Distribution of arsenic hazard in public water supplies in the United Kingdom – methods, implications for health risks and recommendations

Public water supplies in the United Kingdom are highly regulated and monitored and, in particular, have an outstanding compliance with regulatory standards, particularly with respect to the UK PCV (prescribed concentration value) for arsenic of 10 μg L⁻¹. Nevertheless, many UK public water supplies contain arsenic at concentrations within a factor of 10 of the PCV. Given increasing concerns over detrimental health outcomes arising from chronic exposure to drinking water containing arsenic at sub-regulatory concentrations in the 10⁰μg L⁻¹ range, quantifying the distribution of arsenic intake from consumers exposed to arsenic via drinking water in the UK is indicated. Using the limited secondary summary water quality data available in the public domain from the Drinking Water Inspectorate and assuming a log normal distribution, we calculate that, in 2015, on the order of 10⁵ consumers in the UK were supplied with drinking water with arsenic concentrations at or above 5 μg L⁻¹; 10⁶ at or above 2μg L⁻¹ and 10⁷ at or above 1μg L⁻¹. However, examination of much more detailed secondary data kindly supplied by individual UK water supply companies indicates that the overall distribution of arsenic hazard is not log normally distributed and results in an overestimate of the number of the consumers exposed to high As concentrations using that assumption. Our more detailed analysis shows that approximately 130,000 consumers in the UK are supplied with drinking water with arsenic concentrations at or above 5 μg L⁻¹; the equivalent figures for other concentrations being 1,080,000 for at or above 2 μg L⁻¹ and 9,750,000 for at or above 1 μg L⁻¹. Epidemiological evidence seems currently insufficiently powerful to reliably quantify the detrimental health outcomes arising from such sub-regulatory exposures, but arsenic-attributable premature avoidable deaths in the UK on the order of 100 to 1000 per annum are plausibly estimated here from combined cancer and cardiovascular disease causes. There are considerable uncertainties in these estimates due to (i) model (e.g. linearity, threshold) and parameter uncertainties in the dose-response relationships at such low concentrations; (ii) partial reliance on ecological studies, which may be sensitive to the nature of adjustment for socio-economic and other potential confounders of risk and (iii) the lack of explicit consideration of the many other sequela for which arsenic is known, at higher concentrations in drinking water, to contribute. We note that, the estimates here, however, are broadly equivalent to the number of annual fatalities of car occupants in road traffic accidents in the UK.