Last November, the U.S. Food and Drug Administration issued a “cease and desist” order to 23andMe, a major purveyor of direct-to-consumer (DTC) genetic testing. In its letter to the company—issued after three prior warnings—the FDA reiterated its view that 23andMe’s Personal Genome Service (PGS) constitutes a medical device requiring further premarket evaluation:
FDA is concerned about the public health consequences of inaccurate results from the PGS device…we still do not have any assurance that the firm has analytically or clinically validated the PGS for its intended uses.
The FDA’s order, based on potential rather than actual medical harm, has generated a great deal of controversy. In a recent critique published in Nature, Robert Green, MD, MPH, of the Partners HealthCare Center for Personalized Genetic Medicine, and Nita Farahany, PhD, JD, of the Duke Institute for Genome Sciences and Policy, argued against regulating DTC genomic interpretation services as medical devices:
… doing so could put FDA regulations in greater tension with the First Amendment of the US Constitution, which protects the rights of individuals to receive information, and of ‘commercial speech’ ….the agency should avoid restricting consumer genomic testing unless faced with empirical evidence of harm.
Now, a case report in Clinical Pharmacology & Therapeutics provides what may be the first evidence of potential harm caused by errors in test interpretation by a DTC company.
A case of mistaken identity
Researchers Catherine Brownstein, PhD, David Margulies, MD, and Shannon Manzi, PharmD, of the Division of Genetics and Genomics at Boston Children’s Hospital, describe the case of a man they call Dr. J, who had possible Crohn’s disease. Test results provided by a DTC company revealed that he carried mutations in a gene called TPMT that conferred a near-complete loss of the TPMT enzyme, suggesting a high risk for severe toxicity from thiopurine drugs, which include 6-mercaptopurine, commonly used to treat Crohn’s.
This result seemed questionable for several reasons, not least of which was that testing by the same DTC company revealed the same genetic variant in two of Dr. J’s colleagues at Boston Children’s Hospital—a highly improbable result given their very different ethnic backgrounds. Seeking to confirm Dr. J’s results, Brownstein, Margulies and Manzi proceeded to genotype Dr. J and his parents, wife and children in a CLIA–approved laboratory. Claritas Genomics (a company spun off from Boston Children’s last year) sequenced the TPMT gene in each.
There the team found the smoking gun (or lack thereof): None of Dr. J’s family members had the variants reported by the DTC company. Nor did Dr. J himself, although he did have a different variant conferring partial loss of enzyme function.
Had this erroneous finding been used to guide Dr. J’s treatment for Crohn’s disease, clinical guidelines would dictate that he receive just one tenth of the standard thiopurine dose, potentially allowing his disease to progress or his symptoms worsen. Likely, his care team would recognize and address any progression. However, as the researchers note, thiopurines are also used to treat leukemia, where dose restrictions could be life-threatening.
Misleading reassurance
Meanwhile, the DTC company reassured Mrs. J, who has a family history of breast cancer, that she is at “typical risk” for breast cancer, with a four-star confidence rating. In reality, though, she had not been comprehensively tested for known breast cancer mutations, but only for three mutations known to increase breast cancer risk in Ashkenazi Jewish women (Mrs. J is not of Ashkenazi descent).
“It did state in the expanded report that she was only tested for a few mutations, but this wasn’t clear from the summary page,” says Brownstein.
While the DTC company provided disclaimers and offered consultations with a genetic counselor (at a premium), Brownstein and colleagues write: “the most reasonable interpretation of the genotype was not presented clearly, and patients may not realize that they need interpretative support.”
Is this a paternalistic view? In their commentary, Green and Farahany cite survey data (from their own and others’ work) indicating that most consumers handle medical genomic information appropriately, sharing it with health care providers and often using it to improve their health behaviors.
But the Boston Children’s team, which has sought to establish best practices for genomic interpretation, would prefer to err on the side of caution. “How consumers and health care providers handle the information isn’t the relevant question if that information is misleading or presented improperly,” notes Brownstein.
That said, Brownstein and colleagues write, “We are not advocating the end of DTC testing, nor are we universally denouncing DTC genetic testing companies. However, we feel that the interpretations must be accurate and reasonable, with adequate and freely available interpretation support for consumers and physicians.”
