The FDA Versus 23andMe: Does it have to be a Zero Sum Smack Down?

Genetic testing isn’t something we often think about; and it’s certainly not something we regularly indulge in unless having children is involved. Sometimes however, a family history can prompt us to search for answers left previously unnoticed in our DNA. Being told you are “in the clear” would obviously warrant a sigh of relief and possibly a congratulatory hand shake, but what happens when you’re given less than stellar news?

That is exactly what Dr. Uta Francke, a professor of Genetics and Pediatrics at Stanford University’s School of Medicine, and her colleagues hoped to assess in their study “Dealing with the Unexpected: consumer responses to direct-access BRCA mutation testing”. The good doctor and her team looked specifically at the “23andMe Personal Genome Service”—which tests for BRCA1 and BRCA2 (a gene that provides an extremely high risk of breast and ovarian cancer in carriers). Upon revealing the gene-positive results they noted that none of the patients reported extreme anxiety. Even better news: men “burdened” with the gene alerted family, leading to 30 more tests and 13 more positive identifications. With knowledge in hand, positive women were able to take proactive measures to protect themselves. Dr. Francke and Co. concluded that such testing “provided clear benefits for participants” noting the “absence of evidence for serious emotional distress or inappropriate actions”.

The FDA 23andMe

If members of the medical community are all for further genetic testing of women (and men) for a variety of ailments, then why aren’t we hearing about it? Unfortunately, 23andMe hasn’t been the most cooperative with a government force to be reckoned with—the Food and Drug Administration.  While the FDA may not have anything against the average citizen buying a testing kit and using it to “participate in science”, it was forced to crack down on a company that refused to play by their rules, respond to correspondence and concerns and sold itself as a way to test for the breast cancer gene (rather than as the potential database it hopes to be for drug testing and genetic links). Even worse, they never submitted testing results meant to prove the efficacy and consistency of their equipment as a means for revealing genetic information.

Just because 23andMe has chosen to bring its corporate longevity into question by mishandling its relationship with the FDA (which is a much needed organization) does not mean we do not need to develop systems that allow the average person to access important genetic information while providing the tools to deal with real life health data. The FDA doesn’t have to be “the bad guy” here, a compromise can—and should—be reached. Currently there isn’t much opportunity for the public to “opt-in” to studies that test the latest pharmaceutical; instead these companies select their own candidates. While such a method has worked in the past, it is severely limiting and can lead to (whether purposefully or not) swayed data.

Such scientific “snafus” can have detrimental results. Take the published recommendations of Dr. A.B. Goldfine in The New England Journal of Medicine which defended the primary use of a family of drugs known as “statins” for the prevention of cardiovascular disease. Dr. Rita Redberg, M.D., of The University of California San Francisco was alarmed to read such a statement. She then published a rebuttal, providing meta-analysis data (formally collected by Kausik K. Ray, M.D., M.Phil., F.A.C.C., and his colleagues at the University of Cambridge) that concluded: “statins for primary prevention showed no mortality benefit”. Even more troubling: the state of the studies included. After removing 4 studies from the analysis with “serious risk of bias” Redberg and the Cambridge doctors still “noted evidence of selective reporting of outcomes, failure to report adverse events, and inclusion of patients with cardiovascular disease”—leading to potentially skewed data. Researchers also chose to disregard Women’s Health Initiative records showing an increased risk of diabetes (by 48 %!) in women taking statins. The analysis examined a pool of 65,229 participants—of which 2,793 people died (1,447 were assigned the placebo, 1,346 the statins, a statistically insignificant difference). Selecting your own pool may not be done with malicious intent, but as any good scientist knows, everyone is biased. Correcting for your biases and inclinations is key to providing good data. Patient opt-in databases, which can allow anyone to participate quite readily, may be the best way to avoid human selection biases.[1]

Taking unnecessary medication is one thing, but when medications can have hidden complications that lead to death a need for a wider pool of test subjects becomes all the more urgent. Currently in Britain 8,000 people are diagnosed with esophageal or gullet cancer—7,500 of them die of it. Such a dismal prognosis takes on a darker tone when it is revealed that oral bisphosphonates (a common treatment for osteoporosis) double the taker’s risk for these cancers when taken for five years or more. Oxford University’s Cancer Epidemiology Center paired up with the UK General Practice Research Database to review over 6 million cases. Researchers found that the chance for esophageal cancer was 30% higher in people with one or more previous prescriptions for bisphosphonates. Ten or more prescriptions doubled the risk. Because bisphosphonates are used as a preventative prescription for at-risk groups (think menopausal women) doctors are becoming concerned that they are “overprescribed”—with potentially dangerous results. While researchers officially concluded that their findings don’t mean you have to stop taking your bisphosphonates, the power of such meta-analyses continues to point to a need for larger databases with a variety of participants. This is especially true if we desire to gain all the answers—and make the best decisions about our health that we can.[2]

While the FDA was well within its right to “come down upon” 23andMe for not cooperating with an organization whose purpose is to protect American citizens, that doesn’t mean that 23andMe’s premise (a large pool of truly random citizens available for trial, and easily accessible medical information for all) is a bad one. This needn’t be a zero sum game, on the contrary, as we continue to learn how important random trial groups and genetic testing really are, cooperation becomes more important. That is what I advocate for, a meeting of the minds—for the betterment of all.

Yours in Health,

Stella Metsovas