Hum Pathol 1989;20:726-731.
*Present address: Department of Pathology, University of
Mississippi Medical Center, Jackson, MS.
Supported by grant no. HL38079 from the National Institutes of Health and by the Veterans Administration.
Address correspondence and reprint requests to Sherman Bloom, MD, Department of Pathology, University of Mississippi Medical Center, 2500 N State St, Jackson, MS 39216.
There is reason to believe that calcium influx into heart muscle during acute myocardial infarction (AMI) can aggravate myocyte injury. Furthermore, the degree of such influx might correlate with the occurrence of microscopic myocyte calcification observed at autopsy. We have searched for evidence of myocyte calcification in hearts of patients found to have AMI at autopsy at the Veterans Administration Medical Center in Salt Lake City (SLCVA), a region with a low myocardial infection death rate, and at the George Washington University Medical Center in Washington, DC (GWUMC), a region with a high myocardial infection death rate. Of 23 consecutive cases examined under "blind" conditions at the GWUMC in which AMI was found, there were 15 instances of cardiac myocyte calcification observed in von Kossa-stained sections. Not a single example of myocyte calcification was found in 23 comparable cases at the SLCVA. The basis of this difference in myocyte calcification is unknown, but may be related to the fact that the Salt Lake City drinking water contains a higher level of magnesium, which is known to protect against soft tissue calcification, than does that of Washington, DC. This may be the basis for the apparent protection that dietary magnesium exerts against myocardial infarction death.
Key words: myocardial infarction, magnesium, drinking water, geography, myocyte calcification, metastatic calcification.
Myocardial infarction death rates vary widely as a function of geography. Some of this variation has been attributed to variation in dietary intake of magnesium, primarily through drinking water (1-6). Furthermore, there is an inverse relationship between the serum magnesium level and the likelihood of developing a cardiac arrhythmia (7,8), and survival after acute myocardial infarction (AMI) is increased by the administration of magnesium (9). These observations are particularly important in view of the apparent magnesium-deficient status of a large part of the population in the United States (10).
In experimental animals, consumption of a magnesium-deficient diet leads to increased myocardial calcium levels and increased vulnerability to myocardial necrosis. This necrosis is associated with myocyte calcification (11,12). In view of the deleterious effects of calcium on myocyte energy metabolism (13,14), we have postulated that the increase in myocardial calcium is a critical determinant of the deleterious effect of magnesium-deficient diets (15-17). These observations, and the unsubstantiated impression that AMI-associated myocyte calcification is common in Washington, DC, but rare in Salt Lake City, led us to hypothesize that the increased incidence of myocardial infarction death in areas with low magnesium in the drinking water is associated with a greater tendency to develop histologically identifiable myocardial calcification in association with AMI. Such an increased tendency for calcification would reflect the underlying abnormality in calcium metabolism hypothesized to be the basis of the increased AMI death rate in magnesium-poor geographic regions.
To test the above hypothesis, we compared autopsy material from patients in two regions where these factors are divergent. The George Washington University Medical Center in Washington, DC (GWUMC) was used to represent a high myocardial infarction death rate area, and the Veterans Administration Medical Center in Salt Lake City (SLCVA) was used to represent a low myocardial infarction death rate area. These areas are in the highest and lowest quintiles of myocardial infarction death rates, respectively, in the United States (18). We also compared the mineral composition of the drinking water from these two regions, using published information and data from water-testing laboratories.
Autopsy reports from both GWUMC and SLCVA for the year 1986 were examined, and all cases in which recent or healed myocardial necrosis was found were selected for further study. This consisted of examination of all hematoxylin-eosin-stained histologic sections taken from the heart. These sections were obtained from blocks of myocardium, 2 to 4 mm thick, fixed in phosphate-buffered, 4% formaldehyde at neutral pH. Those patients in whom microscopy showed evidence of AMI were selected as study cases. AMI was defined as acute myocyte necrosis evidenced by loss of cross striations and nuclei, with or without the presence of contraction bands, and with or without an inflammatory reaction. Lesions in various stages of repair were also accepted as showing AMI, including those in which there was evidence of prior necrosis with inflammation and early repair (granulation tissue). Healed lesions showing only mature collagen, with or without chronic inflammatory cells, hemosiderin-laden macrophages, or dilated capillaries, were not studied further. Evidence of coronary artery disease was also required: either a statement that there was at least a 75% reduction in luminal diameter of a major coronary artery, a statement that coronary artery narrowing was "severe," or a microscopic slide showing at least 75% reduction of luminal diameter.
For each case, one slide was selected for staining by the von Kossa method (19) with hematoxylin-eosin counter staining. This selection was based on the presence of probable calcium deposits noted in the hematoxylin-eosin-stained slides. If no such suspect deposits were found, sections showing the most extensive AMI were selected, with preference given to sections also showing some adjacent normal myocardium. Additional slides were prepared from these blocks, stained by the von Kossa method, and faintly counter-stained with nuclear fast red. With these sections, von Kossa-positive deposits were especially conspicuous, but other morphologic features were poorly stained and more difficult to identify. Random numbers were assigned to these sections, which were then examined by one investigator, who scored for the presence or absence of calcium deposits. The other investigator examined these same slides, but also had access to hematoxylin-eosin-stained sections from the same paraffin blocks as well as von Kossa stained sections counterstained with hematoxylin-eosin. These other stains were needed to identify the sites at which the von Kossa-stained material was located. Although the von Kossa stain for calcium actually stains phosphate and other precipitated anions, the only known form of such insoluble deposits in human tissues is the calcium salt. Therefore, a positive von Kossa stain was considered to be evidence of calcium depositions We henceforth refer to such von Kossa-positive material as reflecting calcium deposition.
Histologic preparations, autopsy records, and clinical records of all cases were examined for evidence of kidney and parathyroid disease. Cases were identified as having kidney disease if the serum BUN levels exceeded 40 mg/dL, if the serum creatinine level exceeded 1.2 mg/dL, or if significant morphologic evidence of renal parenchymal damage was found. The sex, age, and race of each patient were also documented. Statistical analyses were performed on an IBM-AT computer using a commercial software package (True Epistat, San Antonio, TX). The threshold accepted for significance was P < .05.
Patients in the GWUMC sample were primarily from the greater Washington, DC area. There were 14 patients from the District of Columbia, five patients from adjacent areas in Virginia, and one patient from an adjacent area in Maryland. In addition, there was one patient from North Carolina, one patient from Pennsylvania, and one patient from Pauliabagh, India. For these autopsied cases, 6.3 ± 0.74 sections taken from the left ventricle and 4.0 ± 0.69 sections taken from the right ventricle were examined per case (mean - SD). The GWUMC cases included males and females and blacks and whites (Table 1).
Patients at the SLCVA were primarily from the central inter-mountain area. There were seven patients from Salt Lake City, nine patients from other communities in Utah (six of these from Salt Lake County), two patients from New Mexico, and one patient from each of the following states: Idaho, Montana, Oklahoma, California, and Alaska. All of the SLCVA cases were white males. For the SLCVA cases, 3.52 ± 0.70 sections from the left ventricle and 1.41 ± 0.6 sections from the right ventricle were examined per case.
Following the criteria used in this study, ten cases of kidney disease were found in the GWUMC sample, while seven cases were found in the SLCVA sample. In the GWUMC group, there were two cases of parathyroid hyperplasia and four cases with a malignancy of some type. These malignancies were metastatic leiomyosarcoma of the uterus (one case), well-differentiated adenocarcinoma of the prostate with no demonstrated metastases (one case), renal cell carcinoma without metastases (one case), and adenocarcinoma of the prostate with metastases (one case). In the SLCVA cases, there was one case of parathyroid hyperplasia and no cases of malignancy. However, there was one case of pheochromocytoma, thought to be benign and therefore not considered a malignancy. None of these differences between the SLCVA and GWUMC samples were significant by Fisher's exact test.
Table 2 shows selected features of the composition of drinking water in Salt Lake City and the District of Columbia. Mean values of pH and the content of calcium, magnesium, sodium, and potassium in the six water sources used in Salt Lake City and the three water sources used in the District of Columbia are given for the year 1962. A significant difference in the magnesium content of these water supplies can be seen (P < .0 14). The difference in calcium content of the water supplies of the two areas is not clearly significant (P < .051); calcium, however, cannot be ruled out as a factor of importance from these data. Although the values shown in Table 2 are based on samples taken in 1962, spot checks of data from local water companies in both Salt Lake City and in the Washington, DC area suggest that these differences also existed in 1987, but complete contemporary data for all sources of drinking water in these areas could not be obtained. The more recent data that is available is given in the legend to Table 2. These contemporary data support the prior observation regarding magnesium.
NOTE. Probability values were determined using the two-tailed t test calculated by a commercial software package (True Epistat) on an IBM-AT computer. In 1987, the water content of calcium, magnesium, sodium, and potassium for seven water distribution sites in the Washington, DC area was 36.5 +/- 9.5, 7.0 +/- 3.5, 16.2 +/- 10.4, and 2.4 +/- 0.4 (mean +/- SD). In the Salt Lake City area, the corresponding values for 11 distribution sites were 50.0 +/- 25.6, 14.0 +/- 6.8, 12.0 +/-8.9, and 1.4 +/- 0.6. The differences between these means had probability values of .322, .037, .450, and .008. These contemporary data were obtained directly from the local water districts, and may not be as representative as the data given in the report by Dufor and Becker (20).
The von Kossa-stained slides revealed a striking difference between the SLCVA and GWUMC cases. In the former, there was not a single case of myocardial calcium deposition. In the GWUMC sample, there were 15 such cases of the 23 cases studied (Table 3). Both observers independently classified these 15 cases as positive, and all 23 of the SLCVA cases were classified as negative. There were, however, two cases in the GWUMC sample that showed von Kossa-positive material that could not be scored as positive because it was uncertain whether this positive staining material was within cardiac myocytes. Duplicate sections from the tissue blocks in question were stained by the von Kossa method with heinatoxylineosin counterstain and examined by both observers. Using these sections, both observers agreed that the von Kossa-positive material was not in cardiac myocytes. In one case, the von Kossa-positive material was in a mural thrombus, while in the other case, it was found only in vascular smooth muscle and endothelial cells.
Table 3. Frequency of Myocardial Calcification Among Cases at the SLCVA and the GWUMC ---------------------------------------------------------------- With Calcification Without Calcification ---------------------------------------------------------------- SLCVA 0 23 GWUMC 15 8 ---------------------------------------------------------------- NOTE. The difference between the two samples was highly significant (P < .00002; two-tailed, Fisher's exact test.)
The distribution of the von Kossa-positive material, when present, is of some interest. In some cases, von Kossa-positive granules were found in cardiac myocytes in which cross striations were clearly visible (Fig 1). In these cases, the stained material was finely granular and often distributed in such a way as to accentuate the cross striations. We take it as axiomatic that cells showing calcium deposits are not viable, but we cannot explain the preservation of the normal pattern of cross striations. Further study of these cells was beyond the scope of the present investigation. In other cases, blotches and granules of stained material were found within necrotic myocytes (Fig 2). In yet other cases, positive-staining material was also found in capillary endothelial cells and in vascular smooth muscle cells of otherwise normal-appearing small blood vessels (Fig 3). These differences were not tabulated since it was not always possible to determine if the calcium-containing cells showed cross striations, even when the hematoxylin-eosin-counterstained slides were examined. In one GWUMC case, von Kossa-positive material was found only in endothelial and vascular smooth muscle cells. This case was not counted among the 15 positive cases since our original criteria specified deposits in working cardiac myocytes.
Table 4 shows some characteristics of the GWUMC patient sample. Comparison of cases with myocyte calcification to those without such calcification revealed no significant difference based on race or sex. Analysis of variance (ANOVA) showed no difference in ages among those with and without calcium deposits, even when these groups were subdivided by race. When subdivided by gender, however, ANOVA indicated a clear difference among groups. This was due to the difference between males who were positive for myocardial calcification (52.9 ± 11.9 years) and males who were negative (73.6 ± 3.4 years; P <.0036, two-tailed). There was no demonstrable difference in age between positive and negative cases among females (P = .1045), whites (P = .1354), or blacks (P = .4736). Of the GWUMC cases that were negative for myocardial calcification, four listed the District of Columbia as their site of residence and four listed nearby communities in northern Virginia.
It is possible that the difference in frequency of myocardial calcification between the two populations studied is due to the fact that the GWUMC cases were varied in regard to race and sex, while the SLCVA cases were all white males. There were eight white males in the GWUMC group, and these had an average age of 60.1 ± 14.9 years. This was not significantly different from the average age of the 23 SLCVA cases, which were all white males (61.7 ± 11.1 years). When we only consider the white male GWUMC cases, we still find a significantly increased frequency of myocardial calcification compared with the SLCVA cases (Table 5).
Table 5. Frequency of Myocardial Calcification Among the White Male SLCVA and GWUMC Cases ---------------------------------------------------------------- With Calcification Without Calcification ---------------------------------------------------------------- SLCVA 0 23 GWUMC 5 3 ---------------------------------------------------------------- NOTE. The difference between the two samples was highly significant (P < .00004; two-tailed, Fisher's exact test.)
Table 6 shows the frequency of various ages of infarction as determined histologically. Such infarct dating is obviously imprecise, but the data shown indicate that there was no recognizable difference in infarct age among cases.
Table 6. Age of Myocardial Infarcts Among SLCVA and GWUMC Cases ---------------------------------------------------------------- GWUMC Infarct Age SLCVA Without Calcium With Calcium ---------------------------------------------------------------- Hours 0 1 1 Days 20 8 11 Weeks 11 3 5 Months 16 7 8 ----------------------------------------------------------------
NOTE. Myocardial infarct age was determined as "hours" old if there was a significant lesion comprised of necrotic cardiac myocytes with no associated inflammatory reaction; as "days" old if there was significant inflammation; as "weeks" old if there was granulation tissue; and as "months" old if there was dense scar tissue. Chi square tests showed no difference between groups shown in this table (P = .692). In addition, there was no difference between the SLCVA cases and the combined GWUMC cases (P = .875). The sum of the number of cases with each age of infarction exceeds the actual number of cases because in many cases there were infarcts with more than one age. Among the SLCVA cases, there were only five with infarcts of only one age; all of these were "days" old. Among the GWUMC cases, there were seven with infarcts of only one age. Of these, five were "days" old, one was "weeks" old, and one was "months" old.
It has been known for many years that dystrophic calcification may affect cardiac myocytes. That is, necrotic cardiac myocytes may become calcified, especially in patients who have a high serum calcium level or who are in uremia.(21,22). This, however, is the first report of geographic variation in the incidence of AMI-associated myocyte calcification. The GWUMC and SLCVA samples showed a clear difference in the incidence of AMI-associated myocardial calcification, but a number of potential confounding factors are obvious, including a difference in gender and racial composition. Nevertheless, when only white males were compared, there was still a highly significant difference in the incidence of AMI-associated calcification in the two samples, suggesting that the observed difference was indeed due to a geographic factor. Even if the difference between the two populations studied is due to some unrecognized factor, it may be of great importance. There was no difference in the mean ages of the GWUMC and SLCVA groups. Within the GWUMC group, there was an unexpected and striking age difference between those male patients with AMI-associated myocardial calcification and those without. We have no explanation for this difference, which was not apparent among the females.
The fact that three patients in the GWUMC sample showing myocardial calcification were not from the immediate District of Columbia area suggests that the geography-associated factor that determines this effect does not require long-term residency. We have no information, however, on the water composition in the communities in which these patients resided, nor do we know how long they were in the District of Columbia area before they incurred their AMIs.
Calcification of vascular structures, as noted in several cases (Fig 3), could also be of considerable significance. It has been shown that vascular smooth muscle cells show increased tone when bathed in a medium low in magnesium (23). This may be related to the inverse correlation between serum magnesium and systemic blood pressure (24). Von Kossa-positive granules have been found in the vascular smooth muscle of intramyocardial arteries of hamsters fed a low-magnesium diet, and this was associated with fibrinoid necrosis of some arteries (25). These observations suggest that magnesium deficiency may predispose to coronary artery spasm, and preliminary studies using experimental animals support this hypothesis (26).
The present findings raise the possibility that the factors that influence AMI-associated myocardial calcification may also play a role in the AMI death rate. Previous investigators have shown that reduced dietary intake of magnesium is associated with an increased risk of death caused by myocardial infarction, as previously discussed, and drinking water in the Salt Lake City area has a much higher magnesium content than the water in the Washington, DC area. Furthermore, we have previously shown that reduced levels of dietary magnesium in experimental animals leads to reduced levels of serum magnesium and increased myocardial calcium content, and is associated with increased vulnerability to ischemic and isoproterenol-induced necrosis (12). This property of dietary magnesium has led some investigators to refer to serum magnesium as a physiologic calcium channel blocker (27). This same mechanism has been proposed as the basis for the protective effect of magnesium in the preservation of perfused organs (23).
For the reasons given above, it is tempting to postulate that we found no myocardial infarction associated calcification in the Salt Lake City cases because the high magnesium content of their drinking water led to a low myocardial level of calcium and to a reduced tendency to deposit calcium at the time of infarction. The opposite would be true for the Washington, DC area population. According to this analysis, the calcium-loaded myocardium of the Washington, DC area patients was more vulnerable to necrosis because of the metabolic burden associated with high intracellular levels of calcium (13,29). In both groups, the sympathetic discharge associated with infarction (30) could have been expected to increase calcium influx into heart muscle, but this would have occurred to a greater extent in the Washington, DC area patients because they lacked the protective effects conferred by a high dietary level of magnesium. It is this catecholamine-associated burst of calcium influx, superimposed on an already increased myocardial level, that is presumably responsible for the histologically demonstrable myocardial infarction-associated myocardial calcification in the Washington, DC area patients, since no such calcification is found in patients without myocardial infarction.
The results presented here lead us to predict that persons living in the Salt Lake City area have higher serum magnesium levels and lower myocardial calcium levels than persons living in the District of Columbia area. Although the differences may be small, they may confer a protective effect in myocardial ischemia.
1. Neri LC, Johansen HE: Water hardness and cardiovascular mortality. Ann NY Acad Sci 304:203-219, 1978
2. Karppanen H, Pennanen R, Passinen L: Minerals-coronary heart disease and sudden death. Adv Cardiol 25:9-24, 1978
3. Johnson CJ, Peterson DR, Smith EK: Myocardial tissue concentration of Mg and K in men dying suddenly from ischemic heart disease. Am J Clin Nutr 32:967-970, 1979
4. Punsar S, Karvonen MJ: Drinking water quality and sudden death: Observations from west and east Finland. Cardiology 64:24-34, 1979
5. Marier JR, Neri LC: Quantifying the role of Mg in the interrelationship between human mortality/morbidity and water hardness. Magnesium 4:53-59, 1985
6. Leary WP: Content of magnesium in drinking water and deaths from ischaemic heart disease in white South Africans. Magnesium 5:150-153, 1986
7. Dyckner T: Serum magnesium in acute myocardial infarction: Relation to arrhythmias. Acta Med Scand 207:59-66, 1980
8. DeCarli C, Sprouse G, LaRosa JC: Serum Mg levels in symptomatic atrial fibrillation and their relation to rhythm control by intravenous digoxin. Am Cardiol 57:956 959, 1986
9. Rasmussen HS, Norregard P, Lindeneg O, et al: Intravenous magnesium in acute myocardial infarction. Lancet 1:317-322, 1986.
10. Morgan KJ, Stampley GL, Zabik ME, et al: Magnesium and calcium dietary intakes of the US population. J Am Coll Nutr 4:195-206, 1985
11. Heggtveit HA, Herman L, Mishra RK: Cardiac necrosis and calcification in experimental Mg deficiency. Ann NY Acad Sci 162:758-774, 1964
12. Bloom S: Magnesium deficiency cardiomyopathy. Am J Cardiovasc Pathol 2:7-17, 1988
13. Bloom S: Reversible and irreversible injury: Calcium as a major determinant, in Balazs T (ed): Cardiac Toxicology. Boca Raton, FL, CRC Press, 1981, pp 179-199
14. Kitakaze M, Weisman HF, Marban E: Contractile dysfunction and ATP depletion after transient calcium overload in perfused ferret hearts. Circulation 79:685-695, 1988
15. Chang C, Varghese J, Downey J, et al: Magnesium deficiency and myocardial infarct size in the dog. J Am Coll Cardiol 5:280-289, 1985
16. Chang C, Bloom S: Interrelationship of dietary Mg intake and electrolyte homeostasis in hamsters. j Am Coll Nutr 4:173-185, 1985
17. Bloom S: Effects of Mg deficiency on the pathogenesis of myocardial infarction. Magnesium 5:154-164, 1986
18. Moriyama I, Krueger D, Stamler J: Cardiovascular Diseases in the United States. Cambridge, MA, Harvard University, 1971, pp 49-118
19. Pearse AGE: Histochemistry: Theoretical and Applied. Baltimore, Williams & Wilkins, 1972, pp 1138-1139
20. Durfor CN, Becker E: Public Water Supplies of the 100 Largest Cities in the United States, 1962. Geological Survey Water-Supply Paper. Washington, DC, US Government Printing Office, 1965, pp 136 and 346
21. Gore I, Arons W: Calcification of the myocardium: A pathologic study of thirteen cases. Arch Pathol 48:1-12, 1949
22. Scotti TM: Heart, in Anderson WAD, Kissane JM (eds): Pathology, vol 1. St Louis, Mosby, 1977, pp 745-746
23. Turlapaty PDM, Altura BM: Magnesium deficiency produces spasms of coronary arteries: Relationship to etiology of sudden death of ischemic heart disease. Science 208:198-200, 1980
24. Dyckner T, Wester P: Effect of magnesium on blood pressure. Br Med J 286:1847-1849, 1983
25. Bloom S: Coronary arterial lesions in Mg-deficient hamsters. Magnesium 4:82-95, 1985
26. Yeager JC, Masters TN: Thermographic evidence that an ergonovine-induced coronary artery spasm can be produced in the dog by acutely lowering plasma Mg. Magnesium 1:95-103, 1982
27. Levine BS, Coburn JW: Mg, the mimic/antagonist of calcium. N Engl J Med 310:1253-1255, 1984
28. Shattock MJ, Hearse DJ, Fry CH: The ionic basis of the anti-ischemic and anti-arrhythmic properties of magnesium in the heart. J Am Coll Nutr 6:27-33, 1987
29. Fleckenstein A: Specific inhibitors and promoters of calcium action in the excitation-contraction coupling of heart muscle and their role in the prevention or production of myocardial lesion, in Harris P, Opie L (eds): Calcium and the Heart. San Diego, Academic, 1971, pp 135-188
30. Willerson JT, Buja LM: Beta-adrenergic mechanisms during severe myocardial ischemia and evolving infarction. Postgrad Med 1988, pp 27-32
This page was first uploaded to The Magnesium Web Site on June 12, 1996