THE RISK ASSESSMENT OF ORGANIC COMPOUNDS IN DRINKING WATER
Report No FR0401

A Patel

Sept 1993

SUMMARY

I BENEFITS

As the toxic mechanisms for organic contaminants are better understood, less uncertainty is involved in the risk assessment process. This may lead to drinking water standards that are less conservative than previously established, but which will still a fford adequate protection to consumers. A knowledge of the way in which such standards are calculated will help in understanding why small differences in the numerical values are not usually significant from a health point of view. This in turn will be of value when dealing with enquiries from customers and the media in relation to chemical contamination of drinking water.

II OBJECTIVES

  1. To highlight existing drinking water standards that may need amending because the methods used in the original calculations may now no longer be appropriate.

  2. To review the methodologies used to calculate drinking water standards for organic chemicals.

  3. To investigate the uncertainties and assumptions that go into the calculations and to show how these affect the numerical value of standards.

III REASONS

It is important that drinking water standards are based on sound scientific judgment. A conservative approach is usually taken when standards for organic chemicals are derived, especially for chemicals that have been found, or are suspected to be carcinogenic. This approach reflects uncertainties in the data and our knowledge of toxic mechanisms. As new toxicity data become available and the toxic mechanisms are better understood, less uncertainty is involved in the risk assessment process, and this may lead to revised numerical values for some standards. The calculated values necessarily contain a number of assumptions and a substantial allowance for uncertainties in the extrapolation process. This will of course impact significantly on the final numerical value. Small changes in the values may not be significant from a toxicological or health point of view, but this is difficult to convey to the consumer and, in terms of compliance, such changes could make a large difference in the costs for water treatment in achieving given standards. There are also instances when water suppliers need to know how a standard has been derived. This is necessary to understand the health significance if a standard is approached or exceeded, and will also be of value in dealing with enquiries from consumers and the media on chemical contamination of supplies.

IV CONCLUSIONS

  1. The drinking water standards for some of the organic chemicals listed in the UK Water Supply (Water Quality) Regulations 1989 need updating to incorporate the findings of more recent toxicity data. These include tetrachloroethene and trichloroethene for which there are new data that enable a better understanding of their mechanisms of toxicity. The new WHO (World Health Organization) guideline values for these chemicals have incorporated the more recent data.

  2. In the absence of national standards for many organic chemicals, the guideline values and standards produced by the WHO and the US EPA (US Environmental Protection Agency), or, where these are not available, recommendations made by the WRc are used for guidance. It is important that these are interpreted correctly.

  3. Essentially there are two methods of deriving standards, the `threshold approach' for chemicals that are only toxic when a critical threshold concentration has been reached, and the more conservative `model approach' for chemical carcinogens for which a critical threshold cannot be demonstrated. The organoleptic properties of a chemical should also be considered during standard setting and it should be made clear to water suppliers if the taste or odour threshold concentration is lower than the health-based limit.

  4. Standards produced by various organisations can differ because of differences in approach to risk assessment- The EPA use the model approach for all carcinogens irrespective of the mechanism of carcinogenicity, whereas the WHO (and WRc) use the threshold approach for carcinogens which operate through a mechanism for which there will be a demonstrable threshold, and the model approach for those that do not. The EPA standards for some carcinogens are therefore significantly more conservative (i.e. more stringent) than WHO guidelines for the same chemicals.

  5. Most drinking water standards (e.g. WHO guideline values, EPA standards) are long-term, i.e. derived on the assumption that a person would be exposed to that concentration of the chemical, every day, for a lifetime. Standards in most cases, therefore, tend to be conservative and short-term exceedances do not usually mean that exposure will result in any adverse health effects, although expert toxicological advice should still be obtained in these instances.

  6. The numerical value of a standard can vary by orders of magnitude depending on the methodology used in the risk assessment and on the uncertainties and assumptions made in the calculations. This is particularly true for carcinogens where even the choice of mathematical model can lead to standards that vary by over five orders of magnitude. Standards should not, therefore, be interpreted as absolute values other than for regulatory purposes.

  7. Standards can change as more toxicity data become available for chemicals. As our understanding of toxic mechanisms improves, less uncertainty is involved in the risk assessment. This may lead to some standards changing and, in many cases, to less conservative values.

V RESUME OF CONTENTS

The number of organic chemicals currently regulated are few, and the limits for some of these need updating so that more recent toxicity data are taken into account. For guidance, water suppliers may use WHO guidelines, US EPA standards (MCLs) and health advisories (HAs) or, if these are not available, WRc SNARLs (suggested no adverse response levels) for the many organic chemicals that are not regulated. These organisations can adopt different approaches to standard setting resulting in different evaluations. It is important that the different values produced by each organisation are interpreted correctly and, to do this, an understanding of the methods used to derive standards is necessary.

Animal toxicity data are usually used to derive a safe dose of a chemical for man. The method used to extrapolate animal data to man will depend on the chemical and its mechanism of toxicity. For mohere are new data that enable a better understanding of their mechanisms of toxicity. The new WHO (World Health Organization) guideline values for these chemicals have incorporated the more recent data.

  • In the absence of national standards for many organic chemicals, the guideline values and standards produced by the WHO and the US EPA (US Environmental Protection Agency), or, where these are not available, recommendations made by the WRc are used for guidance. It is important that these are interpreted correctly.

  • Essentially there are two methods of deriving standards, the `threshold approach' for chemicals that are only toxic when a critical threshold concentration has been reachedfirst have to decide what level of risk is acceptable, before standards can be derived for such chemicals.

    Once a safe dose of a chemical for man has been estimated, a drinking water standard is calculated by deciding how much of the chemical could come from drinking water, compared to other sources, such as air and the diet. Assumptions also have to be made, regarding the average daily drinking water consumption. A 60 kg adult drinking 2 litres of water per day is normally used to calculate standards for organic chemicals, although a child or bottle fed infant may be used, if it is considered that the latter are particularly susceptible to the toxic effects of a chemical.

    The methods used to derive health-based standards do not consider the organoleptic properties of chemicals. Some chemicals can, however, cause taste or odour at concentrations much lower than a health-based limit. It should be made clear where this is the case, and it is useful if water suppliers have access to both health-based standards and the organoleptic threshold concentrations.

    Much uncertainty is involved in the methods used to calculate standards and it is not always clear, which method should be used for a given chemical. This uncertainty, together with the assumptions made in the calculations, lead to a standard that can v ary by many orders of magnitude, depending on the criteria used. This uncertainty means that a very conservative approach is usually taken during standard setting. Most drinking water standards include substantial margins of safety and have been calculated, assuming that a person would be exposed to that concentration of chemical, every day, for a lifetime. For this reason, short-term exceedances of drinking water standards do not usually mean that exposure will result in any adverse health effects.

    Copies of the Report are available from FWR, price £15.00 less 20% to FWR Members