Health Stream Literature Summary - Issue 50 - June 2008

Hard drinking water does not protect against cardiovascular disease: new evidence from the British Regional Heart Study.
Morris, R.W., Walker, M., Lennon, L.T., Shaper, A.G. and Whincup, P.H. (2008) European Journal of Cardiovascular Prevention and Rehabilitation, 15(2); 158-189.

Studies published from 1956 to 2004 have generally suggested a protective effect of hard drinking water, in particular high calcium and magnesium intake against cardiovascular disease (CVD). In Britain water hardness varies greatly, with harder water in the south and east of the country and softer water in the north and west. Regional variations in mortality from CVD have been observed in the United Kingdom for many decades and it was noted almost 50 years ago that CVD mortality rates were lowest where water was the hardest. The British Regional Heart Study (BRHS) is a cohort study of men in 24 British towns with a wide variation in water hardness levels. Participants have been followed up for coronary heart disease (CHD) and CVD over 25 years. This study offers the opportunity to examine the relationship between water hardness and CVD.

From each of the 24 towns in the study a single general practice typical of the town’s socio-demographic profile was chosen. A random sample of 400 men aged 40-59 years was taken from each age-sex register. Men were invited for a screening examination during 1978-1980 with 78% attending. At the time of screening, a single measure of water hardness, in calcium carbonate mmol/l equivalent was ascertained for each town. Information was also gathered regarding analytical results on water passing from various sources into the distribution system from relevant water authorities. Three drinking water samples (first draw, random daytime and flushed) were collected from a random sample of approximately 40 men selected from each town and analysed for calcium and magnesium concentration as well as for water hardness. The mineral intake for each participant was based on calcium and magnesium tap water concentrations which were multiplied by a questionnaire estimate of tap water consumption (including consumption of water, tea and coffee at home and locally). Nurses visited each town between January 1978 and May 1980, and the height and weight of participants was measured and body mass index calculated. A serum sample was provided to measure nonfasting total cholesterol. A questionnaire was administered including questions on smoking habits, alcohol intake, physical activity and social class based on the longest held occupation.

All participants were followed up for major nonfatal and fatal CHD events (myocardial infarction including sudden cardiac death) and stroke through death registries and regular biennial review of general practice records throughout the follow-up period. Follow up for more than 25 years was available for mortality and until June 2004 for nonfatal episodes of CHD and stroke. There was data available for 7735 men on water hardness measures at the town level. There were 5796 men meeting the inclusion criteria for analysis of incidence of CHD or CVD. Of these, 998 experienced a major episode of CHD over the 25 years of follow-up (0.74 per 100 person-years); 564 of these episodes were fatal. There were 1371 episodes of CVD (major CHD or stroke). There were 721 men who were initially free of CHD who provided tap water samples. During follow-up of these men 126 events occurred (0.75 per 100 person-years).

Water hardness varied from 0.27 mmol/l in the town with the softest water to 5.28 mmol/l in the town with the hardest water. At the town level there was a weak inverse association between 25-year incidence of CVD and water hardness. However after adjustment for established cardiovascular risk factors there was no statistically significant association found for CVD. Statistical modelling was conducted using individual–level survival data and the hazard ratio (HR) for a two-fold increase in water hardness on 25-year incidence after adjustment for seven confounding variables. The relationships were found not to be statistically significant, especially for CHD (HR 0.99, 95% CI, 0.94-1.04, P=0.62) and CVD (HR 0.96, 95% CI, 0.91-1.01, P=0.083). HRs were of similar magnitude for CHD mortality as for CHD incidence. No significant relationship was found for calcium intake with CHD incidence, CHD mortality and CVD incidence. A positive association was found for magnesium intake with incident CHD. After adjustment for confounding variables, the HR per two-fold increase was 1.10 (95% CI, 1.00-1.20), P=0.045. For CVD after adjustment the HR was 1.06 (95% CI, 0.99-1.14, P=0.087). No significant relationship was found for magnesium intake and CHD mortality. After subdividing magnesium intake by tertiles, the positive association did not display a dose-response relationship.

This long-term study does not provide evidence for an appreciable protective effect of any measure of water quality exposure at the individual level against CHD or CVD in general. The best estimate of CVD mortality reduction associated with a doubling of water hardness in this study was 4%. Therefore any protective effect is likely to be extremely small and of less importance to public health that the well-established cardiovascular risk factors such as high cholesterol and high blood pressure.