Does ultra-processed food cause inflammation? No trial has shown it. People who eat the most of it tend to have higher CRP, a blood marker of inflammation, in 11 of 17 adult analyses, but those are links. What trials did show since 2023 is that ultra-processed diets lead to overeating and weight gain, even with calories matched.
What we thought before
What counts as ultra-processed food? Ultra-processed food (UPF) is a category from the NOVA system, which sorts foods by how much industry has done to them and why. The quick test is the ingredient list: one item you would never keep in a kitchen, or one additive there to make the product more appealing, and it counts.
The trial that built the 2023 picture ran in 2019. A team at the US National Institutes of Health admitted 20 adults to a research ward for two weeks of ultra-processed meals and two weeks of unprocessed ones. The menus were matched for calories on the plate, energy density, protein, fat, carbs, sugar, salt and fiber, and people ate as much or as little as they liked. On the ultra-processed menu they ate 508 more calories a day and gained 0.9 kg. On the unprocessed menu they lost 0.9 kg. Weight change tracked calorie intake closely, with a correlation of 0.8.
So the story was simple. UPF makes you overeat, and the extra calories do the damage. Cohort studies had piled up too: by March 2023 there were 39 meta-analyses linking UPF to one disease or another. But cohort studies watch what people choose to eat. They cannot prove the processing did it.
The closest thing to a mechanism trial was a 2022 study of one additive: 15 g a day of the emulsifier carboxymethylcellulose (CMC) for 11 days. The 7 people on CMC had less diverse gut microbes, and in 2 of them bacteria had pushed into the mucus layer that normally keeps them off the gut wall, a central feature of gut inflammation. A clue about one additive in 16 people. Not a verdict on UPF.
What changed
| Question | What the record said in 2023 | What it says now |
|---|---|---|
| Do people overeat UPF? | Yes, in one 2-week ward trial. 508 extra calories a day, 0.9 kg gained. | Still yes, with calories and macronutrients matched. 813.5 extra calories a day, 1.1 kg more in one week. |
| Is it only the calories? | Weight change tracked calorie intake, correlation 0.8. | Weight and the LDL to HDL cholesterol ratio rose on UPF independent of caloric load. |
| Does "high protein" fix it? | Untested. | No. Protein-enriched UPF did not prevent overeating in a 54-hour chamber trial. |
| Does it matter inside healthy eating guidelines? | Untested. | Both diets lost weight. Minimally processed lost 1.01 percentage points more over 8 weeks. |
| How solid are the disease links? | 39 meta-analyses of cohorts. | 45 pooled analyses, 32 linked. 41 of 45 rated low or very low quality. |
| What about inflammation markers? | One 11-day emulsifier study in 16 people, gut findings only. | 11 of 17 adult analyses: higher CRP with more UPF. 5 of 17: no link. |
In 2024, two teams reviewed the reviews. The BMJ umbrella review gathered 45 pooled analyses covering nearly 10 million people, and higher UPF intake was directly linked to 32 of the 45 outcomes, or 71%. The strongest links, graded "convincing" by the review's own rules, were to death from cardiovascular disease, with a risk ratio of 1.50, and to type 2 diabetes, 1.12 per step up in intake. Then the catch, in the authors' own grading: 22 analyses were rated low quality and 19 very low. That is 41 of 45. Four rated moderate. None rated high.
Not a single study had reported a benefit from ultra-processed food, the 2024 umbrella review authors noted.From a 2024 umbrella review in Clin Nutr
The second umbrella review updated all 39 meta-analyses with 122 newer papers and found links for 25 of 49 outcomes, with a line worth pausing on: not a single study had reported a benefit from UPF.
The trials moved further. In 2025, the UPDATE trial gave 55 adults in England with overweight or obesity two 8-week diets in random order, one minimally processed and one UPF, and both followed the UK Eatwell Guide. Both led to weight loss: 2.06% on the minimally processed diet and 1.05% on UPF, a gap of 1.01 percentage points. A UPF diet that meets national guidelines still came out behind. Not by a landslide.
A 2025 crossover trial in men went after the calorie question directly, testing UPF at matched and excess calorie loads. Switching from unprocessed to ultra-processed food raised body weight and the LDL to HDL cholesterol ratio regardless of calorie load, and a phthalate, a chemical used in plastics, trended up in blood after the UPF diet.
Two smaller trials closed loopholes. At a Tokyo hospital, 9 overweight men on calorie-matched menus ate 813.5 more calories a day on UPF, gained 1.1 kg more in a week, and chewed fewer times per calorie. And in a sealed room that measures the calories you burn, 21 young adults on high-protein UPF for 54 hours still overate.
Does ultra-processed food cause inflammation?
Here the record moved less than the headlines suggest. A 2025 scoping review mapped 24 human studies on UPF and blood inflammation markers; 21 of them measured CRP. In adults, 11 of 17 analyses found higher CRP with more UPF, 5 of 17 found no link, and 1 found a link only in women. IL-6, another inflammation signal, was mostly higher with more UPF. IL-1 beta showed no link at all.
A 2025 analysis of 9,254 people in the US national health survey put a number on it. Compared with people getting 0 to 19% of their calories from UPF, those at 60 to 79% had a risk ratio of 1.11 for raised hs-CRP, after adjusting for smoking and inactivity. About a tenth higher risk at the top, and the middle tiers did not line up in a clean dose curve.
Every one of these is a comparison between people who already eat differently. The review itself calls them associations. A survey cannot adjust for everything that separates a heavy UPF eater from a light one.
What hasn't changed
No feeding trial in our evidence pack reported CRP or IL-6 as a result. The trials measured calories, weight, cholesterol, hormones and chewing. That is the gap. The cohort link to inflammation markers is there, and the trial that would test it is one we did not find.
Which part of "ultra-processed" matters is also still open. Each candidate has a study and none has a verdict: calories, where weight tracked intake in 2019; eating speed, with fewer chews per calorie in Tokyo; additives, with one emulsifier in 16 people for 11 days; chemicals from packaging, with a phthalate trend after the UPF diet; and something in the food itself, since weight and cholesterol moved with calories held equal.
Put the two years together and the question has changed shape. In 2023 it was "is UPF bad for you?" Now four trials say the overeating is real and that some harm survives with calories held equal, while the inflammation link sits exactly where it was: visible in the blood of people who eat a lot of it, untested by a trial. So does cutting UPF reduce inflammation? The honest question today is "which part, and does it reach your CRP?" Still open.
What to do now
Tick what you'll try this week. Your picks stay on this device.
you take medicines for blood sugar or blood pressure, since a change in what you eat can change the dose you need: keep taking them as prescribed and ask before adjusting anything. Also ask first if you are pregnant or breastfeeding, have kidney disease, or have a history of disordered eating.
One tap. Results appear once 50 readers have voted, and the count is always real.
Reader stories
What happened for you?
Studies tell us what works on average. You know what happened in your own body. Share it here, and the next person who reads this page gets something no study can give them.
A reader who shared their story here sent you this article. When you're ready, add yours.
Ask the Research Desk
Heard something about inflammation you want checked? The best questions become articles, credited by first name only if you say yes.
How this article was made
Researched and written from 11 peer-reviewed papers. Our automated check matched every claim against the exact words of its source, confirmed every paper exists in PubMed, and found none retracted. Checked November 3, 2026. Not individually reviewed by a doctor. Photos are illustrative; the people in them are not patients or readers.
Grades run Strong, Moderate, Emerging, Early and Not supported, and the grade decides the words we may use. How we grade evidence · Medical disclaimer
Sources
- Umbrella reviewUltra-processed food exposure and adverse health outcomes: umbrella review of epidemiological meta-analyses.Lane MM, Gamage E, Du S et al. BMJ, 2024.PubMedFull paper
- Umbrella reviewUltra-processed foods and human health: An umbrella review and updated meta-analyses of observational evidence.Dai S, Wellens J, Yang N et al. Clin Nutr, 2024.PubMedFull paper
- Randomized trialUltra-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.PubMedFull paper
- GuidelineUltraprocessed or minimally processed diets following healthy dietary guidelines on weight and cardiometabolic health: a randomized, crossover trial.Dicken SJ, Jassil FC, Brown A et al. Nat Med, 2025.PubMedFull paper
- Randomized trialEffect of ultra-processed food consumption on male reproductive and metabolic health.Preston JM, Iversen J, Hufnagel A et al. Cell Metab, 2025.PubMedFull paper
- Randomized trialShort-term effects of high-protein, lower-carbohydrate ultra-processed foods on human energy balance.Hägele FA, Herpich C, Koop J et al. Nat Metab, 2025.PubMedFull paper
- Randomized trialUltra-processed foods cause weight gain and increased energy intake associated with reduced chewing frequency: A randomized, open-label, crossover study.Hamano S, Sawada M, Aihara M et al. Diabetes Obes Metab, 2024.PubMedFull paper
- ReviewUltra-Processed Food Consumption and Systemic Inflammatory Biomarkers: A Scoping Review.Ciaffi J, Mancarella L, Ripamonti C et al. Nutrients, 2025.PubMedFull paper
- Cross-sectionalUltra-processed foods and increased high sensitivity C-reactive protein.Sajan K, Anthireddy N, Matarazzo A et al. Am J Med, 2025.PubMedFull paper
- Randomized trialRandomized Controlled-Feeding Study of Dietary Emulsifier Carboxymethylcellulose Reveals Detrimental Impacts on the Gut Microbiota and Metabolome.Chassaing B, Compher C, Bonhomme B et al. Gastroenterology, 2022.PubMedFull paper
- StudyUltra-processed foods: what they are and how to identify them.Monteiro CA, Cannon G, Levy RB et al. Public Health Nutr, 2019.PubMedFull paper
