Four ultra-processed ingredients tracked high blood pressure

Ultra-processed food is a category, and a loose one. It holds supermarket bread and energy drinks, packaged soup and candy, sorted by how they were made instead of what they contain.

Cardiologists have accepted the category anyway, because the association keeps appearing. People who eat more ultra-processed food (UPF) have a higher risk of cardiovascular disease and death.

A team working with UK Biobank data has now taken the category apart. Writing in the European Heart Journal, they describe an exploratory prospective cohort study of 101,560 non-hypertensive UK Biobank participants, followed for a decade, and they ask a narrower question: which of the industrial ingredients that define ultra-processing actually track with disease?

Four did. Five did not.

What counts as a marker of ultra-processing

A marker of ultra-processing, or MUP, is a single industrial ingredient of the sort that only appears in formulated products. Ultra-processed foods are made from industrial ingredients and additives: colorings, sweeteners, flavorings, emulsifiers, thickeners. Their presence is what sorts a product into the category, more than its sugar or salt content does.

The team took that literally. Ingredient lists were examined for the presence of 57 pre-defined MUPs, grouped into nine MUP categories.

Each category was then tested on its own against new cases of high blood pressure, one at a time.

Which food markers tracked new high blood pressure

Over a median follow-up period of 10.5 years, 7,633 hypertension cases appeared among people whose blood pressure was normal when they joined.

Positive associations with hypertension risk were found for foods containing four of the nine categories: flavorings, colorings, sweeteners and varieties of sugar. Colorings carried the largest association, at roughly 55% higher risk at the top of the intake range compared with none. Sweeteners followed at about 42% higher, flavorings at 37%, and sugars at 19%.

The absent half is the more informative one. Intake of foods containing flavor enhancer, processing aid, modified oil, protein source, and added fiber showed no significant associations. Five categories, nothing.

Among individual ingredients, the sweeteners came out as a block. Acesulfame, aspartame, erythritol, saccharin, sucralose, xylitol were each positively associated with higher risk, and sweeteners showed a significant increase in hypertension risk measured from the point of lowest risk, which sat at zero intake.

How the food ingredients were worked out

This is where the study needs reading carefully, because nobody photographed anyone’s shopping.

Diet came from the Oxford WebQ, an online questionnaire recording everything eaten in the previous 24 hours, completed at least once. The team then matched up to 10 commercial products to each food item and read those ingredient lists. A Marker Likelihood Index (MLI) was calculated from these commercial products, and that index reflected the proportion of commercial products in which a specific MUP was found.

So a participant’s exposure to colorings is an estimate of how likely their food was to contain colorings, based on what comparable products on the shelf contain. It is a sensible way to get at an unmeasurable thing. It is not a measurement.

The models adjusted for age, sex, ethnic background, smoking status, alcohol and physical activity, plus salt intake derived from nutrient information from commercial food products. Salt is the important one, since it is the dietary factor most firmly linked to blood pressure and it travels with processed food.

Why splitting ultra-processed food into markers matters

Most work on this topic treats ultra-processed food as a single exposure, which makes the advice that follows blunt: eat less of it.

That advice is not wrong. The European Society of Cardiology issued a consensus statement this year reporting that adults with the highest UPF consumption face a 19% higher heart disease risk, and its authors ask that doctors should discuss UPF consumption with patients and recommend limiting intake.

What a category cannot do is tell you where the risk sits. If flavor enhancers and modified oils carry nothing while colorings and sweeteners carry the association, the mechanism is unlikely to be processing itself. The authors put their four categories forward as prime candidates for UPF reduction strategies and mechanistic studies, which asks for experiments before anybody names a cause.

What a decade of food questionnaires cannot show

This is an observational cohort, so it can only report that a pattern travels with a diagnosis. It cannot show that any ingredient raised anyone’s blood pressure.

The exposure is inferred twice over: once from a single day of recalled eating, and again from the ingredient lists of products the participant may never have bought. Errors in that chain would not favor one category over another, but they widen the uncertainty around all of them.

UK Biobank volunteers are also healthier, wealthier and more often white than the British population, which limits how far the numbers travel. The authors label their own work an exploratory prospective cohort study, and that is the right size of claim for a paper testing 57 things at once.

What to do about ultra-processed food at the supermarket

Nothing here overturns the standing advice, which is to choose whole or minimally processed foods instead and to be cautious of foods marketed as healthier that may still be highly processed.

The sweetener result deserves a sentence of its own, because sugar-free products are usually bought as the safer option. Here the artificial sweeteners tracked high blood pressure as clearly as the sugars did. That is one cohort, and it is enough to say the swap has not been shown to help.

Dr Marialaura Bonaccio, one of the authors of that consensus statement, puts the state of the field plainly: “The associations between UPF and heart disease are consistent and biologically plausible.”

Consistent and plausible is not the same as established, and this study narrows the question instead of answering it.