The Caloric Takeover
How ultra-processed food conquered the modern diet
Fifty-five per cent of all calories consumed in the United States now come from ultra-processed food [4]. ✓ Established In the United Kingdom, the figure is approximately 57%. In Australia, 42%. These are not fringe products. They are the foundation of the modern Western diet — and their health consequences constitute an epidemic that has been hiding in plain sight.
In August 2025, the US Centers for Disease Control and Prevention published Data Brief No. 536 — the most comprehensive national assessment of ultra-processed food consumption ever conducted in the United States [4]. The findings were stark. American adults derive 53% of their total caloric intake from ultra-processed foods. Among youth aged 2 to 19, the figure rises to 61.9% [4]. ✓ Established Non-Hispanic white adults consume the highest proportion (56.8%), while Hispanic adults consume the lowest (47.1%) — though even the lowest demographic still derives nearly half its calories from industrially formulated products. The overall national average: 55%. More than half the calories fuelling the world's largest economy come from substances that did not exist in any recognisable form before the mid-twentieth century.
The United States is not an outlier. It is the leading edge of a global phenomenon. In the United Kingdom, ultra-processed foods account for approximately 57% of total caloric intake, with adolescents consuming 65.9% [1]. Australia sits at approximately 42% [5]. France — long considered a bastion of traditional food culture — has reached 35.9%, according to the NutriNet-Santé cohort [6]. Spain has seen UPF's share of calories nearly triple, from 11% to 32%, in three decades [3]. Even in countries with historically low processed food consumption, the trajectory is unmistakable: Brazil rose from 12.6% to approximately 20% [5]; China from 4% to 10% in three decades [3]. The global UPF market is expanding at approximately 9% compound annual growth rate [3]. ✓ Established Ultra-processed food is not merely prevalent. It is ascendant.
The health consequences of this caloric takeover are now documented with a breadth of evidence that is difficult to dismiss. In February 2024, Lane and colleagues published in the BMJ the largest umbrella review ever conducted on ultra-processed food and health outcomes [1]. The review synthesised 45 pooled meta-analyses encompassing approximately 9.9 million participants across multiple continents. It identified 32 health parameters adversely associated with higher UPF exposure — spanning cardiovascular disease, metabolic disorders, cancers, mental health conditions, respiratory disease, and all-cause mortality [1]. ✓ Established The evidence for cardiovascular disease mortality was classified as convincing (relative risk 1.50), as was the evidence for anxiety disorders (odds ratio 1.48) and type 2 diabetes (relative risk 1.12). This is not a marginal association. A 50% increase in cardiovascular mortality risk places ultra-processed food in the same epidemiological category as established risk factors such as hypertension and smoking.
The economic dimensions are equally staggering. The World Obesity Federation projects that the global economic impact of overweight and obesity — conditions to which UPF consumption is a major contributor — will reach $4.32 trillion annually by 2035, representing approximately 3% of global GDP [15]. ◈ Strong Evidence In the United States, the adult obesity rate has reached 42.4% [4]. Seventy-three per cent of the US food supply is classified as ultra-processed [4]. The correlation between these two figures is not coincidental. It is structural.
In 2025, the American Journal of Preventive Medicine published the first multi-country estimate of premature mortality attributable to ultra-processed food consumption [5]. The study, covering eight countries, estimated that 124,000 premature deaths in the United States in 2018 were attributable to UPF consumption [5]. ◈ Strong Evidence That figure exceeds the annual death toll from drug overdoses, firearm violence, and motor vehicle accidents combined. Yet ultra-processed food does not generate the same political urgency as any of these causes. The deaths are diffuse, delayed, and — crucially — profitable for the industries that produce the products responsible.
The CDC's 2025 national data brief represents the most comprehensive assessment of UPF consumption in the United States. The finding that more than half of all caloric intake derives from industrially formulated products is consistent with earlier estimates but benefits from the largest nationally representative sample to date. Among children aged 2 to 19, the 61.9% figure means that nearly two-thirds of the calories fuelling developing bodies and brains come from ultra-processed sources [4].
The 2025 Lancet Series on ultra-processed food — a landmark three-paper collection examining epidemiological evidence, mechanistic pathways, and policy implications — concluded unequivocally that UPF constitutes a major global health risk requiring urgent regulatory intervention [3]. The series noted that the growth of UPF consumption has been most rapid in low- and middle-income countries, where traditional food systems are being rapidly displaced by industrialised alternatives marketed as convenient, affordable, and modern. The globalisation of ultra-processed food is not a natural evolution of dietary preferences. It is the product of aggressive marketing, trade liberalisation, and the structural advantages that industrial food manufacturers hold over local food producers.
The term itself — ultra-processed food — requires definition, and the definitional question is not merely academic. Ultra-processed foods, as classified by the NOVA system developed by Carlos Monteiro at the University of São Paulo, are industrial formulations made mostly or entirely from substances derived from foods and additives, with little or no intact food [9]. They include soft drinks, packaged snacks, reconstituted meat products, instant noodles, pre-prepared frozen meals, and industrially produced breads and confectionery. What distinguishes them from merely processed foods — canned vegetables, cheese, cured meats — is the degree of industrial transformation: UPFs typically contain ingredients not found in domestic kitchens (high-fructose corn syrup, hydrogenated oils, emulsifiers, texturants, flavour enhancers) and are designed to be hyper-palatable, convenient, and shelf-stable. They are, in essence, the products of food engineering rather than food preparation.
The Engineering of Craving
Bliss points, hyperpalatability, and the science of food addiction
Ultra-processed foods are not merely convenient alternatives to home cooking. They are precision-engineered products designed to maximise consumption — exploiting the same neurological reward pathways that make tobacco, alcohol, and gambling addictive [3]. ◈ Strong Evidence The engineering of craving is not a side effect of food manufacturing. It is the business model.
The concept of the "bliss point" — the precise combination of sugar, fat, and salt that maximises sensory pleasure — has been central to processed food development since at least the 1970s. Howard Moskowitz, the psychophysicist who pioneered the concept for the food industry, demonstrated that consumer preference for a product follows an inverted U-curve: too little sugar and the product is bland; too much and it becomes cloying. The bliss point is the apex of that curve — the formulation at which the consumer experiences maximum pleasure and, critically, maximum desire to continue eating [3]. Every major food manufacturer employs sensory scientists whose explicit task is to identify this point for every product in the portfolio. The goal is not nutrition. It is consumption maximisation.
The NOVA classification system, developed by Carlos Monteiro and colleagues at the University of São Paulo and published definitively in Public Health Nutrition in 2019, provides the taxonomic framework for understanding what distinguishes ultra-processed food from other categories [9]. ✓ Established NOVA divides all foods into four groups: Group 1 (unprocessed or minimally processed foods — fruits, vegetables, meat, eggs, milk), Group 2 (processed culinary ingredients — oils, butter, sugar, salt), Group 3 (processed foods — canned vegetables, cheese, freshly made bread), and Group 4 (ultra-processed food products — soft drinks, packaged snacks, instant noodles, reconstituted meat products, industrially produced confectionery) [9]. The critical distinction is not the degree of processing per se but the nature of the industrial transformation: Group 4 products typically contain ingredients not used in domestic cooking and are designed for durability, palatability, and profitability rather than nutritional adequacy.
The neurological basis for compulsive UPF consumption is now being mapped with increasing precision. The 2025 Lancet Series dedicated an entire paper to mechanistic pathways, concluding that ultra-processed foods activate the brain's reward circuitry — specifically the mesolimbic dopamine pathway — in a manner qualitatively similar to addictive substances [3]. ◈ Strong Evidence The combination of rapid caloric delivery, high sugar-fat ratios, and engineered flavour profiles produces a dopaminergic response that exceeds what the brain receives from unprocessed foods of equivalent caloric density. The brain, in effect, responds to a packet of crisps the way it responds to a weak drug — with a burst of reward signalling followed by a refractory period that creates desire for repeated exposure.
The Yale Food Addiction Scale (YFAS), the most widely validated instrument for measuring addictive-like eating behaviour, estimates that approximately 20% of the general population meets criteria for food addiction [3]. ◈ Strong Evidence Among individuals with obesity, the prevalence is substantially higher. The criteria mirror those used for substance use disorders in the DSM-5: continued use despite negative consequences, tolerance (requiring increasing quantities for the same effect), withdrawal symptoms, and loss of control over consumption. The Lancet 2025 series was explicit: "There is now converging evidence from human neuroimaging, behavioural, and clinical studies that ultra-processed foods can trigger addictive-like responses in susceptible individuals" [3].
Ultra-processed foods are engineered to combine sugar, fat, and salt in precise ratios that override the brain's natural satiety signals. The "bliss point" — the optimal formulation for maximum consumption — is not an accident of mass production. It is the product of decades of psychophysical research and sensory optimisation, conducted by food scientists whose explicit objective is to make products as difficult to stop eating as possible. When 73% of the US food supply is classified as ultra-processed [4], the consumer is not choosing between real food and processed food. The consumer is navigating an environment in which the processed option is engineered to be preferred.
The industrial additives characteristic of ultra-processed food may also contribute to harm through mechanisms independent of caloric content. Emulsifiers such as polysorbate 80 and carboxymethylcellulose have been shown in animal models to disrupt the intestinal mucus barrier, alter gut microbiome composition, and promote chronic low-grade inflammation [3]. Artificial sweeteners — marketed as healthier alternatives to sugar — appear to alter glucose metabolism and gut microbiota in ways that paradoxically increase metabolic risk [3]. Advanced glycation end products (AGEs), formed during ultra-high-temperature processing, contribute to oxidative stress and vascular damage. The harm from ultra-processed food is not merely about what it contains — too much sugar, too much fat, too much salt. It is also about what it does: the industrial processing itself creates or introduces compounds that interact with human physiology in ways that whole foods, even nutritionally equivalent ones, do not.
The speed of caloric delivery is a critical and often overlooked factor. Ultra-processed foods are typically engineered for rapid oral disintegration — they dissolve quickly in the mouth, delivering their caloric payload before satiety signals can register. This "vanishing caloric density" means the brain does not process the calories consumed as efficiently as it would from whole foods that require chewing and slower gastric processing [2]. The Hall RCT demonstrated this mechanism directly: participants consuming ultra-processed meals ate faster — measured in grams per minute — than those consuming unprocessed meals of identical nutrient composition [2]. The food is designed to circumvent the body's own regulatory systems. Eating more is not a failure of willpower. It is the engineered outcome.
The Yale Food Addiction Scale uses DSM-5-aligned criteria — continued use despite harm, tolerance, withdrawal, loss of control — to assess addictive-like eating behaviour. The 20% prevalence in the general population suggests that ultra-processed food produces clinically significant compulsive consumption patterns in a substantial minority of consumers. Among individuals with obesity, prevalence is substantially higher. The Lancet 2025 series concluded that converging evidence from neuroimaging, behavioural, and clinical studies supports the biological plausibility of UPF addiction [3].
The scale of industrial food engineering investment dwarfs public health research. The ten largest food and beverage companies — Nestlé, PepsiCo, Unilever, Coca-Cola, Mars, Mondelēz, Danone, General Mills, Kellogg's, and Associated British Foods — collectively spend billions annually on product formulation research, sensory optimisation, and marketing. Their objective is explicitly to maximise what the industry terms "cravability" — a word that concedes the point. Products are not designed to satisfy hunger. They are designed to create craving. The consumer is not a customer to be fed. The consumer is a neurological system to be exploited.
The Causal Evidence
The Hall RCT, mechanistic pathways, and what we actually know
The single most important piece of evidence in the ultra-processed food literature is a randomised controlled trial conducted at the National Institutes of Health by Kevin Hall and colleagues, published in Cell Metabolism in 2019 [2]. It demonstrated, under controlled conditions, that ultra-processed food causes people to eat more — approximately 500 additional calories per day. ✓ Established
The study design was elegant in its simplicity. Twenty adults were admitted to the NIH Clinical Center for four weeks. For two weeks, they received an ultra-processed diet; for two weeks, an unprocessed diet. The diets were matched for total calories offered, macronutrient composition, sugar, sodium, fat, and fibre. Meals were presented ad libitum — participants could eat as much or as little as they chose. The only variable was the degree of processing [2]. ✓ Established The result was unambiguous. On the ultra-processed diet, participants spontaneously consumed approximately 500 additional calories per day compared with the unprocessed diet. They gained an average of 0.9 kilograms in just two weeks. On the unprocessed diet, they lost a comparable amount [2].
The significance of this finding cannot be overstated. It is the only randomised controlled trial — the gold standard of causal evidence — that directly measures the impact of ultra-processing on caloric intake while controlling for nutrient composition. The fact that participants ate more despite matched macronutrients and fibre demonstrates that something about ultra-processing itself — beyond its nutrient profile — drives overconsumption. The mechanisms are multiple and mutually reinforcing: faster eating rate (ultra-processed meals were consumed more quickly), reduced satiety signalling (the rapid oral disintegration of UPF undermines gastric stretch receptors), altered gut-brain communication, and the hyper-palatable flavour profiles that override satiation cues [2].
The gut microbiome represents a critical mechanistic pathway now under intensive investigation. The 2025 Lancet Series identified three distinct routes through which ultra-processed food may harm health via the gut: displacement of fibre-rich whole foods that sustain beneficial microbial populations; direct toxic effects of food additives on the intestinal epithelium; and chronic low-grade inflammation triggered by bacterial translocation across a compromised gut barrier [3]. ◈ Strong Evidence Emulsifiers — present in the vast majority of ultra-processed products — have been shown in mouse models to thin the intestinal mucus layer, allowing bacteria to come into closer contact with epithelial cells and triggering inflammatory cascades that contribute to insulin resistance, adiposity, and colorectal cancer risk [3].
Critics of the UPF hypothesis argue that observed health effects can be explained entirely by the poor nutrient profile of ultra-processed foods — too much sugar, salt, and saturated fat; too little fibre and micronutrients. The Hall RCT directly challenges this argument: even when diets were matched for all major nutrients, the ultra-processed version still drove overconsumption. This suggests that the degree of industrial processing itself — through mechanisms including eating rate, additive effects, and altered gut signalling — constitutes an independent risk factor beyond nutrient composition alone [2]. ⚖ Contested
The inflammation pathway is particularly consequential because it connects ultra-processed food consumption to an extraordinarily broad range of diseases. Chronic low-grade systemic inflammation — characterised by elevated C-reactive protein, interleukin-6, and tumour necrosis factor alpha — is implicated in cardiovascular disease, type 2 diabetes, neurodegenerative conditions, depression, and multiple cancers [1]. If ultra-processed food systematically promotes this inflammatory state — through gut barrier disruption, advanced glycation end products, or the metabolic consequences of chronic overconsumption — it would provide a unifying mechanism for the remarkably diverse array of diseases associated with UPF in the epidemiological literature. The BMJ umbrella review's finding of 32 adverse outcomes associated with UPF becomes more coherent when viewed through this inflammatory lens [1].
Hormonal disruption represents another mechanistic concern. Ultra-processed foods are the primary dietary source of endocrine-disrupting chemicals — including bisphenol A (BPA) from packaging, phthalates from processing equipment, and per- and polyfluoroalkyl substances (PFAS) from food contact materials [3]. These compounds interfere with thyroid function, reproductive hormones, and insulin signalling at concentrations routinely found in human blood. The exposure is cumulative: a person consuming 55% of their calories from ultra-processed sources is exposed to a continuous low-dose endocrine-disrupting cocktail that a person eating predominantly whole foods largely avoids. The regulatory framework treats each chemical individually. The human body encounters them in combination, every day, over decades.
When an ad libitum ultra-processed diet was compared with a nutrient-matched unprocessed diet, participants consumed approximately 500 kcal/day more on the ultra-processed diet, leading to significant weight gain within two weeks.
— Hall et al., Cell Metabolism 2019 — the only randomised controlled trial on ultra-processed food intakeThe metabolic consequences extend beyond caloric overconsumption. Ultra-processed diets have been associated with elevated fasting glucose, increased insulin resistance, and adverse lipid profiles in observational studies and, to a limited extent, in the Hall RCT [2]. The Lancet 2025 series identified evidence that UPF consumption alters hepatic lipogenesis — the liver's production of fat from carbohydrates — in ways that promote non-alcoholic fatty liver disease [3]. The rapidity of caloric delivery from ultra-processed foods may overwhelm hepatic processing capacity, leading to ectopic fat deposition in the liver, pancreas, and visceral compartments. This is not the slow accumulation of excess energy over years. It is a metabolic insult delivered at every meal.
The causal picture, taken in its entirety, is not one of a single mechanism but of a web of mutually reinforcing pathways: overconsumption driven by hyper-palatability and fast eating rate; gut microbiome disruption from additives and fibre displacement; chronic inflammation from gut barrier compromise and advanced glycation end products; hormonal disruption from packaging-derived chemicals; and metabolic derangement from the speed and volume of caloric delivery. Each pathway is supported by evidence of varying strength — from the robust RCT evidence for overconsumption to the emerging animal-model evidence for emulsifier toxicity. Together, they form a mechanistic case that is, at minimum, sufficient to warrant the precautionary action now being taken by the WHO [11].
The Body Count
Cancer, cardiovascular disease, diabetes, dementia, and all-cause mortality
The epidemiological evidence linking ultra-processed food to premature death and chronic disease is now vast, consistent across populations, and — for several outcomes — classified at the highest tiers of evidence certainty [1]. ✓ Established This section presents the disease-specific evidence with the hazard ratios, relative risks, and confidence intervals that the data actually contain.
Cardiovascular disease represents the strongest and most consequential association. The BMJ umbrella review classified the evidence for cardiovascular disease mortality as "convincing" — the highest evidence grade — with a relative risk of 1.50 (95% CI 1.37–1.63) comparing highest with lowest UPF consumers [1]. ✓ Established A February 2026 study reported an even larger association: 47% higher CVD risk in the highest UPF consumption group [13]. The Lancet Regional Health Americas 2024 analysis of three US cohorts — the Nurses' Health Study, NHS II, and the Health Professionals Follow-up Study, encompassing approximately 207,000 participants — confirmed the association and identified specific UPF subcategories driving the risk: processed meats, sugar-sweetened beverages, and industrially produced baked goods showed the strongest individual associations with CVD events [8].
Cancer was the first chronic disease to be associated with ultra-processed food in a major prospective study. The NutriNet-Santé cohort, published in the BMJ in 2018, followed 104,980 French adults and found that each 10% increase in the proportion of UPF in the diet was associated with a 12% increase in overall cancer risk (hazard ratio 1.12, 95% CI 1.06–1.18) [6]. ◈ Strong Evidence Breast cancer showed a particularly strong association (HR 1.11 per 10% increase). The UK Biobank study, published in eClinicalMedicine in 2023, extended these findings to 197,426 participants followed for a median of 9.8 years, confirming associations between UPF intake and cancer incidence and mortality — with ovarian cancer and brain cancers showing the strongest relationships [7].
Type 2 diabetes is associated with UPF consumption at an evidence level the BMJ umbrella review classified as "highly suggestive." Each 10% increase in UPF consumption is associated with a 17% higher incidence of type 2 diabetes [1]. ◈ Strong Evidence The mechanisms are well-characterised: rapid glucose delivery from refined carbohydrates, chronic overconsumption leading to adiposity and insulin resistance, and potential direct effects of food additives on pancreatic beta-cell function. The 42.4% US adult obesity rate — itself driven substantially by UPF overconsumption — creates a population-level metabolic vulnerability that makes the diabetes association particularly consequential in public health terms.
Neurodegenerative disease represents a more recent but rapidly strengthening area of evidence. The Framingham Heart Study — one of the oldest and most respected prospective cohort studies in epidemiology — published findings in January 2025 showing that each additional daily serving of ultra-processed food was associated with a 13% increase in Alzheimer's disease risk [10]. ◈ Strong Evidence Participants consuming more than 10 servings per day had nearly triple the dementia risk compared with the lowest consumers [10]. The proposed mechanisms include neuroinflammation driven by systemic inflammation, cerebrovascular damage from chronic hypertension and dyslipidaemia, and the direct neurotoxic effects of advanced glycation end products that cross the blood-brain barrier.
| Health Outcome | Risk Measure | Evidence Assessment |
|---|---|---|
| CVD mortality | Convincing evidence (BMJ umbrella review). Highest evidence tier. 50% increased risk in highest UPF consumers. [1] | |
| Anxiety disorders | Convincing evidence (BMJ umbrella review). 48% higher odds of anxiety in highest consumers. [1] | |
| All-cause mortality | Meta-analysis of 18 studies, 1.1M participants. 21% increased risk of death from all causes. [14] | |
| Type 2 diabetes | Highly suggestive evidence. Each 10% increase in UPF proportion → 17% higher diabetes incidence. [1] | |
| Overall cancer | Strong evidence from NutriNet-Santé (104,980) and UK Biobank (197,426). 12% increased risk per 10% UPF increase. [6] | |
| Alzheimer's disease | Framingham Heart Study. Each daily UPF serving → 13% higher Alzheimer's risk. >10 servings: nearly 3× risk. [10] | |
| CVD events | Prospective study (Feb 2026). Highest UPF consumers: 47% higher CVD event risk. [13] |
All-cause mortality provides the most comprehensive measure of aggregate harm. A 2025 meta-analysis pooling 18 prospective cohort studies with approximately 1.1 million participants found that the highest UPF consumers had a 21% increased risk of death from all causes compared with the lowest consumers (relative risk 1.21, 95% CI 1.13–1.30) [14]. ✓ Established The AJPM 2025 multi-country analysis translated this epidemiological evidence into population-level impact: an estimated 124,000 premature deaths in the United States, along with significant mortality burdens in the UK, Brazil, Australia, and four additional countries [5].
The BMJ umbrella review also identified mental health associations that have received less public attention than the cardiovascular and cancer data. Anxiety disorders showed a convincing association (OR 1.48), while depression showed a highly suggestive association [1]. Common mental disorders, depressive symptoms, and sleep disturbances were all adversely associated with higher UPF intake. The gut-brain axis — the bidirectional communication pathway between intestinal microbiota and the central nervous system — provides a biologically plausible mechanism: UPF-driven dysbiosis may alter serotonin production (approximately 90% of the body's serotonin is produced in the gut), increase neuroinflammatory signalling, and disrupt the hypothalamic-pituitary-adrenal axis that governs stress responses [3].
The AJPM 2025 study estimated population-attributable mortality across eight countries, using the dose-response relationships established in the meta-analytic literature. The 124,000 figure for the United States exceeds annual deaths from drug overdoses (~107,000 in 2023), firearm violence (~48,000), and motor vehicle accidents (~43,000). Unlike those causes, UPF-attributable mortality receives minimal political attention — in part because the deaths are diffuse, delayed, and commercially advantageous to a powerful industry [5].
The dose-response relationships are particularly informative. They are not threshold effects — there is no safe level of UPF consumption below which risk disappears. Instead, the data consistently show graded increases in risk across the full range of exposure. Each 10% increase in UPF as a proportion of calories is associated with approximately 12% higher cancer risk, 17% higher diabetes incidence, and measurable increases in cardiovascular and all-cause mortality [1] [6]. In a country where the average person derives 55% of their calories from UPF, these incremental risks compound into population-level catastrophe.
The Youngest Victims
Children, toddlers, and the developmental consequences of an ultra-processed diet
American children aged 2 to 19 derive 61.9% of their calories from ultra-processed food [4]. ✓ Established Among toddlers aged 12 to 23 months, an estimated 47% of calories come from UPF [4]. The developing bodies and brains of the youngest members of society are being fed a diet that the growing scientific evidence suggests is systematically harmful.
The CDC's Data Brief No. 536 made one finding that should stop any reader with children: nearly two-thirds of the calories consumed by American youth come from ultra-processed food [4]. The figure is not an artefact of teenage junk food consumption. It extends to the very youngest children — toddlers aged 12 to 23 months already receive approximately 47% of their calories from UPF, and the proportion rises steadily through childhood [4]. By adolescence, UPF dominance is near-total: UK data show that adolescents derive 65.9% of their calories from ultra-processed sources [1].
The implications for childhood obesity are direct and measurable. The US childhood obesity rate has tripled since the 1970s, paralleling the rise of ultra-processed food in the national diet. The Hall RCT demonstrated that UPF drives overconsumption of approximately 500 calories per day in adults [2]. While no equivalent RCT has been conducted in children — for obvious ethical reasons — the mechanistic evidence suggests that children may be even more susceptible to UPF-driven overconsumption. Children's satiety signalling is less mature, their taste preferences are more malleable, and their exposure to food marketing is more difficult to mediate. ◈ Strong Evidence The food industry understands this vulnerability and exploits it systematically: children's UPF products are typically the most aggressively marketed, the most hyper-palatable, and the most heavily fortified with the sugar-fat-salt combinations that maximise consumption.
The neurodevelopmental implications of a UPF-dominant diet during critical growth periods remain underresearched but are biologically concerning. The brain undergoes its most intensive development during the first five years of life and continues significant structural maturation through adolescence. This development requires specific micronutrients — omega-3 fatty acids, iron, zinc, B vitamins, choline — that are poorly represented in ultra-processed diets [3]. When 47% of a toddler's calories come from products engineered for palatability rather than nutrition, the displacement of nutrient-dense whole foods may have consequences for cognitive development that take years or decades to manifest. The absence of evidence is not evidence of absence. It is evidence of neglect in research prioritisation.
By the time an American child reaches their first birthday, nearly half their calories come from ultra-processed food. By age 2 to 19, that figure reaches 61.9%. In the United Kingdom, adolescents derive 65.9% of calories from UPF. These children are not making dietary choices. They are consuming what is available, affordable, and marketed to them — in an environment where 73% of the food supply is ultra-processed [4]. The framing of childhood nutrition as a parental responsibility problem collapses when the food environment itself is the problem.
The marketing environment compounds the exposure. Food and beverage companies spend billions annually marketing ultra-processed products to children, using animated characters, social media influencers, and gamified digital platforms to build brand loyalty from the earliest possible age. The WHO has repeatedly called for restrictions on food marketing to children, but compliance remains voluntary in most jurisdictions [11]. The asymmetry is structural: a child's developing brain is exposed to sophisticated marketing designed to create preference for products that the growing evidence base links to obesity, metabolic disease, and potentially impaired cognitive development. This is not a fair contest. It is not designed to be.
The socioeconomic gradient is particularly stark for children. In lower-income communities, UPF consumption is higher because ultra-processed products are cheaper per calorie than fresh alternatives, more widely available in food deserts, and more heavily marketed in neighbourhoods with fewer retail options [5]. The children most dependent on ultra-processed food are those whose families can least afford the health consequences. Childhood obesity rates are highest in the same communities where UPF consumption is highest and healthcare access is poorest. The epidemic is not merely a health crisis. It is a health equity crisis, with children as its most vulnerable victims.
The Framingham dementia findings carry particularly alarming implications when extrapolated to younger populations. If each additional daily UPF serving increases Alzheimer's risk by 13% [10], the cumulative lifetime exposure of a child who begins consuming UPF at age one and continues through adulthood is substantially greater than that of current elderly cohorts whose UPF consumption began later in life. The children currently eating 61.9% UPF diets are the first generation for whom ultra-processed food has been the dominant caloric source from infancy. Their lifetime neurological outcomes will not be known for decades. By the time the evidence arrives, the exposure will be irreversible.
The Tobacco Playbook
How the food industry manufactured doubt, funded distractions, and delayed regulation
In 2015, the New York Times revealed that Coca-Cola had paid $1.5 million to establish the Global Energy Balance Network (GEBN) — an academic organisation whose purpose was to shift public health messaging away from diet and toward exercise as the primary solution to obesity [12]. ✓ Established The parallels with the tobacco industry's playbook are not metaphorical. They are structural, deliberate, and documented.
The GEBN scandal remains the most fully documented example of food industry influence on public health science. Emails obtained through Freedom of Information requests and published in a 2023 Cambridge analysis revealed the extent of Coca-Cola's involvement [12]. The company did not merely fund the organisation. It helped design its research agenda, reviewed draft publications, and coordinated messaging strategy. Steven Blair, the exercise scientist who served as GEBN's public face, received $3.5 million in personal research funding from Coca-Cola [12]. ✓ Established The core message GEBN promoted — that physical inactivity, not diet, was the primary driver of obesity — was scientifically misleading. While exercise contributes to health, the evidence overwhelmingly shows that dietary intake is the dominant determinant of body weight. Coca-Cola knew this. The messaging was not science. It was strategy.
The tobacco playbook metaphor has specific, identifiable components, each of which has been documented in the food industry context. The first is doubt manufacturing: funding research designed to generate ambiguity about the health effects of the company's products. Coca-Cola's GEBN is the clearest example, but industry-funded nutrition research systematically produces results more favourable to sponsors than independently funded research — a phenomenon documented in multiple systematic reviews [12]. The second is the responsibility pivot: shifting blame from the product to the consumer. "Eat less, move more" — the dominant public health message of the past forty years — frames obesity as a personal choice rather than a product of an engineered food environment. The third is regulatory capture: using lobbying, campaign contributions, and revolving-door employment to ensure that regulatory agencies remain sympathetic to industry interests.
Released emails revealed that Coca-Cola did not merely fund the Global Energy Balance Network — it helped design the research agenda, reviewed draft publications, and coordinated public messaging. The company paid $1.5 million to establish GEBN and $3.5 million to researcher Steven Blair personally. GEBN's core message — that physical inactivity, not diet, drives obesity — contradicted the weight of scientific evidence. When the emails were made public, GEBN was disbanded within months. But the messaging it promoted had already shaped a generation of public health policy [12].
The lobbying expenditure of the food and beverage industry is staggering. In the United States alone, the food and beverage sector consistently ranks among the top five lobbying spenders, deploying hundreds of millions of dollars annually to influence legislation affecting labelling, marketing restrictions, dietary guidelines, and taxation of sugary products [3]. The 2025 Lancet Series documented industry interference in the development of national dietary guidelines across multiple countries, including successful efforts to remove or weaken references to ultra-processed food in official guidance [3]. The revolving door between industry and regulatory agencies ensures a continuous flow of personnel who bring industry perspectives into government positions and government contacts into industry roles.
The NOVA classification system itself has become a target of industry opposition. Food manufacturers have funded critiques of NOVA, arguing that the classification is too broad, internally inconsistent, and unsuitable for regulatory application [9]. ⚖ Contested Some of these critiques have legitimate methodological merit — the heterogeneity within Group 4 is a genuine limitation. But the strategic intent is clear: by challenging the classification system, the industry aims to prevent the regulatory frameworks that would follow from its adoption. If there is no agreed definition of ultra-processed food, there can be no regulation of ultra-processed food. The tactic is identical to the tobacco industry's decades-long challenge to the definition of addiction.
Front groups and third-party advocacy represent another direct parallel. The International Life Sciences Institute (ILSI), funded by major food and beverage companies including Coca-Cola, PepsiCo, and Nestlé, has been documented operating as an industry influence vehicle within the WHO and national health agencies [3]. ILSI-affiliated researchers have served on WHO advisory panels, national dietary guideline committees, and food safety boards — bringing industry perspectives into ostensibly independent scientific processes. The WHO severed ties with ILSI in 2021, but the organisation's decades of influence on global nutrition policy are not easily reversed.
The food industry's strategy is not to prove its products are safe. The strategy is to ensure the question is never definitively answered — because as long as the science is 'uncertain,' regulation can be delayed.
— Analysis based on Cambridge 2023 study of Coca-Cola's funding strategy and GEBN emailsReleased emails revealed that Coca-Cola did not passively fund academic research — it actively shaped the GEBN's messaging, reviewed draft publications, and coordinated strategy. The organisation promoted the message that physical inactivity, not diet, was the primary driver of obesity — directly contradicting the weight of scientific evidence. When the emails were made public, GEBN was disbanded within months. The incident represents the most fully documented example of food industry manipulation of public health science [12].
The "personal responsibility" narrative deserves particular scrutiny because of its rhetorical power and empirical weakness. The food industry's consistent messaging — that consumers should make better choices, eat in moderation, exercise more — presupposes a level playing field that does not exist. When 73% of the US food supply is ultra-processed [4], when products are engineered for maximum "cravability," when the Hall RCT demonstrates that UPF drives an additional 500 calories of consumption per day even when nutrients are matched [2], and when approximately 20% of the population meets criteria for food addiction [3], the "personal responsibility" framework is not an honest assessment of the situation. It is a deflection strategy designed to prevent regulatory action that would reduce industry revenue.
The Lancet 2025 series was explicit about this parallel: "The commercial determinants of the ultra-processed food epidemic mirror those documented in tobacco, alcohol, and fossil fuel industries — doubt manufacturing, regulatory capture, responsibility shifting, and the strategic exploitation of scientific uncertainty to delay policy action" [3]. The question is not whether the food industry has adopted the tobacco playbook. The evidence that it has is beyond dispute. The question is how long the playbook will continue to work.
The Regulatory Response
Chile, Mexico, the WHO, and the emerging global backlash
Regulatory action against ultra-processed food is accelerating across multiple jurisdictions — from Latin America's pioneering front-of-pack labelling to the WHO's first-ever guideline development group on UPF [11]. But the regulatory response remains structurally outmatched by a global UPF market growing at 9% annually. ✓ Established
Chile enacted the most comprehensive anti-UPF legislation to date. Its 2016 Law of Food Labelling and Advertising requires black octagonal warning labels on products high in sugar, sodium, saturated fat, or calories. It bans the use of cartoon characters and toys in marketing products that carry warning labels. It restricts advertising of labelled products to children. And it prohibits the sale of labelled products in schools [3]. The results have been significant: a 24% reduction in purchases of sugar-sweetened beverages and measurable reformulation by manufacturers seeking to avoid the warning labels [3]. ✓ Established The Chilean model has become the template for regulatory action across Latin America and beyond.
Mexico followed Chile with its own front-of-pack warning label system in 2020, applying to a food supply in which ultra-processed products have penetrated deeply. Mexico's labelling system uses black octagonal stamps similar to Chile's and has been accompanied by a tax on sugar-sweetened beverages. Colombia, Peru, Uruguay, and Argentina have adopted similar approaches. Brazil — the country where the NOVA classification system was invented — revised its national dietary guidelines in 2014 to explicitly recommend avoiding ultra-processed foods, the first nation to incorporate processing degree into official dietary advice [3]. Latin America, in short, is leading the global regulatory response — driven by both the scientific evidence and the direct experience of populations undergoing rapid nutritional transitions.
The World Health Organisation's decision in 2025 to convene a Guideline Development Group specifically on ultra-processed foods represents a potentially transformative institutional step [11]. ✓ Established The GDG is tasked with developing the first-ever global guidance on UPF — addressing classification, labelling, marketing restrictions, and recommended consumption limits. If the WHO issues formal guidance recommending reduction of UPF intake, it will provide the evidence-based foundation that national regulators need to justify legislative action. It will also provoke intense industry opposition. The food industry has already begun lobbying against NOVA-based classification at the WHO level, arguing that nutrient profiling is a more appropriate basis for regulation [3].
The United Kingdom is pursuing a regulatory path that combines front-of-pack labelling reform with restrictions on the advertising and promotion of products high in fat, sugar, and salt (HFSS). The UK's planned 9pm watershed on HFSS television advertising and a ban on paid-for online HFSS advertising — repeatedly delayed under industry lobbying — represent a nutrient-profiling rather than NOVA-based approach [3]. ⚖ Contested The debate over whether regulation should target ultra-processing (NOVA) or nutrient content (nutrient profiling) is the central methodological disagreement in UPF policy. Industry strongly favours nutrient profiling because it allows reformulation — removing some sugar or salt while maintaining the ultra-processed matrix. Public health advocates argue that the processing itself, independent of nutrient content, constitutes a risk factor that nutrient profiling fails to capture.
France has adopted a mixed approach, combining the Nutri-Score front-of-pack label — a colour-coded system based on nutrient profiling — with active public health messaging about ultra-processed food. The NutriNet-Santé cohort, which produced the landmark cancer study [6], continues to generate evidence that feeds directly into French and European policy discussions. However, France's UPF consumption has risen to 35.9% despite these measures — suggesting that labelling and education alone are insufficient without structural changes to the food environment.
NOVA-Based Regulation
Targets the industrial transformation that creates hyper-palatability, additive exposure, and rapid caloric delivery — not just the nutrient endpoint.
The Hall RCT, BMJ umbrella review, and Lancet Series all frame their findings in NOVA terms. The epidemiological evidence base is built on NOVA classification.
Industry cannot simply reduce sugar in a product and avoid classification. The processing method, not the nutrient profile, determines the category.
Chile, Mexico, Brazil, and multiple South American countries have incorporated processing degree into regulation with measurable results.
The WHO's decision to convene a UPF-specific GDG signals institutional momentum toward NOVA-based global guidance.
Nutrient-Profiling Regulation
Targets specific nutrients (sugar, sodium, saturated fat) that have well-characterised dose-response relationships with disease. Clear regulatory thresholds.
Group 4 encompasses products with vastly different health profiles — from wholegrain breakfast cereals to energy drinks. Treating them identically may misallocate regulatory effort.
Manufacturers can reduce harmful nutrient content to avoid thresholds, improving the nutrient profile of widely consumed products even if processing continues.
Nutrient profiling is already embedded in food labelling systems worldwide. NOVA-based regulation would require new classification infrastructure.
Nutrient targets are measurable and enforceable. "Degree of processing" is more difficult to define in regulatory terms and may face legal challenges.
The structural challenge for all regulatory approaches is the scale and growth rate of the global UPF market. The industry is expanding at approximately 9% CAGR [3], driven by urbanisation, rising incomes in low- and middle-income countries, and the displacement of traditional food systems by industrialised alternatives. Regulatory action in high-income countries — even if successful — may simply redirect industry growth toward less regulated markets. The Lancet 2025 series documented aggressive UPF market expansion in sub-Saharan Africa, South-East Asia, and South Asia, where regulatory capacity is weakest and nutritional transitions most rapid. A global problem requires a global regulatory response. The WHO Guideline Development Group represents the first step toward such a response. Whether it produces binding recommendations or diluted guidance will depend on whether member states can resist the lobbying pressure that has already begun.
What the Evidence Actually Tells Us
Synthesis — what is established, what is contested, and what would work
The ultra-processed food literature is now vast enough to distinguish with some precision between what is established beyond reasonable dispute, what remains contested, and what the evidence suggests would be effective if implemented. This section provides that synthesis. ◈ Strong Evidence
What is established. Ultra-processed food constitutes the majority of caloric intake in the US (55%), UK (~57%), and Australia (~42%), and its global market share is growing [4]. The BMJ umbrella review of 45 meta-analyses, covering approximately 9.9 million participants, found 32 adverse health outcomes associated with higher UPF consumption, with convincing evidence for CVD mortality (RR 1.50) and anxiety disorders (OR 1.48) [1]. The Hall RCT demonstrated that UPF causes overconsumption of approximately 500 calories per day compared with nutrient-matched unprocessed food [2]. An estimated 124,000 premature deaths in the US are attributable to UPF [5]. The food industry has employed documented tobacco-industry tactics to delay regulatory action [12]. These facts are not in serious dispute.
What is contested. The degree to which ultra-processing itself — beyond the nutrient profile of the products — constitutes an independent risk factor remains debated [9]. ⚖ Contested The Hall RCT provides the strongest evidence that it does, but the trial was small (20 participants) and short (2 weeks per condition). Larger, longer RCTs are needed but face ethical and practical constraints. The NOVA classification system, while widely adopted, is criticised for its heterogeneity — grouping breakfast cereals with energy drinks under the same label. The food addiction framework, while supported by approximately 20% prevalence on the YFAS, is not universally accepted in psychiatry or nutrition science [3]. The precise contribution of specific additives (emulsifiers, artificial sweeteners, AGEs) to human disease remains incompletely characterised in human studies, with much of the mechanistic evidence derived from animal models. These are genuine scientific uncertainties — not manufactured doubts.
What would work. The evidence from Chile, Mexico, and other early-mover jurisdictions suggests that comprehensive regulatory packages — combining front-of-pack warning labels, marketing restrictions (especially to children), taxation of the most harmful product categories, and school food standards — can produce measurable reductions in UPF consumption and industry reformulation [3]. Chile's 24% reduction in sugar-sweetened beverage purchases demonstrates that regulation works when it is comprehensive and enforceable. Single-instrument approaches — labelling without marketing restrictions, or taxation without labelling — are less effective than multi-component strategies. ◈ Strong Evidence
The standard for regulatory action is not proof of causation beyond all methodological dispute. It is the precautionary principle: where credible evidence of harm exists, where the population at risk includes those with diminished capacity for self-protection (children), and where the harm is potentially irreversible (chronic disease, developmental effects), the burden of proof should fall on the industry deploying the product — not on the consumers exposed to it. By this standard, the evidence for UPF regulation is not merely sufficient. It is overwhelming. The BMJ umbrella review's 32 adverse outcomes, the Hall RCT's demonstration of engineered overconsumption, the 124,000 estimated annual US deaths, and the documented industry tactics to suppress and distort the science collectively constitute a case for action that exceeds what was available for tobacco regulation when the first meaningful restrictions were imposed.
The critique that NOVA classification is too broad — grouping whole-grain bread with energy drinks under Group 4 — has genuine methodological merit. Not all ultra-processed foods are equally harmful, and subgroup analyses consistently show that processed meats, sugar-sweetened beverages, and industrially produced baked goods drive the largest associations. However, the industry's strategic use of this legitimate critique to argue against any processing-based regulation should be recognised for what it is: a delay tactic. The existence of methodological debate does not invalidate the evidence base. It means the evidence base requires refinement — not rejection. ⚖ Contested
The structural interventions that the evidence supports can be summarised in a hierarchy of effectiveness: at the top, comprehensive regulatory packages combining labelling, marketing restrictions, taxation, and school food standards (Chile model); in the middle, fiscal measures such as sugar taxes that shift relative prices (Mexico model); below that, front-of-pack labelling alone; and at the bottom, education and awareness campaigns without accompanying structural change. The evidence is clear that informing consumers without changing the food environment produces modest and temporary effects. Changing the food environment itself — through regulation of what can be sold, where, to whom, and at what price — produces durable outcomes [3].
The political economy of UPF regulation is, ultimately, the binding constraint. The evidence base is now sufficient to justify comprehensive regulatory action. The mechanistic understanding is advanced enough to guide intervention design. The models from Chile, Mexico, and Brazil demonstrate that regulation is feasible and effective. What is lacking is political will — and political will is precisely what the food industry spends billions annually to suppress. The $4.32 trillion projected cost of obesity-related illness by 2035 [15] will ultimately be borne by healthcare systems, taxpayers, and the individuals who develop preventable chronic diseases. The profits from the products that cause those diseases will accrue to the companies that manufacture them. This is the fundamental asymmetry. Until it is addressed through regulation proportionate to the scale of harm, the invisible epidemic will continue — because it is invisible only to those who choose not to look.
The Lancet 2025 series concluded with a statement that serves as an appropriate final word: "Ultra-processed foods are a defining public health challenge of the twenty-first century. The evidence is sufficient to act. The cost of inaction is measured in millions of preventable deaths" [3]. The WHO's Guideline Development Group now carries the institutional responsibility to translate that conclusion into global guidance. The food industry will resist. The evidence will not go away. The question — as with tobacco, as with climate, as with every previous confrontation between commercial interests and public health — is how many years of preventable death the delay will cost.