Function

The biochemical role for copper is primarily catalytic, with many copper metalloenzymes acting as oxidases to achieve the reduction of molecular oxygen. Many copper metalloenzymes have been identified in humans (da Silva and Williams, 1991; Harris, 1997).

Amine oxidases participate in important reactions that have markedly different effects. Diamine oxidase inactivates histamine released during allergic reactions. Monoamine oxidase (MAO) is important in serotonin degradation to excretable metabolites and in the metabolism of catecholamines (epinephrine, norepinephrine, and dopamine). MAO inhibitors are used as antidepressant drugs. Lysyl oxidase uses lysine and hydroxylysine found in collagen and elastin as substrates for posttranslational processing to produce cross-linkages needed for the development of connective tissues, including those of bone, lung, and the circulatory system.

Ferroxidases are copper enzymes found in plasma, with a function in ferrous iron oxidation (Fe²⁺→Fe³⁺) that is needed to achieve iron's binding to transferrin (Linder and Hazegh-Azam, 1996). Ferroxidase I, also called ceruloplasmin, is the predominant copper protein in plasma and may also have antioxidant functions. Defects in ceruloplasmin function produce cellular iron accumulation, a result that supports its ferroxidase role (Harris and Gitlin, 1996). Ferroxidase II is found in human plasma, but it may have a role in iron metabolism in specific cellular sites. A transmembrane copper-containing protein (hephaestatin) with ferroxidase activity has been described (Pena et al., 1999; Vulpe et al., 1999). Cytochrome c oxidase is a multisubunit enzyme in mitochondria that catalyzes reduction of O₂ to H₂O. This establishes a high energy proton gradient required for adenosine triphosphate (ATP) synthesis. This copper enzyme is particularly abundant in tissues of greatest metabolic activity including heart, brain, and liver. Dopamine β monooxygenase uses ascorbate, copper, and O₂ to convert dopamine to norepinephrine, a neurotransmitter, produced in neuronal and adrenal gland cells. Dopa, a precursor of dopamine, and metabolites used in melanin formation are oxidatively produced from tyrosine by the copper enzyme tyrosinase. α-Amidating monooxygenase (α-AE), also called peptidylglycine α-AE, uses copper and ascorbate to remove two carbons from a C-terminal glycine of peptides, thus generating an amide. A number of peptide hormones are posttranslationally modified by α-AE (Harris, 1997).

Two forms of superoxide dismutase are expressed in mammalian cells, a mangano and cupro/zinc form (Harris, 1997). Copper/zinc superoxide dismutase (Cu/Zn SOD) uses two copper atoms for conversion of the superoxide anion (O₂^-+) to H₂O₂ and O₂. Zinc atoms have a structural role in the enzyme. The enzyme is localized in the cytosol and, along with the mitochondrial manganese-containing form, provides a defense against oxidative damage from superoxide radicals that, if uncontrolled, can lead to other damaging reactive oxygen species. Mutations in the Cu/Zn SOD gene, which alter the protein's redox behavior, produce amyotrophic lateral sclerosis (Lou Gehrig's disease).

These are the principal copper metalloenzymes found in humans. There is substantial documentation from animal studies that diets low in copper reduce the activities of many of these copper metalloenzymes. Activities of some copper metalloenzymes have been shown to decrease in human copper depletion (Milne, 1994; Turnlund, 1999). Physiologic consequences resulting from copper deficiency include defects in connective tissue that lead to vascular and skeletal problems, anemia associated with defective iron utilization, and possibly specific aspects of central nervous system dysfunction (Harris, 1997; Turnlund, 1999). Some evidence suggests that immune and cardiac dysfunction occurs in experimental copper deficiency and the development of such signs of deficiency has been demonstrated in infants (Graham and Cordano, 1969; Olivares and Uauy, 1996; Turnlund, 1999).

Physiology of Absorption, Metabolism, and Excretion

Metabolism of copper in humans relies on the intestine for control of homeostasis as the capacity for renal copper excretion is limited. Nearly two-thirds of the body copper content is located in skeleton and muscle, but studies with stable isotopes have shown that the liver is a key site in maintaining plasma copper concentrations (Olivares and Uauy, 1996; Turnlund et al., 1998). Copper has a higher binding affinity for proteins than all other divalent trace elements (da Silva and Williams, 1991). Consequently, precise control of intracellular copper trafficking is needed to regulate how it is donated to appropriate sites.

Copper absorption occurs primarily in the small intestine. Some absorption may occur in the stomach where the acidic environment promotes copper solubility by dissociation from copper-containing macromolecules derived from dietary sources (Harris, 1997; Turnlund, 1999). Both saturable-mediated and nonsaturable-nonmediated (possibly paracellular) transepithelial copper movements have been reported. The Menkes P-type ATPase (MNK; ATP7A) is believed to be responsible for copper trafficking to the secretory pathway for efflux from cells, including enterocytes (Harris and Gitlin, 1996). A defective MNK gene causes Menkes' disease, which is characterized by reduced copper absorption and placental copper transport. The extent of copper absorption varies with dietary copper intake (Turnlund, 1998). It ranges from over 50 percent at an intake of less than 1 mg/day to less than 20 percent above 5 mg/day. About 35 percent of a 2 mg/day intake is absorbed and is transported via the portal vein to the liver, bound to albumin, for uptake by liver parenchymal cells.

Biliary copper excretion is adjusted to maintain balance. Copper is released via plasma to extrahepatic sites where up to 95 percent of the copper is bound to ceruloplasmin (Turnlund, 1999). The biological role of ceruloplasmin in copper metabolism has been widely investigated. The autosomal recessive disorder in humans, aceruloplasminemia, does not produce abnormal copper metabolism, thus contradicting a role for the protein in copper delivery to cells. However, this genetic defect results in tissue iron accumulation, supporting the protein's role in cellular iron release. Other P-type ATPases (e.g., Wilson, NND, ATP7B) are responsible for copper trafficking to the secretory pathway for ceruloplasmin synthesis or for endosome formation before transport into the bile (Harris and Gitlin, 1996; Pena et al., 1999). Mutations of this copper-transporting ATPase result in cellular copper accumulation called Wilson's disease. Urinary copper excretion is normally very low (< 0.1 mg/day) over a wide range of dietary intakes (Turnlund, 1999). As with other trace elements, renal dysfunction can lead to increased urinary losses.

Clinical Effects of Inadequate Intake

Frank copper deficiency in humans is rare, but has been found in a number of special conditions. It has been observed in premature infants fed milk formulas, in infants recovering from malnutrition associated with chronic diarrhea and fed cow's milk (Shaw, 1992), and in patients with prolonged total parenteral nutrition (Fujita et al., 1989). In these cases, serum copper and ceruloplasmin concentrations were as low as 0.5 μmol and 35 mg/L, respectively, compared to reported normal ranges of 10 to 25 μmol/L for serum copper concentration and 180 to 400 mg/L for ceruloplasmin concentration (Lentner, 1984). Supplementation with copper resulted in rapid increases in serum copper and ceruloplasmin concentrations.

Symptoms accompanying the copper deficiency included normocytic, hypochromic anemia, leukopenia, and neutropenia (Fujita et al., 1989). Osteoporosis was observed in copper-deficient infants and growing children.

Copper deficiency developed in six severely handicapped patients between the ages of 4 and 24 years who were fed an enteral diet containing 15 μg of copper/100 kcal for 12 to 66 months (Higuchi et al., 1988). Their serum copper concentrations ranged from 0.9 to 7.2 μmol/L and ceruloplasmin concentrations ranged from 30 to 125 mg/L. Two patients had neutropenia, one had macrocytic, normochromic anemia, and some had bone abnormalities including reduced bone density. Neutrophil counts normalized and bone abnormalities improved after copper supplementation. If the copper intake of these patients is extrapolated to adults on the basis of caloric intake, copper deficiency might be expected to develop in adults at an intake of 440 μg/2,900 kcal for men and 290 μg/1,900 kcal for women. This deduction is consistent with a study in which healthy young men who were fed a diet containing 380 μg/day of copper for 42 days had a decline in serum copper and ceruloplasmin concentrations and then an increase with copper repletion (Turnlund et al., 1997). Although serum copper and ceruloplasmin concentrations of these men did not fall to the deficient range in 42 days and clinical symptoms did not appear, these effects might be expected had the low copper diet been continued. In a number of other studies at higher levels of copper intake (i.e., at 600 μg/day and above), serum copper and ceruloplasmin concentrations did not decline significantly (Milne, 1998; Turnlund et al., 1990).

Results of depletion studies in laboratory animals have led to interest in a number of conditions in humans that may be associated with marginal copper intake over a long period. Insufficient data are available at this time to establish whether these conditions are related to dietary copper.

A report of increased blood cholesterol concentrations in one young man consuming 830 μg/day of copper (Klevay et al., 1984) suggested that elevated blood cholesterol concentration may be associated with marginal amounts of dietary copper. This effect was not observed in other subjects or in a number of other studies with this or lower levels of dietary copper. In one study, blood cholesterol concentration decreased with lower dietary copper (Milne and Nielsen, 1996), and in a copper supplementation study investigators found increased blood cholesterol concentrations with supplementation (Medeiros et al., 1991).

Heart beat irregularities were reported in some studies, and investigators linked them to dietary copper intake (Milne, 1998). However, heart beat irregularities are common in normal, healthy people, and other studies with lower copper intake demonstrated that such irregularities, monitored during copper depletion and repletion, were common at all intake levels of dietary copper (Turnlund et al., 1997). Myocardial disease occurs in severely deficient weanling rats, and one investigator has hypothesized that ischemic heart disease is related to marginal copper status (Klevay, 1989). However, the myocardial changes observed in copper-deficient animals are very different from those of ischemic heart disease in humans (Danks, 1988). In severely deficient animals, the myocardium is hypertrophied and may rupture. Coronary artery resistance is decreased in copper-deficient animals, but it is increased in ischemic heart disease.

Several other clinical observations deserve further investigation, but there is insufficient evidence to link them to marginal copper status. Glucose tolerance was lower in two of a group of eight men consuming 80 μg/day of copper than in men consuming higher levels of copper (Klevay et al., 1986), but similar observations have not been reported at lower intakes of copper in other studies. One study reported a negative correlation between ceruloplasmin concentration and blood pressure during a hand grip exercise (Lukaski et al., 1988), but the link between blood pressure and dietary copper has not been investigated further in humans. An index of immune function declined in a depletion study with copper intakes of 380 μg/day that resulted in decreases in indexes of copper status, but other indexes of immune function did not decline and repletion did not result in reversal of the change (Kelley et al., 1995). Changes in blood clotting factors V and VIII were observed in one study with copper intakes of 570 μg/day (Milne and Nielsen, 1996). The role of copper as an antioxidant has led to interest in the possibility that copper deficiency impairs antioxidant status (Johnson et al., 1992). A report of changes in some, but not other, markers of bone metabolism with a dietary copper intake of 700 μg/day deserves further investigation (Baker et al., 1999). Changes in catecholamine metabolism have been investigated, but results are inconsistent (Bhathena et al., 1998).

Serum Copper Concentrations

Serum copper concentration is a reliable indicator of copper deficiency, falling to very low concentrations in copper-deficient individuals. The lower end of the normal range for serum copper concentration is reported to be 10 μmol/L, but serum copper concentrations were considerably lower than this when cases of copper deficiency were discovered. Serum copper concentration returns to normal within a few days of copper supplementation (Danks, 1988). While serum copper concentration is an index of copper deficiency, it does not reflect dietary intake except when intake is below a certain level. Above this level, supplementation with copper does not increase serum copper concentration. Serum copper concentration increases under a number of conditions due to increased concentrations of ceruloplasmin.

Ceruloplasmin Concentration

Ceruloplasmin concentration is also a reliable indicator of copper deficiency. Ceruloplasmin carries between 60 and 95 percent of serum copper, and changes in serum copper concentration usually parallel the ceruloplasmin concentration in the blood. Ceruloplasmin, too, falls to low concentrations with copper deficiency, far below the lower end of the normal range of 180 mg/L, and it responds quickly to repletion (Danks, 1988). Ceruloplasmin does not respond to dietary intake, unless intake is very low. The dietary copper intake at which ceruloplasmin concentration no longer increases in response to increased dietary copper might be considered the copper requirement for ceruloplasmin synthesis. Ceruloplasmin is an acute phase protein and increases markedly with a number of diseases, including liver disease, malignancy, inflammatory diseases, myocardial infarction, and a variety of infectious diseases (Mason, 1979). It also increases with pregnancy and oral contraceptive use. With any of these conditions, copper deficiency might not be diagnosed on the basis of serum copper or ceruloplasmin concentrations.

Erythrocyte Superoxide Dismutase Activity

Erythrocyte superoxide dismutase (SOD) activity, though not as specific as serum copper or ceruloplasmin concentration, may be a reliable indicator of copper status, and some suggest it is more sensitive (Milne, 1998; Uauy et al., 1985). It does not increase with the conditions that increase serum copper and ceruloplasmin concentrations. However, it can increase in situations that produce oxidative stress, and SOD activity is high in some conditions, including alcoholism and Down's syndrome. Methods of analysis are not standardized, and normal ranges for SOD activity are not available. Although SOD activity was measured in fewer studies than were the two indicators above, sufficient data are available to include it as an indicator of change in copper status when it is measured in controlled studies at different levels of dietary copper intake.

Platelet Copper Concentration and Cytochrome c Oxidase Activity

Two studies in women suggest that both platelet copper concentration and platelet cytochrome c oxidase activity may respond more rapidly to low dietary copper than the indicators discussed above. In one study both of these indicators declined when copper intake was 570 μg/day (Milne and Nielsen, 1996). Platelet copper concentration increased after repletion, but platelet cytochrome c oxidase activity did not. In another study, both platelet copper concentration and platelet cytochrome c oxidase activity increased after supplementation of a diet containing 670 μg/day of copper, but baseline measurements were not made, so it is not known whether these parameters declined (Milne et al., 1988). Moreover, an intervening vitamin C supplementation period added another variable to the data interpretation. The fact that serum copper and ceruloplasmin concentrations and SOD activity were not affected at this level of dietary copper suggests the requirement for maintaining serum copper and ceruloplasmin concentration had been met. Therefore, the above research suggests that platelet copper concentration and platelet cytochrome c oxidase activity, when measured in controlled studies, may be more sensitive to changes in copper dietary intake.

Urinary Copper

Urinary copper excretion is extremely low and does not contribute significantly to copper retention, but it has been found to decline when diets are low enough in copper that other indexes of copper status change (Turnlund et al., 1997). Above those levels of dietary intake, urinary copper does not respond to increases in dietary copper. In controlled studies, a decline in urinary copper excretion can be used as supporting evidence for inadequate intake.

Leukocyte Copper Concentration

Leukocyte copper concentration was found to decline along with other indexes of copper status in one study (Turnlund et al., 1997), but it has not been reported in others. Too few data are currently available to use it for establishing dietary recommendations for copper.

Lysyl Oxidase Activity

Lysyl oxidase activity in the skin, which declined with low dietary copper and increased with repletion, is potentially a useful indicator of copper status (Werman et al., 1997). It is not known if lysyl oxidase activity reflects dietary intake at higher levels of dietary copper in humans. Because data are available from only one study, it cannot yet be used as an indicator for estimating copper requirements.

Peptidyl Glycine α-Amidating Monooxygenase Activity

Peptidyl glycine α-amidating monooxygenase (PAM) activity in serum of rats and stimulation of activity are sensitive indicators of copper intake in the rat (Prohaska et al., 1997). Patients with Menkes' disease, who have severe copper deficiency due to a metabolic defect in copper transport, had an increased copper stimulation index of plasma PAM as compared with healthy control subjects. This finding suggests that PAM activity may be a useful indicator of copper status in humans when human dose-response data become available.

Diamine Oxidase Activity

Two copper supplementation studies demonstrated that the activity of serum diamine oxidase (DAO), another cuproenzyme, increases when supplements containing 2 mg (Jones et al., 1997) and 6 mg (Kehoe et al., 2000) of copper were administered daily, a result that suggests the enzyme may be sensitive to increased dietary copper. It has not yet been studied under conditions of copper depletion. Because intestinal damage and a number of conditions also elevate DAO activity, its use as an indicator of copper status is possibly limited.

Copper Balance

Balance studies have been used in the past to estimate dietary recommendations. Numerous copper balance studies in humans have been conducted over a wide range of intakes (Mason, 1979). Unfortunately, there are a number of problems with this approach, as reviewed by Mertz (1987). Copper balance, which can be achieved over a broad range of dietary copper intakes, reflects prior dietary intake; thus long adaptation is required for results to be meaningful. Seldom are studies long enough. Such studies are prone to numerous errors, and data from some studies would suggest that an unacceptable amount of copper would accumulate over time if these levels of retention were continued. In addition, miscellaneous losses, while small, are very difficult to quantify. Therefore, balance studies were not used as an indicator of copper status.

Factorial Analysis

One approach to estimating minimum dietary mineral requirements is by the factorial method. Obligatory losses, the amounts of an element excreted with no dietary intake, are determined, and then the amount needed in the diet to replace these obligatory losses is calculated. Obligatory losses include urinary losses, gastrointestinal losses, sweat, integument, hair, nails, and other miscellaneous losses such as menstrual and semen losses. For copper, as for other elements, reliable values for many of these losses are not available. However, sufficient data are available to make reasonable estimates; therefore, this method can be used in support of estimates of dietary copper requirements made by other methods.

Zinc

Zinc intakes, well in excess of the amount normally found in the diet, can decrease copper absorption in adults (Turnlund, 1999) (see Table 12-7). In one case report, an infant who was given 16 to 24 mg/day of zinc developed copper deficiency (Botash et al., 1992). Very high doses of zinc have been used to treat patients with Wilson's disease, an inborn error of copper metabolism resulting in copper toxicity (Brewer et al., 1983). This zinc-induced inhibition of copper absorption could be the result of competition for a common, apically oriented transporter or the induction of metallothionein in intestinal cells by zinc. Because this protein has a higher binding affinity for copper than for zinc, copper is retained within enterocytes and its absorption is reduced. This response has been used as a therapy to diminish copper absorption in patients with Wilson's disease (Yuzbasiyan-Gurkan et al., 1992). The interaction could also be responsible for reducing copper absorption during consumption of zinc supplements. When zinc-to-copper ratios of 2:1, 5:1, and 15:1 were fed to humans, there were limited effects on copper absorption (August et al., 1989).

Iron

High iron intakes may interfere with copper absorption in infants. Infants fed a formula containing low concentrations of iron absorbed more copper than infants consuming the same formula with a higher iron concentration (Haschke et al., 1986). Such an interaction has been reported to produce reduced copper status in infants (Lonnerdal and Hernell, 1994; Morais et al., 1994).

Fructose

Studies in rats demonstrated that diets very high in fructose were associated with increased severity of copper deficiency in rats (Fields et al., 1984), but a similar effect was not observed in pigs (Schoenemann et al., 1990), which have cardiovascular systems and gastrointestinal tracts more similar to those of humans. The effects were inconsistent in humans (Reiser et al., 1985) but did not result in copper depletion, and the extremely high levels of fructose fed (20 percent of energy intake) suggest the effect would not be relevant to normal diets.

Infants Ages 0 through 12 Months

Method Used to Set the Adequate Intake

No functional criteria of copper status have been demonstrated that reflect response to dietary intake in infants. Thus, recommended intakes of copper are based on an Adequate Intake (AI) that reflects the observed mean copper intake of infants principally fed human milk.

Ages 0 through 6 Months. The AI for infants ages 0 through 6 months was based on the usual intake from human milk. The copper content of human milk is highest during early lactation and then declines during the course of lactation. According to a number of reports (Biego et al., 1998; Raiten et al., 1998; Rossipal and Krachler, 1998), the mean copper content of human milk during the first 6 months of lactation is approximately 250 μg/L (Table 7-1). There are no indications that the copper content of human milk is inadequate to maintain copper status. Liver copper stores are high (Widdowson and Dickerson, 1964) and serum copper and ceruloplasmin concentrations are low (Salmenpera et al., 1986) in newborn infants. During the first 6 months of life, liver stores decline and serum copper concentration increases to adult levels, independent of copper intake.

TABLE 7-1

Copper Concentration in Human Milk.

Based on the copper content of human milk and milk consumption (Chapter 2), the AI for infants ages 0 through 6 months is 200 μg/day (250 μg/L × 0.78 L/day) after rounding. For a 7 kg infant (reference weight for 0 through 6 months, Chapter 2), this would be 28 μg/kg/day (200 μg/day/7 kg), rounded up to 30 μg/kg/day.

Ages 7 through 12 Months. One method for estimating the AI for infants receiving human milk, ages 7 through 12 months, is based on the average intake from human milk plus an added increment for complementary foods (Chapter 2). According to the Third National Health and Nutrition Examination Survey, the median copper intake from weaning food for children aged 7 through 12 months is 100 μg/day (n = 45). The average copper concentration in human milk declines over time, and between 7 and 12 months postpartum the concentration is 200 μg/L or less (Table 7-1). Based on an average volume of 0.6 L/day of human milk that is secreted, the copper intake from human milk is 120 μg/day (0.6 × 200). Therefore the total intake of copper from human milk and complementary foods is 220 μg/day (120 + 100). For a 9 kg infant (reference weight 7 through 12 months, Chapter 2), this would be 24 μg/ kg/day (220 μg/kg ÷ 9 kg).

If the AI were extrapolated from the AI for younger infants by using the calculation in chapter 2, the average intake would be 241 μg/day.

Children and Adolescents Ages 1 through 18 Years

Adults Ages 19 Years and Older

Evidence Considered in Estimating the Average Requirement

Biochemical Indicators. No single indicator was judged as sufficient for deriving an EAR for adults. Results for specific indicators vary between studies. To determine the EAR, a combination of indicators was used, including plasma copper concentration, serum ceruloplasmin concentration, erythrocyte superoxide dismutase activity (SOD), and platelet copper concentration in controlled human depletion/repletion studies using specific amounts of copper. If there were significant decreases in serum copper and ceruloplasmin concentrations and SOD activity when the experimental copper diet was fed, and if this decrease was reversed with added copper, then the experimental diet was considered insufficient to maintain status and therefore deficient in copper. If plasma copper and serum ceruloplasmin concentrations did not change significantly when the experimental diet was fed, but platelet copper concentration decreased, then the experimental diet was judged to be marginally adequate in copper. A lack of change in the copper status indicators indicated that the level of copper in the experimental diet was adequate to maintain status. Because of limited data, data from men and women were combined.

Three studies were used to estimate the average requirement on the basis of copper status. These studies are summarized in Table 7-2. Serum copper and ceruloplasmin concentrations and SOD activity declined significantly in eight of 11 young men fed an experimental, depletion diet containing 388 μg/day and increased with repletion (Turnlund et al., 1997). Although these indicators decreased significantly, they did not fall to the deficient range while the deficient diet was fed for 42 days. However, it is expected that they would have fallen to the deficient range over a longer time. Other changes suggesting copper depletion were observed.

TABLE 7-2

Effects of Copper (Cu) Intake on Copper Status.

When young men were fed 790 μg/day of copper, the above mentioned indicators did not decline significantly (Turnlund et al., 1990). After a decline in copper status, two of the 11 men responded to copper repletion. Therefore, the copper requirement to maintain copper status in half of a group is more than 380 μg/day but less than 790 μg/day. On the basis of these data, a linear model was used to estimate a response curve. The model estimated that half of these men would not maintain copper status at 550 μg/day.

Serum copper and ceruloplasmin concentrations did not decline significantly when ten women were fed 570 μg/day of copper (Milne and Nielsen, 1996). Platelet copper concentration, however, declined significantly for eight of ten women fed 570 μg/day and increased with supplementation. Other indicators did not respond to depletion. Platelet cytochrome c oxidase and erythrocyte SOD activity declined but did not respond significantly to repletion. While an EAR based on the first two studies was estimated at 550 μg/day, the latter study suggests that 600 μg/day may be a marginal intake in over half of the population. Therefore, another increment was added to cover half of the population, and the EAR was set at 700 μg/day.

Factorial Analysis. Another approach for estimating the minimum copper requirement is to estimate obligatory losses of copper and calculate the amount of copper required in the diet to replace these obligatory losses. This approach provides supporting evidence for the EAR based on copper status estimated above. Endogenous losses, estimated from total parenteral nutrition (TPN) data, were estimated to be 300 μg/day by Shike and coworkers (1981). This estimate was based on gastrointestinal losses from patients without excessive gastrointestinal secretions (less than 0.3 L/day) of 191 μg/day and urinary losses of 90 μg/day, which are higher than urinary losses in normal, healthy adults, and would provide an increment for miscellaneous losses. The TPN patients received no copper orally, but copper from TPN ranged from 250 to 1,850 μg/day.

There are no data on obligatory copper losses in healthy people; therefore the study with the lowest copper intake and data on endogenous losses was used to estimate obligatory losses in healthy people (Turnlund et al., 1997, 1998). When copper intake was 380 μg/day, copper status declined significantly. Endogenous fecal losses were calculated to be 240 μg/day, slightly higher than the estimate from TPN data (Shike et al., 1981), and urinary losses were less than 20 μg/day. A careful study of surface copper losses in men reported that these averaged 42 μg/day (Milne et al., 1991). Other losses, such as hair, nails, semen, or menstrual, have not been measured, and it is assumed they are similar to surface losses. Therefore the amount of absorbed copper needed to replace obligatory losses is 344 μg/day (240 + 20 + 42 + 42). Copper absorption at this level of intake is approximately 75 percent. Therefore, 460 μg/day of dietary copper would be the minimum amount required to replace obligatory losses. Endogenous fecal copper was 50 μg/day higher at 380 μg/day than at 460 μg/day, and so 50 μg/day was added to endogenous fecal losses to account for the increase that occurs between 380 and 460 μg/day. Thus 510 μg/day (460 + 50) of dietary copper is required to replace copper losses from all sources and to achieve zero balance. Estimation of the average requirement based on indicators of copper status is similar to, but slightly higher than, the average requirement determined by the factorial approach. The EAR is based on biochemical indicators of copper status of men and women, and there was no basis for a difference in requirement based on gender. There are no data on which to base an EAR for older adults, and no evidence to suggest that the requirements would be different.

Pregnancy

Lactation

Food Sources

Copper is widely distributed in foods. The accumulation of copper in plants is not affected by the copper content of the soil in which they grow. Organ meats, seafood, nuts, and seeds are major contributors of dietary copper (Pennington et al., 1995). Wheat bran cereals and whole grain products are also sources of copper. Foods that contribute substantial amounts of copper to the U.S. diet include those high in copper, such as organ meats, grains, and cocoa products, and those relatively low in copper that are consumed in substantial amounts, such as tea, potatoes, milk, and chicken.

Dietary Intake

Data from nationally representative U.S. surveys are available to estimate copper intakes (Appendix Tables C-15, C-16, D-2, E-3). The median intake of copper for women is approximately 1.0 to 1.1 mg/day, whereas the median intake for men ranges from 1.2 to 1.6 mg/day (Appendix Tables C-15 and D-2).

Intake from Supplements

In 1986, approximately 15 percent of adults in the United States consumed supplements that contained copper (Moss et al., 1989; see Table 2-2). Based on data from the Third National Health and Nutrition Examination Survey provided in Appendix Table C-16, the median dietary plus supplemental copper intake was similar to the intake from food alone. The mean intake of dietary and supplemental copper (1.3 to 2.2 mg/day) was approximately 0.3 to 0.5 mg/day greater for men and women than the mean intake from food (1.0 to 1.7 mg/day).

Hazard Identification

Reviews of the toxicity studies in experimental animals (ATSDR, 1990; EPA, 1987; IPCS, 1998; NRC, 1977) indicate that these studies are not useful for setting a UL for humans. Very few of these studies used chronic exposures, only one or two doses were used, and the reporting of experimental details and results was incomplete. In addition, some studies used routes of exposure that are not relevant to human intake (Toyokuni and Sagripanti, 1994). Finally, animal species vary markedly in their sensitivity to copper (Davis and Mertz, 1987); thus it is difficult to determine the most appropriate model in which to assess human toxicity to copper.

The long-term toxicity of copper is not well studied in humans, but it is rare in normal populations not having some hereditary defect in copper homeostasis (Olivares and Uauy, 1996). Copper homeostasis is affected by the interaction among zinc, copper, iron, and molybdenum. In addition, the level of dietary protein, interacting cations, and sulfate all can influence the absorption and utilization of copper (Davis and Mertz, 1987). Therefore, the derivation of a UL for copper must be made in the context of these interactions. The adverse effects associated with intake of soluble copper salts in supplements and drinking water are reviewed below.

Dose-Response Assessment

Adults

Data Selection. The human data evaluating liver effects after chronic consumption of copper gluconate appear most relevant to setting a UL. The UL derived below does not apply to individuals at increased risk of adverse effects from excess intake of copper. These subgroups are identified under “Special Considerations.”

Identification of a No-Observed-Adverse-Effect Level (NOAEL) and a Lowest-Observed-Adverse-Effect Level (LOAEL). A NOAEL of 10 mg/day of copper was identified on the basis of the results of Pratt and coworkers (1985). In a 12-week, double-blind study, 10 mg of copper as copper gluconate capsules was consumed daily by seven adults. Liver function tests were normal. From a case report, consumption of 30 mg/day as copper tablets for 2 years, followed by 60 mg/day for an additional period of time, resulted in acute liver failure (O'Donohue et al., 1993).

Uncertainty Assessment. The NOAEL of 10 mg/day was considered to be protective of the general population. Therefore, an uncertainty factor (UF) of 1.0 was selected. A larger UF was considered unnecessary in view of the large international database in humans indicating no adverse effects from daily consumption of 10 to 12 mg/day of copper in foods and the rarity of observed liver damage from copper exposures in human populations with normal copper homeostasis.

Derivation of a UL. The NOAEL of 10 mg/day was divided by the UF of 1.0 to obtain a UL of 10 mg/day (10,000 μg/day) of copper intake from food and supplements.

Intake Assessment

Based on data from NHANES III (Appendix Table C-16), the highest median intake of copper from the diet and supplements for any gender and life stage group was about 1,700 μg/day for men aged 19 through 50 years and about 1,900 μg/day for lactating women. The highest reported intake from food and supplements at the ninety-ninth percentile was 4,700 μg/day in lactating women. The next highest reported intake at the ninety-ninth percentile was 4,600 μg/day in pregnant women and men aged 51 through 70 years.

In situations where drinking water that contains copper at the present U.S. Environmental Protection Agency (EPA) Maximum Contaminant Level Goal is consumed daily, an additional intake of 2,600 μg of copper in adults and 1,000 μg in 1- through 4-year-old children is possible. However, as reported by IPCS (1998), data from the EPA indicate 98 percent of flushed drinking water samples had copper levels of less than 460 μg/L. According to these values, most of the U.S. population receives less than 100 to 900 μg/day of copper from drinking water.

Whether total daily intakes of copper will lead to adverse health effects will depend upon the species of copper in the media of concern, its degree of ionization, and its bioavailability.

Risk Characterization

The risk of adverse effects resulting from excess intake of copper from food, water, and supplements appears to be very low in adults at the highest intakes noted above. However, copper intake data indicate that a small percentage of children aged 1 through 8 years are likely to exceed the UL for their age group. Although members of the general population should be advised not to exceed the UL routinely, intake above the UL may be appropriate for investigation within well-controlled clinical trials. Clinical trials of doses above the UL should not be discouraged, as long as subjects participating in these trials have signed informed consent documents regarding possible toxicity and as long as these trials employ appropriate safety monitoring of trial subjects. In addition, the UL is not meant to apply to individuals who are receiving copper under medical supervision.