Why spatial and temporal variations in magnesium in drinking water are important considerations when magnesium in drinking water is used as exposure in health related studies

Abstract

TPS 931: Water and foodborne chemicals, Exhibition Hall, Ground floor, August 28, 2019, 3:00 PM - 4:30 PM Background and objectives The aim of the study was to investigate the stability of magnesium, sodium, potassium and calcium in drinking water and thereby be able to improve exposure estimates in future health related studies. Focusing on magnesium, the objective was to investigate annual temporal and spatial variations in magnesium in drinking water for the period 1980 – 2017 in Denmark. Methods: Public waterworks were linked to water supply areas by spatial joins and the number of addresses within each water supply area was calculated. For each public water supply area (WSA), linear regression and change point analyses were used to analyse the temporal variations in magnesium concentration. Afterwards, each WSA was categorized according to the analysed temporal trend in magnesium concentration. Spatial variations were analysed using spatial scan statistics in SatScan. Results: The study included 62,941 magnesium analyses, covering 2,539 WSAs, 99.5 percent of the Danish addresses, and 417 to 2,106 samples per year. Mean magnesium concentration = 12.1 mg/L and median = 9.8 mg/L. The magnesium concentration was constant for 40 percent of the WSAs, covering 18 percent of the Danish addresses. Two statistically significant clusters were detected, a hotspot (18.2 mg/L, p-value <=0.001) in the eastern part of Denmark and a cold spot (8.1 mg/L, p-value <= 0.001) in the western part of Denmark. Conclusion: A mean magnesium concentration, covering 37 years, is adequate as exposure estimate for two fifths of the WSAs, though only covering about one fifth of the addresses in Denmark. Furthermore, the magnesium concentration in drinking water was clustered at a regional scale. In future health related studies, it is therefore recommended, that magnesium exposure estimates are not only based on a single mean value, but annual temporal and spatial variations should also be included.