Chief Editor: Caleb Acquah (PhD)
GFI | SCIENCE WATCH
A large UK Biobank analysis links higher ultra-processed food intake to greater type 2 diabetes risk. Its more interesting finding is what happened when researchers added objectively measured sleep, physical activity and genetics to the picture.
Key Takeaways
- Among 56,964 UK Biobank adults without diabetes, each 10 percentage-point increase in ultra-processed food (UPF) intake was associated with a 7% higher type 2 diabetes risk (HR 1.07).
- Physical activity and sleep were measured objectively with 7-day wrist accelerometers rather than questionnaires.
- Among participants with both high UPF intake and high genetic risk, sufficient moderate-to-vigorous activity (HR 0.69) and adequate sleep (HR 0.42) were associated with lower diabetes hazard.
- These are observational associations. They do not show that exercise or sleep offsets the effects of UPFs.
- The findings sharpen, rather than settle, the question of which characteristics of UPFs matter for reformulation.
What the Study Measured
The analysis, published online September 17, 2026 in the American Diabetes Association journal Diabetes, followed UK Biobank participants who were free of diabetes at baseline [1]. Diet was assessed with 24-hour dietary records and classified using the NOVA system. Genetic susceptibility was estimated from 424 variants associated with type 2 diabetes.
What sets the study apart is lifestyle measurement. Participants wore accelerometers for seven consecutive days, giving researchers objective data on physical activity and sleep rather than the self-reported estimates most UPF cohort studies rely on.
The Headline Association
Higher UPF intake was associated with higher diabetes risk. Each 10 percentage-point increase corresponded to a 7% higher hazard after multivariable adjustment (HR 1.07, 95% CI 1.03 to 1.12), and the relationship was approximately linear.
That estimate is somewhat smaller than the 12% per 10% increment reported in a 2023 meta-analysis of more than 415,000 participants [2], but the direction is consistent with the wider evidence base.
Where Sleep and Activity Come In
The researchers then examined the highest-risk combination: participants with both high UPF intake and high genetic susceptibility. Within that group, sufficient moderate-to-vigorous physical activity was associated with lower diabetes hazard (HR 0.69, 95% CI 0.50 to 0.96). Adequate sleep showed a stronger association (HR 0.42, 95% CI 0.24 to 0.75). Expressed as relative differences, these correspond to roughly 31% and 58% lower hazard.
What the Study Cannot Tell Us
This was an observational cohort, not a randomized trial. An HR of 0.42 does not mean that sleeping more will cut diabetes risk by 58% in people who eat large amounts of UPFs. Residual confounding is also plausible: people who sleep well and exercise regularly may differ in income, overall diet quality, body composition and other health behaviours in ways statistical models do not fully capture.
The accurate reading is that sleep and activity were associated with lower risk, not that they cancel out diet.
Why It Matters for Manufacturers
The study does not weaken the case that diet matters; UPF intake tracked with diabetes risk. What it challenges is the idea that one feature of the food supply can explain a multifactorial disease on its own. Diet, activity, sleep, genetics and body weight operate simultaneously.
For manufacturers, a harder problem remains. NOVA classifies foods by the nature and purpose of processing, not by nutrient profile. A population-level association with UPF intake therefore does not identify which product attributes to change: processing itself, energy density, added sugars, refined carbohydrates, sodium, low fibre, food structure, or displacement of minimally processed foods.
Earlier U.S. cohort data already hint that UPFs are not uniform. Sugar-sweetened beverages and processed meats were linked to higher risk, while ultra-processed whole-grain breads, breakfast cereals and yogurt showed inverse associations [2]. That heterogeneity is exactly what reformulation strategy depends on resolving.
GFI Watch
Subgroup analyses: Follow-up work that breaks UPF intake into product categories would give manufacturers usable signals.
Intervention trials: Testing sleep or activity changes in high-UPF, high-genetic-risk populations would move the evidence beyond association.
Policy uptake: Watch whether regulators cite combined-exposure studies when designing UPF-focused rules or labelling.
Beyond NOVA: Pressure is building for attribute-based definitions of UPFs that manufacturers can actually formulate against.
The research does not suggest UPFs are harmless. It suggests they are one exposure within a larger metabolic system. The next phase of UPF science will need to show which food characteristics drive risk, and by how much relative to other modifiable behaviours. Those answers will matter far more to product development than another confirmation that the association exists.
Related on GFI: When Labels Change Recipes: New Evidence Links Nutrition Warnings to Food Reformulation and The GLP-1 Grocery Effect Is Becoming Measurable. Food Companies Can No Longer Treat It as a Forecast.
Sources
- Xu W, Chen T, Zhang W, Sakal C, Li X. Associations of Ultraprocessed Food Intake, Wearable-Measured Lifestyle Factors, and Genetic Susceptibility With Incident Type 2 Diabetes. Diabetes. Published online September 17, 2026. doi:10.2337/db26-0395. PubMed
- Chen Z, Khandpur N, Desjardins C, et al. Ultra-Processed Food Consumption and Risk of Type 2 Diabetes: Three Large Prospective U.S. Cohort Studies. Diabetes Care. 2023;46(7):1335-1344. Journal page
Leave a Reply