select search filters
briefings
roundups & rapid reactions
before the headlines
Fiona fox's blog
A study published in the BMJ looks at association between ultra-processed foods and mortality.
Dr Aisling Daly, Lecturer in Nutrition, Oxford Brookes University, said:
“Research on UPFs is very important but can often be difficult to communicate to the wider public, largely driven by the challenges with the definitions of UPFs.
So many foods widely found in our diets are processed and even ultra-processed, but that does not mean they are all “unhealthy”.
“Many people rely on UPFs as they are affordable sources of nutrition, and UPFs are also less risky in terms of food safety and have longer shelf lives.
“Equally, the process of fortifying or adding nutrients like vitamins and minerals to a food, can often push it into the UPF category, and you will rarely find someone claiming nutrient-enriched foods are “unhealthy”.
“With the move towards more sustainable nutrition, many of the meat-alternatives fall into the UPF category, causing further confusion or conflict in the minds of consumers.
“Demonising all UPFs can potentially cause almost as much harm as the foods themselves, if people fear them and cut them out but then can’t replace them with suitable foods.
“As this research rightly notes, the overall quality of the diet is more important than the consumption of UPFs themselves.
“While the main research outcomes relate to all-cause mortality (i.e. death), this can be confusing or difficult to interpret and act on, especially for people who need to rely on UPFs for their food. Ultimately, not all UPFs are “bad” or “unhealthy”, and if we only link UPFs with mortality, it may be interpreted as meaning ‘any UPFs will kill you’.
“We do want to be wary of over-consuming UPFs that are high in saturated fat, sugar and salt, because of their known negative impacts on our health, and at the same time many UPFs are often lower in fibre and protein, therefore people may lose the benefits of these nutrients. Therefore, it is not only the fact that a food is classified as a UPF, but rather the components and quality of the UPF are important, as noted by the authors when they found that diet quality played an interacting or mediating role in their analysis.
“Overall, the research is useful and helpful being from two large, longitudinal populations in the UK and USA, and it highlights the challenges and confusing with demonising all UPFs, however we ultimately need a more useful system for classifying “healthy” and “unhealthy” UPFs, to help the public make health-promoting dietary choices to improve their overall diet quality and subsequent health risks.”
Prof Christine Williams, Emeritus Professor Human Nutrition, University of Reading, said:
“This new UPF study comes from the Harvard group – the leading group worldwide in the area of nutritional epidemiology. This large study includes men and women followed up for cause of death for over 34 years with baseline information dating back to 1984 and 1986. They examined the relationship between UPF consumption levels and risk of mortality (all diseases) as well as mortality from specific causes (cancer, cardiovascular, respiratory, neurodegenerative and other causes).
Findings
“The study showed a modest association with high UPF consumption on the outcome category ‘All deaths’ which were 4% higher in the high UPF group, representing an additional 64 deaths per 100,000 person-years for high UPF consumers. They found no relationship between total UPF consumption levels and deaths from cancers and cardiovascular disease. This is in contrast to most recent research findings for UPF consumption which have shown higher rates of colorectal cancer, cardiovascular diseases and type 2 diabetes in high UPF consumers.
Specific foods
“When the researchers examined the effects of different types of UPF foods separately they found higher rates of mortality for high consumers of ‘meat/poultry and seafood ready-eat dishes’ (processed meats), ‘sugar-sweetened and artificially sweetened beverages’ and dairy desserts. These categories of foods have long been advised to be eaten in modest quantities due to potentially adverse effects on body weight, cholesterol levels, and in the case of processed meats, colorectal cancers.
Methods used that differ from other studies
“They showed that the diet quality score they used (AHEI – based on known effects of foods and nutrients on health) showed a more consistent relationship with overall mortality than did the UPF score. They considered the effects of processing were likely to add little to the categorisation of foods according to their health effects. They quantified the levels of smoking in ‘pack years’ and noted when they adjusted for this measurement the association between UPF consumption and total mortality was attenuated towards zero. May it be that other studies have not sufficiently accounted for (adjusted) smoking levels in their studies? They also discuss the limitations in assessing UPFs intake using Food frequency questionnaires, which were not designed to assess UPFs intake but are the best method to use for large population studies. They considered 24-hour recall measurements to be a weaker approach as this cannot adequately assess habitual intakes.
“Overall the approaches used in the present study are more nuanced than some of the previous studies of UPF. They discuss the well-recognised benefits of whole-grain breakfast cereals and why in their study they excluded this category of breakfast cereals from the level 4 UPF grading. They also excluded alcohol as an UPF as the adverse effects of alcohol on health are well defined and known to be due to the adverse metabolic effects of high alcohol consumption, not related to the processing involved in alcohol production.
“They consider the different approaches they have taken may explain the differences in their findings from others.
“They state that ‘we aim to rectify the misperception that all ultra-processed foods products should be universally restricted’ they also caution health authorities ‘to avoid over-simplification when formulating dietary recommendations’.”
Prof Sir David Spiegelhalter, Emeritus Professor of Statistics, University of Cambridge, said:
“This study shows weak associations of ultra-processed foods with overall mortality; with only 4% higher death rate in the group consuming the most UPFs compared with those consuming least.
“Furthermore, the authors acknowledge that “dietary quality has a predominant influence on long term health, whereas the additional effect of food processing is likely to be limited”, so it is surprising their conclusions focus on processing rather than quality.”
Dr Duane Mellor, Dietitian and Spokesperson for British Dietetic Association, said:
“This is an interesting study which appears to suggest that the risk of an early death that could be associated with the consumption of ultra-processed foods seemed not to be the case in those who consumed an overall healthy diet. It is also noticeable that those who consumed most ultra-processed foods tended to eat few vegetables, fruit, legumes and wholegrain. This appeared to suggest that it might not be as simple as that those who ate more ultra-processed foods were more likely to die earlier – it is quite possible that these foods might displace healthier foods from the diet.
“This study additionally did not include wholemeal ultra-processed foods along with other ultra processed foods, which is different to other analyses of the risk associated with ultra-processed food intake with negative health outcomes. The authors suggested that was because wholegrain foods, even if they were classified as being ultra processed, have previously been shown to be associated with improved health outcomes. This – together with the suggestion in the analysis of this study, that agrees with previous studies – suggests that not all groups of ultra-processed foods are associated with the same health risks, with sugar and artificially sweetened drinks and processed meats being most clearly associated with risk of an early death.
“These types of study are limited as they only asked participants about a limited range of foods and then tried to classify whether they were ultra-processed or not. Also it appears to suggest that the way ultra-processed foods were classified means that it tends to group foods that are known to be less healthy e.g. sweetened drinks and processed meat, with foods which may be associated with better health such as processed wholegrain. Therefore, this paper potentially challenges the view that the simple NOVA classification of ultra-processed foods into four groups (minimally/ unprocessed foods, processed ingredients, processed foods and ultra processed foods) is too simplistic and does not fully explain the health effects associated with foods.”
Prof Janet Cade, Professor of Nutritional Epidemiology and Public Health, University of Leeds, said:
“This analysis of two large, long-term follow up studies from the USA has carefully explored the potential of foods classified as ultra-processed alongside a diet quality score on risk of death from different causes over a 30 year period. Whilst there may be some aspects of some processed foods linked to mortality, the NOVA classification used here was not able to tease this out. The diet quality score was more influential on mortality outcomes. This work backs up a nutrition focus for food classification, showing clearly that diet quality has a stronger role than ultra-processed foods in relation to risk of mortality.
“The authors state ‘The associations between ultra-processed food consumption and mortality were attenuated after we accounted for overall dietary quality’. Providing further evidence that the impact of diet on these mortality outcomes is likely to be driven by a food’s nutritional composition rather than processing per se. Indeed, dark chocolate in the subgroup “packaged sweet snacks and desserts” was actually linked with decreased mortality.
“Classification of foods using NOVA from the relatively short questionnaires was done rigorously, but with acknowledged limitations. It should be noted that this study, like all long-term studies, was not designed to look at food processing. Meaning that there is a strong likelihood that some foods were misclassified. In addition, use of mortality as an end point is complicated and may not reflect factors influencing incidence.
“Overall, these results should encourage clearer recommendations using our knowledge of nutritional composition and diet quality as the driver of effect to limit intakes of specific types of ultra-processed foods.”
Dr Gavin Stewart, Reader in Interdisciplinary Evidence, Newcastle University; and Campbell Collaboration Food security group editor and co-chair, said:
“The conclusion that higher consumption of ultra-processed foods is linked to a slightly higher risk of death is robust. The authors appropriately consider the uncertainties associated with small effect sizes, acknowledge the potential for confounding, and the complexities of ascertaining cause and effect when studying diet. Importantly, the evidence is also consistent with independent evidence synthesis. Pre-registration and consideration of effect magnitudes and precision rather than statistical significance would reduce any potential for spurious false positive results and for incomplete reporting, but the overall findings are appropriately nuanced and can be considered evidence-based.”
Prof Gunter Kuhnle, Professor of Nutrition and Food Science, University of Reading, said:
“This is another study that attempts to investigate the association between ultra-processed foods (UPF) and health. Despite the large sample size, a key limitation of the study is the measurement of UPF and health: the questionnaires used were never designed to determine ultra-processed food intake and it is therefore impossible to know how reliable results are – results of this study therefore should be treated with a lot of caution.
“The results of the study are interesting and show that high UPF intake is not always associated with higher disease risk: indeed, for several food groups and diseases, the authors find that those with a very high intake of UPF have a much lower risk than those with a medium intake. For example for packaged sweet, the group with the highest intake has a lower risk of cancer mortality than medium groups; and for all cause mortality, ready to eat and packaged savoury snacks are even associated with a reduced risk.
“Overall however, the study shows inconsistent and very small associations between UPF intake and disease risk. The authors also find that diet quality has an important effect on the outcome – and it is therefore likely, that it is diet quality, not processing, that drives associations between UPF intake and.
“In the press release the authors say the findings “provide support for limiting consumption of certain types of ultra-processed food for long term health”. We know from previous studies and this one that there is plenty of evidence for limiting foods high in saturated fat, salt and sugar – which some UPS tend to be high in. But I don’t think that this study provides evidence suggesting limiting certain foods just because of their level of processing, as there was no meaningful association with health.
“Public health policy should be informed by evidence, and there is very good evidence about the health effects of foods based on their composition – which is largely confirmed by this study. In contrast, there is still virtually no robust evidence for an effect of “ultra-processing” specifically on health.”
‘Association of ultra-processed food consumption with all cause and cause specific mortality: population based cohort study’ by Zhe Fang et al. was published in the BMJ at 23:30 UK time on Wednesday 8 May 2024.
DOI: 10.1136/bmj-2023-078476
Declared interests
Dr Aisling Daly: “I have no conflicts of interest to declare. I am a registered nutritionist with the Association for Nutrition and I work as a lecturer in nutrition at Oxford Brookes University.”
Prof Christine Williams: “Chair of the Nominations Committee, British Nutrition Foundation.”
Prof Sir David Spiegelhalter: “No conflicts.”
Dr Duane Mellor: “I have discussed and been consulted about the concept of UPF with EUFIC, APPGs and members of the food industry.”
Prof Janet Cade: “I am an employee of the University of Leeds.
I am Director and shareholder of Dietary Assessment Ltd., a spinout company from the University of Leeds, which has produced a software tool for dietary assessment called myfood24.
I chair the Advisory Committee of the British Nutrition Foundation: https://www.nutrition.org.uk/our-work/who-we-are/committees/advisory-committee/ I receive no financial remuneration for doing that role.
(BNF weblink: https://www.nutrition.org.uk/our-work/who-we-work-with/).
I lead a WHO Collaborating Centre in Nutritional Epidemiology.
I am Honorary Secretary of the Association for Nutrition.”
Dr Gavin Stewart: “No conflict of interest to declare.”
Prof Gunter Kuhnle: “COI: https://kuhnle.co.uk/coi.html
Employment
Research funding
Committee membership
Scientific organisations
Others