Applications of toxicokinetic (TK) models on freshwater invertebrates in a regulation perspective

Toxicokinetic (TK) models have been developed to describe the bioaccumulation of chemicals in organisms. They are used as the first step to evaluate the toxicity of a contaminant in environmental risk assessment (ERA) and are developed to provide a theorical framework for understanding the link between exposure and accumulation by the biota, testing hypotheses, and make predictions (e.g. predictions of the chemical concentration in organisms according to environmental concentration or inversely). In France, polycyclic aromatic hydrocarbons (PAH) have generally been analyzed in sediment as part of annual monitoring. However, regulation specifies that for PAHs, the environmental quality standard (EQS) concerns biota, in this case invertebrates. However, modifying the monitoring protocol used for several years would lead to a loss of data continuity. In this context, TK models could be used to predict ( i) concentrations in the sediment equivalent to the EQS biota and ( ii) concentrations in biota, directly from data measured in sediment in situ, then compared to the EQS biota. Thus, the aim of this study was to illustrate how to use TK models to retro-predict chemical concentrations in the sediment leading to the EQS biota. To achieve this purpose, we firstly used experimental data of a TK study available in the literature (e.g. Hyallela azteca and Chironomustentans exposed to benzo(a)pyrene (BaP) spiked sediment) to estimate the distributions of the model parameters and thus to predict the concentration in the sediment that will lead to a concentration in the biota below the corresponding EQS biota (for both BaP and its metabolites). The results raised the issue of taking into account metabolites in regulation, where their concentrations in the organism could exceed the EQS biota defined for the parent compound. Secondly, we used several experimental data of TK studies which reported different amount of organic matter to account for the bioavailability of PAH in the model.