The Science of the Total Environment, 42 (1985)
Elsevier Science Publishers B.V., Amsterdam
Cardiovascular disease (CVD) continues to be the major cause of mortality in developed countries. For the past two-and-a-half decades the inverse relationship between water hardness and CVD mortality has stimulated interest among epidemiologists, clinicians and experimental researchers. Much progress has been made in elucidating which element in the water may account for this situation.
After reviewing those elements found to have a role in cardiovascular function the authors present the epidemiological evidence and its consistency with recent findings: aside from various trace elements emphasis is placed on magnesium which is recognized as having a vital role.
Cardiovascular disease (CVD) is the leading cause of death among men in most industrialized countries, accounting for 37-56% of deaths in 27 countries studied by the World Health Organization (Table 1) (1). Ischemic heart disease (IHD) alone accounts for up to 24% of mortality from all causes in males aged 40-69 despite the decline in death rates from heart disease observed since the mid-1960’s in several countries. There is no fully satisfactory explanation for this recent downward trend, just as there was none for the upward trend prior to the mid-1960s. Factors which might explain this downward trend include a decrease in mortality from and/or a decrease in the incidence of disease. The former could arise from improved treatment such as coronary bypass surgery, coronary care units, and anti-arrhythmic medication, whereas a decrease in incidence could occur as a result of reduction of risk factors through better treatment of hypertension, changing dietary patterns, and lower prevalence of smoking.
Over the last decade a number of very large intervention trials (e.g., MRFIT (2), Oslo Study (3), North Karelia (4), European Collaborative Group (5) have been conducted to discover whether people can be persuaded to change their eating and living habits, and if they do, whether they will be less liable to heart attacks and live longer. Although the evidence tends to favor answering both these questions in the affirmative, it is not entirely clear. Reduction of risk factors was only achieved in a minority of the study populations, and in fact in some of the trials, was achieved to almost the same extent in the control populations. This type of result made it more difficult to answer the life expectancy question and also poses a further question, namely, which of the interventions were responsible for the risk factor reductions?
Trials of this sort are extremely expensive. They require costly equipment, highly specialized manpower and much labour and time to attain high compliance rates. The widespread adoption of the primary prevention techniques employed in these trials would similarly be extremely costly and questionably effective.
Compare this to the prevention of goiter by the introduction of iodized table salt. In the 1950 the prevalence of goiter was high in the low iodine region around the Great Lakes of North America and in much of Europe (6). Death rates were also high and in Ontario were twice as high for natives as compared to immigrants (Fig. 1).
Fifteen years later, iodine in table salt had obliterated these differences and led to a general decline in mortality.
Thus, here we have a disease causing considerable morbidity and mortality during middle age, which is attributable to mineral intake both during early life and in the more recent past. Furthermore, the mineral content of local drinking water provided an index of total intake. With the identification of the mineral came a low cost, highly effective intervention.
The possibility of an analogous situation occurring with ischemic heart disease has intrigued investigators and the cost-effectiveness of this type of primary prevention has motivated the search for elements which could be involved in the complex enzymatic and conduction systems in the myocardium (Table 2).
Improvements in chemical analytical techniques have in the last few decades permitted the study of the role of a number of elements in various aspects of cardiac function.
We propose to discuss the suggested roles and physiologic action of these elements, then look at the epidemiologic evidence in terms of specific hypotheses. We arbitrarily classify the elements according to the completeness of our knowledge of their role in cardiovascular function.
The only evidence that silver plays a role in cardiovascular disease is that increased concentrations of silver have been found in the aortas of atherosclerosis victims (7).
Although there is no direct evidence that molybdenum is essential in man, its presence as a component of several enzymes suggests that it may be. It is rapidly excreted, and there is very little body storage. Tissue determination on myocardial infarction victims has shown increased molybdenum levels in the serum and decreased levels in the injured heart tissue (8).
Nickel is one of the relatively non-toxic trace metals found in human tissues (unless inhaled in a form such as nickel carbonyl). Very few physiological or biological studies of this element have been done, Experimental evidence of nickel’s essential role has been obtained in the chick and the rat, suggesting that nickel may be essential for man (9). However, due to the extremely small amount needed, it is probably not a practical consideration.
Plasma nickel levels increase sharply following myocardial infarction, both in man and in experimental animals, but is not a specific effect (7,10,11,12,13).
Revis (14) has reviewed the physiologic experimental evidence associating vanadium with cardiovascular function. Vanadium inhibits myocardial sarcolemmal NA+ and K+-ATPase, produces a positive inotropic effect, and stimulates adenylate cyclase. However, endogenous vanadium levels in the heart are lower than would be required for the latter two effects, making it difficult to postulate a physiological role for this element. Epidemiologic evidence is also slight, but two studies have found negative correlations between vanadium and atherosclerotic heart disease (15,16).
Cobalt is physiologically active in man only when supplied in one particular form: cyanocobalamin or Vitamin B12. Man is dependent upon the food chain for his supplies of Vitamin B12 as he has no ability to incorporate cobalt into the vitamin by himself. When taken in small amounts over a long period of time and under certain conditions, (e.g., a state of malnutrition and especially if in an alcoholic vehicle) cobalt may induce cardiomyopathy (10). There were several cases of severe and lethal cardiomyopathy in men drinking excessive amounts of beer in Canada, until breweries stopped adding cobalt to beer (17,18).
Cobalt inhibits pyruvate and fatty acid oxidation and induces hyperlipemia (19). These effects can be prevented by Vitamin E and selenium (20).
Lithium is not an essential nutrient for man as far as is known. However, it has been used pharmacologically in the treatment of manic depressive psychosis for more than twenty years.
The evidence for a role in cardiovascular function comes from Voors who found an inverse correlation between the amount of lithium in drinking water and cardiovascular disease (15,21). Other authors have also reported findings that point to a protective effect of lithium against heart disease (22,23,24). This association could be accounted for by lithium's protective effect against several IHD risk factors. However, lithium's concentration in drinking water is very low, so any effect it could have would be minimal (15,25).
While selenium is considered an essential element in animals, evidence for a role in human nutrition is very scarce. Its essential role in man would be in the seleno-enzyme glucation peroxidose. However, the fact that selenium deficiency in animals can manifest itself in a wide variety of symptoms provides ample opportunity for speculation as to its effect on humans. Diseases postulated to be related to selenium deficiency range from cancer to dental caries (26).
The amounts of selenium needed to prevent deficiency are inversely related to dietary levels of Vitamin E. This fact indicates a co-factor role for selenium (27). Both deficiency and excess of selenium can lead to myocardial necrosis, elevation of serum transaminases and degeneration of vascular endothelium (28). However, the evidence of an effect in humans is still almost exclusively indirect and comes from epidemiologic observations. In the United States inverse correlations have been found between death rates for cardiovascular disease and mean levels of selenium in blood banks of 19 states (29). In selenium-deficient regions of China, there is a high prevalence of Keshan disease affecting young women and children. In a four year study conducted in China, in which 36,000 children received .5 to 1 mg sodium selenite per day, the incidence of Keshan disease in the study group was 94% less than in the control group (29). In Sweden, the lowest cardiovascular death rate is in the city of Malmo, which has a high tap water selenium content (30). In Mexico, patients given selenium-tocopherol capsules show a decrease in incidence of angina (31). However, in New Zealand no differences were found between blood selenium concentrations of four groups of hypertensive patients and their healthy controls; but selenium levels are low throughout New Zealand (31).
Selenium interacts strongly with numerous other elements. Its toxicity can be alleviated by antimony, arsenic, copper, germanium and tungsten, and it counters the toxicity of arsenic, cadmium, mercury, silver and thallium (26). Its interaction with magnesium is illustrated by the finding that cardiac-related sudden death in transported pigs can be prevented by supplements of either magnesium (32) or selenium (33). We consider the role of selenium in cardiac function an unresolved question. Because selenium has been so difficult to measure in drinking waters, it has been impossible to carry out large-scale epidemiological studies of this element.
Cadmium is an element toxic to man that is primarily an inhibitor of respiratory processes. Cadmium levels in tissue, particularly kidney cortex, increase with age, being almost nil at birth. Higher levels of cadmium are found in highly industrialized areas where the incidence of CVD is also high (7,34,35).
Revis (14) has reviewed the in vitro and in vivo effects of cadmium. In vitro, it decreases myocardial contractility, and prolongs the PR interval and AV conduction time. In vivo, rats exposed to cadmium develop prolonged PR intervals, and rats, rabbits and pigeons develop hypertension. However, the development of hypertension depends on the dosage: rats given drinking water containing 5 ppm cadmium dropped their blood pressure, while those fed water with 50 ppm cadmium sustained an increase in blood pressure.
In humans, cadmium decreases serum cholesterol but increases deposition of lipids in the walls of the aorta (36,37). A number of studies have shown it to be present in higher concentrations in the kidneys and urine of hypertensive patients as compared to normotensive controls (38,39,40), although at least one author, Morgan (41), failed to confirm these findings. Epidemiologically, a study of 29 North American cities showed a positive correlation between the levels of cadmium in the air and the incidence of hypertension and arteriosclerosis (42). However, neither cadmium workers nor Japanese with "itai-itai" (a cadmium-induced disease giving rise to spontaneous fractures of bone) seem to be at higher risk of developing hypertension (7,10).
Cadmium’s toxicity is potentiated by a low protein diet and counteracted by zinc and selenium (43). Indeed, it may be the cadmium-to-zinc ratio rather than the cadmium concentration which is the determinant of hypertension.
Lead is a persistent bioaccumulative body poison. Its main effects are hemolytic anemia, kidney disease and disturbances of the central nervous system (44,45). Stofen (46), after reviewing the effects of environmental lead on the heart, concluded that lead may play a role in cardiovascular disease. Electrocardiographic changes with prolonged PR intervals were reported on rats perfused with lead (47). Lead has also been reported to uncouple oxidative phosphorylation and inhibit the NA+, K+-ATPase (48). Through this mechanism or perhaps by interfering with the metabolism of cellular calcium, lead has been reported to affect both the electrical and contractile properties of the cardiovascular system (14). Lead has also been shown to induce systolic hypertension in pigeons fed .8 ppm lead for six months, but this effect is not seen in rats (14). Lead toxicity is potentiated by low calcium intake and its absorption from the gut is decreased by the presence of calcium. A recent review by Saltman (49) suggests a role for lead in hypertension because of its damage to renal tissue but the mechanism remains obscure. Decreased levels of plasma renin and aldosterone have been seen in lead poisoning. Increased levels of catecholamines together with decreased responses of baroreceptors have also been reported. Lead may also directly constrict arterioles and increase blood volume. In humans, a lead-induced cardiomyopathy has been described in “moonshine" drinkers (50).
In its trivalent form, chromium is a well-known essential element. Its major role is as a co-factor with insulin to maintain normal glucose tolerance. Evidence of a cardiovascular action came initially from Tipton (51) and Schroeder (52), who showed a significant decrease in tissue levels of chromium with age in populations having a high incidence of cardiovascular disease. Trace amounts fed to rats prevented formation of atheromatous lesions, decreased cholesterol levels and prolonged their overall life-span, while its deficiency caused an increase in the prevalence of aortic plaques (9,53,54). In humans, however, the evidence for a role in cardiovascular function is based on the well-known relationship between diabetes and arteriosclerosis.
Copper is an essential element necessary for optimal absorption and metabolism of iron, normal erythropoiesis and bone collagen formation. Numerous authors report increased serum copper levels in patients and animals with arteriosclerosis, hypertension or myocardial infarction (55-60). Harman (61) has also found that experimental animals who have higher intakes of copper have more arteriosclerosis. He has postulated that individuals prone to coronary artery disease may be identified by a high serum copper level. Rats deficient in copper have abnormal ECG’s, specifically ST changes and bundle-branch block (62). Animals deficient in copper die suddenly and the pathology shows myocardial degeneration, focal necrosis, fatty changes, subendocardial fibroplasia, aneurisms and infarction (62).
It has been hypothesized that an imbalance in zinc and copper metabolism may play a role in coronary heart disease (63). Copper and zinc are biochemical antagonists and both contribute to the control of the blood lipid levels. High zinc to copper ratio and high risk of mortality or hypercholesterolemia have.been found associated with diets high in fat or sucrose and in hypertension (62). Klevay suggests that the mean copper intake described in nutrition texts, of 2-5 mg/day, is out-of-date. He found a geometric mean intake of .82 mg/day. Daily requirements are 1.55-2.1 mg/day (62).
Zinc is essential for the life of all plants and animals. It is both a co-factor and component of many metallo-enzymes and it is present in all the tissues of the human body. Zinc deficiency in humans seems to be associated with poor growth and development, impaired wound healing, impairment to sensory perception and, in all probability, with congenital abnormalities (9,52).
A beneficial effect of zinc therapy has been reported in atherosclerotic patients (64,65,66). It has also been reported that zinc in rats reverses cadmium-induced hypertension. Several investigators have found decreased zinc concentration in plasma or serum after myocardial infarct (66). The levels return to normal values after a few days (67). It is hypothesized that this reflects a migration of the metal from plasma to tissues and that it is a non-specific reaction to myocardial infarction (68). However, Wester found that zinc concentrations decrease in injured heart tissue after myocardial infarction. More recently many of the effects of zinc have been attributed to the ratio of zinc to copper (63).
Calcium is essential for man. It is a main structural element necessary for blood clotting and for normal functioning of nerve tissue. It is possible that certain types of cardiac disease are aggravated by the lack of calcium, as it is required for muscle contraction and it has been shown to decrease serum lipid levels. Its role in cardiovascular function is however an indirect one. Calcium prevents the absorption and transfer of toxins from the intestine into the blood, and acts as a biological antagonist to magnesium (69).
The relationship between sodium and high blood pressure has been confirmed by clinical and experimental studies during the past 40 years. Clinical observations made as early as 1944 by Kemper indicated that a low sodium diet was helpful to hypertensive patients. Animal experimentation results were reviewed by Calabrese and Tuthill (70) who indicated that: 1) the greater ingestion of salt the more severe the hypertension, 2) the younger the animal at the time it is first exposed to a high salt diet the more it develops hypertension, 3) even a brief exposure of two to six weeks to a high salt intake early in life may influence the development of permanently elevated blood pressure, 4) genetic factors influence the individual response to salt. In animal studies as well as in human studies the sodium:potassium ratio rather than sodium concentration alone has been found to be related to hypertension. An association between blood pressure and the amount of salt in the diet has been observed in many different countries (71,72,73). It has, however, been difficult to find a consistent correlation between salt consumption and blood pressure within individuals in industrialized countries. The best evidence to date comes from a study by McGregor et al. (74) in which placebo and real salt tablets were used in a double-blind intervention trial. Blood pressure levels were consistently higher during the period in which the patients were taking the real sodium tablets. Urinary sodium levels confirmed increased sodium intake.
The role of magnesium in cardiovascular function will be discussed later.
As far as the epidemiologists are concerned, our interest in trace elements has been tied to what has become known as ‘the water story’.
Ever since Kobayashi (75), in 1957, noted a parallel between the geographic distribution of the acidity of water in Japanese rivers and the distribution of stroke, a major cause of mortality in Japan, an increasing number of investigators all over the world have attempted to elucidate and confirm the inverse association between water hardness and mortality, particularly mortality from cardiovascular causes.
As remarkable as the geographic diversity of these studies is the great diversity of the hypotheses that have been favored by different investigators, regarding both the identity of the water-borne factor and the nature of the disease or pathologic process induced by it. As well as mortality from cardiovascular disease, mortality from all causes, chronic bronchitis, cancer, and congenital malformations have been found to be associated with soft water.
Numerous mechanisms have been postulated to account for the association between minerals and cardiovascular disease (76). The two most likely are:
1. Minerals may contribute to the incidence of disease by increasing the risk of hypertension, i.e., a toxic effect mediated via hypertension.
2. Minerals may act to lower the case fatality rate by protecting people from sudden death.
The hypertension theory. The elements involved in the hypertension theory are cadmium, sodium and lead, thought to have toxic effects; and calcium, thought to protect from the effects of these toxic elements (77). Thus, there are several different hypotheses within this theory, each implicating different elements or combinations of them.
In 1965, Schroeder (78) hypothesized that the mode of action of soft water was through an increased risk of hypertension due to cadmium, and that cadmium would be leaching from pipes through the corrosive action of soft water. Since then, support for this hypothesis has come from a number of sources: 1) At the experimental level, Schroeder’s findings in rats have been replicated and extended to other animals (79). 2) At the clinical level (autopsies) evidence of increased cadmium or cadmium-zinc ratios in the kidneys of hypertensive patients have been confirmed in most studies (80,81) (although some associate low zinc concentrations more to renal damage than to essential hypertension) (82).
The possibility of cadmium acting through water, how received the strongest support from evidence that would implicate a toxic effect of cadmium with a protective effect of hard water: In this context, Perry (83) has shown that the hypertensive effect of cadmium in rats was inhibited when hard water was used as the vehicle for administration of the metal.
Lead has also been implicated in the hypertension theory. Support for its involvement comes from Beevers et al. (84), who examined the blood and tap-water lead levels of 135 pairs of age-sex-matched hypertensive and normotensive subjects (Table 3).
They found a significant excess of persons with high blood lead levels among hypertensives; among normotensives they found a positive correlation between blood lead and tap-water lead concentrations. Sharrett (85), and Folsom and Prineas (86) have reviewed the evidence implicating lead, and they conclude that if the association between lead and high blood pressure exists, it is probably only in areas with uncommonly high lead levels, such as where lead piping and corrosive waters coexist.
The hypertensive effects of sodium are well-established. What remains unclear is what role sodium may play in the water story. That salt in drinking waters may play a role is suggested from the findings of Steinback et al. (87): when the village of Juilovka was found to have one of the highest prevalences of hypertension in the world — 45% — they searched for an environmental factor and found that the water contained a sodium concentration 26 times greater than in the nearby villages in the Gurghiu Valley. However, it is more likely that dietary sodium is the important factor in inducing hypertension, with another waterborne factor having a protective effect. Crawford (88) has suggested the ratio of magnesium and calcium to sodium may be the important factor while Joossens (71) postulated that calcium itself protects against sodium.
The strongest epidemiologic support for the hypertension theory as an explanation for the water story comes from observations relating water hardness and blood pressure, again suggesting that calcium may have a protective effect against some of the toxic elements just discussed. Stitt (89), in the U.K. reported on a sample of 244 civil servants living in 6 hard water localities and a similar number living in 6 soft water localities. The results of these observations are shown in Figure 2. It can be seen that not only are both systolic and diastolic blood pressures higher in the soft water areas, but that this difference becomes more marked in the older age groups. However, Elwood (90) in Wales had previously found no significant differences in mean blood pressure on similar-sized population samples from hard and soft water areas (see Table 4).
Other types of studies also confirm an association between water hardness (implicating calcium) and hypertensive mortality rates. For example, Masironi (91) reported a study comparing the mortality of residents in four United States river basins, two with hard and two with soft water. He found the death rates from hypertensive heart disease to be 17% to 70% higher in the soft-water basin (Table 5).
It must be remembered that our main interest is not whether these elements can cause hypertension but whether they can explain the water story. Thus, we must consider the intake of the elements from water, relative to the intake from other sources, keeping in mind factors such as chemical state and absorption. Of the three toxic elements discussed, cadmium is perhaps the strongest candidate for a role in the water story, but cadmium intake from drinking waters (Table 6) does not seem to be substantial when compared, for example, to amounts absorbed from cigarettes (92).
Hence, its role would have to be postulated in terms of interaction with some other element, perhaps zinc, more abundant in hard water than in soft water. Epidemiologically, the health effects of two elements, one protective and the other toxic, would be indistinguishable from the effects of the protective element alone.
In summary, although there is good evidence that certain trace elements and sodium may lead to hypertension, they do not seem to adequately explain the water story. Therefore, in Canada we are systematically pursuing the second theory — the idea of an agent, present in hard-water areas, which is protective against premature death, especially sudden death.
The sudden death theory. In Canada we have conducted two studies in which we progressively eliminated candidates from our original list of elements and zeroed-in on those elements more likely to be the water borne factor.
The first is a tissue study which provides for the comparison of the metal content of the heart and of control muscles in residents of hard- and soft-water areas, and specifically those dying from myocardial infarction and from accidents (93). This last group is taken as being representative of healthy subjects. For the interpretation of this study, we established the following criteria to assess the various elements implicated (94):
1. The tissue concentration must differ between cardiac deaths and accidental deaths.
2. The difference in tissue concentration must be consistent with the postulated biological effect of a particular element.
3. The difference in concentration among control patients from each type of area must be consistent with the mineral content of the water consumed.
On this basis, we were able to eliminate zinc and chromium, because the concentrations of these two elements are not different in healthy and cardiac subjects (Table 7).
We also eliminated lead, because the results were not consistent with its postulated toxicologic effect.
Similarly, we eliminated calcium and cadmium, which were found in increasing concentrations in the hearts of subjects dying from myocardial infarction as compared with accident cases, and because the concentration of these elements does not differ among healthy residents in soft- and hard-water areas (Table 8).
However, the increased concentrations of calcium in the myocardium, while inconsistent with the hypothesis of a protective effect of calcium, may substantiate findings such as those of Van Barneveld (95) who noted sudden death among mice receiving magnesium-poor, calcium-rich diets but no deaths among mice receiving diets containing any other combinations of magnesium and calcium.
We are now left with two possible elements, copper and magnesium.
Copper levels are lower in the hearts of people who have had heart attacks but they are also lower in healthy residents of soft-water areas (Table 9).
Here, there is a paradox, because the concentration of copper in soft waters was about double that in hard waters, so one would expect the tissue copper concentrations to be higher in soft water areas.
Tissue magnesium levels showed the greatest difference between the subjects dying from myocardial infarction and accident cases, a difference that is present in both hard- and soft-water areas (Table 10).
Moreover, the hearts of healthy residents in hard-water areas also contained more magnesium than those in soft-water areas. The data from this study can also be rearranged to portray variation in individual risk as a function of individual myocardial magnesium level (Fig. 3). Plotting the estimated relative risks on a logarithmic scale and fitting them to a regression line shows an impressive regularity of the data, despite small sample sizes (only 122 subjects for all six datapoints). Even more remarkable is the range of variation in apparent risk — by a factor of approximately 200.
Thus, the tissue study implicates magnesium as the most likely protective factor in hard water. The second study is an ecologic study in which we examined the mineral content of drinking water in relation to mortality (96). We randomly selected tapwaters from more than 500 localities across Canada, each having a communal water supply, and we measured 15 candidate elements, chosen on the basis of suggestions by previous investigators (97).
Again, we went through a sequence of progressive eliminations, based on fixed criteria:
1. The suspected element must be found in the waters consumed.
2. The trend seen with a particular element must be compatible with the geographical distribution of hardness; that is, if the element is protective, it must be present in larger quantities in hard waters; if the element is toxic, it must be found in larger amounts in soft waters.
3. A particular element must be present in sufficient quantities in the waters consumed so as to make an appreciable contribution to total dietary intake.
On this basis we could eliminate silver, selenium, molybdenum, and antimony — all supposedly protective elements — because they were detected in less than 10 percent of the localities where waters were tested (Table 11).
We also eliminated nickel, lead, zinc, and cadmium, because these elements were not consistently related to the hardness-softness gradient.
In addition we eliminated calcium and cobalt, because intake of these elements from drinking water was too small in comparison with intake from other sources.
Among the remaining five elements, there are three — mercury, chromium, and copper — that can also be excluded because of the lack of consistent supporting evidence from studies in Canada and elsewhere (Table 12).
In fact, mercury, found in 60 percent of the sampled waters, has a gradient of decreasing concentration with increasing hardness. This could fit the hypothesis of a toxic element, contained in soft waters, that provides an average of 10 µg of mercury per day (i.e., five times the amount received from hard waters). However, in the correlation studies, mercury does not correlate significantly with mortality; in the few instances in which it does, it has a negative sign indicating a protective effect.
Chromium, found in 16 percent of the sampled waters, is found in its highest values in hard waters. If the form of chromium found in water is optimal for absorption, this could account for a large proportion of the daily chromium intake in some hard-water areas; however, mortality correlations in most countries do not provide any conclusive evidence of a water-borne chromium effect.
Copper also deserves more attention. It is found in decreasing amounts with increasing hardness, and its highest values are found in the softest groups of water. However, copper concentration is not significantly associated with mortality.
This leaves magnesium and lithium (Table 13).
Magnesium is the strongest correlate in our correlation matrix, that is, stronger than either calcium or hardness, with lithium ranking fourth. Of course, it must be stated that all of these factors are highly intercorrelated (Table 14).
Thus our ecological study also implicated magnesium as the most likely candidate for the water-borne factor. Confirmation comes from other sources as well, the most recent being a report by Leary et al. (98) in South Africa.
Does magnesium then satisfy all the criteria? What is its dietary intake in relation to its requirements? The daily requirement of magnesium is considered to be 300 mg for females, 350 mg for males and 450 mg for pregnant women. Average intakes in various studies, ranges from 6-35% lower than the recommended allowances, and for pregnant women may be as low as 57% of that required (Table 15).
Thus the amount of magnesium ingested from water could make a critical difference. Does it? Figure 4 shows magnesium intake as a function of water hardness in four geographical areas of the United States. The association appears striking, adding evidence that sufficient magnesium can be ingested from hard water to make the difference between inadequate and adequate intake. As well, Binnerts et al. (99) have suggestive evidence that magnesium from water is better and more rapidly absorbed than that from food. Thus in terms of intake, magnesium seems to satisfy all the necessary criteria.
What then, is the clinical and experimental evidence that magnesium is likely to reduce cardiovascular mortality? The literature in this field is voluminous and beyond the scope of this paper but certainly supports the epidemiological evidence (100). Some of the recent more important findings are that magnesium supplementation in sub-pharmacologic doses stabilizes arrhythmias (69), lowers blood pressure in hypertensives (101), increases cardiac output (69) and reduces infarct size (102). Subacute magnesium deficiency increases digitalis toxicity and the multiplicity and severity of side effects from diuretics (103) as well as causing changes in ECG patterns (104,105). At the cellular level magnesium alters sodium transport (106) and seems to safeguard against the influx of sodium and calcium into cells (69). It also prevents catecholamine effects (107), being protective against cardiotoxic substances and stress (108-110).
Because of the long-recognized crucial importance of magnesium as a vital cofactor for various enzymatic and metabolic processes, a myocardium that is deficient in magnesium is likely to be particularly vulnerable to a variety of stressful situations. In fact, myocardial magnesium depletion can be regarded as the "sensitizer" (111) in Hans Selye's (112) "sensitizer-challenger" concept in which the "challenger" is most probably a sudden surge of stimulus at the cardiac site.
Thus, epidemiological considerations, when viewed together with the clinical and experimental work conducted in the area of stress (we refer here to stress imposed on cardiac function), suggest an hypothesis capable of tying together most of the findings. This is the concept of the unstable, vulnerable heart, which — possibly because of some metabolic deficiency (likely related to magnesium deficiency) — is incapable of responding to a sudden upsurge of functional demand. This can conceivably happen in several situations in which the cause of death would not be classified as cardiac on the death certificate particularly in cases such as chronic bronchitis and maybe even some of the cancers. The selective magnesium depletion found in the myocardia of our tissue studies, in the study by Behr and Burton (113) in England, and by various other studies in experimental animals, is in line with this theory. One cannot expect that the epidemiologic evidence collected up to now would be able to confirm an hypothesis such as this. Studies will have to be designed to test the specific hypothesis, and the ultimate answer will hopefully emerge during a magnesium intervention trial.
What then are our conclusions? If the myocardium can 1) become selectively depleted in magnesium, and 2) can thereby become impaired in its responses to a sudden demand for increased cardiac output, we have the prerequisites for an increased likelihood of fatalities, probably sudden fatalities, whether or not they will be ultimately ascribed to the cardiac domain.
In terms of the Canadian experience, such an hypothesis is fully applicable to the so-called "water story". It can also be related to the inadequate intake of magnesium from dietary sources, the so-called "empty-calories diet" of the modern-day world.
1 Z. Pisa and K. Vemura, World Health Organization Quarterly Report, 35(1982) 12-21.
2 MRFIT Research Group, JAMA, 248(1982) 1465-1477.
3 I. Hjermann, I. Holmes, K.V. Byre, P. Leren, Lancet, 2(1981) 1301-1310.
4 J.T. Salonen et al., Br. Med. J., 2(1979) 1179-1183.
5 WHO European Collaborative Group. Europ. Heart I., 3(1982) 184-190.
6 L.C. Neri and D. Hewitt, Hardness of Drinking Water and Public Health.
7 R. Masironi, Bull. WHO, 40(1969) 305-312.
8 Trace Elements in Human Nutrition. WHO Techn. Rep. Sev., (1973) 532.
9 J.C. Reinhold, Clinical Chem., 21(1975) 476.
10 R. Masironi, Europ. Colloquium on the Hardness of Drinking Water and Public Health, Abstracts of Papers, Part 8, Luxembourg, 1975.
11 V. Ryabova, Vop. Biol. Med. Khim., Mater. Nauch. Brockhim. Konf. 1st, 52(1968).
12 F.W. Sunderman et al., Geoch. Env. in Health and Disease, N.Y. Acad. Sci., 199(1972) 300.
13 V.M. Sakharchuk et al., Kardiologiya, 12(1972) 131.
14 N.W. Revis, in E,W. Van Stee (Ed.) Cardiovascular Toxicology, New York, 1982, pp. 365-375.
15 A.W. Voors; Am. J. Epid., 92(1970) 164.
16 H. Schroeder, Arch. Environ. Health, 28(1974) 303.
17 Y. Morin and P. Daniel, Am. J. Cardiol., 19(1967) 143-145.
18 C.S. Alexander, Am. J. Med., 53(1972) 395-417.
19 G.S. Wiberg, I.C. Munro and A.B. Morrison, Can. J. Biochem., 45(1967) 128.
20 J.F. Van Vleet, A.H. Relon and V.J. Ferrans, Am. J. Vet. Res., 38(1976) 991-1002.
21 A.W. Voors, Lancet, 2(1969) 1337-1339.
22 E.B. Dawson and W.J. Mcganity, 9th Int.. Congress of Nutr. Abstracts, Mexico City 1972 p. 51.
23 M.L. Sievers and H.L. Cannon, Trace Subst. in Environ. Health, 7(1973) 57.
24 R.A. Polumbo et al., Proc. Soc. Exp. Biol. Med., 142(1973) 1200.
25 P.A. Blachly, New Eng. J. Med., 281(1969) 682.
26 J.R. Marier and J.F. Jaworski, NRCC, No. 20643.
27 W. Hoekstra, in Hoekstra et al. (Eds.), Trace element metabolism in animals, Baltimore, p. 61-77.
28 H. Vokal-Borek, Selenium USIP Report No. 79-16, Institute of Physics, Univ. of Stockholm (1979).
29 R.J. Shamberger, Sci. Total Environ., 17(1981) 59-74.
30 H. Bostrom and P.O. Wester, Acta Med. Scand., 181(1967) 465.
31 C.D. Thomson and M.F. Robinson, Am. J. Clin. Nutr., 33(1980) 303-323.
32 H.G. Classen et al., Arch. Pharmacol. (Suppl.), 307(1979) 41.
33 D.E. Ullrey, J. Anim. Sci., 51(1980) 645-651.
34 H.D. Livingston, Trace Subst. Environ. Health, 5(1971) 399.
35 H.M. Perry Jr. et al., Trace Subs. Environ. Health, 8(1974) 51.
36 H.A. Schroeder, J. Nutr., 86(1965) 51-66.
37 H.A. Schroeder and J.J. Balassa, Am. J. Physiol., 209(1965) 433-437.
38 J.N. Mackenzie and S. Kay, New Zealand Med. J., 78(1963) 68.
39 J. Iener and B. Bibr, Lancet, 1(1971) 970.
40 V. Karlicek et al., Cas Lek ces, 110(1971) 756.
41 J.M. Morgan, Arch. Int. Med., 123(1969) 403.
42 R.E. Carrol, JAMA, 198(1966) 267-269.
43 W.K. Tai, Water Hardness, Toxicity of Metals and Their Relationship to Cardiovascular Disease. Division of Public Health Engineering, Department of National Health and Welfare, Ottawa, 1970.
44 Biologic Effects of Atmospheric Pollutants — Lead: Airborne Lead in Perspective, National Academy of Sciences, Washington, D.C., 1972.
45 J.J. Chisolen Jr., Sci. Amer., 1971, 335.
46 D. Stofen, J. Mol. Cell Cardiol., 5(1974) 285-290.
47 S.J. Kopp et al., Toxicol. Appl. Pharmacol., 46(1978) 475-487.
48 B. Nechay and J.P. Saunders, J. Environ. Path. Toxicol., 2(1978) 283-292.
49 D. Saltman, Annals Int. Med., 98(1983).
50 S.K. Asokan, M.J. Frank and A.C. Wilham, Am. Heart J., 84(1972) 13-18.
51 I.H. Tipton et al., Health Phys., 2(1965) 40-451.
52 H. Schroeder, Circulation, 35(1967) 570-582.
53 H.W. Staub et al., Science, 166(1969).
54 Schroeder et al., J. Chronic Dis., 23(1970) 123-142.
55 S.S. Adelstein et al., NEJM, 255(1956) 105-109.
56 A. Hanson and G. Biorch, Acta. Med. Scand., 157(1957) 493-502.
57 E.L. Kanabrocki et al., J. Nucl. Med., 8(1967) 166-172.
58 V.A. Kondurtsev, Ter. Arkh., 41(1968) 62.
59 I.D. Rachinskii, Kardiologya, 9(1969) 84.
60 A.V. Aronov, Kardiologya, 13(1973) 43.
61 D. Harman, Circulation, 38, Supp. 6 (1968) 8.
62 L. Klevay, J. Environ. Path. and Toxicol., 4-2, 3(1980) 281-287.
63 L. Klevay, Amer. J. Clin. Nutr., 28(1975) 764-774.
64 J.H. Henzel et al., Trace Subst. Environ. Health, 4(1971) 336.
65 J.P. Isaacs et al., Trace Subst. Environ. Health, 5(1972) 313.
66 L.D. McBearn, Clin. Chem. Acta., 50(1974) 43-51.
67 A.M. Handjani et al., Chest, 65(1974) 185-187.
68 P.O. Wester, Acta. Med. Scand., 178(1965) 765-788.
69 O. Lehr, Magnesium Bull., 3(1981) 178-191.
70 E.J. Calabrese and R.W. Tuthill, Arch. Environ. Health, 32(1977) 200-202.
71 J.V. Joossens, Triangle (Engl. Ed.) 12(1973) 9-16.
72 A.G. Shaper et al., Afr. Med. J., 46(1969) 282-286.
73 N. Sakaki, Geriatrics, 19(1964), 735-744.
74 G. MacGregor et al., Lancet, 2(1982), 351-355.
75 J. Kobayashi, Ber. Ohara. Inst. Landwirtsch. Biol. Okayama Univ., 11(1957) 12-21.
76 L.C. Neri, D. Hewitt, G.B. Schreiber, Am. J. Epidemiol., 99(1974) 75-88.
77 L.C. Neri, H.L. Johansen, N.Y. Ann, Acad. of Sciences, 304(1978) 203-219.
78 H. Schroeder, J. Chronic Dis., 18(1965) 647-656.
79 G.S. Thind, J. Air Pollut. Control Assoc., 22(1972) 267-270.
80 H.M. Perry, J. Am. Dietetic Assoc., 62(1973) 631-637.
81 J.B. Lener, Lancet, 1(1971) 970.
82 G.S. Thind and G.M. Fischer, Clin. Sci. Mol. Med., 46(1974) 137-141.
83 H.M. Perry et al., 8th Annual Conference on Trace Substances in Environ. Health, Columbia, Mo. 1974.
84 O.G. Beevers et al., Lancet, 2(1976) 1-3.
85 A.R. Sharrett, Amer. J. Epid., 110(1979) 401-419.
86 A.R. Folsom and R.J. Princes, Amer. J. Epid., 115(1982) 818-832.
87 M. Steinbach et al., Rev. Roum. Med., 13(1975) 261-263.
88 M.D. Crawford, Proc. Nutr. Soc., 31(1972) 347-353.
89 F.W. Stitt et al., Lancet, 1(1973) 122-126.
90 P.C. Elwood et al., Br. Med. J., 2(1971) 362-363.
91 R. Masironi, Bull. WHO, 43(1970) 687-691.
92 G.P. Lewis et al., J. Chronic Dis., 25(1972) 717-726.
93 T.W. Anderson et al., Can. Med. Assoc. J., 113(1975) 199-203.
94 L.C. Neri, J. Am. Water Works Assoc., 67(1975) 403-409.
95 A.A. Van Barneveld, C.J.A. Van den Hamer and J.P.W. Houtman, Trace Subst. Environ. Health, 16(1982) 196-204.
96 L.C. Neri, J.S. Mandel, D. Hewitt, Lancet, 1(1972) 931-934.
97 L.C. Neri, H.L. Johansen, F.D.F. Talbot, Chemical Content of Canadian Drinking Water Related to Cardiovascular Health, Ottawa, 1977, p. 223.
98 W.P. Leary and A.J. Reyes, SA Med. J., 64(1983) 697-698.
99 W.T. Binnerts et al., In Trace Element — Analytical Chemistry in Medicine and Biology, Vol. 2, Berlin 1983, 87-93.
100 J.R. Marier, L.C. Neri, T.W. Anderson, Water Hardness, Human Health and the Importance of Magnesium, NRCC No. 17581, 1979.
101 T. Dyckner, O.P. Wester, Br. Med. J., 286(1983) 1867-1849.
102 B. Morton, Magnesium Bull., 3(1981) 192-194.
103 L. Cohen and R. Kitzis, JAMA, 251(1984) 730.
104 K.S. Drasner et al., Can. Anesth. Soc. J., 28(1981) 329-333.
105 C. Wan-Chun et al., Am. Heart J., 104(1982) 1115-1116.
106 P. Fisher and A. Giroux, Nutr. Res., 4(1984) 51-57.
107 A. Ebel and T. Gunther, J. Clin. Chem. Clin. Biochem., 21(1983) 249-265.
108 B. Malviel-Shapiro, S. Afr. Med. J., 32(1958) 1211-1214.
109 R.S. Parsons et al., Med. Proc., (1960) 479-481.
110 A. Thurnherr and J. Kock, Med. Wochens Chr., 92(1962) 949-956.
111 L.C. Neri, J.R. Marier, H. Nato. (Ed.) Nutrition and Heart Disease, N.Y. 1982, 81-96.
112 H. Selye, Anaphylactoid Edema, St. Louis Missouri WH Green Inc., 1968.
113 G. Behr and P. Burton, Lancet, 2(1973) 450.
114 J.R. Marier, Rev. Canad. Biol., 37(1978) 115-125.
115 J.R. Marier, Magnesium, 1(1982) 3-15.
116 D. Hewitt, L.C. Neri, J. Environ. Path. Toxicol., 4(1980), 51-63.
This page was first uploaded to The Magnesium Web Site on September 18, 2002