Click on the links below to access all the ArticlePlus for this article.
Please note that ArticlePlus files may launch a viewer application outside of your web browser.
Zinc, copper, and magnesium are minerals involved in many homeostatic mechanisms of the body, including specific immunity, inflammation, and oxidative stress.1–5 Previous animal studies and cell cultures have shown the role of these minerals in atherogenesis and carcinogenesis.6–11 However, the evidence linking zinc, copper, and magnesium to cancer or cardiovascular disease in humans is far from conclusive, and little is known about their potential interactions.
There are few prospective epidemiologic studies on the relationship between these essential elements and cancer or cardiovascular morbidity and mortality that have controlled for the effect of other major risk factors12–18 and, to our knowledge, only one study has examined simultaneously these 3 elements on mortality risk.18 We used data from the Paris Prospective Study 2, a cohort study of 4035 men, to investigate the associations between serum levels of these minerals and all-cause, cancer, and cardiovascular disease mortality.
Subjects and Study Variables
The Paris Prospective Study 2 is an epidemiologic investigation of cardiovascular risk factors.19,20 The study population involved male employees of 3 large public organizations in Paris who volunteered for a cardiovascular screening. At baseline, participants answered a standardized questionnaire administered by 2 specially trained technicians, including information on clinical history (hypertension, ischemic heart disease, arterial occlusive disease, stroke), smoking status (never, former, current), alcohol consumption, and physical activity. Participants also underwent a 24-hour dietary recall interview. After the nutritional data were coded by a single dietitian, a measure of nutrient intake was obtained from a computer program applying a food composition table.
Daily tobacco consumption was estimated as a weighted sum of the kinds of tobaccos (cigarettes, cigars, pipe) consumed by a subject, expressed in grams of tobacco per day. Alcohol consumption was estimated as a weighted sum of the types of alcoholic beverages (wine, beer, cider, liqueur, other aperitifs) consumed by a subject during a typical current week, expressed in milliliters of alcohol per day and classified into 3 categories (≤20, 21–40, >40 mL/d).
Participants were asked about their regular leisure physical activity, which was then classified into 3 categories: sedentary, moderate (eg, walking, gardening), and vigorous (eg, regular sport, cycling).
Systolic and diastolic blood pressures (in millimeters of mercury) were measured twice at baseline, and mean values were used in the statistical analyses. Hypertension was defined as systolic blood pressure of ≥140 mm Hg or diastolic blood pressure of ≥90 mm Hg, or current use of antihypertensive medication. Weight (in kilograms) and height (in meters) were measured for each participant and body mass index (BMI) was computed as weight (in kilograms)/height2 (in meters squared).
A fasting blood sample was drawn for biochemical measurements, including serum zinc, copper, and magnesium concentrations, total cholesterol, low-density lipoprotein (LDL) and high-density lipoprotein (HDL) cholesterol, triglycerides, glycemia, blood cell count, and carboxyhemoglobin concentration (as a marker of tobacco consumption20). Diabetes was defined as fasting blood glucose of ≥1.26 g/L or current use of diabetes medication.
Blood samples were collected in propylene tubes without anticoagulant agent to limit contamination. Serum zinc, copper, and magnesium concentrations were measured by flame atomic absorption spectrometry. Details of the biochemical method have been published elsewhere.21–23
From December 1980 to December 1985, 4547 men age 20–60 years entered the study. Participants born abroad (n = 212) and men age <30 years at baseline (n = 80) were subsequently excluded. Mortality and cause-of-death code were determined for each subject through linkage with the death certificate obtained from the birth municipal corporation and cause of death registered by the National Institute of Health and Medical Research (INSERM).
Mortality follow up for the cohort was through December 2001, and after an average ± standard deviation (SD) of 17.8 years ± 2.9 years, the vital status was obtained for 97.9% of the 4255 native Frenchmen included. Persons missing any data on serum zinc, copper, and magnesium or on other covariates (n = 133) were excluded from the analysis.
Finally, a full health history with clinical and biologic evaluation was achieved for 4035 employees, with a total of 339 participants dying during the follow up. Determination of the study sample is summarized in Figure 1.
Causes of death were established according to the International Classification of Diseases, 9th revision (ICD-9) for deaths occurring up to 1998 and according to the ICD-10 after 1998.
Of the 339 deaths, 176 were from cancer (140–239 ICD-9 and C00–D48 ICD-10 codes), including 50 lung cancer, 48 digestive cancers, 16 hematopoietic cancers, and 62 other. Fifty-six deaths were of cardiovascular origin (390–459 ICD-9 and I00–I99 ICD-10 codes), including 30 deaths from coronary heart disease, 5 sudden deaths, 6 strokes, and 15 other. The remaining 107 deaths were from violent death (n = 47; 800–988 ICD-9 and V01–Y98 ICD-10 codes), digestive origin not cancers (n = 26; 555–569/570–579 ICD-9 and K70–77/K92 ICD-10 codes), other causes (n = 23), and nonspecified cause (n = 11; 799 ICD-9 and R99 ICD-10 codes).
Past studies suggest that low serum zinc and magnesium and high serum copper may lead to atherosclerosis and carcinogenesis.12–18,24–26 However, there are no recognized cutoff values derived from apparently healthy populations above which these essential elements have been shown to be harmful.
Consequently, and to evaluate the effects of marginal values of exposure, we divided serum zinc, copper, and magnesium concentrations into quartiles based on their distribution among the total population at baseline.
Each mineral serum concentration was then classified into low (1st quartile, the referent), medium (2nd–3rd quartiles), and high (4th quartile).
Student t- and χ2 tests were used to describe and compare the characteristics of dead and alive. Kaplan-Meier survival estimates were graphed over the follow-up period according to serum zinc, copper, and magnesium levels. Relative risks (RRs) and 95% confidence intervals (CIs) of death from all-cause, cancer, and cardiovascular disease for zinc, copper, and magnesium were examined using Cox's proportional hazard model. Multivariate adjustment was made for age, BMI, smoking status (never, former, and current, using 2 dummy variables), alcohol consumption (≤20, 21–40, and >40 mL/d, using 2 dummy variables), physical activity (sedentary life, moderate, vigorous exercise, using 2 dummy variables), hypertension, LDL and HDL cholesterol, diabetes (yes/no), and cardiovascular disease history (ischemic heart disease, arterial occlusive disease, stroke) (yes/no). In multivariate models, zinc, copper, and magnesium were included simultaneously.
The proportionality assumption of Cox's model was tested by including an interaction term for follow-up time and each mineral variable separately with no evidence that the assumption was violated. Multiplicative interaction terms between mineral categories were tested. Interaction term between each mineral variable and each covariate were also tested in separate multivariate models.
In addition to the multivariate analysis for the whole cohort, the proportional hazard model was performed after excluding subjects with neoplasms (n = 11) or cardiovascular disease at baseline (n = 60) or those who died within the first 5 years of follow up (n = 39). The relations between minerals and all-cause, cancer, and cardiovascular mortality were not markedly modified.
Baseline characteristics by mortality status are given in Table 1. During the 18-year follow up, 339 deaths (8.4%) were registered. Deceased participants were, at baseline, older, more likely to report tobacco and alcohol consumption, sedentary life, and cardiovascular disease history. They also had higher frequency of hypertension and diabetes, and higher HDL cholesterol, triglycerides, and leukocyte serum values. Subjects who died had lower carbohydrate intake on their 24-hour dietary recall than those who survived. All these factors were associated with cancer deaths (data not shown). In addition, men who died of cancer had lower BMI (24.9 ± 3.1 vs 25.3 ± 2.8 kg/m2) than those who did not die of cancer. Men who died of cardiovascular disease were older and more likely to report tobacco consumption and cardiovascular disease than subjects who did not die of cardiovascular disease. They also had lower HDL cholesterol and higher LDL cholesterol and leukocyte serum values (data not shown).
The serum mineral concentrations in the whole cohort ranged from 6.5 to 35 μmol/L for zinc (14.6 ± 1.9 μmol/L), 6 to 27.5 μmol/L for copper (15 ± 2.2 μmol/L), and 0.24 to 1.78 mmol/L for magnesium (0.80 ± 0.06 mmol/L). Serum zinc was positively correlated with serum magnesium (correlation coefficient [r] = 0.10) and with serum copper (r = 0.12). Serum copper and serum magnesium were slightly less correlated (r = 0.08).
Deceased participants had lower baseline serum zinc and magnesium values, and higher serum copper levels than surviving participants (Table 1). Men who died of cancer had lower zinc and higher copper values at baseline, whereas cardiovascular deaths had higher copper and lower magnesium values (data not shown).
Age was positively related to serum zinc and serum copper but negatively to serum magnesium (Table available with the online version of this article). Serum copper values were higher among tobacco users. Alcohol consumption (>40 mL/d) was negatively linked to serum zinc and serum magnesium but positively to serum copper. Cardiovascular disease history was positively associated with serum copper.
Zinc, copper, and magnesium were positively associated with serum total cholesterol. These associations were mainly the result of LDL cholesterol, and negative associations were found between HDL cholesterol and both serum copper and serum magnesium (data not shown). Blood leukocytes were associated with low serum zinc and high serum copper.
Mortality rates, by baseline levels of serum zinc, copper, and magnesium (low, medium, high), are shown in Table 2. High serum zinc levels (4th quartile) tended to be slightly associated with decreased relative risks for all-cause and cancer mortality compared with low levels (1st quartile). Subjects with high copper values had increased relative risks for all-cause mortality (RR = 1.5; 95% CI = 1.1–2.1) and cancer mortality (1.4; 0.9–2.2) compared with subjects with low levels. Further adjustment on blood leukocyte count decreased the relative risk by 20% for all-cause (1.3; 1.0–1.8) and cancer deaths (1.2; 0.8–1.8).
In contrast, high serum magnesium values were negatively associated with mortality. Men with high magnesium levels had relative risks decreased by 40% to 50% for all-cause and cancer mortality compared with men having low magnesium levels. Cardiovascular mortality tended to be negatively associated with high serum zinc and magnesium levels and positively with high serum copper values (Table 2). Mortality risks for zinc and magnesium were not altered after further adjustment for blood leukocytes count (data available from authors). All-cause, cancer, and cardiovascular mortality risks for each mineral did not change after further adjustments for nutrient intake.
Serum carboxyhemoglobin concentration was strongly associated with smoking status and daily tobacco consumption.20 In all analyses, substitution of carboxyhemoglobin concentration or daily tobacco consumption for smoking status did not modify the mortality risks related to serum minerals (data available from authors).
When minerals were used as continuous variables, multivariate relative risks of all-cause mortality associated with approximately one standard deviation increase in zinc (1.9 μmol/L), copper (2.3 μmol/L), and magnesium (0.06 mmol/L) were, respectively, 0.9 (95% CI = 0.8–1.0), 1.2 (1.0–1.3), and 0.8 (0.7–0.9). For cancer mortality, the corresponding relative risks were, respectively, 0.9 (0.8–1.0), 1.2 (1.0–1.4), and 0.8 (0.7–0.9). These results support the categorical analyses reported in Table 2.
Interactions between serum zinc and serum copper or serum magnesium were observed, suggesting that mortality relative risks associated with these 2 minerals differed according to the level of serum zinc. Table 3 shows the results from the interaction between serum zinc and serum copper levels and between serum zinc and serum magnesium levels. Combined low zinc (1st quartile) and high copper serum values (4th quartile) were associated with an increased risk of dying from all causes (2.6; 1.4–5.0) and from cancer (2.7; 1.0–7.3) as compared with combined low zinc and low copper values (Table 3). Adjustment for blood leukocytes count slightly decreased the mortality risk for cancer deaths (2.2; 0.8–6.2).
In the same way, among persons with low zinc, high magnesium was related to a decrease in 80% of the relative risks for all-cause (0.2; 0.1–0.5) and cancer (0.2; 0.1–0.8) mortality compared with low magnesium. Kaplan-Meier survival curves demonstrate these results for all-cause mortality (Fig. 2) and for cancer (data not shown). No interaction between serum copper and serum magnesium on mortality risks was observed (data not shown).
Results from an 18-year follow up of more than 4000 men in the Paris Prospective Study 2 suggest lower risks of all-cause and cancer mortality with serum magnesium and higher risks with serum copper. The association between cardiovascular mortality and these 2 elements remained marginal, perhaps as a result of a lack of statistical strength. In contrast, the relation between serum zinc and mortality was inconclusive. Furthermore, a markedly increased risk of dying from all-cause and cancer was observed in men having both low serum zinc and high serum copper levels independently of well-known risk factors. Similarly, men having both low zinc and high magnesium serum values were at 80% lower risk of all-cause and cancer mortality.
Few prospective epidemiologic surveys have examined the relationship between these elements and mortality. After a 12- to 16-year follow up of 6244 persons, and with adjustment for potential confounding factors (excepting alcohol consumption), Wu et al14 found that cancer mortality was positively associated with serum copper and negatively related to serum zinc. In the First National Health and Nutrition Examination Survey (NHANES I), including 12,340 subjects, Ford17 have shown that the risk of dying from all-cause and ischemic heart disease was 20% to 30% lower for participants within the third and fourth quartiles of serum magnesium concentration compared with the first quartile. In a 12-year follow-up survey (NHANES II), a study sample of 9252 subjects, Ford16 observed that after adjustment for several confounding factors, serum copper was strongly associated with coronary heart disease mortality, particularly in men.
Ito et al15 examined the zinc/copper serum ratio on 507 subjects aged 40 or more years living in a rural area. After 18-year follow up, they found that cancer mortality risk was higher for subjects with the lowest zinc/copper ratio. Neither serum cholesterol nor systolic blood pressure was included in the multivariate model.
Marniemi et al18 have studied a random sample of 480 community-living elderly aged 65 years and older. After a 13-year follow up, they found that serum copper was independently associated with increased risk of vascular death. However, neither serum zinc nor serum magnesium was associated with mortality when adjusting for confounding risk factors.
Only a few interventional studies have addressed the effects of dietary intakes of various minerals and trace elements on the incidence rate of cardiovascular disease and cancer. Hercberg et al27 have recently shown in a large randomized, placebo-controlled trial that after 7.5 years of daily low-dose antioxidant supplementation, including vitamin C, vitamin E, beta carotene, selenium, and zinc, total cancer incidence and all-cause mortality were lowered in men. Regular use of vitamins and other micronutrients were not registered in our study, but this kind of supplementation remains uncommon in France, especially during the time this cohort was established.
In our study, and as found earlier,28,29 serum zinc was positively linked to serum total cholesterol. We observed that high serum levels of copper were strongly associated with age, tobacco consumption, and serum total cholesterol, 3 major atherosclerosis risk factors, confirming previous results.25,28 Additionally, as shown by He et al,29 serum copper was negatively associated with HDL cholesterol.
As found earlier,12,26 magnesium was inversely associated with age, hypertension, and diabetes. In addition, we noticed a negative association between serum magnesium and HDL cholesterol. We know of no previous studies that have considered interactions between serum zinc and serum copper or magnesium, and their associations with mortality.
The mechanisms linking minerals to mortality risk are not well known and several hypotheses can be formulated. First, moderate or mild zinc deficiency is associated with depressed immune function,30 which can enhance inflammatory cytokines and depress cytokines affecting lymphocytes, promoting apoptosis, angiogenesis, and metastasis.31 Copper is involved in oxidative damage. On the one hand, copper is essential to copper/zinc superoxide dismutase (SOD), a key antioxidant enzyme. A lowered amount of SOD has been found in many tumors, although not all.32,33 On the other hand, free copper ions participate in the formation of reactive oxygen species, able to react in particular with LDL, a key event in human atherogenesis.7,24 Copper may also be implicated in inflammatory processes.16 Our results support this hypothesis. We found a positive association between blood leukocytes count and serum copper. Moreover, adjustment for leukocytes tended to decrease the association between copper and mortality, indicating that part of this relation might be the result of inflammation. Nevertheless, the persistence of a positive association between serum copper and mortality, independent of leukocytes, suggests that other mechanisms may be also involved.
Low serum magnesium seems also to be associated with enhanced LDL oxidation, and it has been suggested that a magnesium decrease could initiate the inflammatory response, which might contribute to atherogenesis and carcinogenesis.5,11 Furthermore, magnesium, like zinc, has DNA-stabilizing properties.3,4
The observed synergistic effects on mortality of low serum zinc and high serum copper, as well as the effects of low serum zinc and low serum magnesium, may be physiopathologically plausible. Zinc (as a component of SOD) protects against free radicals and by competing for binding sites,1,3 as opposed to free copper ions and to low magnesium that could promote lipid peroxidation. In this way, decreased zinc and either increased copper or decreased magnesium might synergistically enhance oxidative damage and the inflammatory response.
High intakes of zinc inhibit the absorption of copper in the intestine, liver, and kidney by stimulating the formation of metallothionein,2 whereas an excess of copper reduces zinc absorption.34 Moreover, zinc and magnesium are both involved in the conversion of essential dietary fatty acids (linoleic acid and α-linolenic acid) to prostaglandins and prostacyclins, which are potent vasodilators and platelet antiaggregators.35 Pure or oxidized fatty acids could promote apoptosis, particularly in endothelial cells.9
Some potential limitations of our study need to be considered. The cohort from the Paris Prospective Study II, based on employed men, was not a representative sample of the French population. This may limit the generalizability of our results. Some authors also have proposed that the relation between serum copper and cancer or coronary heart disease mortality might be sex-dependent.14,16
Minerals were measured in serum at baseline. Serum zinc, copper, and magnesium are perhaps not the best indices of total body store. Also, serum status at entry does not take into account their possible variability overtime. Furthermore, micronutrient serum status may not accurately reflect dietary intake. Metabolic and pathologic processes relating diet and serum status are complex, reflecting, for example, individual genetic susceptibility. Thus, we should not expect simple links between serum status or diet and cancer or cardiovascular risk.36–38 Finally, incidence data were not available in our survey, and we cannot exclude the possibility that serum zinc, copper, and magnesium values are prognostic factors rather than etiologic factors.
In summary, serum zinc, copper, and magnesium were related to subsequent mortality in men. Our findings suggest that combined low serum zinc with either high serum copper or low serum magnesium values contribute synergistically to an increased mortality risk. Whether these minerals are partly responsible for, or are biomarkers of, cancer and cardiovascular disease remains to be established. Further studies are needed to confirm the interactions between serum zinc and serum copper or serum magnesium and their potential contribution to the prediction of all-cause, cancer, and cardiovascular disease mortality in clinical practice.
1. Prasad AS, Bao B, Beck FW, et al. Antioxidant effect of zinc in humans. Free Radic Biol Med
2. Gaetke LM, Chow CK. Copper toxicity, oxidative stress, and antioxidant nutrients. Toxicology
3. Ho E. Zinc deficiency, DNA damage and cancer risk. J Nutr Biochem
4. Anastassopoulou J, Theophanides T. Magnesium–DNA interactions and the possible relation of magnesium to carcinogenesis. Irradiation and free radicals. Crit Rev Oncol Hematol
5. Tam M, Gomez S, Gonzalez-Gross M, et al. Possible roles of magnesium on the immune system. Eur J Clin Nutr
6. Klevay LM. Coronary heart disease: the zinc/copper hypothesis. Am J Clin Nutr
7. Ferns GA, Lamb DJ, Taylor A. The possible role of copper ions in atherogenesis: the Blue Janus. Atherosclerosis
8. Harris ED. A requirement for copper in angiogenesis. Nutr Rev
9. Hennig B, Meerarani P, Ramadass P, et al. Zinc nutrition and apoptosis of vascular endothelial cells: implications in atherosclerosis. Nutrition
10. Kasprzak KS, Waalkes MP. The role of calcium, magnesium, and zinc in carcinogenesis. Adv Exp Med Biol
11. Maier JA. Low magnesium and atherosclerosis: an evidence-based link. Mol Aspects Med
12. Liao F, Folsom AR, Brancati FL. Is low magnesium concentration a risk factor for coronary heart disease? The Atherosclerosis Risk in Communities (ARIC) Study. Am Heart J
13. Salonen JT, Salonen R, Korpela H, et al. Serum copper and the risk of acute myocardial infarction: a prospective population study in men in eastern Finland. Am J Epidemiol
14. Wu T, Sempos CT, Freudenheim JL, et al. Serum iron, copper and zinc concentrations and risk of cancer mortality in US adults. Ann Epidemiol
15. Ito Y, Suzuki K, Sasaki R, et al. Mortality rates from cancer or all causes and SOD activity level and Zn/Cu ratio in peripheral blood: population-based follow-up study. J Epidemiol
16. Ford ES. Serum copper concentration and coronary heart disease among US adults. Am J Epidemiol
17. Ford ES. Serum magnesium and ischaemic heart disease: findings from a national sample of US adults. Int J Epidemiol
18. Marniemi J, Jarvisalo J, Toikka T, et al. Blood vitamins, mineral elements and inflammation markers as risk factors of vascular and non-vascular disease mortality in an elderly population. Int J Epidemiol
19. Cambien F, Warnet JM, Vernier V, et al. An epidemiologic appraisal of the associations between the fatty acids esterifying serum cholesterol and some cardiovascular risk factors in middle-aged men. Am J Epidemiol
20. Zureik M, Ducimetiere P, Warnet JM, et al. Fatty acid proportions in cholesterol esters and risk of premature death from cancer in middle aged French men. BMJ
21. Chappuis P, Jacqueson A, Ducimetiere P, et al. Magnesium, copper and zinc: relationships with ischemic heart disease risk factors. In: Halpern, Durlach, eds. Magnesium Deficiency
. First EurCongr.
Magnesium, Lisbon; 1983:142–145.
22. Speich M, Chappuis P, Robinet N, et al. Se, Zn, Mg, Ca, K, cholesterol, and creatine kinase concentrations in men during the 12 days after an acute myocardial infarction. Clin Chem
23. Arnaud J, Bellanger J, Bienvenu F, et al. Recommended method for assaying serum zinc with flame atomic absorption. Ann Biol Clin (Paris)
24. Salonen JT, Salonen R, Seppanen K, et al. Interactions of serum copper, selenium, and low density lipoprotein cholesterol in atherogenesis. BMJ
25. Reunanen A, Knekt P, Marniemi J, et al. Serum calcium, magnesium, copper and zinc and risk of cardiovascular death. Eur J Clin Nutr
26. Ma J, Folsom AR, Melnick SL, et al. Associations of serum and dietary magnesium with cardiovascular disease, hypertension, diabetes, insulin, and carotid arterial wall thickness: the ARIC study. Atherosclerosis Risk in Communities Study. J Clin Epidemiol
27. Hercberg S, Galan P, Preziosi P, et al. The SU.VI.MAX Study: a randomized, placebo-controlled trial of the health effects of antioxidant vitamins and minerals. Arch Intern Med
28. Menditto A, Morisi G, Alimonti A, et al. Association of serum copper and zinc with serum electrolytes and with selected risk factors for cardiovascular disease in men aged 55–75 years. NFR Study Group. J Trace Elem Electrolytes Health Dis
29. He JA, Tell GS, Tang YC, et al. Relation of serum zinc and copper to lipids and lipoproteins: the Yi People Study. J Am Coll Nutr
30. Finamore A, Roselli M, Merendino N, et al. Zinc deficiency suppresses the development of oral tolerance in rats. J Nutr
31. Philpott M, Ferguson LR. Immunonutrition and cancer. Mutat Res
32. Oberley LW, Buettner GR. Role of superoxide dismutase in cancer: a review. Cancer Res
33. Magalova T, Bella V, Brtkova A, et al. Copper, zinc and superoxide dismutase in precancerous, benign diseases and gastric, colorectal and breast cancer. Neoplasma
34. Johnson S. The possible role of gradual accumulation of copper, cadmium, lead and iron and gradual depletion of zinc, magnesium, selenium, vitamins B2, B6, D, and E and essential fatty acids in multiple sclerosis. Med Hypotheses
35. Das UN. Nutritional factors in the pathobiology of human essential hypertension. Nutrition
36. Riboli E, Slimani N, Kaaks R. Identifiability of food components for cancer chemoprevention. IARC Sci Publ
37. Bingham S, Riboli E. Diet and cancer—the European Prospective Investigation into Cancer and Nutrition. Nat Rev Cancer
38. Ames BN, Wakimoto P. Are vitamin and mineral deficiencies a major cancer risk? Nat Rev Cancer