What is echinacea? This much we can say with some certainty: it is a pretty, pink-flowered member of the daisy family—a native of the American prairie. It is also, according to a small body of (mostly German) scientific research, an immunostimulant. Its proponents contend that, consumed in tea, tincture, or capsule form, it can treat AIDS, boils, and the common cold and purge the liver of toxins. These claims have not been evaluated by the U.S. Food and Drug Administration, which declines to regulate the herb as a medicine, consigning it instead to the fluffier category of dietary supplement. Echinacea is also—or ought to be—the thing inside bottles labelled “ECHINACEA.” Last week, however, New York’s attorney general, Eric Schneiderman, shattered consumer faith in this seemingly self-evident truth. In cease-and-desist letters sent to Walgreens, Walmart, Target, and GNC, Schneiderman announced that his office had commissioned tests on a range of echinacea supplements and found them to contain mostly rice and buttercup DNA, if they contained any plant material at all. Other store-brand supplements appeared to fare no better: Target’s valerian root showed traces of garlic and wild carrot but no valerian; Walgreens’ St. John’s wort consisted of garlic, rice, and dracaena, a houseplant; Walmart’s ginkgo biloba had only dracaena, mustard, wheat, and radish.
The task of analyzing the supplements had fallen to Beckman Coulter Genomics, a company based in Massachusetts, and a Clarkson University biologist named James Schulte, whose academic specialty is in snake and lizard evolution. The group used a method known as DNA bar coding, which identifies an organism by a small segment of its mitochondrial DNA. (The segment is the same for members of the same species, but it mutates quickly enough, over generations of evolution, that even closely related species have distinct bar codes.) In 2004, Paul Hebert, a biologist at the University of Guelph, in Ontario, launched a bar-code database, making it possible, at least in theory, for scientists to identify mysterious species relatively quickly and easily, by extracting a specimen’s DNA, finding and amplifying its identifying sequence, and matching the sequence against the database. (There are between ten and a hundred million species on earth, according to current estimates. So far, Hebert’s project has logged bar codes corresponding to about a hundred and fifty thousand known animals and about sixty thousand known plants. It may take some time for the library to achieve Borgesian completeness.)
It was a 2013 Times article about the Guelph group’s research that inspired Schneiderman’s investigation. Though developed to assist frustrated conservationists, confused botanists, and anxious seed-bank administrators, DNA bar coding has proved useful in other areas where species identification is necessary, including pest control and bioterrorism monitoring, and it has recently become especially prevalent in food-related inquiries. In 2011, the F.D.A. formally adopted DNA bar coding as its standard method for identifying mislabelled fish, and just last month I.B.M. announced a partnership with the Mars company to “sequence the food supply chain” in order to improve food safety. The results of Schulte’s dietary-supplement analysis—lots of bar codes belonging to rice and houseplants, few belonging to curative herbs—would seem to be a slam dunk for Schneiderman. The attorney general’s cease-and-desist letters required that Target, Walmart, and the others furnish documentation about their supplier agreements and in-house authentication processes; they were also instructed to remove the suspect products from their shelves until the investigation is concluded. Meanwhile, consumers in New York and beyond were left with the feeling that they’d been taken for a ride, forking over money for pills that provided, at best, a placebo effect. The Times editorial board issued a recommendation for “stronger oversight.”
But there are some problems with the method that the attorney general’s scientists used. First, what we might call inert ingredients—wheat, rice, even wild carrots—are perfectly legal in dietary supplements, provided that they are kept below a certain level. According to Markus Lipp, a senior director at the United States Pharmacopeial Convention (U.S.P.), a nonprofit scientific organization that sets standards for food, medicine, and dietary-supplement ingredients, there is no indication that Schulte and his team took quantity into account when they performed their analysis. Though Lipp said that it is possible to gauge ingredient amounts using DNA testing, the process is more complicated than standard bar coding. Second, industrial processing can destroy a plant’s genetic material without affecting its supposed health benefits. “DNA is very fickle,” Lipp said. “It breaks apart as soon as it gets a little stressed.” (On Monday, GNC made this same argument in a press release that disputed the attorney general’s claims.) John Spink, the director of the Food Fraud Initiative, at Michigan State University, pointed out that standard manufacturing processes such as heating and irradiation, which is used to kill potentially harmful microbes, can destroy DNA. Lipp’s colleague Gabriel I. Giancaspro told me that bar coding is “not particularly useful for highly processed materials," adding that "for botanicals, we know that the solvents and the process to extract and purify the desirable compounds can exclude DNA.”
When I shared these concerns with a representative of the attorney general’s office, who spoke to me on the condition of anonymity, citing the ongoing investigation, she noted that more than seventy published papers have shown the efficacy of DNA bar coding in identifying ingredients accurately. She also asked a rhetorical question: Why would rice and radish DNA show up in the supplements while all the echinacea DNA had been destroyed? This response fails to address either the problem of quantity or that of the method of preparation—rice powder, for instance, is a filler, typically added to a supplement after most of the processing of the active ingredient has taken place, meaning that its DNA would likely remain intact. Nevertheless, Spink, Lipp, and Giancaspro were keen to stress that dietary-supplement fraud is a very real and widespread problem—though normally they expect to find it in the kinds of products that are advertised in junk e-mail rather than in the aisles of major retailers. Spink, in particular, said that he was thrilled to see a government agency taking an active role in attempting to ferret out fraud, which is in contrast to the F.D.A.’s largely hands-off approach. But, Lipp concluded, “Without details on what the attorney general’s office did, all we can say that we know is that they didn’t follow our standards.”
By federal law, all drugs that are sold in the United States must meet identity standards laid out by the U.S.P. These standards are developed by joint panels of academics, pharmaceutical-company representatives, and other experts. It is a lengthy, collaborative process that also incorporates a mandatory three-month public comment period. The same is not true of dietary supplements, which need only to be tested using, in the F.D.A.’s phrasing, “appropriate, scientifically valid methods.” The attorney general’s office seems to have decided that DNA bar coding fit the bill. The U.S.P., however, recommends a different method. Certifying an echinacea supplement, for instance, requires a battery of chemical tests that measure the presence or absence of particular phenols, alkylamides, and acids. Together, these can conclusively demonstrate the purity and strength of the active ingredient, irrespective of whether its genes are intact.
The U.S.P. does not currently recommend DNA bar coding in any of its standards. Alhough Giancaspro told me that he and his colleagues may add it to their tool kit in the future, he said that they “would never use it alone.” Dietary supplements aren’t always what they say they are, nor do they always do what they say they do. In this case, though, the attorney general's office has picked the wrong weapons for the right fight.
*Note: This post has been updated to correct the stated year in which Paul Hebert launched the bar-code database—it was in 2004, not 2010—and to clarify the types of species data that has so far been logged.
