Ultra-processed diets tracked with 40% more type 2 diabetes in 25,214 Australian adults

Across eight European countries, every extra tenth of daily food, by weight, that was ultra-processed went with a 17% higher rate of type 2 diabetes. Now researchers have applied the same yardstick to a large cohort of Australian adults and arrived at 16%.

Until now, long-term evidence on the association between ultra-processed food and type 2 diabetes had been lacking in Australia. Priscila Machado of Deakin University led the new study, published in August in the Journal of the Academy of Nutrition and Dietetics. It followed 25,214 participants aged 40 to 69, none with diabetes when they joined between 1990 and 1994.

Over an average of 11.6 years, the team counted 1,734 cases of type 2 diabetes, roughly one person in 15. The quarter of the group whose diets leaned hardest on ultra-processed food had a 40% higher risk of type 2 diabetes than the quarter whose diets leaned least. The researchers also found evidence of non-linearity, meaning the risk did not rise in a straight line as the ultra-processed share grew.

Counting ultra-processed food by weight in 25,214 Melbourne adults

The Melbourne Collaborative Cohort Study was set up in the early 1990s to examine how diet and other habits relate to common chronic diseases, especially cancers. Its planners deliberately over-sampled Southern European migrants to widen the range of lifestyles on record, and they make up about 30% of the cohort. Each recruit answered a food frequency questionnaire, a checklist of how often they ate each item on a long list. This one was developed specifically to measure dietary intake in the cohort.

Machado’s team sorted the reported foods by how processed each one was. Then they worked out what share of each person’s daily food, by weight, was ultra-processed. The category itself comes from Carlos Monteiro of the University of Sao Paulo, one of the paper’s authors. He coined the term when he developed the Nova classification system, which groups foods into four categories based on the degree of processing. Under it, ultra-processed foods are industrial formulations built from substances extracted from foods, plus additives to enhance taste, texture, appearance, and durability. Packaged snacks, carbonated soft drinks, instant noodles and ready-made meals all qualify.

Weight is a common measure in this research. French researchers who used it explained that they chose a weight ratio rather than an energy ratio to take into account products that do not provide energy, such as artificially sweetened beverages. A share by weight also gives heavy things, drinks especially, more pull in the score than dry snacks.

Diabetes was checked twice after enrollment. The first follow-up gathered people’s own reports by mailed questionnaire and by telephone. Of the new type 2 diabetes cases found that way, 76% had their diagnosis confirmed by their doctor. At the second, a case could come from a self-report, elevated plasma glucose on a blood test, or medication use. That test can catch diabetes nobody had diagnosed.

That setup earns a fair amount of trust. Diets were recorded years before any diagnosis, the group was large, and the later check did not rest on memory alone. It is still an observational study, which can show an association but cannot show cause. People who ate the most ultra-processed food may have differed in other ways that affect diabetes.

Body weight carried less than a fifth of the ultra-processed food and diabetes link

The team also asked how much of the link runs through body size. Body mass index, the standard weight-for-height score, carried 18% of the association. The authors describe it as an important mediator, meaning a factor on the path between diet and disease. How big that share looks may depend partly on the ruler. The eight-country study measured waist-to-height ratio instead, a gauge of fat around the middle. It potentially mediated 46.4% of the association, though its authors add that the estimate rests on strong assumptions.

Weight is the easiest route to picture. Reviewers writing in The BMJ, a major medical journal, cite experimental evidence linking ultra-processed diets to increased energy intake and weight gain, about 500 extra calories a day in one trial. To explain the overeating, they point to the faster eating rate and hyper-palatability of these foods, along with their calorie density. In other words, the foods go down fast and are hard to stop eating.

The rest of the link is harder to pin down. Some of it may be differences between people that the analysis could not remove. Reviewers have also proposed that intensive processing changes the food matrix, the physical structure of food, in ways that may affect digestion, nutrient absorption, and feelings of satiety, or fullness. Other suspects are additives, including non-sugar sweeteners and emulsifiers, and contaminants that migrate from packaging materials. Research into how these effects work is still in its infancy, and the Melbourne abstract reports no test of them.

Ultra-processed food studies agree on the diabetes link but split on which foods

Melbourne’s result fits a pattern. In France, a study of 104,707 participants aged 18 years or older linked each 10-point rise in the ultra-processed share to a 15% higher risk. The same BMJ reviewers graded type 2 diabetes as one of only four health outcomes with convincing evidence, their top grade, based on a combined analysis of seven cohorts.

The eight-country team split the category further, and the pieces did not move together. Savory snacks, animal-based products, ready-to-eat or heat-up mixed dishes, and drinks sweetened with sugar or artificial sweeteners went with more diabetes. Breads, biscuits and breakfast cereals, sweets and desserts, and plant-based alternatives went with less.

“Simply saying that the entire group of UPFs can have bad health consequences is just too simple,” said Kevin McConway, an emeritus professor of applied statistics at the Open University, commenting on that study. The Melbourne abstract does not break its score down by food type, so it cannot show whether the same split holds in Australia.

A 1990s food questionnaire, answered before ultra-processed food had a name

Each person’s diet was recorded once, when they joined. Eating habits can shift over a decade or more, and so can the food on sale. Sales of ultra-processed food and obesity rates have climbed together, as researchers have observed worldwide, including in Australia.

The questionnaire also predates the category it was used to measure. The Nova classification system, which introduced the concept of ultra-processed foods, was first published in 2009. So the answers had to be sorted into groups defined 15 to 19 years after they were given. Janet Cade, a nutritional epidemiologist at the University of Leeds, raised the same issue about the eight-country study. Its data were collected before the Nova system had been created, she noted, so applying it required a lot of assumption.

The published abstract, which this account rests on, leaves one more gap. It does not list which factors the team adjusted for, so there is no way to judge here how fully those differences between people were handled.

Where cutting back on ultra-processed food fits in type 2 diabetes prevention

Whether eating less ultra-processed food would prevent type 2 diabetes remains unsettled. When the BMJ reviewers searched the literature, they found no existing pooled analyses of randomized trials on ultra-processed food and health.

Weight loss, by contrast, has been tested. You can help prevent or delay type 2 diabetes by losing a modest amount of weight and being physically active most days of the week, according to the US National Institute of Diabetes and Digestive and Kidney Diseases. Its advice also includes a step that matches the sweetened-drink result above: drink water instead of sweetened beverages.

Diabetes Australia suggests starting with additions, by putting more whole, nutrient-dense foods into your diet. For spotting ultra-processed foods, it points to the ingredients list, since they usually have five or more ingredients, many of which are not found in a normal kitchen.

Australia now joins the places where more ultra-processed food has gone with more type 2 diabetes, still without a settled answer on which foods matter.