MAGNESIUM DEFICIENCY IN THE PATHOGENESIS OF DISEASE

Early Roots of Cardiovascular, Skeletal
and Renal Abnormalities

Mildred S. Seelig, M.D., M.P.H., F.A.C.N.

Goldwater Memorial Hospital
New York University Medical Center
New York, New York
1980

Introduction: Chapter 1

1

Introduction: Consideration of Epidemiologic Factors

1.1. Ischemic Heart Disease

The most alarming trend in the past half-century has been the sharp increase in sudden deaths from ischemic heart disease (IHD), particularly in middle-aged men, and the increasing number of younger men who suddenly develop myocardial infarctions, cardiac arrhythmias, or arrests. That men in the prime of life are thus afflicted is the dramatic and tragic tip of the iceberg. Underlying these catastrophes is the widespread increase in incidence of atherosclerosis in young age groups, and in myocardial hyperexcitability and cardiomyopathy without notable coronary atherosclerosis. It is proposed that magnesium deficiency or loss may be a common etiologic factor in the increased incidence of sudden infant deaths, infantile myocardial infarction and arteriosclerosis, and the disease that becomes manifest later in life. It is also suggested that magnesium deficiency might also cause or predispose to some skeletal and renal diseases, all of which can coexist.

The cardiac problem in men has been deemed of sufficient magnitude as to be termed an epidemic that has been increasing, particularly since the middle 1930s. It has led to widespread institution of therapeutic and prophylactic regimens on the basis of suggestive findings. For example, young women have a significantly lower incidence of ischemic heart disease than do young men (Fig. 1-1). Because their α/β-lipoprotein ratios differ from those of the more susceptible young men and especially from those of patients with peripheral or coronary atherosclerosis, there was a period during which estrogens were widely used in the treatment of patients with myocardial infarctions and given prophylactically to high-risk (hyperlipidemic) men and postmenopausal women. This approach has been largely discontinued, predominantly because of the resultant increase in risk of thrombosis Another approach that was given a trial period was administration of excesses of unsaturated fatty acids; the incidence of atherosclerosis and IHD is lower in countries where more vegetable oils than saturated animal fats are consumed. A modification of the fatty-acid-supplement regimens that has been receiving extensive clinical trial is to replace saturated with unsaturated fats. This approach has lowered blood lipids, but not the incidence of IHD. Because altering fat intakes of patients with established hyperlipidemia and atherosclerosis has not reduced the mortality from IHD, it has been recommended that the time to institute such a dietary modification might be in early infancy, a suggestion that has been disputed.

1.2. Concomitant Cardiovascular, Skeletal, and Renal Diseases

Among women, the incidence of atherosclerosis and IHD increases with age, especially after the menopause, often in association with osteopenia or with calcific renal disease. The combined problem of bone wasting and extraskeletal calcification (particularly renal and cardiovascular) is also encountered in renal osteodystrophy and in other conditions associated with hyperparathyroidism and phosphate treatment of hypercalcemia.

Rarer forms of osteopenia, usually found in association with cardiac anomalies, arteriosclerosis, and renal calcinosis, are seen in infants of low birth weight, or who have ontogenesis imperfecta or hypophosphatasia. The more common, but not widely known, arteriosclerosis and IHD of early infancy is also usually accompanied by renal calcinosis, as is the later form that is accompanied by hyperlipidemia, hypertension and atherosclerosis. The latter type-some forms of which are associated with aortic and pulmonary stenoses and atresias, and with endocardial fibroelastosis-has been attributed to hypervitaminosis D (Seelig, 1969b; Seelig and Haddy, 1976/1980) which contributes to loss of magnesium. These conditions are stressed in this volume because they support the supposition that atherosclerosis (and some renal and skeletal diseases) have their roots early in infancy and have put the onus on the absolute or conditioned magnesium deficiency that has become a problem during this century.

Magnesium plays an important role in maintaining the integrity of the myocardium, kidneys, and bone. Its deficiency has been shown to cause cardiomyopathy in several animal species, and to intensify myocardial lesions caused by a variety of modalities. Its deficiency has caused arteriosclerosis and has intensified formation of atheromata, or arteriosclerosis, thrombosis, and even myocardial infarction, induced by atherogenic diets, high intakes of vitamin D, calcium, phosphate, and fat. Its deficiency has caused renal lesions and intensified damage produced by vitamin D, calcium, and phosphate. And its deficiency has been implicated in some forms of bone damage. Magnesium supplementation has prevented or reversed some of the lesions in the experimental models and been used clinically in cardiovascular disease and urolithiasis.

1.3. Changing Magnesium, Vitamin D, and Phosphate Intakes

Examination of the changing nutritional intakes in America, particularly from the middle 1930s is disconcerting in light of these experimental findings. Although magnesium intakes have been gradually falling since the beginning of the century, there were sharply increased intakes of nutrients that increased its requirements [particularly high vitamin D and phosphorus intakes (Seelig, 1964, 1971) subsequently (Fig. 1-2)]. The rise in vitamin D intake began when the addition to each quart of milk of a sufficient amount (400 IU) to cure, rather than merely to prevent, rickets became widespread from the mid 1930s and was made mandatory in most states from the 1940s to 1950, either replacing cod liver oil, or taken in addition to it (Baldwin, 1953; Seelig, 1969b, 1970b). Fortification of many foods in addition to milk, including milk flavoring, oleomargarine, breakfast cereals, or "substitutes," led the Committee on Nutrition of the American Academy of Pediatrics to express concern about the total daily intake of vitamin D in the United States, which they calculated might range from 600 to 4000 IU/day from marketed fortified products (Table 1-1). A survey of 1000 Canadian children from 1 week to 51/2 years of age showed that 70% consumed more than 400 IU, and 30% consumed over 1000 IU of vitamin D daily (Broadfoot et al., 1972). Table 1-2 depicts the sources of vitamin D among those receiving over 1000 to 1800 IU of vitamin D per day. The major source of phosphorus derives from soft drinks that contain phosphoric acid, the consumption of which has been rising markedly in the last quarter of a century (Henderson, 1972; Lutwak 1974).

Although it is generally believed that the rise in blood lipids is due to increased intakes of saturated fats during this century, and that sugar consumption has also increased substantially, comparison of per capita intakes from 1909 to 1965 shows relatively minor changes (Fig. 1-3). The average daily fat intake rose from 112 to 132, but most of the increase has been in unsaturated fatty acids. The total carbohydrate intake dropped from 492 to 374, so that the greater percentage increase of sugar in 1965 reflects an increase of about 40 grams daily. Probably the sugar intake has risen more since the 1965 value (Fig. 1-2) among those who drink larger quantities of sugar-sweetened, phosphorus-containing soft drinks.

Largely disregarded is the possibility that the hyperlipidemia associated with atherosclerosis might be caused by hypervitaminosis D, which also causes hypertension (Linden, 1977; Seelig and Haddy, 1976/1980), as well as the more widely recognized complications; cardiovascular and renal damage, and hypercalcemia (Seelig, 1969b). Much of the clinical data on the cardiovascular, skeletal, and renal damage caused by vitamin D derives from the use of massive doses of vitamin D a quarter of a century ago in the treatment of such diseases as rheumatoid arthritis, and from the lesser overdosage of European children at a time when administration of up to 4000 IU/day was not uncommon (Table 1-3; Seelig, 1969b). The sharp rise in vitamin D intake depicted for the 1947-1957 segment of Fig. 1-1 is presumed because of the probable consumption of large quantities of milk by the college students studied-an impression suggested by their high calcium intake (Scoular et al., 1957), in contrast to the lower intake noted in a general diet survey (Friend, 1967). Since the amount of vitamin D needed by most adults is considered so small as to be met by exposure to sunlight and by ingestion of natural (unfortified) foods (Food and Nutrition Board, 1968), such high intakes must be considered well into the toxic range. As long ago as 1932, L. Harris reported that in the human, the toxic dose of vitamin D is not far removed from the therapeutic (antiricketic) dose. Stewart (1964) reported that there is a narrow toxic-therapeutic ratio. Furthermore, even most infants are protected against rickets by as little as 100 IU of vitamin D daily (Fraser, 1967), whereas a survey of young Americans showed that 50% ingested 400-800 IU daily, 10% usually consumed over 1000 IU daily, and occasionally as much as 2900 IU were taken (Dale and Lowenberg, 1967). Epidemiologic data have correlated moderately high vitamin D intake with increased incidence of myocardial infarction, renal calcinosis, and urolithiasis (Linden, 1974a,b). In northern Norway, where intake of natural foods rich in vitamin D is common, the incidence of hypercholesterolemia and susceptibility to sudden death from ischemic heart disease and to calcific renal diseases, two conditions which are often found in the same patient (Linden, 1972, 1975/1977; Westlund, 1973), seems to be related to the amount of vitamin D ingested and to the individual sensitivity to solar irradiation. Since magnesium deficiency is also associated with abnormal lipid distribution, and vitamin D excess causes magnesium loss, interrelations of protracted high intakes of vitamin D with magnesium requirements, and with the cardiovascular and renal lesions of each imbalance, deserve study (Seelig, 1977).

Like magnesium deficiency and hypervitaminosis D, excess phosphate has also been implicated in cardiovascular, skeletal, and renal damage. The nature of the pathologic changes produced by dietary excesses of phosphorus depends upon its ratios to both calcium and magnesium. Figure 1-2 shows that the phosphorus intake increased sharply in the college studies during the periods analyzed in 1947-1957 (Scoular et al. 1957), and in the most recent survey of college diets (Walker and Page, 1977). The lower phosphorus level entered in the 1960-1967 block of columns derives from an extensive metabolic balance study in several colleges (Leverton et al., 1962). One can speculate that during these strictly controlled periods there was likely to have been less consumption of soft drinks containing phosphoric acid than during the self-selected dietary intakes reflected in the college diet surveys.

The recommended phosphorus/calcium (P/Ca) ratio is 1.5/1 (U.S. Department of Agriculture Report, 1972). In 1932-1939, the P/Ca ratio was about 1.2/1; it was estimated to be rising to as much as 4/1 among those who substitute sodas for milk (Lutwak, 1974). This shift in ratios was stressed as potentially harmful to bones, as a result of secondary hyperparathyroidism, on the basis of the effect of the osteopenia produced by comparable P/Ca dietary ratios in several species of animals, up to the monkey (Krook and Barrett, 1962; Krook et al., 1963, 1971; Henrikson et al., 1970; Draper et al., 1972; Krishnarao and Draper, 1972; Krook et al., 1975).

However, the most recent dietary survey of college diets from fifty colleges (M. Walker and Page, 1977) showed that the mean P/Ca ratio was about 1.5/1, both phosphorus and calcium intakes having risen to 1200 and 1700 mg/day, respectively. What had dropped was the magnesium intake-to a mean of 250 mg/day. Such diets provide dietary ratios of Ca/Mg and P/Mg of almost 5/1 and almost 7/1, respectively. Since an excess of either phosphorus or calcium has been shown to increase magnesium requirements and to intensify signs of magnesium deficiency (Reviews: Seelig, 1964, 1971), such a dietary pattern-particularly when accompanied by high vitamin D and phosphate intakes by many-can be expected to produce either absolute or relative magnesium deficiency.

1.4. Sex Difference in Magnesium Retention

Analysis of published metabolic balance studies (such as are done to establish a nutritional requirement, an amount sufficient to maintain equilibrium) has shown that young men require more magnesium in mg/kg/day than do young women (Fig. 1-4) (Seelig, 1964). The studies analyzed had been obtained from throughout the world, and showed that young Americans tended to ingest less magnesium on self- selected diets than did Orientals and, on average, tended to be in negative balance. This was particularly so for the young men, who on the average excreted more magnesium than they ingested on the typical American intake of 4-4.9 mg/kg/day. Young women on that typical intake, on the other hand, tended to remain in equilibrium. The typical magnesium intake of the Orientals studied was between 7 and 10 mg/kg/day, and positive balance or equilibrium was the rule. In deriving the recommended magnesium intake from the data analyzed, the intake was selected at which equilibrium or positive balance was reached in at least three-fourths of the subjects. On this basis, the minimal daily requirement is 6 mg/kg/day. For a 140-lb woman, this comes to 385 mg of magnesium daily; for a 185-lb man, at least 500 mg/ day. Americans, and others in industrialized countries, tend to ingest diets rich in other nutrients (fat, protein, sugar, phosphorus, and vitamin D), all of which increase magnesium requirements (Seelig, 1964, 1971; Lindeman, 1976/1980). In addition, moderate to heavy ingestion of alcohol (even as "social" drinking) is not uncommon, and alcohol is magnesuretic (McCollister et al., 1958, 1963; Kalbfleisch et al., 1963). Thus, a magnesium intake of 7-10 mg/kg/day might be preferable. On this basis, a 185-lb man might require 580-800 mg/day of magnesium, probably approximately twice as much as his diet normally delivers. Possibly a woman (unless she is pregnant or lactating) requires somewhat less. The most recent survey of college students (from 50 colleges) shows that less than the modest officially recommended amount [300 mg for women; 350 mg for men (Food and Nutrition Board, 1974)] is the amount usually ingested (M. Walker and Page, 1977). Actually, the mean daily magnesium intake of the college students (250 mg) may well be no more than half the amount required by the young women; it may be as little as one-half to one-third the amount needed by large, athletic young men. In contrast to their inadequate magnesium intake, they ingest one and a half times the recommended amount of calcium and twice the phosphorus allowance. Consumption by young college women of diets that provide suboptimal amounts of magnesium is not unique to the 50 colleges surveyed. N. Johnson and Philipps (1976/1980) surveyed the diets of pregnant women from different economic brackets, and found that their magnesium intakes ranged from 103 to 333 mg/day, with an average of 204 mg daily, an amount grossly inadequate for pregnant women. Ashe (1979) confirmed the inadequacy of prenatal magnesium intakes of 10 healthy white women from private practices in Tennessee by 7-day metabolic balance studies done at intervals throughout pregnancy. Their mean daily magnesium intakes were only 60% the recommended 450 mg/day, and mean balances were -40 mg/day. Only 3 of the 47 periods were positive. The investigators suggested that high calcium, phosphorus, and protein intakes might have intensified the severity of the negative magnesium balances. The significance of such low magnesium intakes during gestation, as regards the cardiovascular, skeletal, and renal status of infants of women with gestational magnesium deficiency, is considered in Part I of this volume.

Now that high fiber- (and phytate-) containing diets are increasingly being recommended, the effect of such diets on a magnesium intake that is otherwise meager should be explored. Review of metabolic studies of magnesium utilization by subjects on diets rich in phytates-brown bread, brown rice, oatmeal, or white bread to which phytate had been added-showed poor percentage absorption of the magnesium, particularly when the diet was first changed (Seelig, 1964). After several weeks on the phytate-rich diet, the absorption of magnesium tended to improve (A. Walker et al., 1948; Cullumbine et al., 1950; Hathaway, 1962). McCance and Widdowson (1942a,b) found that addition of phytate to white bread caused greater fecal magnesium excretion, and removing phytate from brown bread greatly improved magnesium absorption. Reinhold et al. (1976) have recently confirmed these observations, not only for magnesium but for trace metals. Thus, the higher magnesium content of phytate-containing whole grain products may not be a reliable source, in terms of availability of magnesium. Whether adaptation to the phytate ingested, on its continued inclusion in the diet, will result in better utilization (as suggested in the early cited studies) remains to be investigated systematically.

Long-term metabolic studies provide a more valuable index of adequacy of intake than do short-term studies. Figure 1-5 shows that on very low intakes (< 4 mg/kg/day) the young men remained in negative balance for the average of 52 days of study, whereas the young women retained sufficient magnesium at the end of their 30 days to maintain equilibrium, even taking into account probable sweat loss. On the usual American intake of 4-4.9 mg/kg/day, the young men went into equilibrium at the end of the study; the young women were in magnesium balance throughout. Why there was less magnesium retention by the young men whose intakes were slightly higher (5-5.9 mg/kg/day) is puzzling. Perhaps that group happened to have higher intakes of nutrients that interfere with magnesium absorption or increased renal magnesium excretion. Continuation of strong positive balances after a month on supplements that raised the magnesium intakes of young men to 9.7-12.7 mg/kg/day suggests restoration of a deficit. A subsequent study by Irwin and Feeley (1967) showed sustained strongly negative magnesium balances (-77, -74, and -38 mg/day) in 15 healthy women evaluated for 3 consecutive 20-day periods that delivered 230-300 mg of magnesium daily. They concluded that the recommended daily intake of magnesium (300 mg) is insufficient to maintain magnesium equilibrium in 140-lb women, and suggested that the proposed intake of 385 mg/day (Seelig, 1964) might be a preferable amount. In a long-term study (50 and 20 weeks) of 3 men on magnesium intakes of 1 .8 mg/kg/day to 5 mg/kg/day, Tipton and Stuart (1970) found that the young man who weighed 100 kg who was on the diet delivering the least magnesium (180 mg/day) or 1.8 mg/kg/day lost an average of 90 mg of magnesium daily during the 50-week study. A smaller (71 kg) young man given twice as much magnesium (that provided 5 mg/kg/day) retained an average of 70 mg/kg/day. An 85-kg middle-aged man who was fed a diet containing 310 mg of magnesium daily (3.8 mg/kg/day) lost an average of 40 mg daily during the 20 weeks on study. In a long-term study of men (in a Veterans Administration Hospital Metabolic Unit) Spencer et al. (1976/1980) found that increasing the magnesium intake about fourfold over the amount supplied (about 250 mg) in the basic diet did not consistently increase the amount of magnesium retained. About two-thirds of the supplement was excreted in the feces. The amount of calcium and phosphorus in the diet and the duration of the metabolic periods influenced the results. On low to high daily calcium intakes, magnesium-supplemented (about 500 mg/day patients retained about 49 to 58 mg of magnesium daily on low calcium intakes (200 mg daily). Patients on 1400-mg calcium intakes remained in slightly negative magnesium balance (-8 mg/day) when they were supplemented with magnesium; when they were not given the extra magnesium their daily magnesium loss was 20 mg. Adding the magnesium supplement to a diet plus calcium supplements providing 2000 mg of calcium raised the magnesium balance from + 2 to + 85 mg/day. Increasing the phosphorus intake to close to 1500 mg from 975, converted a positive magnesium balance (+29) to a negative one (-19 mg/day) during a period of low calcium intake, but not when the calcium intake was also increased. Spencer et al. (1979) suggested that the different amounts of magnesium retained by the different supplemented patients might have reflected their prior magnesium status. This impression is supported by the high retentions of magnesium by supplemented subjects who had previously been subjected to magnesium deprivation (Fitzgerald and Fourman, 1956; Shils, 1964, 1969a,b). They (Spencer et al. 1976/1980) also stressed the importance of the duration of the study, noting that, during the early phase of their studies, the positive magnesium balances were strong; several weeks later, the patients were in equilibrium or even in slightly negative balance. Perhaps this reflects repletion of an insufficiency, such as had been postulated might occur with sufficiently sustained magnesium supplementation (Seelig, 1964).

The cited dietary surveys and metabolic balance studies support the contention that magnesium supplied by the American diet-and most likely by that of most industrialized countries, particularly those populated by Europeans or by those with comparable eating habits-is likely not to be optimal. Such intakes, which are at best marginal, can be frankly deficient when there are concomitant high intakes of nutrients that increase magnesium requirements. Manifestly, although the incidence of abnormalities that resemble those produced in experimental or conditioned magnesium deficiency has increased during the years that the dietary pattern has changed to one that leads to at least conditioned magnesium deficiency, such abnormalities are not found in the entire population. Individual (or familial or group) differences in dietary habits can be partially responsible. (Table 1-4 gives magnesium content of foods.) Also probably contributory are genetic differences in utilization or retention of magnesium and in vitamin D metabolism (Seelig, 1969b, 1970a,b). It is hoped that future investigation will resolve whether the familial instances of parathyroid dysfunction and of some congenital cardiovascular or renal diseases are related to basic genetic variants in the handling of magnesium and vitamin D, and whether those two recognized genetic variants are interrelated.

One wonders whether the demonstrated better retention of magnesium by women than men on marginal magnesium intakes can contribute to the dramatic sex difference in incidence of IHD in young adults (Table 1-5; Figure 1-1) and to the rise of incidence in death rates in Canada from 1926 to 1961 (Fig. 1-6, T. Anderson, 1973). The sharp increase that occurred only in middle-aged men was entirely in the IHD category; cardiac death rates from other causes dropped. Among the women, the cardiac death rate remained the same, but the proportion due to IHD rose. There was a lesser sex difference in the proportion of deaths that occurred suddenly in the middle-aged groups in hard- and soft-water cities in Ontario, and still less in the 65 to 74 year-old groups (T. Anderson et al. 1976/1980). Whether the observation of this group that myocardial magnesium levels were lower in women who had died suddenly (accident or suicide) than they were in comparable men, both in hard- and soft-water areas (T. Anderson et al., 1978), bears on this question requires resolution. On the surface it would seem to militate against the concept that women's better retention of magnesium explains the sex difference in the rise of IHD. Additional factors must be considered. Among such factors are those diagrammed by Raab (1972), who had earlier provided experimental evidence that stress causes decreased myocardial magnesium levels (Raab et al. 1968). Does this imply that women are more subject to stress-induced decreased myocardial magnesium? This seems dubious. More likely, women normally have less myocardial magnesium than do men. Does the amount of muscular exertion play a role? The higher myocardial magnesium levels in left than in right ventricles (Holtmeier, l969a; Szelenyi, 1973) might be germane to this point.

The evidence that dietary magnesium is generally insufficient and that under those conditions women retain more than do men, is clear, however-wherever the magnesium goes. It provides some insight into the provocative epidemiologic studies that demonstrate that the cardiovascular death rates are higher in areas supplied with soft water than they are in hard water areas. N. Goldsmith (1969) and Hankin et al. (1970) have calculated that 12% of the daily intake of magnesium can be derived from water. Among those using only hard water, as much as 18% of the daily magnesium intake may derive from water. Among those whose magnesium intakes from food are marginal, these amounts might well be critical.

1.5. Hard/Soft Water and Cardiovascular Disease

J. Kobayashi (1957) first noted that the nature of drinking water might influence death rates from cardiovascular disease; the incidence of strokes is high in areas with acid (soft) water. Schroeder (1960a,b, 1966) surveyed the hardness of drinking water in each of the United States, and correlated the death rates with state-wide water hardness or softness (Fig. 1-7). He found that death rates from cardiovascular diseases (particularly from "coronary" heart attacks in white men 45-64 years old) were significantly higher in states with soft water than in states with hard water (Fig. 1-8). The death rate in South Carolina, a state with the softest water, was 983/100,000; that in Nebraska, a hard-water state, was 712/100,000. Deaths from cerebrovascular accidents followed a similar pattern. Complicating interpretation of these findings is the fact that ischemic heart disease death rates are higher in urban than in rural communities. To eliminate this factor, the coronary death rates from three cities with hard-, intermediate-, and soft-water supplies are compared (Fig. 1- 9), and reveal a startling contrast between the rates of fatal ischemic heart disease in cities with hard and soft water. Since this observation, there have been many confirmatory studies, although there has not been complete accord that it is the magnesium, rather than the calcium, that is protective, or whether there might be a toxic element in the soft water that is to blame, e.g., cadmium (Perry, 1973), copper (Harman, 1975/1977), or others (Editorial, Lancet, 1969c).On the other hand, Klevay (1975, 1977, 1978) has presented provocative evidence that a high zinc/copper dietary ratio might contribute to ischemic heart disease as a result of relative copper deficiency, which causes a decrease in high-density lipids and an increase in low-density lipids. He has suggested that hard water might be protective by lowering the ratio of zinc to copper. Where hard water is artificially softened (i.e., by sodium chelates), the role of increased sodium should also be considered.

M. Crawford et al. (1968) like others working in areas where calcium is by far the major water factor (J. Morris et al., 1962; Biorck et al., 1965; J. Robertson, 1968, 1969), favored calcium as the probable protective factor. In England, the average calcium content of hard water accounts for about 84% of the hardness and is more than 11 times greater than the magnesium content (M. Crawford et al. 1968). Nonetheless, it was her impression that the water "factor" probably involves interrelationships of the "bulk" ions: calcium, magnesium, and sodium (M. Crawford, 1972). Those who had had favorable experience with the use of magnesium salts in treating patients with acute or chronic IHD (R. Parsons et al., 1961; Berberian, 1962; Browne, 1961, 1963, 1964a,b) favored magnesium as the hard-water protective factor. So also did those who had done or evaluated animal work that showed magnesium to be protective and excess calcium either not protective or harmful in experimental cardiomyopathies or soft-tissue calcification (e.g., cardiovascular and renal) (Neal and Neal, 1962; Marier, 1963; Bajusz, 1967; Marier et al., 1968; Seelig and Bunce, 1972; Seelig and Heggtveit, 1974). In his consideration of the part played by hard water, Bajusz (1967) suggested that the higher content of magnesium might protect the myocardial cell against damage caused by ischemia and improve its ability to resist the effects of cardiotoxic agents. In 1974, Seelig and Heggtveit considered the experimental and clinical evidence that calcium and magnesium exert reciprocal effects on myocardial irritability. High calcium levels stimulate and high magnesium levels suppress hyperexcitability. They then suggested that magnesium might be useful in maintaining normal cardiac rhythmicity, in the face of ischemia or digitalis or in acute (i.e., alcohol- or diuretic-induced) hypomagnesemia [The antiarrhythmic attribute of magnesium is again being utilized in the United States (Chadda et al., 1973a,b, 1976/1980; Iseri et al., 1975; Iseri and Bures, 1976/1980) after a hiatus of 30 years (Boyd and Scherf, 1943).]

The dietary surveys presented here show that magnesium, but not calcium, intakes have been gradually falling. Coinciding in time with the sharp increase, first of vitamin D and then of phosphorus intakes (Fig. 1-1), there has been a steep increase in incidence of sudden deaths from IHD (T. Anderson and LeRiche, 1970). The recognition of this increase in IHD death rates derived from an extensive study of death certificates of men 45-64 years of age (Ontario, 1901-1961). As many as 5000 certificates a year had to be examined when the incidence was low (in 1901). Where deaths from IHD were clearly specified, as compared with all cardiac deaths, it was the IHD category that had risen, more than doubling from 1931 to 1961 (Anderson and LeRiche, 1970). The death rates from other major forms of heart disease in that age group had fallen during the same period of time. The minor changes in cardiac death rates from 1901 to 1931 are not as readily interpreted, because of changes in terminology and possible incompleteness of reporting. Selecting 1931 as the earliest key date (sudden-death coroner reports being available from about 1931 on Toronto), these investigators found that only about half of the non- traumatic sudden deaths were attributed to IHD in 1931, whereas 99% were deemed due to IHD in 1961. Spain et al. (1960), on the basis of an autopsy survey, considered such events the commonest cause of death of middle-aged men, at about the same time. T. Anderson et al. (1969) postulated that there might be an environmental factor that could, by altering myocardial excitability, cause an increase in the incidence of sudden death from ventricular fibrillation and other cardiac arrhythmias, and noted that the sudden death rate (but not the nonsudden IHD death rate) varies inversely with the hardness of the water. T. Crawford and M. D. Crawford 1967), who had noted that despite a much higher frequency of cardiac death rates in Glasgow (a soft-water area) than in London (a hard-water area) degrees of coronary atherosclerosis were not dissimilar, had also suggested that the water factor might affect the way the myocardium reacts to ischemia. They found that the coronary magnesium content was higher in young men (under 40) who had died as a result of accidents in London than in Glasgow, and that the Scottish young men had more small myocardial scars than did the Londoners.

From Ontario, where the magnesium content of hard water is much higher than it is in England, has come much of the definitive data implicating magnesium rather than calcium as the protective factor in hard water, and ruling out most of the potentially toxic trace minerals found in soft water as the harmful soft-water factor. In their surveys of cardiac death rates, T. Anderson et al. (1969) found that there were many more (sudden) cardiac deaths reported in soft- than in hard-water areas (T. Anderson et al., 1978). This supported T. Crawford and M. D. Crawford's (1967) and Bajusz's (1967) suggestion that the hard-water factor was likely to be a myocardial protective factor. They speculated that it probably affected cardiac rhythmicity (T. Anderson et al., 1969, 1973, 1976/1979; T. Anderson and LeRiche, 1970, 1971). Comparable findings were reported by Fodor et al. (1973) from Newfoundland, where there is a much higher death rate for IHD in a city with very soft water than in two communities with hard water, particularly for men, 62% of whom died before they could be brought to the hospital (considered probable sudden deaths). They commented that IHD death rate in men in the soft-water city (702/100,000) is comparable to that in the "high mortality belt" of the southeastern portion of the United States.

At first Anderson et al. (1969) adhered to the English premise that calcium was likely to be the protective water factor. When they became aware of the evidence that Western diets provided marginal amounts of magnesium (Seelig, 1964) and that persons dying of heart attacks have low myocardial magnesium levels, even in non-infarcted segments (Heggtveit et al., 1969; Seelig, 1972), they had pathologists from hard- and soft-water areas secure myocardial specimens from routine autopsies, and had them analyzed for magnesium, calcium, and trace elements (T. Anderson et al., 1973, 1975, 1976. Magnesium was the only element with a significant difference in myocardial concentration, which was higher in hearts of accident victims from hard-water areas (982/918 µ/g dry tissue). IHD disease victims had 22% lower myocardial magnesium levels in soft- than in hard-water areas (697/744). In England, there has been an apparently contradictory pattern (Chipperfield et al., 1976a), with lower levels of myocardial magnesium in hard- than in soft-water cities. T. Anderson et al. (1978) point out that since in the two English cities that were compared the magnesium levels are quite low both in the hard and in the soft water (Chipperfield et al., 1976b), the difference between them represents only 1% or 2% of the probable total intake, and that another factor might be operative.

In Finland, which has a very high death rate from IHD, there is a clear relationship with the amount of magnesium in the soil (Karppanen and Neuvonen, 1973). In eastern and in northern Finland, where the soil content is about a third that found in southwestern Finland (Karppanen et al., 1978) the mortality from ischemic heart disease is twice as high as is that in the southwest. Ho and Khun (1976/1980) surveyed factors that might be contributory both to the rising incidence of cardiovascular disease in Europe, and the falling levels of magnesium both in the soil and in the food supply. They commented that in Finland, which has the highest cardiovascular death rate in Europe, the dietary supply of magnesium had decreased by 1963 to a third of the intake common in 1911 (H. Katz, 1973). In contrast, in Japan with its low cardiac death rate, the daily magnesium intake was cited as 560 mg (Holtmeier, 1969a, 1973). Karppanen et al. (1978) have depicted the steep rise in ischemic heart disease that coincides with increasing dietary calcium/magnesium ratios (Fig. 1-10).

In view of the possibility that sudden deaths of infants might similarly be mediated by magnesium deficiency, and be analogous to the adult cardiac arrhythmic sudden deaths that are prevalent in soft-water areas, the preliminary report by Godwin and Brown (1973) of a somewhat higher incidence of sudden infant deaths in soft-water counties in California than in hard-water counties is provocative. It must be noted that the following year a conflicting report was published (Allwright et al., 1974) that failed to confirm the higher incidence of either IHD of adults or of infant mortality rates with soft water. However, these investigators point out that this "soft" water is approximately as hard as is the "hard" water in some of the English studies, where higher infant-death rates were reported in soft- than in hard-water areas (M. Crawford et al., 1972). These tentative findings call to mind the instance of sudden infant death that was associated with coronary arteriosclerosis (Meurman et al., 1965) from eastern Finland, and the report by Pesonen et al. (1975) on more severe and more frequent infantile coronary arteriosclerosis in eastern Finnish children than in those from the southwest (where the magnesium level is higher).

1.6. Epidemiologic Factors in Calcific Urinary Calculi

There has been an increase, during this century, in the incidence of calcium oxalate stones in Finland (Sallinen, 1960), central Europe and Sweden (Grossman, 1938; Hedenberg, 1951) and Japan (Inada et al., 1958; Editorial, Brit Med J, 1965) that coincides with the rising incidence of cardiovascular diseases in those countries, and with the use of magnesium-poor soil fertilizers in the northern and central parts of the European continent (Holtmeier and Kuhn, 1976/1980). The geographic difference in frequency of calcific urolithiasis in the United States (Landes, 1975/1977; Finlayson, 1974; Landes et al., 1977) coincides with the geographic difference in incidence of sudden death from ischemic heart disease, and with the degree of water hardness. A map of the United States, indicating the distribution of water hardness in 1963 (Fig. 1-11) and one showing the incidence of urinary calculi (Fig. 1-12) clearly shows that in most states where the water is softest the frequency of urolithiasis is highest. Boyce et al. (1956), who pointed out this geographic overlap, reported that the highest incidences of kidney stones were in South Carolina and other southeastern states, an area that has been called "the kidney stone belt," and the lowest incidences were in midwest and southwestern states, where the water is hard. Melnick et al. (1971, 1973) and Landes et al. (1977) reaffirmed this observation, basing their conclusions on hospital diagnoses, obtained from a survey done in 1972. South Carolina again came in first, with the highest frequency (17/1000 discharges). Nebraska, the state shown earlier to have the lowest incidence of sudden death from ischemic heart disease, also had the lowest frequency of urinary calculi (2.6/1000 hospital discharges). Accepting the limitations of such state-wide surveys of stone incidence and water quality, the authors nonetheless felt justified in concluding that the differences were statistically significant, indicating that the incidence of urinary calculi is inversely related to the hardness of the water (Landes et al. 1977).

Prien (1971), who had reported that magnesium therapy in conjunction with pyridoxine (Prien, 1965; Gershoff and Prien, 1967), was useful as prophylactic therapy for recurrent calcium oxalate stone-formers (in northern New England, another soft-water area), presents "the riddle of urinary stone disease." He referred to Grossmann's (1938) evidence that, starting in 1924, the incidence of small calcific stones in young adults rose in central and northern Europe, and was puzzled as to why the incidence should have dropped during World War II, only to rise again thereafter (Boshamer, 1961). It is possible that the work of Linden (1972, 1974a, 1977) correlating concurrent urolithiasis, hyperlipidemia, and ischemic heart disease with only moderately high intakes of vitamin D might be germane to the rise in incidence of kidney stones after 1924. Linden (1977) mentioned that after Mellanby (1920) had demonstrated that cod liver oil could prevent rickets; it was soon widely used for lesser ailments, such as failure to thrive and poor appetite. He referred to reports, in the late 1920s, of infantile fatalities due to hypervitaminosis D. It is possible that inappropriate and widespread use of vitamin D, which increases magnesium requirements, might have intensified magnesium deficiency, the predisposition for which might have derived from the decreased magnesium-availability from the soil, especially in those parts of central Europe where fertilizers high in potassium and low in magnesium were commonly used after World War I (Aleksandrowicz and Stachura, 1976/1980; Holtmeier and Kuhn, 1976/1980). Perhaps, during World War I the vitamin D supplements and soil fertilization were less widely used, only to be taken up again after the war. In the last year of World War I and for more than a decade thereafter, in the British Isles, excessive vitamin D was provided in infant formulas and other foods, with a resultant epidemic of supravalvular aortic stenosis syndrome (SASS) and increased incidences of renal tubular acidosis, infantile nephrocalcinosis, and osteosclerosis. In Germany, "Stosstherapie" with huge parenteral doses of vitamin D also caused SASS and related "congenital" abnormalities (Review: Seelig, 1969b). To what extent the long-term use of therapeutic doses of vitamin D in infants and children with low requirements or hyperreactivity to vitamin D (Seelig, 1970a,b), and to what extent its continued use throughout life, and especially during adolescence and early adulthood when milk consumption tends to be high (in America), might predispose to a high urinary calcium/magnesium ratio and to a conditioned magnesium deficiency should be systematically investigated. Possibly it might be part of the answer to Prien's kidney-stone riddle (1971), as well as to the continued high sudden-cardiac death rate.

The inverse relationship between the tendency toward calcium oxalate urinary tract stones and the tendency toward osteoporosis (McGeown and Oreopolis, 1969) is provocative. There are fragmentary data indicating that magnesium deficiency contributes to several pediatric osteopenias and to osteoporosis, all of which are characterized by loss of matrix. A high Ca ratio might favor hypermineralization of bone with defective matrix. High P/Ca and P/Mg ratios might favor osteomalacia. Correlation of these mineral ratios with hormonal responses might shed some light on the high rate of osteoporosis in postmenopausal women, who might have a high parathyroid/estrogen ratio, in addition to loss of the estrogen-induced capacity to store magnesium in bone Whether low magnesium and high vitamin D intakes during pregnancy contribute to osteogenesis imperfecta, hypophosphatasia, and fragile bones of low-birth-weight infants should be studied.

1.7. Genetic Factors in Cardiovascular, Skeletal, and Renal Diseases

Even though the dietary factors (high vitamin D and phosphate intakes and declining magnesium intakes) have been widespread, and in the case of vitamin D unavoidable for milk-drinkers, the increased incidence of frequency of some cardiovascular, skeletal, and renal diseases has not been distributed equally in the population. Except for osteoporosis, which is most prevalent in white postmenopausal women (McGeown, 1969; Meema et al., 1973, 1975; N. Goldsmith and Johnston, 1978), and hypertension, which is most prevalent in black women (Kuller et al., 1973), most of the diseases for which evidence is presented in this volume, of relationships to low magnesium, to vitamin D, and to phosphate intake, are most prevalent in white males. Furthermore, there is evidence of familial predisposition to what may be risk factors: (1) specific magnesium malabsorption and renal wasting, and (2) hyperreactivity to vitamin D (Seelig, 1969b, 1970a,b). It is suggested that the familial instances of calcium oxalate urolithiasis (McGeown, 1960; Resnick et al., 1968), of pseudohypoparathyroidism with vitamin-D-resistant rickets (DeLuca et al., 1967; Falls et al., 1968; Reitz and Weinstein, 1973), and possibly of hyperparathyroidism (Cutler et al., 1964; Cholod et al., 1970; Marx et al., 1973) might also be secondary to a primary abnormality in magnesium metabolism, leading to magnesium deficiency. Several forms of neonatal or infantile cardiovascular disease, possibly including juvenile hyperlipidemia and hypertension, might also be related to abnormalities in magnesium or vitamin D metabolism, or both, as might vitamin-D-resistant rickets. The familial instances of these diseases and of other osteopenias, which are not infrequently associated with renal disease, might have an underlying defect: magnesium malabsorption or renal tubular wastage or both.

The renal and skeletal disorders contribute to significant morbidity. The cardiovascular complications lead both to morbidity and sudden mortality. An editorial (JAMA, 1972) entitled "A Magic Carpet Is Not Enough" calls urgent attention to the fact that as many as over 50% to 73% of sudden deaths from lethal arrhythmias (Kuller et al. 1967; Armstrong et al., 1972) occur before the patients reach the hospital. Among almost a thousand cases of medically untreated deaths from IHD in which autopsies were done, 60% of the men and 47% of the women died within 15 minutes of onset of symptoms (Wikland, 1971). The reference to the inadequacy of the "magic carpet" pertained to the finding that even if the patient is resuscitated, death is usually merely somewhat delayed by a period of invalidism (Geddes et al., 1967). In a confirmatory study, Kuller et al. (1973) showed that 75% of those who died suddenly had had no serious disability; only 12% had been unable to work. Among those who died in a hospital, only 17% survived more than 24 hours. Of the almost 500 who died within 24 hours of onset of symptoms who were autopsied, only 13% had evidence of a recent infarct. The investigators concluded that their findings indicated that no current community health approach will be effective. They state that a fundamental change in therapeutic and preventive approach is needed.

The correlation of data implicating interrelationships between absolute magnesium deficiency and the magnesium-losing excesses of vitamin D and phosphates in so many of the diseases that have increased in incidence during the time that these dietary imbalances developed might point toward the new approach that Kuller et al. (1973) said was required. Such imbalances are a particular risk during pregnancy (Fig. 1-13). Possibly they might lead to abnormal pregnancy, to congenital abnormalities, and to infantile and later morbidity and mortality.

Now that vitamin D has been proved to have a steroid hormone mode of action (Norman, 1968; Norman et al., 1975/1977; DeLuca, 1969, 1976), and the active antirachitic metabolites have been isolated, synthesized, and made available, vitamin-D-resistant rickets can be treated specifically, in preference to putting an entire population at risk of hypervitaminosis D by fortifying milk and other foods (Seelig and Mazlen, 1977). Perhaps the popular soft drinks that provide so much of the excess phosphate can be reformulated to be bubbly by other means than by use of phosphate salts. And, finally, physicians should evaluate their patients, particularly those with the cited familial disorders-for abnormal magnesium metabolism, and they should prescribe supplements when needed.

This volume places major emphasis on the early establishment of cardiovascular, skeletal, and renal lesions-possibly during gestation, infancy, and early childhood-that can either cause early manifestations of disease or death, or lay the groundwork for disease processes that become overt later in life. Thus, the first part deals with prenatal, perinatal, and infantile disorders, and the subsequent parts deal more generally with the abnormalities of the three systems to which magnesium deficiency might well be contributory.