We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
Ultra-Processed Food Literature Synthesis
Summary
This review pulls together the current science on "ultra-processed foods" and finds that researchers don't even agree on what counts as processed, which makes the science messier than headlines suggest. A couple of small but real experiments show that processed foods can cause people to eat more calories, and there are plausible reasons why (they're easier to overeat, may mess with gut signals and gut bacteria, and hijack our brain's reward system), but most of the strongest evidence linking processed food to disease is still just an association, not proof. One hot topic, plastic particles found in human brain tissue, is still too uncertain and unconnected to diet to draw conclus
Research on food processing and health has grown faster than the field’s capacity to synthesize it, and faster than itscapacity to agree on what “processing” means. This narrative review has three aims. First, it maps the operationaltaxonomies now competing to define processing as a dietary exposure — NOVA, Siga, the Hyper-Palatable Food(HPF) criteria, the University of North Carolina scheme, EPIC-Soft/GloboDiet, and the recently proposed CHIPSframework — and shows that they measure partially non-overlapping constructs, so that “ultra-processed” is not asingle exposure variable. Second, it summarizes the mechanistic literature proposed to link processing to metabolicoutcomes: food-matrix disintegration and elevated eating rate, blunted distal-gut enteroendocrine signalling,emulsifier-associated changes in the gut microbiome, and reward-circuit engagement. Third, it sets that mechanisticcase against the principal epistemological objection — that the epidemiological base is almost entirely observationaland confounded — and against the small number of controlled feeding trials now available.The review’s position is deliberately narrower than the literature’s rhetoric. Two randomized crossover feeding trials(Hall et al., 2019; Dicken et al., 2025) support a causal contribution of processing to energy intake independent ofnutrient composition, but both are small, short, and contested in their effect size and policy relevance. Mechanisticpathways are plausible and partially demonstrated, but the strongest supporting studies are in mice. Observationalassociations are large and consistent across dozens of outcomes but graded low to very low certainty by GRADE. Therecent finding of micro- and nanoplastic accumulation in human brain tissue is included as an open question ratherthan a conclusion: it is methodologically contested, no dietary exposure was measured in the source study, and theauthors explicitly disclaim causality. The review concludes that the case for treating processing as an exposuredistinct from nutrient content is strong enough to justify measurement standardization and further controlled trials,and not yet strong enough to support the stronger causal and toxicological claims often made in its name. Keywords: ultra-processed food; NOVA classification; food matrix; hyper-palatable food; enteroendocrine signalling;dietary exposome; evidence synthesis; nutrition policy