You Can't Handle the Truth

Do genetic tests need more federal regulation?

I have alleles that suggest that I have a lower than average risk of suffering from rheumatoid arthritis. Other alleles, however, indicate that I am slightly more likely to experience age-related macular degeneration than other people of European ancestry. In addition, if I need to use the blood thinner warfarin in the future, I should let my physician know that I have a version of the CYP2C9 gene, suggesting that I have a greater sensitivity to the drug and probably should start with a low dose. How do I know this genetic information? Because I paid for a direct-to-consumer genotype test from the California start up 23andMe.

I was especially eager to learn whether or not I carried the APOE4 allele that dramatically increases one's risk for Alzheimer's disease. 23andMe does not provide that information in its research reports, but the raw data it supplies can be exported to a wiki-like tool called Promethease for further exploration. Good news for me, Promethease finds that I don't have that deleterious allele. My failing mind must have other causes.

Genotype scanning is in a very preliminary stage of development. "With only a few exceptions, what the genomics companies are doing right now is recreational genomics," Duke University geneticist David Goldstein told the New York Times in April. "The information has little or, in many cases, no clinical relevance." As one goes through the mass of information—allelic variations linked to various studies of their effects in the peer-reviewed literature—it becomes clear that most of the current genotype information provides little guidance about future health.

But does such direct-to-consumer (DTC) genotyping need to be regulated—or banned? Last month, Germany banned direct-to-consumer genetic testing. In 2007, the Genetics and Public Policy Institute found that 24 states limited or restricted direct-to-consumer genetic testing in some ways. Last year, both New York and California heightened their regulatory scrutiny of genotype scans such as those done by 23andMe, deCode, and Navigenics. The good news is the genotype scanning companies obtained licenses from California authorities, and so can now still offer their services to California residents. Several companies are still navigating through New York State's regulatory maze. Last year, the American Medical Association adopted recommendations opposing direct-to-consumer genetic testing, stating that a health care professional should be involved with any genetic testing.

More recently, Genetics and Public Policy Institute director Kathy Hudson organized and sent a letter urging further federal oversight of all advanced diagnostic tests. The letter was signed by representatives of 108 companies, patient advocacy groups, and health non-profits. The goal is to "get personalized medicine right." The letter urges that oversight of diagnostic tests be risk-based and that regulation of those tests be implemented by means of a Food and Drug Administration rulemaking procedure. The letter further advocates the creation of a mandatory registry for all tests which would collect information supporting the analytical and clinical validity of each test.

When asked via email if the regulatory scheme envisioned by the letter would apply to direct-to-consumer genotype scanning and genomics companies, Hudson replied, "The letter is silent on the mode of sale. It is the risk of the test, not the way it is sold, that the group of signatories were focused on. That means that if a test is of high risk, it would be subjected to heightened regulatory scrutiny whether it was offered DTC or not. Current DTC genetic testing services include those that are of no medical significance (low risk) and some that are of substantial medical significance (e.g. cancer)." She then referred me to the GPPC's list of DTC testing services. On the list, tests for genes for hair loss (presumably low risk) are offered along with tests that suggest risks for cancer, heart disease, and Alzheimer's disease (presumably high risk).

The high risk in this case seems to mean two things: how likely is the test to be accurate in what it tells customers, and how likely are customers to find out that they are at higher risk for some dread genetic malady. For example, on the GPPC list one finds that the genotyping companies Navigenics and deCode offer results from a scan for alleles that contribute to the risk of Alzheimer's disease.

Are they accurate in finding such alleles? I asked Princeton University molecular biologist Lee Silver via email how accurate the DTC genotype scanning tests are. "I ran an analysis on personal genome results obtained from 23andMe and DeCODE for me," Silver replied. "There were about 300,000 data points that overlapped between the two tests. There was not a single data point (among 300,000) that was scored positive in one test and negative in the other." Silver is satisfied with accuracy of such screening tests.

But what about the second risk—that customers of DTC genetic testing services will find out that they have a high risk of being susceptible to some horrible disease? "Granted, there's the possibility that you'll be upset with what you find out about yourself, but that's never been a sufficient reason to keep personal information out of a person's hands, when the person is asking to see it," argues Silver. Currently there are few tests of that sort, the APOE4 Alzheimer's allele test being one of them.

So should people be protected against learning such information without the guidance of a knowledgeable physician or genetic counselor? "Actually, I suspect that most consumers of the 23andMe service know more about modern genetics than do most American physicians and most genetic counselors," asserts Silver. I really hope Silver is wrong about that, but agree wholeheartedly when he says, "And I don't see how anyone can be harmed any more by personal genetic information than by some other kind of personal medical information."

There may well be some inaccurate tests and there will certainly be people who mislead customers about the meaning of certain tests. But do we really need additional federal regulation to weed out bad actors? Most evidence suggests that the current tests are fairly accurate, and that customers are not being misled by the results that are reported. All new technologies involve a societal learning process in which some early adopters try it out, explain to others how it works, and find out its flaws—which newer innovators then fix.

In fact, the way the genetic testing industry will evolve is that the companies that tease out and explain useful information about disease risks and pharmaceutical interactions will be the ones to succeed. The bigger risk is that increased federal regulation will slow down beneficial genetic testing innovations.

Ronald Bailey is Reason magazine's science correspondent. His book Liberation Biology: The Scientific and Moral Case for the Biotech Revolution is now available from Prometheus Books.