On April 6, 2017 the FDA Center for Devices and Radiological Health formally approved 23andMe’s Personal Genome Services Test as a Class II Medical Device. Approved tests include assessment of an individual’s genetic risk of developing Parkinson’s disease, late-onset Alzheimer’s disease, celiac disease, alpha-1 antitrypsin deficiency, early-onset primary dystonia, factor XI deficiency, Gaucher disease type 1, glucose-6-phosphate dehydrogenase deficiency, hereditary hemochromatosis, and hereditary thrombophilia. The FDA did not approve the 23andMe tests for direct diagnosis of any disease or condition, and prohibits use of the test for prenatal testing, determining most cancer risks, identifying gene variants that affect drug metabolism, and assessing autosomal dominant conditions.
The approval was granted through the FDA’s de novo pathway, used for devices presenting low-to-moderate risk of harm to patients and where no substantially equivalent device is currently on the market. This approval follows 23andMe’s initial decision to enter the market without seeking FDA approval, which caused the FDA to issue the company warning letters in 2010 and again in 2013. The FDA demanded that 23andMe cease offering unapproved genetic testing services to the public and classified the 23andMe test as a device, rather than a laboratory-developed test, for regulatory purposes. Following this, 23andMe sought approval to sell the test for specific individual indications, obtaining approval in 2015 for a test on gene variants associated with Bloom Syndrome. Importantly, the 2015 approval established Direct-to-Consumer (DTC) screening tests to identify carriers of autosomal recessive disorders as Class II devices, clarifying what had been a key ambiguity in the approval of additional DTC genetic diagnostics.
The most recent FDA approval was limited to a narrow set of specific conditions. However, the approval also lays out a series of highly specific “special controls,” that may reflect requirements for approval of similar products going forward. The special controls include requiring statements to consumers detailing, in plain language, limitations on the predictive value of each individual test, the availability of genetic counseling, the role of non-genetic factors in the development of disease, and the desirability of discussing test results with a medical professional. The controls also require disclosure of technical information relating to the particular gene variants tested, as well as the accuracy and clinical performance of each individual test. More interesting, perhaps, are provisions requiring consumers to “opt-in” in order to receive results that may indicate a risk of developing a life threatening or highly debilitating condition, and a requirement that 23andMe provide data from a user comprehension study, described in detail in the approval letter.
The opt-in requirement is unusual, reflecting the FDA’s special concern with the question of whether genetic data in untrained hands is more anxiety-provoking than it is useful. Requiring a highly specific user comprehension study takes this concern a step further. In detailing the required user comprehension data, the FDA seems to be pointing out the importance of understanding the degree to which consumers are able to make sense of the genetic information that will be provided, rather than simply assuming that users lack the ability to understand or contextualize test results. This requirement may lead to both better-educated consumers, and a more sophisticated understanding of how best to present genetic and genomic information directly to consumers.
The coming wave of DTC genetic and genomic testing services will place an unprecedented amount of detailed health information into the hand of individual patients. While the path to market for these diagnostics has not been easy, and the present approval is tightly limited, it seems that we have finally seen the basic structure of what the FDA will require for approving these devices.
A copy of the approval letter can be found here.
