NIVA Risk Assessment database (NIVA RAdb)
NIVA has developed a database tool that facilitates rapid and consistent Hazard and Risk Assessment of single chemicals and mixtures of these. The Database tools utilize concepts outlined by Adverse Outcome Pathways (AOPs) to compile, assemble, integrate and visualize data from different levels of biological organization. Potential outcome from different modules of the database include identification of risk drivers (most toxic chemicals), relevant toxic endpoints (i.e. mode of action), susceptible species and species sensitivity distributions for a given aquatic exposure scenario.
Use of available exposure and effect data are key to performing hazard and risk assessment of pollutants, and compiling different sources of data are often done in a case-by-case manner. Processing data is thus highly cumbersome and time-consuming, whereas the availability of data is a large source of uncertainty in resulting assessments. The NIVA Risk Assessment database (NIVA RAdb) has been developed as a module-based tool to facilitate the assembly, organization, integration, visualization and quality assurance of available exposure and effect information in order to speed up and perform consistent handling of relevant data (fig. 1).
Fig. 1. Overview of the data compilation, integration and calculations performed on exposure and effect data to perform risk assessment of single compounds and mixtures. Exposure data is provided as Measured Environmental Concentrations (MEC) and 95% percentile of the MECs (MEC95) are calculated as a “worst-case” exposure scenario in cases where sufficient data are available. Effect data is compiled as Predicted No Effect Concentrations (PNEC) and Environmental Quality Standards (EQS) or as Effect Concentrations (EC) for a given toxic effect x (ECx). The ECx values are obtained from the ECOTOX (™) database or open literature, and complemented by predicted data generated by Quantitative Structure Activity Relationship (QSAR) models when experimental data was not available. The 5% percentile of the ECs (ECx5) from single species data is calculated as a representative of the “most sensitive” species within a species-group (e.g. algae, crustaceans, fish and amphibians). The resulting exposure and effect data are used to calculate a risk quotient (RQ) based on PNEC/EQS data (RQMEC/(PNEC/EQS)) and a species group-specific RQ based on Toxicity Units (TUspecies) as a risk prediction of single compounds. Summation of the RQs is performed as a prediction of the cumulative risk of mixtures assuming principles of Concentration Addition.
The NIVA RAdb compile available experimental and predicted (computational) effect data that range from molecular and cellular responses characterizing the mode of action (MOA), typically derived from high-throughput and high-content (in vitro) bioassays, to (apical) adverse data derived from whole organism bioassays of potential regulatory relevance. These effect data are assembled within the context of Adverse Outcome Pathways (AOPs) by anchoring data to initial cellular responses referred to as molecular initiating events (MIE), to downstream key events (KE) at the cellular/organ level and finally to adverse outcomes (AO) at the individual or organism level (fig. 2).
Fig. 2. Principles of an Adverse Outcome Pathway to assemble, organize and visualize effect data from different levels of organization ranging from the Molecular Initiating Event (MIE) to the Adverse Outcome (AO). Key events (KE) represent intermediary events linking the MIE to the AO and the different levels of organization are connected by a number of key event relationships (KER) that describe the nature and weight of evidence for the individual cause-effect relationships.
The resulting multi-level assemblies of data can be used within hazard assessment to identify the MOA of one or more stressors, to link molecular responses to higher organization level effects and to identify potential stressors among large assemblies of pollutants that can give rise to a given AO. The NIVA RAdb also support risk assessment by calculating risk quotients (RQs) of single pollutants and mixtures of these on basis of exposure (typically measured or predicted environmental concentrations) and effect data (typically NOEC, ECx, PNEC or EQS values) and can identify risk drivers (most toxic chemicals), relevant toxic endpoints (i.e. MIE, KE and AO), susceptible species and species sensitivity distributions for a given exposure scenario. Recent development includes integration of non-chemical stressors such as ionizing radiation into the hazard and risk assessments procedures. Although the current version covers aquatic species, effort to expand this to terrestrial species and mammals are ongoing.