4. Risk Assessment
The EU initially focused on the hazard assessment of new chemicals put on the market after 1981 which were required to have some pre-market toxicity testing. In 1993, the EU began to assess the risks of the 100,000 existing and 2,000 new chemicals that have no, limited, or adequate toxicity and/or eco-toxicity data, starting with those whose production exceeds 1,000 tonnes a year (Fig.5). There are 2,500 of these high production volume chemicals (HPVCs) which are currently being assessed by the European Chemicals Bureau (ECB). These risk assessments are performed by the EU and Member States, and they require comprehensive information and data, which are often not available.
For about 75% of the 3,000 or so chemicals in large-scale use, the "minimal" toxicity data required by the OECD for a preliminary assessment of health hazards to humans is not publicly available (NRC, 1984; EDF, 1997). Similarly, little is known about eco-toxicity.
Although searching more databases reveals more data, the results from recent surveys
by the US EPA and the Chemical Manufacturers Association of America confirms that about 75% of chemicals have insufficient toxicity or eco-toxicity data for preliminary OECD risk assessments (CMA,1998).
More toxicity and eco-toxicity data is held by companies and regulatory authorities in the EU and elsewhere. Some of this is confidential and not available to the public for peer review, but some is being made available to the authorities and the public.
It is difficult to even broadly classify chemicals as "dangerous", under labelling requirements, for example, without minimal toxicity data (Temanord, 1997). Box 1 illustrates these main toxicity data gaps. Filling these gaps is a priority, but a costly one. Costs vary from ECU 100,000 for a basic set of toxicity data to an estimated ECU 5 million for the comprehensive toxicity testing of one substance, and up to ECU 15 million for exceptional cases where field testing and monitoring are needed (Teknologi-Rådet, 1996).
|Box 1 Some
toxicity and exposure data gaps
No adequate toxicity data for about 75% of substances in use
No adequate ecotoxity data for 50-75% of the priority (HPVC) chemicals reviewed by the EU
A "major lack" of human health and exposure data for these priority chemicals
Chemical structure data (QSARs) may provide only a reliable estimate of the aquatic toxicity of 15-20% of HPVCs.
Source: European Environment Agency, based on NRC (1984), EDF (1997), van Leeuwen et al. (1996), and van der Wielen (1996)
There is also a need for the toxicity testing of mixtures (EHP, 1997a), since current risk assessments are usually based on single substances. Mixtures may be more or less toxic than the additive effects of single substances. Thousands of animals are needed for these tests, a practice which itself is also a controversial issue. Steps are being taken to minimise use of animals in toxicity testing by the European Centre for the Validation of Alternative methods of the Joint Research Centre. The relevance of animal test results for risks to humans is also difficult to establish, and is complicated by species dependent effects. For example, humans are 100 times more sensitive to the birth-defect impact of thalidomide than rats (Epstein, 1978). Some chemicals may be harmful to rats at high doses, but not to humans at lower doses.
There are also data gaps concerning the pathways and use of chemicals and related exposures (Box 2), which makes risk assessment difficult, especially when the identity of the chemicals is difficult to establish (Johnston et al., 1996). The time taken to complete risk assessments is causing concern (Greenpeace, 1996), and the quality of most of the toxicity data submitted by businesses to the regulatory authorities is not checked by the authorities.
|Box 2 Some other
chemical data deficiencies
Production and use of marketed substances and their presence in consumer products;
Pathways, fate and concentrations of chemicals in the environment;
Human and ecosystems total exposure including multiple exposures and mixtures;
Identity of sensitive subgroups of people, other species and ecosystems;
Nature and costs of impacts on people and the environment and their distribution;
"Eco-efficiency" ratios for the production/use of chemicals.
Source: European Environment Agency
More than 10,000 existing chemicals will be on the ECB database (IUCLID) by the end of 1998. This is a data source for the authorities of Member States and for risk assessments, and ECB will make the non-confidential part available to industry and the public. However, available financial resources are only sufficient for an adequate assessment of about 20-30 chemicals per year. Assessments of 10 chemicals were completed under the EU existing chemicals risk assessment programme by the end of 1997 (Fig. 5 ). Progress will accelerate in 1998, during which another 25 risk assessments are expected to be completed.
The EUs risk assessment programme is part of the legal scheme to reduce the risks of use of dangerous chemicals for man and the environment. Besides that it fits into the OECD programme on the co-operative testing and assessing of HPV Chemicals. The OECD and their member countries are currently working on 240 chemicals and have completed 109 initial risk assessments over the last decade. Of these, 86 substances were considered of "low" risk, 13 needed more testing or exposure information, and, for 10 substances, further risk management measures were considered necessary, in some exposure situations (OECD, 1998a). The World Health Organization, through its Inter-national Programme on Chemical Safety (IPCS) and partners, UNEP and the International Labour Organisation (ILO), have also conducted peer reviewed risk assessments of chemicals since 1976. Over 200 assessments have been completed and published in the WHO Environmental Health Criteria Monographs, with most providing numerical "guideline values" for exposure limits designed to protect people and the environment from damage. Summaries of key scientific studies are published in their new Concise International Chemical Assessment Documents (CICAD), of which the first 6 were published in 1998
(IPCS, 1998). In addition, IPCS also evaluates chemicals in food and pesticides, producing
"acceptable daily intakes" limits for over 1,000 food additives and 220 pesticides. Measures are being taken to harmonise these approaches to risk assessment and to minimise duplication between the EU, the OECD and IPCS programmes.
There is also an increasing focus on the chemical properties of groups of chemicals, such as those that persist and bio - accumulate, rather than on the specific toxicity of single substances (Teknologi-Rådet, 1996; Swedish Ministry of Environment, 1997), and this may help to speed up and focus the process of risk assessment and risk reduction. The problem of accumulative exposures to chemicals with similar biological effects has recently been addressed by the Food Quality Protection Act, 1996, in the USA. This requires the government to consider total risk from several pesticide exposures, rather than from single exposures, when setting "acceptable pesticide residues and daily intakes."
An exposure-based assessment that uses the persistence and spatial range of the chemical as an indicator of environmental threat requires less data, and can usually be performed faster, than a toxic effects based risk assessment (Berg and Scheringer, 1994). It can also help to identify any gaps between those who benefit from chemicals and those who bear the environmental or health damage, as chemicals with a high persistence and spatial range can distribute costs over a much wider area than that which receives the benefits, as, for example, with CFCs and ozone layer damage. It has been suggested that exposure-based threat assessment could be used for the initial screening of chemicals, complemented by toxic effects risk assessment, where this is likely to be cost effective and where data is available (Scheringer, 1997).
Exposure potential is also important for assessing the toxic risks from chemical emissions those assessment methods that incorporate the most comprehensive human exposure data seem to produce the best estimates of risk, which can vary by 3 orders of magnitude (i.e. by 1,000 times), depending on the method chosen (Hertwich et al., 1998).