DTC Genetic Testing: 23andme, DNA Direct and Genelex

Figure 1: The state of California is clamping down on companies that offer direct-to-consumer testing in a move that threatens the burgeoning industry.

© 2008 Nature Publishing Group Wadman, M. Gene-testing firms face legal battle. Nature 453, 1148 (2008). All rights reserved.

You can't actually "google" your DNA... at least not yet. But recently, Google has received a great deal of press for its heavy investment in two start-up genetic screening companies: 23andMe and Navigenics. Google invested $4.4 million in 23andMe, which was initiated partly by Anne Wojcicki, the wife of Google cofounder Sergey Brin, and another undisclosed but undoubtedly large sum of money in Navigenics. Both 23andMe and Navigenics offer genetic testing services direct to the consumer. In other words, you don't need a physician order or any other authorization to be tested. All you need to do is write a hefty check (about $1,000 for 23andMe and $2,500 or more for Navigenics) and send your saliva through the mail; then, several weeks later, you will receive your results. But what do those results mean?

Moreover, while Google and others clearly see huge economic potential in so-called direct-to-consumer (DTC) testing, could DTC testing be dangerous to your health? Some experts say yes. For example, in June 2008, California's Department of Public Health sent letters to 13 genetic testing companies, informing them that they were "in violation of California law... which prohibits the offering of a clinical laboratory test directly to the consumer without a physician order" (Figure 1; Wadman, 2008). One of the 13 companies receiving this letter was 23andMe.

DTC or consumer genomic tests typically involve analyzing hundreds of thousands to over a million single nucleotide polymorphisms (SNPs). Some of these SNPs are known to be associated with a disease, drug response, or other phenotype that consumers might want to know about.

23andMe, for example, provides consumers with information on 78 gene-trait associations. According to information posted on the company's website, 23andMe analyzes only those associations that have been published in peer-reviewed scientific literature. They divide the associations into two categories: established research and preliminary research. Established research associations are those that have been confirmed in at least two large studies or "have gained widespread scientific acceptance in the scientific community." Preliminary research associations are those that "still need to be confirmed by the scientific community" and "may not stand the rigors of scientific replication." The company relies on an editorial team, which includes three Ph.D. scientists with university affiliations, to make decisions about which gene-trait associations to include in its analysis.

Table 1 shows the gene-trait information that 23andMe provides its customers, including the levels of research confidence for each trait. Traits and diseases in red are in the established research category of gene-trait associations; those in black are in the preliminary research category. The number of asterisks for each trait or disease indicates the level of confidence in the related research; four asterisks is representative of the highest confidence level (23andMe, 2008).

Table 1: Gene-Trait Information Available from 23andMe

Navigenics, on the other hand, provides information on only 18 gene-trait associations. Unlike the phenotypes that 23andMe examines (Table 1), which range from earwax type to muscle performance, all of Navigenics's traits are serious disease phenotypes. In addition, you must become a member of what the company calls its "Health Compass" before undergoing or having access to your results, and the company provides genetic counselors for members. The membership fee is $2,500 for the first year and $250 per year for each subsequent year.

The information that 23andMe, Navigenics, and other DTC companies provide is neither definitive nor intended to be definitive. The purpose of most of the tests is simply to provide individuals with information about their risk of developing a disease or some other trait, or their odds of responding in a particular way to a certain drug. For the vast majority of these tests, there is no way to definitively determine whether an individual will actually ever develop the disease in question or respond in a certain way to a drug.

For example, suppose your genetic profile reveals that you have a genetic marker known to be associated with obesity. This marker, identified on the 23andMe website as RS3751812, is an SNP that is located near a gene (FTO) known to be overexpressed in the fat cells of many obese people. In other words, studies have shown that the difference between the number of obese people that have that SNP marker in their genome and the number of obese people without the SNP marker is significant enough to associate the marker with obesity. But having this genetic marker does not necessarily mean that you are or will ever become obese. It simply means that your risk of obesity is greater than that of individuals who lack the marker. (Similarly, this does not mean that unmarked individuals will necessarily stay trim.) Also, if you look at the fine print on the 23andMe website, you will notice that the association has only been confirmed in people of European ancestry (i.e., the data on the association were collected using samples from France and Germany), so the information may mean nothing for people of other ancestries.

Still, there are few people who would deny that this type of genetic information is, at the very least, interesting. On its website, 23andMe welcomes you to "[u]nlock the secrets of your own DNA" and "shed new light on yourself." But is it helpful? In theory, yes. For example, a man who learns that he has a higher risk of prostate cancer based on the presence of a DNA marker known to be associated with the disease might be more vigilant about prostate screening. However, it is not at all clear whether people who receive positive genetic test results are, in fact, taking a more active role in maintaining their health and preventing disease. And even if they are, there remain questions about whether these individuals are more careful with their health because of the DTC test results or because those people who sign up for DTC testing are typically more careful about their health anyway. Perhaps the more serious question is whether or not DTC testing could be harmful.

A March 2008 article in Science magazine stirred some heated controversy about the value of DTC tests, particularly in cases in which the science is still being debated (Katsanis et al., 2008). The Science authors, all three of whom are affiliated with the Johns Hopkins University Genetics and Public Policy Center, focused on DTC testing for cytochrome P450 (CYP450), a family of genes that neither 23andMe nor Navigenics analyzes. Rather, the article points to two companies that do offer DTC testing for CYP450: DNA Direct and Genelex. Genelex, for example, offers a "Drug Reaction Panel" test for $690, which includes a genetic analysis of CYP2D6, CYP2C9, and CYP2C19, all of which are members of the CYP450 family. Consumers can simply go to Genelex's online store, add the test to their cart, and proceed to checkout.

CYP450 polymorphisms have been associated with the effectiveness of SSRIs (selective serotonin reuptake inhibitors), a type of antidepressant drug. In theory, therefore, the CYP450-SSRI association information could potentially be used to make better decisions about which antidepressants to prescribe to patients and at what doses. But data from clinical studies show that the genetic information does not really make a difference with respect to patient outcome. In fact, a 2007 review of the studies commissioned by the U.S. Centers for Disease Control concluded that patients prescribed a certain SSRI at a particular dose, based on CYP450 testing, fared no better nor worse than patients who chose not to have the test (Evaluation of Genomic Applications in Practice and Prevention Working Group, 2007).

The fact that DNA Direct and Genelex offer CYP450 DTC services to consumers is "troubling," the Science authors write. "There is no health-care provider or ‘gatekeeper' to prevent inappropriate test ordering or misinterpretation of test results... For example, a patient informed of his or her cytochrome P450 (CYP450) profile might independently change the dose of antidepressant medication with adverse health outcomes." Antidepressants have been associated with suicidal thoughts and suicide attempts, among other adverse effects, and physicians are warned to carefully monitor their patients, particularly when adjusting dosage, which may prove difficult with increased access to DTC testing.

The Hopkins scientists argue that this lack of gatekeeping raises questions about whether federal agencies with the authority to do so, like the U.S. Food and Drug Administration (FDA) and the Federal Trade Commission (FTC), need to step in and enact measures to ensure that consumers are not making inappropriate, perhaps even dangerous, decisions based on their DTC test results. Unless appropriate measures are taken, the scientists argue, DTC testing could actually backfire, by causing consumers to lose their trust in the information provided by such services.

Not everybody agrees with the argument for more oversight of DTC testing, and there has been considerable public discussion around the type of oversight needed, which is why the Science article—and, more recently, the California Department of Public Health letter—have prompted so much heated debate. Some people think that genetic tests simply provide information, unlike most medical procedures, and therefore should not be regulated. Others think that tests providing information about disease risk should be regulated differently than tests providing "fun" information about your genetic makeup, such as some of those that 23andMe offer (Wadman, 2008). What do you think?