select search filters
roundups & rapid reactions
before the headlines
Fiona fox's blog

expert reaction to two studies looking at ultra-processed foods and heart disease, bowel cancer and death

Two studies published in the BMJ look at ultra-processed food consumption and risk of colorectal cancer, cardiovascular disease, and mortality.


Prof Gunter Kuhnle, Professor of Nutrition and Food Science, University of Reading, said:

“If these studies and the editorial are interpreted as ‘ultra-processed foods cause these diseases’ then that would be misleading – a more correct summary would be that people who consume food classed as ultra-processed are at higher risk of disease.  The studies do not show whether there is a causal relationship or not – and as with all other observational research, “correlation does not imply causation” does apply here as in other studies.

“The definition of ultra-processed food is fairly vague and ambiguous.  There are different interpretations of the definition: some refer to specific processes, the type or number of ingredients or even the intent behind the processing.  For example, most supermarket breads are considered “ultra-processed”, even though many are not fundamentally different from other breads.

“This ambiguous definition makes it difficult to assess intake, especially with limited information on diet such as in those two studies: sausages for example were classed as “ultra-processed”, even though many sausages are not.  Both studies show that participants with a certain dietary pattern have a high risk of diseases – these dietary patterns include sweets and carbonated drinks, which are both independently associated with increase disease risk.  There are no data on the socio-economic status of participants, but foods classed as ultra-processed have often longer shelf-lives and are therefore cheaper.  This might also contribute to the observed associations.

“The term “ultra-processed foods” has become very popular despite very limited data on underlying mechanisms beyond the “hyper-palatability” of some of these foods.  Many supporters of the term call for a ban of these foods and often even compare it with smoking – but ignore the problems that this causes.  Ultra-processed foods can often make much better use of resources (e.g. fish fingers) or have a much longer shelf-life than other foods and are therefore cheaper.  Banning ultra-processed foods would increase the cost of foods and many people rely on ultra-processed foods.

“In many ways, this focus on the “NOVA” classification prevents the development of healthier foods: it incentivises the reformulation of foods to meet the arbitrary criteria of “processed” foods instead of creating healthier foods – e.g. by reducing salt, saturated fat or sugars.  It is therefore important to understand what causes the observed association before calling for or implementing radical measures.”


Dr Duane Mellor, Registered Dietitian and Senior Teaching Fellow, Aston Medical School, Aston University, said:

“There is a lot of interest in the concept of ultra-processed foods, however it is not clear if looking at ‘ultra-processed food’ as a definition is any better than looking at established dietary factors associated with chronic disease and risk of early death including higher intakes of added fat, salt and sugar and lower intakes of vegetables, fruit, wholegrains and legumes.

“In one study from the US three historic studies were combined to look at ultra-processed food and the risk of colorectal cancer.  This involved looking at food questionnaires collected over 30 years ago with researchers recently classifying which foods were ultra-processed – this risks an interpretative bias as the foods available currently are different to the food eaten 30 years ago.  This study suggested a diet high in some ultra-processed foods were associated with colorectal cancer in men but not women – these foods were sugar sweetened beverages and processed meat products.  This is not particularly surprising as these foods have been associated with increased risk of this type of cancer previously (using this data).  This could suggest that in this large cohort a global definition of ultra-processed food did not add any extra information beyond what was already known.  Interestingly, women in the study had a slightly increased risk of colorectal cancer when they ate more ready to eat meals – and a slightly reduced risk when they consumed yoghurt.  However, this data is based on one questionnaire completed more than two decades before the risk of colorectal cancer was assessed, and people may have changed their diets over this time.

“In the study from Italy, the authors used a method based on Nutri-Score, which was designed to see if a food can be advertised on TV in the UK before 9pm (FSA-NPS dietary index), which is a form of nutrient profiling, and compared this with how processed the food was again using the NOVA system.  This study looked at nearly 23,000 people and followed them up for about 14 years.  Both the modified labelling measure and the amount of processed food was linked to a higher risk of death during the follow up period.  The authors try to say that the food labelling method did not explain all the excess mortality seen with eating ultra-processed food and that meant it was better than just looking at food quality.  It is important to note that although this approach has been used to look at diets of populations before, it may not be the best method as it only looks at nutrients in a positive or negative way to give a score, as it was originally defined to help reformulate individual food products to make them more healthy.  It perhaps would be better to look at a holistic dietary pattern using scores or measurements which are more applicable to dietary patterns?

“Although the idea of ultra processed food is interesting, these studies only provide weak evidence to support the use of the definition, it still is unclear if it is any better than trying to limit foods with added fat, salt and sugar whilst trying to eat more vegetables, fruit and fibre rich foods such as wholegrains, lentils, beans and peas.  The risk with the ‘ultra processed foods’ definition is that it would suggest a shop bought sponge cake would be ultra-processed, but one made at home is not, even though the average consumer is aware that both contain sugar and fat and are not ideal as a regular part of most people’s diet.”



Paper 1: ‘Association of ultra-processed food consumption with colorectal cancer risk among men and women: results from three prospective US cohort studies’ by Lu Wang et al. was published in the BMJ at 23:30 UK time on Wednesday 31 August 2022.

DOI: 10.1136/bmj-2021-068921

Paper 2: ‘Joint association of food nutritional profile by Nutri-Score front of-pack label and ultra-processed food intake with mortality: Moli-sani prospective cohort study’ by Marialaura Bonacci et al. was published in the BMJ at 23:30 UK time on Wednesday 31 August 2022.

DOI: 10.1136/bmj-2022-070688



Declared interests

Prof Gunter Kuhnle: “No COI.”

Dr Duane Mellor: “I have not conflicts of interest to declare for this.”


in this section

filter RoundUps by year

search by tag