Ultra-Processed Food and 8.8 Million Adults: What the Review Found

On September 22, 2026, researchers published a dose-response meta-analysis of ultra-processed food consumption across 51 prospective cohorts encompassing 8,819,894 adults. The study tracked chronic disease outcomes across global populations, linking higher intake with elevated relative hazards for cardiovascular events, digestive diseases, and all-cause mortality.

Key takeaways:

• A September 22, 2026 meta-analysis of 51 prospective cohorts encompassing over 8.8 million adults linked each additional 100 grams per day of ultra-processed food to a 14% higher relative hazard of cardiovascular events (HR 1.14).
• A companion September 16, 2026 meta-analysis of randomised controlled trials showed excess calorie intake nearly disappears when ultra-processed diets are matched for energy density.
• The cohort findings show relative associations rather than direct causation, with authors grading evidence certainty as low to moderate due to self-reported diet metrics and lifestyle confounders.
• No mandatory label identifies ultra-processed food; shoppers must rely on Nutrition Facts calorie density and ingredient proxies such as emulsifiers, flavourings, and modified starches.

What happened

On September 22, 2026, lead author Liu X, alongside Luo S, Zeng Y, and colleagues, published a systematic review and dose-response meta-analysis in the journal Family Medicine and Community Health (DOI: 10.1136/fmch-2026-003925). The research team synthesized prospective cohort data gathered from literature searched up to April 2026, assembling records from 8,819,894 adult participants across 51 distinct cohorts to assess how chronic health conditions correlate with dietary exposure.

Dietary intake across the aggregated studies was measured using food-frequency questionnaires, 24-hour recalls, or comprehensive diet histories. When comparing the highest intake categories to the lowest intake categories across the cohorts, the researchers observed elevated relative hazards across multiple health endpoints.

The calculated hazard ratios for the highest versus lowest intake categories included:

• Digestive diseases: 1.31 (95% CI 1.16 to 1.48)
• Depression and anxiety: 1.27 (95% CI 1.15 to 1.40)
• Cardiovascular events: 1.24 (95% CI 1.14 to 1.36)
• Metabolic syndrome and diabetes: 1.24 (95% CI 1.08 to 1.42)
• Obesity and overweight: 1.23 (95% CI 1.17 to 1.30)
• All-cause mortality: 1.18 (95% CI 1.10 to 1.26)
• Cancer: 1.12 (95% CI 1.04 to 1.21)

In the dose-response analysis, each additional 100 grams per day of ultra-processed food intake was associated with a hazard ratio of 1.14 for cardiovascular events, 1.04 for cancer, 1.03 for all-cause mortality and 1.02 for metabolic syndrome or diabetes.

What changed?

This publication shifts the conversation from broad categorical associations to quantitative dose-response modeling. However, the authors explicitly grade the certainty of the observational evidence as low or moderate. They noted that ultra-processed food consumption frequently clusters with broader lifestyle patterns that statistical models struggle to fully isolate, and that dietary measurements depended on self-reported questionnaires.

Crucially, a separate meta-analysis published the same week adds a clinical counterpoint. On September 16, 2026, Robinson E and colleagues published a meta-analysis of 10 randomised controlled trials in Critical Reviews in Food Science and Nutrition. It found that people ate faster on ultra-processed diets (standardised mean difference [SMD] 0.96) and gained more weight (SMD 0.65); the diets compared were not matched for nutrient profile.

Yet the intervention trials revealed a key mechanism: the extra energy intake held strong when ultra-processed diets possessed higher energy density (SMD 0.71), but that surplus nearly vanished when diets were experimentally matched for energy density (SMD 0.02). The trial review authors rated the certainty of their evidence as low and wrote that the results may be explained by the energy density of the foods used in trials so far.

Who is affected?

The September 22 meta-analysis pooled adults without restriction to specific health conditions at baseline, so its findings speak to general adult populations rather than to one patient group.

However, it is vital to distinguish relative hazard from absolute risk. A relative hazard ratio of 1.24 for cardiovascular events indicates a comparative difference between intake extremes within observational cohorts, not an individual guarantee of disease. These observational studies do not prove that ultra-processed food directly causes chronic illness, as confounding factors such as physical activity, socioeconomic status, and overall nutrient balance influence real-world outcomes.

From a regulatory standpoint, ultra-processed foods remain fully legal across the United States and global markets, with no federal bans or categorical commercial exclusions in force.

What it means for shoppers?

For everyday shoppers standing in a grocery aisle, no mandatory packaging label currently reads "ultra-processed." Regulators do not require front-of-pack classifications based on processing categories, meaning consumers must evaluate standard packaging elements.

Two practical proxies can assist shoppers in assessing products:

• Calorie density: Review the Nutrition Facts panel to check calories per gram or calories per serving relative to product volume. In the controlled trials above, the extra calories appeared when ultra-processed diets were more energy dense.
• Cosmetic additive markers: Check the ingredients list for ingredients rarely used in home kitchens. These include emulsifiers, artificial or natural flavourings, non-nutritive sweeteners, modified starches, and added colourings.

Shoppers monitoring functional ingredients can explore our guides on Emulsifiers on Ingredient Lists Explained: Lecithin to Polysorbate 80 and Preservatives on Ingredient Lists Explained: Benzoate to BHA to understand how specific formulation additives are reviewed by safety authorities.

The background

Public interest in food classification systems has grown steadily over the past decade, driven in large part by the NOVA classification model developed in Brazil, which categorizes food items into four tiers based on industrial processing extent. Despite wide academic usage, global food regulators have approached the system with caution.

As explained in guidance published by the U.S. Food and Drug Administration, there is currently no uniform federal definition of ultra-processed food. The agency notes that the concept includes diverse food matrices and that scientific consensus on precise category bounds remains developing.

A proposed federal definition from the FDA and USDA is under review at the Office of Management and Budget; our explainer US to Define 'Ultra-Processed Food': What the Proposal Means covers what it would do. The Dietary Guidelines Name 4 Things No Food Label Shows, and California is weighing a state seal in California's Non-Ultraprocessed Seal: Newsom Decides by Sept 30. Meanwhile, international bodies like the European Food Safety Authority continue to assess specific food additive categories individually rather than establishing category-wide bans.

What to watch next

Several upcoming regulatory and research milestones will influence how ultra-processed foods are handled on store shelves and in official policy:

• OMB review and publication: Watch for the formal release of the joint FDA/USDA federal definition following executive review, which will establish standardized criteria for future research and label discussions.
• Dietary Guidelines integration: Review upcoming cycles of the federal nutrition recommendations to see how committee debates unfold on whether processing tiers should be incorporated alongside traditional nutrients.
• State-level policy proposals: Monitor state legislative developments, including emerging initiatives evaluating certification marks and regional retail procurement standards, such as the California seal decision due by September 30.
• Mechanistic feeding trials: Additional controlled clinical research will clarify whether health risks stem from food structure, sensory texture, additive combinations, or basic energy density.

Evidence limitations remain important to state plainly: existing observational data cannot establish individual health responses, cannot resolve unstudied demographic subgroups, and cannot separate every correlated lifestyle behavior from food intake. Evaluating ingredients item by item remains the most reliable path to understanding personal nutrition.