select search filters
briefings
roundups & rapid reactions
Fiona fox's blog
A study presented at the International Congress on Obesity (ICO 2026) and published in the American Journal of Preventive Medicine estimates deaths that could be prevented by reducing ultra-processed food (UPF) consumption.
Dr Adam Jacobs, Executive Director and Strategic Consultant, Biostatistics, Ergomed, said:
“Hamel et al have tried to estimate how much cardiovascular disease (CVD) in Canada is attributable to ultra-processed food (UPF) consumption, using a modelling study in which they used as their inputs current levels of UPF consumption in Canada, current levels of CVD in Canada, and associations between UPF consumption and CVD found in other studies (not in Canada, but in France, Italy, and USA). They estimate that between 23% and 38% of all CVD cases in Canada could be avoided if there were no UPF consumption at all, corresponding to 96,000 out of 258,550 CVD cases and 17,400 out of 46,153 CVD-related deaths annually.
“A key input to this model is the estimate of the relative risk of CVD associated with high UPF consumption relative to low UPF consumption. They use as their estimate a relative risk of 1.29 (in other words, a 29% increase in CVD risk) for high UPF consumption relative to low UPF consumption from a paper by Pagliai et al published in 2020. That paper reports a meta-analysis of existing studies, and included 3 studies in their estimate of the relative risk for CVD.
“An important assumption of Hamel et al’s analysis is that the relative risk represents a causal relationship. If it is not causal, then reducing UPF intake would not be expected to reduce CVD cases, at least not to the extent that Hamel et all estimate. It seems unlikely to me that Pagliai et al’s estimate of a relative risk of 1.29 represents an entirely causal relationship.
“An important confounding factor is socioeconomic status. Poor people generally eat more UPF than rich people, and poor people have worse CVD outcomes. So it is possible that the observed association between UPF and CVD risk may simply be a proxy measure for socioeconomic status rather than a direct causal effect of UPF consumption.
“Pagliai et al’s estimate included estimates from 3 studies. Only one of those adjusted for income level (Kim et al 2019), and that study found a risk ratio of 1.13: substantially lower than the estimate from the other studies included in Pagliai et al’s analysis or their combined estimate. Even Kim et al’s adjustment for income level was relatively crude, dividing participants into just 3 income categories, and it is possible that the risk ratio would have been even lower with more precise adjustment.
“It therefore seems likely that the estimate of 1.29 for the relative risk used in Hamel et al’s study is an overestimate of the true causal effect of UPF, and therefore their estimates of the number of CVD cases or deaths that could be avoided by reducing UPF consumption are similarly overestimated. Also, as Hamel et al acknowledge in their paper, reducing UPF consumption to zero is unlikely to be realistic, and they give more realistic figures including a 20% reduction in energy intake from UPF, which they estimate would avert 16,800 new CVD cases and 3,100 deaths annually, but those figures are also likely to be overestimates as they also rely on the assumption that the observed relationship between CVD and UPF is causal.
“While it is entirely possible that a diet rich in UPF is less healthy than a diet rich in fresh fruit and vegetables, I do not believe Hamel et all have given realistic estimates of the extent to which CVD could be avoided by switching dietary patterns away from UPF.”
Prof Kevin McConway, Emeritus Professor of Applied Statistics, Open University, said:
“This is an interesting attempt at modelling potential health effects of UPF consumption in one country (Canada) – but I really wouldn’t want to put much trust in the detailed estimates it presents. That’s why I’m pleased that the wordings in the research paper express some doubts: “Reducing UPF consumption could decrease the burden of cardiovascular disease” and “UPF consumption may account for a substantial and potentially preventable burden of CVD in Canada.” (My emphases.)
“I’m not trying to claim that reducing UPF consumption in Canada would have no effect on cardiovascular disease and death. It might well do so, but any reduction could be substantially smaller than the figures in this paper, or indeed larger. And the measure used in the research, deaths and disease attributable to one specific risk factor (UPF consumption here), can be awkward to interpret. It sounds as if it means ‘deaths caused by UPF consumption’, but it doesn’t mean that, particularly not for a disease like cardiovascular disease (CVD) that has many potential causes that can interact with one another.
“The reasons are that the figures are calculated from statistical models. Like all such models, the researchers had to work with a simplified version of what happens in the real world, and some of the simplifications seem to me to be pretty gross. The paper refers to some of them in its ‘Limitations’ section:
The data on the rates of CVD in people consuming different amounts of UPFs weren’t based on data from Canada, because no such data was available. Instead data from three studies in France, Italy and the USA, was used. Things might (or might not) be different in Canada.
The relative risks of CVD were assumed not to depend on people’s sex or age, which seems very unlikely to me – but better data that provide information by age and sex was not available. I don’t have any such data either, so can’t comment on how big the effect of this simplification could be.
The data on food consumption is from what’s now a rather old survey in Canada (from 2015) and consumption patterns may or may not have changed since then.
“But there are other points along these lines:
The paper reports that the relative risk they used to estimate the changes in CVD risk if the consumption of UPFs changed came from 3 previous cohort studies. But it also mentions (rightly) that the estimate comes from comparing in each of the three studies, the rate of CVD in the group with the highest UPF consumption with the rate in group with the lowest consumption. But the UPF consumption in those groups in the three studies was different, because they split up amounts of UPF consumption in different ways, and indeed did not all measure UPF consumption in the same way as one another. Looking at the overall estimate of the relative risk, despite these differences in consumption, makes sense in the context of the review paper where the researchers in the new study found the information, but that context is different from the estimation in the new paper, and this might add an unknown amount of extra uncertainty.
As far as I can make out, the estimates of reductions in CVD events and deaths, attributable to UPF consumption, have been worked out by assuming that the pattern of actual types of UPFs consumed, if the consumption fell by (for example) 50%, would be unchanged from current levels of consumption. But I’m not sure anybody seriously claims that every type of UPF is equally important for CVD risk, so this is another simplification that might increase uncertainty in the estimates.
“And there are two overall issues about doing this sort of estimation at all:
The calculations depend on relative risks calculated from the three studies in France, Italy and the USA. These studies are all, inevitably, observational, and observational studies always raise questions about what causes what. The studies are also not particularly recent. They were published in 2019, 2021 and 2019 respectively, and for two of them at least, the measures of food consumption go back quite a long way (2005-2010 for the Italian study, and 1988-1994 for the American one; some of the French consumption data is a bit newer). This goes back before the introduction of the notion of ultra-processed foods, so converting what people said they ate to UPF percentages is yet another source of uncertainty. More recently, there have been improvements in collecting this kind of data, but that would generally not apply to the studies used for this new paper, because of their age. I’m not claiming that there’s no evidence that UPFs can harm health, but this is another possibly important source of uncertainty.
What’s presented in the new paper are estimates of CVD events (deaths and illness) attributable to UPF consumption. But ‘attributable’ in this context is a technical term that can be confusing. It sounds as if it means that these events were caused by UPF consumption. But that isn’t what it means. The events might not be, or more likely, not all be caused in that way, because of the remaining doubt over cause and effect in observational studies. More importantly, though, attributable counts of events can work oddly in diseases like CVD that have several causes, or potential causes – amounts of exercise, one’s genetics and ancestry, smoking, and so on, to name but a few. The calculation of attributable events doesn’t directly take these other potential causes into account. So there will be some people who smoke heavily, and also have a high UPF consumption. A study like this, estimating CVD deaths attributable to UPFs, would count deaths of such people as being attributable to UPFs. But a different study, estimating CVD deaths attributable to smoking, would count their deaths as being attributable to smoking. The calculation doesn’t account for multiple causes.”
Prof Tom Sanders, Professor emeritus of Nutrition and Dietetics, King’s College London, said:
“This study has attempted to model the impact of ultraprocessed food on cardiovascular disease in Canada. The authors suggest that between 23 and 38 percent of cardiovascular disease might be attributable to ultraprocessed food. The conclusions regarding dietary intake are based on 2015 national survey. Participants were asked to recall what they ate on the previous day. Estimates of cardiovascular disease were on different participants. This analysis is based on associations not on data from controlled trials.
“It is well established that observational studies can be confounded by association with other life-style factors. Of relevance is the fact that a high intake of ultraprocessed food is associated with low socioeconomic status, which in turn is strongly associated with unhealthy behaviours such as smoking and social deprivation. Although ultraprocessed food accounted for 43% in this study, it probably has not increased over last few decades in Canada. In contrast, death from cardiovascular disease has fallen by about 80 percent over the past twenty years in Canada and the consensus view is that this has been due to improvements in diet as well as well as the better management of cardiovascular disease (blood pressure and cholesterol lowering medication. In my opinion, the claim that halving the intake of ultraprocessed food would have substantial impact on cardiovascular disease is speculative.”
Prof Alberto Fiore, Professor in Food Chemistry and Technology, Abertay University, said:
“This is a modelling study, not a clinical trial — it does not measure what actually happened to people who ate more or fewer ultra-processed foods. It takes a 2015 dietary snapshot, applies a risk multiplier borrowed from studies in France, Italy, and the US, and projects how many CVD events might be attributable to UPF consumption. The authors’ own sensitivity analysis reduces the headline figure of 96,000 avoidable CVD cases by nearly 40% depending on which risk estimate is used — that is a very wide uncertainty range for a number being put in front of the public.
“But the deeper problem is one this study cannot resolve: are we actually measuring the effect of industrial processing, or are we simply measuring the well-known harms of a poor diet that happens to come in a packet? The paper itself tells us the answer. It acknowledges that “ultra-processed dietary patterns” are characterised by excess free sugars, saturated fats, and sodium, and low fibre — and it separately estimates that targeting free sugars and sodium alone could prevent thousands of CVD deaths per year in Canada. If standard nutritional harms already explain the observed risk, then the concept of “ultra-processing” is doing no independent scientific work whatsoever.
“This matters because the NOVA classification — which underpins the entire study — groups an enormous range of foods into a single “harmful” category. When the UPF-CVD associations reported in the underlying cohort studies are broken down by food subtype, the signal is overwhelmingly driven by sugar-sweetened beverages and processed meat products. These are foods whose harmfulness has been established for decades on purely nutritional grounds — high free sugar, high saturated fat, high sodium, low fibre — with no need to invoke the concept of industrial processing at all. A can of cola and a fortified wholegrain breakfast cereal are both NOVA category 4. Treating them as equivalent health risks, as this model implicitly does, is not scientifically defensible, and it means the model is likely attributing to “ultra-processing” risks that belong entirely to specific nutrient profiles.
“The conclusions — that diets high in nutritionally poor foods increase CVD risk — are sound and consistent with decades of evidence. But that is precisely the point: this is not new knowledge dressed up in new clothing. In my view it would be more accurate to interpret this study as a modelling exercise that reinforces what we already know about poor diet and heart disease, not as evidence that industrial food processing is itself a distinct cardiovascular hazard. Those are meaningfully different claims, and this study provides no data to support the stronger one.”
Prof Gunter Kuhnle, Professor of Nutrition and Food Science, University of Reading, said:
“This modelling study is based on the assumption that so-called “ultra-processed food” consumption increases the risk of cardio-vascular diseases. However, there is very little reliable evidence that this is the case. Some clinical trials investigating the impact of UPF did find an increase food intake, although others (https://www.nature.com/articles/s41591-025-03842-0) found that a diet that meets the current UK dietary recommendations did not result in increased food intake and had no adverse effect on health.
“Observational studies can only show associations between UPF intake and health, and have two major limitations:
“1) Estimating individual UPF intake is very difficult, and many studies use methods that do not collect sufficient information to characterise the food consumed properly. For example, bread can be processed or ultra-processed, but the amount of detail required to make such a classification is not part of standard dietary assessment methods. This results in unreliable results.
“2) Observational studies can only provide information about an association, but not causality. UPF are often more affordable than less processed foods, which means that people who consume larger amounts of ultra-processed foods might already be disadvantaged, which can result in poorer health.
“The authors have used data from observational studies to estimate the impact of changes in UPF intake on health in a Canadian population. Apart from the limitations outlined above, this has another important limitations as it assumes that all ultra-processed foods have the same risk profile. However, UPF is a very broad category, and a high consumption of sugar-sweetened beverages is more likely to increase the risk of heart disease than increasing the consumption of bread or fish fingers – indeed, those might even be protective due to their content of fibre and essential fatty acids.
“In my opinion, the estimates are therefore not reliable. Not only because the alleged adverse health effects are based on weak data, but also because the analysis does not take differences in the nutrient profile of UPF into consideration.
“While diet is an important risk factor for diseases, it is important to provide the public with advice that is based on the best evidence available and communicates limitations and uncertainty. Claims that a reduction of UPF intake could have prevented more than 15,000 deaths could create unnecessary panic and concern, and could detract from public health nutrition messages that are based on more robust evidence, such as an increase in fibre intake.”
‘Estimating the preventable burden of cardiovascular disease attributable to ultra-processed dietary patterns in Canada: A modeling study’ by Hamel V et al. will be published in the American Journal of Preventive Medicine at 23:01 UK time on Wednesday 15 July 2026, which is when the embargo will lift.
DOI: 10.1016/j.amepre.2026.108363
Declared interests
Dr Adam Jacobs: “No conflicts to declare.”
Prof Kevin McConway: “I have no conflicts of interest to declare.”
Prof Tom Sanders: “I was Chair of the British Nutrition Foundation Report on Nutrition and Development: Short and Long term consequences for Health.
I have received grant funding for research on vegans in the past. I have been retired for 10 years but during my career at King’s College London, I formerly acted as consultant for companies that made artificial sweeteners and sugar substitutes.
I am a member of the Programme Advisory Committee of the Malaysia Palm Oil Board which involves the review of research projects proposed by the Malaysia government.
I also used to be a member of the Scientific Advisory Committee of the Global Dairy Platform up until 2015.
I did do some consultancy work on GRAS affirmation of high oleic palm oil for Archer Daniel Midland more than ten years ago.
My research group received oils and fats free of charge from Unilever and Archer Daniel Midland for our Food Standards Agency Research.
I was a member of the FAO/WHO Joint Expert Committee that recommended that trans fatty acids be removed from the human food chain.
Member of the Science Committee British Nutrition Foundation. Honorary Nutritional Director HEART UK.
Before my retirement from King’s College London in 2014, I acted as a consultant to many companies and organisations involved in the manufacture of what are now designated ultraprocessed foods.
I used to be a consultant to the Breakfast Cereals Advisory Board of the Food and Drink Federation.
I used to be a consultant for aspartame more than a decade ago.
When I was doing research at King’ College London, the following applied: Tom does not hold any grants or have any consultancies with companies involved in the production or marketing of sugar-sweetened drinks. In reference to previous funding to Tom’s institution: £4.5 million was donated to King’s College London by Tate & Lyle in 2006; this funding finished in 2011. This money was given to the College and was in recognition of the discovery of the artificial sweetener sucralose by Prof Hough at the Queen Elizabeth College (QEC), which merged with King’s College London. The Tate & Lyle grant paid for the Clinical Research Centre at St Thomas’ that is run by the Guy’s & St Thomas’ Trust, it was not used to fund research on sugar. Tate & Lyle sold their sugar interests to American Sugar so the brand Tate & Lyle still exists but it is no longer linked to the company Tate & Lyle PLC, which gave the money to King’s College London in 2006.”
Prof Alberto Fiore: “I do not have any conflict interest along the side, have not received any funding from industry or other research, have no roles as committee member.”
Prof Gunter Kuhnle: “I am a former member of the UK Committee on Toxicity of Chemicals in Food, Consumer Products, and the Environment; a current member of the Advisory Committee on Novel Foods and Processes; the Director of the Chemical Analysis Facility at the University of Reading, which provides analytical services to academic and commercial clients; have received research funding (2010–20) from Mars for work on flavanols; and have received consultancy payments from RSM UK and EQT, paid to the University of Reading. As a member of the EFSA ANS panel (2018-2019) and the UK’s COT (2019-2025) I was involved in the evaluation of food additives.”