In summary

The Western diet often is low in Cu⁽ Reference Klevay ^{22 )} with substantial numbers of individuals eating less than recommended amounts; Cu intakes in nutrition surveys are falsely high because calculated values greatly exceed values from chemical analyses. Cu deficiency is the only nutritional insult to experimental animals that elevates cholesterol, blood pressure, homocysteine, isoprostanes and uric acid, has adverse effects on arteries and electrocardiograms, decreases dehydroepiandrosterone, impairs glucose tolerance and paraoxonase activity, promotes thrombosis and oxidative damage, and to which males respond differently than females.

More than eighty anatomical, chemical and physiological similarities between animals deficient in Cu and people with IHD have been identified. Most of these similarities have been tabulated⁽ Reference Klevay ^{23 ,} Reference Klevay ^{24 )}. Data also were tabulated from more than sixty medical publications about more than 2500 people with poor Cu nutriture in cardiovascular, musculoskeletal and neural diseases with more than 1000 people benefitting from supplements containing Cu⁽ Reference Klevay ^{22 ,} Reference Klevay ^{25 )}.

Anatomy, chemistry, pathology and physiology

More than ten of the more obvious clinical aspects of IHD are enumerated above. Examination of the anatomy, chemistry, pathology and physiology of atherosclerosis and IHD reveals a bewildering array of other, more subtle signs. Arteries show smooth muscle cell proliferation and hearts have ventricular aneurysms. There is decreased leucocyte Cu and increased cardiac Na along with heart blocks and hearing loss, inter alia. Some of these findings are exacerbated by salt or pregnancy; some are mitigated by aspirin or beer⁽ Reference Klevay ^{23 ,} Reference Klevay ^{24 )}. Cu deficiency has produced all of these findings in experimental animals. No other nutritional insult has produced such multifarious pathology in experimental animals.

Cholesterol probably is the dietary component tested most widely in the study of atherosclerosis and lipid metabolism. All of the thousands of experiments done since 1913⁽ Reference Anitschkow and Chalatow ^{26 ,} Reference Wacker and Hueck ^{27 )} have Cu metabolism as a hidden variable because dietary cholesterol, or cholesterol plus cholic acid, induces Cu deficiency⁽ Reference Klevay ^{28 ,} Reference Klevay ^{29 )} according to several experiments from four, independent laboratories.

Copper nutriture

According to the Oxford Textbook of Medicine ⁽ Reference Golden ^{30 )}, low nutrient intake can reduce nutrient concentrations in tissues and compromise metabolic pathways. Tabulations⁽ Reference Klevay ^{22 )} of medical articles are available on impaired Cu nutriture in cardiovascular, mostly ischaemic, diseases. Eleven articles on 254 individuals reveal low tissue Cu. Eight articles on 876 individuals reveal low activities of Cu-dependent enzymes indicating compromised metabolic pathways.

In addition, abnormal cardiac physiology is associated with poor Cu status. Oster et al. ⁽ Reference Oster, Dahm and Oelert ^{31 )} reported that cardiac output correlates positively with cardiac Cu in patients with CHD. Erythrocyte superoxide dismutase was decreased in patients examined 6 months after myocardial infarction⁽ Reference Pytel, Olszewska-Banaszczyk and Koter-Michalak ^{32 )} and in patients with coronary slow flow⁽ Reference Kaplan, Meric and Acar ^{33 )}. Cu/Zn superoxide dismutase protects against oxidative damage and depends on Cu for activity⁽ Reference Owen ^{34 )}.

Nutrition surveys

Nutrient intake data in large nutrition surveys are based on the memory of individuals interviewed about amounts and types of foods recently consumed. Their intakes are calculated from data obtained from food tables⁽ Reference Klevay ^{22 )}. Eleven, peer-reviewed articles have been found in which the authors obtained the foods identified by survey subjects and measured Cu in them by chemical analysis. Intakes calculated from food tables exceeded these measured intakes in ten of the articles (P=0·0054 by sign test)^{( 35 )}. Thus there is a systematic, or determinate, error in the estimation of Cu intakes by calculation giving results that are falsely high by 77 % for Western diets⁽ Reference Klevay ^{22 )}. Chiplonkar & Agte⁽ Reference Chiplonkar and Agte ^{36 )}, Nakasuta et al. ⁽ Reference Nakatsuka, Shimbo and Watanabe ^{37 )} and Rahmdel et al. ⁽ Reference Rahmdel, Abdollahzadeh and Mazloomi ^{38 )} published the more recent reports since these concepts were considered in greater detail⁽ Reference Klevay ^{22 ,} Reference Klevay ^{39 )}.

The RDA for the USA and Canada and estimated average requirement (EAR) for Cu are 0·9 and 0·7 mg daily, respectively^{( 40 )}. Correction of published intakes (Third National Health and Nutrition Examination Survey (NHANES III)) for the 77 % error reveals that nearly one fourth of the women do not achieve the EAR and only half the men achieve the RDA⁽ Reference Klevay ^{22 )}. Even fewer individuals achieve the higher recommendations (>1·0 mg/d) of the UK, European Commission, or Australia and New Zealand⁽ Reference Klevay ^{22 )}.

Decreasing dietary copper

Measurements of Cu in common foods published in 1942 were compared with measurements published in 1966; it was concluded that Cu had decreased in a quarter of a century⁽ Reference Klevay and Forbush ^{41 )}. The most robust data on this decline in food Cu⁽ Reference Fan, Zhao and Fairweather-Tait ^{42 )} are from analyses made on archived wheat grain; a single analytical method revealed that Cu in the grain decreased since the mid 1960s.

In addition there was a decline in Cu in fruits and vegetables in the UK between 1940 and 1991⁽ Reference Thomas ^{43 )}, in fruits and vegetables in the UK between the 1930s and the 1980s⁽ Reference White and Broadley ^{44 )} and in the USA since the 1930s⁽ Reference White and Broadley ^{44 )}. Cu also declined in meat and some milk and cheese products in the UK⁽ Reference Thomas ^{43 )} between 1940 and 1991. Fan et al. ⁽ Reference Fan, Zhao and Fairweather-Tait ^{42 )} also quote two surveys from the UK showing that Cu intakes fell in the last 15 years of the 20th century in agreement with tabulated percentile data from three surveys in the USA⁽ Reference Klevay ^{22 , 40 )}. Even though some of these comparisons are separated widely in time, competent chemists probably produced accurate data earlier because analytical standards of excellence have not changed in a century. Thus IHD has increased while dietary Cu has decreased in agreement with Hill’s viewpoint on temporality.

People deficient in copper

Approximately thirty-five men and women of middle age have been depleted of Cu in metabolic wards where experimental conditions were controlled carefully and dietary Cu was measured by analytical chemistry. Cardiac function was monitored with many Holter cardiograms. Data⁽ Reference Lukaski, Klevay and Milne ^{45 –} Reference Reiser, Powell and Yang ^{48 )} are shown in Table 1.

Table 1 Insufficient amounts of dietary copper for adult humansFootnote *

* Modified from Klevay⁽ Reference Klevay ^{108 )}.

Hypertensive responses to sustained handgrip exercise in women depleted of Cu fell below control values after repletion with Cu⁽ Reference Lukaski, Klevay and Milne ^{45 )}. Cu depletion increased the concentration of plasma glucose by an average of 38 mg/dl (2·11 mmol/l) at each point on intravenous glucose tolerance tests done on two men; supplementation generally decreased these values⁽ Reference Klevay, Canfield and Gallagher ^{46 )}.

Cholesterol in plasma of a young man was increased significantly by Cu depletion. He was supplemented with Cu when we noticed a short run of ventricular tachycardia; cholesterol then fell significantly and precipitously to less than the initial value⁽ Reference Klevay, Inman and Johnson ^{47 )}.

In a similar experiment, Cu depletion increased LDL-cholesterol and decreased HDL-cholesterol. Four men were removed from the study because of one myocardial infarction, two severe tachycardias and an intermittent, second-degree heart block that were recorded⁽ Reference Reiser, Smith and Mertz ^{49 )}. No problems related to heart function were found in the previous 337 subjects observed during nutritional studies not involving Cu at that centre. Overall, lipoprotein cholesterols improved over control values on supplementation⁽ Reference Reiser, Powell and Yang ^{48 )}.

These lipid effects probably were driven by hydroxymethylglutaryl-coenzyme A reductase (EC 1.1.1.34). Three independent laboratories have found its activity to be increased in rats deficient in Cu⁽ Reference Valsala and Kurup ^{50 –} Reference Kim, Chao and Allen ^{52 )}. Activity of other enzymes germane to heart disease mechanisms generally is decreased⁽ Reference Klevay ^{53 )}; some of these are mentioned here. No data have been found related to mechanisms for altered glucose metabolism or increased blood pressure. Some other interrelationships between Cu and heart disease are shown in Fig. 1.

Fig. 1 Metabolic pathways affected by copper deficiency. For metabolic pathways involving homocysteine, also see the text. * Pathways inhibited by copper deficiency, in the figure and below. † Copper metalloenzyme, in the figure and below. The storm cloud and lightening bolts identify the inhibition of lysyl oxide† by homocysteine thiolactone. Superoxide dismutase*† (EC 1.15.1.1) defends against oxidative damage and is inhibited by high homocysteine. Homocysteine thiolactone and methyltetrahydrofolate are synthesised from homocysteine and 5,10-methylenetetrahydrofolate by methionyl-tRNA synthase (EC 6.1.1.10) and 5,10-methylenetetrahydrofolate reductase (EC 1.1.99.15), respectively. This figure was first published in Klevay⁽ Reference Klevay ^{92 )} and is reproduced here with permission.

More than 70 % of chemically analysed, daily diets of randomly selected people in Baltimore, MD, contained less than 1 mg Cu⁽ Reference Pang, MacIntosh and Ryan ^{54 )}. The people above responded to diets similarly low in Cu with potentially hazardous changes in risk factors that disappeared on Cu repletion. Production and elimination of abnormalities in depletion experiments are the sine qua non of measuring nutritional requirements.

Supplementation with copper

Cu supplementation⁽ Reference Klevay ^{22 )} of people has abolished cardiac arrhythmia⁽ Reference Spencer ^{55 )}, protected erythrocytes from oxidation⁽ Reference Rock, Mazur and O’Connor ^{56 )}, increased activity of superoxide dismutase⁽ Reference Bügel, Harper and Rock ^{57 )} and decreased plasma homocysteine (see below). Recently DiSilvestro et al. ⁽ Reference DiSilvestro, Selsby and Siefker ^{58 ,} Reference DiSilvestro, Joseph and Zhang ^{59 )} found increased serum caeruloplasmin in young women and middle-aged men and women given Cu. In the latter study superoxide dismutase in erythrocytes also was increased. A micronutrient supplement containing Cu improved ventricular function in people with heart failure⁽ Reference Klevay ^{60 )}. Cu supplementation (along with Zn) improved survival of people in a long-term, double-blind study of eye disorders⁽ Reference Klevay ^{22 ,} Reference Clemons, Kurinij and Sperduto ^{61 )}. According to Golden⁽ Reference Golden ^{30 )}, a beneficial effect on a metabolic pathway or functional system from nutrient replacement is a sign of deficiency.

There are no large trials of Cu supplementation to reduce heart disease similar to those with vitamins to reduce homocysteine. It is possible that the failure of these trials to produce clinical benefit is because only vitamins that lower homocysteine were tested⁽ Reference Klevay ^{62 )}. If Cu had been included in the supplemental mixture, it may have improved health by an unforeseen mechanism (see text below and Fig. 1).

Iron as a copper antagonist

Fe is known to interfere with Cu utilisation. Some early data were reviewed and an experiment revealed that higher dietary Fe can increase the dietary requirement for Cu⁽ Reference Klevay ^{63 )}. High Fe intakes interfere with Cu absorption in infants^{( 40 )} and adults⁽ Reference Klevay ^{21 )}. Evidence that Fe overload interferes with Cu utilisation is increasing⁽ Reference Videt-Gibou, Belliard and Bardou-Jacquet ^{64 ,} Reference Videt-Gibou, Belliard and Rivalan ^{65 )}. Haemochromatosis decreases immunoreactive caeruloplasmin⁽ Reference Cairo, Conte and Bianchi ^{66 )}. Sometimes this interference is beneficial as when haemochromatosis delays the onset of Wilson’s disease to the age of 50 years⁽ Reference Dib, Valsesia and Malinge ^{67 )}.

The onset of Fe supplementation of pregnant women was established by examination of Williams Obstetrics and advertisements in the American Journal of Obstetrics and Gynecology from the 1930s. John W. Williams was the founder of academic obstetrics in the USA⁽ Reference Dunn ^{68 )}. His textbook first was published in 1903; several editions were examined.

The 7th edition recommends ‘iron medication’ as being ‘very satisfactory’ for the ‘microcytic hypochromic anaemia of pregnancy’. Because the 6th edition does not mention Fe, one can infer that Fe supplementation during pregnancy began between 1930 and 1936. The 8th edition mentions ferrous sulfate specifically. The 24th edition⁽ Reference Cunningham, Leveno and Bloom ^{69 )} recommends ‘27 mg of elemental iron supplement be given daily to pregnant women’. Thus while dietary Cu was falling (above), use of a Cu antagonist in pregnancy was increasing, from no women in 1930, to some women in 1936, to all women now. Advertisement of Fe supplements began in 1933; a Cu–Fe supplement advertised for the anaemias of pregnancy in 1934 seems to have been abandoned.

Iron overload

Sullivan suggests that Fe overload contributes to heart disease risk⁽ Reference Sullivan ^{71 –} Reference Sullivan ^{73 )}. It was suggested that people with Fe overload may benefit from Cu supplements because pathology related to CVD is resistant to phlebotomy⁽ Reference Klevay ^{63 )}. Finland has abandoned Fe fortification of food, probably because of the associations of Fe with heart disease risk⁽ Reference Klevay ^{63 )}. High intakes of haeme Fe were associated with increased risk of fatal, myocardial infarction in comparison with low intakes⁽ Reference Kaluza, Larsson and Hakansson ^{74 )}.

Pregnant women must consume an extra 0·75 mg of Cu daily to avoid depletion while supplying the newborn, placenta, etc., with Cu⁽ Reference Klevay and Medeiros ^{75 )}. This amount plus the EAR (0·70 mg) is nearly 1·5 mg; approximately 60 % of 849 chemically analysed diets (from Belgium, Canada, UK and USA) contain less than this daily amount⁽ Reference Klevay, Buchet and Bunker ^{76 )}. Supplementation with large amounts of Fe (a Cu antagonist, above), particularly if dietary Cu is too low, may harm both the mother and the baby later in life. Perhaps Fe supplements for pregnancy should contain Cu as in the mid-1930s (see above).

Fetal programming

Barker suggests that undernutrition in utero leads to IHD⁽ Reference Barker ^{77 )}. In brief, being small at birth and not catching up by the first birthday (or then being too fat) predisposes to diabetes mellitus or hypertension in middle age.

Barker suggests that ‘deficiencies in specific nutrients which influence fetal growth, including vitamins A, C, and D, folate, Fe and Zn’ may be important. Studies of nutrients as related to developmental models for adult disease seem to be few, but some data on Cu are available (see below). He cites O’Dell et al. ⁽ Reference O’Dell, Kilburn and McKenzie ^{78 )}: rats made Cu deficient during gestation developed increased alveolar size, which persisted even with Cu repletion for several months⁽ Reference O’Dell, Kilburn and McKenzie ^{78 )}.

Recent experiments on animals

Reproductive failure in animals from Cu deficiency has been known for more than a half century^{( 79 )}. In female rats deficiency results in reproductive failure from fetal death on the 13th day of pregnancy⁽ Reference Davis and Mertz ^{80 )}. Thus finding a low concentration of dietary Cu sufficiently high to provide offspring for study may be difficult. W. T. Johnson is prominent in this area; some anatomical, chemical and physiological abnormalities in pups born of deficient dams are resistant to extra dietary Cu across generations.

Female rats were fed a diet low in Cu for 3 weeks before mating and during gestation and lactation. Cardiac cytochrome c oxidase activity remained decreased in pups despite consumption of a diet adequate in Cu for 6 weeks⁽ Reference Johnson and Brown-Borg ^{81 )}. According to Owen, the cardiac enzyme is particularly resistant to Cu therapy of deficiency⁽ Reference Owen ^{82 )}.

A diet low in Cu was fed to female rats from 3 weeks before conception until 3 weeks after birth. Cytochrome c oxidase activity was decreased in cardiac mitochondria of adult offspring despite their being fed a diet with normal Cu from age 21 d to 290 d. Electron microscopy of cardiac myocytes was altered and H₂O₂ generation was increased in the replete animals⁽ Reference Johnson and Newman ^{83 )}.

A diet low in Cu was fed to female rats for 3 weeks before conception, during gestation and lactation⁽ Reference Anderson and Johnson ^{84 )} in a multigenerational study designed to evaluate effects of dietary remediation. Vascular functional responses in mesenteric arteries were examined using several agents; some sexual differences were noted. Impaired vascular function occurred in the second generation. Some effects were direct, from parents exposed to postnatal deficiency, and some were indirect from the initial in utero deficiency in spite of seemingly adequate Cu intakes in the first generation.

Although trans-generational effects of nutritional deficiency may not be unique to Cu, these experiments show that adverse effects of Cu deficiency in utero can lead to persistent pathology in offspring fed seemingly adequate amounts of Cu. Some adverse effects can pass to the next generation. The following observations on people are consonant with Cu deficiency in utero in people that Barker suggests will be at risk for heart disease in middle age⁽ Reference Barker ^{77 )}.

Observations on people

Measurement of Cu in maternal plasma during pregnancy is assumed to be nutritionally irrelevant to fetal development because it increases markedly when the unborn are receiving substantial amounts of Cu from the mothers. Measurements on cord or infant blood, however, may be useful.

Krishnamachari & Rao⁽ Reference Krishnamachari and Rao ^{85 )} found that Cu and caeruloplasmin were halved in cord blood of infants born of malnourished mothers. Small, newborn infants had less serum Cu than normal⁽ Reference Artal, Burgeson and Fernandez ^{86 )}. Plasma Cu also was decreased in cord blood of infants small for gestational age at term and in premature infants whose size was appropriate for gestational age⁽ Reference Goel and Misra ^{87 )}.

Superoxide dismutase activity was decreased in cord blood of small-for-gestational-age and preterm infants⁽ Reference Phylactos, Leaf and Costeloe ^{88 )}. Premature infants have lower superoxide dismutase activity in erythrocytes and plasma after 100 d of life than term infants⁽ Reference Nassi, Ponziani and Becatti ^{89 )}; premature placentas are low in Cu and Cu-dependent enzymes⁽ Reference Zadrozna, Gawlik and Nowak ^{90 )}.

These small infants, with impaired Cu status, are those that Barker suggests are likely to become hypertensive and/or diabetic in middle age. Diabetes mellitus and hypertension are major risk factors for IHD.

Homocysteine

McCully⁽ Reference McCully ^{91 )}, cited by Klevay⁽ Reference Klevay ^{92 )}, suggests hyperhomocysteinaemia destroys arteries and promotes IHD. High homocysteine concentrations lead to increased homocysteine thiolactone, an irreversible inhibitor of lysyl oxidase, which depends on Cu to initiate the cross-linking of collagen and elastin in arteries⁽ Reference Klevay ^{92 )}. Homocysteine thiolactone hydrolase, also known as paraoxonase, destroys this lactone. Its activity is decreased in rats deficient in Cu and low activity is associated with increased risk of IHD⁽ Reference Klevay ^{1 )}. Cu supplements can decrease homocysteine concentrations in men⁽ Reference Tamura and Turnlund ^{93 )}.

Feeding homocysteine to rats decreases the activity of superoxide dismutase and disrupts Cu utilisation; rats deficient in Cu have increased plasma homocysteine and decreased activity of methionine synthase⁽ Reference Klevay ^{92 )}. Methionine synthase may be a Cu enzyme. Some of these interrelationships are shown in Fig. 1.