In the last month, headlines have been touting the newest diet fad: Using genetic testing to personalize a diet plan based on food sensitivities and metabolic responsiveness.
The idea may sound like a good one—genes play a role in obesity, satiety, and overall health. So why not incorporate that information into a diet and health plan? However, the evidence that using this information to guide eating and lifestyle behavior makes dieters more successful is mixed. And the results of the studies only tell us whether people changed their behaviors within about a year, not that they reduced BMI, had better glucose levels or reduced their incidence of cardiovascular disease or cancers.
A major piece of evidence is a study of more than 1,000 Europeans, which found that personalizing diets helped people adopt healthier behaviors. But the level of personalization needed was surprising. Asking about food preferences was as successful as taking into account people’s health biomarkers and their genes. From the study, published in the International Journal of Epidemiology:
The main findings of this study were that, overall, PN advice was more effective in improving dietary behaviours when compared with conventional ‘one size fits all’ population-based advice. However, we found no evidence that including phenotypic or phenotypic plus genotypic information.
In the study, the control group was given generic nutritional information—the European equivalent of U.S. My Plate recommendations. The second group filled out a food frequency survey and their diet advice was tailored to the foods they already reported eating. The third group was given advice based on food preferences alongside body metrics like BMI and waist circumference and results of blood test like their fasting glucose levels. The researchers called this group the ‘phenotypic’ group. The final group of people were given a diet plan based on all the same information as the phenotypic group plus the test results of the five genes: MTHFR (folate, vitamin B metabolism), FTO (body weight and fat composition), TCF7L2 (blood sugar regulation), APOE ε4 (choleseterol) and FADS1
(fatty acid metabolism).
When the research team analyzed the results they found that the groups that got personalized dietary plans did better than those who received generic diet information. But they also found no statistical differences between groups who received good personalized nutrition. Adding blood test results or gene test results did no better than asking what foods people generally ate and tailoring a diet to suit those preferences. These results were in line with other papers that tested fewer genes, but found the same results. Adding genetic info didn’t change people’s behaviors or health outcomes.
But the same study group did find an effect for knowing one’s genetic information in a second study that tracked people’s adherence to a Mediterranean diet. When people were told of their genetic risks of the same five genes, they did improve their diets along a 14-point scale that measures adoption of the high-fiber, low saturated fat diet. The results were published in the American Journal of Clinical Nutrition. But the effect was small, 0.6 points on the scale. The researchers noted that it takes 2 points on the scale to lead to better health outcomes that are clinically significant.
So what about the growing number of companies that offer this kind of genetic testing?
Nutrigenomix is a Toronto-based startup with the largest public presence of these emerging companies. They test a set of 45 genes to help determine diet and athletic training plans to optimize nutrition and athletic performance. Nutrigenomix’s marketing materials say they offer personalized advice regarding coffee consumption, gluten sensitivity, protein balance for weight loss and physical activity guidelines.
Nutrigenomix founder Ahmed El-Sohemy told GQ that his research shows that when people are told of their genetic predispositions they are more likely to comply to dietary changes:
In a randomized controlled trial, the gold standard for scientific evidence, we gave people advice to limit their sodium intake a certain amount, and they didn’t do it. But when we said, “You have a gene that increases your risk of salt-sensitive hypertension”—just that one extra line of text then the same advice to limit their sodium intake to a certain amount—they did it and maintained it a year later. From a behavioral science standpoint that’s a pretty profound effect.
In the study, published in PLOS in 2014, 15 percent of patients who were told they had risk genes for sodium-sensitive hypertension reduced their intake of sodium to the daily recommendation of 1,500 mg. People who were told they didn’t have the risk gene didn’t change their behavior. Some studies find this effect, and some don’t.
Nutrigenomix requires a physician or other healthcare provider to order the tests, rather than go directly to consumers to fit with the FDA’s requirements. Their website helps patients locate a doctor who offers testing. The test costs roughly $400, tests for 45 metabolism-related genes and is not covered by insurance El-Sohemy said.
El-Sohemy and his company are far from the only health experts who are betting that personalized nutrition will be crucial to ending the obesity epidemic. Early this year another group found that people’s physiological response to eating the same food can be wildly different. From the New York Times:
“The same dietary advice cannot be good for everyone, because we are all different,” said Eran Elinav, an immunologist at the Weizmann Institute of Science in Rehovot, Israel. “This is why we have failed so miserably at controlling the obesity epidemic.” Dr. Elinav and his colleagues are hoping to build a new kind of diet-counseling business from findings they published recently in the journal Cell. Their study found a startling variation in the glucose responses of 800 subjects fed the same foods. Some participants had sharp increases in blood sugar when they ate ice cream and chocolate, while others showed only a flat or moderate response.
That study didn’t include genetic information but did include the participants’ microbiomes, the bacterial makeup in their guts.
Critics argue that the science is still too new to completely understand the relationships between our genes and our diets. Of the approximately 40 genes that have been associated with metabolism of nutrients or caffeine, for example, are causing deficiencies or obesity. This is related to a gene variants penetrance: how many people with a specific gene ever develop the condition that gene causes? From the Wall Street Journal:
Some experts say genetic science is too young to be able to provide nutritional advice. “We still have a long way to go in the development of practical tools,” says José M. Ordovás, director of nutrition and genomics at the Jean Mayer USDA Human Nutrition Research Center on Aging at Tufts University and a leading researcher in nutrigenomics. Dr. Ordovás says most nutritional genetic tests look at just a handful of genes, which isn’t enough to guide dietary decisions. His own research is aimed at evaluating the entire human genome, which consists of 20,000 to 25,000 genes.
It’s extremely likely that in ten years the science will have advanced enough to provide patients with a great deal of information about their personal metabolism and how they might use diet to optimize for their bodies. But right now, the information nutrigenomics tests can provide is limited. And the field in general suffers from mixed messages. If personalized diets are the key to solving the public health problem of obesity, we’ve got to find a way to personalize for millions of people. That means keeping costs low. And maybe, as the European study showed, simply listening to people’s preferences is enough personalization to make a difference.
This story was updated on 9/25/2016 based on further reporting and comment from Ahmed El-Sohemy.
Meredith Knight is a contributor to the human genetics section for Genetic Literacy Project and a freelance science and health writer in Austin, Texas. Follow her @meremereknight.