Copper and Cholesterol: The Copper-Deficiency Theory of Heart Disease

Since 1973 a small group of nutrition scientists, led by Leslie Klevay, has argued that heart disease is partly a disease of too little copper — or too much zinc relative to copper. Morley Robbins builds on that idea: in his telling, raised cholesterol, oxidised LDL, iron in artery plaque and the calcium that follows it are downstream of a copper shortage. The honest summary is mixed. Copper-deficient rats reliably develop high cholesterol, and a carefully controlled copper-poor diet did lower copper enzymes in people. But the only human depletion study to report a rise in cholesterol involved one man, a later and longer study in twelve women found no change, and large population studies link high blood copper, not low, to more heart disease. This page lays out the theory fairly, then sorts every link into human, animal or laboratory evidence.

Table of Contents

  1. The Zinc/Copper Theory in Brief
  2. Robbins’s Framing: Cholesterol, Oxygen, Iron and Calcium
  3. The Proposed Chain, Link by Link
  4. What the Animal Studies Show
  5. The Human Copper-Depletion Studies
  6. Oxidised LDL and Foam Cells: The Mainstream Part
  7. Iron in Plaque and the Iron Hypothesis (Contested)
  8. Where Copper Cuts the Other Way
  9. Evidence Map: What Is Human, Animal or Lab
  10. Practical: Food Copper, Labs and Safety
  11. Key Research Papers
  12. Connections
  13. Featured Videos

1. The Zinc/Copper Theory in Brief

In 1973 Leslie Klevay, a nutrition researcher who later worked at the US Department of Agriculture’s human nutrition research centre in Grand Forks, North Dakota, published a short rat experiment with a long title: raising the ratio of zinc to copper in the feed raised the animals’ blood cholesterol. Two years later, in 1975, he widened that finding into a full theory in the American Journal of Clinical Nutrition. His argument was that a metabolic imbalance between zinc and copper — either an outright shortage of copper or too much zinc relative to it — is “a major factor” in coronary heart disease, working through high cholesterol. He went so far as to suggest that, apart perhaps from cholesterol itself, no other factor had been tied so closely to risk.

The idea has an appealing logic. Zinc and copper compete for absorption in the gut, so a diet (or a supplement habit) heavy in zinc can push copper down. Refined Western diets are relatively low in copper. And copper is built into enzymes that matter for blood vessels: the copper-zinc form of superoxide dismutase (SOD), one of the body’s main antioxidant enzymes; ceruloplasmin, which carries most of the copper in blood and helps move iron; and lysyl oxidase, which cross-links the elastic fibres of artery walls.

What the theory needs, though, is evidence in people — and that is where it has always been thinner than the rat work. The rest of this page goes through it piece by piece.

Back to Table of Contents

2. Robbins’s Framing: Cholesterol, Oxygen, Iron and Calcium

Morley Robbins, the founder of the Root Cause Protocol, takes Klevay’s copper theory and joins it to his own central idea, the copper–iron dysregulation hypothesis. In his talks and interviews the argument runs roughly like this:

Some of this lines up with mainstream cardiology (oxidised LDL matters; iron is found in advanced plaques). Some of it is Robbins’s own synthesis without direct testing (the “oxygen sink” description of cholesterol, and calcium following iron). The sections below separate the two.

Back to Table of Contents

3. The Proposed Chain, Link by Link

Put together, the copper-deficiency theory of heart disease is a chain of five links. The diagram labels each arrow with the best kind of evidence that exists for it. Two things stand out. The first link — a low-copper diet lowering the body’s copper enzymes — has been shown in controlled human studies. The last two links — oxidised LDL being taken up by macrophages to form foam cells, and foam cells building plaque — are standard, mainstream atherosclerosis biology. The weak point is the middle: no human study cited here shows that a shortage of copper is what produces more oxidised LDL. And the side branch, the cholesterol rise that started the whole theory, is consistent in animals but rose in one man and did not move in 12 women.

A vertical chain of five boxes runs from copper deficiency, to impaired iron handling and lower antioxidant enzymes, to oxidised LDL, to macrophage foam cells, to plaque and calcification; the first arrow is tagged human evidence, the middle arrow is tagged mechanistic only, the third mechanistic and mainstream, and the last human pathology, while a side branch from copper deficiency to raised blood cholesterol is tagged consistent in animals but mixed in people, and a right-hand column concludes that the ends of the chain are solid while the copper-specific middle is untested in people. THE COPPER-DEFICIENCY CHAIN, LINK BY LINK each arrow carries the best kind of evidence behind it copper deficiency impaired iron handling, lower copper-zinc SOD oxidised LDL macrophage foam cells plaque, then calcification HUMAN controlled diets MECHANISTIC ONLY MECHANISTIC mainstream HUMAN autopsy pathology raised blood cholesterol ANIMAL: CONSISTENT people: 1 man rose, 12 women unchanged WHERE THE CHAIN HOLDS The first link is real low-copper diets lowered ceruloplasmin and SOD in people The middle is the gap the depletion studies did not measure oxidised LDL at all The end is textbook oxidised LDL, foam cells and plaque are mainstream atherosclerosis Calcium following iron not shown: plaque iron did not track plaque calcium in the one study that looked Both ends are solid; the copper-specific middle is untested in people.

Back to Table of Contents

4. What the Animal Studies Show

Evidence tier: animal. The theory was born in rats, and in rats it holds up well. In Klevay’s 1973 experiment, raising the zinc-to-copper ratio of the diet raised blood cholesterol. In 1978 Klevay and Kenneth Allen reported that copper-deficient rats developed highly significant cholesterolaemia, with plasma cholesterol tracking liver copper; their hearts were enlarged, with areas of haemorrhage, inflammation and focal necrosis, and their aortas showed distorted, depleted elastic fibres (lysyl oxidase, the enzyme that cross-links elastin, needs copper). Notably, the heart’s own arteries looked normal — the damage was to heart muscle and the aorta, not coronary plaque.

The rat work also produced one finding that matters for modern diets. USDA researchers showed in 1983 that in copper-deficient rats the type of carbohydrate mattered. Copper deficiency raised cholesterol and triglycerides whatever the carbohydrate, but on sucrose or fructose diets it was far more severe: of the copper-deficient rats, 1 of 10 fed starch died, against 7 of 20 fed fructose, apparently from rupture of the heart. That observation is why a human study of fructose and copper was run in the 1980s (see the next section).

Two caveats keep the animal evidence in its place. First, the diets used were severely deficient — much lower in copper, relative to need, than anything a person eating ordinary food is likely to reach. Second, rats handle cholesterol differently from people (most of their cholesterol travels in HDL, not LDL), so “cholesterol went up in rats” does not translate directly into “LDL goes up in humans.” Animal results show that a mechanism can exist; only human studies show whether it does at real-world intakes.

Back to Table of Contents

5. The Human Copper-Depletion Studies

To test copper in people, researchers at USDA nutrition centres fed volunteers carefully weighed diets in metabolic wards, lowering copper for weeks or months and then restoring it. These are small but tightly controlled experiments. Three are central to the cholesterol question.

Klevay, 1984: one young man, cholesterol up

Evidence tier: human experiment, single participant. A healthy young man ate a diet supplying 0.83 mg of copper a day, an amount the authors noted is similar to some ordinary diets. He lost more copper in urine and stool than he took in. His plasma copper, ceruloplasmin and red-cell superoxide dismutase all fell, and his plasma cholesterol rose, while his blood counts did not change. The authors concluded that lipid metabolism may be a more sensitive sign of copper status than anaemia, and that the result supported the copper theory of heart disease. It is a real and careful observation — but it is one man.

Reiser, 1985: 24 men, a fructose diet, and four removed for heart findings

Evidence tier: human controlled feeding study. Twenty-four men ate a diet comparatively low in copper (1.03 mg a day) containing either 20% fructose or starch, for 11 weeks. Fructose did not change serum copper or ceruloplasmin but did lower red-cell superoxide dismutase compared with starch; a later copper repletion raised it again in the fructose group. During the study, four of the men developed heart-related abnormalities and were removed. Supporters of the copper theory cite those four withdrawals often. What the paper’s abstract reports is the withdrawals and the enzyme changes; it does not establish that copper deficiency caused the heart findings, and it is not a cholesterol result.

Milne and Nielsen, 1996: 12 women, 105 days, no cholesterol change

Evidence tier: human controlled feeding study. Twelve postmenopausal women ate a diet supplying 0.57 mg of copper a day — lower than in Klevay’s study — for 105 days, then had copper restored. Several copper-dependent enzymes clearly fell (red-cell superoxide dismutase, platelet cytochrome c oxidase, glutathione peroxidase), so the women were genuinely becoming copper-depleted. Yet, in the authors’ words, the low copper intake “did not induce the changes in serum cholesterol and hematology generally found in copper-deficient animal models.” Plasma copper and ceruloplasmin barely moved either; the authors called them insensitive markers.

Taken together: controlled human copper depletion reliably lowers copper enzymes, which supports the first link of the chain. The cholesterol rise that defines the theory was seen in 1 man and not in 12 women eating less copper for longer. That is not a replicated human finding. It is also not a disproof — the groups differed in sex, age and study length — but it means “low copper raises cholesterol in people” is best described as unconfirmed.

Back to Table of Contents

6. Oxidised LDL and Foam Cells: The Mainstream Part

Evidence tier: laboratory mechanism, supported by human and animal tissue findings; mainstream. This is the part of Robbins’s framing that most cardiologists would recognise. In a widely cited 1989 review, Daniel Steinberg and colleagues set out why native LDL alone does not easily turn macrophages into the fat-laden “foam cells” seen in early plaques, whereas oxidatively modified LDL is taken up rapidly through scavenger receptors. Foam cells are the building blocks of the fatty streak, the earliest visible lesion of atherosclerosis. A 1992 review by Hermann Esterbauer’s group gathered the evidence that oxidised LDL is present in human and animal arteries and plaques, as well as the chemistry of how it forms.

So “oxidised LDL is worse than ordinary LDL” is solid ground. What it does not show is that copper deficiency is the reason LDL gets oxidised. Oxidation has many drivers — smoking, high blood sugar, inflammation, low intake of dietary antioxidants — and the oxidised-LDL model sits comfortably alongside, not in place of, the mainstream view that the amount of LDL in the blood matters too.

Back to Table of Contents

7. Iron in Plaque and the Iron Hypothesis (Contested)

Evidence tier: human tissue studies (iron is there); human cohorts (whether body iron predicts heart disease) — contested.

Iron is found in plaques. A 2004 study measured metals directly in human carotid plaques removed at surgery and compared them with healthy artery tissue. Iron in the inner artery wall of the plaques was markedly higher (0.370 versus 0.022 nmol per mg of tissue by one method), and plaque iron rose with plaque cholesterol. A 2003 autopsy study of people who died suddenly of coronary causes found iron deposits and red-cell membrane remnants clustered in advanced, unstable plaques, and traced them to small bleeds inside the plaque — red blood cells leaking in from fragile new vessels. That is the mainstream explanation of where most plaque iron comes from: hemorrhage inside the lesion, not a body-wide iron excess.

Does calcium follow iron? The same 2004 study looked, and found that plaque iron did not correlate with plaque calcium. That is one study, but it is the only direct test cited here of Robbins’s “calcium follows iron” claim, and it did not support it.

The iron hypothesis. In 1981 Jerome Sullivan proposed in The Lancet that stored iron explains why premenopausal women have less heart disease than men, since menstruation keeps their iron stores lower; he suggested regular blood removal as both a test and a possible prevention. The idea was influential, but it has not held up cleanly. A 1999 meta-analysis of 12 prospective studies with about 7,800 coronary cases found no good evidence of a strong link: people with ferritin of 200 micrograms per litre or more had essentially the same coronary risk as those below it (combined risk ratio 1.0), and transferrin saturation, serum iron and dietary iron showed no meaningful association either. The iron hypothesis is best described as contested and unproven. Iron in plaque is real; whether higher body iron causes coronary disease is not established. For more on Robbins’s view of iron, see Iron Overload and Hidden Toxicity.

Back to Table of Contents

8. Where Copper Cuts the Other Way

An honest look at the theory has to include the evidence that points in the opposite direction, because there is a good deal of it.

Free copper oxidises LDL in the test tube (tier: laboratory). The standard laboratory method for making oxidised LDL is to add copper ions. Esterbauer’s 1992 review describes LDL “oxidized by exposure to copper ions” as the workhorse model of the field. Copper bound in enzymes such as superoxide dismutase is protective; loose copper is a pro-oxidant, just as loose iron is. That does not prove copper causes plaque in the body, but it means “more copper means less oxidation” is not automatically true.

High blood copper goes with more heart disease, not less (tier: human observational). In a US national survey cohort of 4,574 adults aged 30 and over, with 151 deaths from coronary heart disease, people in the highest quarter of serum copper had about 2.87 times the risk of coronary death of those in the lowest quarter, after adjusting for cholesterol, smoking, blood pressure and other factors. A 2018 BMJ meta-analysis of 37 studies of toxic metals found that people in the top third of copper levels had a relative risk of 2.22 for coronary heart disease compared with the bottom third.

How the copper camp answers this. Serum copper is mostly ceruloplasmin, and ceruloplasmin is an acute-phase protein that rises with inflammation — and atherosclerosis is an inflammatory disease. So high serum copper may be a marker of inflammation rather than a sign of too much copper in the tissues; the US survey author himself said it remains to be established whether copper affects atherosclerosis directly or simply marks the inflammation that goes with it. That is a fair point, and it is close to Robbins’s own argument that serum copper is a poor measure of usable copper. But it cuts both ways: if serum copper cannot show excess, it also cannot show deficiency, and there is no large human cohort cited here showing that low copper status predicts heart disease. The observational evidence that exists does not support the theory.

Back to Table of Contents

9. Evidence Map: What Is Human, Animal or Lab

Each row is a claim from this page. A filled dot means the evidence of that kind supports the claim; an open ring means evidence of that kind exists and does not support it, or is null; an empty cell means this page cites no evidence of that kind. The pattern is the whole story: the claims that are specific to copper are filled mainly in the animal and lab columns, while the human columns carry open rings.

An evidence map with nine claims as rows and four columns for animal, human trial, human observational and laboratory evidence, where filled dots mark support and open rings mark null or contrary evidence: copper-specific claims such as low copper raising cholesterol are filled in the animal column but mixed or open in the human columns, while mainstream claims about oxidised LDL, foam cells and iron in plaque are filled, and the claims that high body iron predicts heart disease, that low copper causes heart disease in people and that calcium follows iron in plaque are open rings in the human observational column. THE COPPER THEORY, BY KIND OF EVIDENCE filled dot = supports · open ring = null or contrary · blank = no evidence cited ANIMAL HUMAN TRIAL HUMAN OBSERVATIONAL LAB low copper raises blood cholesterol high zinc-to-copper ratio raises cholesterol low copper lowers copper enzymes (SOD, ceruloplasmin) oxidised LDL turns macrophages into foam cells free copper oxidises LDL iron accumulates in plaque high stored iron predicts heart disease low copper causes heart disease in people calcium follows iron in plaque 1 man / 12 women The copper-specific claims fill the animal and lab columns; the human columns hold the open rings.

How each cell was filled, row by row: cholesterol — rats (Klevay 1973; Allen and Klevay 1978; Reiser 1983), the 1984 single-participant rise and the 1996 null result in twelve women; zinc-to-copper ratio — rats only; copper enzymes — the 1984 and 1996 human studies; oxidised LDL and foam cells — the Steinberg and Esterbauer reviews (laboratory work, and oxidised LDL found in human and animal plaques); free copper oxidising LDL — laboratory; iron in plaque — the 2004 surgical-plaque study and the 2003 autopsy study, which also induced plaque bleeding in rabbits; stored iron — the 1999 meta-analysis (null); low copper causing heart disease — the US survey cohort and the 2018 meta-analysis, which point the other way; calcium following iron — the 2004 plaque study (no correlation). No human trial has tested whether giving copper lowers cholesterol or heart disease, which is why that column is mostly blank.

Back to Table of Contents

10. Practical: Food Copper, Labs and Safety

Get copper from food first. The adult Recommended Dietary Allowance for copper is 900 micrograms (0.9 mg) a day, and the Tolerable Upper Intake Level — the most the US National Academies consider safe to take every day long-term — is 10 mg a day. A USDA metabolic-ward study found that 0.57 mg a day for 105 days was enough to deplete copper enzymes in older women, so very low intakes are a real possibility on a narrow diet. Rich whole-food sources, per 100 grams (USDA FoodData Central values): pan-fried beef liver about 14.6 mg; cooked eastern oysters about 4.4 mg; raw cashews about 2.2 mg. Sesame seeds, dark chocolate, lentils and mushrooms add useful amounts. A balanced diet built on these foods supplies copper without any pill.

Liver is a once-a-week food, not a daily one. A single 100-gram portion of beef liver already carries more copper than the 10 mg upper limit, and liver is also extremely high in preformed vitamin A (retinol). Retinol in large amounts can cause birth defects, so anyone who is pregnant or trying to conceive should keep liver portions small and occasional and not combine liver with vitamin A supplements or cod liver oil without medical advice.

Watch zinc supplements. The one part of Klevay’s theory that is firmly established in people is that long-term high-dose zinc can cause copper deficiency, with anaemia and low white cells. If you take zinc above food amounts for months, copper status is worth checking.

Do not dose copper on a theory. Copper supplements beyond food can push intake toward the upper limit, and people with Wilson’s disease — an inherited inability to excrete copper — must avoid extra copper altogether. A ceruloplasmin and serum copper test can be ordered, but as the 1996 study showed, both are insensitive to mild depletion, and both rise with inflammation.

Do not drop proven heart care. High LDL cholesterol remains a well-established, treatable cause of heart disease. Nothing in the copper literature shows that copper replaces a statin, blood pressure control or stopping smoking. And if a blood test shows true iron-deficiency anaemia, the right response is to find the cause (for example bleeding from the gut or heavy periods), not simply to withhold iron because of the iron hypothesis.

Back to Table of Contents

Key Research Papers

  1. Klevay LM (1973). Hypercholesterolemia in rats produced by an increase in the ratio of zinc to copper ingested. Am J Clin Nutr. — PubMed PMID: 4754665
  2. Klevay LM (1975). Coronary heart disease: the zinc/copper hypothesis. Am J Clin Nutr. — PubMed PMID: 1146731
  3. Allen KG, Klevay LM (1978). Cholesterolemia and cardiovascular abnormalities in rats caused by copper deficiency. Atherosclerosis. — PubMed PMID: 629827
  4. Reiser S, Ferretti RJ, Fields M, Smith JC Jr (1983). Role of dietary fructose in the enhancement of mortality and biochemical changes associated with copper deficiency in rats. Am J Clin Nutr. — PubMed PMID: 6881079
  5. Klevay LM, Inman L, Johnson LK, Lawler M, Mahalko JR, Milne DB, Lukaski HC, Bolonchuk W, Sandstead HH (1984). Increased cholesterol in plasma in a young man during experimental copper depletion. Metabolism. — PubMed PMID: 6503710
  6. Reiser S, Smith JC Jr, Mertz W, Holbrook JT, Scholfield DJ, Powell AS, Canfield WK, Canary JJ (1985). Indices of copper status in humans consuming a typical American diet containing either fructose or starch. Am J Clin Nutr. — PubMed PMID: 4025196
  7. Milne DB, Nielsen FH (1996). Effects of a diet low in copper on copper-status indicators in postmenopausal women. Am J Clin Nutr. — PubMed PMID: 8602593
  8. Steinberg D, Parthasarathy S, Carew TE, Khoo JC, Witztum JL (1989). Beyond cholesterol. Modifications of low-density lipoprotein that increase its atherogenicity. N Engl J Med. — PubMed PMID: 2648148
  9. Esterbauer H, Gebicki J, Puhl H, Jürgens G (1992). The role of lipid peroxidation and antioxidants in oxidative modification of LDL. Free Radic Biol Med. — PubMed PMID: 1398217
  10. Sullivan JL (1981). Iron and the sex difference in heart disease risk. Lancet. — PubMed PMID: 6112609
  11. Danesh J, Appleby P (1999). Coronary heart disease and iron status: meta-analyses of prospective studies. Circulation. — PubMed PMID: 10027804
  12. Kolodgie FD, Gold HK, Burke AP, Fowler DR, Kruth HS, Weber DK, Farb A, Guerrero LJ, Hayase M, Kutys R, Narula J, Finn AV, Virmani R (2003). Intraplaque hemorrhage and progression of coronary atheroma. N Engl J Med. — PubMed PMID: 14668457
  13. Stadler N, Lindner RA, Davies MJ (2004). Direct detection and quantification of transition metal ions in human atherosclerotic plaques: evidence for the presence of elevated levels of iron and copper. Arterioscler Thromb Vasc Biol. — PubMed PMID: 15001454
  14. Ford ES (2000). Serum copper concentration and coronary heart disease among US adults. Am J Epidemiol. — PubMed PMID: 10905530
  15. Chowdhury R, Ramond A, O’Keeffe LM, et al. (2018). Environmental toxic metal contaminants and risk of cardiovascular disease: systematic review and meta-analysis. BMJ. — PubMed PMID: 30158148
  16. Trumbo P, Yates AA, Schlicker S, Poos M (2001). Dietary reference intakes: vitamin A, vitamin K, arsenic, boron, chromium, copper, iodine, iron, manganese, molybdenum, nickel, silicon, vanadium, and zinc. J Am Diet Assoc. — PubMed PMID: 11269606

PubMed Topic Searches

  1. PubMed: Copper deficiency and cholesterol
  2. PubMed: Zinc/copper ratio and coronary heart disease
  3. PubMed: Serum copper and cardiovascular disease

Back to Table of Contents

Connections

Back to Table of Contents