Dietary recommendations often call for a restriction of the intake of animal source foods, in particular red and processed meats and animal fat, due to their presumed impact on public health. Overall there is no clear scientific unanimity on the matter, as such advice is based on partial data and prone to problematic interpretations. Observational studies typically suffer from confounding and bias, whereas there is insufficient evidence from intervention trials and experimental models to assume causality. In the case of red and processed meat, recent comprehensive quality assessments of the data have classified the certainty of the evidence as low.
This article contains the following elements to illustrate the controversy:
- What are dietary recommendations saying with respect to animal source foods?
- Why are recommendations to minimize animal source foods overstating their case?
- Why is advice to specifically limit red meat not supported by strong evidence?
- How should we evaluate the risk of animal source food intake, including red meat?
_________________________________
Dietary recommendations regarding the consumption of animal source foods (ASFs) often argue for restriction, particularly of red and processed meats and animal fat. Some suggest minimal ASF consumption, emphasizing the benefits of increasing plant food intake for better health. Certain authors even classify ASFs as 'less healthy' or equate red meat to unhealthy foods like sugar. These recommendations, however, lack unanimous scientific support and raise concerns, especially for populations with high nutritional needs. Some argue that these recommendations are based on incomplete data and problematic interpretations, and that overly restricting ASFs, which have evolutionary and nutritional value, may pose risks.
Dietary recommendations to moderate or restrict animal source foods (ASFs) refer to a presumed harmful impact on public health, particularly so for red and processed meats and for animal fat. Some authors argue that only small amounts of ASFs should be consumed [e.g., Willett et al. 2019; see elsewhere]; others even suggest that they should be totally eliminated from the human diet [Barnard & Leroy 2020]. An increased intake of plant foods is almost systematically promoted to improve health outcomes, whereas this is much less emphasized (or even omitted) for ASFs, despite their high nutritional value. Some authors go as far as clustering ASFs with 'less-healthy' plant foods [Satija et al. 2019; Asfura-Carrasco et al. 2022], or defining red meat as an 'unhealthy food choice' together with sugar [Willett et al. 2019], arguing that no intake level of red meat is without harm [Murray et al. 2020].Considering the potential implications that dietary recommendations may have on populations with elevated needs [see elsewhere], one would assume that they are supported by solid scientific evidence. They are, however, far from representing scientific unanimity. This is not only a problem for AFS [Barnard & Leroy 2020], but also more generally valid for nutritional sciences [see elsewhere]. It has been argued that calls for a decreased intake of ASFs are based on partial data and problematic interpretations [Leroy & Barnard 2020]. Given that they are evolutionary foods to which human physiology has adapted [see elsewhere], which are rich in otherwise limiting nutrients [see elsewhere], a too severe restriction of ASFs may entail risk [see elsewhere].
Why are recommendations to minimize animal source foods overstating their case?
Observational studies suggest that high consumption of some animal source foods may be associated with slightly increased health risks, but these findings are debatable and suffer from confounding and bias. Human intervention studies do not strongly support these claims, and biochemical studies lack clear mechanisms to explain them. Despite these uncertainties, observational studies often embed causal messages and recommendations in their findings, which can undermine the credibility of nutritional sciences. Researchers, however, tend to be more nuanced than policymakers, by acknowledging the limitations and potential biases in their findings and cautioning against overinterpretation due to heterogeneity and residual confounding. The WHO/IARC panel investigating the link between red meat and colorectal cancer also emphasized the possibility of chance, bias, or confounding and did not establish red meat consumption as a cancer cause.
Further reading (summary of the scientific literature):
While there are (debatable) indications from observational studies that high consumption of processed meats - and to a lesser extent red meat - may be accompanied by slightly increased risks, other ASFs mostly display neutral or protective associations; moreover, these data suffer from serious confounding and bias [see elsewhere]. In addition, human intervention studies do so far not support these findings on a more causal basis [see elsewhere], whereas biochemical studies do not convincingly provide clear mechanisms at present [see elsewhere]. For instance, the German Nutrition Society (GNS) has concluded that there is insufficient evidence for a risk increase in type-2 diabetes with higher intakes of animal protein and a risk decrease with plant protein intake, referring to a lack of biological plausibility, inconsistency, and contradictory messages coming from intervention studies [Schulze et al. 2023].Causal messages and recommendations for acting on the findings are nonetheless commonly embedded in observational studies [Olarte Parra et al. 2021] or the policy documents based thereon, which reflects a systemic problem and undermines credibility of the nutritional sciences [Cofield et al. 2010]. Although nutritional epidemiology should not be dismissed as beyond repair, it needs to be serious reformed [Brown et al. 2023]. The Nordic Nutritional Recommendations (NNR) are an example of how poor methodological practice is used at policy level, leading to problematic public health advice [Pouzou & Zagmutt 2023]. Key problems seem to be the lack of rigour, transparency, and consistency, poor handling of statistical uncertainty. Furthermore, there seems to be excessive flexibility in data collection, analysis, and reporting, making it 'unacceptably easy to accumulate (and report) statistically significant evidence for a false hypothesis' [cf. Simmons et al. 2023]. Such 'false positives' then become persistent dogma, amplified by publication pressure and eventually leading to a waste of resources and reputational damage of the research field. As highlighted by Simmons et al. [2023]: 'Our goal as scientists is not to publish as many articles as we can, but to discover and disseminate truth'To be fair, researchers are usually more nuanced than policy makers. As stated in a highly cited study on meat and mortality [Abete et al. 2014], results 'should be interpreted with caution due to the high heterogeneity observed in most of the analyses as well as the possibility of residual confounding'. The WHO/IARC [2015] panel looking into the red meat and colorectal cancer link likewise declared that 'other explanations for the observations (technically termed chance, bias, or confounding) could not be ruled out'. The same panel further stated that 'consumption of red meat has not been established as a cause of cancer' (emphasis added).
Why is advice to specifically limit red meat not supported by strong evidence?
Various experts have criticized the recommendations made by organizations such as the WCRF, IARC, EAT-Lancet Commission, Nordic Nutritional Recommendations, and Global Burden of Diseases consortium. Some concerns include issues with data quality and weak evidence for associations with diseases like cancer, diabetes, and heart disease. The GRADE system, a widely respected method for evaluating epidemiological data, suggests there is 'low' to 'very-low' certainty evidence that reducing red and processed meat intake leads to benefits in terms of cancer mortality, cardiometabolic outcomes, and all-cause mortality. As a result, it weakly recommends that adults maintain their current consumption of these meats, emphasizing individual decision-making based on absolute effect estimates. Additionally, the PURE study, a large global dietary cohort, suggests that unprocessed red meat should not be a primary target for health concerns and may be part of healthy dietary patterns. In summary, there is ongoing debate within the nutritional sciences about the recommendations regarding red and processed meat consumption, with some challenging the prevailing views and citing limitations in the available evidence.
Further reading (summary of the scientific literature):
Based on the above, the common case against red and processed meat has been criticized and contextualized by various authors [Alexander & Cushing, 2011; Alexander et al. 2015; Feinman 2018; Leroy & Cofnas 2020; Xie et al. 2022], for instance with respect to the WCRF2 [Truswell 2009] and IARC reports on colorectal cancer [Klurfeld 2018; Kruger & Zhou 2018], the Flemish Food pyramid [Leroy et al. 2018], the EAT-Lancet Commission's Planetary Health Diet [see elsewhere; Leroy & Cofnas 2020; Zagmutt et al. 2019a, 2019b, 2020; Stanton 2024], the Nordic Nutritional Recommendations (NNR) [Pouzou & Zagmutt 2023], and the Global Burden of Diseases, Injuries, and Risk Factors Study (GBD) [Stanton et al. 2022; Gordon Dseagu et al. 2022; Stanton 2024]. With respect to the GBD study, it was eventually admitted in the Author's Reply that there are issues with the use and interpretation of some of the data [Murray et al. 2022]. This was followed by a more detailed elaboration by the same group of researchers, stating that the evidence for colorectal and breast cancer, type-2 diabetes, and ischemic heart disease is 'weak', and non-existent for stroke, so that 'stronger and conclusive recommendations should not be made' [Lescinsky et al. 2022]. This report nonetheless came with a series of additional concerns [outlined in Stanton 2024].
Such severely criticized reports, including the GBD and EAT-Lancet studies, have not been retracted despite admitted shortcomings and keep on influencing dietary research and policies worldwide (including the UK National Food Strategy, the World Wildlife Fund’s Livewell Diet, and the Nordic Nutritional Recommendations) [Stanton 2024].GRADE analysis of the evidenceThe most comprehensive quality assessment of the evidence, using the authoritative GRADE system for epidemiological data, concluded that there is only 'low-' to 'very-low-' certainty evidence that higher intake of red and processed meats may result in very small reductions in cancer mortality and incidence [Han et al. 2019], or cardiometabolic outcomes and all-cause mortality [Zeraatkar et al. 2019]. Similar conclusions were obtained when assessing dietary patterns with less intake of red or processed meats [Vernooij et al. 2019]. The authors therefore suggested (as a 'weak recommendation') that adults continue (rather than reduce) current consumption of both red and processed meat [Johnston et al. 2019]. It is suggested, therefore, that guideline panels 'refrain from making strong recommendations' and should instead 'encourage individual decision making based on absolute estimates of effect' [Vernooij et al. 2021]. Similarly, another GRADE-based analysis concluded that the recommendation to reduce the consumption of processed meat and meat products in the general population to lower the risk of cancer 'seems to be based on evidence that is not methodologically strong' [Händel et al. 2021]. The use of Bradford Hill criteria for causality also led to the conclusion that red and processed meat intakes are 'not causally related to cardiovascular disease outcomes due to consistently weak associations and a lack of coherence with experimental evidence', even if for type-2 diabetes some potential causality was inferred in the case of processed meats [Hill et al. 2022].Because the GRADE analysis provided the most inclusive review of the evidence to date, it has been argued that 'those who seek to dispute it will be hard pressed to find appropriate evidence with which to build an argument' [Carroll & Doherty, 2019]. Moreover, the GRADE system is in use by over 110 organizations, including leading health organizations (e.g., World Health Organization, Cochrane Collaboration, National Institute for Health and Care Excellence) [Vernooij et al. 2021]. Those that indeed sought to dispute it (e.g., members of the True Health Initiative and the animal rights front PCRM) tried to discredit the authors based on alleged conflicts of interests in what is to be seen as a vitriolic campaign [Dyer 2020; Rubin 2020]. All this is indicative of a divide within the nutritional sciences, whereby 'the most vocal nutritional epidemiologists argue that GRADE is applicable only to contexts in which randomized trials are feasible', thereby accepting lower standards of evidence for nutritional epidemiology [Vernooij et al. 2021]. The usually cited alternative is the less strict NutriGrade method, even if the lead author of that method now endorses GRADE over his own approach [Vernooij et al. 2021; for further discussion on nutritional exceptionalism and lower standards in NutriGrade vs. the rigor of GRADE, see Brown et al. 2023].Explanatory video 💬 Science Animated 2021
Global data analysis (PURE study)
The PURE study, being the most comprehensive global dietary cohort, provided further evidence that unprocessed red meats should not be seen as a priority target for health, with higher amounts typifying the most healthy dietary setups (55 g/d), compared to the least healthy ones (24 g/d), while indicating no mortality or cardiovascular benefit for meat restriction to low amounts [Mente et al. 2023; Mozaffarian 2023].
How should we evaluate the risk of animal source food intake, including red meat?
The reporting of 'relative' risks associated with the intake of animal source foods in relation to health outcomes often leads to overstatements and sensationalism. These values are frequently presented without mentioning the 'absolute' risk, which can create exaggerated interpretations, especially among non-experts. Dietary guidelines also tend to overlook or downplay the very small absolute risk reductions that come with dietary changes. For example, an 18% relative risk increase for colorectal cancer linked to high processed meat consumption may sound alarming, but its corresponding absolute risk increase is only 1%. To put it differently, the likelihood of someone not developing colorectal cancer during their lifetime would only shift from 94% to 93% with such a dietary change.
Further reading (summary of the scientific literature):
Even if one assumes that the relative risks (RR) that quantify the association between the intake of ASFs and morbidity/mortality are not mere artefacts created by confounding, their usual reporting is typically leading to overstatements. In most publications, they are reported without mention of absolute risk (AR), potentially giving rise to hyperbolic interpretations by non-experts [Noordzij et al. 2017; Feinman 2018]. It is also the metric of choice in mass media, allowing for more sensationalist headlines [see elsewhere]. Given that AR is normally the standard approach for reporting magnitudes of treatment or exposure effects in guidelines, even dietary guidelines often fail to highlight, or even report, the very small absolute risk reductions involved in dietary scenarios [Vernooij et al. 2021].For instance, an 18-% RR increase (i.e., from 1.00 to 1.18) for colorectal cancer linked to high processed meat intake sounds more impressive than its corresponding AR increase of 1%. Yet, put differently, the AR that someone will not develop colorectal cancer during a lifetime would only shift from 94% to 93%.
