- A long-running cohort study found that cancer survivors who ate the most ultra-processed foods had a higher risk of dying from any cause and from cancer specifically.
- The link remained even after accounting for overall diet quality, suggesting processing itself may play a role beyond nutrients alone.
- Higher inflammation and a raised resting heart rate appeared to partly explain the association.
Ultra-processed foods are now a dominant part of many people’s diets, but “processed” is not a single category.
Some processing is harmless – freezing vegetables, pasteurising milk, baking bread.
Ultra-processed foods, however, are typically industrial formulations made with additives, flavourings, emulsifiers and preservatives, often alongside high levels of added sugar, refined starches, and fats the body may not handle well over time.
A study of long-term cancer survivors has linked higher consumption of ultra-processed foods with a greater risk of death, including deaths from cancer.
The research draws on the Moli-sani Study, a prospective cohort that followed adults in the Molise region of southern Italy over many years.
Within the wider cohort, researchers identified 802 cancer survivors who provided detailed diet information using a validated food frequency questionnaire. Foods were categorised using the NOVA system, which groups items based on how and why they are processed.
The team then calculated how much of each person’s diet came from ultra-processed foods in two ways:
- A weight ratio – the weight of ultra-processed foods and drinks consumed each day divided by the total weight consumed.
- An energy ratio – the calories from ultra-processed foods divided by total daily calories.
Participants were grouped into thirds based on the weight ratio, from lowest to highest.
The researchers adjusted for a wide range of factors that could influence outcomes, including age, smoking, body mass index, physical activity, medical history, cancer type, and overall diet quality measured by a Mediterranean diet score.
Over a median follow-up of 14.6 years, 281 deaths occurred among the 802 cancer survivors.
People in the highest third of ultra-processed food intake (by weight) had a substantially higher risk of death than those in the lowest third – around 48% higher for deaths from any cause and about 57% higher for cancer-specific deaths.
When the researchers used the energy ratio measure, the pattern was similar for cancer deaths, but the link with all-cause deaths was less consistent.
One important detail is that the association with all-cause mortality persisted even after taking overall diet quality into account.
That suggests the observed risk is not explained purely by a diet being “unhealthy” in the conventional nutrient sense, and that the nature of industrial processing may carry additional effects.
To explore possible mechanisms, the team looked at biological markers linked to inflammation, metabolism, and cardiovascular health.
When they adjusted the analysis for inflammatory scores and resting heart rate, the strength of the association between ultra-processed food intake and all-cause mortality dropped by about 37%. That does not prove cause and effect, but it suggests inflammation and cardiovascular strain may be part of the pathway.
The researchers also broke ultra-processed foods into seven broad groups, including sweetened drinks, processed meats, salty snacks, sugary foods, sauces and spreads, and other categories. Some groups showed clearer links with mortality than others, but the authors argued it is difficult to interpret single items in isolation.
Ultra-processed foods tend to cluster together as a dietary pattern, so the overall proportion in the diet may matter more than any one product.
- Cognitve decline associated with some ultra-processed foods
- Eating lots of ultra-processed food linked to higher levels of inflammation marker
- Ultra processed foods linked to more fat gain even with the same calories
As with all observational research, there are limits. Diet was self-reported, eating habits may have changed over time, and the sample size – especially the number of deaths within each subgroup – restricts precision.
There was also survival bias, because diet was assessed years after diagnosis for many participants, and cancer stage at diagnosis was not available.
Even so, the takeaway is practical: reducing ultra-processed foods overall is likely more meaningful than obsessing over one “bad” item.
For many people, a useful rule of thumb is to rely more on minimally processed ingredients and home-cooked meals where possible, and to scan labels for long ingredient lists and additives.
For people living with diabetes, that shift often brings an added benefit: fewer highly refined carbs, less hidden sugar, and a diet that is easier to manage for blood glucose and weight.
