Home/Issues/Issue 06/Study

Exploratory prospective cohort study Publication date verified at source

Markers of ultra-processed food and incident arterial hypertension in the UK Biobank

AI narration, generated on first listen
Journal
European Heart Journal
Authors
Krost KM
Institution
Justus-Liebig University of Giessen
Published
3 August 2026
Source
PMID 42545878 · DOI 10.1093/eurheartj/ehag614
Design
Exploratory prospective cohort study with Cox proportional hazards regression and penalised cubic splines, exposure derived by matching 24 hour dietary recalls to commercial product ingredient lists.
Sample
UK Biobank participants without hypertension at baseline

What markers of ultra-processing is

Specific additives and ingredients, such as flavourings, colouring agents and sweeteners, that appear on the ingredient lists of ultra-processed products. They are used here to ask which components of the category might carry a risk, rather than blaming the whole category at once.

Drawn from the paper's introduction.

Why they ran it

The authors note that intake of ultra-processed food is positively associated with arterial hypertension, but that the association with specific markers of ultra-processing had not been assessed before.

Their stated reason for decomposing the category is that the mechanisms linking ultra-processed food to hypertension remain unclear, and that specific ingredients are where a mechanism could be found.

Drawn from the paper's introduction.

The team took 101,560 UK Biobank participants who did not have hypertension at baseline and had completed at least one 24 hour dietary recall. Each recalled food item was matched to up to ten commercial products, whose ingredient lists yielded 37 specific markers of ultra-processing grouped into nine categories. Cox models with penalised cubic splines then related intake of foods containing each category, as a percentage of total food intake, to incident hypertension.

Over a median 10.5 years there were 7,633 new cases. Four categories were positively associated: flavour, colouring agent, sweetener and varieties of sugar. Colouring agents carried the strongest association, roughly half again the risk at the top of the range compared with the nadir. Among specific markers, acesulfame, aspartame, erythritol, saccharin, sucralose, xylitol, bulking agent, dextrose, fructose, hydrolysed protein and whey protein were each positively associated.

Five categories showed no significant association at all: flavour enhancer, processing aid, modified oil, protein source and added fibre. Firming, foaming and glazing agents were negatively associated. The authors describe the analysis as exploratory and name the four flagged categories as candidates for mechanistic study and for reformulation, rather than as established causes.

The numbers

undefinedundefined
undefinedundefined
undefinedundefined
undefinedundefined
undefinedundefined
undefinedundefined

Why this might happen

Open question The authors do not claim to know why.

The authors state plainly that the mechanisms linking ultra-processed food to hypertension remain unclear. They do not offer an explanation for why flavours, colouring agents, sweeteners and sugars should carry the association where other categories do not.

What they do instead is name those four categories as prime candidates for mechanistic studies and for reduction strategies, which is a proposal for the next study rather than an account of this one.

Drawn from the paper's introduction and conclusion.

What this does not show

  • No additive is shown to cause hypertension here. This is an observational cohort and the authors label it exploratory. People who eat more of these foods differ from those who eat fewer in many ways, and dietary pattern is not randomly assigned.
  • Additive intake was never measured directly. Exposure was inferred by matching each reported food to up to ten commercial products and reading their ingredient lists. Whether any individual actually consumed a given additive is an inference, not an observation.
  • It does not show that removing these additives would lower blood pressure. An association between foods containing a marker and later hypertension does not establish what happens if the marker is taken out, particularly when the marker travels with everything else in the product.
  • The inverse associations should not be read as protective. Firming, foaming and glazing agents ran negatively. In an exploratory analysis of 37 markers, associations in both directions are what you would expect, and the authors do not interpret these as benefits.

Where this leaves us

It is now more likely that the ultra-processed category is doing coarse work, and that only some of its components track with hypertension. That refines a blanket association into a shorter list, without yet showing that the shorter list is causal.

UK Biobank volunteers without hypertension at baseline who completed at least one 24 hour dietary recall, with additive exposure inferred from matched commercial products.

Mechanistic studies of the four flagged categories, which the authors name as the needed next step, and exposure measurement that does not rely on matching foods to products.

Caveats worth holding

  • The authors describe the analysis as exploratory, which is the right frame for 37 markers tested against one outcome.
  • Exposure is inferred from matched commercial products rather than measured, so misclassification is likely and probably not random.
  • A single baseline dietary assessment is used to represent a decade of eating.
  • UK Biobank volunteers are healthier and less deprived than the general population.

Newsletter

Each issue by email, when the newsletter launches. Leaving your address puts you on the list, nothing is sent yet.