Annu. Symp Proc. Univ. Missouri. Trace substances in environmental health. 16: 189-195 (1982)
Magnesium in drinking water has been calculated to constitute some 10% of the daily population intake. But the effect on public health must be much larger. This we conclude from results of self experimentation in a pilot study (M.R.H.L.) with diets of different Mg levels and a contribution by water Mg. Urinary analysis revealed that water Mg is absorbed 30% better and much faster than dietary Mg. There is a considerable diurnal variation in the amount of Mg absorption, leading to a 50-70% lower Mg status during the morning hours. When little or no breakfast is taken, the low Mg period is extended until the coffee break, and the Mg content of water will then have a crucial importance. The effect of omitting a meal can be observed during the weekend; it results in a depression to 70% or less of the usual urinary excretion. Magnesium status is induced favorably by sports and exercise, since the necessary high energy diets usually contain much Mg. In statistical calculations on cardiovascular mortality due note should be taken of the precipitation of calcium (and much less so of Mg) during boiling of water. Hence the relative influence of Mg is larger. Statistics on Mg should be based on the Mg water content of the working rather than the living area.
There is a long history of Mg deficiency in cattle, mainly high-producing milk cows, with frequent occurrence of tetany and heart disease. The clinical picture is not uniform: sometimes symptoms of tetany dominate, sometimes there is a labile situation with muscle paresis, but also sudden heart death may result. Siollema and co-workers (11,12) were the first investigators to connect the condition with Mg. It is remarkable that not all of the cows with extremely low serum Mg are affected, and the variety of symptoms also suggests a combination of factors, of which Mg deficiency is a very important one.
In the human the relation of Mg and heart disease is somewhat different, at least quantitatively. The recommended dietary allowance is a factor of three larger than the minimum needed to produce urinary excretion (7). Conformably, serum Mg generally is in the normal range of .75 to 1.05 mmol/l, whereas in tetany-prone cows it comes down to one fifth of this value, with zero urinary excretion. In the so-called "water story" (2,8) the incidence of human heart disease is inversely related statistically to the hardness of the drinking water in the living area. The relationship does not hold in all countries, e.g. in Britain (5,6) where Mg is generally low, even in hard water. The question then arises: could it be that Mg is the real driving force behind the water story? Magnesium has all the required properties to protect the heart. It ensures the normal quiet heart rhythm. If this is unduly accelerated by toxic factors or other stress, Mg may reverse the effect. Magnesium prevents the so-called "calcium overload". Usually clinicians fight this condition with pharmaceutical preparations ("calcium blockers") but Mg would probably do as well, or even better. It also counteracts digitalis, epinephrine and other toxic agents used to induce heart disease experimentally in animals. In all of these conditions Mg restores the original rhythm and effective functioning of the heart. On the other hand, active Mg is lost from the ischaemic heart, even before loss of sodium occurs, and it is replaced later by inactive Mg, probably in the form of insoluble phosphate. Much of the active Mg is in combination with ATP, and in the enzyme Na/K-ATPase which stabilizes the Na/K gradient over the membranes and generates the resting potential (1,10).
Why is it then that Mg is so often disregarded? As we have seen, the abnormal human heart condition may occur well above the deficiency range. Secondly, in the water story, Mg contributes only little to the daily intake. The contribution has been estimated as between 1 and 20% (9), or an average 10%, or lower, in Europe (13). It is difficult to see that such a minor contribution should have the described important statistical effect. The aim of this study is to check this low percentage contribution. It is conceivable that, due to factors like fiber and phytate, dietary Mg is less efficiently absorbed.
The research was mainly based on self experimentation. Self experimentation on one person (M.R.H.L.), while producing only tentative results, has the advantage of reliability and flexibility. This was much enhanced by the well established and fast methods of analysis by atomic absorption spectroscopy. Urine was collected quantitatively and analyzed in mostly 1:100 dilution. Faeces and food were digested with nitric and perchloric acid and similarly analyzed. In 4 experimental periods of 10-14 days duration 3 diets were taken, low, moderate and high in Mg, whereas in the fourth period the low Mg diet was reinforced with MgCl2 dissolved in water to the level of the high Mg diet. The diets were variable, and had only approximately the calculated content. The exact quantity of Mg taken daily was determined by duplicate sampling - one tenth scale. Experimental details will be given elsewhere.
The daily urinary excretion of Mg related to daily intake is presented in Figure 1. We have added, for this purpose, each morning's first sample as "night urine" in the total of the preceding day. It will be noted easily that we did not attempt to prepare uniform diets, and that most of the daily intakes were above the 300-350 mg recommendation. Although a line has been drawn through the point swarm, it only roughly indicates the relationship, especially because of adaptation effects. As can readily be seen, the black dots (low Mg + water MgCl2) lie higher than the black crosses (high Mg diet).
This can be observed more clearly when we omit the preliminary period, and average the 4 experimental days which start at day 7. In Figure 2 the averages and the standard error of the mean are given. Prolonging the experiment by averaging over days 11-14 did not result in different averages. Our interpretation of the higher level of urinary Mg excretion with the lowMg + MgCl2 diet is that Mg is better absorbed from water than from high Mg food. The difference is 30%.
In the beginning of the experiment it was noted that separate urine voidings of the same day had different Mg contents. Therefore the experiment was carried out with analysis of the weighed voidings, usually 4 in 24 hr, and the daily total urinary excretion was calculated. The subsamples had a diurnal rhythm, which for the different Mg intakes is presented in Figure 3. We postulate that these short term effects are only secondarily generated by variations of the kidney excretion efficiency, and primarily by variations of the Mg status. (This is confirmed by Figure 4.) The quantities of Mg excreted per hour generally increase in the following order: morning, afternoon, evening. The morning urine then has a value which is a little below that of the night urine. Only in the experiment with water Mg was the morning urine higher in Mg than the night urine. This difference is significant (p<0.05). Obviously Mg from water taken with breakfast is absorbed much faster than food Mg,
That the morning period has a crucial Mg status is schematically given in Figure 4. For the construction the relative Mg contribution of the normal (and low) Mg meals have been used, as well as literature values on the speed of absorption and excretion.
During the night hours the Mg status declines considerably, so that it depends on the following breakfast or morning coffee to resupply Mg. Magnesium needs, however, are probably high in the morning hours. We have indicated this with the dashed line in Figure 4, which we take (3) from the diurnal rhythm of the hormone somatomedin. Calculations of the levels and the known relation of somatomedin and the N balance (4) permit computation of the Mg needs for the morning's resynthesis of cellular material and plasma proteins. We intend to perform such calculations elsewhere.
Omitting a meal causes a depression of the daily Mg supply, easily as much as 30%. Omission of breakfast then has a very pronounced effect on Monday morning, after the careless weekend. An example is given in Figure 5, with urinary Mg below 60% of the mean. Though we did not look for short-term effects, we are convinced that over short intervals of the morning urinary Mg excretion may be very low, and when the overall basal Mg status is low, urinary Mg could be temporarily zero, accompanying a depressed serum Mg content. The effect of morning coffee, consumed during such a low Mg period, could be much greater than reflected by its Mg contribution to the daily supply. Coffee and tea are made with boiled water. Hard water, upon boiling, is rendered soft, but its Mg content is hardly altered (Fig . 6).
As seen in Fipure 5, sports decrease urinary Mg excretion, and we have calculated routine corrections for this, amounting to 25% under our conditions. But at the same time sports, by their need for energy, increase Mg intake. In modern times no athlete will choose fat and sugar for his high energy supply, but much more probably nuts, raisins and chocolate, all with a high to very high Mg content. Formerly people doing hard work might have taken beans, peas, etc. ; but also wheat and other grains would have provided much Mg. Finally we pose the following thesis: benefit of the training of athletes (and also of heart patients) produced by exercise is the lowering of the heart frequency by Mg from the extra food Mg intake. Magnesium, of course, is not the sole protector of normal heart function, and we do not wish to over-estimate its effect, but it certainly is one of the factors that continue to deserve our special attention.
The short range Mg status of an individual has an appreciable diurnal variation, with a minimum during the morning hours. This necessitates hour-to-hour studies instead of the customary balance studies of several days or weeks duration. Magnesium from water is absorbed faster and to a higher degree than Mg from (high-Mg) food. This has an important all-or-none effect on the morning Mg status, depending on the Mg content of the water. Although boiling of water softens it, this does not affect the Mg content, which remains nearly the same. When sport or exercise is advised to people who are endangered by heart disease, it involves consumption of high energy foods, which generally contain much Mg. Part of its effect could be obtained by drinking water with a high Mg content.
We thank J. Rijken for the pleasant cooperation in the analytical part. Further we are indebted to experts, especially C. Kalkman of Amsterdam University, who kindly provided us with his recent literature studies on Mg and heart/vascular disease, including the water factor.
1. Aikawa, J.K. 1981. Magnesium, Its Biological Significance, CRC Press, Boca Raton, P. 72.
2. Arden, T.V. 1977. Cardiovascular Mortality, An Analysis of the Water Effect. Fed. Europ. du Traitement de l'eau, Genéve.
3. Binnerts, W.T., P.W.M. van Adrichem, C.P.J. Oudenaarden, J.E. Vogt and J.E. Wassenaar. 1982. Plasma somatomedin in dairy cows: effect of management and feeding level. Neth. Milk Dairy J. 36:149-152.
4. Binnerts, W.T. and P.M.W. van Adrichem. 1982. Some observations on plasma somatomedin activity in dairy cows: a higher correlation with protein than with energy metabolism. Neth. J. Vet. Sci. (Submitted).
5. Chipperfield, B., J.R. Chipperfield, G. Behr and P. Burton. 1976. Magnesium and potassium content of normal heart muscle in areas of hard and soft water. Lancet (1) 121-122.
6. Dauncy, M.J. 1972. Urinary excretion of calcium, magnesium, sodium and potassium in hard and soft water areas. Lancet (1) 711-774.
7. Marshall, D.H., B.C.E. Nordin and R. Speed. 1976. Calcium, phosphorus and magnesium requirement. Proc.Nutr. Soc. 35:163-173.
8. Neri, L.C., D. Hewitt and C. B. Schreiber. 1974. Can epidemiology elucidate the water story? Am. J. Epid. 99:75-88.
9. Neri, L.C. and H.L. Johansen. 1978. Water hardness and cardiovascular mortality. Ann. N. Y. Acad. Sci. 304:203-219.
10. Seelig, M.S. 1972. Myocardial loss of functional magnesium. In: Recent Advances in study of Cardiac Structure and Metabolism I, University Park Press, Baltimore, pp. 615-637.
11. Sjollema, B. 1930. On the nature and therapy of grass staggers. Vet. Res. 10:425 439, 450-454.
12. Sjollema, B. and L. Seekles. 1930. On disturbance of the mineral regulatory mechanisms in diseases of cattle. Biochchem. Z. 229:358-380 (in German).
13. Zoetman, B.C.J. and F.J.J. Brinkmann. 1976. Human intake of minerals from drinking water in the European Community, In: Hardness of Drinking Water and Public Health, Pergamon Press, Oxford, pp. 173-202.
Inquirer: M. Van Hardenbrock, Dillon, CO
Q. Correlation of Dutch protein food intake and magnesium depletion and correlation with hypoxia?
A. This is essentially unknown, and partly so because the
Netherlands Table of Nutrients, 1978, does not list Mg. The level
of protein intake is generally considerable, but some groups are
exposed to "cafeteria type diets", among which are the groups of
"managers". There are Dutch statistics on coronary heart disease
and income (more income, more disease).
Inquirer: John R.J. Sorenson, College of Pharmacy, University of Arkansas, Little Rock, AR
Q. Did you say that there was Mg deficiency in England?
A. I have said that hard water in England contains remarkably
little Mg, compared to elsewhere.
Inquirer: Herta Spencer, Veterans Administration Hospital, Hines, IL
Q. There may possibly be an adverse relationship between protein and magnesium, i.e. that the magnesium requirement may be higher due to a higher protein metabolism. However, practically all high protein foods are high in Mg.
A. Thank you for your comment. I have calculated Mg over
protein (in mg/g) and this results in: meat 0.6; fish 2; dairy
products 3 and cereals, nuts, beans, green vegetables and grass
15. (The latter with probably a considerable reduction of uptake
Inquirer: Carl Marienfeld, University of Missouri, Columbia, MO
Q. Could it not be an excess of manganese which is the factor?
A. Statistically, Mg is only one of the protective factors, with the unique advantage that it has a causal relationship to ischaemic heart disease. Other trace elements, like Cu and Mn, are considered to be independent risk factors. But this study could not be substantiated in at least one recent study (Manthey et al. 1981. Magnesium and trace metals: Risk factors for coronary heart disease? Circulation 64:722-729.
This page was first uploaded to The Magnesium Web Site on June 12, 1996