Avoiding red meat

• Lower colorectal cancer risk, especially by cutting processed meat.
• Reduce the risk of coronary heart disease and type 2 diabetes.
• Lower LDL cholesterol when red meat is replaced with plant proteins.
• Reduce TMAO and other gut-derived metabolites linked to heart disease.
• Reduce the environmental footprint of the diet.

The question has two parts, and most confusion comes from mixing them.

Definitions from IARC and WHO [1] [19].

What the evidence says, in one line each:

• Processed meat and colorectal cancer: sufficient evidence in humans; each 50 g a day raises risk by about 18% [1] [2].
• Unprocessed red meat and colorectal cancer: limited evidence in humans, strong mechanistic evidence [1].
• Heart disease and diabetes: consistent associations for processed meat, weaker and less consistent for unprocessed red meat [6] [5].
• Mortality: each extra daily serving linked to 13% (unprocessed) to 20% (processed) higher death rates in US cohorts [15].
• Randomized trials: none tests red meat against hard outcomes well; the best available found little or no effect [11].

The effects are relative and modest. Avoiding red meat matters most for people who eat it daily, and most of all for processed meat.

The IARC working group of 22 experts from 10 countries based its decision mainly on large prospective cohorts from the previous 20 years [1] [20].

• Processed meat (Group 1): sufficient evidence that it causes colorectal cancer. Each 50 g eaten daily, about two slices of ham, raises risk by 18% [1].
• Red meat (Group 2A): limited human evidence, mainly for colorectal cancer, with associations also seen for pancreatic and prostate cancer, plus strong mechanistic evidence [1].

Same group as tobacco, not the same risk. WHO's explainer is explicit that IARC groups describe how sure scientists are that something can cause cancer, not how much cancer it causes [19]. An 18% relative increase in a cancer with a lifetime risk of a few percent is a small absolute change for one person, though a meaningful one across a population, as IARC's Kurt Straif noted [1].

Dose response. A meta-analysis of prospective studies found colorectal cancer risk rose 17% per 100 g a day of fresh red meat and 18% per 50 g a day of processed meat, roughly linearly up to about 140 g a day [2]. The World Cancer Research Fund's update of 111 cohorts put the combined figure at 12% per 100 g a day [3].

The dissent. The 2019 NutriRECS review of the same cohorts found that cutting processed meat by 3 servings a week was associated with only very small decreases in cancer, on low-certainty evidence [4]. A 2022 Burden of Proof analysis rated the evidence for unprocessed red meat and colorectal cancer as weak [5]. Both groups agree on the direction; they disagree on how sure we can be.

Cohorts. An early meta-analysis split the two types. Processed meat was linked to 42% higher coronary heart disease risk per 50 g a day; unprocessed red meat showed no association (RR 1.00 per 100 g a day) [6]. Later, longer cohorts found smaller links for both: in 43,272 US men followed for 30 years, each daily serving was associated with 11% higher risk for unprocessed and 15% for processed red meat [7]. In UK Biobank, each 70 g a day of red and processed meat was linked to 15% higher ischaemic heart disease risk [12].

Trials on cholesterol. What replaces red meat matters:

• Against all comparison diets combined, red meat made no difference to LDL, HDL, ApoB or blood pressure [8].
• Against high-quality plant proteins (legumes, soy, nuts), red meat produced smaller decreases in total cholesterol (by 0.26 mmol/L) and LDL (by 0.20 mmol/L, about 8 mg/dL) [8].
• Eating at least 35 g a day did not worsen lipids or blood pressure compared with eating less [10].
• In a tightly controlled feeding trial, red meat and white meat raised LDL and ApoB by similar amounts compared with nonmeat protein [9]. On lipids alone, swapping beef for chicken does not help; swapping it for plants does.

So if you have high cholesterol, swapping red meat for legumes, soy or nuts helps; swapping it for refined carbohydrates does not. See LDL cholesterol and ApoB.

Mortality. In 121,342 US health professionals and nurses, each additional daily serving was linked to 13% higher total mortality for unprocessed red meat and 20% for processed. Replacing one serving a day with fish, poultry, nuts, legumes, low-fat dairy or whole grains was associated with 7% to 19% lower mortality [15].

Diabetes. Processed meat was linked to 19% higher diabetes risk per 50 g a day in an early meta-analysis [6]. In UK Biobank, each 70 g a day of red and processed meat was linked to 30% higher diabetes risk, though higher BMI accounted for a substantial part of the associations [12].

Two ways to grade the same data. NutriRECS applied GRADE, which starts observational evidence at low certainty, and concluded that cutting 3 servings a week was associated with very small reductions in mortality, heart attack, stroke and diabetes [14]. Frank Hu's group applied NutriGRADE, designed for nutrition studies, and rated the diabetes evidence high quality and the mortality evidence moderate [13].

The randomized evidence. Only one large trial is informative: in 48,835 women, a diet lower in red meat (as part of a low-fat diet) had little or no effect on mortality (HR 0.99) [11]. The reviewers themselves noted few trials and small differences in red meat intake [11], and Harvard critics pointed out that the key trial tested a low-fat diet, not red meat reduction [21].

What NutriRECS did. The consortium ran four systematic reviews of health effects and one of people's values and preferences, rated certainty with GRADE, and had a 14-member panel from 7 countries vote [22]. It found low-certainty evidence that eating 3 fewer servings a week was associated with very small benefits [14] [4]. Weighing those small, uncertain benefits against people's attachment to meat, it issued a weak recommendation to continue current intake [22].

The editorial. Aaron Carroll and Tiffany Doherty argued that repeated warnings based on small relative risks were not changing behavior, and that environmental and animal welfare arguments might persuade more people [23].

The responses.

• Harvard's Nutrition Source said the reviews' own meta-analyses largely confirmed links with poorer health, that GRADE was built for drug trials rather than diet, that the trial review leaned on a low-fat diet trial, and that excluding environmental impact was a mistake [21].
• Frank Hu and colleagues argued in Diabetes Care that the guidance rested on flawed methodology and misread nutritional evidence [13].
• JAMA reported on the backlash and on questions about the lead author's earlier industry-linked research funding [24].
• Annals later published corrected versions of the guideline and its reviews [22].

What to take from it. Neither side found that red meat is good for health. The disagreement is about certainty and about how to turn uncertain, modest relative risks into advice. On the numbers, both readings support eating less processed meat; for unprocessed red meat, reasonable experts disagree on how strongly to recommend cutting back.

Heme iron. People with the highest heme intake had 18% higher colon cancer risk, and heme catalyzes the formation of N-nitroso compounds and cytotoxic aldehydes from fat oxidation [18]. In rats and mice at dietary doses, heme increased precancerous lesions and tumor load, while heterocyclic amines from high-heat cooking and added nitroso compounds did not [25]. IARC described the mechanistic evidence for red meat as strong even though the human evidence was limited [1].

Processing. Curing, smoking and salting define processed meat [1]. Heme iron, heterocyclic amines from high-heat cooking and N-nitroso compounds are the three proposed explanations for the cancer link [25], and nitrite curing adds to N-nitroso formation, a leading hypothesis for why processed meat carries more risk.

TMAO. Gut bacteria convert L-carnitine, abundant in red meat, into trimethylamine, which the liver turns into TMAO. In mice, this accelerated atherosclerosis, and omnivores made more TMAO from carnitine than vegans or vegetarians [17]. In a controlled feeding trial in 113 healthy adults, 4 weeks of red meat, but not white meat or non-meat protein, more than doubled plasma and urine TMAO and reduced the kidneys' clearance of it [16]. TMAO is a marker with plausible links to heart disease, but no trial has shown that lowering it prevents heart attacks.

LDL cholesterol. Red meat also leaves LDL higher than plant proteins do [8], a separate pathway to heart disease from TMAO.

NHS and WCRF figures from their published guidance (see Examples).

For scale, the NHS lists a quarter-pound burger at 78 g, a 5 oz rump steak at 102 g and a cooked breakfast of two sausages and two rashers of bacon at 130 g.

Is red meat bad for you once a week?

Apart from EAT-Lancet's planetary reference, no major guideline sets a limit below about 3 portions a week. Mortality associations in cohorts are expressed per daily serving [15], and in trials, eating at least 35 g a day did not worsen lipids or blood pressure [10]. A weekly steak is inside the NHS and WCRF limits.

Replacement decides the result. In US cohorts, swapping a daily serving of red meat for fish, poultry, nuts, legumes, low-fat dairy or whole grains was linked to 7% to 19% lower mortality [15]. In trials, swapping it for legumes, soy or nuts lowered LDL; swapping it for other foods combined did not [8].

Iron. Red meat is a rich source of well-absorbed heme iron. In UK Biobank, more unprocessed red meat was linked to less iron deficiency anaemia (HR 0.80 per 50 g a day) [12]. Menstruating, pregnant and young people cutting red meat should plan iron from legumes, fortified grains and other sources, and can check ferritin.

Alpha-gal syndrome. Some people must avoid red meat entirely. After certain tick bites, people can develop IgE antibodies to a sugar called alpha-gal, causing delayed anaphylaxis or hives 3 to 6 hours after eating beef, pork or lamb [28].

The skeptics. Frédéric Leroy and Nathan Cofnas argue that claims of harm are improbable given human evolutionary history and lack robust support [27]. They are right that the evidence for unprocessed red meat is weak by trial standards [5]; they disagree with most guideline bodies about what to do with weak but consistent evidence.

For whole patterns that cut red meat, see the Mediterranean diet and plant-based diets. For the opposite approach, see the carnivore diet.

• Treating unprocessed steak and processed meat as the same risk.
• Replacing red meat with refined carbohydrates, pastries or processed plant products.
• Assuming chicken nuggets or processed poultry are a healthy swap.
• Cutting red meat without replacing its iron, zinc and B12, especially for menstruating women and teenagers.
• Reading a Group 1 cancer classification as meaning bacon is as dangerous as smoking.

• You have iron deficiency, heavy periods or are pregnant: cutting red meat is fine, but plan iron and check ferritin with your clinician.
• You are an older adult with poor appetite or low protein intake: make sure replacement foods keep protein high enough.
• You have a history of an eating disorder: rule-based food avoidance can be risky.
• You have chronic kidney disease: the protein source and amount may need guidance from your care team.
• You have alpha-gal syndrome: you must avoid mammalian meat, and sometimes dairy and gelatin, under an allergist's guidance.

• Lower iron and zinc intake if replacements are poorly planned.
• Lower vitamin B12 if all animal foods are cut.
• Lower protein intake if red meat is replaced with refined carbohydrates.
• More gas and bloating when legumes increase quickly.