disingenuous way."
Under the Trump Administration, the agency for a time backed off its attempts to regulate LDTs, and former Commissioner Scott Gottlieb struck a favorable chord with industry by assuring that regulatory reforms for diagnostics would be established through legislation. But, in recent months the agency surprised and worried labs by suggesting legislative language that would essentially codify FDA's authority to regulate all clinical tests, even LDTs, and remove them from the market when there are public health concerns.
The FDA appears to be underscoring this point in the warning letter, and Gibbs opined that it may be a tactic to prod more labs to voluntarily bring their tests in for FDA review. "It may be a negotiating position to show that if you don't come to the table we're going to claim that we can regulate LDTs any time that we choose to," Gibbs said.
Since the FDA insists it has the authority to regulate LDTs as medical devices and protect the public health, one might wonder why the agency would expend the time and resources to try to create a brand new regulatory framework for diagnostics through legislation? "While FDA has the authority to regulate LDTs under existing medical device provisions, we believe a new framework tailored to in vitro diagnostics would better serve stakeholders," an agency spokesperson said.
The agency outlined its vision for a new regulatory framework for diagnostics in technical assistance to legislators last year, and those ideas have been incorporated into a draft bill, called the VALID Act. However, the FDA's hardline stance in the letter to Inova may not inspire collaboration among stakeholders as they negotiate the bill language.
"People in the lab industry will certainly take this [warning letter] into consideration when they're assessing what the FDA says about legislation and how they will implement or interpret the law," Gibbs said, noting that the strong language asserting the agency's authority over LDTs could make lab industry players less flexible and wary of giving the agency autonomy over their space. "It's not going to breed confidence or trust."
Encroaching on the practice of medicine?
The FDA has been inconsistent in its actions against LDTs being marketed without its approval, but in recent months it has turned its attention to pharmacogenetic tests. The warning letter follows a safety alert the FDA issued last year to convey its concerns to the medical community about the way pharmacogenetic tests were being sold. Subsequently, the agency reached out to several firms, including Inova, about their marketing of PGx tests. According to the FDA, most of the firms agreed to its demands to remove mentions of specific drugs from promotional materials and test reports, but Inova didn't.
The lab replied to the agency that its MediMap products are LDTs and that it is "properly operating within the scope of FDA's LDT exemption and thus is not subject to FDA's premarket review or labeling requirements." To this, the agency countered that "enforcement discretion" doesn't mean there is a regulatory exemption for LDTs and that it can change its discretion.
Industry insiders opined that Inova's unwillingness to address FDA's concerns initially and its aggressive marketing practices may have triggered the agency to take action. PGx testing has been a centerpiece of Inova's marketing messages to the Northern Virginia community it serves. The health system has invested heavily in its Center for Personalized Health and hired leading doctors, experts, and researchers. In 2016, Inova made PGx testing available to every baby born at its new women's hospital. The health system had also been running ads for PGx testing that may have been particularly visible to the FDA, located in the Washington, DC metropolitan region.
In the warning letter, the agency took issue with Inova's claims on its website that its MediMap tests can reduce trial-and-error prescribing, predict whether patients will respond to certain drugs, if they are likely to experience adverse reactions to treatments, and what the right dose is. Since Inova is making these claims for tests without FDA review, the agency is worried that doctors could make inappropriate treatment decisions that could harm patients.
The FDA also flagged the direct-to-consumer-type model Inova was operating — where MediMap tests were ordered by "lab physicians" and test results were provided directly to patients — which the agency said could lead them to change medications without a doctor's involvement. A 2018 presentation (see slide 20 in link) from an Inova executive contains a graphic describing the PGx test ordering process, and shows that the lab could provide results directly to patients, but they also have access to genetic counselors in the pre- and post-test setting and doctors.
The FDA's concerns about the potential for patient harm in the way Inova is marketing its tests are underscored by studies suggesting that primary care doctors have limited understanding of genetics. Even so, the language in the warning letter sounds to McLeod, Gibbs, and other experts who spoke on the condition of anonymity, like the FDA is trying to encroach on the practice of medicine.
While the FDA claims to have authority to regulate LDTs as devices, it does not regulate the practice of medicine, meaning doctors can prescribe a drug or use a device that they know works in other settings from published studies or other data sources, but not in FDA-approved labeling. If a lab or a group were to challenge the agency's efforts to regulate pharmacogenetic tests, based on the statements in the warning letter to Inova, "this idea of overreaching into the practicing medicine would be a central issue in such a dispute," said Gibbs, who at one time was FDA's associate general counsel for enforcement. He couldn't recall another instance when the agency had sent a warning letter about an LDT performed by a healthcare system's lab.
McLeod doesn't see much risk for patient harm at a place like Inova, which has invested in medical professionals with expertise in pharmacogenomics. "I don't mind the FDA regulating this area, but pick a place that's doing something wrong," said McLeod, noting that Inova's marketing only aims to make the community it serves aware of its PGx services. "How are you supposed to convey options to the people that use your health system? It's not like they're trying to hunt for new patients. They're trying to inform their existing patients."
Labeling as truth
What really got industry insiders worked up, however, is the agency's emphasis in the warning letter that FDA-approved drug labeling is the only source of "established" information on drug/gene relationships. Even when tests aren't approved by the FDA, labs can only make claims supported by FDA-approved labeling, the agency seems to be saying.
In answers to questions for this article, the FDA elaborated that in its November safety alert it was warning doctors about pharmacogenetic tests without sufficient evidence of clinical validity, but not about FDA-approved or -cleared companion and complementary diagnostics. The agency considers these PGx tests to have sufficient data supporting their use in clinical decision making because it has reviewed the evidence on the underlying drug/gene relationships.
"During FDA review of tests intended to predict a patient's response to specific drugs, FDA reviews scientific and clinical data to determine if the provided data support the claims being made about the relationship of the test and the listed drug(s) such that the use of the test is consistent with the safe and effective use of the listed drug(s)," an agency spokesperson said over email. "FDA would consider the relationship between genetic variants detected by the test and a claim for predicting drug response of specific drugs established if it is demonstrated that the test is safe and effective for its intended use for each listed drug."
As to whether guidelines from CPIC or another expert body can also be a source of "established" drug/gene relationships for doctors, the spokesperson said that the FDA "doesn't typically endorse consensus guidelines," but companies can use information in such guidelines as part of the data package they submit for test approval.
The agency's reasoning in this regard reminded Gibbs of a 1999 Washington Legal Foundation lawsuit against the FDA, challenging its ability to restrict drugmakers from disseminating information regarding off-label or unapproved uses of FDA-approved drugs. A district court determined that the FDA's policy of barring drugmakers from sharing truthful information with doctors about off-label use of drugs was an unconstitutional restriction of commercial speech. The district court judge deciding this case, Royce Lambert, even chided the FDA, saying the agency “exaggerates its overall place in the universe," because the "First Amendment is premised upon the idea that people do not need the government's permission to engage in truthful, non-misleading speech about lawful activity."
"I think people can go back and look at that Washington Legal Foundation case, and other subsequent First Amendment cases involving the FDA," Gibbs said. "I think that same kind of allegation could be made here since FDA is saying that only the agency can determine these drug/gene variant relationships and that CPIC or other authorities cannot. I think that someone could say that that, too, is an exaggeration of FDA's place."
Some other legal questions raised by the warning letter, include whether the FDA has the authority to regulate labs under existing laws and whether the agency runs afoul of the Administrative Procedures Act by taking action against a lab without having a policy in place and based on unarticulated criteria. "That can be seen as arbitrary and capricious conduct," Gibbs said.
I hope they put [the tests] right back up, and make CDRH substantiate that what they're doing is harmful.
More letters?
It's not clear if this letter is a one-off or other labs offering PGx testing without FDA approval will also get warning letters. A number of healthcare systems have implemented preemptive pharmacogenetic testing and incorporated clinical decision support within electronic medical records that alert doctors when a patient might be at risk for an adverse event or when the patient metabolizes a drug too slow or too fast and should be put on another drug. Just as Inova was doing (see footnote in link) with its MediMap tests, most healthcare systems offering PGx testing decide which tests to offer based on information in FDA-approved labeling and guidelines from pharmacogenetic consortia available to doctors.
Within the NorthShore University HealthSystem, which serves patients in the Chicago metropolitan area, doctors can order a multi-gene PGx panel for patients if they deem it to be medically necessary. The healthcare system has also been running a pharmacogenomics clinic since 2015, within which interested patients can learn how testing can be useful, as well as its limitations, and then decide with their healthcare providers if testing is right for them.
NorthShore hasn't been contacted by the FDA regarding its PGx testing. According to Mark Dunnenberger, director of pharmacogenomics at NorthShore's Neaman Center for Personalized Medicine, the healthcare system's aims to personalize patient care by integrating insights on drug response from PGx testing with a variety of other clinical tools and information. "To expect that pharmacogenomics alone is going to be a crystal ball is irresponsible," he said, noting that for example, in the case of mental health conditions, combining effective drugs with counseling has shown to improve patient outcomes.
Dunnenberger didn't claim to have any knowledge of Inova's marketing practices that prompted the FDA's warning letter, but he acknowledged that some labs are making marketing claims that lack scientific support. "I get where the FDA is coming from, and we need to make sure that we put pharmacogenomics in the appropriate light," Dunnenberger said. "There's no room for hype in this setting, and there's a lot of hype out there."
NorthShore, for its part, is careful in the way it presents pharmacogenetic information to doctors. Currently, information on specific drug/gene responses are incorporated into the healthcare system's clinical decision support tools. Such tools include system alerts based on drug/gene relationships in FDA-approved drug labels, which guide providers through the decision-making process with regard to ordering a genetic test or determining the appropriate therapeutic strategy based on existing test results. Healthcare providers can also access guidelines from expert bodies like CPIC on drug/gene relationships in the decision support system.
While FDA labeling and expert guidelines on PGx testing are incorporated into NorthShore's clinical decision framework, doctors ultimately rely on their experience, as well as the health system's shared protocols and best practices, to decide what testing to order and when. "At NorthShore, we leave it up to the practicing physician to practice medicine in ways that he or she sees fit, and if they want to use pharmacogenomics data to take care of the patient, they have access to it." Dunnenberger said. "We focus on facilitating the responsible use of PGx data."
He acknowledged, however, that it is challenging for doctors to sift through the rapidly changing published literature and discern which drug/gene relationships have sufficient evidence to be incorporated into patient care. "In our eyes, we think CPIC is the great gold standard for that," said Dunnenberger, who has been a member of CPIC since 2010 and is an author on a number of guidelines. "Obviously, the FDA label is another great place to go, but CPIC does a lot of great work and shouldn't be ignored."
As far as McLeod knows, Moffitt also hasn't been contacted by the FDA about the PGx tests that it offers its cancer patients. The center draws on FDA labeling, evidence from the published literature, and expert guidelines in deciding which tests to offer, and he said that the cancer center doesn't plan on changing what it is doing in reaction to the agency's action against Inova. "But,it has made us reach out to our colleagues at Inova and ask how we can help," McLeod said
For the time being, Inova has stopped selling MediMap tests and removed marketing language from its website. McLeod guessed that the agency's actions, at a minimum, could result in Inova changing its marketing language for MediMap tests, but he also wants the healthcare system to push back against the agency. "I hope they put [the tests] right back up," he said, "and make CDRH substantiate that what they're doing is harmful."