On Thursday, April 6th, the FDA announced that it will allow the direct-to-consumer (DTC) genetic testing company 23andMe to market “Genetic Health Risk” (GHR) tests for 10 diseases or conditions including early-onset Alzheimer’s and Parkinson’s Diseases. This is in addition to 23andMe’s current offering of ancestry, wellness (e.g., lactose intolerance), trait (e.g., hair color), and autosomal recessive carrier screening (e.g., sickle cell anemia) test reports.
The decade since 23andMe entered the market has been a regulatory labyrinth of twists and turns. But what direction are we headed now?
The way we were
23andMe was a pioneer of the field, entering the DTC genetics market in 2007 with a product offering 13 health-related reports for $999. By December 2013, it was offering more than 250 reports; including carrier status, drug response, and over 100 GHRs. In response to a set of FDA Untitled Letters that went out in 2010, 23andMe filed for de novo 510(K) premarket clearance for some tests… but also concurrently marketed them in a national television and web campaign.
In November 2013, 23andMe received the dreaded FDA Warning Letter in which the Agency berated the company that despite “more than 14 face-to-face and teleconference meetings, hundreds of email exchanges, and dozens of written communications” it continued to market an adulterated and misbranded product. (Note that 23andMe General Counsel had left about 6 months prior, and the legal team was reporting to Andy Page MBA, the former President of Gilt Groupe. Hard to prove that was causal, but still a good example for your next pitch).
In February 2015, 23andMe re-entered the health report market with FDA-authorization for the autosomal recessive carrier-screen (i.e. doesn’t affect parent, and both parents would have to be carriers for child to have ¼ chance of getting the disease) for Bloom Syndrome, as well as a Proposed Rule/Notice of Proposed Rulemaking (NPRM) to to exempt all such carrier screens from needing premarket authorization in the future. Since both parents would have to receive false-positive test reports in order to even worry that their child would have a ¼th risk of disease, FDA concluded that warning statements were likely enough for such a low-risk test. And that’s where it’s been since.
The way we are
There are a couple of important things to note about this new authorization. First, the FDA authorization is tailored to predictive risk reports, not diagnosis. That sounds like an important distinction, but for the fact that almost all genetic testing is about risk—how your genetic predisposition operates in conjunction with things like health behavior, environment, and clinical symptomatoloy. For example, 23andMe just sent me a new health report regarding my weight, telling me that my “genes predispose you to weigh about 8% more than average.” I don’t, which—23andMe informs me in a handy graphic—is likely due to lifestyle choices such as exercise, abstaining from fast food, and eating my leafy greens. Risk avoided.
But, arguing that testing for genetic variation affiliated with Parkinson’s Disease and late-onset Alzheimers tests are predictive tests for risk, and that testing for genetic variants associated with hereditary breast and ovarian cancer (BRCA1 & BRCA2) are diagnostic tests, is a bit oversimplified.
The purpose of predictive tests is not to definitively “predict” what is going to happen. They are probabilistic in nature, and therefore can only tell you what might happen by recognizing mutations that can increase your risk of disease. Predictive tests are ordered before a person has symptoms. Generally, you would only receive clinical treatments (non-curative) and management for Parkinson’s or Alzheimer’s once you demonstrate actual clinical symptoms of the disease—the genetic variations are just a component of that diagnosis.
Genetic “diagnostic” tests on the other hand can give a more definitive “yes” or “no” answer. Diagnostic testing is often used identify a suspected genetic condition or confirm a clinical diagnosis. Unlike results from predictive tests, results from diagnostic tests can be used to influence health decisions, because the person generally already has physical signs and symptoms of the disease at issue. In the case of BRCA1 & BRCA2 testing however, the risk of getting breast or ovarian cancer is very high if the variants are present; and clinical interventions (e.g. mastectomy) are available before the onset of disease. So it’s considered diagnostic.
The crux of the problem here is that the FDA medical device classification system, based on safety and effectiveness risk to the patient, is hard to apply to a GHR testing product based on predictive genetic risks to the consumer. Essentially FDA is saying that it’s not as risky to give consumers risk information about conditions they might (or might not) get, as giving them DTC information about diseases that they already (or don’t) have.
A second important point is that—much like it did with carrier testing and Bloom Syndrome (where 23andMe only got Bloom Syndrome authorized, but then put over 30 other like-tests on the market)—FDA concurrently announced a limited authorization for the 10 GHRs along with the intent to exempt additional such tests from premarket review in the future. This is the buried lede. Matthew Herper (Forbes) reported that another NPRM will follow, which will hopefully lend clarity to the scope of that and some of our questions above.
Three, none of this is actually about DTC genetic testing anyway. Despite the inspiring enabling consumers with their information-kind of rhetoric, this is about research. 23andMe’s bio/databank, the largest of its kind in the world, is a major component of its $1.1 billion valuation. Not that 23andMe is hiding the ball. 23andMe participant data (note it has a research consent process) has been used for several recent high profile genetic breakthroughs including for depression and skin cancer. The more tests 23andMe can offer, the more consumer interest—the more consumer interest, the more donors and data for its bank. The more donors and data for its bank, the more interest from industry and other researchers.
In conclusion, this is only the latest turn in FDA’s saga over how to best regulate genetic testing, and there is more still to come. Time to restart all those unfinished drafts…