Stopping 23andMe Will Only Delay the Revolution Medicine Needs

Genetic testing is a powerful tool. Two years ago, with the help of my colleagues, it was this tool that helped us identify a new disease. The disease, called Ogden Syndrome, caused the death of a four-month old child named Max. But the rules and regulations for genetic testing in the US, laid down in the CLIA (Clinical Laboratory Improvement Amendments), meant I could not share the results of the family’s genetic tests with them.

Since that time, I have advocated performing all genetic testing involving humans such that results can be returned to research participants. This I believe should extend beyond research, and some private companies, like 23andMe, are helping to do just that.

For as little as US $99, people around the world can send a sample of their saliva to 23andMe to get their DNA sequenced. Their Personal Genome Service (PGS) analyses parts of a person’s genome. This data is then compared with related scientific data and 23andMe’s own database of hundreds of thousands of individuals to spot genetic markers, which the company claims “reports on 240 health condition and traits”.

Earlier this month, however, as I had feared, the US Food and Drug Administration (FDA) has ordered 23andMe to stop marketing their service. In a warning letter, FDA said: “23andMe must immediately discontinue marketing the PGS until such time as it receives FDA marketing authorisation for the device.” By calling PGS “a device”, the FDA fears that people may self-medicate based on results they receive from 23andMe.

Somehow the US and UK governments find it acceptable to store massive amounts of data about their own citizens and that of the rest of the world. They are happy spending billions on such mass surveillance. But if the same people want to spend their own money to advance genomic medicine and possibly improve their own health in the process, they want to stop them.

There are many diseases that appear to occur in the presence of genetic mutations, with large effect in certain populations. A case in point is that of deltaF508 mutation in the CFTR gene, which is known to predispose people to cystic fibrosis, which causes scarring inside organs.

The expression of cystic fibrosis in each of these people is highly variable, but the presence of the mutations can certainly raise suspicion for this illness in individuals with any such symptoms. This is particularly the case when there is an already known instance of cystic fibrosis in the immediate family.

This is why carrier screening in families with diagnosed cases of such diseases is advocated. And yet, such screening is not commonly performed, even though it could decrease prevalence of affected infants.

Genetic data (or genotype) on its own is of little use. It is the correlation of how those genes manifest in people, which is their phenotype, that makes genotypes useful.

I dream of a world in which we have phenotype and genotype data on millions of individuals, so that we can really begin to better understand genotype-phenotype relationships.

Instead, we still live in the medical world described in the Pulitzer prize-winning novel Arrowsmith published in 1925, where doctors pretend to know far more than they actually do. The sad fact is that there is no way the FDA can evaluate and regulate each and every genetic variant in the billions of letters which make up the human genome that get variably expressed in trillions of cells in every human body.

We need to collect billions of data points for analysis by computers. The only company in major contention to do this soon is 23andMe. With FDA’s latest attempt to stop 23andMe, all it is really doing is delaying, or worse stopping, the revolution that today’s medicine desperately needs.

Gholson Lyon is an assistant professor in human genetics at Cold Springs Harbour Laboratory. This article was originally published at The Conversation.

13 replies »

  1. Simply put, the FDA ordered cease and desist because it was a fraud. I would not rely on any testing from them. There is great mainstream genetic testing available for cystic fibrosis, BRCA and such. The understanding of the results can be complicated, and it is crucial that testing is ordered and interpreted by experts, whether they be interested physicians or other well-trained people.

  2. “… have confidential results, that insurers didn’t know about, and have peace of mind”

    The very definition of “Moral Hazard.”

  3. I turned to 23andMe a few years ago when my insurance company refused to cover a BRAC1 test. I am adopted and failed my mammogram. I had a first degree relative with breast cancer and a grandmother with ovarian cancer. For $99 I was able to have the screening completed, have confidential results, that insurers didn’t know about, and have peace of mind. Needless to say this was a great cost reduction from a traditional BRAC1 study. One may share this with health care providers for further direction, research, and so on as is the case with any of their test results. Knowledge is power. Genetics is here to stay and we must learn how best to utilize this in our quest to best serve our patients, communities, and the world health needs.

  4. The information that genetic testing provides is just a further refinement of what has been done for years, when patients are asked about their family history. Genetic testing just provides more information. Genetic test results indicate the mutations and the latest that science knows that they mean. Some of the knowledge we have on mutations is more known than others, but it is the latest we.

  5. A tremendous abuse of power.

    On the other hand let the buyer beware. Also, let the company be liable for any damages, if any.

    It sounds like fortune telling with a scientific twist. Reminds me of a joke about tennis elbow.

  6. “prognosis”

    “pro” – in the future

    “gnosis” – of knowledge

    risk estimation = likely knowledge of adverse future outcome, based on current evidence.

  7. “even prognosis”

    Really? What is giving people the odds of illness x based on your genetic analysis?

    From their home page:

    “Health risks
    Understand your genetic health risks. Change what you can, manage what you can’t.”

    Risks. Empirical prognoses. Which should be based in good science.

  8. People have access to all sorts of information that they didn’t 15-20 years ago. They can look up anything on the web, can get medical advise by calling in, going to the local pharmacy to see practioner, etc. Most of these sources are not vetted, so why not 23andme? They don’t advocate or prescribe treatment, even prognosis, just provide the latest information in a personalized way.

  9. You’ve cherry picked the best examples, but what about the false positives and overdiagnosis?

    And what does a man do when told he has a 5 % lower chance of getting lung cancer and a 8.2 % higher chance of getting prostate cancer? Start smoking after his first episode of terminal dribbling?

    Not all regulations are equal. FDA is spot on here.

  10. The FDA is correct in holding this business accountable. It should do the same for HIT devices including decision support and electronic ordering. No proof or evidence that it improves outcomes.