1Supported by grants from the Bureau of Chronic Disease, California State Department of Public Health, and the Research Committee, Bay Area Heart Association. Received for publication January 31, 1969.
2Present address: School of Public Health, University of Hawaii, Honolulu.
3The authors are grateful to Dr. Nemat Borhani, former Chief, Bureau of Chronic Disease, and to the Bureau of Sanitary Engineering, California State Department of Public Health, for their assistance; and to the food and beverage processors for food composition data concerning their products. We also appreciate the constructive suggestions from this JOURNAL'S reviewers.
ASSOCIATIONS BETWEEN hardness of drinking water and mortality from cardiovascular diseases have been reported for the United States, Great Britain, and Sweden (1-4). Correlations were negative and showed lower death rates in areas with hard water supplies. These associations could be due to: (a) the concentration of calcium or magnesium, the principal cations contributing to water hardness; (b) trace elements or other constituents common to hard water; or (c) spurious factors. If the relationships were due to calcium or magnesium, we would expect persons consuming hard water to have significantly higher calcium and magnesium intakes, possibly higher serum and urinary levels of these minerals, and lower cardiovascular disease rates than those drinking soft water.
Before comparing hard and soft water consumers, we conducted a pilot study among adult volunteers in a hard-water community in California during December 1967 to learn if water made a significant contribution to the total calcium and magnesium intakes. This paper describes our preparatory activities for analyzing reported food intakes, the data collection and results, and some implications of the findings.
Although the limitations of food composition tables are known, standard values are needed for estimating usual dietary intakes in studies of large groups, A primary problem for our survey was the dearth of published data on the magnesium content of foods, Because magnesium has been included in the 1968 revision of the Recommended Dietary Allowances (5), it seems likely that other investigators may share our concern.
For this study, we compiled the values for calcium and magnesium of selected foods from several sources. Official documents of the United States (6), Great Britain (7), and Germany (8) were our primary references. Additional resources were the publicatiions of the California Agricultural Experiment Station (9) and of private or commercial organizations (10-13). We also obtained unpublished data from the U.S. Bureau of Commercial Fisheries and from fifteen manufacturers through personal correspondence. Table 1 classifies the foods into broad categories, such as "meat and poultry," "eggs and dairy products," and so on, and includes mean food intakes, the range of values, and the references for the published values. Using atomic absorption methods, we analyzed samples of carbonated beverages for calcium and magnesium.
We designed a seven-day measured record for obtaining quantitative estimates on food and water intakes and developed standard methods for converting the reported data into equivalent gram weights. We also conducted preliminary experiments on spaghetti and rice for calculating the changes in weight and volume in cooking. The results indicated a ratio of 1 part dry to 2 parts water for spaghetti or other pastes and a ratio of 1 part dry to 2.5 parts water for rice.
We selected the community of Brentwood because its water supply was reported to have relatively high values of calcium and magnesium, 8.6 and 3.6 mg. per 100 ml., respectively (14). With the assistance of the local public health and civic leaders, we arranged a community meeting to recruit adult volunteers. After presenting our objectives, we demonstrated methods of measuring and recording intakes of food and water and then distributed seven-day booklets, directions for recording food preparation and intakes, household measuring equipment, and stamped envelopes for returning the records.
Our standards for acceptability of a record were that the measurements of all items consumed at home be complete. Some omissions for restaurant dishes, food purchases, or preparation methods occurred, in which case we telephoned the participants for clarification. From the reported data, we calculated the proportion of water in all beverages and in all home-prepared cereals, soups, stews, casseroles, and legume dishes.
A frequency tally indicated that canned or frozen vegetables were more popular than fresh cooked varieties during the study week. Since we could not quantitate evaporation losses in cooking or the extent of mineral exchange between the water and fresh produce, we did not calculate the contributions of local water to these items. It also was not practical to perform experiments on fresh vegetable cookery due to large individual variations in product selections and in cooking times. Another omission pertained to desserts and bread items. We found that most persons purchased packaged and bakery products or used commercial mixes. Because the magnesium values of mixes and several baking ingredients were unknown, we decided to code all items in these categories according to the available data of comparable prepared products.
The daily records showed that home water supplies represented three main categories: hard city, hard well, and softened or treated city and well waters. We collected random samples of tap water from the three categories, performed calcium and magnesium analyses by atomic absorption methods, and used group means for calculating the contributions of the particular source of water to the daily intakes. The State Department of Public Health provided information on the mineral contents of the water consumed in other California communities. The individual and group intakes of calcium and magnesium from food and from all sources of water and the percentage of the total intakes contributed by water were calculated. We also classified participants according to their intakes of hard or soft water at home and compared the percentage contributions of water to intakes of calcium and magnesium for the two groups.
The laboratory analyses of water samples are shown in Table 2.
Our data for calcium in hard water were considerably less than the value of 8.6 mg. per 100 ml reported by the California State Department o Public Health (14). The difference could be due to sampling or laboratory methods. We obtained samples from kitchen taps and used atomic absorption methods, whereas the Department of Public Health sampled at a central distribution point and used oxalate precipitation and EDTA or permanganate titration (15). The analyzed and reported data for magnesium, however, were similar.
Of the sixty-five adults attending the community meeting, fifty-six volunteered for the survey. Of these, forty-nine persons returned the records, while two refused and five did not respond after follow up phone calls. Four records were rejected due to incomplete information. The group with acceptable records included twenty-seven men and eighteen women, representing thirty-five households. The age of the participants ranged from twenty-three to seventy-four years.
Water contributed 6.9 ± 5.8 per cent calcium and 12.5 ± 9.8 per cent magnesium to the total intakes for the forty-five participants. Classification of subjects by hard or soft home water supplies yielded twenty-three in the hard and twenty in the soft water groups. Two persons consumed both types of water at home and were omitted from these calculations. For the two sub-groups, the mean contributions of water for calcium and magnesium were, respectively: hard water supply, 9.7 ± 6.2 per cent and 18.1 ± 9.2 per cent; soft water supply, 3.4 ±3.2 per cent and 5.4 ± 4.7 per cent.
Although these results include the errors of measurement, interpretation, and food composition, the findings suggest that food, not water, was the major determinant of the calcium and magnesium intakes of the forty-five volunteers. We also may assume that the contribution of hard water to magnesium intakes was overestimated, because all records included one or more items of unknown magnesium content. The magnesium values for fresh meat, poultry, fish, fruits, and vegetables were generally available, but we found minimal data for items such as cereals, breads, processed entrées, candies, and desserts. The diversity and frequency of unknowns made it impractical to select any particular items for laboratory analyses.
Our compilation of reported food composition values also revealed large variations for several items. Because we had little information on the variability of the calcium and magnesium contents of foods, we were forced to use the means of the published values. Without knowledge concerning the variation in these means, the significance of studies such as ours becomes difficult to interpret. We were pleased that Zook and Lehmann recently reported both the means and standard deviations in their analyses of the mineral contents of selected fresh fruits (16). Their data, published after our study, seem particularly useful, and we recommend that future publications on food composition follow similar procedures.
Like other investigators conducting dietary studies in California (17), we observed large variations in the daily eating patterns of the Brentwood subjects. The socio-economic status of these volunteers was probably high, since they were either high school or college graduates and were engaged in professional, technical, or business careers. The heterogeneity of their food habits may be typical of similar groups in a Western society who volunteer to keep accurate dietary records. Before conducting this study, we recognized that the frequent use of new food products and meals eaten in restaurants could increase measurement and interpretation errors in the food diaries. Nevertheless, we selected a field study, rather than a metabolic investigation, to obtain a picture of the usual eating practices of a larger sample living in a hard water community.
Variation in food patterns and lack of food composition data were not the only problems in this survey. We found that residence in a hard water area did not insure an intake of only hard water. Approximately half of our subjects drank softened or treated water at home. The participants also were highly mobile, and several reported drinking water or coffee in six or seven different localities during the week. This suggests that other factors, such as the use of home water softeners, place and kind of employment, and location of restaurants, may be as important as the reported hardness of the municipal water supply.
Our plans for further research on the relationship of water constituents and cardiovascular diseases will depend on the identification of a stable population that consumes primarily hard, untreated water. We also need additional reliable values concerning the mineral contents of available food items and hope that our preliminary survey will stimulate others to perform laboratory analyses on food composition and on mineral exchanges in water cookery.
A pilot study among forty-five adult volunteers living in a hard-water community of California was conducted to determine the contribution of water to total calcium and magnesium intakes. A food composition table was compiled and local water supplies were analyzed for calculating seven-day calcium and magnesium intakes. The contribution of water to the total mineral intakes averaged 7 per cent for calcium and 12 per cent for magnesium. The magnesium contributions were probably overestimated due to lack of nutrient data for several reported foods.
The frequent use of water softeners and the mobility of the subjects decreased the expected contributions of hard water to total calcium and magnesium intakes. Further studies on the possible relationship of water constituents to cardiovascular diseases will depend on the identification of a population that consumes primarily hard, untreated water and on the availability of additional information concerning the content of foods.
(1) SCHROEDER, H. A.: Relation between mortality from cardiovascular disease and treated water supplies. J.A.M.A. 172: 1902, 1960.
(2) SCHROEDER, H. A.: Municipal drinking water and cardiovascular death rates. J.A.M.A. 195: 81, 1966.
(3) MORRIS, J. N., CRAWFORD, M. D., AND HEADY, .J. A.: Hardness of local water-supplies and mortality from cardiovascular diseases in the county boroughs of England and Wales. Lancet 1: 860, 1961.
(4) BIORCK G., BOSTROM, H., AND WIDSTROM, A.: On the relationship between water hardness and death rate in cardiovascular diseases. Acta. Med. Scand. 178: 239, 1965.
(5) FOOD & NUTR. BD.: Recommended Dietary Allowances, Seventh Revised Edition, 1968. Natl. Acad. Sci. Pub. 1964, 1968.
(6) WATT, B. K., AND MERRILL, A. L.: Composition of Foods-Raw, Processed, Prepared. Revised. USDA Agric. Handbook No. 8, 1963.
(7) MCCANCE, R. A., AND WIDDOWSON, E. M.: The Composition of Foods. 3rd ed. Medical Research Council Special Rept. Series No. 297, 1960.
(8) SOUCI, S. W., FACHMAN, W., AND KRAUT, H.: Die Zusammensetzung Der Lebensmittel, Nahrwert-Tabellen. Vol. I & II. (German). Stuttgart: Wissenschaftliche Verlagsgesellschaft M.B.H., 1962.
(9) HOWARD, F. D., MACGILLIVRAY, J. H., AND YAMAGUCHI, M.: Nutrient Composition of Fresh California Grown Vegetables. Calif. Agric. Exper Sta. Bull. No. 788, 1962.
(10) CHURCH, C. F., AND CHURCH, H. N.: Bowes and Church Food Values of Portions Commonly Used. 10th ed. Philadelphia: J. B. Lippincott, 1966.
(11) Newer Knowledge of Milk. 3rd ed. Chicago: Natl. Dairy Council, 1965.
(12) Nutritional Data. 5th ed. Pittsburgh: H. J. Heinz Co., 1963.
(13) Nutritive Composition of Products. Camden, N. J.: Campbell Soup Co., 1966.
(14) California Domestic Water Supplies. Berkeley: Bur. of Sanitary Engin., Calif. Dept. of Pub. Health, 1962.
(15) Personal communication.
(16) ZOOK, E. G., AND LEHMANN, J.: Mineral composition of fruits. 2. Nitrogen, calcium, magnesium, phosphorus, potassium, aluminum, boron, copper, iron, manganese, and sodium. J. Am. Dietet. A. 52: 225, 1968.
(17) HANKIN, J. H., REYNOLDS, W. E., AND MARGEN S. A short dietary method for epidemiologic studies 2. Variability of measured nutrient intakes. Amer. J. Clin. Nutr. 20: 935, 1967.
This page was first uploaded to The Magnesium Web Site on August 28, 2002