A good diet for you may be bad for me

A cookie can give one person a sugar rush while barely affecting another person, a new study finds, indicating that a food’s glycemic index is in the eater.

People’s blood sugar rises or falls differently even when they eat the exact same fruit, bread, desserts, pizza and many other foods, researchers in Israel report November 19 in Cell. That suggests that diets should be tailored to individuals’ personal characteristics.

The researchers made the discovery after fitting 800 people with blood glucose monitors for a week. The people ate standard breakfasts supplied by the researchers. Although the volunteers all ate the same food, their blood glucose levels after eating those foods varied dramatically. Traits and behaviors such as body mass index, sleep, exercise, blood pressure, cholesterol levels and the kinds of microbes living in people’s intestines are associated with blood glucose responses to food, the researchers conclude.

Those findings indicate that blood sugar spikes after eating depend “not only on what you eat, but how your system processes that food,” says Clay Marsh, an epigenetics researcher at West Virginia University in Morgantown.

Such individual differences have been noticed in previous studies, says study coauthor Eran Elinav, an immunologist at the Weizmann Institute of Science in Rehovot, Israel. While previous studies dismissed the differences as flukes, “we’re actually quantifying it,” Elinav says. For instance, eating bread produced a postmeal blood sugar level rise of 44 milligrams per deciliter of blood per hour on average. But some people’s blood sugar rose as little as 15 mg/dl*h, while others had a spike as high as 79 mg/dl*h after eating the same amount of bread.

A team led by Elinav and Weizmann computational biologist Eran Segal created a computer algorithm that used 137 personal measurements to predict how much a person’s blood sugar would rise or fall after eating a certain food. When tested on a new group of 100 people, the algorithm correctly predicted the response about 70 percent of the time.

A third group of 26 participants were then given personalized meals. The computer algorithm analyzed each person and then picked diets for 12 of them. A nutritionist chose a “good” and “bad” diet for the remaining participants. Good diets were ones that that minimized blood sugar spikes after eating. Bad diets sent blood sugar skyrocketing. The diets contained the same amount of calories.

It turned out that foods on the “good” diet for one person were sometimes on another participant’s “bad” list, Segal says. For instance, one woman’s blood sugar spiked when she ate tomatoes. But tomatoes were on other people’s healthy list.

“What our data suggest is that relying on population averages is not only inaccurate, but may even be dangerous in some cases,” Elinav says.

For 10 of the 12 people, the computer algorithm correctly predicted responses to the good and bad foods. Nutritionists were equally good at predicting how a person would fare on a given diet, the team found. But the computerized approach could reach more people, the researchers say.

Mixes of microbes living in people’s guts, known as the gut microbiome, also changed with the good and bad diets. Bacteria help break down food and have been implicated in causing obesity and diabetes. This study can’t distinguish whether the microbiome is causing differences in blood sugar responses or being influenced by how a person responds to certain foods, says Peter Turnbaugh, a microbiome researcher at the University of California, San Francisco.

While Turnbaugh agrees that personalized diets would be better than blanket recommendations for improving health, he sees some caveats. “The frustrating thing about all this is that we can learn how to optimize the diet for a given person, but ultimately, you have to stick to that diet.”

---

Editor’s note: This article was updated on January 11, 2016, to clarify the measurement of the postmeal blood sugar rise.