Key studies

There is no large randomised controlled trial of strict carnivore in humans. There are zero long-term follow-up studies. What exists is a mix of historical case work, adjacent ketogenic and low-carb literature, recent self-reported survey data, and a small number of observational studies.

That sparseness is honestly the most important fact about the evidence base. Treat anyone claiming carnivore is "proven" by science with the same skepticism you would treat anyone claiming it is "proven dangerous" by science. The data on which to make either claim does not exist.

What follows is the most relevant published work, organised by what it does and does not show.

Historical: the Bellevue Hospital experiment (1928)

Vilhjalmur Stefansson and Karsten Andersen spent a year on a meat-only diet under medical supervision at Bellevue Hospital in New York. The protocol was driven by Stefansson's observations of the Inuit, who lived on near-exclusively animal foods for much of the year.

Results published in 1930 (McClellan and Du Bois, Journal of Biological Chemistry):

• No scurvy, despite the standard prediction.
• No deficiencies of any other vitamin or mineral measured.
• Body weight stable.
• Cardiovascular and renal function within normal range at the end of the year.
• Both subjects in good health at study completion.

This remains the only long-form (one year) controlled meat-only observation in adult humans in the published literature. It is nearly a century old, the sample size is two, and the dietary specifics differ from modern strict carnivore — but it establishes that the strongest predicted harms (vitamin C deficiency, gross malnutrition) did not occur.

Recent: the Harvard survey (2021)

Lennerz et al., 2021, published in Current Developments in Nutrition. A survey of 2,029 adults who had been on a self-reported "carnivore diet" for at least six months. Not a controlled trial, not a clinical study — a survey.

What was reported:

• High satisfaction (95% reported the diet improved their health).
• Mean weight loss of 7 kg in those who started overweight.
• Self-reported improvements in chronic illnesses including type 2 diabetes, hypertension, autoimmune conditions, and psychiatric symptoms.
• LDL cholesterol elevated in many respondents. HDL also elevated. Triglycerides generally lower than baseline.
• No deaths or serious adverse events reported in the follow-up window.

Important caveats: this is a self-selected sample. The people who tried carnivore and felt worse are systematically less likely to still be on it after six months and therefore less likely to fill in a survey about it. Survivorship bias is the obvious large confound. The data is suggestive, not conclusive.

Adjacent: ketogenic and low-carbohydrate diet trials

Carnivore is in practice always a ketogenic diet. The substantial ketogenic literature applies in part. The most relevant strands:

Volek and Phinney — the metabolic case

Jeff Volek and Stephen Phinney have produced decades of work on low-carbohydrate physiology, much of it on athletic performance. The FASTER study (2016) compared elite ultra-endurance athletes on high-carb versus low-carb diets and found the low-carb cohort matched performance while running on dramatically higher fat oxidation. This sits adjacent to carnivore rather than directly on it but informs the metabolic-flexibility argument.

Virta Health type 2 diabetes trial

Virta Health (Sami Inkinen, Stephen Phinney, Jeff Volek) has published several papers on a clinical low-carb intervention for type 2 diabetes. The 2018 and 2019 results showed substantial diabetes reversal, weight loss, and medication reduction in the intervention arm, with positive cardiovascular markers. This is one of the better-conducted modern low-carb trials. Carnivore is more restrictive than Virta's protocol but operates on the same metabolic principles.

Ludwig group on metabolic adaptation

David Ludwig's group at Harvard has published widely on the metabolic effects of carbohydrate restriction. The work on the "carbohydrate-insulin model" of obesity (Ludwig et al., American Journal of Clinical Nutrition, 2021) reframes the energy-balance debate in a way that makes very-low-carb diets like carnivore mechanistically defensible.

The cholesterol question: LMHR work

A specific subset of people on ketogenic and carnivore diets show very large LDL cholesterol increases. These are typically lean, athletic, insulin-sensitive individuals — "lean mass hyper-responders" (LMHRs).

Nick Norwitz, Adrian Soto-Mota, David Feldman and others have published a series of papers on LMHR phenotype. The 2022 BMJ Open Heart paper and subsequent work characterise the response, propose mechanisms (the "Lipid Energy Model"), and report that LMHRs do not appear to have higher cardiovascular event rates despite the elevated LDL — though follow-up windows remain short.

This is the most active area of carnivore-adjacent research. If you have a dramatic LDL response on the diet, this literature is what your clinician should be reading before prescribing a statin reflexively.

What is not in the literature

The gaps are larger than the present:

• No long-term (10+ year) follow-up of strict carnivore practitioners.
• No controlled trial of carnivore versus another elimination diet for autoimmune outcomes, despite this being the most common reason people start.
• No published microbiome trajectory studies past about 12 months.
• No randomised data on carnivore for inflammatory bowel disease, despite many anecdotal remissions.
• No data on pregnancy and lactation outcomes — these populations are essentially untested.
• No data on childhood development on the diet.
• No formal trial of the Lion Diet specifically as an autoimmune protocol.

How to read all of this

A reasonable position based on the current evidence:

• The historical evidence (Bellevue) suggests strict carnivore does not produce the predicted gross deficiencies.
• The adjacent ketogenic literature is broadly favourable on metabolic markers, weight, and type 2 diabetes outcomes.
• The carnivore-specific evidence is weak by clinical standards but consistently positive in self-reports.
• The biggest individual risks (LDL response) are characterised enough that practitioners can monitor and respond.
• The long-term unknowns are real and should temper claims of "proven safety" in either direction.

The strongest defensible claim is: the diet is reasonably safe for healthy adults to run as a controlled experiment, with monitoring, for some months at a time, with the right baseline labs and follow-up. It is not the magic universal solution some advocates claim. It is also not the cardiovascular disaster some critics assume.

Reading list (worth your time)

• McClellan WS, Du Bois EF. "Clinical Calorimetry. XLV. Prolonged Meat Diets with a Study of Kidney Function and Ketosis." J Biol Chem (1930). The Bellevue paper.
• Lennerz BS et al. "Behavioral characteristics and self-reported health status among 2029 adults consuming a 'carnivore diet'." Curr Dev Nutr (2021). The Harvard survey.
• Volek JS, Phinney SD. "The Art and Science of Low Carbohydrate Performance" (2012). Practical reference.
• Volek JS et al. "Metabolic characteristics of keto-adapted ultra-endurance runners." Metabolism (2016). The FASTER study.
• Athinarayanan SJ et al. "Long-Term Effects of a Novel Continuous Remote Care Intervention Including Nutritional Ketosis for the Management of Type 2 Diabetes." Front Endocrinol (2019). Virta two-year data.
• Norwitz NG et al. "Elevated LDL Cholesterol with a Carbohydrate-Restricted Diet: Evidence for a 'Lean Mass Hyper-Responder' Phenotype." Curr Dev Nutr (2022).
• Ludwig DS et al. "The carbohydrate-insulin model: a physiological perspective on the obesity pandemic." Am J Clin Nutr (2021).

Where to go next

• The mechanistic case: Why it works
• The inflammation story specifically: Inflammation
• The people behind much of this work: Key figures