← All articles
Nutrition

Ultra-Processed Foods: What 45 Meta-Analyses Say About Health Risks

The BMJ umbrella review (2024) pooled 45 epidemiological meta-analyses and approximately 10 million people. The associations identified with mortality risk, heart disease, and diabetes are observational, but are consistent across different populations and warrant attention.

7 min readNutrition16.06.2026
Quick answer

The BMJ umbrella review (2024, 45 meta-analyses, ~10 million people) found: high ultra-processed food consumption is associated with +21% all-cause mortality risk and +50% cardiovascular disease mortality risk. The data are observational — causality has not been established — but the strength of the association is consistent across different cohorts.

Ultra-processed foods (UPF) is a term coined by Brazilian epidemiologist Carlos Augusto Monteiro of the University of São Paulo within the NOVA classification system (2009). This system divides all foods into four groups: from natural and minimally processed foods to industrial formulations with five or more ingredients not typical of home cooking — emulsifiers, stabilizers, colorings, flavor enhancers. It is this fourth group — UPF — that has become the subject of large-scale epidemiological research in recent years.

How does the NOVA classification work?

NOVA groups foods by degree of industrial processing rather than nutrient composition. Group 4 includes mass-produced bread products with additives, frozen ready meals, processed meats, sugary drinks, packaged snacks, instant noodles, and similar products. What they share is the presence of industrial ingredients absent from ordinary home cooking: modified starch, protein isolate, hydrogenated fats, artificial colorings, sweeteners. The NOVA classification is recognized by the Food and Agriculture Organization of the United Nations (FAO) as an analytical tool.

What does the BMJ 2024 umbrella review show?

In February 2024, the BMJ published an umbrella review pooling data from 45 epidemiological meta-analyses involving approximately 10 million people. The authors — Lane M.M., Gamage E., Du S., and co-authors — systematically assessed associations between ultra-processed food consumption and health outcomes. Associations identified for the highest UPF consumption group:

  • all-cause mortality: RR = 1.21 (95% CI: 1.15–1.27), i.e. +21%;
  • cardiovascular disease mortality: RR = 1.50 (95% CI: 1.37–1.63), +50%;
  • coronary heart disease mortality: HR = 1.66 (95% CI: 1.51–1.84), +66%;
  • type 2 diabetes (dose-dependent): RR = 1.12 per serving (95% CI: 1.11–1.13);
  • obesity: OR = 1.55 (95% CI: 1.36–1.77);
  • depression: HR = 1.22 (95% CI: 1.16–1.28).

The authors emphasize: the overwhelming majority of the associations identified are observational in nature. This means it is impossible to establish causality: high UPF consumption may correlate with other risk factors — lower socioeconomic status, less physical activity, poorer sleep quality. Adjustments for these variables were made in the original studies, but residual confounding cannot be fully excluded.

45 meta-analyses, approximately 10 million people — and the association replicates: +21% all-cause mortality risk with high UPF consumption.

Dose-dependence: data from the 2025 meta-analysis

A systematic review and meta-analysis from 2025 (Liang S. and co-authors, Systematic Reviews) pooled 18 prospective cohort studies with 1 148 387 participants and 173 107 deaths over the observation period. The authors evaluated both absolute UPF intake levels and dose-dependent associations. Results:

  • comparing highest vs. lowest UPF consumption: HR = 1.15 (95% CI: 1.09–1.22);
  • for each 10 percentage point increase in the share of UPF in the diet: HR = 1.10 (95% CI: 1.04–1.16).

The dose-dependent nature of the association is one of the criteria epidemiologists consider as supporting the biological plausibility of a relationship. It does not prove causality, but substantially strengthens the argument for the reality of the observed pattern.

What is happening at the population level?

A study published in the American Journal of Preventive Medicine (April 2025) estimated the proportion of premature deaths among adults aged 30–60 attributable to UPF consumption in eight countries. In the US and UK, where UPF makes up a significant portion of the diet, this figure was approximately 14%. In Brazil and Colombia, where the share of natural and minimally processed foods is higher, it was approximately 4%.

The gap between countries underscores that this is not about individual habits, but about the food environment — availability, marketing, and pricing policies for different food groups.

Why are biological mechanisms plausible but not proven?

Researchers propose several mechanisms through which UPF may affect health: industrial additives with pro-inflammatory potential, disruption of satiety signals due to high palatability, high glycemic load, displacement of fiber- and nutrient-rich foods from the diet. None of these mechanisms has been confirmed in large interventional trials in humans — this is an area of active research, not established fact.

What this means in practice
  • To identify UPF, read ingredient labels: if they contain emulsifiers, stabilizers, artificial colorings, sweeteners, or modified starch — you are looking at NOVA Group 4.
  • Replacing UPF with minimally processed alternatives automatically improves the nutrient profile of your diet: more fiber, less added sugar, and fewer industrial additives.
  • Aim for a diet predominantly from NOVA Groups 1–2: fresh produce, legumes, whole grains, fish, unprocessed meat, cold-pressed vegetable oils.
  • The data are observational: reducing UPF intake does not guarantee a specific reduction in individual risk — but the association is robust enough to take into account when shaping your diet.
  • Dose matters: the association is dose-dependent. This is not an all-or-nothing principle — a gradual reduction in the share of UPF in the diet is biologically plausible and practically achievable.

Frequently asked questions

Are all processed foods equally harmful?
No. NOVA identifies four groups. Significant associations with increased mortality risk are specific to Group 4 — industrial formulations with five or more ingredients not typical of home cooking. Frozen vegetables with no additives, canned tuna, and hard cheese belong to Groups 1–3 and do not show similar associations in research.
How do you identify an ultra-processed food in practice?
NOVA Group 4 indicators: ingredients not found in ordinary home cooking — emulsifiers (E471, E472), maltodextrin, protein isolate, carrageenan, modified starch, artificial colorings, sweeteners (aspartame, sucralose). If the ingredient list reads like a roster of industrial compounds, you are most likely looking at a UPF.
How large a share of the modern diet do UPFs represent?
According to cohort studies, UPFs account for 50–60% of daily caloric intake in the US and UK. In the American Journal of Preventive Medicine study (2025), approximately 14% of premature deaths among adults aged 30–60 in those countries were attributed to high UPF consumption.
Does reducing UPF consumption automatically improve health?
The data are observational: causality has not been established. Replacing UPFs with minimally processed alternatives automatically improves the nutrient profile of the diet — more fiber, less added sugar, and fewer industrial additives. The probability of benefit is high, but the precise effect size for any individual is unknown.

Sources

  1. Lane MM, Gamage E, Du S et al. «Ultra-processed food exposure and adverse health outcomes: an umbrella review of epidemiological meta-analyses». BMJ, 2024; 384:e077293. doi.org/10.1136/bmj-2023-077293
  2. Liang S, Zhou Y, Zhang Q, Yu S, Wu S. «Ultra-processed food intake and risk of all-cause mortality: a systematic review and dose-response meta-analysis of cohort studies». Systematic Reviews, 2025. link.springer.com/article/10.1186/s13643-025-02800-8
  3. Monteiro CA, Cannon G, Lawrence M, Louzada MLC, Machado PP. «Ultra-processed foods, diet quality, and health using the NOVA classification system». Food and Agriculture Organization of the United Nations (FAO), Rome, 2019.
  4. «Ultra-processed food consumption and premature mortality: a comparative study in 8 countries». American Journal of Preventive Medicine, April 2025. ajpmonline.org/article/S0749-3797(25)00072-8
This material is for educational purposes only and does not constitute medical advice.

Nutrition is part of a system, not a set of rules

Anvil helps you build diet, training, and recovery into a unified whole — without hype or oversimplification.

Open in Telegram