Because the link between certain animal source foods and chronic disease risk is mostly based on low-certainty evidence from observational studies, there is a need to clearly establish plausible mechanisms for harm. Saturated fat is typically mentioned as a contributor to cardiovascular disease, but the risk, if any, is contingent on the dose, food matrix, dietary context, and interindividual differences and vulnerabilities. These nuances are also valid for the other alleged harmful factors in animal source foods, including dietary cholesterol, haem iron, Neu5Gc, TMAO, 'animal' protein, and the effects of processing. Moreover, classifying red and processed meats as 'hazards' does not necessarily equate to them being risks, and hazard classification systems (e.g., the WHO's IARC system) have been criticized for their potential negative consequences.
This subsection contextualizes the various potential mechanisms for harm that have been proposed to explain potential associations between the intake of ASFs and disease [Hammerling et al. 2016; Jeyakumar et al. 2017; Barnard & Leroy 2020; Nordhagen et al. 2020]. As shown below, these mechanisms are largely unproblematic, except for some predisposed populations or when the application of excessive food processing exerts a clear harmful effect. In all cases, a distinction needs to be made between potential hazards and actual risks.
- Saturated fat and dietary cholesterol
- Haem iron
- Sialic acid N-glycolylneuraminic acid
- Trimethylamine N-oxide
- Gut microbiota
- Protein excess
- Omega-6 fatty acids and arachidonic acid
- Food processing
- A 'hazard' is not necessarily a 'risk'
_________________________________
Saturated fat and dietary cholesterol
Saturated fat and dietary cholesterol in animal source foods have often been vilified due to concerns about their potential to increase serum cholesterol levels and the associated risk of cardiovascular disease. However, the relationship between these factors and disease risk is complex and has been questioned. Meta-analyses have either negated or downplayed the impact of saturated fat on cardiovascular and mortality risk. The evidence is not as strong - and certainly more nuanced - as once thought. Replacing saturated fat with other nutrients is not necessarily beneficial and may even lead to worse outcomes, whereas advice to choose low-fat over whole-fat dairy is not well-supported by evidence. Overall, the relationship between saturated fat, dietary cholesterol, and health outcomes is more intricate than previously thought, and dietary recommendations should consider the complexity of these factors.
Saturated fat and dietary cholesterol in animal source foods have often been vilified due to concerns about their potential to increase serum cholesterol levels and the associated risk of cardiovascular disease. However, the relationship between these factors and disease risk is complex and has been questioned. Meta-analyses have either negated or downplayed the impact of saturated fat on cardiovascular and mortality risk. The evidence is not as strong - and certainly more nuanced - as once thought. Replacing saturated fat with other nutrients is not necessarily beneficial and may even lead to worse outcomes, whereas advice to choose low-fat over whole-fat dairy is not well-supported by evidence. Overall, the relationship between saturated fat, dietary cholesterol, and health outcomes is more intricate than previously thought, and dietary recommendations should consider the complexity of these factors.
Further reading (summary of the literature):
Saturated fat and dietary cholesterol are among the most vilified components in animal source foods (ASF), as they may increase the levels of total serum cholesterol and LDL-cholesterol, which have in turn been associated with the risk of cardiovascular disease [Sacks et al. 2017]. The case against ASFs is thus often one against saturated fat [Jensen et al. 2016], building on both observational [e.g., Zong et al. 2016] and intervention studies [e.g., Hooper et al. 2020]. However, the linearity of the relationships (e.g., saturated fat -> LDL-C -> cardiovascular disease) and the state of the science from observational studies and randomized controlled trials to underpin these accusations has been questioned as too simplistic [e.g., Kolby Zinöcker et al. 2021; Givens 2024]. Resistance to the introduction of the current evidence into dietary policy has been speculatively ascribed to conflicts of interest and longstanding biases [Teicholz 2022].
What does the summary evidence tell us?
Numerous meta-analyses, systematic reviews, narrative reviews, and expert opinions have downplayed or dismissed the impact of dietary saturated fat on cardiometabolic disease and cancer risk or mortality [Astrup et al. 2011, 2019, 2020, 2021; Lawrence 2013; Chowdhury et al. 2014; Lamarche & Couture 2014; Ravnskov et al. 2014; de Souza et al. 2015; Harcombe 2017; Harcombe et al. 2015, 2016, 2017a, b; Mente et al. 2017; Gershuni 2018; Zhu et al. 2019; Heileson 2020; Kang et al. 2020; Cortese 2022; Valk et al. 2022]. Taken together, systematic reviews almost universally suggest very small absolute changes in risk, if any, and the data is based primarily on low/very low certainty evidence [Talukdar et al. 2023]. The most recent position of the WHO [2023] should also be seen as a case of low to very-low certainty evidence due to low strength of association, unexplained inconsistency, and a lack of clear dose-response [as discussed here]. Similarly, the role of increased dietary cholesterol intake is much less clear than often assumed (depending on baseline levels), or may even be inexistent [Berger et al. 2015; Soliman 2018].In support of the old paradigm, and to counter the criticism related to the use of observational data, a few influential meta-analysis of randomized controlled trials (RCTs) are still often referred to. A 2015 Cochrane meta-analysis has suggested that reduced saturated fat intake could create a small but potentially important reduction in cardiovascular risk [Hooper et al. 2015], but its conclusions have been questioned and publication bias was reported after a methodological re-evaluation [Thornley et al. 2019]. Its 2020 update has failed to show evidence that reducing dietary saturated fat lowers mortality, myocardial infarction, and stroke, only showing a significant outcome for 'combined CVD events' [Hooper et al. 2020]. According to Astrup et al. [2021], however, the latter become nonsignificant upon sensitivity analysis when including only those trials that actually did reduce saturated fat (and excluding those that intended to reduce saturated fat but were not successful). For more details, an in-depth look at the problems with the 2020 Hooper study can be found here. In addition to these Cochrane reports, other meta-analyses have suggested that the replacement of saturated fat with polyunsaturated fat could lead to better health outcomes [Mozaffarian et al. 2010; Mensink 2016], but this conclusion has been criticized based on the selection of included trials or lack of clinical end-points [Nordhagen et al. 2020, p.8]. Focus is usually on a limited set of biomarkers, while other food-based risk factors are overlooked [Micha & Mozaffarian 2010; Bier 2016], potentially distracting from the main causes of dysfunctional metabolism [Ruiz-Núñez et al. 2016; Tsoupras et al. 2018].
Effects on LDL cholesterol are complex
The response of LDL-C shows inter-individual variation [Griffin & Lovegrove 2024]. Although saturated fat may increase LDL cholesterol in most (but not all) individuals, this may as well be a healthy response relating to cholesterol homeostasis, which becomes disrupted in the case of inflammation-driven metabolic disorders [Kolby Zinöcker et al. 2021]. Moreover, the increase in LDL cholesterol induced by saturated fat is predominantly due to an increase in large, buoyant LDL particles. The latter are much less strongly related to cardiovascular risk than the small, dense particle variants (sdLDL) [Zambon et al. 1999; Astrup et al. 2020; Ikezaki et al. 2021, 2023; Krauss 2022; Schaefer et al. 2023]. The association between large, buoyant LDL particles and atherosclerosis is much lower to non-existent [Aneni et al. 2019; Ceponiene et al. 2021; McGarrah et al. 2022; Sekimoto et al. 2022; Jung et al. 2023; Sæther et al. 2023]. Another way to approach LDL heterogeneity is to distinguish between native and modified (i.e., 'oxidized', 'electronegative') LDL, with the latter creating the damage [Nishi et al. 2002; Allison & Wright 2004; Chen et al. 2008; Chan et al. 2013; Sánchez-Quesada et al. 2017; Chu et al. 2020; Kim et al. 2020; Wang et al. 2020; Singla et al. 2021; Benitez et al. 2023; Itabe & Obama 2023; Jayaraman et al. 2023; Law et al. 2023]. Such oxidative modifications cannot only be induced by in vivo factors, such as systemic inflammation and insulin resistance [Bale et al. 2022], but also by oxidized lipids in the diet [Spiteller 2007; Ahotupa et al. 2009; Addis & Grootveld 2021; Gonçalinho et al. 2023]. For more info on the argument that modified LDL drives the disease, and not LDL as such, see: The Dietary Diary [2024].
Saturated fatty acids differ in effects
Structural and functional heterogeneity within the group of saturated fatty acids is underappreciated [Bhavsar & St-Onge 2016; Praagman et al. 2016; Perna & Hewlings 2022]. Stearic acid, with beef and dairy as its main sources, is associated with lower cardiovascular and cancer risk in contrast to palmitic acid, causing a drop in circulating long-chain acylcarnitines and likely increasing fatty acid beta-oxidation in vivo [Senyilmaz-Tiebe et al. 2018]. In mice, dietary stearic acid leads to visceral fat reduction, possibly through apoptosis of preadipocytes [Shen et al. 2014]. As another example of in-group heterogeneity, increased levels of circulating very long-chain saturated fatty acids are associated with lower risks of cardiometabolic disease and with better aging [Lemaitre & King 2022]. In a Dutch study, higher intakes of short- to medium-chain saturated fatty acids that are mainly derived from dairy were associated with a reduced risk of ischemic heart disease, whereas no association was seen with lauric, palmitic, and stearic acids [Praagman et al. 2016]. It has even been argued that pentadecanoic acid (C15:0), an odd-chain saturated fat found in butter, is to be considered as an emerging essential fatty acid [Venn-Watson & Butterworth 2022]. High levels of circulating very long-chain fatty acids (C22:0 and C24:0, found in nuts and seeds) display protective associations with all-cause mortality and cardiovascular outcomes [Tao et al. 2023].
Substitution may not bring obvious benefits
When it comes to dietary strategies, replacement of dietary saturated fat does not seem to be the straightforward rewarding dietary strategy that it is sometimes presumed to be [Heileson 2020]. Replacement by carbohydrates is unlikely to reduce major cardiovascular events and mortality [Siri-Tarino et al. 2010, Hooper et al. 2015]. Globally, cardioprotective diets tend to be higher in saturated fat and lower in carbohydrates [PURE cohort; Mente et al. 2023]. More controversially, this also may hold true for replacement with vegetable oils rich in omega-6 fatty acids, at least when controlling for confounding [Hamley 2017] or after re-examination of the data [Ramsden et al. 2016]. Also, a meta-analysis has shown that short-term substitution of saturated with unsaturated fat does not improve insulin sensitivity or β-cell function [Lytrivi et al. 2023].
Whole-fat dairy and saturated fat
The case of full-fat dairy exemplifies how the current dietary fat recommendations lag behind the scientific state-of-the-art [Torrez-Gonzalez 2023]. This problem is due to a reductionist view that presents dairy as a delivery system for individual nutrients, such as saturated fatty acids, rather than as a complex food matrix [Torrez-Gonzalez & Rice Bradley 2023; Dunne et al. 2024]. Calls for restriction of whole-fat dairy, based on the saturated fat rationale, to reduce mortality [Cavero-Redondo et al. 2019], colorectal cancer [Larsson et al. 2005], or cardiometabolic issues [Kratz et al. 2013; Astrup et al. 2016; Morio et al. 2016; Drouin-Chartier et al. 2016a, b; Mena-Sánchez et al. 2019; Mente et al. 2023] are not supported by evidence. Also, reduced-fat over whole-fat dairy does not lead to increased adiposity or worsened cardiometabolic risk markers in children [O'Sullivan et al. 2020]. Moreover, an inverse association of regular-fat dairy consumption with obesity risk has been found [Kratz et al. 2013; Astrup et al. 2016]. In the Framingham Offspring Study, males with higher intakes of dairy-derived saturated fats had a less atherogenic profile than males with lower intakes of these fats [Yuan et al. 2022]. While association between dairy intake and cardiovascular risk is found in some studies, this does not seem to be valid across dairy categories. Because dairy is a heterogenous food group, health effects associated with saturated fat content may of course potentially diverge per type of product [Van Parys et al. 2022]. Be that as it may, systematic review and meta-analysis has revealed that even intake of butter, an archetypal saturated-fat ASF, shows a negligible association with all-cause mortality, no association with any cardiovascular disease, and an inverse association with diabetes [Pimpin et al. 2016]. When looking at global populations data, the PURE cohort study makes a case for the inclusion of whole-fat dairy as a protective element in healthy dietary patterns, with a suggestion of 14 weekly servings [Mente et al. 2023; Mozaffarian 2023].
Eggs and cholesterol
Consumption of eggs, often unreasonably discouraged because of their cholesterol content [Soliman 2018], does not elevate serum cholesterol, except in a small minority of hyperresponders [Kim & Campbell 2018]. Also, it usually does not parallel worsened coronary heart disease or mortality and may even show a protective association with stroke mortality [Rong et al. 2013; Alexander et al. 2016; Marventano et al. 2018; Mazidi et al. 2019; Xu et al. 2019; Dehghan et al. 2020; Mah et al. 2020; Xia et al. 2020]. Although cholesterol from eggs intake may show a weak association with increased mortality in some US studies [Zhong et al. 2019], an inverse relationship has been found in Asia [Zhuang et al. 2020]. Results are similarly different between regions for the effects of egg intake on type-2 diabetes [Drouin-Chartier et al. 2020]. Take together, dietary recommendations to restrict egg intake to low intake levels are poorly supported by the evidence [Zagmutt & Pouzou 2023].
Haem iron from red meat has been associated with risks. However, this is not a concern for healthy individuals within a wholesome dietary context. Iron is usually kept within safe limits by homeostasis, except for individuals who have genetic predispositions or metabolic disturbances. In this case, a precautionary principle of haem restriction may be advisable. Excessive iron consumption can elevate plasma ferritin levels, which may then be linked to adverse outcomes like oxidative stress and ferroptosis, leading to disturbances in insulin metabolism and atherosclerosis. However, the association between ferritin and cardiometabolic disease is inconsistent. Another point of concern is that in vitro and rodent models seem to suggest that high iron exposure could promote colorectal cancer by inducing lipid peroxidation, inflammation, cellular damage, and genotoxicity. Such studies, however, usually rely on unrealistically high levels of exposure and a dietary context that is not typical in human healthy diets, insufficiently taking into account the presence of chemoprotective compounds. Colorectal cancer risk related to iron intake might also differ by dietary source, sex, and specific individual vulnerabilities. According to the WCRF, 'the evidence suggesting that the consumption of foods containing haem iron increases the risk of colorectal cancer is limited'.
Further reading (summary of the literature):
Haem iron, which is related to red and processed meat intake, has been associated with higher mortality [Etemadi et al. 2017], cardiovascular disease [Van der A et al. 2005; Kaluza et al. 2013; Fang et al. 2015], type-2 diabetes [Bao et al. 2012], and colorectal adenoma [Bastide et al. 2016] in several observational studies.
Cardiometabolic diseases: unclear outcomes
The concern is that high iron intake could lead to elevated plasma ferritin, which can in turn be associated with pathological developments, such as the risk of type-2 diabetes [Montonen et al. 2012; Wittenbecher et al. 2015]. Iron overload can trigger oxidative stress and the development of ferroptosis, leading to decreased insulin production and insulin resistance [Simcox & McClain 2013; Miao et al. 2023], the deposition of iron in macrophages which may lead to the formation of foam cells and atherosclerosis [Cai et al. 2020], and accelerated aging [Wang et al. 2023]. Yet, high iron status has not been consistently associated with the risk of cardiovascular diseases. Taken together, observational data shows no significant association between markers of iron status, such as high ferritin levels, and coronary heart disease [Das De et al. 2015; Reyes et al. 2020]. A Mendelian randomization study even supported the hypothesis that higher iron status may reduce coronary artery disease risk [Gill et al. 2017].
Colorectal cancer: limited evidence
It has been suggested that haem iron promotes colorectal cancer [Bastide et al. 2011]. Iron is needed for nitrate reductase activity, involved in bacterially mediated N-nitrosylation in the colon. However, no differences were found when comparing interventions with red meat with white meat, although both increased colonic N-nitrosation [Bingham et al. 2002]. If causal relations between colorectal cancer risk associations and higher iron consumption would exist, they may be complicated. In the EPIC cohort, for instance, such associations differ by dietary source and sex [Aglago at al. 2023]. In women, risk was not associated with intakes of total, haem, or non-haem iron. In men, risk was not associated with total iron intake, non-significantly associated with haem iron, and inversely associated with non-haem iron. Similarly, in the EPIC-Spain cohort, no evidence was found to support a link between nitrosyl-heme or heme iron intake from processed meats and colorectal cancer risk in Spanish subjects [Rizzolo-Brime et al. 2024].
In addition to what has been observed in human studies, in vitro models and rodent models have shown a variety of effects, in particular with relationship to the promotion of colorectal cancer, because of the induction of lipid peroxidation, which then leads to inflammation, cellular damage, and genotoxicity [Bastide et al. 2011; Gamage et al. 2018; WCRF 2018; Martin et al. 2019]. Whilst some studies suggest no effect of haem iron on colon carcinogenesis [Parnaud et al. 1998], the haem exposure in experimental animal studies that do so is usually orders of magnitude above normal dietary intake by humans, whilst the diets are often both low in calcium and high in fat [Kruger & Zhou 2018]. Moreover, detrimental effects are counterbalanced by the presence of antioxidants and chemoprotective substances in the diet, for instance due to the presence of green vegetables, olive oil, and calcium [Gamage et al. 2018; Sasso & Latella 2018; Yang et al. 2019]. According to the WCRF, 'the evidence suggesting that the consumption of foods containing haem iron increases the risk of colorectal cancer is limited' [WCRF 2018].
Vulnerable populations may need to moderate
Iron overload (hemochromatosis) should not occur in most healthy subjects due to homeostasis, yet it may be of concern in certain individuals. This predisposition may have originated as a genetic adaptation to cereal-based diets during the Neolithic [Naugler 2008]. However, the condition is more relevant in Northern Europe (up to 10% allele frequency) than for populations of Mediterranean and Near East origin, possibly because of a higher dietary intake of vitamin C in the latter regions (an iron uptake cofactor). In addition, it may affect people that suffer from metabolic disturbances, such as in type-2 diabetic patients due to the suppression of liver hepcidin synthesis [Wang et al. 2014]. In the elderly Framingham Heart Study cohort (68-93y old), the risk of high iron stores was higher in subjects consuming more than four portions of red meat per week [Fleming et al. 2002]. For those subjects vulnerable to iron overload, dietary iron restriction is advisable, and therefore a low consumption of red meat; alternatively (or additionally), iron chelation and regular bloodletting can be applied [Cai et al. 2020].
Sialic acid N-glycolylneuraminic acid
The idea that Neu5Gc, a signalling molecule in mammalian cells and present in meat and milk, would trigger human antibodies and thereby cause inflammation is speculative. Given that red meat has been a part of human evolutionary diets, such speculation seems highly unlikely.
Further reading (summary of the literature):
The sialic acid N-glycolylneuraminic acid (Neu5Gc) is present in mammalian cells as a signalling molecule, which after consumption of mammal flesh and milk may generate human anti-Neu5Gc antibodies, creating chronic inflammation []. However, this remains mere speculation and seems counterintuitive given that red meat is an evolutionary relevant food within the human diet [see elsewhere]. Moreover, evidence from human hunter-gatherers (Hadza) and animals (mice) has indicated that diets rich in Neu5Gc select for gut microbiota that can mitigate pro-inflammatory effects based on the enzymatic release of Neu5Gc from meat, preventing its incorporation into colonic tissue [Zaramela et al. 2019].
Trimethylamine N-oxide
The conversion of L-carnitine (from meat) or choline (from eggs) into TMAO has been proposed as a mechanism leading to atherosclerosis, but this is likely a red herring. The biological role of TMAO needs to be cautiously considered within a context of complex interactions between diet, gut microbes, and the host. The case becomes more complex when considering the protective effects of fish, which contributes significantly more TMAO than other sources such as meat. Moreover, some studies found no connection between meat and TMAO, especially in a healthy dietary context, while TMAO levels rose in parallel with less healthy plant-based diets. In pigs, a prudent diet with red and processed meat resulted in lower TMAO excretion compared to the same meat-based approach combined with a Western-type background diet. The link between diet, TMAO, and disease is intricate and not as straightforward as it may appear. Factors like cardiometabolic and kidney diseases can elevate TMAO levels, suggesting that observational evidence may be influenced by confounding or reverse causality.
Further reading (summary of the literature):
The conversion of L-carnitine (meat) or choline (eggs) in trimethylamine by the gut microbiome and further in the liver to trimethylamine N-oxide (TMAO) has been pointed at by various authors as a potential culprit to explain the association between meat eating and atherosclerosis [Koeth et al., 2013; Abbasi 2019; Wang et al. 2019; Crimarco et al. 2020; Wang et al. 2022].However, the role of TMAO as robust disease marker can be ruled out based on the generally neutral or even beneficial health responses obtained in studies with carnitine [Shang et al. 2014; Samulak et al. 2019] and dietary choline/betaine [Meyer & Shea 2017]. The case becomes particularly troubling when looking at the protective effects of fish [Wang et al. 2006; Zheng et al. 2012; Leung Yinko et al. 2014; Alhassan et al. 2017; Jayedi et al. 2018; Qin et al. 2018; Rimm et al. 2018; Zhao et al. 2019], which is the largest supplier of TMAO by orders of magnitude [Zhang et al. 1999]. TMAO levels may even serve as a biomarker for the consumption of fish and seafood [Zhang et al. 1999; Cho et al. 2017; Hamaya et al. 2020].As far as meat is concerned, a harmful relationship seems absent as long as a healthy eating pattern is adopted. In a Paleo dietary setting, high meat intake did not lead to increased TMAO compared to a regular diet with lower amounts of red meat and eggs [Genoni et al. 2019]. Another study also failed to find a link between meat and TMAO levels, while concentrations went up in parallel with an 'unhealthful plant-based index' [Hamaya et al. 2020]. In pigs, a lower urinary TMAO excretion was found for a prudent diet with red and processed meat than when this meat-based approach was combined with a Western-type background diet [Thøgersen et al. 2020].Taken together, the relationship between diet, TMAO, and disease seems intricate and not as linear and causal as often assumed. A moderate increase in the plasma TMAO of hypertensive rats did not negatively affect the circulatory system and increased dietary TMAO even reduced diastolic dysfunction [Huc et al. 2018]. Also, it has been shown that cardiometabolic disease and kidney disease increase TMAO levels, so that observational evidence is likely due to confounding or reverse causality [Jia et al. 2019; Papandreou et al. 2020]. In conclusion, TMAO is - most likely - no more than a red herring [Landfald et al. 2017] and should not be used as a marker for unhealthy diets [Hamaya et al. 2020]. All this suggests that the issue is one of complex interactions between diet, gut microbiota, and host.
Gut microbiota
While the impact of animal source foods on the gut microbiota has been sometimes been portrayed as negative, compared to plant-based diets, this notion has been challenged by an intervention study. Adopting healthy diets lead to a positive shift in gut microbial composition and improved blood lipid profiles, regardless of the inclusion of unprocessed or processed lean red meats. This suggests that the relationship between animal source foods, gut microbiota, and health outcomes is less harmful (if at all) and certainly more complex than often assumed.
Further reading (summary of the literature):
It has been suggested that certain animal source foods affect the gut microbiota negatively [Xiao et al. 2023]. However, an intervention study has shown that independent of the addition of unprocessed or processed lean red meats, US-style healthy diets shift the gut microbial structure in a favourable direction and improve blood lipid profiles [Wang et al. 2023]. Further links with the gut microbiome are discussed in the paragraphs on Neu5Gc, TMAO, and nitrate/nitrite curing elsewhere on this page.
Protein excess
The notion that high-protein diets harm kidney and bone health in humans and may promote cancer and type-2 diabetes is not supported by robust evidence in the case of healthy humans. If anything, a generous intake of high-quality protein seems to promote rather than deteriorate health, especially in populations with higher needs, such as older adults. Although animal source foods are rich in certain amino acids like leucine and methionine,
which - as such - can theoretically accelerate oxidative stress and aging in excess, there is no clear evidence that this would have harmful effects on cancer risk in healthy humans eating wholesome diets. Moreover, potential negative effects of high-methionine diets can be balanced by higher glycine intake, available through nose-to-tail eating. With respect to type-2 diabetes, there is insufficient evidence for a risk increase with higher intake of animal protein or a risk decrease with plant protein intake, which suggests a lack of effect of protein per se.
Further reading (summary of the literature):
It has been suggested that high-protein diets lead to adverse health outcomes for kidney and bone health, as well as through mediation by (pro-oncogenic signalling) and dietary amino acid-induced stimulation of the mTOR pathway [Zhang et al. 2020]. In mice, ingestion of protein in excess (above a threshold of 22% of energy requirements) was found to activate mTOR and drive atherosclerosis, ascribed to higher circulating levels of leucine [Zhang et al. 2024]. ASFs are also high in methionine, an amino acid that accelerates oxidative stress and aging in animal models, when fed in excess [López-Torres & Barja 2008; McCarty et al. 2009]. Consumption of animal protein may decrease glycine levels in the circulation due to higher utilization in methionine metabolism, which may be associated with diabetes risk [Wittenbecher et al. 2015]. Because glycine plays an essential role in gluconeogenesis and the formation of glutathione, this may be related to insulin resistance and oxidative stress [Wang et al. 2013].However, there is no evidence that diets that are high in (animal) protein content have harmful effects [Wolfe et al. 2008; Traylor et al. 2018]. This is neither the case for kidney function in healthy subjects [Friedman et al. 2012; Devries et al. 2018; Van Elswyk et al. 2018; Remer et al. 2023] nor - according to the acid-base hypothesis - for bone loss, well on the contrary [Tucker et al. 2001; Promislow et al. 2002; Cao 2017]. Rabbit starvation (syn. protein poisoning or fat starvation) is usually not an issue provided that protein stays below 25% of the energy requirements at approximately 2.0-2.5 g/kg/d [Bilsborough & Mann 2006]. As a matter of fact, kidney injury seems to be triggered more readily by chronically low intakes of protein, at least in animal models [Fotheringham et al. 2021].Allegations that high protein intake would increase cancer risk have been disputed [Wolfe et al. 2008], despite speculations based on studies in experimental animal models. The proposed mechanism relates to the potential impact of the growth hormone receptor (GHR) and insulin-like growth factor (IGF-1) axis on tumor growth when protein intake is elevated [Levine et al. 2014]. In humans, however, higher intake seems mostly associated with benefits, such as better survival for women with breast cancer [Borugian et al. 2004]. Others have suggested, based on the IGF-1 rationale, that it may be prudent to keep protein intake lower during middle age but increase consumption for the elderly [Levine et al. 2014].Also, the effects of high-methionine diets on accelerated aging in animal models, are counterbalanced by glycine supplementation [Wang et al. 2013; Razak et al. 2017; Liu et al. 2019; Miller et al. 2019]. In addition, there is support for the benefits of glycine supplementation on mechanisms involved in healthy aging from human intervention trials [Kumar et al. 2022]. If this is indeed sufficiently relevant for humans, glycine levels can be increased by bone broth or nose-to-tail eating. Collagen has a glycine/methionine ratio of 25:1, so that 1 g of collagen for every 10 g of non-collagen animal protein would suffice [Masterjohn]. Concentrations of methionine are higher in fish eaters and vegetarians, followed by meat eaters, and lowest in vegans; vegans also had the highest concentration of glycine and meat eaters the lowest [Schmidt et al. 2016]. Compared to omnivores, higher serum homocysteine levels can be found [Majchrzak et al. 2006; Elmadfa & Singer 2009; Obersby et al. 2013; Vanacore et al. 2018].
In its evidence-based guidelines for protein intake, 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, which indicates a lack of effect [Schulze et al. 2023]. Even if animal protein intake is associated with increased risk of type-2 diabetes in observational cohort studies, the evidence is uncertain and there is a lack of effects of animal protein per se on major glycaemic traits during human intervention studies. The Society further concluded that the majority of systematic reviews on plant protein do not support a link between higher plant protein intake and lower type-2 diabetes risk, while there is also a lack of clear biological plausibility.
Omega-6 fatty acids and arachidonic acid
Omega-6 fatty acids present in eggs and meats have been (speculatively) linked to inflammation and carcinogenicity. However, the content and types of omega-6 fatty acids, as well as the omega-6/omega-3 ratio, depend on animal feeding methods. Moreover, how all this translates into actual and generalizable health outcomes is uncertain and a topic of debate. Arachidonic acid (AA), an omega-6 fatty acid present in eggs and meats,
is sometimes speculatively linked to inflammation and carcinogenicity, although it can also be anti-inflammatory. Consumption of red meat has been associated with circulating AA
levels in Chinese adults, which is attributed to pre-formed AA intake
rather than its precursor linoleic acid.
Further reading (summary of the literature):
Arachidonic acid (AA), an omega-6 fatty acid found primarily in eggs and to some degree in meats, is sometimes accused of being pro-inflammatory and pro-carcinogenic. Metabolization of AA by cyclooxygenase (COX) and lipoxygenase (LOX) leads to pro-inflammatory eicosanoids (prostaglandins and leukotrienes), but AA can also be metabolized into lipoxins, which have anti-inflammatory properties. Moreover, it seems that circulating AA is involved in the anti-inflammatory effects of fasting [Pereira et al. 2024].
In the context of colorectal cancer, red meat is cited as a potentially significant source of dietary AA [Phinney 1996]. Red meat consumption was associated with circulating AA in Chinese adults, which is said to be determined by intake of pre-formed AA rather than by AA's precursor linoleic acid [Seah et al. 2017]. However, the amount of omega-6 fatty acids in red meat and the omega-6/omega-3 ratio are highly dependent on the animal feeding strategies, since meat from range-fed ruminants or wild ungulates supports reduced pro-inflammatory prostaglandin PGE₂ synthesis compared to grain-fed variants [Broughton et al. 2011]. In older adults, the intake of red and processed meats is not associated with inflammation to begin with, when correcting for excess body mass [Wood et al. 2023]. Taken together, there is no reliable evidence to link consumption of animal source foods within healthy dietary patterns to morbidity risk based on AA intake.
Food processing
Meat processing may create potential carcinogens due to intense smoking or heating (HCA and PAH) or curing (N-nitroso compounds). Yet, evidence on their impact through human diets is mixed. The case against 'processed' meats may be less alarming than often stated. It stigmatizes a very heterogenous group of foods, many of which offer a lot of valuable nutrition per serving. The 'ultra-processed' variants, however, do require caution. Sodium, often high in
processed meats, conflicts with dietary guidelines, but the link between
salty foods and health risk is not conclusive for normotensive populations.
Further reading (summary of the literature):
The effects of some food processing interventions have raised concern, particularly in the case of processed meats. It remains unclear if the most commonly mentioned mechanistic speculations (heterocyclic amines, polycyclic aromatic hydrocarbons, N-nitroso compounds, sodium) can causally explain the associations between the intake of processed meats and colorectal cancer risk [Key et al. 2020].
Intense heating or smoking
Polycyclic aromatic hydrocarbons [PAH; Farhadian et al. 2010; Lee et al. 2016] and heterocyclic amines [HCA; Zhang et al. 2013; Góngora et al. 2019; Olalekan Adeyeye & Ashaolu 2021] are potentially carcinogenic chemicals that are formed during the heating or smoking of meat [Cross & Sinha 2004]. This is especially the case when using high heat, in direct contact with a flame, or upon badly performed barbecuing. According to the IARC/WHO (2015), however, there were not enough data 'to reach a conclusion about whether the way meat is cooked affects the risk of cancer'. In general, processing can certainly have detrimental effects on the toxicological safety of meat. Nevertheless, this is unlikely to be the case when processing is not excessive and when the products are consumed within a normal dietary context, not confounded by the effects of the Western ultra-processed food culture. That being said, caution is warranted, especially in the case of intense heating or smoking.
Curing with nitrite/nitrate
N-nitroso compounds have been mentioned as a particular concern for processed meats. Nitrate or nitrite are indeed used in the production of most commercial cured meats (with the exception of some variants, such as Parma ham or San Daniele ham), but usually in much lower concentrations (100-200 ppm) than is the case for leafy vegetables (up to 4000 ppm). In an Italian context, consumption of cured meats did not contribute to nitrate ADI exceedance [Roila et al. 2018]. Nonetheless, curing may cause the formation of N-nitrosamines [Crowe et al. 2019], especially during high temperature processing such as baking and grilling (T > 160 °C). In addition, the strong acidic environment of the stomach may result in endogenous formation of various N-nitroso compounds during the digestion of cured meat products [De Mey et al. 2017]. Some of these N-nitroso compounds could be carcinogenic when eaten in substantial amounts and within inappropriate dietary patterns [Cross & Sinha 2004; Etemadi et al. 2017]. Be that is may, the suggestion that nitrate/nitrite curing is harmful has been disputed [AAF 2020]. Straightforward mechanistic evidence is mostly lacking under proper dietary context and storage conditions [Turner & Loyd 2017]. In one rat study, bacon was even shown to be protective towards colon carcinogenesis [Parnaud et al. 1998]. Feeding cured meat to rats did not affect oxidative stress and inflammation, but affected the microbiome (increasing Ruminococcaceae) and increased stomach protein carbonylation and cecal propionate, while decreasing cecal butyrate, fecal phenol, and dimethyl disulfide levels [Van Hecke et al. 2021]. It is unclear what these findings may mean for human health.
Salting: excessive sodium
Salt levels are often high in processed meats, cheeses, or fish-derived products. Sodium has been listed as a main driver of the associations between processed meat intake and cardiometabolic morbidity, including type-2 diabetes risk [Männistö et al. 2010]. Moreover, the evidence for the link between salty foods and stomach cancer risk has been classified as 'suggestive' (above 'weak' but below 'highly suggestive' or strong') [Papadimitriou et al. 2021]. The salting of processed animal source foods indeed conflicts with dietary guidelines in favour of uniform and generalized low-salt approaches [Hu et al. 2011], but the validity of the latter to reduce clinical events or mortality is being challenged [Feldman & Schmidt 1999; O'Donnell et al. 2014; Mente et al. 2016; Graudal et al. 2017; Graudal & Jürgens 2018; Lelli et al. 2018; Mente et al. 2021; Ezekowitz et al. 2022; Kwon et al. 2022].
'Processing' level
Depending on the type of processing, animal-derived foods can be categorized as either 'processed' (e.g., traditional dry-aged ham) or 'ultra-processed' (e.g., deep-fried nuggets with reconstituted meat) [see also elsewhere for a discussion on ultra-processed foods]. Whereas ultra-processed food consumption is associated with an increased risk of colorectal cancer precursors in large cohort studies, these results remained essentially unchanged after excluding processed meat from total ultra-processed food intake [Hang et al. 2022].
A 'hazard' is not necessarily a 'risk'
There is a lack of consistent evidence establishing a
direct causal link between the consumption of red and processed meats
and chronic diseases, such as colorectal cancer. Despite the
International Agency for Research on Cancer (IARC) classifying red and
processed meats as (probable) carcinogenic hazards, this classification does not necessarily equate to equal levels of risk as other known carcinogens. Indeed, risk assessment suggests that hazards may not
translate into significant risks against reasonable dietary backgrounds. Whether or not a dietary hazard, such as red meat, also poses an actual health risk cannot be readily answered without a dedicated risk assessment. A risk is the complex outcome of a specific hazard combined with exposure (intake dose, preparation method, dietary context, etc.) and vulnerability (individual predispositions). Harms should also be weighed against benefits (e.g., nutritional value). Criticisms have emerged against IARC-type hazard classifications,
highlighting issues with methodology, expert conflicts, and outdated
approaches that can lead to unnecessary health concerns and adverse
consequences.
Further reading (summary of the literature):
Taken together, a consistent causal link between red and processed meat intake and chronic diseases, such as colorectal cancer, is missing [Oostindjer et al. 2014; Turner & Lloyd 2017]. Even as far as red meat has been classified as a 'probably carcinogenic' hazard by the IARC 2015, this has not been substantiated as an actual risk. According to WHO/IARC (2015), classifying processed meat in the same category as causes of cancer such as tobacco smoking and asbestos (IARC Group 1) 'does not mean that they are all equally dangerous. The IARC classifications describe the strength of the scientific evidence about an agent being a cause of cancer, rather than assessing the level of risk'. In fact, risk assessment indicates the opposite [Kruger & Zhou 2018]. Whether hazards translate into risks will depend on myriad factors, including dietary ones that may attenuate potentially harmful effects [Thøgersen & Bertram 2021].Lately, IARC-type hazard classification systems have been criticized for several reasons. Criticism relates to the methodology and involvement of conflicted experts [Lancet Oncology Editorial 2016], as has been advanced by one of the members of the IARC panel looking into red and processed meats [Klurfeld 2018], but also to the fact that such approaches are seen as outmoded and leading to undesirable consequences. The latter include avoidable health scares, public funding of unnecessary research and nutritional programs, loss of beneficial foods, and potentially increased health costs [Boyle et al. 2008; Anonymous 2016; Boobis et al. 2016].
