A large European study has added new nuance to one of the most heated debates in modern nutrition: whether all ultra‑processed foods should be treated as equal threats to metabolic health.

Drawing on dietary and health data from more than 300,000 adults across eight European countries, researchers report that higher consumption of ultra‑processed foods is linked with a significantly greater risk of developing type 2 diabetes.

Yet the picture is far from black and white. Some ultra‑processed foods appear neutral. Others are even associated with a lower risk.

The findings, published in a leading international medical journal The Lancet and led by scientists working with the European Prospective Investigation into Cancer and Nutrition (EPIC), suggest that public health advice may need to move beyond blanket warnings.

Instead, the focus may shift towards identifying and reducing specific types of ultra‑processed products that carry the greatest metabolic harm.

Type 2 diabetes remains one of the fastest‑growing global health challenges. More than half a billion people are currently affected worldwide, with numbers rising sharply in Malaysia and elsewhere.

The condition increases the risk of heart disease, stroke, kidney failure and premature death, while also placing enormous pressure on healthcare systems. Diet is a major modifiable risk factor, yet debate continues over what matters most: nutrients, calories, or how foods are processed.

Ultra‑processed foods, often abbreviated as UPFs, have become central to this debate. These products are typically industrial formulations made from refined ingredients, additives, flavourings and emulsifiers. They often bear little resemblance to their original food sources.

Common examples include sugary drinks, packaged snacks, processed meats, ready meals and many breakfast cereals. Consumption has risen steadily over recent decades, driven by convenience, affordability and aggressive marketing.

To investigate how food processing relates to diabetes risk, researchers turned to the EPIC cohort, one of the largest and most detailed nutrition studies ever conducted in Europe. Participants were recruited in the 1990s from centres across the United Kingdom, France, Germany, Italy, Spain, the Netherlands, Denmark and Sweden. Most were aged between 35 and 70 at enrolment and were followed for more than a decade.

Dietary intake was assessed at baseline using country‑specific questionnaires. Each food item was classified using the Nova system, which groups foods by degree of processing rather than nutrient content. Unprocessed or minimally processed foods include fresh fruit, vegetables, meat, fish and grains. Processed culinary ingredients cover items such as oils, sugar and salt. Processed foods include bread, cheese and canned vegetables. Ultra‑processed foods form the fourth and most industrialised category.

Over an average follow‑up of nearly 11 years, more than 14,000 participants developed type 2 diabetes. When researchers examined overall patterns, a clear trend emerged. Every 10 per cent increase in the proportion of the daily diet coming from ultra‑processed foods was associated with a 17 per cent higher risk of developing diabetes. This association remained robust after accounting for age, sex, education, smoking, physical activity, alcohol intake and total energy consumption.

In contrast, higher intakes of less processed foods told a different story. Greater consumption of unprocessed or minimally processed foods, together with culinary ingredients used in home cooking, was linked with a lower risk of diabetes.

The same was true for processed foods such as bread, yoghurt and cheese. Substitution analyses added further insight. Replacing 10 per cent of daily intake from ultra‑processed foods with foods from lower processing categories was associated with a meaningful reduction in diabetes risk.

At first glance, these findings appear to support the growing body of evidence warning against ultra‑processed diets. Previous meta‑analyses have linked high UPF intake with obesity, cardiovascular disease and diabetes. International health agencies have called for stronger policy measures, including marketing restrictions and reformulation targets. However, the new EPIC analysis complicates the narrative.

When researchers broke ultra‑processed foods down into sub‑groups, striking differences appeared. Not all UPFs behaved in the same way. Sugary and artificially sweetened beverages were strongly associated with a higher diabetes risk. So were savoury snacks, ultra‑processed animal‑based products such as processed meats, and ready‑to‑eat or ready‑to‑heat mixed dishes.

Yet other ultra‑processed foods showed the opposite pattern. Breads, biscuits and breakfast cereals classified as ultra‑processed were associated with a lower risk of developing diabetes. Sweets and desserts also showed an inverse association, as did plant‑based meat and dairy alternatives.

Some categories, including sauces, spreads, condiments and ultra‑processed alcoholic drinks, showed no clear link at all.

These contrasting findings highlight the complexity of modern food systems. Ultra‑processed foods are a broad and heterogeneous group. Their nutritional profiles, energy density, fibre content and additive use vary widely. Some are fortified with vitamins and minerals. Others contain whole‑grain ingredients. Many breakfast cereals and breads, for example, contribute fibre intake in European diets, despite their industrial processing.

Experts note that these results do not suggest sweets or biscuits should be actively encouraged. Instead, they raise questions about whether processing alone is a sufficient marker of health risk. Reverse causality may also play a role. People at higher risk of diabetes may reduce their intake of sugary treats, leading to an apparent protective association in observational data.

The study also explored why ultra‑processed foods might increase diabetes risk. Body fat distribution emerged as a key factor. Measures of central adiposity, particularly waist‑to‑height ratio, explained nearly half of the association between UPF intake and diabetes. This finding aligns with previous research linking ultra‑processed diets to weight gain and abdominal obesity. These foods tend to be energy‑dense, highly palatable and easy to overconsume.

Importantly, the association between ultra‑processed foods and diabetes was not fully explained by traditional nutritional factors. Adjusting for saturated fat, sugar and sodium intake only modestly reduced the risk estimates. Nor did overall diet quality, measured using Mediterranean diet scores or front‑of‑pack nutrient profiling systems, eliminate the link. This suggests that aspects of processing itself may play an independent role.

Several mechanisms have been proposed in the scientific literature. Industrial processing can alter the physical structure of food, affecting satiety signals and glucose absorption. Additives may influence gut microbiota or inflammatory pathways. Packaging chemicals and neo‑formed contaminants have also been raised as potential contributors, although evidence remains limited.

Despite its scale and depth, the study has limitations. Dietary data were collected at a single time point, meaning changes in eating habits over time were not captured. Food processing methods have evolved since the 1990s, when participants were recruited. Ultra‑processed food consumption was lower then than it is today, particularly in southern European countries. The findings may therefore underestimate current risks.

There is also the possibility of residual confounding. Lifestyle factors such as stress, sleep and income, which can influence both diet and diabetes risk, are difficult to measure fully. The researchers addressed this by calculating so‑called E‑values, which suggest that an unmeasured confounder would need a moderate association with both diet and diabetes to completely explain the results for ultra‑processed foods.

Even so, the consistency of the findings across multiple analyses lends weight to the conclusions. Sensitivity tests, alternative exposure definitions and exclusion of early cases produced similar results. The large sample size and rigorous case verification strengthen confidence in the outcomes.

For policymakers and clinicians, the implications are significant. Current dietary guidance in many countries focuses on nutrients such as fat, sugar and salt. Some governments have begun to incorporate food processing concepts, while others remain cautious. The new evidence suggests that a simple message to avoid all ultra‑processed foods may be too crude.

Instead, experts argue for a more targeted approach. Reducing consumption of sugar‑sweetened and artificially sweetened drinks appears consistently beneficial. Limiting processed meats and ready meals also aligns with existing guidance. At the same time, recognising that some industrially processed foods can fit within a balanced diet may help avoid unnecessary confusion.

Public health messaging must balance clarity with nuance. Overly simplistic rules risk undermining trust or pushing people towards less practical diets. Food environments matter too. Ultra‑processed products dominate supermarket shelves and are often cheaper and more accessible than fresh foods. Structural changes, not just individual choice, will be needed to shift consumption patterns at scale.

The EPIC findings add an important layer to the ongoing conversation. They reinforce concerns about heavy reliance on ultra‑processed foods while challenging the idea that processing alone determines health impact.

As researchers continue to unravel how modern diets influence chronic disease, one message remains steady: what we eat, and how it is made, both matter.