The top of the Nourish3d homepage cycles through images of three young adults shown next to the seven ingredients in their own personalized daily supplements. Olivia’s Nourish3d includes ginger, marine collagen, and an “omega blend.” Mark needs tart cherry and vitamins A and E. Sophie gets green tea extract and something called CoQ10.
Like Olivia, Mark, and Sophie, consumers in the UK will be able to order supplements of their own that are customized to their goals and lifestyle beginning later this month, says Nourish3d CEO Melissa Snover. Each person’s selected ingredients, as determined by a questionnaire, will be “conveniently 3D printed into a tasty fruity, gummy stack,” according to the website, and mailed to them for £40 (about $50 US) per month or £360 per year.
Nourish3d’s 3D-printed spin on supplements has garnered some publicity, including a profile of Snover in Inc. and...
Personalized supplements: worth the money?
Vitamins and other dietary supplements are a $30 billion industry in the US. In addition to the bottles lining drugstore shelves, those products include a niche that promises to meet customers’ unique needs based on their specific data.
Those data come in a few different forms. Like Nourish3d, some other companies, including care/of and Persona, use an online questionnaire that covers questions such as goals and current diet. Persona and care/of then recommend personalized vitamin packs for customers to order.
Other companies use DNA testing to make their recommendations. PureGenomics, available only through healthcare providers, uses the results of 23andMe’s direct-to-consumer SNP test to sell its supplements. For example, the company states on its website, people with certain alleles of the gene BCMO1 may require a vitamin A supplement, if they aren’t eating foods high in the vitamin. TeloYears, by contrast, zeros in on telomere length, testing customers and delivering supplements it claims result in “measurably younger cells.”
Using a questionnaire to make supplement recommendations “cannot be particularly accurate . . . because it’s self reported, and it’s hard to know what it means,” says Marion Nestle, an emerita food researcher at New York University. “Many of the components of [Nourish3d’s] supplements have no known, scientifically based value for anything having to do with human health. Or if there is science, it’s really iffy.”
You’re still not entirely sure what science was used to get those particular variants onto their panel of variants that they’re testing for.—David Mutch, University of Guelph
As for the DNA-based services, nutitional genomicist José Ordovás of Tufts University in Boston credits PureGenomics for going through healthcare providers with its offerings, but says the research basis for its service is shaky. “[W]hat they are doing is to extrapolate findings from studies with low scientific evidence (observational studies) to clinical practice without any hard evidence that it will work,” he writes in an email to The Scientist.
Take the BCMO1 alleles used by PureGenomics to suggest a possible need for vitamin A supplements. “As far as I can see, there is not enough information out there in the moment to support such a claim,” writes Georg Lietz, a nutrition researcher at Newcastle University who’s studied the link between the gene and vitamin A levels in the body, in an email. “[T]he interaction of nutrients with genotypes is complex and involves a lot of different genetic variations, so it is difficult to pin this down to one or two variations only,” he adds. Ohio State University nutrition researcher Ouliana Ziouzenkova, who also studies vitamin A, writes to The Scientist that she would recommend that people with some BCMO1 mutations take supplements of the vitamin for one month each year, particularly if they are vegans who don't eat dairy products and eggs, which are high in vitamin A. That recommendation “is my personal opinion based on body of evidence but it is not supported by a specific study,” she says.
In the case of telomere-lengthening supplements, the very premise may be faulty. Mary Armanios, who studies telomere-linked disease at Johns Hopkins University School of Medicine, noted in an article last year that the view that short telomeres are bad and equal aging is oversimplified; she calls direct-to-consumer telomere tests “a form of molecular palm reading.”
Individualized diet advice
Discerning the value of companies that offer dietary advice instead of (or in addition to) supplements is somewhat murkier. David Mutch, a nutritional biochemist at the University of Guelph in Canada, says he sees such companies as “paving the way for precision nutrition or personalized nutrition,” an area he sees as promising. But, he adds, “I do think it’s, in many ways, a bit premature.”
Among the companies using SNP testing are Vitagene (which makes recommendations for diet, exercise, and supplements); InsideTracker, which makes nutrition, fitness, and lifestyle recommendations based on DNA testing, blood biomarkers, and a questionnaire; and Nutrigenomix, which uses a DNA test and is only available through healthcare providers.
Although he says he thinks there isn’t yet a solid research foundation for such services, Mutch adds that the specific studies underlying their recommendations aren’t always apparent. “I’d also even say that the scientists working in the field, such as myself, would struggle [to evaluate the evidence behind personalized nutrition companies’ products] because it’s really a bit of a black box,” and even for those companies that disclose what exactly they test for, “you’re still not entirely sure what science was used to get those particular variants onto their panel of variants that they’re testing for.”
Mutch also says it’s important to present information to consumers in a way that’s understandable and likely to lead to action, and that few studies have been conducted on how best to do that. With multiple SNP results and recommendations in a report, he says, “At what point do we run the risk of saying, we’re just giving people too much information and it’s almost now that too much information is leading to inaction?”
Asked for examples of the variants behind Nutrigenomix’s recommendations, and the published studies underlying them, Nutrigenomix founder Ahmed El-Sohemy referred The Scientist to a video of a talk he gave at a November 2018 conference. In it, he cites results from a case-control study suggesting that people with a “fast-metabolizer” variant in a gene for an enzyme that breaks down caffeine have a lower risk of a heart attack if they drink several cups of coffee per day, while those with slow-metabolizer variants increase their risk by drinking coffee. El-Sohemy also gave the example of a variant in the FTO gene that another group associated with a propensity to lose weight when on a high-protein diet.
El-Sohemy, who is a nutrition professor at the University of Toronto in addition to his role at Nutrigenomix, has a response to critics who point out that the recommendations of companies like his are based on association studies, rather than on the gold standard for medical evidence, randomized controlled trials (RCTs): “What many of these skeptics fail to recognize is that the vast majority of dietary advice that we currently give is not based on evidence from RCTs . . . we base it on the totality of evidence, and oftentimes, that comes from observational studies,” he says in his talk.
The emerging data on personalized nutrition advice
A few personalized nutrition companies have published studies on their products’ results. El-Sohemy and a colleague found that compared with dietary advice not based on genetics, advice that incorporated subjects’ genotype information made people more likely to reduce their salt intake if told they had a variant that put them at high risk of developing hypertension as a result of sodium consumption. Yet for other dietary components analyzed in the study, such as caffeine and added sugars, there was no significant difference between the behaviors of the genetic-advice group and of controls.
How about food? That would be my starting position, is eat a healthy diet for heaven’s sakes.—Marion Nestle, New York University
InsideTracker has also published results—in this case, of a study of 1,032 customers showing that levels of blood biomarkers that were considered outside the normal range at initial testing had a “trend toward normalcy” over time with the company’s advice. Ordovás, whose former PhD student works at InsideTracker, says it’s commendable the company is publishing research, but notes that the study “was retrospective rather than prospective, and it was not randomized. . . . In the published paper the company does not make any use of genetic information to tailor the dietary/lifestyle recommendations.” (Ordovas is nominally a member of Nutrigenomix’s scientific advisory board, although he says he has never met with the company or received payment from it. He is also involved in the PREDICT 2 study, which is funded by the personalized nutrition startup Zoe.)
In general, Ordovás writes, “we still have so many holes in our knowledge that using a SNP to make recommendations is like if we are building a jigsaw puzzle for which we don’t have a picture of the final product. . . . We need to integrate that piece with other pieces of the puzzle to see where it fits and what it means.”
DNA isn’t the only tool companies are using to tailor dietary advice; emerging insights into the many effects of the gut microbiome have led to an interest in using its composition to make recommendations. In 2015, a research group led by Eran Elinav and Eran Segal of Israel’s Weizmann Institute reported using artificial intelligence to predict, based on healthy people’s microbiome and other parameters, whether their blood glucose would spike after a particular meal—a risk factor for diabetes. That technology has since been commercialized as a program, DayTwo, that aims to help people with type 2 diabetes control their blood sugar.
While he says he doesn’t know the details of the algorithms used for DayTwo and doesn’t endorse the company, Paul Franks, a genetic epidemiologist at Lund University in Sweden, says the study that preceded its launch was “head and shoulders above the quality of much of the other research that’s been done around nutritional genomics.” Franks, like Ordovás, is involved in the PREDICT 2 study.
Rossi, whose own research focuses on tailoring diet to an individual’s microbiome to treat irritable bowel syndrome, writes in an email that “the commercialisation of these predictive tools is an example of industry outpacing the science.” She says that DayTwo has not been clinically validated; a study described in a recent meeting abstract on its results lacked a control arm, she notes, and “it is likely that standard dietary advice could have also achieved” the improvements seen.
Despite her impressions of current personalized nutrition offerings, Rossi says she thinks that with further research, approaches will emerge that can treat or prevent disease. Similarly, Mutch emphasizes that he’s “very optimistic for the for the field, for sure,” while at the same time, “I think there’s just a lot of room still for scientific evidence to really substantiate some of the products out there.”
Future possibilities aside, Nestle’s advice to people looking to optimize their nutritional intake right now isn’t complicated. “How about food?” she suggests. “That would be my starting position, is eat a healthy diet for heaven’s sakes.”
Clarification (August 21): The summary of Ouliana Ziouzenkova's comments on vitamin A supplementation has been amended to specify that in her view, vitamin A supplementation is most important for vegans with certain genetic variants.