Technical Report
29.10.1995

TR 067 – The Role of Bioaccumulation in Environmental Risk Assessment: The Aquatic Environment and Related Food Webs

TR 067 : The Role of Bioaccumulation in Environmental Risk Assessment: The Aquatic Environment and Related Food Webs | October 1995

In this report the assumptions and equations used to determine bioconcentration and bioaccumulation of substances are reviewed and a possible approach for the integration of these processes into a risk assessment is discussed. The approach has been evaluated with the use of different bioaccumulation models for several representative substances with different properties and use patterns.

A distinction is made between bioconcentration and bioaccumulation. Bioconcentration is defined as the net result of the uptake, distribution, and elimination of a substance in an organism due to water-borne exposure, whereas bioaccumulation includes all routes of exposure (i.e. air, water, soil, food). Biomagnification is defined as accumulation and transfer of substances via the food web. The report describes the critical mechanisms underlying these processes with a brief description of some of their implications on environmental exposure and effects. It also addresses methods for predicting and determining the potential of substances to bioconcentrate or bioaccumulate.

It was concluded that bioaccumulation of a substance into an organism is not an adverse effect or hazard in itself. Bioconcentration and bioaccumulation may lead to an increase in body burden which may cause toxic effects due to direct (water) and/or indirect (dietary) exposure. Bioaccumulative substances characterised by high persistence and toxicity, negligible metabolism and a log Kow between 5 and 8 may represent a concern when widely dispersed in the environment. Therefore, when appropriate, the potential of a substance to bioaccumulate in the aquatic environment should be included as an exposure related parameter in risk assessment. Biomagnification, resulting in an increase of the internal concentration in organisms at succeeding levels in the trophic chain, is not as widespread as commonly believed; it has only been demonstrated for a very limited number of substances.

The potential of a substance to bioaccumulate is related primarily to its lipophilicity. A surrogate measure of lipophilicity is the n-octanol-water partition coefficient (Kow), which is correlated with bioconcentration potential. Therefore, Kow often used as predictor in quantitative structure-activity relationships (QSARs) for bioconcentration factors (BCF) of organic non-polar substances. Such QSARs, however, are not universally applicable. While predictions of BCF in aquatic organisms for lipophilic, nonionic substances undergoing minimal metabolism or biotransformation may be satisfactory, there are exceptions, and the equations to predict BCF are best used only within the chemical class for which the QSAR was developed. Substances with a low lipid solubility, substances with a molecular weight well above 700, or substances which are considered as highly lipophilic will not be taken up as predicted from simple QSARs for BCFs. Non-linearity of BCF versus log Kow for highly lipophilic substances has been demonstrated. Similarly deviations have been reported also for other chemical classes such as surface-active, ionisable and polar substances.

When biotransformation of the substance by the organism occurs, elimination may increase significantly thus reducing bioconcentration. The major drawback of the various models proposed for the prediction of bioconcentration and bioaccumulation is that they assume no active biotransformation. In addition, specific physico-chemical properties of the substance may reduce availability and possibly exclude uptake. Therefore, when available, measured BCF values based on the analysis of parent substance should be used rather than predicted values.

A step-wise approach is recommended to integrate bioaccumulation in an environmental risk assessment scheme for substances which are widely distributed in the environment due to wide dispersive use and effective intermedia transfer, and which potentially can be taken up by biota. Substances which are persistent, bioaccumulative and exhibit negligible metabolism will be selected by this scheme for a more detailed evaluation.

For those substances which reach a steady-state body burden within the organism during the toxicity test, direct effects of bioconcentration are included, and thus the PNEC derived from this testing is appropriate for use in risk assessment. However, for lipophilic substances which are taken up and depurated very slowly by fish, the steady-state body burden may not have been achieved during the test. Hence, environmental effects assessments should consider the "time to reach steady-state" in evaluating PNEC values for these substances. In a preliminary assessment, it is recommended to evaluate the time to reach 95% of steady-state (T95).

It is concluded that dietary uptake by aquatic organisms is significant only if the substance has low water solubility, high lipid solubility and is slowly metabolised or eliminated by the prey organism. Initially, the BCF may be estimated as described above. When the predicted BCF value is above 1,000 (corresponding to a log Kowof 4.3) a PEC/PNEC assessment for predators is made and refined as deemed necessary.

The EU Technical Guidance for environmental risk assessment requires that bioaccumulation and secondary poisoning be assessed when the substance has a log Kow of 3 or more. In this scheme, the approach does not explicitly include the dietary pathway at lower tiers and may underestimate the body burden of prey organisms for substances with higher lipophilicities. The EU method could be overconcerned with substances of little relevance for secondary poisoning, while underestimating actual exposure for substances in the log Kow range of 4.5 to 8 which are potentially the most critical for dietary exposure.

Bioconcentration and bioaccumulation models will generally overestimate the potential for bioaccumulation and therefore may trigger unjustifiable concerns about transfer of substances up the food web. Additional work is needed to provide further insights into the limitations and uncertainties of QSARs used at the screening stage in the bioaccumulation assessment. Validation and/or reformulation of these QSARs for a wider range of chemical classes is needed if these QSARs are to be used with confidence in the risk assessment process. Furthermore, additional work is needed to incorporate both knowledge and prediction of metabolic processes to account for its influence on bioaccumulation. The development of empirical QSARs for metabolism would aid in the prioritisation, assessment and regulation of apparently persistent and bioaccumulative substances.