Ultra-Processed Food Is Linked to Earlier Death and Diabetes. A Controlled Trial Shows It Can Also Make People Eat More
Evidence review & editorial synthesis
Ultra-processed food has become one of nutrition science's most controversial categories. It includes obvious products such as soft drinks, packaged candy and instant meals, but can also include mass-produced bread, breakfast cereal, flavored yogurt and protein products that many people would not immediately consider junk food.
A new study of 6,563 Korean adults adds another clue. People who ate the most ultra-processed food were not just more likely to report one questionable eating habit. They were much more likely to accumulate several at once, including breakfast skipping, overeating and nighttime snacking.
That raises a bigger question: is ultra-processed food simply a marker of an unhealthy lifestyle, or can properties of the food itself change how much we eat and ultimately affect long-term health?
What actually counts as ultra-processed food?
The most widely used research definition comes from the NOVA classification.
NOVA does not classify food primarily by calories, sugar or fat. It asks how extensively the food has been industrially transformed and why.
Fresh fruit, vegetables, eggs, milk, rice, oats and fresh meat are unprocessed or minimally processed. Butter, oil, sugar and salt are considered processed culinary ingredients. Simple foods made by combining these, such as traditional bread, cheese or canned vegetables, can be categorized as processed rather than ultra-processed.
Ultra-processed foods occupy NOVA's fourth group. They are typically industrial formulations containing ingredients rarely used in a home kitchen, such as modified starches, protein isolates, hydrogenated or interesterified oils, maltodextrin or high-fructose corn syrup. They frequently contain cosmetic additives such as flavors, colors, sweeteners, emulsifiers and texture modifiers.
Typical examples include many soft drinks, packaged sweet and savory snacks, candy, mass-produced baked goods, instant noodles, ready meals, chicken nuggets, processed meat products and many sweetened cereals and bars.
But this is where the category becomes controversial. Two foods with similar calories and nutrients can receive different NOVA classifications because one contains an industrial additive. And some nutritious products, including certain whole-grain breads or fortified cereals, can still qualify as ultra-processed.
High UPF intake travels with a different pattern of eating
The new Korean study examined adults aged 43 to 74 in the Korean Genome and Epidemiology Study.
Compared with people in the lowest quarter of UPF intake, those in the highest quarter had 3.42 times the odds of frequently skipping breakfast, 2.22 times the odds of frequent nighttime snacking and 33% higher odds of frequent overeating.
Fast eating was the exception. UPF intake showed essentially no association with that behavior.
The accumulation was particularly striking. Participants with the highest UPF consumption had 3.24 times the odds of reporting three or more dietary risk behaviors rather than none.
That does not tell us which came first. Someone who regularly skips meals or eats late at night may reach for convenient packaged food more often. Conversely, the properties of those foods could make irregular or excessive eating easier. Work schedules, income, sleep, stress and other lifestyle factors could influence both.
Still, the finding fits a broader idea: UPF consumption may be part of an entire eating environment rather than an isolated nutrient exposure.
The health associations now extend across millions of people
A 2024 BMJ umbrella review assembled 45 meta-analyses involving almost 9.9 million participants.
Greater UPF exposure was associated with adverse outcomes in 32 of the 45 analyses examined. The strongest signals involved cardiovascular mortality, type 2 diabetes, obesity, depressive outcomes, common mental disorders and all-cause mortality.
The magnitude varied considerably by outcome and the certainty was often limited because much of the evidence came from observational research.
Mortality has since been examined in an updated meta-analysis of 18 prospective studies involving more than 1.14 million people and 173,000 deaths. People consuming the most UPF had a 15% higher risk of death from any cause than those consuming the least. Each additional 10% of the diet coming from UPF was associated with roughly 10% higher mortality risk.
Those numbers should not be translated directly into years of life lost. They show population-level relative risk, not an individual countdown clock.
But they make UPF difficult to dismiss as merely another temporary nutrition trend.
A controlled experiment found that people simply ate more
Observational studies always leave a major question: perhaps people who consume more UPF also exercise less, smoke more, sleep differently or have lower overall diet quality.
A small but unusually rigorous NIH experiment removed some of that uncertainty.
Twenty adults lived inside a research center and received either an ultra-processed or minimally processed diet for two weeks before switching to the other diet. The meals were designed to be broadly matched for presented calories, macronutrients, sugar, sodium and fiber, and participants could eat as much or as little as they wanted.
On the ultra-processed diet, they spontaneously consumed about 508 additional calories per day.
They also gained roughly 0.9 kilograms during the ultra-processed phase and lost roughly the same amount during the unprocessed phase.
That trial was too short and too small to tell us whether UPF causes diabetes or cardiovascular disease. But it established something much more immediate: under controlled conditions, an ultra-processed dietary pattern changed how much people chose to eat and changed body weight within weeks.
Why remains unsettled. Possible explanations include faster eating, softer texture, higher effective energy density, stronger reward properties, altered food structure and combinations of refined carbohydrate and fat that make it easier to consume large amounts before satiety catches up.
But not every UPF appears equally harmful
This is one of the most important nuances in the entire debate.
NOVA places very different products in the same category. A sugary soft drink, processed sausage, fortified whole-grain breakfast cereal and high-fiber packaged bread can all qualify as ultra-processed under particular formulations.
Their nutritional profiles and health associations are not identical.
Recent reviews suggest that sugar-sweetened beverages and processed meats are among the UPF categories most consistently associated with adverse cardiometabolic outcomes. By contrast, some breads, cereals and dairy products classified as UPF have shown neutral or even favorable associations in observational datasets.
This does not invalidate the UPF concept. It means processing level is probably one dimension of food quality rather than the only one.
For practical decisions, the strongest signal may come from combining questions: Is the product ultra-processed? Is it energy dense? Does it contain a large amount of refined carbohydrate, added fat, sodium or sugar? Does it contain little intact food or fiber? Is it easy to eat rapidly and in large quantities? And is it replacing foods such as vegetables, legumes, fruit, whole grains, nuts or minimally processed protein sources?
The real question is becoming more interesting than 'junk food is bad'
The new Korean study does not prove that UPFs create unhealthy eating habits. Instead, it shows that high consumption sits inside a broader cluster of behaviors, including skipped breakfasts, overeating and nighttime snacking.
The larger evidence base goes further. High UPF consumption is associated with obesity, diabetes, cardiovascular disease, mental-health outcomes and mortality across enormous populations. And one controlled human experiment demonstrated that an ultra-processed diet can make people eat substantially more calories even when they are free to stop whenever they want.
The unresolved scientific question is no longer simply whether many packaged foods contain too much sugar, salt or fat.
It is whether modern industrial food design changes eating itself.
That distinction matters. If processing, texture, food structure and formulation help determine how quickly we eat, how full we feel and how many calories we consume before stopping, then the health effects of food cannot always be understood from the nutrition label alone.
Scientific References & Further Reading
- Anchor Study: “Higher Ultra-Processed Food Intake is Associated with Co-occurrence and Accumulation of Multiple Dietary Risk Behaviors in Korean Adults.” (Lee J, Oh H. Journal of Nutrition. 2026. doi:10.1016/j.tjnut.2026.101851. PMID:42790705.)
- Health Outcomes Umbrella Review: “Ultra-processed food exposure and adverse health outcomes: umbrella review of epidemiological meta-analyses.” (Lane MM, Gamage E, Du S, et al. BMJ. 2024;384:e077310. doi:10.1136/bmj-2023-077310.)
- Randomized Feeding Trial: “Ultra-Processed Diets Cause Excess Calorie Intake and Weight Gain: An Inpatient Randomized Controlled Trial of Ad Libitum Food Intake.” (Hall KD, Ayuketah A, Brychta R, et al. Cell Metab. 2019;30:67-77.e3. doi:10.1016/j.cmet.2019.05.008.)
- Mortality Meta-Analysis: “Ultra-processed foods and risk of all-cause mortality: an updated systematic review and dose-response meta-analysis of prospective cohort studies.” (Liang S, Zhou Y, Zhang Q, Yu S, Wu S. Syst Rev. 2025;14:53. doi:10.1186/s13643-025-02800-8.)
- UPF Definition: “Ultra-processed foods: what they are and how to identify them.” (Monteiro CA, Cannon G, Levy RB, et al. Public Health Nutr. 2019;22:936-941. doi:10.1017/S1368980018003762.)
- NOVA Classification Debate: “Are all ultra-processed foods bad? A critical review of the NOVA classification system.” (Louie JCY. Proceedings of the Nutrition Society. 2025. doi:10.1017/S0029665125100645.)
Disclaimer: This article is for informational purposes only and does not constitute medical advice. It is not a substitute for diagnosis, treatment, or individualized medical care.

