Health Policy$ense

The FDA, 23andMe, And What We Really Want From Genetic Testing

FDA Faults Company's DNA Product on Both Analytic and Clinical Validity

In a harshly worded letter released Monday, the Food and Drug Administration (FDA) ordered the company 23andMe to “immediately discontinue marketing” its saliva spit test (from which a consumer’s DNA is isolated) and personalized genome service until it provides the FDA with requested information on safety and effectiveness. I asked Reed Pyeritz, MD, PhD, a medical geneticist at Penn and former president of the American College of Medical Genetics, to comment on this development. He wrote:

The FDA 'Warning Letter' essentially shuts down 23andMe, although the company could potentially carve out an ancestry/genealogy product. There is an obvious conflict between those who argue a libertarian perspective that any person has a right to know his or her genome sequence and its implications and those who argue that misinformation or accurate information taken out of context can be misused in a medical context.

In my view, because the company is viewed by the FDA as offering a medical service, the FDA has a responsibility (under Federal statute) to require the laboratory component of 23andMe to perform to the standards of the Clinical Laboratory Improvement Act (CLIA). In addition to various quality measures, each test must have demonstrated analytic validity and clinical validity.  The FDA appropriately takes 23andMe to task on both measures.

Ironically, the FDA letter does not mention clinical utility, which is what consumers (and their health professionals) are really interested in.  In our studies of consumers who utilized direct-to-consumer genetic testing (not through 23andMe), the vast majority indicated that they would share the results with their primary health care providers, and want the genotype data put in their health record. When we surveyed the primary health care providers, most were unprepared to deal with the data. So, let's assume that the genotype data and their interpretation emanating from 23andMe are accurate. We are still left with the question of, 'so what?' There are precious few instances in which the kind of results available from 23andMe (BRCA1 and BRCA2 mutations being exceptions) should prompt definitive management decisions.  For the vast majority of results available from 23andMe genotyping, at best the consumer might be stimulated to adopt some sensible lifestyle and health maintenance adjustments that would be appropriate for all of us. Here, too, there are scant data that suggest long-term adherence to such adjustments, whether stimulated by a genotype report or the advice of a physician.