Red meat sugar, ultra-processed foods linked to type 2 diabetes risk in new studies

Two new studies add evidence that diet may shape type 2 diabetes risk. One found that a sugar molecule in red meat is independently associated with higher risk. The other found that people who ate fewer ultra-processed foods and got regular moderate-to-vigorous physical activity had lower risk, including those with a strong genetic predisposition. Both were presented this week at the annual meeting of the European Association for the Study of Diabetes (EASD) in Milan, Italy.

The first study, from researchers in France, found that adults with the highest intake of meat-derived N-glycolylneuraminic acid (Neu5Gc) had a 63% higher risk of developing type 2 diabetes than those with the lowest intake. The authors say this points to chronic inflammation as a possible mechanism linking red meat to the disease.

The second study, from researchers in Hong Kong using U.K. data, found that among people at high genetic risk, combining low ultra-processed food intake with guideline-level exercise was associated with a 49% lower risk of type 2 diabetes. That study was also published in the journal Diabetes.

Both studies are observational and cannot establish cause and effect. Authors of the Neu5Gc study said their results are preliminary, have not yet been validated and may change once the analysis is finalized.

A red meat sugar the immune system treats as foreign

Most mammals produce Neu5Gc, but humans lost the ability because of an ancient mutation in the CMAH gene. The sugar is found in red meat, particularly beef and lamb, and in smaller amounts in dairy products. When people eat these foods, the Neu5Gc is absorbed and incorporated into human tissues, where it acts as a xeno-autoantigen. The body produces antibodies against it, and the resulting chronic immune activation, known as xenosialitis, has previously been linked to cardiovascular disease and cancer. Until now, however, no large prospective study had examined whether it might also promote type 2 diabetes.

Researchers analyzed data on 89,953 adults in NutriNet-Santé, a long-running French prospective cohort. Participants had an average age of 42, 79% were female, and none had type 2 diabetes or cancer at baseline. Each person completed three 24-hour dietary records. The researchers estimated daily Neu5Gc intake by matching those records against a validated database of the Neu5Gc content of foods.

The researchers defined red meat as beef, lamb, pork, rabbit, goat and game meats. They analyzed processed meats and dairy as separate food groups. Participants were then divided into quintiles of meat-derived Neu5Gc intake, calculated separately for men and women.

Over a median follow-up of 7.2 years, 966 participants developed type 2 diabetes. The researchers adjusted for age, sex, physical activity, dietary quality, smoking status and underlying health conditions. After those adjustments, the risk comparisons with the lowest quintile were:

• Highest quintile: 63% higher risk (hazard ratio [HR], 1.63; 95% CI, 1.29–2.06).
• Fourth quintile: 41% higher risk (HR, 1.41; 95% CI, 1.11–1.77).
• Second and third quintiles: no statistically significant difference.

The researchers said the steep rise in risk at higher exposure levels suggests a possible threshold mechanism. In that scenario, the immune system becomes overwhelmed once tissue incorporation of Neu5Gc passes a certain level.

"Our findings support chronic inflammation as a specific mechanism linking red meat consumption with T2D, partly beyond the amount of red meat a person eats and partly independent of body fat," said lead author Leopold Fezeu of Université Sorbonne Paris Nord in Bobigny, France.

The association appeared in both men and women, with no statistically significant difference between the sexes, and it was consistent regardless of energy intake. Adjusting for body mass index weakened the association, but it remained significant. The researchers estimated that obesity accounted for about 13% of the association. That suggests most of the effect works through mechanisms independent of weight gain, most likely direct inflammatory effects on insulin-producing cells and insulin-sensitive tissues, though the authors said more research is needed to confirm the pathway.

When the researchers also accounted for total red meat intake, the association pointed in the same direction but was no longer statistically significant. The authors said this suggests Neu5Gc may explain part, but not all, of the link between red meat and type 2 diabetes.

The researchers found no association between dairy-derived Neu5Gc and diabetes. They noted that cow's milk contains less Neu5Gc than beef or lamb and is typically eaten in smaller servings. People in the highest quintile of total dietary Neu5Gc, from all sources combined, had a 37% higher risk than those in the lowest, an association the researchers said was largely driven by red meat.

"Together, these findings suggest that the association between Neu5Gc and T2D may depend on its dietary source, with meat-derived Neu5Gc showing a stronger association with T2D risk than total dietary Neu5Gc," Fezeu said.

The researchers also reported typical meat intake in the top and bottom quintiles:

• Men, highest quintile: a median of 97 grams of red meat per day, or 136 grams including processed meat.
• Women, highest quintile: 73 grams of red meat, or 96 grams including processed meat.
• Lowest quintile: 5 grams (21 grams) for men and 0 grams (9 grams) for women.

The authors cautioned that these figures describe what people in each group typically ate. They should not be read as a threshold or a "safe" amount of meat.

Fezeu said future research should explore whether reducing exposure to meat-derived Neu5Gc could help prevent type 2 diabetes, for example by favoring poultry or plant-based proteins over beef and lamb, which have the highest Neu5Gc content. The team also plans to directly measure anti-Neu5Gc antibodies in large cohorts to test the xenosialitis mechanism in humans. They also plan to examine how Neu5Gc and other red meat components, such as heme iron, jointly relate to diabetes risk.

The authors acknowledged several limitations. They could not rule out confounding from unmeasured lifestyle and social factors. Estimating Neu5Gc intake from dietary records, without tracking dietary changes during follow-up, may also have introduced measurement error.

Lifestyle factors linked to lower risk across genetic groups

The second study, led by Wenxin Xu of City University of Hong Kong, examined whether lower intake of ultra-processed foods and regular exercise were associated with lower type 2 diabetes risk across levels of genetic susceptibility.

"Unlike genetic susceptibility, diet and physical activity can be modified — they are things people can change," Xu said.

The researchers analyzed 56,964 adults in the UK Biobank study, who had an average age of 69.5, and 42.3% of them were male. Ultra-processed food intake was assessed from 24-hour dietary recalls and classified using the Nova system. Moderate-to-vigorous physical activity was measured by wrist-worn activity trackers over one week. Genetic susceptibility was estimated from genetic data.

During an average follow-up of 7.8 years, 926 participants developed type 2 diabetes. For every 10% of a participant's diet by weight that came from ultra-processed foods, diabetes risk was 7% higher. Participants who met the World Health Organization recommendation of at least 150 minutes of moderate-to-vigorous activity per week had a 39% lower risk than those who exercised less.

Among participants at high genetic risk, two findings stood out:

• Low ultra-processed food intake (about 11% or less of total food by weight) was associated with a 35% lower risk.
• Meeting activity guidelines was associated with a 33% lower risk.

Physical activity was also associated with lower risk among people whose diets were high in ultra-processed foods, about 33% or more by weight. In that group, meeting activity guidelines was associated with a lower diabetes risk at every level of genetic susceptibility:

• High genetic risk: 31% lower risk.
• Intermediate genetic risk: 57% lower risk.
• Low genetic risk: 58% lower risk.

The researchers also compared people who ate fewer ultra-processed foods and met activity guidelines against those who did neither. The combination was associated with a lower risk across all genetic groups:

• High genetic susceptibility: 49% lower risk.
• Intermediate genetic susceptibility: 70% lower risk.
• Low genetic susceptibility: 59% lower risk.

These figures compare groups of people at a single point in time. They do not represent the reduction an individual patient could expect from changing their habits.

"Our study indicates that our genes are not our destiny," Xu said. "Even among people with a high genetic susceptibility to type 2 diabetes, lower intake of ultra-processed foods and regular moderate-to-vigorous physical activity were associated with a substantially lower risk of developing the condition."

The authors said the results highlight the value of addressing lifestyle factors together rather than one at a time. They acknowledged that the findings are observational, but noted that they align with existing advice to eat a healthier diet and stay physically active to prevent or delay type 2 diabetes.

Independent experts urge caution on interpretation

Independent researchers who reviewed the study through the U.K.'s Science Media Centre called it a large, useful addition to the evidence base. They cited its prospective design, objectively measured physical activity and adjustment for confounders, including overall diet quality. Several also raised limitations relevant to how physicians discuss the findings with patients.

• No behavior change was measured. Each participant contributed a single measure of ultra-processed food intake and physical activity. The study therefore could not assess whether changing those behaviors alters diabetes risk.
• One subgroup drove the food association. When the researchers broke ultra-processed foods into subgroups, only ultra-processed drinks remained significantly associated with diabetes after full adjustment. Reviewers said this suggests overall diet quality, including sugar-sweetened beverages, may matter more than processing level itself.
• Limited ancestry. The genetic analyses included only people of European ancestry, so it is unclear whether the findings apply to other populations.
• Small absolute effect. One statistician estimated that the 7% relative increase per 10% of the diet from ultra-processed foods works out to roughly one additional diabetes case per 1,000 people over about eight years.
• Physical activity was the stronger signal. Reviewers noted that activity was measured more reliably than diet and showed larger associations with diabetes risk.

Taken together, the two studies add to a growing body of evidence supporting lifestyle intervention in diabetes prevention. They also point physicians to specific factors to discuss with at-risk patients: the type of protein on the plate, overall diet quality and whether weekly activity meets guideline levels.