{"slug":"ultra-processed-food","name":"Ultra-processed food","aliases":["UPF","NOVA group 4","Ultraprocessed food"],"kind":"lever","system":"exposomics","unit":"percent of daily energy","unitNote":"The share of a day's calories that comes from NOVA group 4, the ultra-processed group, measured by dietary recalls or food records. NOVA sorts foods by the extent and purpose of industrial processing, not by nutrients. Some cohorts counted servings a day or percent of food weight instead; those are approximations of the same exposure and are quoted as the source states them.","answer":"Ultra-processed food supplies about 57 percent of the energy US adults eat. Across 45 pooled analyses covering almost 10 million people, a higher intake went with higher risks of cardiovascular death, type 2 diabetes, obesity, depression and all-cause mortality, though most of that evidence was rated low or very low quality. In an inpatient trial, adults ate about 500 kcal a day more on an ultra-processed diet than on an unprocessed one matched for presented nutrients, and gained 0.9 kg in two weeks. In an eight-week trial where both diets met national guidance, the minimally processed diet produced about twice the weight loss.","summary":"Ultra-processed foods are industrial formulations built from substances extracted or modified from whole foods, plus additives that make them palatable, durable and ready to eat. They now supply the majority of calories in the US and UK. The cohorts consistently link them with cardiometabolic disease and earlier death, two feeding trials show they drive excess intake within weeks, and the open question is how much of the effect is processing itself rather than energy density, eating rate and what they displace.","whatItIs":[{"text":"Ultra-processed food is the fourth group of the NOVA classification, which sorts foods by the extent and purpose of industrial processing rather than by nutrients. The four groups are unprocessed or minimally processed foods, processed culinary ingredients (oils, butter, sugar, salt), processed foods, and ultra-processed foods and drinks. The group's authors describe ultra-processed foods as products made by fractioning whole foods into substances, chemically modifying them, assembling them with cosmetic additives, and packaging them, designed to be profitable, convenient and hyper-palatable. Their practical test is the ingredient list: a product is ultra-processed if it contains at least one ingredient rarely or never used in kitchens (such as high-fructose corn syrup, hydrogenated or interesterified oils, or hydrolyzed proteins) or an additive class meant to make it palatable or appealing (flavors, flavor enhancers, colors, emulsifiers, sweeteners, thickeners, and foaming, gelling or glazing agents).","cites":["monteiro2019","gibney2017"],"grade":"B"},{"text":"It is most of what people eat. In 40,937 US adults across nine national surveys, the share of energy from ultra-processed food rose from 53.5 percent in 2001 to 2002 to 57.0 percent in 2017 to 2018, while minimally processed foods fell from 32.7 to 27.4 percent. Among 33,795 US youths aged 2 to 19 it rose from 61.4 to 67.0 percent between 1999 and 2018. In the UK national diet survey, ultra-processed food supplied 56.8 percent of energy. In the US, intake fell with age and income and was lower in college graduates than in people with less education.","cites":["juul2022","wang2021","rauber2019","baraldi2018"],"grade":"B"},{"text":"A physician cares because it is the dietary exposure with the broadest set of associations in the literature. An umbrella review of 45 pooled analyses with 9,888,373 participants found direct associations between ultra-processed food and 32 of the health parameters studied (71 percent), spanning mortality, cancer, and mental, respiratory, cardiovascular, gastrointestinal and metabolic health. That breadth is itself a reason for caution: an exposure tied to everything may be standing in for a way of living as much as for a food.","cites":["lane2024"],"grade":"B"}],"origins":[{"text":"Processing food is older than our species in its modern form. Humans use more energy than chimpanzees for their body size yet have smaller structures for chewing and digesting, a suite of traits that points to adaptation to a high-quality diet, and the case made by Carmody and Wrangham is that cooking supplied it: cooking makes food easier to chew, more digestible and higher in net energy, and was probably adopted more than 250,000 years ago. Cooking, fermenting, drying and salting changed whole foods. Ultra-processing is a different kind of step: it breaks foods into substances and reassembles them into products. By 2013 such products dominated the food supplies of high-income countries and were rising quickly in middle-income ones, in sales and purchase data from 79 countries, characteristically energy-dense, fatty, sugary or salty, cheap and ready to consume.","cites":["carmody2009","monteiro2013","monteiro2019"],"grade":"C"},{"text":"The classification is young. Monteiro and colleagues set out the NOVA definition and its ingredient-list test in 2019, and it is contested: critics argue that the group is too broad to build dietary guidelines from and that its foods are not uniquely hyper-palatable. Palatability has its own measure. A quantitative definition of hyper-palatable foods, three clusters of fat with sodium, fat with simple sugars, and carbohydrate with sodium above set thresholds, applied to the US food composition database, classified 62 percent of foods (4,795 of 7,757) as hyper-palatable, including some labeled reduced fat or low fat. The two ideas overlap but are not the same thing.","cites":["monteiro2019","gibney2017","fazzino2019"],"grade":"C"},{"text":"The appetite system evolved for whole foods that take time to eat, and formulated products are eaten faster. People tend to eat a consistent weight of food, so a diet of higher energy density delivers more calories in the same bite count. Forde and colleagues combined energy density with eating rate into an energy intake rate: across 327 foods it averaged 35.5 kcal a minute for unprocessed foods, 53.7 for processed and 69.4 for ultra-processed, with wide spread inside each group. In the inpatient trial, the ultra-processed meals were eaten 17 kcal a minute faster, the non-beverage foods were about 85 percent more energy dense, and the participants rated the two diets no differently for pleasantness or familiarity. On this reading the mismatch is speed and density, an intake signal arriving after the calories, more than flavor alone.","cites":["rolls2009","forde2020","hall2019"],"grade":"C"},{"text":"On the shelf, ultra-processed food sits upstream of several markers at once. In US national data it supplied 89.7 percent of the energy from added sugars, and in controlled feeding, fructose-sweetened drinks at a quarter of energy for 10 weeks raised liver fat production, after-meal triglycerides, ApoB, small dense LDL, fasting glucose and insulin and visceral fat, where glucose-sweetened drinks did not; 200 g of fructose a day for two weeks raised fasting triglycerides by 55 mg/dL, blood pressure and fasting insulin, and lowering uric acid with a drug prevented the blood pressure rise. Salt is one of the base ingredients of the group, and salt raises blood pressure. In mice, two common emulsifiers at relatively low concentrations thinned the separation between gut bacteria and the gut lining and produced low-grade inflammation and metabolic syndrome, through the microbiota. And as the ultra-processed share rose in US diets, the content of fiber, protein, potassium, magnesium and vitamins fell: part of the effect is what the products displace.","cites":["martinezsteele2016","stanhope2009","perezpozo2010","he2013","chassaing2015","martinezsteele2017"],"grade":"C"}],"labRange":{"typical":"No laboratory or guideline range exists. US adults get about 57 percent of their energy from ultra-processed food (national surveys, 2017 to 2018) and UK adults about 57 percent (2008 to 2014); several national dietary guidelines advise preferring unprocessed foods and freshly made meals to ultra-processed ones, without a numeric limit.","claims":[{"text":"There is no target number. The advice to prefer unprocessed or minimally processed foods and freshly made meals to ultra-processed foods is being adopted in official dietary guidelines issued by national governments and health associations, according to the classification's lead author; the opposing side of the same debate argues that NOVA adds little to nutrient profiling and labels several nutrient-dense foods as unhealthy. Neither side proposes a percentage of energy that is safe.","cites":["monteiro2022","astrup2022"],"grade":"C"},{"text":"What is typical is high. Ultra-processed food supplied 57.9 percent of US energy intake in 2009 to 2010, almost 60 percent across 2007 to 2012, and 57.0 percent in adults in 2017 to 2018; in the UK it supplied 56.8 percent. The intake of the cohorts below is often lower: a mean of 22.1 percent of energy in the UK Biobank diabetes analysis and 29.1 percent in French adults over 45.","cites":["martinezsteele2016","baraldi2018","juul2022","rauber2019","levy2021","schnabel2019"],"grade":"B"},{"text":"Intake tracks circumstance. In French adults over 45, a higher ultra-processed share went with younger age, lower income, lower education, living alone, higher body mass index and less physical activity; in the US it was higher in younger, lower-income and less-educated groups. Every cohort below adjusted for some of these, but no adjustment removes them entirely, which is the central limit on reading the associations as causes.","cites":["schnabel2019","baraldi2018"],"grade":"B"}]},"evidence":{"headline":"The cohorts agree: across 45 pooled analyses a higher intake went with more cardiovascular death, type 2 diabetes, obesity, depression and all-cause mortality, convincing in strength of association for cardiovascular death and diabetes but mostly low or very low in GRADE quality. Two feeding trials show the effect within weeks: about 500 kcal a day more and 0.9 kg gained in two weeks, and half the weight loss over eight weeks when both diets met national guidance. Whether processing itself, or energy density and eating rate, does the work is not settled.","scale":{"min":0,"max":100,"ticks":[0,20,40,60,80,100]},"bands":[{"label":"Typical in the US and UK: about 54 to 58 percent of energy in national surveys","low":53.5,"high":57.9,"tone":"lab","source":"Juul 2022; Martínez Steele 2016; Rauber 2019"},{"label":"Cohort means in older, health-conscious samples: 22.1 percent (UK Biobank) to 29.1 percent (French adults over 45)","low":22.1,"high":29.1,"tone":"neutral","source":"Levy 2021; Schnabel 2019"},{"label":"US youths aged 2 to 19: rose from 61.4 to 67.0 percent between 1999 and 2018","low":61.4,"high":67,"tone":"caution","source":"Wang 2021"},{"label":"Trial population: habitual intake of 50 percent or more, mean 67.4 percent; the minimally processed diet doubled weight loss in 8 weeks","low":50,"high":67.4,"tone":"caution","source":"Dicken 2025"}],"claims":[{"text":"The umbrella review pooled 45 meta-analyses of cohort, case-control and cross-sectional studies. Convincing evidence (class I) supported higher risks of cardiovascular death (risk ratio 1.50, GRADE very low), type 2 diabetes (1.12 per dose step, GRADE moderate), anxiety (odds ratio 1.48) and common mental disorders (1.53). Highly suggestive evidence (class II) supported higher all-cause mortality (1.21), heart disease death (hazard ratio 1.66), type 2 diabetes (odds ratio 1.40), depression (hazard ratio 1.22), poor sleep (1.41), wheezing (1.40) and obesity (1.55). By GRADE, 22 of the pooled analyses were low quality, 19 very low and four moderate. The strength-of-association class and the quality grade measure different things, and the entry grades this sentence B on the quality.","cites":["lane2024"],"grade":"B"},{"text":"Mortality has been studied in four countries and the direction is the same in each. In 44,551 French adults over 45 followed a median 7.1 years, each 10-point rise in the ultra-processed share of food weight carried a hazard ratio of 1.14. In 19,899 Spanish university graduates with 335 deaths, more than four servings a day carried a hazard ratio of 1.62 against the lowest quarter, and each added daily serving 1.18. In 11,898 US adults followed a median 19 years, the highest quarter of frequency carried a hazard ratio of 1.31 for all-cause mortality and no association with cardiovascular death. Pooling seven cohorts of 207,291 adults gave a hazard ratio of 1.21 for all-cause mortality, 1.50 for cardiovascular death and none for cancer death, and 1.15 for each 10-point rise in share of energy, a linear dose-response.","cites":["schnabel2019","ricocampa2019","kim2019","suksatan2021"],"grade":"B"},{"text":"The longest follow-up gave the smallest number. In 74,563 women and 39,501 men from two US health-professional cohorts, with 48,193 deaths over a median 34 and 31 years, the top quarter of ultra-processed intake carried a hazard ratio of only 1.04 for all-cause mortality and no association with cancer or cardiovascular death. The signal sat in particular subgroups: ready-to-eat meat, poultry and seafood products (hazard ratios 1.06 to 1.43) and sugar-sweetened and artificially sweetened drinks (1.09). Within each quarter of overall diet quality, ultra-processed intake showed no consistent association with mortality, while better diet quality was protective within each quarter of ultra-processed intake. That is the strongest single argument that what people eat overall may matter more than the processing label.","cites":["fang2024"],"grade":"B"},{"text":"For cardiovascular disease, 105,159 French adults followed a median 5.2 years showed a hazard ratio of 1.12 for each 10-point rise in share, with 1,409 cases, and 1.13 for coronary and 1.11 for cerebrovascular disease; incidence was 277 against 242 per 100,000 person-years in the highest and lowest quarters, and the association held after adjustment for saturated fat, sodium, sugar, fiber and a healthy dietary pattern. In 3,003 adults of the Framingham Offspring cohort, who averaged 7.5 servings a day, each added daily serving carried 7 percent higher risk of hard cardiovascular disease and 9 percent higher cardiovascular mortality. In 14,790 Spanish graduates free of hypertension, the highest third of intake carried a hazard ratio of 1.21 for new hypertension over about nine years.","cites":["srour2019","juul2021","mendonca2017"],"grade":"B"},{"text":"Type 2 diabetes is where the cohorts are most consistent. In 104,707 French adults with 821 new cases over a median 6.0 years, each 10-point rise in share carried a hazard ratio of 1.15, and 1.13 after adjustment for weight change; absolute rates were 113 and 166 per 100,000 person-years in the lowest and highest consumers. In 21,730 UK Biobank participants, the highest quarter carried a hazard ratio of 1.44 and each 10-point rise 1.12. The umbrella review rated the diabetes dose-response its only cardiometabolic finding of moderate quality.","cites":["srour2020","levy2021","lane2024"],"grade":"B"},{"text":"Weight follows the same direction. Among 8,451 Spanish graduates without overweight at the start, the highest quarter of intake carried a hazard ratio of 1.26 for becoming overweight or obese over a median 8.9 years. In 22,659 UK Biobank adults followed a median 5 years, the highest quarter carried a hazard ratio of 1.79 for obesity and 1.30 for abdominal obesity. In the UK national diet survey, the highest intake went with a body mass index 1.66 higher and a waist 3.56 cm larger than the lowest. Cross-sectional pooling found higher odds of low HDL cholesterol (102 percent) and metabolic syndrome (79 percent), but no significant association with high blood pressure, high blood sugar or high triglycerides.","cites":["mendonca2016","rauber2021","rauber2020","pagliai2021"],"grade":"B"},{"text":"Cancer and mood are less settled. In 104,980 French adults, each 10 percent rise in share carried a hazard ratio of 1.12 for overall cancer and 1.11 for breast cancer, held after adjustment for nutrient intake. In three US cohorts with 3,216 colorectal cancers over 24 to 28 years, men in the highest fifth had a hazard ratio of 1.29, concentrated in the distal colon (1.72), while women showed no association overall, and yogurt and dairy-based desserts were associated with lower risk in women (0.83). In about 31,000 middle-aged US women free of depression at the start, the highest fifth of intake carried a hazard ratio of 1.49 for depression under a strict definition and 1.32 when exposure was lagged four years, which argues against people eating differently because they were already low.","cites":["fiolet2018","wang2022","samuthpongtorn2023"],"grade":"B"},{"text":"The first trial was small, short and tightly controlled. Twenty weight-stable adults (mean age 31, body mass index 27) lived on an inpatient metabolic ward and ate an ultra-processed diet and an unprocessed diet for two weeks each, in random order, with meals matched for presented calories, energy density, macronutrients, sugar, sodium and fiber and eaten as freely as they liked. They ate 508 kcal a day more on the ultra-processed diet, from carbohydrate and fat but not protein, gained 0.9 kg against a 0.9 kg loss, and gained 0.4 kg of fat against a 0.3 kg loss; weight change tracked intake closely (r 0.8). The appetite-suppressing hormone PYY rose and the hunger hormone ghrelin fell on the unprocessed diet. The trial had no run-in or washout period, and glucose tolerance did not differ between the diets.","cites":["hall2019"],"grade":"B"},{"text":"The second trial asked the harder question. Fifty-five adults in England with a body mass index of 25 to under 40 and a habitual ultra-processed intake of 50 percent or more (mean 67.4 percent; 91 percent women) ate two eight-week diets in random order, both following national healthy-eating guidance, one made of minimally processed food and one of ultra-processed food, delivered and eaten freely. Both produced weight loss, 2.06 percent on minimally processed food against 1.05 percent on ultra-processed, a difference of 1.01 percentage points in the 50 analyzed. Fat mass fell 0.98 kg more, triglycerides fell more (difference 0.25 mmol/L, about 22 mg/dL) and self-rated craving control improved more on the minimally processed diet; LDL cholesterol fell more on the ultra-processed diet (0.25 mmol/L, about 10 mg/dL), and blood pressure changes did not differ. The ultra-processed diet was more energy dense, and a diet-order effect was detected.","cites":["dicken2025"],"grade":"B"},{"text":"A third crossover reached the same answer in nine overweight Japanese men: one week on each diet, matched for presented energy and macronutrients, with a two-week washout. They ate 813.5 kcal a day more and gained 1.1 kg more on the ultra-processed diet, and chewed less per calorie.","cites":["hamano2024"],"grade":"C"},{"text":"Adjusting for diet quality does not make most associations disappear. A review of prospective cohorts found that the majority of associations between ultra-processed food, obesity and health outcomes stayed significant and similar in size after adjustment for diet quality or dietary pattern, and the French cardiovascular, cancer and diabetes analyses each held after adjustment for nutrient markers. The large US mortality analysis above found the opposite pattern. The two readings are not yet reconciled.","cites":["dicken2021","srour2019","fiolet2018","srour2020","fang2024"],"grade":"B"}]},"levers":[{"name":"More meals cooked from whole ingredients","effect":"lowers energy intake and weight within weeks in trials","claim":{"text":"In both controlled trials, the same people ate less and weighed less when the diet was built from unprocessed or minimally processed foods: 508 kcal a day less and a 0.9 kg loss over two weeks on the ward, and twice the weight loss over eight weeks with diets that both met national guidance. The second trial matters most for the kitchen, because its ultra-processed diet was built from products that already carried healthier labels; following the guidance with formulated products bought half the result.","cites":["hall2019","dicken2025"],"grade":"B"}},{"name":"The ingredient-list test","effect":"identifies the products; no trial has tested the test","claim":{"text":"The classification's authors give a practical rule: a product is ultra-processed if its ingredient list includes at least one substance rarely or never used in kitchens, or an additive class whose purpose is to make it palatable or appealing, such as flavors, colors, emulsifiers, sweeteners or thickeners. The nutrition panel does not do this job: a product can meet nutrient targets and still be ultra-processed, which is the point the second trial tested. No study has measured whether teaching the test lowers intake.","cites":["monteiro2019","dicken2025"],"grade":"C"}},{"name":"Drinks first","effect":"the largest single source of added sugar in the group","claim":{"text":"Ultra-processed foods supplied 89.7 percent of the energy from added sugars in the US diet; each 5-point rise in their share raised added-sugar energy by 1 point, and 82.1 percent of people in the highest fifth of intake exceeded the 10 percent added-sugar limit against 26.4 percent in the lowest. In the UK, removing ultra-processed foods could cut the prevalence of excessive free-sugar intake by an estimated 47 percent. Sugar-sweetened and artificially sweetened drinks were among the subgroups tied to higher mortality in the longest cohort (hazard ratio 1.09), and sugar-sweetened drinks to colorectal cancer in men (1.21). Fructose in drink form is the most direct line from this lever to triglycerides, ApoB and fasting insulin in feeding studies.","cites":["martinezsteele2016","rauber2019","fang2024","wang2022","stanhope2009"],"grade":"B"}},{"name":"Slower, firmer food","effect":"lowers intake at the meal","claim":{"text":"Across 22 experiments that changed how fast people ate, slower eating lowered energy intake (standardized mean difference 0.45) without leaving people hungrier at the end of the meal or up to 3.5 hours later. Texture is one way to slow it: in 50 young adults, harder meals were eaten more slowly and cut energy intake by 26 percent, from 789.4 kcal for a soft ultra-processed lunch to 482.9 kcal for a hard minimally processed one, and the difference carried through to total intake that day. Firmer, less processed foods and meals not eaten in a hurry use the same mechanism the inpatient trial measured in reverse.","cites":["robinson2014","teo2022","forde2020"],"grade":"B"}},{"name":"Lower energy density: water-rich foods on the plate","effect":"fewer calories in the same amount of food","claim":{"text":"People tend to eat a consistent weight of food, so lowering the energy density of what is served lowers energy intake, and trials that lowered dietary energy density by adding water-rich fruits and vegetables produced substantial weight loss without telling patients to restrict calories. The non-beverage foods of the ultra-processed diet in the inpatient trial were about 85 percent more energy dense than the unprocessed diet's, and the ultra-processed diet of the second trial was also the more energy dense.","cites":["rolls2009","hall2019","dicken2025"],"grade":"B"}},{"name":"Protein and fiber at meals","effect":"counters the dilution that comes with a higher ultra-processed share","claim":{"text":"In US national data, protein fell from 18.2 to 13.3 percent of energy between the lowest and highest fifths of ultra-processed intake while total energy rose and absolute protein stayed roughly constant, consistent with protein leverage: people keep eating until they reach their protein. Fiber, potassium, magnesium and several vitamins fell across the same gradient. In the inpatient trial the extra 508 kcal came from carbohydrate and fat, not protein. Building meals around a protein food and plants restores what the products displace.","cites":["martinezsteele2018","martinezsteele2017","hall2019"],"grade":"C"}},{"name":"What less salt does to blood pressure","effect":"lowers it, by about 4 mm Hg systolic for 4.4 g a day less salt","claim":{"text":"Salt is one of the base ingredients the definition names alongside sugar, oils and fats. Across 34 trials with 3,230 participants, reducing salt by about 4.4 g a day for four weeks or more lowered blood pressure by 4.18 over 2.06 mm Hg, by 5.39 over 2.82 in people with hypertension and by 2.42 over 1.00 in people without. In the guidance-matched trial, blood pressure fell from baseline only on the minimally processed diet, though the difference between diets was not significant.","cites":["he2013","martinezsteele2016","dicken2025"],"grade":"A"}}],"caveats":[{"text":"The cohorts are observational and the exposure is mostly self-reported. Most used food frequency questionnaires or a few 24-hour recalls; intake follows income, education, age and living alone; and the umbrella review rated 41 of 45 pooled analyses low or very low quality. Consistency across countries and designs is why the associations are graded B here, but a 12 percent higher risk per 10 points of share in people who eat this way is not the same as a 12 percent lower risk for a person who stops.","cites":["lane2024","schnabel2019","baraldi2018"],"grade":"B"},{"text":"The definition has soft edges. Its critics write that NOVA rests on poorly defined processes and a chemically mixed group of additives, which invites misclassification; that experts assigning the same foods to processing levels disagree; and that it labels several nutrient-dense foods as unhealthy. Breads, yogurts, plant milks and breakfast cereals fall on either side of the line depending on their ingredient lists, and the categories inside the group do not behave alike: yogurt and dairy-based desserts were associated with lower colorectal cancer risk in women, while ready-to-eat meat products carried the strongest mortality signal. A whole-grain loaf with an emulsifier and a pastry are both ultra-processed, and the evidence does not say they are equal.","cites":["astrup2022","gibney2017","gibney2023","wang2022","fang2024"],"grade":"C"},{"text":"The trials are few, short and small. The inpatient trial ran two weeks per diet in 20 adults with no run-in or washout; the guidance-matched trial ran eight weeks per diet with 50 people analyzed, 91 percent of them women, and detected a diet-order effect; the Japanese trial was nine men for a week each. None measured disease, and in the longer trial LDL cholesterol fell more on the ultra-processed diet. Effects on weight are well shown; effects on events are inferred.","cites":["hall2019","dicken2025","hamano2024"],"grade":"B"},{"text":"The mechanism may be energy density and eating rate rather than processing as such. Critics argue that the inpatient trial's difference can be explained by energy density, intrinsic fiber, glycemic load and added sugar; the trial's own authors note that the ultra-processed meals were eaten faster and were more energy dense outside the drinks; energy intake rate varies widely within every NOVA group; and in the guidance-matched trial the ultra-processed diet was the denser one. If speed and density are the active ingredients, some ultra-processed foods may be harmless and some minimally processed ones not.","cites":["astrup2022","hall2019","forde2020","dicken2025"],"grade":"C"},{"text":"The additive findings are signals, not verdicts. In about 95,000 French adults, higher intakes of some emulsifiers carried small increases in cardiovascular risk (hazard ratio 1.05 per standard deviation for celluloses, 1.03 for carboxymethylcellulose), and in about 103,000, total artificial sweetener intake carried a hazard ratio of 1.09 for cardiovascular disease; both are single observational cohorts. The emulsifier mechanism comes from mice. A systematic review of non-sugar sweeteners in trials and cohorts found no differences for most health outcomes, with low confidence because the studies were few, small and short.","cites":["sellem2023","debras2022","chassaing2015","toews2019"],"grade":"C"},{"text":"Cost and convenience are real. The inpatient trial's authors note that ultra-processed foods are less expensive and more convenient than meals made from whole foods, and that their study, where food was free and prepared for participants, could not address how cost and convenience shape choices. Intake is higher in lower-income groups. Advice that ignores the price and the time of cooking will not be followed by the people it most concerns.","cites":["hall2019","baraldi2018"],"grade":"C"}],"seeAPhysician":["Unintended weight loss, loss of appetite, trouble swallowing or a change in bowel habit, which call for an evaluation before any change in diet.","Prescribed formulas, tube feeds, oral nutrition supplements or other medical nutrition products: these are formulated products by design and are part of treatment, so change them only with the physician or dietitian who prescribed them.","Diabetes treated with insulin or with tablets that can cause low blood sugar, before a large or sudden change in how much carbohydrate and sugar is eaten, because doses may need to come down.","A history of an eating disorder, or food rules that are starting to cause distress or isolation, where a rule about processing can do more harm than the food.","Gout, kidney stones or a raised uric acid, very high triglycerides, or a raised ALT or GGT, which have causes beyond diet and deserve their own workup.","Food allergy, celiac disease or inflammatory bowel disease, where some processed products are the safe choice and a blanket rule against them does not fit."],"openQuestions":["No randomized trial has tested cutting ultra-processed food against disease events or death. The longest controlled trial ran eight weeks per diet and measured weight, body composition and markers.","Whether processing itself causes harm, or whether energy density, eating rate, texture, added sugar and the foods displaced explain the effect, is the field's central dispute. A trial that matches diets for energy density and eating rate and still finds a difference would settle much of it.","Which parts of the group matter. The longest US cohort tied mortality to ready-to-eat meat products and sweetened drinks while finding little for the group as a whole once diet quality was considered, and some subgroups, such as yogurt, run in the other direction.","Whether the emulsifier and sweetener signals are causal in people. The emulsifier mechanism rests on mice, and the human data are single observational cohorts with small hazard ratios.","Whether a NOVA classification applied to food frequency questionnaires measures intake reliably enough for the dose-response curves to be trusted, given that experts disagree when assigning foods to groups."],"questions":[{"q":"What counts as an ultra-processed food?","a":"In the NOVA classification it is an industrial formulation made from substances extracted or modified from whole foods, plus additives that make it palatable or durable. The authors' own test is the ingredient list: if it contains at least one ingredient rarely or never used in kitchens, such as high-fructose corn syrup, hydrogenated oils or hydrolyzed proteins, or an additive class such as flavors, colors, emulsifiers, sweeteners or thickeners, the product is ultra-processed. Cooking, freezing, fermenting, canning with salt or drying do not by themselves make a food ultra-processed."},{"q":"Is ultra-processed food bad for you, or is that just people with worse diets overall?","a":"Both readings have evidence. Across 45 pooled analyses, higher intake went with higher risks of cardiovascular death, type 2 diabetes, obesity, depression and death from any cause, and most associations held after adjustment for diet quality. But the longest US cohort found only a 4 percent higher mortality in the top quarter and no consistent association within levels of diet quality. The trials settle the narrower question: in a controlled ward, people ate about 500 kcal a day more on ultra-processed food, and over eight weeks, guideline-compliant ultra-processed food produced about half the weight loss of minimally processed food."},{"q":"How much of the US diet is ultra-processed?","a":"About 57 percent of the energy US adults ate in 2017 to 2018 came from ultra-processed food, up from 53.5 percent in 2001 to 2002, and about 67 percent of what US children and teenagers ate in 2017 to 2018. The UK figure is similar, about 57 percent. No guideline sets a numeric limit."},{"q":"Why do ultra-processed foods make people eat more?","a":"The leading explanations are energy density and eating rate. In the inpatient trial, the ultra-processed meals were eaten 17 kcal a minute faster and the solid foods were about 85 percent more energy dense, while participants rated both diets equally pleasant; the hormone PYY, which signals fullness, rose on the unprocessed diet. Across many foods, ultra-processed ones deliver about twice the calories per minute of unprocessed ones, and slowing eating lowers intake in experiments. Lower protein density and less fiber probably contribute."}],"citations":[{"id":"monteiro2019","authors":"Monteiro CA, Cannon G, Levy RB, et al.","title":"Ultra-processed foods: what they are and how to identify them","journal":"Public Health Nutrition","year":2019,"pmid":"30744710","doi":"10.1017/S1368980018003762","type":"review","finding":"Defines ultra-processed foods within NOVA by the extent and purpose of industrial processing; a product is ultra-processed if its ingredient list contains at least one substance never or rarely used in kitchens or an additive class used to make it palatable or appealing (flavors, colors, emulsifiers, sweeteners, thickeners and others).","url":"https://pubmed.ncbi.nlm.nih.gov/30744710/"},{"id":"gibney2017","authors":"Gibney MJ, Forde CG, Mullally D, Gibney ER","title":"Ultra-processed foods in human health: a critical appraisal","journal":"American Journal of Clinical Nutrition","year":2017,"pmid":"28793996","doi":"10.3945/ajcn.117.160440","type":"review","finding":"Sets out NOVA's four categories and argues that there is no evidence ultra-processed foods are uniquely hyper-palatable or quasi-addictive, that the definition is too broad to build food-based dietary guidelines from, and that NOVA offers little advantage over nutrient-based epidemiology.","url":"https://pubmed.ncbi.nlm.nih.gov/28793996/"},{"id":"gibney2023","authors":"Gibney MJ","title":"Ultra-processed foods in public health nutrition: the unanswered questions","journal":"Public Health Nutrition","year":2023,"pmid":"36514812","doi":"10.1017/S1368980022002105","type":"review","finding":"Different definitions of processing give different associations on the same data, assignment of foods to processing levels by experts is variable, and coding is prone to error; the relative importance of processing versus formulation is unresolved.","url":"https://pubmed.ncbi.nlm.nih.gov/36514812/"},{"id":"astrup2022","authors":"Astrup A, Monteiro CA","title":"Does the concept of \"ultra-processed foods\" help inform dietary guidelines, beyond conventional classification systems? NO","journal":"American Journal of Clinical Nutrition","year":2022,"pmid":"35670128","doi":"10.1093/ajcn/nqac123","type":"review","finding":"Argues NOVA rests on poorly defined processes and invites misclassification; that effects beyond nutrient composition, energy density and food matrix are poorly supported; that the inpatient trial's result can be explained by energy density, intrinsic fiber, glycemic load and added sugar; and that NOVA labels several nutrient-dense foods as unhealthy.","url":"https://pubmed.ncbi.nlm.nih.gov/35670128/"},{"id":"monteiro2022","authors":"Monteiro CA, Astrup A","title":"Does the concept of \"ultra-processed foods\" help inform dietary guidelines, beyond conventional classification systems? YES","journal":"American Journal of Clinical Nutrition","year":2022,"pmid":"35670127","doi":"10.1093/ajcn/nqac122","type":"review","finding":"The recommendation to prefer unprocessed or minimally processed foods and freshly made meals to ultra-processed foods is increasingly adopted in official dietary guidelines and is supported by meta-analyses of surveys and cohorts; different attributes probably act through different mechanisms.","url":"https://pubmed.ncbi.nlm.nih.gov/35670127/"},{"id":"juul2022","authors":"Juul F, Parekh N, Martinez-Steele E, Monteiro CA, Chang VW","title":"Ultra-processed food consumption among US adults from 2001 to 2018","journal":"American Journal of Clinical Nutrition","year":2022,"pmid":"34647997","doi":"10.1093/ajcn/nqab305","type":"cross-sectional","finding":"In 40,937 US adults across nine national surveys, ultra-processed food rose from 53.5 to 57.0 percent of energy between 2001 to 2002 and 2017 to 2018, while minimally processed foods fell from 32.7 to 27.4 percent.","url":"https://pubmed.ncbi.nlm.nih.gov/34647997/"},{"id":"wang2021","authors":"Wang L, Martínez Steele E, Du M, et al.","title":"Trends in Consumption of Ultraprocessed Foods Among US Youths Aged 2-19 Years, 1999-2018","journal":"JAMA","year":2021,"pmid":"34374722","doi":"10.1001/jama.2021.10238","type":"cross-sectional","finding":"In 33,795 US youths aged 2 to 19, the share of energy from ultra-processed foods rose from 61.4 to 67.0 percent between 1999 and 2018.","url":"https://pubmed.ncbi.nlm.nih.gov/34374722/"},{"id":"rauber2019","authors":"Rauber F, Louzada MLDC, Martinez Steele E, et al.","title":"Ultra-processed foods and excessive free sugar intake in the UK: a nationally representative cross-sectional study","journal":"BMJ Open","year":2019,"pmid":"31662351","doi":"10.1136/bmjopen-2018-027546","type":"cross-sectional","finding":"In 9,364 UK survey participants, ultra-processed foods supplied 56.8 percent of energy and 64.7 percent of free sugars; eliminating them could cut the prevalence of excessive free-sugar intake by an estimated 47 percent.","url":"https://pubmed.ncbi.nlm.nih.gov/31662351/"},{"id":"baraldi2018","authors":"Baraldi LG, Martinez Steele E, Canella DS, Monteiro CA","title":"Consumption of ultra-processed foods and associated sociodemographic factors in the USA between 2007 and 2012: evidence from a nationally representative cross-sectional study","journal":"BMJ Open","year":2018,"pmid":"29525772","doi":"10.1136/bmjopen-2017-020574","type":"cross-sectional","finding":"In 23,847 US survey participants, almost 60 percent of calories came from ultra-processed foods; intake decreased with age and income and was lower in college graduates.","url":"https://pubmed.ncbi.nlm.nih.gov/29525772/"},{"id":"lane2024","authors":"Lane MM, Gamage E, Du S, et al.","title":"Ultra-processed food exposure and adverse health outcomes: umbrella review of epidemiological meta-analyses","journal":"BMJ","year":2024,"pmid":"38418082","doi":"10.1136/bmj-2023-077310","type":"meta-analysis","finding":"Across 45 pooled analyses (9,888,373 participants), ultra-processed food was directly associated with 32 health parameters; convincing evidence for cardiovascular death (RR 1.50) and type 2 diabetes (dose-response RR 1.12, moderate quality); highly suggestive for all-cause mortality (1.21), depression (1.22) and obesity (1.55); 22 analyses low and 19 very low quality by GRADE.","url":"https://pubmed.ncbi.nlm.nih.gov/38418082/"},{"id":"carmody2009","authors":"Carmody RN, Wrangham RW","title":"Cooking and the human commitment to a high-quality diet","journal":"Cold Spring Harbor Symposia on Quantitative Biology","year":2009,"pmid":"19843593","doi":"10.1101/sqb.2009.74.019","type":"review","finding":"Humans have high energy use but reduced chewing and digestive structures for their size, consistent with adaptation to a high-quality diet; cooking eases chewing, raises digestibility and net energy, and was probably adopted more than 250,000 years ago.","url":"https://pubmed.ncbi.nlm.nih.gov/19843593/"},{"id":"monteiro2013","authors":"Monteiro CA, Moubarac JC, Cannon G, Ng SW, Popkin B","title":"Ultra-processed products are becoming dominant in the global food system","journal":"Obesity Reviews","year":2013,"pmid":"24102801","doi":"10.1111/obr.12107","type":"review","finding":"In sales and purchase data from 79 high- and middle-income countries, ultra-processed products, characteristically energy-dense, fatty, sugary or salty, hyper-palatable, cheap and ready to consume, dominate the food supplies of high-income countries and are rising rapidly in middle-income ones.","url":"https://pubmed.ncbi.nlm.nih.gov/24102801/"},{"id":"fazzino2019","authors":"Fazzino TL, Rohde K, Sullivan DK","title":"Hyper-Palatable Foods: Development of a Quantitative Definition and Application to the US Food System Database","journal":"Obesity","year":2019,"pmid":"31689013","doi":"10.1002/oby.22639","type":"cross-sectional","finding":"A data-derived definition of hyper-palatable foods (fat and sodium, fat and simple sugars, or carbohydrate and sodium above set thresholds) classified 62 percent of 7,757 foods in the US food composition database as hyper-palatable, including some labeled reduced or low fat.","url":"https://pubmed.ncbi.nlm.nih.gov/31689013/"},{"id":"rolls2009","authors":"Rolls BJ","title":"The relationship between dietary energy density and energy intake","journal":"Physiology and Behavior","year":2009,"pmid":"19303887","doi":"10.1016/j.physbeh.2009.03.011","type":"review","finding":"People tend to eat a consistent weight of food, so lowering dietary energy density lowers energy intake; trials lowering energy density with water-rich fruits and vegetables produced substantial weight loss without instructions to restrict calories.","url":"https://pubmed.ncbi.nlm.nih.gov/19303887/"},{"id":"forde2020","authors":"Forde CG, Mars M, de Graaf K","title":"Ultra-Processing or Oral Processing? A Role for Energy Density and Eating Rate in Moderating Energy Intake from Processed Foods","journal":"Current Developments in Nutrition","year":2020,"pmid":"32110771","doi":"10.1093/cdn/nzaa019","type":"cross-sectional","finding":"Across 327 foods from five studies, energy intake rate averaged 35.5 kcal/min for unprocessed, 53.7 for processed and 69.4 for ultra-processed foods, with wide variability within each group; in the inpatient trial it was 48 against 31 kcal/min.","url":"https://pubmed.ncbi.nlm.nih.gov/32110771/"},{"id":"hall2019","authors":"Hall KD, Ayuketah A, Brychta R, et al.","title":"Ultra-Processed Diets Cause Excess Calorie Intake and Weight Gain: An Inpatient Randomized Controlled Trial of Ad Libitum Food Intake","journal":"Cell Metabolism","year":2019,"pmid":"31105044","doi":"10.1016/j.cmet.2019.05.008","type":"rct","finding":"In 20 inpatient adults eating each diet freely for two weeks, with meals matched for presented calories, energy density, macronutrients, sugar, sodium and fiber, energy intake was 508 kcal a day higher on the ultra-processed diet and weight rose 0.9 kg against a 0.9 kg loss; meals were eaten 17 kcal a minute faster and rated no more pleasant.","url":"https://pubmed.ncbi.nlm.nih.gov/31105044/"},{"id":"martinezsteele2016","authors":"Martínez Steele E, Baraldi LG, Louzada ML, et al.","title":"Ultra-processed foods and added sugars in the US diet: evidence from a nationally representative cross-sectional study","journal":"BMJ Open","year":2016,"pmid":"26962035","doi":"10.1136/bmjopen-2015-009892","type":"cross-sectional","finding":"In 9,317 US survey participants, ultra-processed foods supplied 57.9 percent of energy and 89.7 percent of energy from added sugars; added sugars rose from 7.5 to 19.5 percent of energy across fifths of ultra-processed intake, and 82.1 percent of the top fifth exceeded the 10 percent limit against 26.4 percent of the lowest.","url":"https://pubmed.ncbi.nlm.nih.gov/26962035/"},{"id":"stanhope2009","authors":"Stanhope KL, Schwarz JM, Keim NL, et al.","title":"Consuming fructose-sweetened, not glucose-sweetened, beverages increases visceral adiposity and lipids and decreases insulin sensitivity in overweight/obese humans","journal":"Journal of Clinical Investigation","year":2009,"pmid":"19381015","doi":"10.1172/JCI37385","type":"rct","finding":"In overweight and obese adults drinking glucose- or fructose-sweetened beverages at 25 percent of energy for 10 weeks, fructose specifically raised hepatic de novo lipogenesis, postprandial triglycerides, fasting ApoB, LDL and small dense LDL, fasting glucose and insulin and visceral fat, and lowered insulin sensitivity.","url":"https://pubmed.ncbi.nlm.nih.gov/19381015/"},{"id":"perezpozo2010","authors":"Perez-Pozo SE, Schold J, Nakagawa T, et al.","title":"Excessive fructose intake induces the features of metabolic syndrome in healthy adult men: role of uric acid in the hypertensive response","journal":"International Journal of Obesity","year":2010,"pmid":"20029377","doi":"10.1038/ijo.2009.259","type":"rct","finding":"In 74 men given 200 g of fructose a day for two weeks, ambulatory blood pressure rose 7 over 5 mm Hg, fasting triglycerides rose 55 mg/dL and fasting insulin and HOMA rose; lowering uric acid with allopurinol prevented the blood pressure rise but not the triglyceride or HOMA rise.","url":"https://pubmed.ncbi.nlm.nih.gov/20029377/"},{"id":"he2013","authors":"He FJ, Li J, Macgregor GA","title":"Effect of longer term modest salt reduction on blood pressure: Cochrane systematic review and meta-analysis of randomised trials","journal":"BMJ","year":2013,"pmid":"23558162","doi":"10.1136/bmj.f1325","type":"meta-analysis","finding":"Across 34 trials (3,230 participants), reducing salt by about 4.4 g a day for four weeks or more lowered blood pressure by 4.18 over 2.06 mm Hg; 5.39 over 2.82 in hypertension and 2.42 over 1.00 in normotension.","url":"https://pubmed.ncbi.nlm.nih.gov/23558162/"},{"id":"chassaing2015","authors":"Chassaing B, Koren O, Goodrich JK, et al.","title":"Dietary emulsifiers impact the mouse gut microbiota promoting colitis and metabolic syndrome","journal":"Nature","year":2015,"pmid":"25731162","doi":"10.1038/nature14232","type":"laboratory","finding":"In mice, relatively low concentrations of carboxymethylcellulose and polysorbate-80 caused microbiota encroachment on the gut lining, low-grade inflammation and obesity or metabolic syndrome, and colitis in predisposed mice; germ-free and fecal-transplant experiments showed the microbiota changes were necessary and sufficient.","url":"https://pubmed.ncbi.nlm.nih.gov/25731162/"},{"id":"martinezsteele2017","authors":"Martínez Steele E, Popkin BM, Swinburn B, Monteiro CA","title":"The share of ultra-processed foods and the overall nutritional quality of diets in the US: evidence from a nationally representative cross-sectional study","journal":"Population Health Metrics","year":2017,"pmid":"28193285","doi":"10.1186/s12963-017-0119-3","type":"cross-sectional","finding":"In 9,317 US survey participants, the dietary content of protein, fiber, vitamins A, C, D and E, zinc, potassium, phosphorus, magnesium and calcium fell, and carbohydrate, added sugar and saturated fat rose, across fifths of ultra-processed energy share.","url":"https://pubmed.ncbi.nlm.nih.gov/28193285/"},{"id":"martinezsteele2018","authors":"Martínez Steele E, Raubenheimer D, Simpson SJ, Baraldi LG, Monteiro CA","title":"Ultra-processed foods, protein leverage and energy intake in the USA","journal":"Public Health Nutrition","year":2018,"pmid":"29032787","doi":"10.1017/S1368980017001574","type":"cross-sectional","finding":"In 9,042 US survey participants, protein fell from 18.2 to 13.3 percent of energy between the lowest and highest fifths of ultra-processed intake while total energy rose and absolute protein stayed roughly constant, consistent with protein leverage.","url":"https://pubmed.ncbi.nlm.nih.gov/29032787/"},{"id":"levy2021","authors":"Levy RB, Rauber F, Chang K, et al.","title":"Ultra-processed food consumption and type 2 diabetes incidence: A prospective cohort study","journal":"Clinical Nutrition","year":2021,"pmid":"33388205","doi":"10.1016/j.clnu.2020.12.018","type":"cohort","finding":"In 21,730 UK Biobank participants (mean intake 22.1 percent of energy) with 305 new cases of type 2 diabetes over 5.4 years, the highest quarter carried a hazard ratio of 1.44 and each 10-point rise 1.12.","url":"https://pubmed.ncbi.nlm.nih.gov/33388205/"},{"id":"schnabel2019","authors":"Schnabel L, Kesse-Guyot E, Allès B, et al.","title":"Association Between Ultraprocessed Food Consumption and Risk of Mortality Among Middle-aged Adults in France","journal":"JAMA Internal Medicine","year":2019,"pmid":"30742202","doi":"10.1001/jamainternmed.2018.7289","type":"cohort","finding":"In 44,551 French adults over 45 (ultra-processed food 29.1 percent of energy) with 602 deaths over a median 7.1 years, each 10 percent rise in the share of food weight carried a hazard ratio of 1.14; higher intake went with younger age, lower income and education, living alone, higher BMI and less activity.","url":"https://pubmed.ncbi.nlm.nih.gov/30742202/"},{"id":"ricocampa2019","authors":"Rico-Campà A, Martínez-González MA, Alvarez-Alvarez I, et al.","title":"Association between consumption of ultra-processed foods and all cause mortality: SUN prospective cohort study","journal":"BMJ","year":2019,"pmid":"31142450","doi":"10.1136/bmj.l1949","type":"cohort","finding":"In 19,899 Spanish university graduates with 335 deaths, more than four servings a day carried a hazard ratio of 1.62 for all-cause mortality against the lowest quarter, and each added daily serving 1.18.","url":"https://pubmed.ncbi.nlm.nih.gov/31142450/"},{"id":"kim2019","authors":"Kim H, Hu EA, Rebholz CM","title":"Ultra-processed food intake and mortality in the USA: results from the Third National Health and Nutrition Examination Survey (NHANES III, 1988-1994)","journal":"Public Health Nutrition","year":2019,"pmid":"30789115","doi":"10.1017/S1368980018003890","type":"cohort","finding":"In 11,898 US adults followed a median 19 years, the highest quarter of ultra-processed intake frequency carried a hazard ratio of 1.31 for all-cause mortality; no association with cardiovascular mortality.","url":"https://pubmed.ncbi.nlm.nih.gov/30789115/"},{"id":"suksatan2021","authors":"Suksatan W, Moradi S, Naeini F, et al.","title":"Ultra-Processed Food Consumption and Adult Mortality Risk: A Systematic Review and Dose-Response Meta-Analysis of 207,291 Participants","journal":"Nutrients","year":2021,"pmid":"35011048","doi":"10.3390/nu14010174","type":"meta-analysis","finding":"Across seven cohorts (207,291 adults, four countries), ultra-processed intake carried hazard ratios of 1.21 for all-cause, 1.50 for cardiovascular and 1.66 for heart-disease mortality, none for cancer mortality, and 1.15 per 10 percent of energy, with a linear dose-response.","url":"https://pubmed.ncbi.nlm.nih.gov/35011048/"},{"id":"fang2024","authors":"Fang Z, Rossato SL, Hang D, et al.","title":"Association of ultra-processed food consumption with all cause and cause specific mortality: population based cohort study","journal":"BMJ","year":2024,"pmid":"38719536","doi":"10.1136/bmj-2023-078476","type":"cohort","finding":"In 114,064 US health professionals with 48,193 deaths over about three decades, the top quarter carried a hazard ratio of 1.04 for all-cause mortality and none for cancer or cardiovascular death; ready-to-eat meat products (1.06 to 1.43) and sweetened drinks (1.09) drove the signal; no consistent association within quarters of diet quality.","url":"https://pubmed.ncbi.nlm.nih.gov/38719536/"},{"id":"srour2019","authors":"Srour B, Fezeu LK, Kesse-Guyot E, et al.","title":"Ultra-processed food intake and risk of cardiovascular disease: prospective cohort study (NutriNet-Santé)","journal":"BMJ","year":2019,"pmid":"31142457","doi":"10.1136/bmj.l1451","type":"cohort","finding":"In 105,159 French adults with 1,409 cardiovascular events over a median 5.2 years, each 10-point rise in ultra-processed share carried a hazard ratio of 1.12 for cardiovascular, 1.13 for coronary and 1.11 for cerebrovascular disease, holding after adjustment for diet-quality markers.","url":"https://pubmed.ncbi.nlm.nih.gov/31142457/"},{"id":"juul2021","authors":"Juul F, Vaidean G, Lin Y, Deierlein AL, Parekh N","title":"Ultra-Processed Foods and Incident Cardiovascular Disease in the Framingham Offspring Study","journal":"Journal of the American College of Cardiology","year":2021,"pmid":"33766258","doi":"10.1016/j.jacc.2021.01.047","type":"cohort","finding":"In 3,003 adults averaging 7.5 servings a day, each added daily serving of ultra-processed food carried 7 percent higher risk of hard cardiovascular disease, 9 percent of hard coronary disease and 9 percent of cardiovascular mortality.","url":"https://pubmed.ncbi.nlm.nih.gov/33766258/"},{"id":"mendonca2017","authors":"Mendonça RD, Lopes AC, Pimenta AM, et al.","title":"Ultra-Processed Food Consumption and the Incidence of Hypertension in a Mediterranean Cohort: The Seguimiento Universidad de Navarra Project","journal":"American Journal of Hypertension","year":2017,"pmid":"27927627","doi":"10.1093/ajh/hpw137","type":"cohort","finding":"In 14,790 Spanish graduates free of hypertension, with 1,702 new cases over a mean 9.1 years, the highest third of ultra-processed intake carried a hazard ratio of 1.21 for hypertension.","url":"https://pubmed.ncbi.nlm.nih.gov/27927627/"},{"id":"srour2020","authors":"Srour B, Fezeu LK, Kesse-Guyot E, et al.","title":"Ultraprocessed Food Consumption and Risk of Type 2 Diabetes Among Participants of the NutriNet-Santé Prospective Cohort","journal":"JAMA Internal Medicine","year":2020,"pmid":"31841598","doi":"10.1001/jamainternmed.2019.5942","type":"cohort","finding":"In 104,707 French adults with 821 new cases over a median 6.0 years, each 10-point rise in ultra-processed share carried a hazard ratio of 1.15 for type 2 diabetes, 1.13 after adjustment for weight change; absolute rates 113 and 166 per 100,000 person-years in lowest and highest consumers.","url":"https://pubmed.ncbi.nlm.nih.gov/31841598/"},{"id":"mendonca2016","authors":"Mendonça RD, Pimenta AM, Gea A, et al.","title":"Ultraprocessed food consumption and risk of overweight and obesity: the University of Navarra Follow-Up (SUN) cohort study","journal":"American Journal of Clinical Nutrition","year":2016,"pmid":"27733404","doi":"10.3945/ajcn.116.135004","type":"cohort","finding":"In 8,451 Spanish graduates initially without overweight, with 1,939 new cases over a median 8.9 years, the highest quarter of ultra-processed intake carried a hazard ratio of 1.26 for overweight or obesity.","url":"https://pubmed.ncbi.nlm.nih.gov/27733404/"},{"id":"rauber2021","authors":"Rauber F, Chang K, Vamos EP, et al.","title":"Ultra-processed food consumption and risk of obesity: a prospective cohort study of UK Biobank","journal":"European Journal of Nutrition","year":2021,"pmid":"33070213","doi":"10.1007/s00394-020-02367-1","type":"cohort","finding":"In 22,659 UK Biobank adults followed a median 5 years, the highest quarter of ultra-processed intake carried hazard ratios of 1.79 for obesity and 1.30 for abdominal obesity against the lowest.","url":"https://pubmed.ncbi.nlm.nih.gov/33070213/"},{"id":"rauber2020","authors":"Rauber F, Steele EM, Louzada MLDC, et al.","title":"Ultra-processed food consumption and indicators of obesity in the United Kingdom population (2008-2016)","journal":"PLoS One","year":2020,"pmid":"32357191","doi":"10.1371/journal.pone.0232676","type":"cross-sectional","finding":"In 6,143 UK adults, the highest ultra-processed intake went with a 1.66 kg/m2 higher BMI, a 3.56 cm larger waist and 90 percent higher odds of obesity than the lowest.","url":"https://pubmed.ncbi.nlm.nih.gov/32357191/"},{"id":"pagliai2021","authors":"Pagliai G, Dinu M, Madarena MP, et al.","title":"Consumption of ultra-processed foods and health status: a systematic review and meta-analysis","journal":"British Journal of Nutrition","year":2021,"pmid":"32792031","doi":"10.1017/S0007114520002688","type":"meta-analysis","finding":"In 10 cross-sectional studies, the highest intake went with higher odds of low HDL cholesterol (102 percent) and metabolic syndrome (79 percent) but not hypertension, hyperglycemia or high triglycerides; in 13 cohorts (183,491 participants), all-cause mortality RR 1.25 and cardiovascular disease RR 1.29.","url":"https://pubmed.ncbi.nlm.nih.gov/32792031/"},{"id":"fiolet2018","authors":"Fiolet T, Srour B, Sellem L, et al.","title":"Consumption of ultra-processed foods and cancer risk: results from NutriNet-Santé prospective cohort","journal":"BMJ","year":2018,"pmid":"29444771","doi":"10.1136/bmj.k322","type":"cohort","finding":"In 104,980 French adults with 2,228 cancers, each 10 percent rise in ultra-processed share carried a hazard ratio of 1.12 for overall and 1.11 for breast cancer, holding after adjustment for nutrient intake and a Western pattern.","url":"https://pubmed.ncbi.nlm.nih.gov/29444771/"},{"id":"wang2022","authors":"Wang L, Du M, Wang K, et al.","title":"Association of ultra-processed food consumption with colorectal cancer risk among men and women: results from three prospective US cohort studies","journal":"BMJ","year":2022,"pmid":"38752573","doi":"10.1136/bmj-2021-068921","type":"cohort","finding":"In 46,341 men and 159,907 women with 3,216 colorectal cancers over 24 to 28 years, men in the highest fifth had a hazard ratio of 1.29 (distal colon 1.72); no association in women overall; yogurt and dairy-based desserts were associated with lower risk in women (0.83).","url":"https://pubmed.ncbi.nlm.nih.gov/38752573/"},{"id":"samuthpongtorn2023","authors":"Samuthpongtorn C, Nguyen LH, Okereke OI, et al.","title":"Consumption of Ultraprocessed Food and Risk of Depression","journal":"JAMA Network Open","year":2023,"pmid":"37728928","doi":"10.1001/jamanetworkopen.2023.34770","type":"cohort","finding":"In about 31,000 US women aged 42 to 62 free of depression, the highest fifth of ultra-processed intake carried a hazard ratio of 1.49 for depression (strict definition) and 1.34 (broad); 1.32 with a four-year exposure lag.","url":"https://pubmed.ncbi.nlm.nih.gov/37728928/"},{"id":"dicken2025","authors":"Dicken SJ, Jassil FC, Brown A, et al.","title":"Ultraprocessed or minimally processed diets following healthy dietary guidelines on weight and cardiometabolic health: a randomized, crossover trial","journal":"Nature Medicine","year":2025,"pmid":"40760353","doi":"10.1038/s41591-025-03842-0","type":"rct","finding":"In 55 adults in England with habitual intake of 50 percent or more (50 analyzed), eight-week guideline-compliant diets eaten freely produced 2.06 percent weight loss on minimally processed food against 1.05 percent on ultra-processed food (difference 1.01 points); fat mass and triglycerides fell more on minimally processed food, LDL cholesterol more on ultra-processed food.","url":"https://pubmed.ncbi.nlm.nih.gov/40760353/"},{"id":"hamano2024","authors":"Hamano S, Sawada M, Aihara M, et al.","title":"Ultra-processed foods cause weight gain and increased energy intake associated with reduced chewing frequency: A randomized, open-label, crossover study","journal":"Diabetes, Obesity and Metabolism","year":2024,"pmid":"39267249","doi":"10.1111/dom.15922","type":"rct","finding":"In nine overweight Japanese men eating each diet for one week with a two-week washout, the ultra-processed diet produced 813.5 kcal a day more intake and 1.1 kg more weight gain, with fewer chews per calorie.","url":"https://pubmed.ncbi.nlm.nih.gov/39267249/"},{"id":"dicken2021","authors":"Dicken SJ, Batterham RL","title":"The Role of Diet Quality in Mediating the Association between Ultra-Processed Food Intake, Obesity and Health-Related Outcomes: A Review of Prospective Cohort Studies","journal":"Nutrients","year":2021,"pmid":"35010898","doi":"10.3390/nu14010023","type":"review","finding":"Across prospective cohorts, the majority of associations between ultra-processed food, obesity and health outcomes remained significant and unchanged in size after adjustment for diet quality or dietary pattern.","url":"https://pubmed.ncbi.nlm.nih.gov/35010898/"},{"id":"robinson2014","authors":"Robinson E, Almiron-Roig E, Rutters F, et al.","title":"A systematic review and meta-analysis examining the effect of eating rate on energy intake and hunger","journal":"American Journal of Clinical Nutrition","year":2014,"pmid":"24847856","doi":"10.3945/ajcn.113.081745","type":"meta-analysis","finding":"Across 22 experiments, slower eating lowered energy intake against faster eating (SMD 0.45), with no difference in hunger at the end of the meal or up to 3.5 hours later.","url":"https://pubmed.ncbi.nlm.nih.gov/24847856/"},{"id":"teo2022","authors":"Teo PS, Lim AJ, Goh AT, et al.","title":"Texture-based differences in eating rate influence energy intake for minimally processed and ultra-processed meals","journal":"American Journal of Clinical Nutrition","year":2022,"pmid":"35285882","doi":"10.1093/ajcn/nqac068","type":"rct","finding":"In a crossover of 50 healthy-weight adults, harder meals were eaten more slowly and cut energy intake by 26 percent; intake ranged from 482.9 kcal for a hard minimally processed lunch to 789.4 kcal for a soft ultra-processed one, and total daily intake followed.","url":"https://pubmed.ncbi.nlm.nih.gov/35285882/"},{"id":"sellem2023","authors":"Sellem L, Srour B, Javaux G, et al.","title":"Food additive emulsifiers and risk of cardiovascular disease in the NutriNet-Santé cohort: prospective cohort study","journal":"BMJ","year":2023,"pmid":"37673430","doi":"10.1136/bmj-2023-076058","type":"cohort","finding":"In 95,442 French adults with 1,995 cardiovascular events over a median 7.4 years, higher intakes of celluloses (HR 1.05 per SD), carboxymethylcellulose (1.03), and mono- and diglycerides of fatty acids were associated with higher cardiovascular risk.","url":"https://pubmed.ncbi.nlm.nih.gov/37673430/"},{"id":"debras2022","authors":"Debras C, Chazelas E, Sellem L, et al.","title":"Artificial sweeteners and risk of cardiovascular diseases: results from the prospective NutriNet-Santé cohort","journal":"BMJ","year":2022,"pmid":"36638072","doi":"10.1136/bmj-2022-071204","type":"cohort","finding":"In 103,388 French adults, total artificial sweetener intake carried a hazard ratio of 1.09 for cardiovascular disease and 1.18 for cerebrovascular disease.","url":"https://pubmed.ncbi.nlm.nih.gov/36638072/"},{"id":"toews2019","authors":"Toews I, Lohner S, Küllenberg de Gaudry D, Sommer H, Meerpohl JJ","title":"Association between intake of non-sugar sweeteners and health outcomes: systematic review and meta-analyses of randomised and non-randomised controlled trials and observational studies","journal":"BMJ","year":2019,"pmid":"30602577","doi":"10.1136/bmj.k4718","type":"meta-analysis","finding":"Across 56 studies, most health outcomes did not differ between people exposed and unexposed to non-sugar sweeteners; studies were few, small and short, so confidence is limited.","url":"https://pubmed.ncbi.nlm.nih.gov/30602577/"}],"written":"2026-10-23","updated":"2026-10-23","reviewed":true,"reviewedOn":"2026-10-24","related":{"essays":[],"entries":["fiber","protein-intake","triglycerides","fasting-insulin","homa-ir","hba1c","uric-acid","apob","hs-crp","alt","ggt"]},"url":"https://taggehealth.com/labs/ultra-processed-food","author":"Daniel Tagge, MD","publisher":"Tagge Health","license":"Free to read and cite with attribution to the URL above.","inPractice":"I do not moralize about food, and I do not hand anyone a list of forbidden products. Before I read a panel I ask about the kitchen: who cooks, how many meals in a week come from a box, a bag or a drive-through, what gets drunk across the day, and how fast meals are eaten. Then I read the markers against each other, because one marker is a clue and three are a story. Triglycerides drifting up with HDL drifting down, a fasting insulin that is high for its glucose, a uric acid near the top of its range, an ALT or GGT creeping up, a blood pressure that will not settle and a weight that keeps moving the wrong way each have many causes; together they often point at a diet built from formulated food, especially sweet drinks and refined starch, and ApoB and hs-CRP fill in the rest. The rule I teach is the one the classification's own authors use: read the ingredient list, and if it carries things you would not keep in a kitchen, it is ultra-processed. I do not ask anyone to cook every meal. I ask for more of them, eaten more slowly, with protein and plants on the plate, and for the drinks to change first. I will not tell someone that one additive is poisoning them, and I will not set a family rules it cannot afford in money or time. I will say plainly where I am ahead of the evidence: the cohorts are consistent but observational, the trials lasted weeks, and no trial has shown that eating less ultra-processed food prevents a heart attack or a cancer. I act on the trials, the cohorts and the mechanism together, and I say so."}