TR 056 – Aquatic Toxicity Data Evaluation

TR 056 : Aquatic Toxicity Data Evaluation | December 1993

A database, termed the ECETOC Aquatic Toxicity (EAT) database, consisting of original published information on the toxicity of substances to aquatic species in fresh and saline waters has been compiled. The principal quality criteria for acceptance of data were that test methods should be well described and the toxicant concentrations must be measured. On this basis 42% of the 530 papers examined were found to be suitable for inclusion in the EAT database.

The EAT database input software is easy to use and has been prepared for Personal Computers operating in a DOS environment. For each entry there are 21 fields of information on the substances, test species, test details, results and source references. All the references are held at ECETOC.

The EAT database includes information on 368 substances for 122 aquatic test species. Publications from 1970 to 1991 have been assessed, giving 2200 entries; organohalogens and heavy metals make up most of the data entries which reflects historical concern over these chemicals. The toxicity of all substances was evenly distributed in a log normal scale. Some groups of substances were found to be more toxic than others. The higher toxicity is not necessarily linked with substances of historical concern. Sensitivity of the test organisms to chemicals was evenly distributed. On the scale of sensitivity bacteria seem to be least sensitive and invertebrates most sensitive as a general rule.

Analysis of the data can be performed using a number of specially written routines. These include the ability to select data using various options, to prepare simple counts, frequency distributions, ratios (e.g. between acute and chronic results) and regression analyses. The regression analyses take account of the fact that in comparing toxicity test end-points there are no dependent or independent axes in the strictest sense.

In order to provide a scientific basis for application factors used in risk assessment, the ratio of acute EC₅₀:chronic NOEC was assessed for 12 different selections of data. The median ratios varied from 3.6 to 28.0. When 19 substances typical of those which could be notified under the provisions of the Seventh Amendment Directive (92/32/EEC) were considered the range of ratios was 1.25 to 28.3. The maximum acute EC₅₀:chronic NOEC value of 28.3 (for 100% of substances) is an indication that the factor of 40 (for 90% of substances) given in ECETOC (1993) may be rather conservative. The latter was derived from the same database using a different statistical approach, i.e. not allowing for the separate assessment of individual species.

This report includes a summary of approaches to hazard assessment, especially the use of different application factors and describes the result of these using three substances.