[사이언스미디어센터] 전문가 의견 '초가공식품의 세계적 증가, 심각한 공중 보건 위협' (랜싯)

*2025.11.18.

*2025.11.18.

Comments focusing on paper 1 but also giving some comments more widely on the series:

Prof Martin Warren, Chief Scientific Officer and Group Leader, the Quadram Institute, said:

“This paper provides a thoughtful overview of the history of UPFs and the development of definitions used to characterise them, while acknowledging where these classifications could be improved.  It draws on an extensive body of evidence, from purchasing data and national surveys to cohort studies, interventional trials, and meta-analyses, to evaluate three major hypotheses.  Although the conclusions are broadly aligned with the direction of the available evidence, the strength of the causal claims exceeds perhaps what the underlying data can fully support.

“In particular, the paper gives insufficient weight to socioeconomic confounding, which remains a major challenge in untangling the effects of UPF consumption from the broader social, dietary, and behavioural contexts in which these foods are consumed.  Many of the observed associations may therefore reflect structural disadvantage as much as product-specific harms.  The mechanistic evidence described is suggestive and worthy of further investigation, but it is not yet definitive, something the authors themselves acknowledge.

“The evidence presented convincingly demonstrates that UPFs have displaced traditional dietary patterns in many countries and that diets high in UPFs are characterised by poorer nutrient profiles.  However, this second point is conflated with the broader issue of poor overall diet quality, making it difficult to isolate the specific role of processing itself.  The link between UPFs and chronic disease is consistently observed across cohorts, but remains associative; the causal chain is incomplete.  The current evidence supports treating UPFs as likely contributors to poor health, but it does not establish that UPFs, as a category, are uniquely or mechanistically responsible.  It remains unclear whether the observed risks derive from components common to energy-dense UPFs, from what UPF-heavy diets lack, or from the combination of both.

“The paper makes a compelling public health case for action, and the data justify heightened concern.  However, the assertion that UPFs are a key driver of chronic disease should still be regarded as reasoned rather than conclusively proven.

“These three papers make a timely and valuable contribution to the ongoing debate about ultra-processed foods (UPFs) and their continued dominance in the global food system.  They underscore the strong associations between UPF consumption, obesity, and a wide range of related co-morbidities, while also drawing attention to the need for more nutritionally balanced diets.  Greater intake of fibre, vitamins, and minerals will be essential for tackling the growing problem of “hidden hunger”, a phenomenon that is reflected in rising NHS admissions in the UK for micronutrient deficiencies over the past decade.

“The series of papers rightly emphasises that improving diets cannot rely on consumer behaviour change alone.  Meaningful progress will require coordinated policies that reduce UPF production, marketing, and consumption, while simultaneously addressing the excessive presence of salt, sugar, and unhealthy fats in the food supply and improving access to affordable, minimally processed foods.  The final paper argues that UPFs are the product of a food economy driven primarily by corporate profit rather than nutrition or sustainability, and that a coordinated global response is needed to counter this dynamic.

“While many of these points are persuasive, international action ultimately depends on national commitment.  For such commitment to emerge, much greater domestic attention must be given to education, food literacy, and public awareness.  Critical questions remain unaddressed: How do we effectively improve food literacy, such as label interpretation, portion awareness, and basic cooking skills?  How can long-term public awareness campaigns help people recognise UPFs and understand their health implications?  And how do we build school-based interventions that foster healthier eating patterns from an early age?

“A further omission in the papers is a deeper acknowledgement of socioeconomic inequalities and the practical difficulty of influencing dietary behaviours in lower-income groups.  Although the papers touch on issues of affordability and access, they do not fully capture the structural realities: UPFs are cheap, convenient, aggressively marketed, widely available, and often the only realistic option for individuals with limited time, money, or cooking facilities.  Policymakers may understand the problems within the food system, but designing interventions that meaningfully shift eating patterns, particularly for those most constrained by circumstance, remains an immense challenge.

“Addressing these social and economic barriers is crucial, not only for improving individual health outcomes, working life expectancy, and mental well-being, but also for reducing the broader economic burden associated with lost productivity and escalating demands on the NHS.  A comprehensive response to UPFs must therefore pair regulatory action with sustained investment in education, community support, and structural change, ensuring that healthier choices become both accessible and achievable for all.”

Dr Hilda Mulrooney, Reader in Nutrition & Health, London Metropolitan University, said:

“I find the papers very interesting and I think they add a great deal to the literature.  Mixed methods have been used which I think strengthens the case being made.  None of the methods alone would be sufficient but together they make a compelling picture.

“Much of the evidence in relation to health effects is epidemiological in nature and therefore cannot show causality, and the randomised controlled trials cited are very short term (1-8 weeks duration) and with small group sizes (9-20 individuals).  However, as the authors highlight, carrying out RCTs in this area is difficult.  Using diverse data from multiple countries (e.g. sales data, national dietary intake data, epidemiological data and RCTs) strengthens their arguments.  The systematic review and meta-analysis carried out in paper 1 is carried out to a high standard.  There is biological plausibility for the potential impacts of these products.  A key criticism of the Nova classification is that it relies solely on the degree of processing rather than the nutritional value of the food.  The authors have addressed this - and other - criticisms.  They are not suggesting that nutritional content should be ignored in the rush to examine the nature of processing in different foods.  Action to address high fat, salt and sugar (HFSS) and UPFs combined in my opinion sounds sensible.

“The case they make is that the evidence is compelling and growing, and that action to address UPFs cannot and should not wait until all the evidence is complete.  In my view taking a precautionary approach is justified.  Given the disproportionate risks of chronic disease to the most disadvantaged groups and the costs of poor diet to individuals, healthcare systems and finances, it is beyond time to act.  Individual countries have taken fragmented and individual approaches to improve the food system but since it is global, a much more coordinated and ideally global approach is needed to address it.

“At this stage I do not think that taxation alone will be sufficient - much stronger, ideally coordinated effort is needed.  For example, it cannot be accepted that the harms of a westernised diet are allowed to be promoted and marketed aggressively without consequences in less developed countries where access to public health information or indeed to adequate healthcare may be lacking.  The information available on the multiple negative effects of these foods - including pollution from packaging and potential health impacts - should be paid for by the companies causing the problems using a 'polluter pays' approach but nothing like this will work unless the action taken is global.  Taxes will not work unless the action taken is coordinated across multiple jurisdictions.  Whether there is political will to identify, acknowledge and fight vested interests (of companies and countries) is unclear.”

Declared interests

Dr Hilda Mulrooney: “I am a committee member of the BDA Obesity Group, council member for Public Health for the Nutrition Society, committee member of the European Specialist Dietitians Network for Obesity and of the Obesity Management Collaborative.  All these roles are unpaid.”

Comment on paper 3 looking at global corporations:

Prof Richard Smith, Professor of Public Health Economics, University of Exeter, said:

“The papers offer a comprehensive review of existing research from which to draw their position, rather than conduct major new empirical work or more formal ‘systematic review’.

“This series presents a much needed, timely and compelling picture of one of the major challenges facing global public health today.  Together with foods that are high in fat, salt and sugar, known as HFSS foods, ultra-processed foods are a core driver of diet-related disease, and are woven into all corners of the fabric of the modern, heavily industrialised, food system.

“The series rightly points to the further research that would be helpful, but also urges that we do not need to wait for that research to act now – there is sufficient evidence of the drivers of production and consumption, the extensive harm done to human health and available measures to address this urgent issue.

“As we know from studies in reducing HFFS, major, significant and meaningful change in consumption will not come from targeting consumers.  It will only arise from addressing global production, retail and marketing machines.  As the series outlines, there are many options to do this, but one critical question left unanswered, and where research is very clearly urgently needed, is the balance to be struck between confrontation and collaboration with industry to achieve this.”

Declared interests

Prof Richard Smith: “I’ve got not conflicts – nothing form industry.  I have obviously received funding for research on food systems but that is from public sources (UKRI, NIHR, WT) and I am on Science Council of FSA.”

Prof Kevin McConway, Emeritus Professor of Applied Statistics, Open University, said:

“I’m going to concentrate on the three hypotheses examined in the major part of the first of the three papers in this series – the part on ‘the evidence’.  That’s where my own expertise as a statistician mainly lies, and I wouldn’t have had time to deal properly with the other two papers anyway.

“However, the first paper begins with an introduction that covers what it describes as the ‘main thesis’ of the series, and there is at least one aspect that doesn’t fit in too well with points made in the ‘evidence’ part of the paper.  I’ll come back to that later.

“On the three hypotheses, then, I’d say the overall picture on evidence is somewhat mixed.  There are aspects where the authors have, in my view, done a pretty good job, and there are others where, sometimes because the necessary evidence is still rather lacking, there remain holes in what has been established.

“One thing that I find particularly useful is the content of Panel 3, where the authors of the paper address some important scientific criticisms of the Nova classification and of research on UPFs.  I’ve made some of these criticisms in the past, and it’s good to see them being addressed and suggestions being made for research that could be done to alleviate problems.  But some of that research hasn’t been done yet, or was started only recently.  This means that some of the critical points that have been made do still apply to the evidence presented earlier in the paper.

“In particular, the first point in Panel 3 addresses the criticism that UPFs have not been defined clearly enough.  The authors do concede that assigning some food items to the correct groups can be challenging and requires training and standardised methods, and it describes methods that have been developed specifically to measure the consumption of UPFs and other Nova food groups.

“That’s good, but the publications that the paper references about the new measurement methods have all appeared since 2020.  So this does not deal with the criticism that many previous studies of associations between UPFs and poor health outcomes, particularly for chronic diseases which take a long time to develop, used dietary measures that date back well before the idea of UPFs and the Nova food classification first appeared in 2009.  This must cast some doubt on studies that used older methods of measuring diet, and that relates to some of the evidence for all three hypotheses.

“Further research using better dietary measurement might well not lead to different conclusions – but actually we don’t know whether that’s the case until and unless the research is carried out using the new measures.

“The final point in Panel 3 is the longest, and deals with the criticism that “UPFs vary in nutritional composition, so Nova might overlook health-relevant distinctions and misclassify certain foods as being unhealthy”.  As the paper points out, some studies have found differences in associations with the risks of different chronic diseases and the consumption of different types of UPF.  It also points out that it’s not particularly easy to pick out specific associations like this.

“But, interestingly, the final paragraph in the panel proposes that future studies should directly compare the health effects of specific UPFs with their non-UPF counterparts, mentioning comparing flavoured with plain yoghurts and comparing extruded (UPF) and wholegrain cereals.  That’s interesting because, in Panel 1 (in the introductory part of the paper) it’s confidently asserted that ultra-processed yoghurts are inferior to plain yoghurts with fresh fruit, and UPF cereals with extruded starches are inferior to steel-cut oats.

“Since Panel 3 says that research to establish whether the health associations of these different types of yoghurt or cereal haven’t been done yet, the statement in Panel 1 can only be an informed opinion, rather than something based on data.  Informed opinion has its place in public health, where action quite often has to be taken before all the data is in, but the paper should have been more transparent about what’s known and what’s not.

“The first hypothesis is that a dietary pattern based on ultra-processed foods (UPFs) “is globally displacing long-established diets centred on whole foods and their culinary preparation as dishes and meals”.  The authors address this by carrying out a narrative review of studies, for which they chose relevant studies based on their knowledge and expertise.  That’s in contrast to the review carried out on studies to assess the association between UPF intake and chronic diseases for Hypothesis 3 (see below), which described and used a strategy for searching databases for relevant studies, as recommended in the standard guidelines for carrying out reviews.  Without such a search procedure, one can’t be so confident that all relevant studies have been included.

“Though this is not an area in which I have strong expertise on the underlying data, I, personally, see no reason to doubt that UPF consumption has increased in the share of diets in the general way they describe.

“Whether the data establishes a change to ‘long-established diets’, as the hypothesis suggests, is perhaps less certain.  How long-established?

“It’s interesting that, with one exception (Canada), the time series of trends in UPF purchases in nine countries (in Figure 1) go back only to the 1980s at most, and the Euromonitor data (in Figure 2) starts in 2007.  In many of the countries, this would not go back to a period where the diet was largely “diets centred on whole foods and their culinary preparation as dishes and meals”, in the words of the hypothesis.

“It's interesting and, I think, relevant that the one country that did not provide good statistical evidence of an increase in UPFs in the national diet in Figure 1, the UK, is also the country that began industrialising first, over 200 years ago, and in which nowadays agriculture is the lowest of all the countries shown in terms of the share it contributes to the economy1, well under 1%, compared to between 5% and 7% in China, Argentina and Brazil.

“Maybe this is saying something about the links between the whole economic and social position in a country and the diet that its inhabitants consume.  And turning round a whole economy and society isn’t an easy thing to do.

“But it’s the second paper in this series that discusses policies to change levels of UPF consumption, and I am not commenting on that one.

“The second hypothesis, that the changing pattern of UPF consumption “results in deterioration of diet quality”, is supported by another narrative review of studies, which again were identified “based on the authors’ experience and expertise in this area”.  This again raises a question of whether all the relevant studies have been included, though I have no personal reason to doubt that they have.

“Obviously the hypothesis depends on a shared understanding of how diet quality should be measured.  Again, I’m no dietician, so I can’t comment on this, but again I see no obvious reason to doubt that the authors have looked at appropriate aspects of diet quality.

“I do have a concern that the conclusion on this hypothesis depends so much on a meta-analysis of data from national food studies in 13 countries (reference 46 in the paper).  This analysis was conducted competently, in my opinion, but it is observational and does not look at changes over time.

“This makes it difficult to assess how patterns of food quality might change if some action was successful in reducing the proportion of UPFs in a diet.  Would UPFs be replaced by other foods that are of higher quality in terms of fibre, calories, fats and sugar, or not?

“The other studies reviewed for this hypothesis, particularly the small randomised trials, do give some indication that reducing the UPF consumption would indeed improve diet quality, but those trials are very small and do not evaluate population-level actions.

“The third hypothesis is the one that has seen the most empirical research so far.  It is that the pattern of UPF consumption “increases the risk of multiple diet-related chronic diseases through various mechanisms.”

“The largest number of studies in this hypothesis that are reviewed in this paper are prospective observational studies on associations between UPF intake and the risk of chronic diseases in adults.

“Here the authors carried out a generally appropriate systematic review, which was pre-registered so that it’s possible to check their intentions before the review was actually done, and which, they state (and I don’t doubt), was conducted according to standard international guidelines for this type of review.  It involved a comprehensive search for studies to include in the review, and found 104 of them.

“The systematic review does seem to have been done properly, and does not (for example) just include studies that the authors chose from their experience.  I have to point out that I could not assess the review’s methodology in full detail.  The details are given at some length in an appendix to the main paper, but there has not been time for me to assess all the details of its methodology and its assessments of the quality of the studies by the deadline for these comments.

“I do have two issues about the findings of the review.  One is a generic doubt.  All the studies that were reviewed are observational, and single observational studies of this kind can never show beyond doubt that exposure to a risk factor, in this case consuming larger amounts of UPFs, actually causes changes in the outcomes, in this case having a chronic disease such as heart disease or diabetes later.  A study like this can find a correlation, but it can’t be certain about cause and effect.  That’s because many other factors could be associated with UF consumption, and they might also be independently associated with a disease outcome, so that the correlation could be caused by one of these other factors.  Such factors can be, and generally are, taken into account statistically, but those statistical adjustments can’t generally be shown to allow for everything.

“However, in this case, there are so many correlations of this sort, as well as some limited evidence from small randomised trials and from so-called mechanistic studies of how UPF consumption could lead to disease in the body, that it seems to me likely that at least some UPFs could cause increases in the risk of some chronic diseases (certainly not all – there’s little evidence of an increase in cancer risk, for instance).  But this certainly doesn’t establish that all UPFs increase disease risk.  There’s still room for doubt and for clarification from further research.

“My other issue worries me rather more.  It’s to do with the quality of the studies included in the review.  As is standard in such a review, the authors assessed the quality of each of the 104 included studies, using the standard framework for such assessments, the Newcastle-Ottawa Scale.  They report that most of the studies, indeed 85 of the 104, were rated as being of high quality.

“They also carried out a pretty standard additional analysis, by leaving out all the studies that were not classed as being of high quality.  This didn’t change their findings much, they report – except for one chronic disease, Crohn’s disease, which happened to be the one where the increased chance of the disease outcome in the group of people who consumed the most UPFs was biggest of all the chronic diseases they considered.  For Crohn’s disease, that reduced the number of studies analysed from four to two, and the statistical evidence for an increased risk became inconclusive.  This illustrates the potential impact of including poorer quality studies in the review.

“But my concern isn’t so much about Crohn’s disease, but about the quality ratings in general.  I’ve mentioned that there has been concern about the accuracy of recording UPF consumption, particularly in studies using dietary measures that were designed before the introduction of the Nova classification and of the term UPF itself.  I do wonder whether most of the studies in this review used these old measures, and if so, how they were classified in quality terms.  Although many of the reviewed studies were published after the publication of the work on better measures of UPF consumption, referred to in Panel 3 of the paper, they are likely to have used measures from older databases that did not use these new methods.

“In principle, given full information about the studies in the review, this could be checked (though that would be quite time-consuming).  And it might turn out not to make much difference to the findings anyway – but one can’t say for sure at this moment.  We do need more research, using the better UPF measures that have been developed recently.

“The picture is being put together, but there are still gaps.  I’m certainly not advocating that no public health action about UPFs is taken until all the gaps have been filled.  But I am advocating transparency and clarity about what we have good evidence for and what we don’t.  It would be futile to claim that public health shouldn’t be political – it involves important public decisions.  But people should be informed about where the evidence gaps are and what’s actually being done to fill them.”

1 See https://ourworldindata.org/grapher/agriculture-share-gdp

Declared interests 

Prof Jules Griffin: “I am a consultant for Sitryx, a company specialising in designing drugs to target immunometabolism.  The company is looking at drugs for irritable bowel disease and psoriasis.

I have received funding from the European Union to investigate endocrine disrupting chemicals and hold a grant from UK Research and Innovation examining the health benefits of a fish diet.

I hold shares in GlaxoSmithKline and Haleon plc.”

• 한국과학기술미디어센터 초대 이사장에 노정혜 서울대 교수(조선일보)
• 한국과학기술미디어센터장에 이근영 사회평론아카데미 기획위원(연합뉴스)
• 언론-과학자 잇는 ‘가교’ 되겠다…한국과학기술미디어센터 개소(매일경제).

• 과학자, 연구자에게는 의견과 해설이 온전한 맥락과 함께 제공되는 안전한 발언 공간이 돼줍니다. 선의를 위해 한 논평이 기사화 과정에서 왜곡되거나 부정확하게 변질될 우려를 줄이는 완충 작용을 합니다. 
• 기관 홍보 담당자에게는 기관의 성과를 기자들에게 보다 객관적이고 정교하게 알리고, SMC 글로벌 네트워크를 통해 영향력을 높일 기회를 제공합니다. 
• 기자에게는 사안을 해석하는 데 도움이 될 치우침 없는 종합적인 정보를 빠르고 풍성하게 제공하고, 이를 통해 기사에서 과학과 기술을 보다 자유롭고 편리하게 활용하도록 돕습니다.