Health Stream Literature Summary - Issue 50 - June 2008

A systematic review of analytical observational studies investigating the association between cardiovascular disease and drinking water hardness.
Catling, L.A., Abubaker, I., Lake, I.R., Swift, L. and Hunter, P.R. (2008) Journal of Water & Health, 6(4); 433-442.

This paper summarises a comprehensive and systematic review and critical assessment of analytical observational epidemiological studies examining drinking water hardness, calcium or magnesium and cardiovascular disease. Electronic databases were searched for both medical subject headings (MeSH) and text words where appropriate. The titles and abstracts of the papers were screened and papers were retained if they presented primary data of human studies, were related directly to the research question, involved a comparison of populations or individuals at different levels of exposure using a case control or cohort design and were in English. Papers were reviewed to extract data and the methodological quality of studies was assessed.

After exclusions only 14 epidemiological studies presenting analytical results were included in this review from the 2,906 studies initially identified. There were nine case control studies identified. Only one of the case control studies considered drinking water hardness as the exposure variable and examined its association with deaths from atherosclerotic cardiovascular disease. No significant association was reported in this study. One study included individuals both alive and deceased following first acute myocardial infarction (AMI) and therefore differential effects on morbidity or mortality could not be examined. The seven remaining case control studies assessed concentrations of specific drinking water constituents, namely magnesium and calcium and cardiovascular disease mortality. Of these, five studies reported no evidence of a statistically significant association between calcium concentrations and cardiovascular mortality and no consistent direction of association. Two of the seven studies reported a protective effect of drinking water calcium on mortality from AMI for females and males and females combined. Five of the seven studies found a statistically significant protective effect of drinking water magnesium against mortality from AMI, hypertensive disease and stroke for males and females.

There were three cohort studies reported in five publications. One of the cohort studies examined mortality from stroke and arteriosclerotic heart disease in a large cohort of US males and females. There was no significant association found between water hardness and stroke mortality and only a weak suggestion that soft water was harmful in females and possibly associated with a slightly greater risk of sudden death. One study examined coronary heart disease (CHD) incidence and mortality in Finnish males. Drinking water calcium and magnesium concentrations were not associated with individual level risk factors for CHD. A lower, but non-significant drinking water magnesium concentration was described for males dying from CHD during a follow up period of 10 and 15 years. The final cohort study presented findings of a 15 year follow up of the British Regional Heart Study. The occurrence of both fatal and non-fatal CHD was examined. There was an inverse but non-statistically significant association between water hardness and CHD incidence found after controlling for age alone, or all individual level risk factors.

Where appropriate data were available, studies were summarised according to exposure type and meta-analysis was used to pool odds ratios. It was not possible to summarise relative risks from the cohort studies using meta-analysis because of differences in the outcome measures used. Seven of the case control studies examining cardiovascular mortality and drinking water magnesium or calcium were included in the meta-analysis. Drinking water magnesium concentrations in the highest exposure category (range 8.3 to 19.4 mg/L) were significantly associated with a decreased likelihood of cardiovascular mortality (OR 0.75 (95% CI 0.68, 0.82), p less than 0.001), compared with the baseline (range 2.5 to 8.2 mg/L). It was not found to be appropriate to derive a single summary estimate for calcium and cardiovascular mortality.

This review supports an association between cardiovascular mortality and magnesium levels in water. Better designed epidemiological investigations are required to examine this association as a comparatively small reduction in cardiovascular risk seen with increasing magnesium concentrations in drinking water may translate into substantial benefits for public health at the population level. The challenge for future studies is to accurately measure individual consumption of minerals in drinking water and diet and the mineral nutritional status of the study population, and to determine the biologically relevant dose.