Subscribe to eTOC

FDA Warns Companies about DTC Genetics Tests


Neurologists with expertise in genetics discuss reasons why the FDA warning that companies marketing genetic tests to consumers need premarket approval is necessary.

In June, the FDA warned five companies that are marketing genetic testing to consumers that their products are medical devices requiring the federal agency's pre-market approval. The products are considered “devices” under the Federal Food, Drug and Cosmetic Act because they are “intended for use in the diagnosis of disease or other conditions or in the cure, mitigation, or prevention of disease or intended to affect the structure or function of the body.”

The five companies that received nearly identical “warning” letters are deCode Genetics, NaviGenics, Illumina, Inc., 23andMe, Inc., and Knome, Inc. All of the FDA letters are posted on the FDA Web site.

Some of the neurological conditions for which at least some of the companies are offering tests include Alzheimer disease, brain aneurysm, Canavan disease, multiple sclerosis, amyotrophic lateral sclerosis, and restless leg syndrome.

“The FDA has been reviewing the online marketing practices of these companies for over a year,” FDA spokesman Dick Thompson told Neurology Today. “The agency is taking a more active role now because they are seeing them increasingly market their products to consumers. To protect the public's health, we believe it is important that these tests have been shown to be safe and effective.”


Neurologists who spoke with Neurology Today say they believe the FDA action was necessary because much of the information the products provide to consumers is about risk for disease that can be easily misinterpreted.

“The biggest concern in the genetic medicine community has to do with quality control of the information that is being given out to consumers,” said Katherine Mathews, MD, director of the division of pediatric neurology at the University of Iowa. “In many of the conditions for which they are testing, the genetic component of risk is smaller than the risk you could identify by doing a simple family history.”

“By taking a history and BMI you can determine the risk for stroke or heart attack with much greater accuracy than you can by looking at genetic variation,” Dr. Mathews told Neurology Today. “So people are likely to misunderstand the value of the genetic information they are paying for and making decisions about their health based on that.”

Dr. Mathews' perspective was echoed by two other neurologists with expertise in genetics, Thomas Bird, MD, and Jeffery Vance, MD, PhD. Commercially available genetic tests might inform someone that they have an “increased” risk for a disease, though the gene may confer only a marginally increased risk, they said, for a condition that may actually require multiple genetic mutations in combination with environmental factors.

For example, Dr. Bird, professor of medicine, neurology, and medical genetics at the University of Washington, noted that a gene known as DPP6 is believed to be associated with risk for AKS. “But the risk is very low, so that even if you have it, it remains unlikely that you would develop the disorder,” he said.

Yet he cited a March 2009 editorial in Annals of Neurology by editors Stephen L. Hauser, MD, and S. Claiborne Johnston, MD, PhD, entitled “Personalized Genetic Scans: With Gifts Like These…,” which recounted the story of a healthy 27-year-old man who had been given as a Christmas gift a personalized genetic scan from one of the five companies cited by the FDA, revealing that he was a carrier of genetic variants that increased risk for ALS. “The young man was frightened and the family devastated,” the editorial noted, though in fact the variant had only been identified in a single, not-yet-replicated, case-control study.

“In that study, the variant increased ALS risk by about 30 percent, typical of a low risk disease-associated single nucleotide polymorphism (SNP) marker,” Drs. Hauser and Johnston noted. “The test is less ‘predictive’ of ALS risk than by simply determining if the individual is an athlete or a smoker.”

The editorial described another case involving a patient who used personalized genetic screening to find a similarly misleading “risk” for MS.

And just as ”risk” can be inflated by some tests, the failure to find a genetic mutation can be equally misleading. “People can learn from these products that their risk for, say, diabetes is low and then on the basis of that decide that they don't have to worry about weight, diet, or exercise,” Dr. Mathews said.

“Typically, companies provide a kit consumers can use to send in a saliva sample, which is then analyzed; results from the analysis are usually accessible six to eight weeks later in a password-protected Web site.

Rachel Nagler, a spokesperson for 23andMe issued a statement from the company stating that it had received the FDA letters and would be responding to the agency directly.

“It is important to note that we disagree with the FDA's conclusion,” the company said in its statement. “We are sensitive to the FDA's concerns, but we believe that people have a right to know as much about their genes and their bodies as they choose.”

“We carefully explain to our customers that the data we provide is informational and educational in nature,” the company stated. “We also caution our customers that if they want to use the data and information to learn about their own genetic make-up for medical decision making, they do so in consultation with a physician or other healthcare professional. ”


But Dr. Bird told Neurology Today people may be using test results for different purposes. “Some of the results are genetic findings which are viewed by the consumer as risk factors for various health problems, and some are for attributes such as intelligence or height,” Dr. Bird said. “I know there is at least one company marketing tests for muscle strength or speed, and consumers are making decisions about which sports their children should participate in.”

“All of these results can have different sorts of implications,” he said. “Some are known risk factors for disease, some are inadequately verified associations, and some are mutations that can directly cause a disease. It's difficult for both physicians and the public to differentiate between those kinds of tests and I am not sure the companies do a good job of making clear which is which.”

Dr. Vance, professor of neurology and chair of the department of genetics at the University of Miami, suggested that much of the public still thinks of genetic risk for disease in terms of single-gene disorders, when in fact many conditions are complex disorders requiring multiple genetic mutations.

“The genes companies test for may be associated with a disorder in the research literature, but they have different weights,” he explained. “If you think of a scale, you may need a hundred pounds to get, say, Parkinson disease. But many of the genes being tested are five pound or two pound or 10 pound weights. So it's true they are risk alleles, but they really don't provide much real knowledge about the actual risk, and the average person is not going to be able to decipher that.”

Dr. Vance added that genomic medicine is coming but such knowledge as has been garnered to date is mostly useful to researchers striving to understand disease processes; it's not yet ready for mainstream medical use, he said.

“These companies are correct that [the use of genetic information for personal medical and health decision-making] is probably the future,” he said, “but we are not there yet.”


Hauser SL, Johnston SC. Message from the Editor: Personalized genetic scans: With gifts like these… Ann Neurol 2009;65:A7–A9.