The assessment of chemical risks is increasingly challenged by insufficient toxicological data available for the large number of marketed substances. Recent advancements in artificial intelligence (AI) and machine learning (ML) offer promising avenues to bridge these data gaps through model-based predictions. However, the integration of these methodologies into risk assessment practices remains limited, largely due to issues surrounding (regulatory) trust in the model outcomes. We developed an ML model to predict the ecotoxicity of chemicals across a broad range of aquatic species. These predictions are utilized to construct Species Sensitivity Distributions (SSDs), which constitute a key tool in environmental risk assessment. Trust of risk assessors in the SSDs is essential for use in chemical risk assessment, especially for modeled SSDs. Hence, besides quantitative performance evaluation, we also performed an extensive qualitative validation encompassing aspects such as interpretability and transparency. However, to truly advance the development of ML models for risk assessment, it is crucial to foster interdisciplinary collaboration to enhance the applicability of these technologies within regulatory frameworks. Therefore, this study emphasizes the importance of regulatory acceptance when developing new ML models for SSD predictions. We highlight future opportunities for ML in SSD prediction, while also addressing the challenges of qualitative model validation. As such, this work aims to stimulate the discussion on advancing in silico methodologies beyond the current state of the art and bridging the gap between the technological efforts made in the field of AI and the regulatory needs of chemical risk assessors.

Assessment of the carcinogenic potential of chemicals is considered an important element of human health risk assessment. However, the approaches currently used for different regulatory sectors have some shortcomings. To overcome these, an alternative testing strategy, like an IATA (Integrated Approach to Testing and Assessment), for the detection of non-genotoxic carcinogens (NGTXCs) is in demand. Such an IATA should be mechanism-based and, wherever possible, consist of New Approach Methodologies (NAMs) to avoid testing in experimental animals. To explore which type of NAMs (in silico and in vitro) should be included in the first tier of an IATA for NGTXCs we performed a case study, in line with various international initiatives focusing on this need. The case study comprised a diverse set of 29 chemicals that together cover different modes of action relevant for non-genotoxic carcinogenesis. Different NAMs, including in silico tools (e.g., QSARS, ADME predictions) and high-throughput in vitro assays such as ToxCast and CALUX, were explored. The findings from the case study reveal the complementarity of the NAMs studied as well as the need for additional NAMs to be included, to ensure a broader coverage of MOAs relevant for carcinogenicity. As such, the case study nicely contributes to a more defined composition of a first tier for an IATA for NGTXC.

New approach methodologies (NAMs) are promising for refining, reducing, and replacing animal experiments for hazard characterization. Quantitative in vitro-in vivo extrapolation (qIVIVE) is essential to extrapolate an in vitro-based point of departure to an in vitro-based human equivalent dose and subsequently to an in vitro-based health-based guidance or threshold value. The use of NAMs for hazard characterization leads to the need for various new extrapolations and linked uncertainties that preferably are quantified. Currently, qIVIVE is often performed without addressing these uncertainties. A clear description and, if possible, quantification of extrapolations and uncertainties when using NAMs for risk assessment will aid the regulatory implementation of NAMs for risk assessment. A case study of a qIVIVE-based assessment on the risk of liver steatosis from dietary exposure to imazalil is reported, using a human cell line in vitro test method as an example of a NAM to replace animal experiments. We consider the uncertainties related to the extrapolations from in vitro to in vivo effects, from in vitro nominal concentrations to in vitro intracellular concentrations, from in vitro concentrations to external doses (reverse dosimetry), from in vitro exposure durations to in vivo exposure situations, and from the average human to a sensitive individual. The case study addresses these uncertainties in a mainly quantitative approach, using available data and the Monte Carlo Risk Assessment platform.

The number of chemicals being placed on the market is increasing. As such, there is an increased need for screening and evaluation of chemical hazards and risks. Particularly, chemicals with intrinsic properties that are considered of very high concern are ideally identified and regulated before wide-spread use and exposure. The use of in silico tools can help to identify potential substances of very high concern (SVHCs). Earlier, predictive models have been developed that identify potential SVHCs based on global structural similarity to known SVHCs. Here in this study, these read-across similarity models have been extended with other similarity modules, including toxicophore, biological and physicochemical similarity. The newly developed SVHC similarity profiles do individually not outperform the existing global similarity model. However, combining these new modules in an extended similarity approach results in more comprehensive predictions and allows for improved interpretability and applicability to the broader chemical universe. As such, this new approach is thought to support model users in interpretation of the model-prediction, and can thereby contribute to better screening and prioritization of potential SVHCs.

The use of bisphenol A (BPA), a substance of very high concern, is proposed to be banned in food contact materials (FCMs) in the European Union. To prevent regrettable substitution of BPA by alternatives with similar or unknown hazardous properties, it is of importance to gain the relevant toxicological information on potential BPA alternative substances and monitor them adequately. We created an inventory of over 300 substances mentioned as potential BPA alternatives in regulatory reports and scientific literature. This study presents a prioritization strategy to identify substances that may be used as an alternative to BPA in FCMs. We prioritized 20 potential BPA alternatives of which 10 are less familiar. We subsequently reviewed the available information on the 10 prioritized less familiar substances regarding hazard profiles and migration potential obtained from scientific literature and in silico screening tools to identify a possible risk of the substances. Major data gaps regarding the hazard profiles of the prioritized substances exist, although the scarce available data give some indications on the possible hazard for some of the substances (like bisphenol TMC, 4,4-dihydroxybenzophenone, and tetrachlorobisphenol A). In addition, very little is known about the actual use and exposure to these substances. More toxicological research and monitoring of these substances in FCMs are, therefore, required to avoid regrettable substitution of BPA in FCM.

Ecotoxicological safety assessment of chemicals requires toxicity data on multiple species, despite the general desire of minimizing animal testing. Predictive models, specifically machine learning (ML) methods, are one of the tools capable of solving this apparent contradiction as they allow to generalize toxicity patterns across chemicals and species. However, despite the availability of large public toxicity datasets, the data is highly sparse, complicating model development. The aim of this study is to provide insights into how ML can predict toxicity using a large but sparse dataset. We developed models to predict LC50-values, based on experimental LC50-data covering 2431 organic chemicals and 1506 aquatic species from the ECOTOX-database. Several well-known ML techniques were evaluated and a new ML model was developed, inspired by recommender systems. This new model involves a simple linear model that learns low-rank interactions between species and chemicals using factorization machines. We evaluated the predictive performances of the developed models based on two validation settings: 1) predicting unseen chemical-species pairs, and 2) predicting unseen chemicals. The results of this study show that ML models can accurately predict LC50-values in both validation settings. Moreover, we show that the novel factorization machine approach can match well-tuned, complex, ML approaches.

Background Lately scientific and societal concern has emerged about persistent (P), mobile (M) and toxic (T) chemicals. Such chemicals, like some polyfluoroalkyl acids (PFAAs), are of concern due to their high mobility and persistence in aquatic compartments which relates to long-term biotic exposure and difficult removal from drinking water. In this study, a screening approach for identification of PMT chemicals was developed and applied to 6158 diverse chemicals. Results Chemicals are given a continuous score for P, M and T potential based on the modelled indicators (low to moderate potential is a score of 0–0.33, high potential is a score of 0.33–0.5 and very high potential a score of 0.5–1). The P score was based on the estimated aquatic environmental half life and the M score on the chemical’s organic carbon/water partition coefficient ( K oc ) using respectively the BIOWIN3 and KocWIN QSAR models of EPISuite™. The T score was based on the indicators for five human health endpoints: carcinogenicity ( c ), mutaganicity ( m ), reprotoxicity ( r ), endocrine disruption (ED) and general repeated dose systemic toxicity. Structural alerts for these endpoints taken from the OECD QSAR Toolbox™ and Toxtree™ were used as indicators of potential (human) toxicity. Chemical similarity values to Substances of Very High Concern (SVHC) with c, m and/or r properties were also included. Value functions were developed to translate the presence of alerts and similarity to the existing SVHCs to values between 0 and 1. Subsequently, all values were also aggregated to an overall PMT score, again ranging from 0 to 1. Applying the approach to chemicals from the Inventory of Existing Commercial chemical Substances, which are also REACH registered, resulted in 15% of the chemicals receiving high scores (≥ 0.33) for all three (P-, M- and T-) indicators and 4% getting very high scores (≥ 0.5) for both the P- and M-indicators. Conclusions The approach confirmed the properties of chemicals classified as SVHC due to PMT properties (e.g. 1,4-dioxane), illustrating the ability of the approach to identify PMT chemicals of concern. Water regulators, drinking water suppliers and others can use this approach to identify potential PMT/vPvM chemicals that need further investigation.

Toxicological information as needed for risk assessments of chemical compounds is often sparse. Unfortunately, gathering new toxicological information experimentally often involves animal testing. Simulated alternatives, e.g., quantitative structure-activity relationship (QSAR) models, are preferred to infer the toxicity of new compounds. Aquatic toxicity data collections consist of many related tasks─each predicting the toxicity of new compounds on a given species. Since many of these tasks are inherently low-resource, i.e., involve few associated compounds, this is challenging. Meta-learning is a subfield of artificial intelligence that can lead to more accurate models by enabling the utilization of information across tasks. In our work, we benchmark various state-of-the-art meta-learning techniques for building QSAR models, focusing on knowledge sharing between species. Specifically, we employ and compare transformational machine learning, model-agnostic meta-learning, fine-tuning, and multi-task models. Our experiments show that established knowledge-sharing techniques outperform single-task approaches. We recommend the use of multi-task random forest models for aquatic toxicity modeling, which matched or exceeded the performance of other approaches and robustly produced good results in the low-resource settings we studied. This model functions on a species level, predicting toxicity for multiple species across various phyla, with flexible exposure duration and on a large chemical applicability domain.

BACKGROUND:Humans are exposed to combinations of chemicals. In cumulative risk assessment (CRA), regulatory bodies such as the European Food Safety Authority consider dose addition as a default and sufficiently conservative approach. The principle of dose addition was confirmed previously for inducing craniofacial malformations in zebrafish embryos in binary mixtures of chemicals with either similar or dissimilar modes of action (MOAs). OBJECTIVES:In this study, we explored a workflow to select and experimentally test multiple compounds as a complex mixture with each of the compounds at or below its no observed adverse effect level (NOAEL), in the same zebrafish embryo model. METHODS:Selection of candidate compounds that potentially induce craniofacial malformations was done using in silico methods-structural similarity, molecular docking, and quantitative structure-activity relationships-applied to a database of chemicals relevant for oral exposure in humans via food (EuroMix inventory, n=1,598). A final subselection was made manually to represent different regulatory fields (e.g., food additives, industrial chemicals, plant protection products), different chemical families, and different MOAs. RESULTS:A final selection of eight compounds was examined in the zebrafish embryo model, and craniofacial malformations were observed in embryos exposed to each of the compounds, thus confirming the developmental toxicity as predicted by the in silico methods. When exposed to a mixture of the eight compounds, each at its NOAEL, substantial craniofacial malformations were observed; according to a dose-response analysis, even embryos exposed to a 7-fold dilution of this mixture still exhibited a slight abnormal phenotype. The cumulative effect of the compounds in the mixture was in accordance with dose addition (added doses of the individual compounds after adjustment for relative potencies), despite different MOAs of the compounds involved. DISCUSSION:This case study of a complex mixture inducing craniofacial malformations in zebrafish embryos shows that dose addition can adequately predicted the cumulative effect of a mixture of multiple substances at low doses, irrespective of the (expected) MOA. The applied workflow may be useful as an approach for CRA in general. https://doi.org/10.1289/EHP9888.

Background Safe and clean drinking water is essential for human life. Persistent, mobile and toxic (PMT) substances and/or very persistent and very mobile (vPvM) substances are an important group of substances for which additional measures to protect water resources may be needed to avoid negative environmental and human health effects. PMT/vPvM substances do not sufficiently biodegrade in the environment, they can travel long distances with water and are toxic (those that are PMT substances) to the environment and/or human health. PMT/vPvM substance research and regulation is arguably in its infancy and in order to get in control of these substances the following (non-exhaustive list of) knowledge gaps should to be addressed: environmental occurrence; the suitability of currently available analytical methods; the effectiveness and availability of treatment technologies; the ability of regional governance and industrial stewardship to contribute to safe drinking water while supporting innovation; the ways in which policies and regulations can be used most effectively to govern these substances; and, the identification of safe and sustainable alternatives. Methods The work is the outcome of the third PMT workshop, held in March 2021, that brought together diverse scientists, regulators, NGOs, and representatives from the water sector and the chemical sector, all concerned with protecting the quality of our water resources. The online workshop was attended by over 700 people. The knowledge gaps above were discussed in the presentations given and the attendees were invited to provide their opinions about knowledge gaps related to PMT/vPvM substance research and regulation. Results Strategies to closing the knowledge, technical and practical gaps to get in control of PMT/vPvM substances can be rooted in the Chemicals Strategy for Sustainability Towards a Toxic Free Environment from the European Commission, as well as recent advances in the research and industrial stewardship. Key to closing these gaps are: (i) advancing remediation and removal strategies for PMT/vPvM substances that are already in the environment, however this is not an effective long-term strategy; (ii) clear and harmonized definitions of PMT/vPvM substances across diverse European and international legislations; (iii) ensuring wider availability of analytical methods and reference standards; (iv) addressing data gaps related to persistence, mobility and toxicity of chemical substances, particularly transformation products and those within complex substance mixtures; and (v) advancing monitoring and risk assessment tools for stewardship and regulatory compliance. The two most effective ways to get in control were identified to be source control through risk governance efforts, and enhancing market incentives for alternatives to PMT/vPvM substances by using safe and sustainable by design strategies.

Current regulatory chemical safety assessments do not acknowledge that ambient exposures are to multiple chemicals at the same time. As a result, potentially harmful exposures to unintentional mixtures may occur, leading to potential insufficient protection of the environment. The present study describes cumulative environmental risk assessment results for European fresh water ecosystems, based on the NORMAN chemical surface water monitoring database (1998-2016). It aims to characterize the magnitude of the mixture problem and the relative contribution of chemicals to the mixture risk, and evaluates how cumulative risks reduce when the acceptable risk per single chemical is fractionally lowered. Available monitoring data were curated and aggregated to 26,631 place-time combinations with at least two chemicals, of which 376 place-time combinations had at least 25 chemicals identified above the Limit of Detection. Various risk metrics were based on measured environmental concentrations (MECs). Mixture risk characterization ratio's (ΣRCRs) ≥ 1 were found for 39% of the place-time combinations, with few chemicals dominating the ΣRCR. Analyses of mixture toxic pressures, expressed as multi-substance Potentially Affected Fractions of species based on No Observed Effect Concentrations (msPAFNOEC), showed similar outcomes. Small fractional reductions of the ambient chemical concentrations give a steep increase of the percentage of sufficiently protected water bodies (i.e. ΣRCR < 1 and msPAFNOEC < 5%). Scientific and regulatory aspects of these results are discussed, especially with reference to the representativeness of the monitoring data for characterizing ambient mixtures, the robustness of the findings, and the possible regulatory implementation of the concept of a Mixture Allocation Factor (MAF) for prospective chemicals risk management. Although the monitoring data do not represent the full spectrum of ambient mixture exposures in Europe, results show the need for adapting policies to reach European Union goals for a toxic-free environment and underpin the utility and possible magnitude of a MAF.

This report evaluates the training courses delivered under the contract OC/EFSA/SCER/2017/01 - Lot 1. Within the period of January 2018 to February 2022 a total of 21 training courses were provided, eight on-site training courses in Parma at EFSA, six virtual training courses during the Covid-19 pandemic and seven eLearning courses comprising various numbers of modules. The courses covered different aspects of chemical and biological risk assessment and related tools, namely i) harmonisation of risk assessment methodologies for human health and ecological risk assessment of combined exposure to multiple chemicals (mixture assessment), ii) risk assessment of the application of nanoscience and nanotechnologies in agro/food/feed (nanotoxicity); iii) science-based criteria for identifying endocrine disruptors in the context of EU legislation on pesticides and biocides (endocrine disruption); iv) principles on genotoxicity on scientific assessment (genotoxicity) and v) computational toxicology approaches and tools (in silico). All tutors were experts in their field and had previously performed training courses on these topics. The target participants of the training courses were members of EFSA’s Scientific Committee/Panels and their working groups as well as employees from national and international regulatory agencies associated with risk assessment of feed and food compounds. Members of the EFSA Networks as well as EFSA scientific staff also participated in the training courses. Courses were evaluated based on the feedback of participants and continuously improved also by integrating updated or new EFSA guidance documents.

Next generation risk assessment is defined as a knowledge-driven system that allows for cost-efficient assessment of human health risk related to chemical exposure, without animal experimentation. One of the key features of next generation risk assessment is to facilitate prioritization of chemical substances that need a more extensive toxicological evaluation, in order to address the need to assess an increasing number of substances. In this case study focusing on chemicals in food, we explored how exposure data combined with the Threshold of Toxicological Concern (TTC) concept could be used to prioritize chemicals, both for existing substances and new substances entering the market. Using a database of existing chemicals relevant for dietary exposure we calculated exposure estimates, followed by application of the TTC concept to identify substances of higher concern. Subsequently, a selected set of these priority substances was screened for toxicological potential using high-throughput screening (HTS) approaches. Remarkably, this approach resulted in alerts for a selection of substances that are already on the market and represent relevant exposure in consumers. Taken together, the case study provides proof-of-principle for the approach taken to identify substances of concern, and this approach can therefore be considered a supportive element to a next generation risk assessment strategy.

We have studied the effect of relatively high concentrations of extracellular N-acetylaspartate (NAA) on the intracellular free calcium concentration [Ca2+](i) in NTera2-neurons, While low concentrations of extracellular NAA (0.1, 1 mM) had no effect on the [Ca2+](i), high concentrations of extracellular NAA (3, 10 mM) elicited sharp and statistically significant elevations of [Ca2+](i). Different classes of antagonists of the N-methyl-D-aspartate (NMDA) receptor abolished the NAA induced elevations of the [Ca2+](i), indicating the involvement of the NMDA receptor in NAA-induced elevations of [Ca2+](i).

Abstract In this paper we have explored the possibilities and difficulties of integrating data between ERA and HRA, in particular how toxicity in fish could inform the toxicity in other (aquatic) vertebrates and how their relative sensitivity can be used to perform the risk assessment of chemical more efficiently. With respect to acute toxicity of chemicals with a nonspecific mode of action, the sensitivity of fish seems to correlate well with that of rats. For substances interfering via a specific mode of action, extrapolation between species might be feasible provided that the molecular features of the receptors involved are known. Overall, fish species appear to be rather sensitive to chemicals interacting with the hypothalamic–pituitary–gonadal/thyroid axes. Fish could therefore be a good model to identify estrogenic and androgenic and thyroid active substances and to determine whether or not developmental and/or reproductive effects in vertebrates via these mode of actions can be expected. It is therefore recommended for hazard assessment of endocrine disruptors to include histopathology of the thyroid gland and analysis of thyroid hormones into a fish reproduction assay, to reduce testing on other aquatic vertebrates.

While the burden of disease from well-studied drinking water contaminants is declining, risks from emerging chemical and microbial contaminants arise because of social, technological, demographic and climatological developments. At present, emerging chemical and microbial drinking water contaminants are not assessed in a systematic way, but reactively and incidence based. Furthermore, they are assessed separately despite similar pollution sources. As a result, risks might be addressed ineffectively. Integrated risk assessment approaches are thus needed that elucidate the uncertainties in the risk evaluation of emerging drinking water contaminants, while considering risk assessors' values. This study therefore aimed to (1) construct an assessment hierarchy for the integrated evaluation of the potential risks from emerging chemical and microbial contaminants in drinking water and (2) develop a decision support tool, based on the agreed assessment hierarchy, to quantify (uncertain) risk scores. A multi-actor approach was used to construct the assessment hierarchy, involving chemical and microbial risk assessors, drinking water experts and members of responsible authorities. The concept of value-focused thinking was applied to guide the problem-structuring and model-building process. The development of the decision support tool was done using Decisi-o-rama, an open-source Python library. With the developed decision support tool (uncertain) risk scores can be calculated for emerging chemical and microbial drinking water contaminants, which can be used for the evidence-based prioritisation of actions on emerging chemical and microbial drinking water risks. The decision support tool improves existing prioritisation approaches as it combines uncertain indicator levels with a multi-stakeholder approach and integrated the risk assessment of chemical and microbial contaminants. By applying the concept of value-focused thinking, this study addressed difficulties in evidence-based decision-making related to emerging drinking water contaminants. Suggestions to improve the model were made to guide future research in assisting policy makers to effectively protect public health from emerging drinking water risks.

Due to the large amount of chemical substances on the market, fast and reproducible screening is essential to prioritize chemicals for further evaluation according to highest concern. We here evaluate the performance of structural similarity models that are developed to identify potential substances of very high concern (SVHC) based on structural similarity to known SVHCs. These models were developed following a systematic analysis of the performance of 112 different similarity measures for varying SVHC-subgroups. The final models consist of the best combinations of fingerprint, similarity coefficient and similarity threshold, and suggested a high predictive performance (>= 80%) on an internal dataset consisting of SVHC and non-SVHC substances. However, the application performance on an external dataset was not evaluated. Here, we evaluated the application performance of the developed similarity models with a 'pseudo-external assessment' on a set of substances (n = 60-100 for the varying SVHC-subgroups) that were putatively assessed as SVHC or non-SVHC based upon consensus scoring using expert elicitations (n = 30 experts). Expert scores were direct evaluations based on structural similarity to the most similar SVHCs according to the similarity models, and did not consider an extensive evaluation of available data. The use of expert opinions is particularly suitable as this is exactly the intended purpose of the chemical similarity models: a quick, reproducible and automated screening tool that mimics the expert judgement that is frequently applied in various screening applications. In addition, model predictions were analyzed via qualitative approaches and discussed via specific examples, to identify the model's strengths and limitations. The results indicate a good statistical performance for carcinogenic, mutagenic or reprotoxic (CMR) and endocrine disrupting (ED) substances, whereas a moderate performance was observed for (very) persistent, (very) bioaccumulative and toxic (PBT/vPvB) substances when compared to expert opinions. For the PBT/vPvB model, particularly false positive substances were identified, indicating the necessity of outcome interpretation. The developed similarity models are made available as a freely-accessible online tool. In general, the structural similarity models showed great potential for screening and prioritization purposes. The models proved to be effective in identifying groups of substances of potential concern, and could be used to identify follow-up directions for substances of potential concern.

Physiologically-based toxicokinetic (PBTK) models are important tools for in vitro to in vivo or inter-species extrapolations in health risk assessment of foodborne and non-foodborne chemicals. Here we present a generic PBTK model implemented in the EuroMix toolbox, MCRA 9 and predict internal kinetics of nine chemicals (three endocrine disrupters, three liver steatosis inducers, and three developmental toxicants), in data-rich and data-poor conditions, when increasingly complex levels of parametrization are applied. At the first stage, only QSAR models were used to determine substance-specific parameters, then some parameter values were refined by estimates from substance-specific or high-throughput in vitro experiments. At the last stage, elimination or absorption parameters were calibrated based on available in vivo kinetic data. The results illustrate that parametrization plays a capital role in the output of the PBTK model, as it can change how chemicals are prioritized based on internal concentration factors. In data-poor situations, estimates can be far from observed values. In many cases of chronic exposure, the PBTK model can be summarized by an external to internal dose factor, and interspecies concentration factors can be used to perform interspecies extrapolation. We finally discuss the implementation and use of the model in the MCRA risk assessment platform.

ADVERTISEMENT RETURN TO ISSUEPREVLetters to the Edito...Letters to the EditorNEXTPractical Application of Next Generation Risk Assessment of Chemicals for Human HealthMirjam Luijten*Mirjam LuijtenCentre for Health Protection, Centre for Safety of Substances and Products, Centre for Nutrition, Prevention and Health Services, National Institute for Public Health and the Environment, P.O. Box 1, Bilthoven 3720 BA, The Netherlands*Email: [email protected]. Phone: +31 30 2743628.More by Mirjam Luijtenhttp://orcid.org/0000-0002-5277-1443, Emiel RorijeEmiel RorijeCentre for Health Protection, Centre for Safety of Substances and Products, Centre for Nutrition, Prevention and Health Services, National Institute for Public Health and the Environment, P.O. Box 1, Bilthoven 3720 BA, The NetherlandsMore by Emiel Rorije, R. Corinne SprongR. Corinne SprongCentre for Health Protection, Centre for Safety of Substances and Products, Centre for Nutrition, Prevention and Health Services, National Institute for Public Health and the Environment, P.O. Box 1, Bilthoven 3720 BA, The NetherlandsMore by R. Corinne Sprong, and Leo T. M. van der VenLeo T. M. van der VenCentre for Health Protection, Centre for Safety of Substances and Products, Centre for Nutrition, Prevention and Health Services, National Institute for Public Health and the Environment, P.O. Box 1, Bilthoven 3720 BA, The NetherlandsMore by Leo T. M. van der Venhttp://orcid.org/0000-0002-7659-0962Cite this: Chem. Res. Toxicol. 2020, 33, 3, 693–694Publication Date (Web):March 5, 2020Publication History Published online5 March 2020Published inissue 16 March 2020https://pubs.acs.org/doi/10.1021/acs.chemrestox.0c00074https://doi.org/10.1021/acs.chemrestox.0c00074letterACS PublicationsCopyright © 2020 American Chemical Society. This publication is available under these Terms of Use. Request reuse permissions This publication is free to access through this site. Learn MoreArticle Views1337Altmetric-Citations8LEARN ABOUT THESE METRICSArticle Views are the COUNTER-compliant sum of full text article downloads since November 2008 (both PDF and HTML) across all institutions and individuals. These metrics are regularly updated to reflect usage leading up to the last few days.Citations are the number of other articles citing this article, calculated by Crossref and updated daily. Find more information about Crossref citation counts.The Altmetric Attention Score is a quantitative measure of the attention that a research article has received online. Clicking on the donut icon will load a page at altmetric.com with additional details about the score and the social media presence for the given article. Find more information on the Altmetric Attention Score and how the score is calculated. Share Add toView InAdd Full Text with ReferenceAdd Description ExportRISCitationCitation and abstractCitation and referencesMore Options Share onFacebookTwitterWechatLinked InRedditEmail PDF (686 KB) Get e-AlertscloseSUBJECTS:Environmental pollution,Molecular modeling,Risk assessment,Testing and assessment,Toxicity Get e-Alerts

Mixtures of substances to which humans are exposed may lead to cumulative exposure and health effects. To study their effects, it is first necessary to identify a cumulative assessment group (CAG) of substances for risk assessment or hazard testing. Excluding substances from consideration before there is sufficient evidence may underestimate the risk. Conversely, including everything and treating the inevitable uncertainties using conservative assumptions is inefficient and may overestimate the risk, with an unknown level of protection. An efficient, transparent strategy is described to retain a large group, quantifying the uncertainty of group membership and other uncertainties. Iterative refinement of the CAG then focuses on adding information for the substances with high probability of contributing significantly to the risk. Probabilities can be estimated using expert opinion or derived from data on substance properties. An example is presented with 100 pesticides, in which the retain step identified a single substance to target refinement. Using an updated hazard characterisation for this substance reduced the mean exposure estimate from 0.43 to 0.28 μg kg-bw-1 day-1 and reduced the 99.99th percentile exposure from 24.9 to 5.1 μg kg-bw-1 day-1. Other retained substances contributed little to the risk estimates, even after accounting for uncertainty.