FDA Once Again Changes Course on LDTs; Eases Up on Pharmacogenomic Test Restrictions

Last week, FDA once again changed course in its approach towards regulating pharmacogenomic (PGx) tests.  We have blogged on this story several times before (see past blog posts here and here).  A brief recap follows.

On October 31, 2018, FDA issued a Safety Communication regarding PGx tests.  In this safety notice, FDA made sweeping statements regarding the risks these tests allegedly posed.  The safety notice barely noted that PGx testing could be extremely helpful to physicians in guiding decisions about what drug to prescribe.  In early 2019, FDA began contacting individual companies offering PGx services, asking them to cease including information about specific medications that could be affected by variants identified in the PGx test report.  One laboratory – Inova Health Systems – declined to do so.  FDA sent them a scathing Warning Letter, which included a broad assertion of authority to regulate all LDTs.

FDA continued to contact labs and software providers, creating massive confusion and uncertainty in the process.  In our experience, FDA told individual companies that they could not even include gene-drug information in PGx test reports that was derived directly from the FDA-approved drug labeling.  FDA also rejected the inclusion of PGx information from well-established authoritative third parties, such as the Clinical Pharmacogenetics Implementation Consortium (CPIC).  For those companies targeted, FDA refused to engage in meaningful dialogue or identify any specific violation of the law that the companies had committed.

FDA’s actions were heavily criticized (see, e.g., letter from the National Alliance on Mental Illness).  Multiple concerns were expressed, including interference with the practice of medicine; impeding dissemination of information of clinical value (as recognized by inclusion in approved drug labeling); inconsistency with FDA’s expressed goal of furthering personalized medicine through better understanding of individual genetic variation; and the utter lack of transparency in FDA’s process, in violation of the Administrative Procedure Act.  On January 9, 2020, we filed a Citizen Petition with FDA on behalf of the Coalition to Preserve Access to Pharmacogenomics Information requesting FDA reverse its position.  The Citizen Petition raised multiple legal and policy arguments, including that FDA’s suppression of truthful, non-misleading information violated the First Amendment.  The Citizen Petition also noted the significant dilemma FDA had created for laboratory directors, who are required under CLIA to ensure that test reports include pertinent information for test interpretation; under FDA’s restrictive approach, labs could report out the genetic variants that had been identified but could not include information about the potential clinical relevance of such findings to specific patients.

On February 20, FDA unveiled a new approach towards PGx.

FDA described the revised thinking as the result of a new “collaboration between FDA’s Center for Devices and Radiological Health and Center for Drug Evaluation and Research intended to provide the agency’s view of the state of the current science in pharmacogenetics.”  The announcement again asserted that some PGx tests are potentially dangerous.  At the same time, FDA acknowledged that PGx tests could play a useful role, stating “this type of testing offers promise for informing the selection or dosing of some medications for certain individuals.”  This document also announced that FDA was releasing a Table of Pharmacogenetic Associations (“PGx Table”), which lists gene-drug interactions that the agency believes are supported by FDA-approved drug labeling and/or “sufficient scientific evidence based on published literature.”

The agency opened a docket for public comment on the PGx Table.  FDA invited feedback on “specific pharmacogenetic associations that should or should not be included as the agency continues to update this table,” noting that the feedback should include the supporting rationale and underlying evidence that supports any new pharmacogenetic association proposed to be included in the list.

FDA’s announcement is a welcome change in FDA policy.   Even if there are instances in which companies make unsubstantiated gene-drug association claims, FDA’s effective ban on gene-drug information went way too far.  The agency’s acknowledgment that this information can be useful to doctors is a step forward, and a change from FDA’s prior communications, as is FDA’s acknowledgement that drug labeling is not the sole repository of scientifically valid PGx data.  FDA admitted that “not all supported gene-drug interactions may be found in current FDA labeling,” and specifically recognized that gene-drug interactions could be adequately supported by professional guidelines, such as CPIC, and scientific publications.

Nevertheless, FDA’s most recent corrective actions do not fully address the problems FDA unleashed back on October 31, 2018.  The agency once again has acted without first seeking public consultation or making formal requests for stakeholder input.  One consequence of this seemingly ad hoc and nontransparent approach is that the information in the new PGx Table reportedly includes some gaps that may have been addressed with public input.  For example, as noted by Teri Klein of Stanford University, a co-principal investigator at CPIC, there are instances in which the PGx Table advises physicians to refer to FDA labeling for PGx-specific dosing recommendations that are not in fact contained in the labeling (e.g., azathioprine).  In sum, the regulatory landscape for PGx testing is better than it was when we submitted the Citizen Petition about 45 days ago.  However, it is still worse off than it was on October 30, 2018.