Read our COVID-19 research and news.

Even foods traditionally viewed as “healthy” cause blood sugar spikes in some people.

Even foods traditionally viewed as “healthy” cause blood sugar spikes in some people.

Gregory Dubus/iStockphoto

Why you shouldn’t always listen to dietary guidelines

Sure, an ice cream sundae will probably cause your blood sugar levels to spike. But a plate of tomatoes? Or a piece of toast? A new study reveals that different people have surprisingly different reactions to the foods they eat—a finding that may prompt doctors to rethink their one-size-fits-all diet advice.

“We tend to think that this obesity and diabetes epidemic is because people are not compliant with what you tell them,” says study author Eran Segal, a computational biologist at the Weizmann Institute of Science in Rehovot, Israel. “When you see these data, you think, well, maybe part of the problem is in what we’re telling them to do.”

Most current dietary advice is based on what seems to be healthiest on average for large groups of people. Nutritionists devise diets that reduce spikes in blood glucose, or blood sugar, by limiting peoples’ carbohydrate intake or by steering them away from foods with a high “glycemic index”—a measure of how much the average person’s blood glucose rises after eating a given food. But carbohydrate content and glycemic index can only offer vague guesses as to how any food will affect a particular person.

To get a better understanding of how diet and metabolism interact, Segal and his colleagues looked at the metabolisms of 800 people, collecting data on their blood composition, body measurements, lifestyle, and gut microbiome. For a week, they also monitored how participants’ blood glucose fluctuated in response to both simple meals provided by the researchers and meals they made for themselves.

What the researchers found was striking: Two peoples’ glucose responses to the same food could be as different as their reactions to entirely different meals, even though their own responses remained consistent over time, the team reports online today in Cell. The researchers analyzed the data using a machine learning algorithm, which they then developed into a system for predicting how someone would respond to any given food.

Segal and his team then used the findings to develop customized diet plans for a group of 26 people. The researchers chose foods based on which ones caused the smallest and largest spikes in blood glucose. Large glucose spikes are linked to a wide range of chronic diseases, including type 2 diabetes, a condition in which high blood glucose makes the body resistant to insulin. The diets varied widely, with foods on one person’s “good” diet sometimes showing up on another person’s “bad” diet. As predicted, “good” diets evened out peoples’ glucose spikes, whereas “bad” diets made them more extreme.

Not everyone agrees that researchers failed to appreciate peoples’ unique metabolisms in the past. “We've known for decades that one person will experience a much greater rise in blood sugar than another for many reasons unique to the individual,” says David Ludwig, an endocrinologist and nutrition researcher at Boston Children’s Hospital, who was not involved in the study. “This provides interesting quantitative data about the individual factors, but the basic point isn't new.”

The analysis of diverse health and lifestyle factors is valuable, Ludwig says, because it will help researchers understand why glucose responses vary. But customized diets based on the new algorithm aren’t ready for the clinic, he says. For one thing, the researchers have yet to pit their customized diets against generic low carbohydrate or low glycemic index diets. Until they do, it’s impossible to know how much the labor-intensive customization really helps. 

Such comparisons are the next stage of the project, Segal says. The researchers are planning a long-term study to see whether customized diets are more effective than traditional diets at managing blood glucose in patients at risk for diabetes. They are also working to simplify the customization process so that they can assign diets based on fewer types of information. 

Eleanor Scott, a diabetologist who manages patients at the University of Leeds in the United Kingdom, says she would welcome such a diet customization tool. Like Ludwig, Scott—who was not involved in the study—isn’t surprised that people vary in their responses to food; she struggles with such variation every day in her practice. She says that custom-tailored diets could be hugely helpful for groups of people with extra motivation to stick to them, such as recent heart attack victims and pregnant women with diabetes.

“We know that there is quite wide interindividual variability in responses to meals, and at the moment we tackle that in a very blunt way,” Scott says. By using body metrics, behavior, and gut bacteria to predict how people process food, doctors may some day be able to treat metabolic woes with more finesse.