Higher consumption of ultra-processed food was associated with 32 adverse health outcomes across nearly ten million participants, spanning cardiovascular mortality, common mental disorders, type 2 diabetes, and all-cause death, according to an umbrella review of 45 meta-analyses by Lane and colleagues published in BMJ in 2024. The framework it uses, the NOVA classification developed by Carlos Monteiro’s group at the University of São Paulo and described in Public Health Nutrition in 2019, defines ultra-processed food not by nutrient content but by the industrial processes and additives used to make it. The distinction matters, because the associations found by Lane and colleagues held even after controlling for individual nutrients.
The NOVA classification is about process, not nutrients
Traditional food guidelines rank foods by their nutrient content: how much saturated fat, added sugar, sodium. NOVA takes a different axis. It classifies foods into four groups based on the extent and purpose of industrial processing. Group 1 is unprocessed or minimally processed foods: fresh, frozen, or dried whole foods with nothing added. Group 2 is processed culinary ingredients such as oils, butter, sugar, and salt, used to season and cook. Group 3 is processed foods, made by adding group 2 ingredients to group 1 foods (canned tomatoes, salted nuts, fresh bread). Group 4 is ultra-processed foods: industrial formulations built largely from substances extracted from foods or synthesized in laboratories, with additives designed for shelf life, appearance, and palatability rather than culinary purpose.
The definition matters because it captures things nutrient-based approaches miss. A frozen fish fillet with a preservative and a soda made from cane sugar, water, and citric acid are both processed by the older definition, but only one is engineered to be consumed in quantities that would be difficult with home-prepared food.
The epidemiological signal is now large
The Lane umbrella review synthesized 45 pooled analyses from 14 review articles, covering nearly 9.9 million participants, and applied graded evidence classification criteria. Direct associations were found for 32 of 45 health outcomes examined. The strongest evidence pointed to a roughly 50% increase in cardiovascular disease-related mortality, a 48 to 53% increase in anxiety and common mental disorders, and a 12% increase in type 2 diabetes with higher UPF intake. Weaker but suggestive evidence covered obesity, sleep problems, depression, and all-cause mortality (21% higher). Evidence for asthma, some cancers, and gastrointestinal disorders remains more limited. These are associations from observational studies, and the authors are explicit that residual confounding cannot be ruled out.
One trial pinned down a mechanism for overconsumption
The clearest experimental evidence comes from a small randomized crossover trial by Kevin Hall and colleagues at the National Institutes of Health, published in Cell Metabolism in 2019. Twenty adults spent two weeks each on an ultra-processed diet and a matched minimally processed diet, presented so they could eat as much as they wanted. The two diets were closely matched for calories, sugar, fat, sodium, and fiber offered. On the ultra-processed diet, participants ate about 500 calories more per day and gained about 0.9 kg over two weeks. On the minimally processed diet, they ate less and lost about the same amount. The finding suggests that something about ultra-processing itself, not just the nutrient profile, drives overconsumption. The trial is small, but it is the only randomized experimental evidence of its kind, and the authors concluded that “limiting consumption of ultra-processed foods may be an effective strategy for obesity prevention and treatment.”
Several mechanisms are plausible, none is definitive
Ultra-processed foods share features that could plausibly drive the observed harms. They are typically low in fiber and dietary fluid, high in energy density, and engineered for what the food science literature calls hyperpalatability, the combination of fat, sugar, salt, and texture that overrides the sensory-specific satiety usually triggered by whole foods. They are often rapidly digested, which alters glycemic response. Some emulsifiers used in processed foods have shown adverse effects on gut barrier function and microbiome composition in animal studies. Substitution effects also matter: a diet high in UPFs is by definition low in whole foods, so some of the harm may come from what is missing rather than what is present.
The definition itself has critics
Not everyone in nutrition science accepts the NOVA framework. Critics point out that the classification depends on judgments that can be inconsistent between raters. A study cited in the American Journal of Clinical Nutrition found that trained evaluators using NOVA disagreed on classification of many common foods. Others argue that the framework lumps together very different products (a slice of packaged whole-grain bread and a soda) in ways that obscure meaningful nutritional differences. Defenders reply that the associations remain strong across studies using different implementations of NOVA, and that no better framework has emerged with equivalent predictive power.
The practical guidance is unchanged and boring
The dietary pattern that emerges from this evidence is not novel. It looks like the Mediterranean pattern, the DASH pattern, and most other evidence-supported approaches: mostly whole and minimally processed foods, with UPFs occasional rather than foundational. The specific brands and cuisines vary; the structure does not. Building meals around vegetables, fruits, whole grains, legumes, unprocessed meats and fish, plain dairy, and eggs delivers the profile that most cohort studies associate with lower risk of the outcomes Lane and colleagues examined.
What the evidence cannot yet answer
The observational evidence cannot prove causation, and the size of any true causal effect is not settled. Larger controlled feeding trials of the kind Hall and colleagues ran are difficult and expensive to conduct at scale. Interrater reliability of the NOVA classification remains a concern for research and for individual application. And whether specific components of ultra-processing (the additives, the physical form, the calorie density, the marketing environment) drive the associations, or whether the effect is the combination of all of them, is not yet resolved.
References
- Monteiro CA, Cannon G, Levy RB, et al. Ultra-processed foods: what they are and how to identify them. Public Health Nutrition, 2019; 22: 936-941. DOI: 10.1017/S1368980018003762
- Lane MM, Gamage E, Du S, et al. Ultra-processed food exposure and adverse health outcomes: umbrella review of epidemiological meta-analyses. BMJ, 2024; 384: e077310. DOI: 10.1136/bmj-2023-077310
- Hall KD, Ayuketah A, Brychta R, et al. Ultra-Processed Diets Cause Excess Calorie Intake and Weight Gain: An Inpatient Randomized Controlled Trial of Ad Libitum Food Intake. Cell Metabolism, 2019; 30: 67-77. DOI: 10.1016/j.cmet.2019.05.008
- Srour B, Fezeu LK, Kesse-Guyot E, et al. Ultra-processed food intake and risk of cardiovascular disease. BMJ, 2019; 365: l1451. DOI: 10.1136/bmj.l1451