Nutrition, Diet & Fasting - Food foundations
Cholesterol and LDL, without the shouting
By Ryan Lowe. Last reviewed 3 August 2026. How we write this.
The argument that decides how to read the keto and carnivore pages. What LDL is, why it is treated as causal rather than merely associated, and what is genuinely contested versus what is settled.
This lesson exists because two other lessons in this classroom need it. Ketogenic and carnivore diets both raise LDL cholesterol in a substantial proportion of people, and what that means is the single most consequential unresolved question either of them faces. Answering it requires knowing what LDL actually is and what kind of evidence supports the mainstream position - which is stronger than its online critics usually acknowledge and narrower than its defenders sometimes imply.
What the numbers mean
Cholesterol does not dissolve in blood, so it travels inside particles called lipoproteins. LDL - low-density lipoprotein - carries cholesterol out to tissues; HDL is involved in returning it. "LDL cholesterol" on a blood test is not a measure of particles but of the cholesterol carried inside them, which matters because two people with the same LDL-C can have different numbers of particles. ApoB, which counts one per particle, is a better measure of particle number and is increasingly preferred; Lp(a) is a separate genetically determined particle worth measuring once in a lifetime and rarely mentioned online.
Total cholesterol on its own is close to useless as a number, since it sums components that behave differently.
Why LDL is treated as causal
The strongest evidence is not the observational studies that critics usually attack. It comes from genetics and from drug trials, which is a different and much harder-to-explain-away kind of evidence.
- Mendelian randomisation: people who inherit gene variants that give them lifelong lower LDL have lower rates of heart disease. Because the variants are allocated at conception, they are not confounded by lifestyle in the way an observational study is.
- Familial hypercholesterolaemia: an inherited condition causing very high LDL from birth, with dramatically elevated heart disease risk from a young age. One variable, one outcome.
- Drug trials across multiple unrelated mechanisms - statins, ezetimibe, PCSK9 inhibitors - all lower LDL by different routes and all reduce cardiovascular events, with the size of the benefit tracking the size of the LDL reduction.
That combination is why cardiology treats LDL as causal rather than correlated. Anyone arguing otherwise has to account for all three lines, not just the observational one.
What is genuinely contested
Several things, and they are real disagreements rather than fringe positions.
The first is whether LDL raised on a low-carbohydrate diet, in a lean and insulin-sensitive person with low triglycerides and high HDL, carries the same risk as the same number in someone with metabolic syndrome. This is the "lean mass hyper-responder" debate. Imaging work in this group has been offered as evidence that plaque burden does not follow the LDL number as predicted; published analyses have questioned whether the phenotype is distinct at all and argued the LDL rises are clinically significant across the board. Both positions appear in peer-reviewed journals. It is unresolved.
The second is the appropriate treatment threshold in people at otherwise low risk, and how to weigh a single elevated number against overall risk - a question of clinical judgement where guidelines differ between countries.
What is not seriously contested in the literature, whatever the internet suggests, is that LDL is causally involved in atherosclerosis. The live argument is about context, magnitude and thresholds, not about whether the relationship exists.
What to actually do with this
The practically useful conclusion is that this is measurable, and measuring it converts an argument into information. Anyone starting a diet known to move lipids - keto or carnivore in particular - has a strong case for a lipid panel before and a few months after, ideally including ApoB, and for taking the result to a doctor rather than to a forum. That is true whichever side of the argument turns out to be right, which is precisely what makes it good advice.
Educational content only and not medical advice. Interpreting a lipid panel requires knowing your full cardiovascular risk profile and family history - that is a conversation with a qualified healthcare professional, not something to settle from an article.
Related in this classroom
The studies behind this page
- Low-density lipoproteins cause atherosclerotic cardiovascular disease. 1. Evidence from genetic, epidemiologic and clinical studies (EAS consensus statement) - European Heart Journal, 2017
- Behavioral characteristics and self-reported health status among 2029 adults consuming a "carnivore diet" - Current Developments in Nutrition, 2021
Sources for this classroom
Educational content only. Not medical advice, diagnosis or treatment. Always consult a qualified healthcare professional before changing your health regimen.
