After customers hit the “buy” button on one of these sites and enter their payment data, they get a test kit in the mail for submitting their DNA as a saliva sample or a buccal swab. A few weeks later, an e-mail notifies them that their password-protected results are available on the Web site.
Results are based on one-letter DNA variations (single nucleotide polymorphisms, or SNPs). All three vendors will conduct an overall assessment of your genome and an analysis of your SNPs relative to published research on SNPs known to be associated with common diseases.
For example, Navigenics lists, among many other conditions, Alzheimer’s disease, breast cancer, type 2 diabetes, heart attack, lung cancer, lupus, macular degeneration, multiple sclerosis, and obesity under “Conditions we cover.”
Estimated disease risks come with many qualifications, however. For example, an SNP may be associated with a disease, but so may one or more still unidentified SNPs or gene mutations. Environmental factors also influence risk. Also, someone whose results show a higher risk for a disease may never get it, and someone with a lower risk already may have the disease or may get it one day.
Beth Peshkin, senior genetics counselor at the Lombardi Comprehensive Cancer Center of Georgetown University Medical Center, makes a distinction between testing for SNPs and testing for gene mutations known to be associated with such diseases as breast cancer or Huntington’s disease. Right now, a highly predictive test for hereditary breast and ovarian cancer is available only from the company holding patents on the BRCA 1 and BRCA2 genes.
“In the adult genetics arena, the most common testing being done is for breast, ovarian, and colon cancer susceptibility, and unlike the tests being offered by companies like Navigenics, 23andMe and deCODEme, these are highly predictive genetic tests,” says Peshkin. “SNPs may be associated with increased or decreased risk of disease, but in many cases, these are very minimal changes in risk, like a 15 percent increase in risk, which is not enough to impact medical management. That number sounds high, but when you calculate the percent increase in risk, it is very small.”
Crenson, at 23andMe, concedes Peshkin’s point. “We’re aware that the information we provide is more educational and informational,” he says. “I certainly wouldn’t take medical action based on the information in my account.”
But Stefansson, for one, disagrees.
“Many of these variants confer risk that is greater than the risk of heart attack when you are in the top quintile of cholesterol,” he says. “Forget that they are genetic variants; look at them as risk factors we can measure. And there are all kinds of things that can be done to diminish the genetic risk. This is bound to have a tremendous impact on how we switch from interventional to preventive medicine.”
Peshkin maintains that SNP data cannot be properly interpreted in the absence of a family history: “The danger of interpreting this information without context of family history is that people who have a family history and test negative for gene alterations may be falsely reassured.”
Navigenics says it includes only conditions that are “actionable,” meaning prevented, detected early, or treated.
Peshkin thinks that for many consumers, “Spending that [money] on a gym membership that’s actually used may be more valuable than spending it on getting information that we don’t know how to interpret.”
Elissa Levin, director of Navigenics’ genetic counseling program, has a different viewpoint. “Knowing that they are potentially predisposed to certain conditions may be the push people need to engage in their own health,” she says. “They are gaining insight into how to make better health choices about screening, diagnosis, and prevention.”
Stefansson, for his part, is convinced that services like deCODEme will be heavily subscribed. “I think [genetic tests] are going to conquer the world within the next 3 to 5 years, and we are fighting for our share of that.”