washingtonpost.com
Correction to This Article
The byline on the article about companies that sell genetic analyses directly to consumers misspelled the writer's name. He is Chris Berdik.
Genetic tests give consumers hints about disease risk; critics have misgivings

By Chris Berdik
Special to The Washington Post
Tuesday, January 26, 2010; HE01

Last fall, Sgt. Timothy Gall, an Army medic stationed at Fort Belvoir, sought clues to the multiple sclerosis and heart disease that ran in his family by looking into his DNA. All it took was some spit and about a thousand bucks.

He didn't go to a doctor. Instead, Gall, 30, joined the growing number of consumers ordering scans of their DNA directly from private companies. A handful of companies such as 23andMe, Navigenics and Decode Genetics offer customers a personal peek at their genetic code, finding variations linked to certain traits, diseases and drug sensitivities -- a process known as genotyping. As the cost of genetic scanning has dropped and the pace of genetic discovery quickened, these companies began springing up about three years ago. Now, as they're attracting more and more customers, they're also drawing more scrutiny.

Once Gall decided to try it, he persuaded his father to be genotyped, too. As a quality check, both men mailed saliva samples to two testing firms: 23andMe, based in California's Silicon Valley, and Decode, of Reykjavik, Iceland. "I figured I'd only pay attention to the results where both companies agreed," Gall said.

Gall found some of the results "useful," including a warning that he might be unusually sensitive to blood-thinning medications and an alert that some of his ancestors were Ashkenazi Jews, who have an increased risk of certain diseases. Other findings aligned well with his own experiences: One of his genetic variants makes him less able to taste bitter flavors and another means he's more likely to go bald.

But the results related to his most important concerns were perplexing: Despite the fact that his father and other close relatives have multiple sclerosis, both companies agreed that Gall and his father had a below-average likelihood of getting the disease. And they found an average predisposition for heart disease, even though heart disease runs in his family and even though his father had a heart attack at age 50 despite eating right, exercising and having very low cholesterol.

How much it costs

Gall's experience illuminates the controversy around direct-to-consumer genotyping. Advocates say these services can guide people toward appropriate preventive medical care, help them choose medications and motivate them to make lifestyle changes. But others criticize the companies for overselling their supposed insights and producing reports that untrained consumers might easily misunderstand. The American Medical Association recommends that a physician always be involved in genetic testing, and the lack of physician involvement has made the tests technically illegal in several states.

The only current federal oversight of direct-to-consumer genetic testing is truth-in-labeling enforcement by the Federal Trade Commission. But an advisory committee within the Department of Health and Human Services has been studying the companies and plans to send its recommendations on regulation to HHS Secretary Kathleen Sebelius within the next few weeks.

For the time being, it's up to consumers to be careful -- beginning with understanding what genetic testing companies are out there. There's a wide range: Some test risk for a single health condition, such as celiac disease, or are focused on telling the customer something about his ancestry; their services typically cost between $100 and $500. Others offer full genome sequencing at prices up to about $100,000.

In between are genotyping services, which look at just a fraction of the customer's DNA but provide a broad range of genetic information at prices of $400 to $1,000. Rather than reading an entire genome, which contains about 3 billion base pairs of chemical building blocks, genotyping companies scan hundreds of thousands of base pairs looking for "single nucleotide polymorphisms" (known as SNPs, pronounced "snips"). A SNP is a variation of a single DNA base pair that studies have shown to be more common in a certain population: A particular SNP might be more common among people who develop breast cancer, for example; another, among schizophrenics; yet another, among individuals who live to be 100.

Genotyping companies test for SNPs associated with dozens of diseases, traits and drug reactions -- everything from Parkinson's to lupus to earwax type. They typically predict a customer's genetic risk by calculating how often that condition occurs among people of the customer's general age, sex and ethnicity, then factor in the presence or absence of the relevant SNP. The customers explore their personalized results via secure Web sites. Some companies provide updates to the customer's profile as scientists uncover more SNP associations.

How much it matters

Third-party audits by independent geneticists have found the major direct-to-consumer genotyping companies to be more than 99 percent accurate when it comes to locating the DNA variants. But not all variants are equally meaningful. Some extremely rare genetic variations can increase disease risk by several hundred percent, but most SNPs are much less significant, and many of those tested by genotyping companies raise risk by as little as 2 or 3 percent.

Leslie Biesecker, a senior investigator in the National Human Genome Research Institute of the National Institutes of Health, says SNPs alone are of limited practical use, because the additional disease risk associated with any bit of DNA may be multiplied or completely overridden by other DNA variations that aren't known, by the location of a given genetic sequence within the genome, or by environmental and lifestyle factors.

"If you could take the integrated influences of a bunch of genetic variants and a number environmental inputs, and you put that into a formula, then you'd have the potential to make a good prediction," Biesecker says. "But the SNPs themselves are just inputs."

Another criticism is that the same person sometimes gets very different results from DNA submitted to more than one genotyping company. In October 2009 , a group of scientists writing in the journal Nature said they submitted five DNA samples to both 23andMe and Navigenics and had them tested for risk of 11 disease. For seven of the diseases, at least one sample got contradictory risk results from the two companies. Many geneticists suggest that a cheaper way to judge your chances of developing a disease is to know your family history.

Representatives from the leading direct-to-consumer genetics companies say they're upfront about what their tests' algorithms can and can't reveal. They point out that most company Web sites have an extensive Q&A section about SNPs and what the numbers show, as well as disclaimers. Navigenics, for example, notes that reports and risk estimates "are for informational purposes only and are not intended to substitute for professional medical advice, diagnosis or treatment."

But Navigenics also puts this sales pitch on its home page: "We use the latest science and technology to give you a view into your DNA, revealing your genetic predisposition for important health conditions and empowering you with knowledge to help you take control of your health future."

Robert Green, a Boston University professor of neurology, genetics and epidemiology, says it's important to distinguish fly-by-night companies using genetic tests to push nutritional supplements or skin-care products from "the leading companies, which are very responsible about writing clear and accurate material on their Web sites." Nevertheless, he adds, there's a potential conflict between total honesty and maximum profit, "because these companies are selling the idea that genetics matters."

'It calmed my fears'

Customers often say they understood exactly what they were buying.

"I actually had markers decreasing my risk for breast cancer," says Alysen Plennert, who signed on with 23andme about a year ago. "It calmed my fears a little, but they make it very clear that just because you're at reduced risk, doesn't mean you're not going to get the disease."

Plennert, of Chicago, had a reason for probing her genes: Adopted as a baby, she had no information on her birth family. She'd had several unusual health problems as a young woman: early indications of glaucoma and MS, a lump in her breast at age 27 (it turned out to be dense tissue) and then unexplained chest pains last year.

"These kinds of things kept happening to me, and it made me wonder why they were happening and what else I should look out for," she says. According to 23andMe, Plennert has a higher than average genetic risk for glaucoma, psoriasis, multiple sclerosis, colitis and thyroid cancer.

Some customers say a personal genetic risk profile provided extra motivation to change their lifestyles. After being told he has a genetic susceptibility for skin cancer, for example, Gall says he'll definitely be slathering on sunscreen. Meanwhile, the SNPs of Doe Rieger, a 23andMe customer from northern New Jersey, indicated a higher risk for several cancers. So she asked her doctor to perform extra tests, which revealed a large tumor hiding under her stomach. It was benign, but because of its size and its potential to become cancerous, it was removed. (23andMe provided Rieger and Plennert's contact information for this story.)

But an ongoing study indicates that any motivational benefits might just be self-selection. Since 2006, researchers at the National Human Genome Research Institute have looked at the goals and follow-up patterns of individuals who volunteered to undergo genetic testing of their risk for eight common health conditions. According to the study's principal investigator, Colleen McBride, one major finding is that "the people who were most likely to get tested were already the most motivated to make health behavior changes."

So why do they get the test? McBride says, "I honestly believe it's pure curiosity and information for information's sake."

Value added, or not?

Why shouldn't people be able to pay for every bit of personal health information that science can provide, with or without a doctor's approval, no matter what the real value?

One worry is privacy: that a person's genetic risk profile might somehow get into the hands of employers or insurance companies. That concern has been somewhat alleviated by last year's passage of the Genetic Information Nondiscrimination Act.

Another danger is the potential for customers to be falsely reassured by "low-risk" SNP results and to make poor health choices as a result. There's also the worry that high-risk results could lead to extreme anxiety. But a study reported last July in the New England Journal of Medicine found that participants who were told the results of a test for genetic links to Alzheimer's were in no more distress than those who were kept in the dark.

Nevertheless, Siobhan Dolan, an OB-GYN and geneticist at New York's Albert Einstein College of Medicine, says there's still the question of utility. After all, she says, most of the options available to somebody whose SNPs reveal a predisposition for common health ailments such as heart disease or skin cancer are already well-established public-health mantras such as eat right, exercise and wear sunscreen. One day, our DNA may help us delineate a personalized approach to preventive medicine, says Dolan, "but if the question is what's the value added of genetic information as of now, then the answer is: not much."

As for the direct-to-consumer issue, with no physician involved, many genotyping companies acknowledge that a well-trained guide can be useful for interpreting their information. Navigenics, for instance, has five licensed genetic counselors on staff who call customers who have received SNP reports to ask if they need assistance in analyzing their results.

The flood of genetic associations available for consumers to ponder is rising quickly. The number of SNPs included in a Wikipedia-style database called SNPedia doubled in 2008 and nearly doubled again in the first six months of 2009. The cost of genotyping continues to drop. And while direct-to-consumer companies have taken their lumps in the down economy -- 23andMe had layoffs in September, as did Decode Genetics, which filed for bankruptcy in November -- even skeptics expect such services to survive and grow as consumers take increasingly active roles in managing their health information.

"These are fascinating new examples of approaching health outside of the conventional medical care system," says Green. "I see this as an evolving process, and these companies are shaking up our way of thinking about genetics and health in a way I think is ultimately going to be very positive."

View all comments that have been posted about this article.

© 2010 The Washington Post Company