DNA tests have their limits
When last we visited with the genetic testing firm 23andMe, it was in big trouble.
That was in December 2013. The Food and Drug Administration had just ruled that the company had been breaking the law in marketing its Personal Genome Service. For $99, customers spit into a tube, mailed it to 23andMe and got back a report on their carrier status or susceptibility to breast cancer, heart disease and some 250 other diseases or conditions, based on genetic information extracted from their saliva.
The FDA described the service as an unlicensed medical device and ordered it off the market. Its warning letter implied that the Mountain View, Calif., company hadn’t been straight with regulators and was ignoring demands for data proving its service was accurate. Experts in the genetics testing field wondered whether the tests — and the company itself — were ready for prime time.
As of mid-October, 23andMe is back in the direct-to-consumer genetic testing business. (The firm’s name refers to the 23 chromosome pairs in the human genome.) It’s selling a different product, one that doesn’t raise the consumer education or health and safety issues that motivated the FDA’s action. Its relationship with the agency appears to be much better too.
That doesn’t mean that the company’s new businesses are entirely free of consumer and safety concerns. Its efforts to exploit its growing database of customer genomes, now numbering 1 million subjects, by selling it to drug companies and launching a drug-development program of its own could raise privacy concerns.
The tradition-bound FDA’s approach to regulating a novel health service also may be evolving, partially as a result of its interaction with 23andMe.
“The FDA has been diligently working toward clearing a kind of product they haven’t cleared before,” Andy Page, the president of 23andMe, told me last week.
The agency, meanwhile, has been seeking expert advice on how to ensure that genetic tests marketed to the public are “accurate, reliable and clinically relevant” without stifling innovation, as it said in a dis-
cussion paper last December. Its challenge is only going to grow, as new, innovative companies enter the market.
Consider Theranos, a Palo Alto start-up with an ostensibly new way of drawing blood whose effectiveness and reliability have come under FDA scrutiny. Theranos and 23andMe are in different businesses but share an evangelical marketing pitch built around individuals’ right to access their personal medical data without the intervention of a physician. The attitude can cast regulators and the medical establishment in the role of hidebound protectors of entrenched interests, rather than guardians of public safety — which makes the regulator’s job that much more complicated.
Yet the change in the consumer service now being sold by 23andMe — for the higher price of $199 — is at least an implicit acknowledgment that the FDA was right to shut down the old one.
Originally, customers were given a report detailing their relative susceptibility to hundreds of diseases based on the presence of gene variants detected in their DNA.
This raised several concerns: The role of genetics in some conditions may be marginal or conjectural. With such multifaceted conditions as cancer or heart diseases, one’s behavior and lifestyle probably have much more to do with disease risk than one’s genes. These considerations might have evaded many consumers despite lengthy 23andMe disclaimers. The FDA feared that consumers might be misled into undergoing expensive tests or surgeries for conditions they didn’t really face.
The company’s new program, which has received the FDA’s general blessing, focuses chiefly on “carrier status” — the presence of a gene variant that could affect the health of one’s offspring, especially if they inherit the same variant from a second parent. The service covers 36 conditions known to be linked to the inheritance of recessive genes, including Tay-Sachs disease, cystic fibrosis and sickle-cell anemia.
Instead of assessing the customer’s own health or disease risk — which are uncertain — the service has been remade into a tool for “adults contemplating having a family or looking at what could be passed on to future generations,” Page says. That’s a much simpler statistical calculation — the child of two parents each carrying a specific recessive gene is known to have a 25% chance of inheriting the gene from each parent and thus inheriting the disease, a 25% chance of not inheriting the gene at all, and a 50% chance of inheriting the gene from one parent and thus becoming a carrier, but not contracting the disease.
That turns the service into a family-planning tool similar to that routinely provided to prospective parents by genetic counselors. Indeed, the FDA required 23andMe to give customers information on finding genetic counseling as part of its reports. The firm’s reports also give customers more innocuous data, including windows into their genetic ancestry, their predisposition to four “wellness” conditions such as lactose intolerance, and such personal traits as eye color, baldness and dimples.
Then there’s the firm’s entry into the sales of its customer database to drug companies and its own efforts to develop treatments for gene-based conditions. These ventures conceivably could come to dominate the firm’s business; its revenue from a deal with the biotech company Genentech will yield as much as $60 million. The company says all customers are asked to opt in to an anonymized and aggregated database, and more than 80% do so.
The narrower focus of the company’s new consumer service may not quell the debate over how much medical data is healthy for individuals to have about themselves, and who should interpret it.
“I disagree with the culture that says more information is always good,” says Brian Zikmund-Fisher, an expert on bioethics at the University of Michigan. “Providing people with more information is not helpful if they can’t do anything about it, or it leads them to focus on the wrong thing” — on their genes rather than their lifestyles, for example.
Yet “it makes no sense to have a physician as a barrier” to obtaining such information on one’s genetic makeup, says David B. Agus, a professor of medicine and engineering at USC and co-founder of the genetic testing company Navigenics. Even information about marginal genetic risks for common conditions such as heart disease can persuade patients to take better care of themselves, says Agus, author of “The Lucky Years,” a forthcoming book about new healthcare technologies.
The FDA has cast its vote for more access. “In many circumstances it is not necessary for consumers to go through a licensed practitioner to have direct access to their personal genetic information,” the agency ruled in February.
Nor should anyone be surprised if 23andMe or its competitors eventually return to providing customers with a lot more data. The current service, Page says, is “a launching-off point for lots of information we’d hope to provide in the future.”