Additional Deliverable AD6.3 Roadmap on aggregated exposure strategy through different sources and routes related to general and occupational environments WP 6 – T6.2

ADDITIONAL DELIVERABLE P-A-R-C HORIZON-HLTH-2021-ENVHLTH-03 CONTRACT N. 101057014 Partnership for the Assessment of Risks from Chemicals Additional Deliverable AD6.3 Roadmap on aggregated exposure strategy through different sources and routes related to general and occupational environments WP 6 – T6.2 This partnership has received funding from the European Union’s Horizon Europe research and innovation programme under Grant Agreement No 101057014. ADDITIONAL DELIVERABLE AD6.3 2 Technical reference Work package WP 6 - Innovation in regulatory risk assessment Task T 6.2 – Integrative exposure and risk assessment Dissemination level1 PU Lead Beneficiary/ Responsible AE ANSES, UNISANTE Contributing Participants P: Participation in the work of PARC A6.2.1. R: Participation in the revision of this additional deliverable. Philippe Palmont / Anses / [email protected] / (P) Margaux Sanchez / Anses / [email protected] / (P) Géraldine Carne / Anses / gé[email protected] / (R, P) Nancy Hopf / Unisanté / [email protected] / (P) Hélène Hirzel / Unisanté / hel[email protected] / (P) Jordi Minnema / RIVM / [email protected] / (P) Martine Bakker / RIVM / mart[email protected] / (R, P) Bas Bokkers / RIVM / bas.bokk[email protected] / (P) Kaatje Touchant / VITO / kaatje.touchan[email protected] / (P) Mirja Van Holderbeke / VITO / mirja.vanholderb[email protected]e / (P) Arno Vanderbeke / VITO / arno[email protected]e / (P) Lisa Chedik / INRS / lisa.chedi[email protected] / (P) Anneli Julander / IVL / anneli.ju[email protected] / (P) Sarka Langer / IVL / sarka.lang[email protected] / (P) Tomas Rydberg / IVL / tomas.[email protected]e / (P) Ioannis Liagkouridis / IVL / Ioannis.Liagkourid[email protected] / (P) Paul Scheepers / RUMC / [email protected] / (P) Milija Koponen / TTL / milja.k[email protected] / (P) Sanni Uuksulainen / TTL / sanni.uuksula[email protected] / (P) Jonna Weisell / TTL / jonna.w[email protected] / (P) Mikko Poikkimäki / TTL / mikk[email protected] / (P) Tiina Santonen / TTL / tiina.santon[email protected] / (R, P) Wouter Fransman / TNO / [email protected] / (R, P) Ruby Vermoolen TNO / ruby[email protected] / (P) Hasnae Benjeddi / TNO / hasnae.benjedd[email protected] / (P) Laetitia Six / OVAM / laetitia.[email protected] / (P) Sonia Namorado / INSA / [email protected]in-saude.pt / (P) Henriqueta Louro INSA / Henriq[email protected]-saude.pt / (R, P) Vivi Schlünssen / AU / [email protected] / (P) Trine Husøy / NIPH / trine.hus[email protected] / (P) Jasper Engel / WR-BIOM [email protected] / (R, P) Mirjam Blokker / KWR / mirjam.blok[email protected]l / (P) Natalie von Götz / FOPH / [email protected] / (R, P) Joanna Jurewicz / NIOM / joan[email protected].pl / (P) Lucija Perharic / NIJZ / [email protected] / (P) Stanislava Kirincic / NIJZ / sta[email protected] / (R, P) Viviana Golja / NIJZ / viviana.g[email protected] / (P) ADDITIONAL DELIVERABLE AD6.3 3 1 PU = Public PP = Restricted to other programme participants (including the Commission Services) RE = Restricted to a group specified by the consortium (including the Commission Services) CO = Confidential, only for members of the consortium (including the Commission Services) 2comments do not necessarily represent the corporate EFSA view Justine Pincemaille / LNS / [email protected] / (R, P) Ruth Moeller / LNS / ruth.mo[email protected] / (R, P) Špela Bavec / GeoZS / [email protected] / (P) Teja Čeru / GeoZS / [email protected] / (P) Mateja Gosar / GeoZS / [email protected] / (P) Miloš Miler / GeoZS / [email protected] / (P) Martin Gabersek / GeoZS / [email protected] / (P) Sarah Habran / ISSEP / s.hab[email protected] / (R, P) Susana Viegas / ENSP-UNL / su[email protected]t / (R, P) Carla Martins / ENSP-UNL / carla.[email protected]t / (R, P) Frank Welle / IVV / frank.well[email protected] / (P) Joan Grimalt / CISC / [email protected] / (P) Agathi Charistou / BPI / a.char[email protected] / (R, P) Stuart Harrad / UOB / s.j.harr[email protected]c.uk / (R, P) Vikas Kumar / IISPV / [email protected] / (P) Deepika / IISPV / [email protected] / (P) Ana Virgolino / FMUL / avirgol[email protected]lisboa.pt / (R, P) Osvaldo Santos / FMUL / [email protected] / (R, P) Responsible author(s) Amélie Crépet / ANSES / amelie.[email protected] Clément Blassiau / ANSES / [email protected] David Vernez / UNISANTE / [email protected] Katleen De Brouwere / VITO / [email protected] Radu Duca / LNS / Radu.DU[email protected] Co-authors Hilko van der Voet / WR-BIOM / hilko.vander[email protected]l Sebastiaan L. Zoutendijk / RIVM / [email protected] Calvin Ge / TNO / [email protected] José Tarazona / ISCIII / jtaraz[email protected]s Wenjuan Wei / CSTB / [email protected]r Reviewers Peter Fantke / DTU / pef[email protected] Claudia Cascio / EFSA / [email protected] Luc Mohimont / EFSA/ [email protected] Due date of deliverable 30 April 2023 Actual submission date 26 June 2023 ADDITIONAL DELIVERABLE AD6.3 4 Document history “Funded by the European Union. Views and opinions expressed are, however, those of the author(s) only and do not necessarily reflect those of the European Union or the Health and Digital Executive Agency. Neither the European Union nor the granting authority can be held responsible for them.” Version Date Reviewer name/Institutions Short description of changes 1 03/04/2023 Initial version, see the Technical References. 2 25/04/2023 Peter Fantke, Claudia Cascio, Luc Mohimont. Initial version with comments from reviewers. 3 06/05/2023 Christophe Rousselle Previous version with Christophe Rousselle's comments. 4 14/06/2023 Co-authors Final version with modifications linked to reviewers' comments, for subsmission to the European Commission ADDITIONAL DELIVERABLE AD6.3 5 Abstract In the context of a compartmentalized view of risk assessment, the activity A6.2.1 “Aggregated exposure assessment from multiple sources and routes for general population and workers” of task T6.2 on “Integrative exposure and risk assessment” aims to advance knowledge on the combination of various exposure sources and routes related to general and occupational environments. It will help to propose a more integrative risk assessment and management crossing regulatory silos, in line with recent incentives from European agencies (EFSA, ECHA) and the European Commission. This deliverable presents the first steps of the developed strategy and roadmap to assess aggregate exposure through different living environment, sources and routes. In section 1, the scientific, societal, and regulatory questions which this work aims to answer are presented, as the proposed roadmap to develop the strategy to aggregate exposure. In section 2, the inventory of general and occupational population exposure, source-to-dose and aggregate models started on the basis of the previous work performed by the working group on exposure models of the International Society of Exposure Science (ISES), and Working Party on Exposure Assessment of the Organization of Economic Co-operation and Development (OECD) are presented. The different data sources which can be used to feed the exposure models on general and occupational environments are identified. In section 3, the strategy to assess aggregate exposure from various living environments is outlined. In the final section 4, the case studies that are proposed to start the second year to apply and feed the methodology are presented. More developments, recommendations and results will come during the next years of the project. Key Words Aggregated exposure, human exposure, multi-source exposure, multi-route exposure, risk mitigations, occupational exposure, lifetime exposure, consumer exposure, environmental exposure ADDITIONAL DELIVERABLE AD6.3 6 2.1.1. Description of the model inventory steps ........................................................................................ 21 2.1.2. Results from inventoried source-to-dose models............................................................................. 23 2.1.3. Results from inventoried exposure models for general and occupational environments ............... 24 2.1.4. Results from inventoried aggregate models (within and between general and worker populations) ................................................................................................................................................ 25 3.2.1. Methods used to perform source-to-dose modelling....................................................................... 35 3.2.2. Methods used to aggregate exposure in the general environment ................................................. 36 3.2.3. Methods used to aggregate exposure in occupational population .................................................. 39 3.3.1. Guidance to model source-to-dose .................................................................................................. 40 3.3.2. Guidance to aggregate exposure in the general environment ......................................................... 41 3.3.3. Guidance to aggregate exposure for workers .................................................................................. 57 3.4.1. Approach to perform source-to-dose modelling .............................................................................. 57 3.4.2. Approach to aggregate exposures from different sources and routes for the general environment ............................................................................................................................................... 60 Table of contents Technical References .............................................................................................................................................. 2 Document history.................................................................................................................................................... 4 Abstract ................................................................................................................................................................... 5 Key Words ............................................................................................................................................................... 5 Table of contents .................................................................................................................................................... 6 Authors and Acknowledgments .............................................................................................................................. 8 Glossary ................................................................................................................................................................... 8 1. Introduction ...................................................................................................................................................... 13 1.1. Scope of the activity A6.2.1 and the associated projects P6.2.1.a and P6.2.1.b ....................................... 14 1.2. Regulatory, scientific and societal questions ............................................................................................ 15 1.3. Roadmap to aggregate exposure from different sources and living environments and to model the migration and transfer from emission sources ................................................................................................ 19 2. Model and data inventory ................................................................................................................................. 20 2.1. Elaboration of a model inventory and first results .................................................................................... 21 2.2. Elaboration of an exposure data inventory and first results ..................................................................... 26 3. Strategy to aggregate exposures through different living environments, sources and routes ........................ 32 3.1. Model selection for the functional design of exposure model connection ............................................... 32 3.2. General presentation of methods used to aggregate exposure in general and occupational populations ....................................................................................................................................................... 34 3.3. Summary of guidance on source-to-dose models and aggregate exposure ............................................. 40 3.4. Developing an approach to aggregate exposures from different sources and routes and living environments ................................................................................................................................................... 57 ADDITIONAL DELIVERABLE AD6.3 7 3.4.3. Approach to aggregate exposures from different sources and routes for occupational population ................................................................................................................................................... 61 3.4.4. Approach to aggregate exposures from general and occupational environments .......................... 61 3.4.5. Use of HBM data to evaluate external exposure models ................................................................. 62 3.4.6. Exposure model connections as part of the PARC integrative model network ................................ 62 4.2.1. General information on the case studies .......................................................................................... 64 4.2.2. Case study construction process....................................................................................................... 65 4.2.3. Description of proposed case studies ............................................................................................... 65 4.2.4. Prioritization of case studies and partner organization in working groups ...................................... 67 4. Case studies for application of the strategy ...................................................................................................... 64 4.1. Case studies for source-to-dose modelling ............................................................................................... 64 4.2. Case studies for aggregate exposure ......................................................................................................... 64 5. Conclusion ......................................................................................................................................................... 68 6. References ......................................................................................................................................................... 69 Appendix ............................................................................................................................................................... 83 Appendix 1: Short name and extended name of the tasks/activities/projects cited in the document ........... 83 Appendix 2: Overview of the descriptors and the structure of the model inventory ...................................... 84 Appendix 3: Overview of the descriptors and the structure of the data inventory ......................................... 87 Appendix 4: Overview of the partners involved in the different scoring groups ............................................. 90 Appendix 5: Overview of the descriptors of the case studies tables................................................................ 91 ADDITIONAL DELIVERABLE AD6.3 8 Authors and Acknowledgments The authors would like to thank the reviewers for their valuable comments that highly improve the clarity of the document regarding the definition of the objectives and addressed questions as well as the wording used. They would like to also address the fact that this document is a first step to develop global exposure assessment considering all exposure sources, routes and environments. It gather scientific knowledge from scientists working in different risk assessment field (occupational risk assessment, risk assessment for general population, HBM data collection, external exposure modelling, statisticians, toxicologists, chemists, analysts, etc.). Common vocabulary and approaches are needed to be developed, first lines are presented here and will evolve with the progress of the project. Glossary To ensure consistency of wording throughout this document, and generally in activity A6.2.1, a terminology document was developed by the group members. This terminology is firstly based on that proposed by ISES Europe (Heinemeyer et al. 2022), seen by the authors as a preliminary glossary of key terms, which will be developed into a living document to be enriched with further terminology and evaluated with more international experts (Heinemeyer et al. 2022). The ISES Europe terminology is based upon that developed through previous initiatives from authoritative documents: WHO/IPCS (IPCS and IOMC 2004; WHO, IPCS 2008, 2009, 2011), OECD (OECD 2003, 2004; OECD, IOMC 2018), IUPAC (Duffus, Nordberg, and Templeton 2007), ILSI/HESI, US EPA (US EPA 2011; US EPA, ORD, and CPHEA 2003), ECHA (ECHA 2013b, 2016b, 2016e), EFSA (EFSA 2012), EEA (EEA and EC 2023), EU (European Parliament and The Council 2022), and other scientific publications (Bennett et al. 2002; Fantke et al. 2020; ISO/TS 21623 2017; Johnson 2009; Wild 2005) . Some of the definitions were adapted by the working group, such as ‘exposure sources’ and ‘exposure pathways’, which were considered insufficiently precise for the purposes of the activity. Note that this work is common to both projects: P6.2.1.a ‘source-to-dose' project and P6.2.1.b. ‘strategy for aggregate exposure’ project. This is an initial proposal for definitions that may evolve as the project progresses. Absorption barrier. [adapted from (Heinemeyer et al. 2022)]. Any exposure surface that may retard the rate of penetration of a substance into a target. Examples include the skin, respiratory tract lining, and gastrointestinal tract wall. (Single chemical) aggregate exposure. [adapted from (Heinemeyer et al. 2022), and (OECD, IOMC 2018:11)]. Exposure to a single substance from multiple sources (i.e.: water, air, dust, diet, etc.) and from multiple routes (ingestion, inhalation, dermal contact). ADDITIONAL DELIVERABLE AD6.3 9 Aggregate exposure assessment. [from PARC A6.2.1]. Exposure assessment to one or multiple substances from multiple sources (i.e.: water, air, dust, diet, etc.) and from multiple routes (ingestion, inhalation, dermal contact). This can comprise exposures from general and occupational environments. Aggregate exposure model. [from PARC A6.2.1]. Mathematical model whose purpose is to estimate exposure at a relevant target, modelling all relevant sources and routes. If the effect is local, then in general the relevant target is an outer exposure surface, and aggregation is carried out only at exposure source level. If the effect is systemic, then aggregation is also modelled at exposure routes level, and the target is a biological matrix. Aggregate oral exposure assessment. [from PARC A6.2.1]. Assessment of exposure to one or multiple substances from multiple sources (i.e.: water, air, dust, diet, etc.) via ingestion. This can comprise oral exposures from general life and occupational environments. Biological matrix, or Biological media, or targeted biological matrix. [from PARC A6.2.1]. All biological media relevant to the assessment of internal exposure (urine, blood, hair placenta, cordon, etc.). Biomonitoring. [adapted from (Heinemeyer et al. 2022)]. The measurement of one or several substances and/or their metabolites, and/or markers of subsequent health effects, in biological media such as tissues, cells or fluids. Consumer products [adapted from GEneral Multilingual Environmental Thesaurus)) and (US EPA, 2011, Exposure Factors Handbook, Chapter 17, Consumer products)]. Economic good that directly satisfies human wants or desires. This refers in particular to food, cosmetics hygiene products, household furnishings, garment conditioning products, household maintenance products, home building-improvement products, automobile-related products, and personal materials. In this report, for practical reasons, we will isolate food from consumer products, so that by abuse of language we may write 'consumer product' to refer to consumer products other than food. The definition covers substances, mixtures and articles for consumer use (as defined under REACH). Consumer product environment. [from PARC A6.2.1]. Environment related to a consumer product, such as: environment linked to contact between the skin and a cosmetic product, linked to the spraying of a biocidal product or cosmetic, or linked to human proximity to a surface impregnated with chemical substances. Cumulative exposure. [from PARC A6.2.3]. Exposure to a single exposure dose of a mixture estimated using the dose addition model to cumulate substance exposure converted using relative toxicity/potency factors (RPF, TEF). ADDITIONAL DELIVERABLE AD6.3 16 Furthermore, the EU Chemicals Strategy for Sustainability Towards a Toxic-Free Environment from the European Commission provides incentives for harmonization and policy coordination on the level of sectorial legislations, which creates the need to further develop modelling approaches (EU Commission and Parliament 2020). Based on these considerations, the ISES Europe working group on exposure models identified exposure modelling as one of the priority areas to an overarching European Exposure Science Strategy (Schlüter et al. 2022). In this context, the ISES Europe working group on exposure models positions itself as a facilitator to make progress in this area (Schlüter et al. 2022). ISES Europe, and in particular the working group on exposure models identifies the PARC program, and subsequently the activity A6.2.1, as an opportunity to make progress in the field of exposure science at European level, both through the development of new harmonised aggregate realistic modelling methods, organised within a European tool network, but also through the development of a data hub according to the FAIR principles, recognized as a key element in developing this type of approach (Schlüter et al. 2022). This refers to the links that will be established between the present activity A6.2.1. and the PARC T4.1. and T4.2, T7.3. concerning data gaps and their organization; the other activity of T6.2: A6.2.2., A.6.2.3. & A6.2.4. for the establishment of methodological collaboration; and the T8.3 for the development of the PARC network model (see Appendix 1). More recently, EFSA prioritized the need for further developing the methodology and launched a project (ExpoAdvance) for developing a roadmap for future actions in order to achieve a harmonized crosscutting methodology and regulatory guidance for aggregate exposure assessments covering all relevant sources and routes of exposure to chemicals, in close cooperation with European and international partners (EFSA, Cascio, et al. 2022). This priority is under implementation, following an open call launched by EFSA for outsourcing this activity, and is expected to continue until 2030 (EFSA, Cascio, et al. 2022). Synergies with PARC and in particular with the activity A6.2.1 will be developed. More specifically the activity A6.2.1 will develop methods, tools and data to answer the following questions. Aggregate exposure from multiple sources and routes within one living environment (either general or occupational environment) With the development of chemical risk assessment, plenty of models and tools have been proposed to perform risk assessment (Schlüter et al. 2022). However, in the general environment setting, most of them consider only one exposure source related to one type of living environment (general / occupational) (Schlüter et al. 2022). Some modelling approaches consider several sources, pathways or routes, of which a few propose aggregation strategies, usually on a limited number of sources, pathways and routes (Schlüter et al. 2022). In the occupational setting, there is a frequent segregation of exposure models dealing with different routes (e.g., inhalation, dermal), that is rather technical or contextual in nature. Some routes are assumed a priori, sometimes incorrectly, for certain pollutants or exposure circumstances. Inhalation and dermal are often considered as only exposure routes in occupational exposure but there is growing body of evidence showing the relevant contribution of inadvertent ingestion to the total combined exposure in occupational settings (see e.g. (Cherrie et al. 2006; Gorman Ng et al. 2012)). There are currently limited possibilities for the modelling of inadvertent ingestion exposure in occupational context. Moreover, the exposure determinants to be considered often differ, which leads to the construction of structurally different models. ADDITIONAL DELIVERABLE AD6.3 17 To assess aggregate exposure from multiple sources within one living environment, one way to build on current knowledge is to connect the most relevant exposure models and their output. Thus, there is a need to develop a strategy to inventory, select and connect the available exposure models. A structural analysis of the models is required to describe how the variability and the uncertainty are addressed and to identify model connection points. This work, supported by the project P6.2.1.b, aims to provide tools to move away from a compartmentalized view of risk assessment, to facilitate comparisons between entry routes and exposure situations and, ultimately, to prioritize areas for action and prevention. Aggregate exposure integrating both general and occupational environments Most chemicals can be present in both general and occupational environments. For some of them, such as PFAS, heavy metals, pesticides, phthalates, considering exposures from both types of living environment, can make the difference regarding the relative significance of exposure routes, exposure pattern (e.g., chronic vs acute exposure) and consequently, the risk. However, today, due to in particular the specificities of occupational situations and separate regulation, exposure and risk assessment in general and occupational environments are performed independently. Thus, because of this historical separation, methods, models, tools, datasets, goals and practices differ substantially between the two types of exposure. However, the respective assessment processes could converge to common methods, for instance when we come at the absorption point by oral, inhalation of dermal exposure. Moreover, with the development of biomonitoring data that automatically reflects aggregated exposure, there is an increase need to identify the relevant exposure sources, whether the exposure comes from the general or the occupational environment. Finally, combining exposures from both types of living environment is an important challenge that will produce a better understanding of the relative contributions of each source to support more effective risk management measures. This work on the development of models and tools combining general and occupational environments is part of the P6.2.1.b project. Source-to-dose modelling Exposure in the context of the general environment can take place in the outdoor environment, in the indoor environments (dwellings, public indoor spaces, vehicle interior, etc.), and several consumer product environments (contact between skin and cosmetic product, spray diffusion area, proximity to a surface impregnated with chemical substances, etc.). Chemicals produced and used by industry, and having applications in our everyday life products end up (unintentionally) in the indoor environment, outdoor environment, and several consumer product environments. Chemicals present in our daily living environments have on the left side of the exposure chain an emission source (e.g.: a ‘source’), or release factor, and at the right side end an exposure source (e.g.: a ‘concentration’ or ‘dose’) (see the area framed by a blue square in Figure 1). Chemicals follow complex pathways from emission sources to exposure sources, in the general environment. These pathways consist of release, dispersion, transfer and/or redistribution from one to another ‘intermediate’ environmental compartment (soil, water, dust, air, crops, animals). Some of these compartments are likely to be in contact with humans and are therefore sources of exposure, to which humans are exposed through the various possible routes of exposure (i.e. inhalation exposure, oral exposure due to dietary intake, ingestion of soil and dust particles; dermal exposure due to skin contact with e.g. soil, water and dust). Aggregation of the ADDITIONAL DELIVERABLE AD6.3 18 different exposure routes, and exposure sources present in the outdoor and indoor environment is important in order to reflect appropriately the human exposure. Release, dispersion, transfer and redistribution are complex processes depending on intrinsic chemical properties, as well on the properties of the indoor, outdoor and other environmental settings. The chemicals can even react with chemicals already present in this environment and generate new chemicals, normally called secondary pollutants. This work, supported by the project P6.2.1.a, aims to produce appropriate methods, guidance and tools to assess the exposure arising from indoor and outdoor sources, thereby accounting for and parameterizing all relevant exposure pathways. This will feed project P6.2.1.b which will further integrate the environmental (indoor and outdoor) sources in the aggregation accounting for the overall assessment of general and occupational exposures (see figure 1). Output for risk assessors and managers The activity A6.2.1 and its associated projects will provide new knowledge such as:  Exposure models, aggregate models and data inventories.  Methods and tools for aggregation and combination of multiple living environments (general and occupational), sources and routes of exposure with a comparison with HBM data.  Developments of source-to-dose models for indoor environments including consumer products emissions and outdoor environment.  Identification of data and model gaps and needs.  Prioritisation of main sources and routes of exposures.  Aggregate exposure results for prioritized case studies.  Identification and quantification of associated uncertainties.  Interpretation of HBM data in terms of sources and routes of exposure. Prioritized chemical families The methodological developments of the activity A6.2.1 will be applied to case studies (section 4). The PARC strategy prioritised chemical families in considering pre-existing knowledge on exposure, hazardous properties including epidemiological data for some of the compounds and concerns on cumulative risks. The identification of the first set of priorities for PARC started before the launch of the Partnership with the legacy from HBM4EU and with a survey sent to the interim governing board members. In order to have a clearer view on the prioritisation strategy within the different WPs, a survey dedicated to the prioritisation of substances and meetings with WPLs and task leaders to discuss the prioritisation of methods took place after the beginning of PARC. The approach and results are overviewed in PARC deliverable D2.1 Prioritisation criteria report WP2. This process led to a prioritised list of 10 chemical families by PARC stakeholders (academia, HBM4EU workshop, European Commission, Ministries of several Member States). The PARC T6.2 task leaders on integrative risk assessment refined these prioritised chemicals groups with the EFSA and the European Commission to ensure synergies with results reported in EFSA opinions or ongoing discussions within working groups of DG SANTE. From these discussions and other prioritisation exercises from PARC (e.g. T4.1 - Human biomonitoring), PFAS, heavy metals, pesticides, mycotoxins and plasticizers were prioritised in T6.2. They will be studied in priority in activity A6.2.1 as well as other chemical families in considering other ADDITIONAL DELIVERABLE AD6.3 19 priority criteria such as the interest in combining the general and occupational living environments and the level of risk concern from previous projects such as HBM4EU (section 4.24). 1.3. Roadmap to aggregate exposure from different sources and living environments and to model the migration and transfer from emission sources To achieve the objectives of the two projects P6.2.1.a and P6.2.1.b, a roadmap of the different steps to be conducted in the activity A6.2.1 was designed. Four parts were defined. The part A “Inventory and review” aims to develop an inventory of the available exposure and aggregate models, toolboxes, guidance and data to be subjected to detailed review. The step B “Selection and connection” consists of defining appropriate criteria to select the relevant models that will be candidates for connection in the PARC toolbox, the selection of relevant data for the case studies and the identification of recommendations from guidance to design the innovative approach to perform aggregate exposure in the part C “Strategy”. The last part D is the “Application to case studies” of the aggregate approach to relevant case studies selected from the defined criteria and PARC priorities. Connection with other tasks of PARC such as T4.1, T4.2 on collection of human biomonitoring and environmental monitoring data, T7.3 on uncertainty and T8.3 on model integration will be done (see Appendix 1). The transversality between T6.2 projects will be also ensured as output on aggregate exposures from this subtask can be used as input to A6.2.3 Real-life mixtures project and A6.2.4 on case studies for health impact assessment (see Appendix 1). The PBTK models developed in A6.2.2 will be used to compare modelled aggregated systemic exposure from several exposure sources and routes from this activity with HBM data. Figure 2. Roadmap proposed in A6.2.1 to aggregate exposure from different sources and living environments and to model migration and transfer from emission sources. Partners involved in the two projects were divided into working groups. Project P6.2.1.a “Source-to-dose” was divided into two working groups: WG Indoor environment to model chemical migration from consumer products, articles and building materials including all relevant routes. WG Outdoor environment to model the transfer and exposure of chemicals from the ADDITIONAL DELIVERABLE AD6.3 20 environment to human exposure, accounting for all relevant exposure routes (inhalation, consumption of agricultural food products, soil and dust ingestion). Project P6.2.1.b “Aggregate exposure” was divided into four working groups: WG1. Exposure models and data Inventories, WG2. Functional design for model connection, WG3. Strategy for aggregate exposure, and WG4. Case studies. All working groups have started during the year 1, except WG2 waiting for output from WG1. Table 1. Partners in each working group of project P6.2.1.a. Working group Indoor environment LNS + ISPV, RIVM, IVV, IVL, NIJZ, ANSES, UOB, CSTB, Unisanté Working group Outdoor environment VITO + ANSES, IVL, FMUL, NIJZ, GeoZS, KWR, ISVP, UOB Table 2. Partners in each working group of project P6.2.1.b.. Working group 1: Model inventory and data inventory Inventory of existing models and tools for general environment ANSES, +: INSA, ISCIII, RIVM, VITO, NIOM, NIJZ, INSA, UU, IVL, MU, FOPH, ENSP-UNL, CSTB, CSIC, FMUL, BPI, IISPV. Inventory of existing models and tools for occupational environment ANSES and Unisanté, +: INSA, ISCIII, Unisanté, NIOM, TTL, AU, IVL, LNS, ENSP-UNL, STAMI, ANSES, INRS, UU and BPI. Data inventory and gap analysis ANSES, +: INSA, ISCIII, Unisanté, NIOM, TTL, AU, IVL, LNS, ENSPUNL, UU, ANSES, RIVM, VITO, NIOM, NIJZ, INSA, UU, IVL, MU, FOPH, ENSP-UNL, CSTB, CSIC, FMUL, AU, GeoZS, BPI, UOB. Scoring methodology Unisanté, +: all partners. Scoring activity Unisanté, ANSES, +: all partners. Working group 2: Functional design for models' combination (WR-BIOM) Structural analysis and development of a functional design for model connection WR-BIOM, +: Unisanté, NIOM, TTL, AU, IVL, LNS, INSA, ENSPUNL, UU, ANSES, RIVM, VITO, NIOM, NIJZ, INSA, IVL, MU, FOPH, ENSP-UNL, CSTB, CSIC and FMUL. Working group 3: Strategy for aggregate exposure Inventory of aggregate exposure models ANSES, Unisanté, +: NIOM, RIVM, TTL, TNO, INSA, AU, UU, IVL, MU, VITO, INRS, NIJZ, LNS, FOPH, STAMI, GeoZS, FMUL, EFSA, CSTB, ENSP-UNL and BPI. Review of regulatory guidances ANSES, Unisanté, TNO, TTL, + all partners. Working group 4: case studies Propose and implement case studies All partners. 2. Model and data inventory This chapter concerns the inventory of data sources, tools and models commonly used for general and occupational population chemical exposure assessment, in the academic, regulatory and commercial fields, for Europe and the rest of the world. Several models and tools have been developed for the past thirty years for chemical exposure assessment, for general and occupational environments (Schlüter ADDITIONAL DELIVERABLE AD6.3 21 et al. 2022). However, most of them consider only one source of exposure, or consider several sources, but without developing an aggregation strategy, and are dedicated for a single environment. To consider chemical exposure from multiple sources and living environments, one way is to aggregate the output exposures from models with one or a limited number of sources. From the inventory of these models, a subset of relevant models will be selected, according to a methodology specified in section 3.1. Then, a second selection stage, allowing a deepening of the retained models will be conducted. Thereafter, a structural analysis of the selected models will be performed to allow identifying model connection points into a functional design. Finally, these models will be connected to implement composite model(s) for chemical exposure assessment, considering multiple sources and routes, and associating general and occupational environments. Model connection will be carried out in close collaboration with T8.3 on model integration in the PARC Toolbox. Furthermore, an inventory of potentially useful input data for these models and selected case studies (section 4) is also carried out in parallel. This inventory should identify available and missing data and thus, on the one hand, highlight current data needs that could be filled thanks to the synergies developed with T4.1 'Human biomonitoring' and T4.2 'Environmental and multisource monitoring', and on the other hand, specify the modelling opportunities. 2.1. Elaboration of a model inventory and first results An inventory of models and associated tools commonly used for chemical exposure assessment for general and occupational populations was compiled. This inventory activity is common to both projects: P6.2.1.a ‘source-to-dose' project and P6.2.1.b. ‘strategy for aggregate exposure’ project. An inventory session was organised and divided in four steps. The first step was the setting up of the inventory structure. The second step was compilation of the inventory using available information from previous inventories initiated by the International Society of Exposure Science (ISES) and the Organization of Economic Co-operation and Development (OECD). The third step was the completion of the inventory with the missing models and the verification and completion of the information collected on the previous models. The fourth step was the addition of possible missing descriptors to the inventory structure, in line with the scoring methodology developed in parallel. Simultaneously, as part of the working group WG3. Strategy for aggregate exposure, an inventory containing models that explicitly propose a strategy of aggregation of exposure sources was carried out, and progressively fed by the models identified in the first inventory meeting these criteria. This inventory contains some additional descriptors, aiming at a more in-depth description of the modelling method. Steps are described below. 2.1.1. Description of the model inventory steps Step 1: Structure of models and tools inventory. The inventory of exposure models was divided in three sub-inventories on: 1/ general environment exposure including source-to-dose models with human exposure as outputs, 2/ occupational exposure, and 3/ source-to-dose models that do not have human exposure as an output, but concentration of substance(s) in human exposure source(s). (see Appendix 2 for a description of the inventory). The last sometimes referred to as 'environmental models’ (Schlüter et al. 2022). Keeping in mind that the separation between the three domains is not always very clear, especially for platforms offering toolkits, composed of several models or algorithms. ADDITIONAL DELIVERABLE AD6.3 22 The models with an explicit aggregate strategy were extracted and collected in separate two subinventories: one for occupational chemical exposure assessment models, and the other for general environment chemical exposure assessment models. The models' inventories were structured in five excel files containing selected descriptors (Tables S2, Figure S1, and Figure S2, Appendix 2) to provide a general overview of the models in terms of internal characteristics (what is the purpose of the model?), in terms of operational capacity (how can the model be used?), and in terms of external evaluation (how is the model positioned in relation to other models in its category and to the data?). Of these descriptors, the majority are described in all inventories and the rest are specific to the different inventories. Step 2: Filling in of the models and tools inventory with available information from previous inventory work. The sub-inventories for general environment, occupational population, and environmental models and tools were started using inventories carried out by the ISES Europe working group on exposure models (Schlüter et al. 2022), the OECD Working Party on Exposure Assessment (OECD, Environment Directorate chemicals and biotechnology committee, and Working Party on Exposure Assessment 2022), and Concawe workshop on PAH integrated exposure modelling (Prepared for the Concawe Health Management Group’s Special Task Force on Exposure Assessment (STF-29) 2017). Collaboration with the members of these working groups has been established in order to make the best use of the information available. The ISES Europe working group on exposure models (Schlüter et al. 2022) described the strategy in term of exposure modelling in Europe, as part of the European Exposure Science Strategy developed by the European Chapter of the International Society of Exposure Science (ISES Europe). This article provides an inventory, focuses on exposure models and tools used in Europe, but also presents models used outside Europe. It consists of 48 models and tools for general environment exposure, 37 models and tools for ‘occupational exposure’, and 10 models and tools for ‘environmental exposure’. Note that a number of these ‘environmental exposure’ models were not included in our inventory because some of the modelled sources of exposure were not immediate sources of exposure to humans, and therefore fall outside the scope of A6.2.1. Moreover, the ISES Europe inventory builds on previous compilations made by different regulatory bodies (ECHA 2012, 2015; SCCS 2021) and projects such as the OMEGA-NET project (Peters et al. 2022). Moreover, an inventory work was conducted by the Working Party on Exposure assessment (WPEA) of the Organisation for Economic Co-operation and Development (OECD). OECD members were invited to complete a detailed questionnaire on human and environmental exposure assessment models and tools used in the regulatory frameworks of their country. Responses to this survey were then compiled into a table. This work is an update of a previous survey conducted a decade ago (OECD and IOMC 2012). This inventory contains 120 entries. Some of which correspond to the same models but described by different institutions. All three types of models and tools are represented. Finally, the Concawe report (Prepared for the Concawe Health Management Group’s Special Task Force on Exposure Assessment (STF-29) 2017) summarizes the discussions held during the ‘Polycyclic Aromatic Hydrocarbons (PAHs) integrated exposure modelling’ 2-day workshop on 8-9th October 2015 at Concawe in Brussels organized by the Flemish Institute for technological research (VITO) together with Concawe, an association created by a small group of leading oil companies to carry out research on environmental issues relevant to the oil industry. Concawe aimed to address the challenge of assessing the contribution of Petroleum Substances to aggregated PAH exposure. In this context, there was a need to identify integrated multi-source, multi-route exposure model(s) and tool(s) suitable for ADDITIONAL DELIVERABLE AD6.3 23 characterising exposure to PAHs including those that may derive from Petroleum Substances. The focus of the workshop was general environment exposure assessment models and tools, including general environment models and environmental models, provided that the modelled sources of exposure were immediate sources of exposure to humans. Occupational exposure was out of the scope of this workshop. The available inventories contain a total of 29 models. Step 3: Verification of the collected information in the inventory, and completion with missing models and tools. The OECD inventory reports on the responses from each country that participated to the survey. There were therefore models and tools that appear several times. A first step was therefore to merge the lines associated with the same models. Some models also appeared in several of the base inventories. We therefore eliminated the supernumerary models. In some places, the information collected by OECD and ISES was partial, and sometimes not up to date. We therefore checked the information and completed as much as possible where information was missing. This completion work was first limited to the most important descriptors to have a general overview of the model: the owner of the model, several valid URL links for each model, a description of the model, the exposure sources involved, the exposure routes involved, the type of substance considered, the latest version of the model. In parallel, the partners were invited to complete and update the inventory with the models they know. The inventory was also gradually completed with the models targeted in P6.2.1.a., through the scoping documents and the bibliography made in this project. Simultaneously, the models explicitly proposing a strategy of aggregation of sources of exposure were selected and put in specific inventory excel files. Step 4: Final preparation of the inventories for the scoring activity. In this step missing descriptors were first added to the inventory structure, in line with the scoring methodology developed in parallel (see section 3.1.). A quick review of all the models was carried out again, and these descriptors were completed where information was readily available in the literature. As described in section 3.1, the model evaluation grid contains nine criteria. These criteria are all described through the descriptors in the model inventory, but sometimes through several descriptors at the same time. To facilitate the reporting of the scoring work, nine new columns corresponding to the descriptors of the scoring methodology, highlighted by a colour code, have been included in the excel files. These columns will receive the consensus opinions of the reviewers, for each model scored. Results are presented for source to dose models including 6.2.1.a project model description, general and occupational environments models and aggregate models. 2.1.2. Results from inventoried source-to-dose models Outdoor source-to-dose models In total, 44 source-to doses models and tools with human exposure as outputs and 66 source to dose models with concentrations in exposure sources as outputs were inventoried (see Appendix 2 for a description of the inventory). For a selection of outdoor source-to-dose models (18 models), the inventory was extended by a more in depth-description of the models (see section 3.4.1.). The selection of models was based on experiences with models among partners in the working group ‘outdoor environment’ of project ADDITIONAL DELIVERABLE AD6.3 24 6.2.1.a, and was extended by two review papers on models addressing human exposure to soil contaminants and industrial contaminated sites (Hoek et al. 2018; Swartjes 2007). For the selected models an overview was generated for the following aspects: 1) a description of the exposure pathways accounted for in the model, 2) the chemicals for which the model has been parameterized, 3) whether the model is dynamic or steady state, 4) the exposure factors included in the model, 5) land use scenarios considered, 6) type of model input required, 7) model for local, sitespecific versus generic context, 8) accounting for background exposure (not related to local setting), 9) case studies performed based on the model, and 10) opportunities and points of attention for use in case studies. Indoor and consumer products source-to-dose models This inventory corresponds to source-to-dose models with human exposure as output and source-todose models with concentration in exposure sources as output (see Appendix 2 for a description of the inventory). In total, 62 models and tools were inventoried related to the indoor environment. As previously stated, the exercise of identifying the available models has been done jointly with project 6.2.1b, since the ultimate overall aim of the activity is to arrive to a global aggregation from the source to the most suitable cumulative matrix and further to human exposure. These models will feed into a selection process for project 6.2.1.a. which has not yet begun. The modelling and accumulation from different emission and exposure sources (like building material emissions, consumer products utilisation, heating and overall human activities) into indoor dust as a promising matrix for (1) estimating the partitioning among the gas and adsorbed/absorbed phases and (2) evaluating the indoor aggregated exposure via different pathways will be studied. The integrative nature of settled dust represents simultaneously indoor exposure and human activity. Indoor settled dust is considered a promising matrix for studying SVOC indoor exposure and is particular relevant in terms of the need to develop the (aggregated) exposure estimation approach. However, VOC represent a major indoor air quality parameter and the aggregation of general and occupational environment exposures may be given more in depths attention in later stage of the project(s). The result of this exercise will be the subject of further publication in scientific journals and will be included in the future deliverables of the 6.2.1. Furthermore, the generated databases and aggregation results could be included in specific repositories as defined by PARC. 2.1.3. Results from inventoried exposure models for general and occupational environments General environment exposure models In total, 74 models and tools belonging to this category were inventoried. It was noted that 19 of them do not appear to have a ‘source-to-dose' module (see Appendix 2 for a description of the inventory). The remaining 55 models and tools clearly had a source-to-dose module, where the fate of the substance from the emission source to the human exposure source is modelled. Among the models and tools, 6 were clearly tools: Chesar, EASY-TRA, EPA ExpoBox, K-CHESAR, MCRA, and Merlin-Expo tool. 57 models out of 74 model several sources of human exposure, but it was not always clear, given the level of analysis considered for this stage of the process whether human exposure to these exposure sources is aggregated by exposure route, or if the results were given independently from one exposure source to another. Among the 74 models, 45 considered the dietary ingestion route, 44 considered non dietary ingestion routes, 51 considered the dermal route, and 59 considered the inhalation route. This inventory included the SHEDS-Residential version 4 (US EPA, Glen, et al. 2012) ADDITIONAL DELIVERABLE AD6.3 25 and SHEDS-Dietary version 1 (US EPA, Xue, et al. 2012) models, developed by US EPA, which are multisource and multi-route models estimating human exposure to chemicals in the residential environment and through food, respectively. The association of the two modules has been carried out through a case study. These models propose aggregation strategies that will be relevant to study in greater detail in order to enrich the general reflection. The inventory also included INTEGRA, developed by CERTH in the framework of a CEFIC LRI project B 11 (Sarigiannis et al. 2014). The model included a module to predict indoor air and indoor dust concentrations starting from emissions from materials. The environmental exposure model is based on the EUSES model, but with a higher flexibility compared to EUSES since additional food categories can be added, and monitoring data might be inserted in order to overrule the default prediction levels. Personal exposure predictions can consider time-activity patterns. Finally, the external exposure model is extended with a PBPK model for prediction of levels of contaminants in several internal human organs and tissues; hence, the model covers the full chain from emission sources to internal human exposure. Occupational exposure models. In total, 63 models and tools were inventoried. 26 models out of 76, partly or totally concern the dermal route, and 27 out of 76 partly or totally concern the inhalation route (see Appendix 2 for a description of the inventory). With the exception of very specific situations (e.g. spraying of plant protection products), exposure by ingestion is generally not considered in occupational models. None of the most commonly used occupational exposure models consider inadvertent ingestion at workplaces (Gorman Ng et al. 2012). In general, note that the recent report (Schlüter et al. 2022) highlighted the lack of models related to objects and solid materials as emission sources. The models which propose this type of modelling are currently ConsExpo, CEM, and DustEx. However, as noted in (Schlüter et al. 2022) large gaps exist in the availability of appropriate input data for these models, e.g., specifically for diffusion and partition coefficients. Those depend on both the substance and the material and may vary by multiple orders of magnitude between different substance/material combinations (Schlüter et al. 2022). Some approaches were proposed to predict these critical parameters (Schlüter et al. 2022). Despite their usefulness, these approaches cover only a limited portion of the possible substance/material combinations and therefore need to be complemented by modelling approaches for other applicability domains (e.g., non-food polymers, inorganics and polymer additives) (Schlüter et al. 2022). 2.1.4. Results from inventoried aggregate models (within and between general and occupational environments) General environment aggregate exposure models In total, 5 models have been targeted as explicitly proposing a strategy of aggregation of exposure sources: RSExpo, CARES-NG, CEM, MCRA, and IEUBK (see Appendix 2 for a description of the inventory). In addition, from the exposure model inventory (section 2.1.2), 52 models proposed to model several sources of human exposure, but it was not always clear, given the level of reading undertaken at this stage, whether human exposure to these sources is aggregated, or if the results were given independently from one source to another. We will gradually investigate in more detail the strategies developed in these models. ADDITIONAL DELIVERABLE AD6.3 32 3. Strategy to aggregate exposures through different living environments, sources and routes 3.1. Model selection for the functional design of exposure model connection Model selection One of the motivations for the model inventory described in section 2 is the connection of the relevant models for chemical human exposure assessment to consider multiple sources and routes, and to associate general and occupational environments (Figure 5). The selected models and the process of connection will be described in a functional design document. Figure 5. Selection strategy for the connection of exposure models For this purpose, only relevant models will be selected based on defined criteria for each of the model categories. Note that the exact model categories to consider are not yet fully determined. For the time being, we defined 4 categories of general environment exposure models, one for each exposure route (inhalation, dermal, ingestion), and an additional category for models that incorporate more than one exposure route. For occupational exposure models, which are mostly focused on inhalation and dermal exposure routes, only two categories were defined. Finally, for models without human exposure as an output, we may consider grouping models by sources only. In order to reduce the number of models to be studied further, on which a functional and structural analysis will be performed, we have developed criteria for scoring the models. This scoring step allows us to identify the relevant models "eligible" for a functional analysis. The number of models to be retained in each category at this stage still needs to be discussed. It is likely that a second selection will be necessary to determine which of the eligible models will be connected to perform aggregate exposure assessment. There are two reasons for this: (1) functional analysis will be necessary to determine which models will be the most suitable for aggregation (practical limitations of structure accounting or algorithm access may arise), and (2) there is no point in multiplying the tools when we are dealing with models that fulfill the same functions/perimeters. We will evaluate this after obtaining the results of the first selection. ADDITIONAL DELIVERABLE AD6.3 33 The selection methodology is the result of a collective work including all partners of the activity A6.2.1. The general principle of this methodology is described below and in Figure 6. A scoring grid includes 9 criteria grouped into three categories: methodology, scope and efficiency, and usability of the model. The grid was agreed upon between the participants before the scoring exercise. The principle is to evaluate each model by means of a blind review. Two reviewers score the model independently before comparing their evaluations. Minor discrepancies are resolved by consensus, if a major discrepancy appears, a third reviewer is asked to reach a decision. In case of doubt, the final decision rests with the 3rd reviewer. Figure 6. Scoring strategy The organisation of the scoring of the models was as follows (see also Table S3, Annex 4). About twenty people volunteered, including the 2 projects of A6.2.1. Seven scoring groups were formed (see Appendix 4), each consisting of evaluators with similar affinities, distinguishing several model categories: 1. General environment dietary ingestion exposure models mainly and general environment exposure models including multiple routes. 2. General environment inhalation models and general environment exposure models including multiple routes. 3. General environment dermal and non-dietary ingestion exposure models mainly, and general environment models including multiple routes. 4. Occupational inhalation exposure models mainly, and occupational models including multiple routes. 5. Occupational dermal and non-dietary exposure models mainly, and occupational models including multiple routes. 6. Indoor air and dust source-to-dose models, without humans as output. 7. Outdoor, food and consumer product source-to-dose models, without humans as output. ADDITIONAL DELIVERABLE AD6.3 34 Although this division of work induces some between-assessor variability, it was deemed necessary because of the large number of models to be evaluated and the different types of models, which require some experience in their respective fields. To reduce between-assessor variability, a session was organised between the scoring volunteers. Two examples of scoring were discussed in groups, one occupational model and one general environment model. Functional design A general functional design will be made for connecting the selected models and tools in the PARC integrative model network for the purpose of aggregate exposure modelling and (downstream) risk assessment. The model network is being established in a bottom-up, top-down approach [ref: deliverable D8.3]. At a high level, a general framework and guidelines are created for establishing connections between models. This may include recommendations for use of a common (overarching) conceptual and semantic data model (WP7, T8.3), specific data formats (e.g. HBM harmonized data format), and guidelines/recommendations for interfacing and establishing technical connections. At a lower level, bilateral connections between models will be established considering the general framework, but based on the practical possibilities (e.g., driven by development capacity) of the modelling tools between which connections are needed. There will be a focus on harmonization and interoperability. The general functional design will describe (1) the general process for linking models at the conceptual level, (2) the general technical process for connecting models, and (3) user functionality in the PARC model network. For each 6.2.1 case study, these three elements may be detailed further in a specific functional design, in close collaboration between the case study and T8.3 teams. A specific functional design will: 1. Define and describe the conceptual link between multiple models and the intended application or purpose of connecting them. 2. Describe the data requirements for the case study, including needs for data security and protection. 3. Describe the technical links between models, given that the conceptual linking (step 1) is clear and desirable. This description includes a mapping of data formats between tools and the possibilities for automated model interaction (e.g., via a Web API link or via a command line interface). Due to the large heterogeneity between models and their software implementation it is expected that technical linking will often require tailored solutions for each pair of models. 4. Describe the targeted users and user functionality such as the user interface (existing like MCRA or newly developed), data input/output functionality, presentation of the linked model outcomes, and error handling functionality. 3.2. General presentation of methods used to aggregate exposure in general and occupational environments This section presents a non-exhaustive list of methods and tools used by the partners of the 6.2.1.a and 6.2.1.b projects to aggregate exposure from different sources, routes and living environments. ADDITIONAL DELIVERABLE AD6.3 35 3.2.1. Methods used to perform source-to-dose modelling Partners presented during internal PARC 6.2.1.a project meetings and in a written internal report, their methods concerning human exposure to several case studies on chemicals present in the indoor/outdoor environment (general environment), where aggregation of exposure was performed. For example, VITO presented methods for metals and PFAS (De Brouwere et al. 2012; Fierens et al. 2016; Van Holderbeke et al. 2016). ANSES presented methods for Cd in a mass balance model including the transfer from soil to plant (Carne et al. 2021); NIJZ presented their use of the IEUBK model in case study on Pb contamination in Upper Meža Valley. IVL presented the method of USEtox adapted to PFAS (Holmquist et al. 2020). CSTB presented models for indoor plasticisers’ emission and transport prediction (Wei, Ramalho, and Mandin 2019). In these case studies, firstly, an overview was given of the first part of the source-to-dose modelling chain, i.e. the transfer across the environment. Depending on case studies following processes were accounted for:  Transfer from air to soil (deposition)  Transfer from air to crops (deposition)  Loss process on soil: run off, leaching, degradation  Partitioning within soil phases (pore water, solid, air) and impact on transfer to plants and loss processes  Transfer from soil to crops: plant uptake by roots, and role of soil splash to plants. Including differentiation of these processes in function of crop type (fruits, tuber, root crops, foliar crops, other vegetables)  Distribution across water bodies (surface water, groundwater, drinking water),  Permeation from to drinking water pipelines  Transfer from soil, water, crops and feed to animals, and subsequently distribution to edible animal products (meat, milk, eggs)  Transfer from building materials (flooring, furniture, interior wall) to air (gas phase and airborne particles), settled dust and sink surfaces  Etc. A schematic, comprehensive overview of processes playing a role in the source to exposure pathway is given in section 3.4.1 (i.e. Figure 7 for the outdoor environment and Figure 8 for the indoor sources). Concerning the second part of the source-to-dose modelling chain, i.e. exposure calculations: common approach across studies involves in a first step the calculation of the dose per exposure medium (soil, drinking water, air, food) by multiplying the environmental concentrations with exposure factors per medium. Some studies involved a deterministic approach, while in other studies probabilistic calculation were used for this first step. In a second step the exposures per exposure routes are summed (e.g. oral exposure arising from food, dust and soil ingestion). In a subsequent step, exposures across routes are combined. Across the different studies, either PBPK modelling or relative absorption factors (for oral, dermal, inhalation) were used for this last step of aggregation. By preference, substance specific, data-underpinned absorption factors are used. In case of lack of substance specific absorption factors, default values are used. The studies generated as output pie charts depicting the relative contribution from different environmental sources and routes (drinking water, fish, edible plants, milk, soil, dust and air). ADDITIONAL DELIVERABLE AD6.3 36 3.2.2. Methods used to aggregate exposure in the general environment 3.2.2.1. ANSES approach for aggregate exposure for general environment Vanacker, M., et al., (2020) presents the approach that ANSES has developed to estimate the overall exposure of a target population by aggregating exposures from various sources (dust, air, soil, tap water, veterinary drugs, etc.) and routes (ingestion, inhalation, cutaneous contact) using heterogeneous data from different surveys (Vanacker, Tressou, et al. 2020). The approach can be applied to several type of populations: adults, children, toddlers, pregnant women, even workers. The approach is based on the general principle of creating a simulated population of the targeted one from the individual data of the different surveys related to the different exposure sources using Monte Carlo (MC) simulations. Different scenarios regarding MC simulations were tested and the conclusions of the most relevant one are described below in three main steps: Step 1: Selection of a reference population A reference population must be chosen to serve as a basis for the MC simulations and to reproduce its characteristics in the simulated population. To do this, among the available studies, the one on the population most representative of the target population must be chosen, considering the main characteristics (age, region, sex, etc.). Step 2: Selection of stratification variables To link the different surveys, it is proposed to check the need to use stratification variables derived from the characteristics of the population (such as age, sex, region, etc.). First, it is necessary to identify the socio-demographic variables that are shared between the surveys and that may influence the input variables. Then, using statistical analysis, the significance of correlations between sociodemographic variables and concentrations can be tested for each survey. In case of significant correlations, it is necessary to include the stratification variables when drawing the concentration values in the different surveys. Steps 3: Monte Carlo simulation strategies Second-order MC simulations are used to create 100 samples of a new population composed of 100,000 individuals. Exposure factor values (e.g., amount of soil/dust ingested or inhalation rate) are randomly drawn based on the age of each individual in their respective distributions mostly from the literature. Chemical concentration values are assigned to each individual in the newly simulated population by drawing values from observations in the various surveys. An important choice is the timing of the assignment of exposure factors and concentrations in the simulation process. It is recommended to randomly sample 100,000 individuals from the reference population and simultaneously assign exposure factors and concentration values according to important stratification variables. In addition, it is possible that the surveys include several variables of interest that are highly correlated, such as for example lead concentrations in tap water, dust, and soil, within the same dwelling. In this case, the three concentrations must be selected together to be assigned to individuals in the simulated population. To preserve correlations, it is proposed to select a vector of these three concentration variables at each simulation. To quantify the uncertainty associated with the simulation process, it is recommended to create at least 100 samples of 100,000 individuals. Different statistics and their uncertainty intervals can be then calculated. The corresponding algorithm was implemented in a RShiny Toolbox, the RSExpo software available on request to ANSES. ADDITIONAL DELIVERABLE AD6.3 37 This approach was then applied to calculate the aggregate exposure of the French adult population to a mixture of 4 pyrethroids: cyfluthrin, cypermethrin, permethrin and deltamethrin (Vanacker, Quindroit, et al. 2020). It is interesting to observe that according to the percentile studied the contribution to the aggregate exposure of the different sources and routes differ. While for 50% of the most exposed individuals, it is food ingestion that constitutes the majority of the overall exposure, for the 1% most exposed individuals, it is the combination of food ingestion and use of medication, implying dermal exposure, that are the principal contributors to the overall exposure. The same is true for the contribution to risk with a major contribution of deltamethrin in food for 50% of the most exposed individuals and a sharing of contributions with permethrin present in human and veterinary drugs for 1% of the most exposed individuals. Thus, when considering aggregate exposure, it is therefore important to consider the whole distribution and to calculate contributions at different percentiles. HBM data were used to evaluate modelled aggregate exposure to the 4 pyrethroids from the different sources and routes. For that, a PBPK model was applied to each individual exposure estimated from the different sources for the 100 samples. Then, the distribution of simulated exposure estimates and the associated uncertainty intervals were compared with HBM data from the national French ENNS biomonitoring survey. 3.2.2.2. RIVM modelling tools for consumer exposure Two modelling tools were proposed by RIVM to evaluate consumer exposure: ConsExpo and PACEM. These tools have been developed for different purposes and different applications. Both are available as web applications. ConsExpo simulates exposure that arises during single product use, based on mechanistic modelling. ConsExpo by itself is not to be used to estimate aggregate exposure. PACEM assesses aggregate consumer exposure based on product usage information. PACEM is limited in the sense that it contains only a limited set of products, and that it does not have built-in capabilities to estimate exposure from single usage events. The latter is required, but is currently user input, which limits the usability of the tool. The methodology of PACEM was presented by Delmaar, C., J., E., et al., (2022) (Delmaar et al. 2022). PACEM currently contains product usage information for personal care products and household cleaning products, gathered from different surveys inducted in several European countries. To begin, the user selects the product group(s), country, and exposure metric (systemic dose or dermal load) appropriate to their case. Subsequently, the user can choose products from the selected product groups and is asked to enter product information (% products containing substance, concentration in product, and exposure fractions or retention factors, depending on the exposure metric). Each input can be a point value or a parametric distribution. A Monte Carlo simulation is performed with a userselected sample size. For each sample, the exposure of the survey respondents is simulated, combining the user-supplied product information with the product usage information from the surveys relevant to the selected country. Finally, analysis reports can be generated, where the user has the options whether to limit the analysis to a single sex, and whether to exclude non-exposed individuals. The analysis report contains percentiles and a histogram of the desired exposure metric. The user can in some scenarios use ConsExpo to calculate the single-event exposure input for PACEM, but an automatic process would be preferred. Furthermore, combining the wider range of scenarios in ConsExpo with the aggregation framework in PACEM would significantly advance the scope of consumer exposures that can be aggregated. ADDITIONAL DELIVERABLE AD6.3 38 In Activity 6.2.1, an analysis of the input/output flow needed to combine the use of ConsExpo and PACEM, is proposed. Additionally, ConsExpo will be used to develop templates of product and substance specific inputs for PACEM. Also, a method will be proposed to combine exposures estimated with PACEM and ConsExpo for products lacking in PACEM. The results of this activity can be used by the case studies in Project 6.2.1.b and will serve as input for PARC Task 8.3, ‘Integrative Models’. 3.2.2.3. VITO approach for aggregate exposure for the general environment VITO presented the aggregate exposure approach developed under project TAGS (Tiered Approach for aggregate Exposure (CEFIC LRI project; project partners: CERTH, IOM and VITO) (van Tongeren et al. 2011). The TAGS project was a project specifically tailored toward methods for aggregation. The first part of the TAGS project summarized aggregation methods applied in 1) regulatory contexts (i.e. chemicals: REACH; Pesticides and biocides (EU regulations) and contaminated sites assessment (mostly country/region specific approach)) and 2) domain overarching methods in advisory/scientific context (i.e. WHO harmonization project, US EPA general principles for aggregate exposure, and ILSI aggregate exposure workshop report). Subsequently, the methodology for aggregation of exposure developed under TAGs was presented and demonstrated in case studies. The TAGs methodology involved structured criteria and a process flow to decide what triggers the launch of an aggregated exposure assessment, or turned around: “when is a simple (single) exposure assessment appropriate and adequate”. The method includes two stages: 1) approach for aggregation within an exposure scenario (across exposure routes), and, stage 2: aggregation across different exposure scenarios. In each stage, a tiered approach is followed. After the tier 0 step (deciding whether an aggregate exposure assessment is needed), Tier 1 approach is launched, involving deterministic exposure assessment, default exposure models (ECETOC TRA, EUSES), eventually including monitoring data (Tier 1.5), calculation of aggregated Risk Characterization Ratio (RCR). When tier 1 RCR is below 1, no further assessment is needed (risk is under control). When RCR exceeds unity, a Tier 2 step is launched, including probabilistic modelling, uncertainty assessment, PBPK modelling, differences in bioavailability, time activity patterns and human biomonitoring data. The probabilistic exposure assessment for consumer/environmental exposure was done by setting probabilistic distributions on the exposure factors and probabilistic distributions on monitoring data. In case of lack of data to construct distribution functions, exposure assessment for selected scenarios (median, P95 or max) were run. The TAGS guidance describes what type of data and models can be used in each of the tiers. This TAGs methodology was applied in several case studies, i.e. nickel, chlorpyrifos, BPA and BDE-209. The case study on chlorpyrifos included scenarios combining general environment and workers scenarios. From the BDE-209 case study, it was observed that applying slight modification on existing tools (e.g. ECETOC TRA), it is possible to perform aggregate exposure, in view of the purpose of demonstrating compliance. The apparent relative contribution of sources and routes depends on the exposure assessment method used (Cornelis 2012). When using high tier models or monitoring data, it is likely that the relative contribution of sources and routes is adequately addressed. When using a low tier exposure model, exposure is likely to be overpredicted and hence the relative contribution might be overestimated. Therefore, to get a good picture on relative contribution of sources and route, default models (Tier 1) are not suitable. ADDITIONAL DELIVERABLE AD6.3 39 3.2.3. Methods used to aggregate exposure in the occupational environment Various exposure models are available for estimating inhalation and dermal exposures in the workplace for risk assessment purposes. Occupational exposure models are mechanistic and typically provide deterministic exposure estimate outputs, although some models (e.g. ART, dART) are able to provide outputs with variability and uncertainty distributions (McNally et al. 2014, 2019). Tier 1 prediction models (e.g., ECETOC TRA) are typically used for screening purposes and tier 2 models (e.g. ART) are used for estimation of occupational exposures in specific scenarios (ECHA 2016a). Because occupational models were intended for risk assessment of chemical exposure in specific work environments, the models produce exposure estimates for highly specific scenarios and target populations. For instance, the exposure model ART has more than 20 input parameters for users to characterise the working material and environment, including questions on material moisture content, work activity, room size, and exposure control measures. By design, model estimates from risk assessment occupational exposure models are generally not applicable to more general target populations such as “adult workers” or “construction workers.” As a result, aggregation of occupational exposure is currently typically applied in specific occupational contributing scenarios where exposure may arise from various sources and occur via various routes. For instance, for different contributing scenarios throughout a workday, inhalation exposure may be estimated using ART and dermal exposure may be estimated using dART. These estimates may then be combined with known or assumed absorption fractions and aggregated into a daily dose for a typical workday. The contribution of daily doses attributable to occupational exposure may then be combined with exposure estimates from general environment exposures to determine major contributors of total exposure. Most of the models in the occupational field are not specific to one pollutant or one exposure situation but cover a wide range of work situations that can be encountered with a family of pollutants (e.g. metal dusts, vapors, liquid aerosols...). With the exception of those that combine several models in the same tool (e.g. ART and ART), they are not aggregated models because each simulation refers to a single exposure situation. Some models, such as TREXMO and TREXMO+ (Savic et al. 2016, 2020), allow the aggregation of several algorithms. Although these are also non-aggregated exposure situations (TREXMO combines the results of several models for the inhalation route only), these tools are interesting examples of possible connections between different algorithms. In addition to risk assessment exposure models, other occupational exposure models are available. For instance, job-exposure matrices (or JEMs) are exposure assessment tools used in occupational epidemiology (Descatha et al. 2022). JEMs may be applied on a general population level (e.g. all adult workers in Europe) based primarily on job title. Occupational exposure may be aggregated by certain JEMs (e.g. dermal and inhalation exposure by a benzene JEM), but the exposure estimates are typically qualitative or semi-quantitative applied to large groups, plus the aggregation may be implicit and determined largely be expert judgement. While JEMs are valid and valuable tools in epidemiology, they are seldom used in risk assessment due to factors including, but not limited to, their lack of distinction between individual with the same job title and their lack of fully quantitative exposure estimates. ADDITIONAL DELIVERABLE AD6.3 40 3.3. Summary of guidance on source-to-dose models and aggregate exposure The ambition of the PARC A6.2.1 is to develop methods and tools for modelling exposure to chemical substances, considering all relevant sources and routes, and combining general and occupational environments of life to better assess the risk and to propose more targeted management measures. These methods could then provide operational support for future European regulations concerning the protection of populations from chemical substances. Indeed, EU agencies have currently no harmonised methodology/guidance for aggregate exposure assessment to chemicals (EFSA, Cascio, et al. 2022). To develop these methods, it is necessary to review guidance and associated regulations, to highlight the obstacles, levers and recommendations in the field of aggregate exposure and source-todose modelling. This includes research into guidelines for the development of new aggregate exposure modelling methods and source-to-dose modelling methods; identification of relevant data to feed into the models; and identification of promising models on which our developments will be based. To this end, we propose a review of the various European, but also global, regulations and related guidance in connection with aggregate chemical exposure assessment and source-to-dose models. The review of these guidance is divided into three parts: guidance related to source-to-dose models, guidance related to general environment aggregate exposure assessment, and guidance related to occupational exposure assessment. Some guidance addresses several of the above three topics simultaneously and is therefore described in the paragraph where the development is most important. For the time being, the ECHA REACH guidance documents, the SCCS, EMA and EFSA have been reviewed. The review of guidance documents from the following institutions will be carried out: WHO/WHOPES, OECD, FAO, US EPA, US FDA, OSHA, Health Canada, ECHA pesticides and biocides, SCHEER, JRC, EEA, and CEFIC. 3.3.1. Guidance to model source-to-dose The project source-to-dose modelling has not performed an in-depth review of regulatory guidance related to source-to-dose models, because for various aspects (e.g., soil contamination) the regulatory scene is set at a level of regional or national authorities, and an EU regulatory framework is lacking. At a European level, the source-to-dose or human environmental exposure for the general environment exposure to chemicals must be considered in Chemical Safety Assessment to comply with the Regulation (EC) No 1907/2006 on the Regulation, Evaluation, Authorization and Restriction of Chemicals (REACH) (see also section 3.3.2.4. ). First tier screening tools such as the European Union System for the Evaluation of Substances (EUSES) (Vermeire et al. 1997) developed for and used in the context of risk assessment in REACH Registration dossiers, are designed to be conservative and are appropriate for benchmarking against a threshold value (e.g. Derived No Effect Level, DNEL) since the level of conservatism does not play a role as long as the estimated exposure is below the DNEL. However, for human health impact assessments required for REACH authorization, and for other purposes (i.e., a realistic contribution of different exposure routes and sources, comparison with human biomonitoring data) instead of an overly conservative approach for exposure modelling, concrete guidance is lacking. The REACH guidance R16 (ECHA 2016c) describes in general terms options for refinement of source-to-dose modelling (so called ‘indirect exposure’ in R16) considering refinements of release estimates, concentrations in food items, consumption rates of food items and alternative modelling approaches (e.g., for the plant uptake model). However, the guidance does not provide further links or tools for such refinements. ADDITIONAL DELIVERABLE AD6.3 41 General conclusion. Guidance on environmental source-to-dose modelling is set in framework of REACH Regulation (EC) No 1907/2006, but is mainly focused on screening tools. 3.3.2. Guidance to aggregate exposure in the general environment 3.3.2.1. Substances present in medicinal products for human use and for veterinary use (EMA guidance) General conclusion: A lack of guidance in relation to aggregate exposure modelling, but a presence of elements and positive signals on relevant data to aggregate exposure modelling Definitions. Medicinal products for human use are defined in Directive 2001/83/EC (European Parliament, Council 2022a), and veterinary medicinal products are defined in Regulation (EU) 2019/6 (European Parliament, Council 2022b). The legislative body. At the European level, the body of European Union legislation in the pharmaceutical sector is compiled in Volume 1 and Volume 5 of the publication "The rules governing medicinal products in the European Union" (EMA 2018g; EudraLex 2023). Volume 1 is the EU pharmaceutical legislation for medicinal products for human use, and Volume 5 is the EU pharmaceutical legislation for medicinal products for veterinary use. Concerning medicinal products for human use, the principal legislative text corresponds to the Directive 2001/83/EC (European Parliament, Council 2022a), and for medicinal products for veterinary use, it is the Regulation (EU) 2019/6 (European Parliament, Council 2022b). Note that Medicinal products for pediatric use, orphans, herbal medicinal and advanced are governed by specific rules (EudraLex 2023). The body of guidance. To facilitate the effectiveness of the law, guidance reflecting a harmonized approach of the EU Member States and the European Medicines Agency are made available to the applicants and marketing authorization holders of medicinal products (EMA 2018k). Therefore, the basic legislation is supported by a series of guidelines, that are also published in the following volumes of "The rules governing medicinal products in the European Union": Volume 2: Notice to applicants and regulatory guidelines for medicinal products for human use; Volume 3 (no updated): Scientific guidelines for medicinal products for human use; Volume 4: Guidelines for good manufacturing practices for medicinal products for human and veterinary use; Volume 6: Notice to applicants and regulatory guidelines for medicinal products for veterinary use; Volume 7 (no updated): Scientific guidelines for medicinal products for veterinary use; Volume 8: Maximum residue limits; Volume 9: Guidelines for pharmacovigilance for medicinal products for human and veterinary use; and Volume 10: Guidelines for clinical trial (EMA 2018g; EudraLex 2023). Volume 3 and 7 regarding respectively Scientific guidelines for medicinal products for human use and veterinary use are henceforth updated and replaced by dedicated section of the website of the European Medicines Agency (EMA 2018k). Other guidelines, such as regulatory guidelines, good-manufacturing-practice guidelines and pharmacovigilance guidelines, were excluded from this re-organization exercise (EMA 2018k). They continue to be published by the European Commission (EMA 2018k). It should be noted that the European Medicines Agency also publishes scientific guidelines on human medicines that are harmonized at the international level, by the International Council for Harmonization of Technical Requirements for Registration of Pharmaceuticals for Human Use (ICH) (EMA 2018e). ADDITIONAL DELIVERABLE AD6.3 48 General guidance on dermal exposure, and provision of useful data sources and models. For dermal exposure, the guide proposes a simple modeling considering the concentration of a substance in a product category, the amount of product category that is applied/received per day and a retention factor specific to product category (SCCS 2023a). Note that the amount per day can be calculated from the frequency of application and the amount per application. The guide mentions several references for obtaining exposure parameters, compiled in tables 3A, 3B, and 4, p. 27, 28, and 29. Appendix 7 of the guide completes that report with a literature review from 2015 to 2020 (SCCS 2023a). It also lists studies which provide detailed external exposure values to different cosmetic products, for specific countries (SCCS 2023a). Most of these data are derived from a large-scale use study from Cosmetics Europe (SCCS 2023a:202). Note that for deriving the relative amounts and exposures reported in Table 3A of the guide, bodyweight distributions from the European countries included in the study were used in a Monte Carlo approach (SCCS 2023a). It should be noted that concerning in-use conditions of topical products, no comprehensive exposure data for newborns and early infants, representative for Europe, are available in the open literature (SCCS 2023a). The European Commission is preparing aggregate exposure data for babies and children for different baby care cosmetics used in Europe (more information, see Table A.7, Appendix 7 of the guide) (SCCS 2023a). For oral exposure, the same principles as described for dermal exposure can be applied. General guidance on inhalation exposure, and provision of useful data sources and models. For inhalation exposure, when using mathematical models, a tiered approach should be followed. Default equations can be used as a conservative, worst case approach, and as a first estimate (SCCS 2023a). For a more realistic assessment 1or 2-Box models, as well as higher-tier models, can be considered (SCCS 2023a). For higher tier assessment, one of the tools that can be considered for calculating exposure estimates is the ConsExpo model (SCCS 2023a). The guide cites some references where to find values for the key parameters, for example, deposition rates have been determined in an International Commission Radiological Protection project (SCCS 2023a). The SCCS emphasizes that it is not the intention to provide parameter values and exposure estimates for all cosmetic product categories. Only for the most common categories are default values provided. For all other cosmetic product categories, the individual companies and/or the qualified safety assessors need to make a case-by-case assessment of the daily exposure level and/or the frequency of application (SCCS 2023a). General guidance on aggregate exposure assessment: absence of technical details. According to the guide, aggregate exposure is obtained by adding up the exposures to a cosmetic ingredient contained in several single product categories. It needs to be calculated when several product categories contribute (SCCS 2023a). For the calculation of LEDs, the aggregation is specific to the investigated site and if a risk assessment should be conducted for local exposure, the cosmetic ingredient single doses need to be added up for the specific investigated site (SCCS 2023a). If the external aggregate exposure should serve to calculate SEDs, aggregation needs to take into account all product categories that can be taken up by a specific route (SCCS 2023a). For each route a specific aggregate external exposure needs to be provided (SCCS 2023a). If aggregation over routes is necessary, because different routes (e.g. dermal and inhalation route) contribute, aggregation over routes needs to be calculated on the level of internal exposure (SCCS 2023a). The guidance does not ADDITIONAL DELIVERABLE AD6.3 49 provide technical details on how to implement strategies for source-aggregation-by-route, and for route-aggregation. As specified previously, for CMR 1A and 1B substances, according to Art. 15d of the Cosmetic Regulation (European Parliament and the Council 2022), the consideration of aggregate exposure from all sources (including non-cosmetics) is required (SCCS 2023a). A guidance document has been developed by the EU Commission with the aim of enabling a harmonized approach to the development and use of aggregate exposure estimates in assessing the safe use of CMR substances as cosmetic ingredients (SCCS 2023a). This guidance is presented in Appendix 5 of the cosmetic guidance document (SCCS 2023a). This guideline establishes in particular the pivotal role of the SCCS in the dialogue to be set up, when necessary, between the various European authorities: ECHA, EFSA and EMA, potentially holding data that can be used to estimate an aggregated exposure. However, the guideline does not provide technical on how to implement strategies for source-aggregation-byroute, and for route-aggregation. Finally, the guide recalls the value of PBTK models for deriving internal exposure values from external exposure estimates, as well as the benefits of acquiring biomonitoring data to support aggregate exposure modeling (SCCS 2021). Currently, SCCS has issued several public notes expressing its views on various case studies involving aggregate exposure: opinions on the aggregate exposure to aluminium concerning the European population when considering the use of cosmetics and personal care products, medicines (e.g. antacids) and dietary intake (SCCS 2022b); climbazole in cosmetics (SCCS 2017); aggregate exposure of vitamin A in food and cosmetics (SCCS 2020, 2022a); aggregation exposure of Butylphenyl methylpropional (pBMHCA) in cosmetic and non-cosmetic products (SCCS 2019a); aggregate exposure of children on methyl salicylate in multiple cosmetic products (SCCS 2023b); and on skin sensitization quantitative risk assessment for fragrance ingredients (SCCS 2019b). Concerning the opinion SCCS/1644/22 on aluminium, as it does not belong to substances classified as CMR 1A and B, only exposure from cosmetic uses was considered in the safety assessment provided by the committee (SCCS 2022b). However, in a scenario provided by the applicant, exposure from noncosmetic sources of aluminium: food and pharmaceuticals was aggregated with exposure from cosmetics (SCCS 2022b). In short, the applicant has proposed an independent internal exposure calculation for cosmetics and non-cosmetics, and has therefore not aggregated the two sources (SCCS 2022b). Several scenarios were proposed for calculating exposure to aluminium from cosmetics, including a probabilistic scenario considering distributions of aluminium concentrations in products, and a deterministic scenario using maxima for concentrations (SCCS 2022b). All the scenarios used absorption factors to derive internal exposures. In addition, the Creme care model was used to calculate exposure (SCCS 2022b). The SCCS judged the probabilistic approach to be invalid due to the use of a distribution and not maxima on concentrations, as requested in the guidance, and highlighted above (SCCS 2022b). Concerning the opinion SCCS/1639/21 on vitamin A, in short, the applicant proposed a probabilistic exposure calculation for cosmetics, with simulation of a population, and certain parameters taken from distributions, without resimulations (which could have made it possible to model the uncertainty over the choice of values in the distributions) (SCCS 2022a). On the other hand, the same type of ADDITIONAL DELIVERABLE AD6.3 50 probabilistic calculation was carried out on a population of food consumers (SCCS 2022a). Stratification variables: age, sex and country, were used to link the two populations, taking those of cosmetics users as a reference (SCCS 2022a). The Creme care and cosmetics and Creme global nutrition probabilistic models were used (SCCS 2022a). The SCCS did not approve the calculation proposal for reasons unrelated to the statistical approach used (SCCS 2022a). In the case of skin sensitization, quantitative risk assessment for fragrance ingredients, SCCS cites the Creme RIFM Aggregate Exposure Model, which is a probabilistic model (SCCS 2019b). However, the committee does not currently provide any recommendations regarding this type of approach and plans to evaluate it in the future (SCCS 2019b). The other opinions cited above do not propose a probabilistic approach, or the aggregation of cosmetic and non-cosmetic sources. 3.3.2.3. Guidance on nanomaterials (SCCS guidance) General overview. Concerning nanomaterials and other personal care products than cosmetics, the SCCS committee does not appear to have issued a guideline for aggregate exposure. 3.3.2.4. Substances subject to an exposure assessment under REACH General conclusion. Statement on the importance of conducting an aggregate exposure assessment and the lack of guidance currently available. Many recommendations for exposure models and exposure data. General introduction to human exposure assessment under REACH regulation. At the European level, the regulation concerning the Registration, Evaluation, Authorization and Restriction of Chemicals (REACH) explicitly aims to ensure a high level of protection of human health and the environment, including the promotion of alternative methods for the assessment of the hazards of substances (European Parliament and The Council 2022:1). Note that the promotion of methods to characterize exposure is not explicit, but elements reflecting this inclination are present, as developed in the following. Some chemicals are not covered by the regulation and are mentioned in Article 2 of the Regulation (European Parliament and The Council 2022:2). Nevertheless, REACH applies to many chemical substances: those that are used not only in industrial processes and in the workplace, but also in daily life, for example in cleaning products, paints as well as in articles such as clothing, furniture and electrical appliances (ECHA 2023b). The regulation therefore impacts on most businesses in the EU (ECHA 2023b). REACH places the burden of proof on companies (ECHA 2023b), i.e. producer and importers of articles (European Parliament and The Council 2022:7). To comply with the regulation, companies must identify and manage the risks associated with the substances they manufacture and market in the EU (ECHA 2023b): in particular, if the substance fulfils any of the criteria of the Article 14(4) hazard classes, categories or properties for any endpoint (European Parliament and The Council 2022:14 4.), an exposure assessment and risk characterization has to be carried out, and described in the Chemical Safety Report (ECHA 2011:2). The place of aggregate exposure in the REACH regulation. In REACH Regulation, terminology ‘Aggregate exposure assessment’ does not appear. Nevertheless, reference is made to this concept, which is reflected in the term 'combined exposure'. In this regard, paragraph 5.2.4. of Annex 1. (European Parliament and The Council 2022:1-5.2.4.) states that: ‘An estimation of the exposure levels ADDITIONAL DELIVERABLE AD6.3 51 shall be performed for all human populations (workers, consumers and humans liable to exposure indirectly via the environment) and environmental spheres for which exposure to the substance is known or reasonably foreseeable. Each relevant route of human exposure (inhalation, oral, dermal and combined through all relevant routes and sources of exposure) shall be addressed. Such estimations shall take account of spatial and temporal variations in the exposure pattern’. The assessment of aggregate exposure is therefore a regulatory requirement. Each assessment of the aggregate exposure should be based on an exposure scenario, which is, according to the regulation, Annex 1, paragraph 0.7.: ‘the set of conditions that describe how the substance is manufactured or used during its life cycle and how the manufacturer or importer controls, or recommends downstream users to control, exposures of humans and the environment’ (European Parliament and The Council 2022:1-0.7.). Furthermore, according to the regulation, Annex 1, paragraph 0.7., an exposure scenario ‘may […] cover a large range of processes or uses’ (European Parliament and The Council 2022:1-0.7.). These references to the populations to be considered, to spatial and temporal variations in exposure, and to multiple uses, therefore implicitly imply that it may be relevant to combine exposure at work and during general life. The aggregate exposure assessment should be reported in the Chemical Safety Report, at the level of heading 10.x. ‘Overall exposure (combined for all relevant emission/release sources)’, subheading 10.x.1. ‘Human health (combined for all exposure routes)’ (European Parliament and The Council 2022:1-7.). In addition, the regulations, at Annex 1., paragraph 5.2.5., highlight the importance of using appropriate modelling tools to estimate exposure, and the importance of considering all available exposure data (European Parliament and The Council 2022:1-5.2.5.). However, there is no mention of the treatment of uncertainty and variability in the estimation of exposure, which is mentioned in the hazard characterization, at Annex 1., paragraph 1.4.1.(a) (European Parliament and The Council 2022:1-1.4.1.(a)). General information on the ECHA guidance document in relation to REACH. A set of guidelines specifies these regulatory provisions. These guidance documents are produced by the European Chemicals Agency (ECHA), established by (European Parliament and The Council 2022:75) for the purposes of managing and in some cases carrying out the technical, scientific and administrative aspects of this Regulation and to ensure consistency at Community level in relation to these aspects (European Parliament and The Council 2022:75). These guidance documents constitute a corpus of 25 documents, gathered under the name ‘Guidance on information requirements and chemical safety assessment’ (ECHA 2023a). Among these 25 documents are 5 Concise guidance and 20 Reference documents (ECHA 2023a). The purpose of the concise guidance is to support the processes needed to meet the information requirements on intrinsic properties of substances to be registered, and where relevant to carry out a chemical's safety assessment (ECHA 2023a). This includes information collection processes, communication processes and assessment processes. The purpose of the reference guidance is to provide in-depth scientific and technical advice (ECHA 2023a). All these documents were consulted. Relevant information for aggregate exposure was collected in the following Concise guidance: ‘A. Introduction’, ‘D. Exposure assessment’ and ‘E. Risk characterization’. Additional information was collected in the following Reference documents: ‘Occupational exposure assessment (Chapter R.14)’, ’Consumer exposure assessment (Chapter R.15)’, ‘Environmental exposure estimation (Chapter R.16)’, and ‘Uncertainty analysis (Chapter R.19)’. Part A. Reaffirmation of the importance of conducting an aggregate exposure assessment. Part A introduces the guidance for conducting the chemical safety assessment and preparing the chemical safety report for substances manufactured or imported in a quantity of 10 tonnes or more per year. ADDITIONAL DELIVERABLE AD6.3 52 This section details certain regulatory provisions (ECHA 2023a). As observed in the regulations, there is a ‘high degree of flexibility on how to derive results’ (ECHA 2011:2). Further on, it is mentioned that ‘the exposure estimation may be derived from models or from measured data’ (ECHA 2011:6), implicitly referring to data-based models, and to more mechanistic models, such as source-to-dose models. Reference is then made to the tiered modelling approach (ECHA 2011:7). Finally, it is reaffirmed that the risk characterization needs to consider risks from [aggregate] exposure via different routes of exposure or via different sources (ECHA 2011:15). Part D. Suggestion of data sources, and statement of the lack of guidance on aggregate exposure. Part D sets out the principles for carrying out an exposure assessment to determine the conditions of safe use for all the uses of a substance registered under REACH (ECHA 2023a). It covers exposure for environment, workers and consumers (ECHA 2023a). Firstly, this chapter provides information on the available data sources on exposure factors that can be used to estimate exposure, and in particular aggregate exposure (ECHA 2016e:24–26). Of the types of data sources mentioned, the following are open access. First, there is information agreed at sector level made available by industry sector associations as use maps (ECHA 2016e:24–26). These use maps contain relevant and realistic information on uses and conditions of use. There exist use maps for worker exposure, called Specific Workers Exposure Description (SWED), use maps for human exposure via the environment, called Specific Environmental Release Categories (SPERCs), and use maps for consumer exposure, called Specific Consumer Exposure Determinants (SCEDs). ECHA provides a use maps library, available on its website: Use maps - ECHA (europa.eu). Then, there are published documents meant to describe technical processes and/or work processes from the perspective of environmental release or human exposure such as: Emission Scenario Documents (ESDs), which are documents developed under the OECD describing the sources, production processes, pathways and use patterns (ECHA 2016e:24–26). ESDs aim to quantify the releases of a chemical into water, air, soil and/or solid waste (ECHA 2016e:24–26); Best Available Techniques (BAT) which are reference documents (BREFs) developed in the context of the Industrial Emissions Directive (IED, 2010/75/EU) (ECHA 2016e:24–26). They include techniques and processes used in a specific sector, as well as current emission and consumption levels (ECHA 2016e:24–26); Control sheets including basic advice on exposure controls to hazardous substances in the workplace (ECHA 2016e:24–26). It takes the form of straightforward advice in ‘factsheets’ called ‘control guidance sheets’ which are sometimes specific to a certain industry sector e.g. Control of Substances Hazardous to Health (COSHH) sheets developed by the UK HSE authority or the so called VSKs developed by the German authorities. Then, this chapter provides a section dedicated to aggregate risk (ECHA 2016e:24–26). In this regard, it is mentioned that ‘a simple and robust (Tier 1 type) methodology to assess "co-use/exposure" [i.e. aggregate exposure] patterns among consumers is not yet readily available’ (ECHA 2016e:24–26). Finally, this chapter contains a section dedicated to the treatment of uncertainty in a broad sense (ECHA 2016e:24–26). In particular, the importance of considering the uncertainty of the input parameters in the higher tier models is stressed (ECHA 2016e:24–26). Part E. A Tier 1 aggregate exposure strategy. Part E contains the guidance on risk characterization (ECHA 2016f). Firstly, it contains a section dedicated to aggregate exposure (ECHA 2016f:37–39). It is explained that, where relevant, multi-sources and multi-routes exposure must be assessed, and occupational exposure and general environment context of exposure should be associated (ECHA 2016f:37–39). After assessing the risk by route of exposure, taking into account any necessary management measures (van Tongeren et al. 2011), it is recommended to sum the RCRs associated with each relevant route and source, using external values of exposure and reference values, where at least one common ADDITIONAL DELIVERABLE AD6.3 53 target organ is identified, and where the risk is chronic (ECHA 2016f:37–39). Note that in this approach, the exposure is indirectly aggregated: this aggregation is done at the risk level only. Furthermore, uncertainty is not considered. Implicitly, it seems to us that this proposal should be considered as tier 1. Finally, this chapter seems to express an idea that is absent from the regulations, but which is important for the recognition of aggregate exposure modelling methods (ECHA 2016f:38): ‘[…] if biomarkers of exposure can be reliably measured and if reliable information on the biomarkerresponse relationship is available, the assessment of the integrated risk for various routes of exposure is considered more valid and more predictive based on biomonitoring data than on the approach via the route-specific risk characterization ratios.’ (ECHA 2016f:38). This idea is qualified in the following paragraph (ECHA 2016f:38): ‘But even in this data-rich situation knowledge on the relative route-specific contribution of exposure to the overall risk is considered helpful in order to inform risk managers to concentrate on the most effective route-specific risk management measures.’ (ECHA 2016f:38). Chapter R.14. Recommendations in terms of data. Little on aggregate exposure. Chapter R14 contains elements on occupational exposure assessment (ECHA 2016a). It describes how to build the exposure scenario and estimate the exposure. Some tools for occupational estimation are mentioned. It should be noted that other tools than those mentioned in this chapter, as well as more up-to-date versions of these models can be used under REACH, if appropriate (ECHA 2016a:38–42, 48–70). These tools are: ECETOC TRA, MEASE, EMKG-EXPO-TOOL, TREXMO, STOFFENMANAGER, RISKOFDERM, BEAT and ART. The last five are designated to be used for higher tier exposure assessment (ECHA 2016a:38– 42, 48–70). These models are extensively described in the document (ECHA 2016a:38–42, 48–70). Finally, concerning occupational aggregate exposure, a reference to the RCR formula in Part E is made (ECHA 2016a). Chapter R.15. Recommendations in terms of models and data. Little on aggregate exposure. Chapter R15 contains elements on consumer exposure assessment (ECHA 2016b). Good practice for estimating exposure to different sources and routes, taken separately, is explained in detail (ECHA 2016b). In this context, reference is made to several exposure models: ConsExpo, ECETOC TRA consumer tool, AISE, ESIG EGRET, REACT, WPEM, CEM, MCCEM, INTERA, BAMA/FEA Indoor Air model, and RIVM emission model (ECHA 2016b:36–46, 62–69). These models are described in some detail (ECHA 2016b:36–46, 62–69). Furthermore, reference is made to data sources relevant for exposure estimation. In short, here is the type of data that has not yet been cited (ECHA 2016b). Datasets from industry peer-review initiatives such as HERA and Nordic Council for use frequency factors and exposure duration factors; datasets in mouthing times for children in the context of restriction dossier opinion, on lead and its compounds in articles intended for consumer use from the Committee for Risk Assessment (RAC) (RAC and SEAC 2014); in the context of other restriction proposals, mouthing times for articles, including toys, containing phthalates have also been discussed (ECHA 2013a:24; RAC 2013); regarding physiological parameters and activity factors (ECHA 2016b:26), reference is made to: RIVM Fact Sheets (RIVM 2013), JRC ExpoFacts Database (Zenié and Reina 2007), and US EPA Exposure factor handbook (US EPA 2015); regarding exposure data, reference is made to about 20 references, collected in Appendix R.15.2 of (ECHA 2016b:57–61); finally, reference is also made to release coefficients of some articles, given in [OECD, 2013]. All these data sources have been included in our data inventory. Finally, concerning general environment multi-source and multi-route exposure, a reference to the RCR formula in Part E is made (ECHA 2016b:48–49). Contrary to the statement in Part E, it is clearly stated here that this formula is tier 1. Thus, it is stated that: ‘in most cases more sophisticated (e.g. probabilistic) methods and corresponding datasets will be needed, to properly reflect the co-use ADDITIONAL DELIVERABLE AD6.3 54 pattern of products across consumers’ (ECHA 2016b:48–49). Regarding the consideration of uncertainty in these calculations, reference is made to the dedicated guidance chapter (ECHA 2016b:48–49). Chapter R.16. Recommendations in terms of models and data. Little on aggregate exposure. Chapter R16 contains elements on environmental exposure assessment, both source-to-dose models that predict concentrations in environmental compartments, and models (source-to-dose or not) that predict human exposure to environmental exposure sources such as ambient air, drinking water, and food items (ECHA 2016d). Firstly, some tools are described (ECHA 2016d:53,70,87,149-156): EUSES, ECETOC TRA, Chesar, EUTGD excel sheet, ECPA LET, FOCUS recommended models (MACRO, PEARL, PELMO, PRZM_GW, STEPS1-2, and SWASH) and CHARM. Then, there is a section with recommendation regarding release assessment, which is part of sourceto-dose modelling (ECHA 2016d). It follows another section dedicated to exposure estimation (ECHA 2016d). There are recommendations on source-to-dose modelling, and particularly on distribution and fate of the released substances in the environment, and concentration estimation in exposure sources, some of which are in the scope of the project, i.e. sources of exposure for human: aquatic compartment, atmosphere, soil compartment, and groundwater (ECHA 2016d). There is another part of this section with recommendations on human exposure assessment (ECHA 2016d). In this part it is exclusively mentioned the use of exposure factors from the EUSES tool (ECHA 2016d). Then, concerning general environment aggregate exposure, it is noted that ‘a combined assessment for several uses (or techniques for a same use) taking place at a same site is usually not covered in the registration dossier, as the variety of combination across the registrants market may be too wide' (ECHA 2016d:29). Finally, there is a section dedicated to assessment of the quality of measured data, which could be studied in more detail to improve the quality descriptors of our data inventory (ECHA 2016d:168–73). Chapter R.19. Technical details on the evaluation of uncertainty in exposure assessment, but no explicit mention of aggregate exposure. Finally, Chapter R19 provides guidance on dealing with uncertainty in the chemical safety assessment and outlines methods for making an uncertainty analysis (ECHA 2016d:7). The underlying principle is that a tiered approach should be followed and that the amount of detail should be proportionate to the level of uncertainty and its potential impact on the risk characterization (ECHA 2016d:7). The guidance has been written according to the principles outlined in the World Health Organisation's (WHO) "Draft guidance document on characterizing and communicating uncertainty in exposure assessment" (WHO-IPCS, 2006) (ECHA 2016d:7). There is a first section on the role of uncertainty analysis in the chemical safety assessment (ECHA 2016d:11–13). Then, there is a chapter on key concepts in uncertainty analysis (ECHA 2016d:11–13). The following section is related to uncertainty analysis in the chemical safety assessment (ECHA 2016d:11–13). In this section, it is stated that RCR uncertainty is an important output, and uncertainty analysis is recommended when RCR exceeds or approaches 1, and when non-standard and non-guideline approaches have been used (ECHA 2016d:11–13). Aggregate exposure is not clearly mentioned here (ECHA 2016d:11–13). Later in this section, the tiered strategy is detailed (ECHA 2016d:13–15). Level 3 is probabilistic assessment (ECHA 2016d:26–27). It is reported that these methods lack guidance. Some approaches are listed: 1D and 2D Monte Carlo simulations, bootstrapping and Bayesian analysis, fuzzy arithmetic and probability bounds. Uncertainty analysis ADDITIONAL DELIVERABLE AD6.3 55 made in EUSES is cited as a template. The probabilistic approach to exposure assessment is described as follows: 1. ‘Based on the knowledge obtained by the qualitative and/or quantitative deterministic uncertainty analysis, parameters to be treated in a probabilistic approach should be identified. 2. Uncertainty and variability of model input parameters should be described by appropriate distributions. This usually involves the collection of data, expert judgement and fitting distribution functions to data. Dependencies among model input parameters should be also considered. 3. Computations (e.g. Monte Carlo simulations) should be carried out to estimate the propagation of variability and uncertainty through the model. The model output will be also a probabilistic distribution shaped by uncertainty and variability. 4. The estimated exposure can be expressed by a probability distribution […]. 5. Confidence intervals can be also calculated for the cumulative distribution. While the cumulative distribution mainly represents the variability (e.g. spatial and temporal variability of exposure), the width of confidence intervals mainly indicates the contribution of uncertainty sources. The uncertainties associated with scenarios and applied models are usually not treated with probabilistic methods. In principle, the probabilistic approach can be applied to different scenarios or models, and associated uncertainties can be evaluated [by deterministic approaches]. In alternative, different scenarios/models can be also assigned probabilities representing their relative plausibility.’ 3.3.2.5. EFSA guidance documents and associated regulatory activities on aggregate exposure assessment General conclusion: EFSA has developed cross-cutting guidance, applicable to all EFSA sectors, on combined exposure and risk assessments, which has been implemented in guidance documents and assessments, mostly in the area of pesticides. In addition, EFSA is interested in the estimation of the contribution from dietary exposure to the overall exposure combining several sources. EFSA has not developed specific guidance on aggregate exposure, although has conducted ad-hoc aggregated exposure assessments. In 2022, EFSA prioritized the need for further developing the methodology and launched a process for developing a harmonized crosscutting methodology and regulatory guidance for aggregate exposure assessments covering all relevant sources and routes of exposure to chemicals. The vision is to partnership with European institutions and involve international actors for developing guidance by 2030. EFSA role and regulatory needs for dietary exposure. Most EFSA activities regarding chemical risk assessment focus on dietary exposure, covering: a) substances intentionally added to food such as food and feed additives; b) those not intentionally added but present in food due to their intended use, such as food contact materials or pesticide residues; and c) food contaminants. Aggregation of dietary exposure. The same substance may have different uses, resulting in the presence of the substance in different food items. The dietary exposure assessment results from the combination of the diet of the consumers, which varies according to different population groups and habits, and the levels of the substance in the different consumed food items. EFSA has developed extensive guidance and tools on this topic, and is the European reference regarding methodological developments for dietary exposure assessments. The same substance may be present in different foods, and combining exposure from different foods during the EFSA assessment is a common practice for both acute and chronic exposure assessments, and is not considered as “aggregate exposure”. The same substance may have several uses, each contributing to the overall dietary exposure. Some EFSA assessments consider the overall dietary exposure from all different sources, such as the recent assessment of copper (EFSA Scientific Committee et al., 2023). ADDITIONAL DELIVERABLE AD6.3 56 In the specific case of the assessment of pesticide residues, EFSA assesses independently the risk of each single use, on the basis of the Good Agricultural Practices (application rate, number of applications, postharvest interval, etc.) for each crop. The same pesticide active substance may be used for different crops. In line with the sectoral guidance documents, the acute dietary exposure assessments of pesticide residues, considers only the exposure from the treated crop under assessment, while for the chronic exposure assessment levels from all approved or intended uses, are considered, aggregating the exposure from all relevant food items for each diet (EFSA et al. 2019). Recent sectoral guidance documents suggest the aggregation of dietary exposure from other sources not assessed by EFSA if the information is available (e.g., (EFSA Panel on Food Additives and Flavourings (FAF) et al. 2022)). EFSA role and regulatory needs for non-dietary exposure. EFSA conducts non-dietary assessments for pesticides and feed additives; and has developed guidance documents and exposure tools to cover this need. The assessments cover occupational exposure, with a focus on inhalation and dermal exposure, and environmental risks. For pesticides, non-dietary environmental exposure to humans is also included. Currently, these assessments are considered independently from the dietary exposure, without aggregating exposure estimations from dietary and non-dietary routes. In 2016, EFSA published an overview of existing methodologies for estimating non-dietary exposures (EFSA 2016); a main recommendation was to consider the models developed by ECHA. Aggregation of non-dietary exposure. Recent sectoral guidance documents suggest the aggregation of non-dietary exposure from other sources not assessed by EFSA if the information is available; no specific methodology is proposed, just generic references to ECHA and SCCS approaches (e.g., (EFSA Panel on Food Additives and Flavourings (FAF) et al. 2022)). Current EFSA approaches on aggregate exposure assessment. In 2022 EFSA included the development of a framework for aggregate exposure assessment as part of the priorities (EFSA, Cascio, et al. 2022); the vision is to develop harmonized methodology in close cooperation with European and international partners. This priority is under implementation, following an open call launched by EFSA for outsourcing this activity. In addition, aggregate exposure considerations have been included in some risk assessment protocols published by EFSA, such as the one for phthalates (EFSA, Mancini, et al. 2022). Examples of aggregate exposure assessments conducted by EFSA In some specific cases, EFSA has included aggregated exposure considerations in their risk assessments. The approach has been to conduct case-by-case aggregated exposure assessments according to the specific needs. Examples are the risk assessments for carvone, focused on oral exposure and aggregating dietary exposure and oral exposure from cosmetics and personal care products (toothpaste and mouthwash) (EFSA 2014); the risk assessment of bisphenol A (BPA) in foodstuffs, aggregating exposure from diet, dust, cosmetics and thermal paper (EFSA Panel on Food Contact Materials, Enzymes, Flavourings and Processing Aids (CEF) 2015); or the risk assessment for phthalates in food contact materials, aggregating dietary exposure from different sources and including comparisons with other dietary and non-dietary exposure estimations including human biomonitoring (EFSA Panel on Food Contact Materials, Enzymes and Processing Aids (CEP) et al. 2019). In the recent copper opinion, other sources than dietary exposure are considered, but considered negligible for the general population compared with dietary exposure (EFSA Scientific Committee et al., 2023). ADDITIONAL DELIVERABLE AD6.3 57 3.3.3. Guidance to aggregate exposure for the occupational environment There is a general lack of regulatory guidance available for aggregate exposure for workers. For instance, in Chapter R.14: Occupational Exposure Assessment from ECHA's Guidance on Information Requirements and Chemical Safety Assessment, the only exposure aggregation approach described involves combining exposures across work tasks within a single exposure route (e.g., inhalation or dermal) to calculate full-shift exposure (ECHA 2016a). Ch R14 does identify the possibility of using biological monitoring to reflect total exposure to a substance through any relevant route and from any source (i.e. from consumer exposure, via environment in addition to occupational exposure through inhalation, dermal absorption and ingestion). The advice given for the use of biomonitoring is, however, limited as noted already by Louro et al., 2019 (Louro et al. 2019). Although REACH guidance does not give much advice for the estimating aggregated exposure/risks covering all exposure routes, in REACH authorisations for systemic carcinogens and reprotoxicants, risks caused by aggregated exposure covering both inhalation and dermal routes of exposure are typically calculated for human health impact assessment purposes. This is typically made combining estimated risks for all relevant worker contributing scenarios (WCS) and for both inhalation and dermal exposure, i.e. RCRcombined = RCRWCS1_inhal + RCRWCS1_dermal + … + RCRWCSX_inhal + RCRWCSX_dermal (RCR stands for risk characterization ratio, RCR = exposure level/DNEL). Chemical safety report template for authorisation applications contains table 10.1.1 which contains calculation of combined inhalation and dermal risk for all relevant worker contribution scenarios. 3.4. Developing an approach to aggregate exposures from different sources and routes and living environments Based on previous actions on exposure model selection and connection, data inventory, review of common methods and available guidance to aggregate exposure, the first lines of the approach we are developing are presented in this section by type of aggregation (multiple sources and living environments). 3.4.1. Approach to perform source-to-dose modelling The information described in previous sections on source-to-dose models (0, 3.2.1 and 3.3.1) demonstrated that several source-to-dose models and tools are in place, and are being used in scientific and policy case studies addressing 1) the impact of environmental release or contamination on humans living in the neighbourhood of contamination 2) the impact of release from indoor sources to human exposure. While some model elements may overlap across some models, other aspects (e.g. exposure pathways/dynamic versus steady state model, land use scenarios, etc.) are model-specific or elaborated in a more sophisticated or mechanistic way than in other models. Also, the level of parameterization (chemicals, exposure groups) differs across models. The proposed approach will rely on the pool of source-to-dose models (see 0) and will lead to the selection of the most appropriate model (or combination of models) to address the research or policy question. For the outdoor environmental part, a conceptual framework was created, a mapping of relevant transfers, processes and exposure pathways (see Figure 7). Figure 7A presents the first part of the exposure pathway (from emission source to exposure sources), Figure 7B builds further on Figure 7A, ADDITIONAL DELIVERABLE AD6.3 64 For aggregation of exposures from multiple routes of exposure (oral, dermal, inhalation), it is necessary to apply either physiologically based kinetic (PBK) models, or simple absorption or exposure conversion factors. MCRA accommodates both options. In the EuroMix project, a generic PBK model was developed (Tebby et al. 2020), that is now included in MCRA and which can be applied to any substance by providing appropriate parameter values in a model instance data file. This generic model is currently being tested in the PARC P6.2.2a PBPK project by comparing its performance with more specific PBK models for prioritised substances and substance groups. When desired, such specific PBK models can also be included in MCRA using an established protocol for linking external PBK models. 4. Case studies for application of the strategy 4.1. Case studies for source-to-dose modelling The objective of the PARC P6.2.1.a case studies is to test, demonstrate and use the approach and tools for source-to-dose modelling (see previous sections) for addressing science-policy questions regarding impact of environmental release and contamination on human exposure, and hereby addressing the relative contribution from the different routes and sources of exposure, in order to select the most effective risk management options. The complementarity and slightly different focus compared to case studies for aggregate exposure (section 4.2) rests with the following aspects. 1. In source-to-dose modelling, we will rely partly on modelled concentrations in exposure sources (instead of monitoring). Modelling will fill gaps when monitoring data are lacking (e.g. lack of monitoring data in edible crops or animals, milk, eggs around contaminated sites). 2. In source-to-dose modelling, we will not only consider general population exposed to ‘background’ or average environmental concentrations in European populations, but also address exposures at contaminated sites. The contribution of exposure routes and sources of exposure at contaminated sites might differ strongly with respect to pollution level, contamination level and land use scenario. 3. In source-to-dose modelling, the modelling will help to indicate the contribution of pathways to contamination in exposure sources (e.g. is contamination in crops mainly due to aerial deposition, or rather due to uptake from soil). 4. In source-to-dose modelling, we will model impacts of policy option on exposure reduction. Besides these aspects that differ between source-to-dose and aggregate exposure projects, other aspects (chemicals, exposure factors, concentrations/exposure levels at non-polluted sites) are similar across the projects, and collaboration is set up and will continue during case studies. In the phase of drafting this derivable, the case study construction process has started. VITO and LNS proposed case studies (for outdoor environmental exposure: PFAS and metals; for indoor environment: plasticizers including phthalates, and metals), and drafted a structure to elaborate case studies. VITO and LNS launched a call to partners to join case studies, or to propose other case studies. Proposals for case studies will be discussed at upcoming meetings. 4.2. Case studies for aggregate exposure 4.2.1. General information on the case studies The objective of the PARC P6.2.1.b project is to develop an aggregate modelling method to predict human exposure to chemicals considering all relevant sources, routes and living environments, i.e. ADDITIONAL DELIVERABLE AD6.3 65 general and occupational environment of life. The aim is to determine the contribution of the different sources of human exposure, and to propose effective management measures. In parallel with the development of this method, case studies will start at the beginning of the second year as an application stage. The objective of this application stage is to test the developed aggregate method and to feed it with practical questions. The goal of the case studies is also to draw up when possible European panoramas on exposure and contribution profile of exposure sources, for several exposure situations, i.e. several substances and target human populations. A case study will focus on one substance or a mixture of substances from a common chemical family, and to a target population of interest that may be characterized by a particular working activity. Case studies with several European countries, including different partners and with available data will be prioritized. Three types of case studies are possible, and the choice of the type of case study depends on the data available and its relevance to the exposure situation to be investigated. The first type of case study concerns exposure situations involving only the general living environment, i.e. for which occupational exposure is not relevant or difficult to model with available information. The second type of case study concerns exposure situations involving only the occupational environment, i.e. for which exposure involving general living environment is not relevant or difficult to model. Finally, the third type of case study associates the two exposure environments: general living environment and occupational environment. 4.2.2. Case study construction process The construction of the case studies began in parallel with the development of the aggregate exposure modelling method. On the basis on the first list of proposed case studies by the A6.2.1 partners during the construction of the PARC project, a second list was established by calling partners to fill a case study inventory describing the substances, the target populations they would like to work on and the available exposure and HBM data they have access to. The case study inventory is composed of one Excel file for each of the three case study types (general environment, occupational environment and combining both). Each case study sheet is dedicated to one chemical or chemical family and is composed of descriptors presented in Table S4 (Appendix 5). New case studies not mentioned during PARC construction can be proposed, and it is also possible that some partners may withdraw from initially proposed case studies. The objectives of this case study inventory are also to make the partners aware of the need to find relevant data, to detect early on potential data gaps that will have to be filled or that will have to be dealt with in the development of the modelling methods. Finally, it is intended to draw up an overview of possible collaborations between European countries. Note that it is possible for a member to participate in a case study without providing data, but only expertise. 4.2.3. Description of proposed case studies The case studies proposed to date have begun to be described in a table (see Appendix 5 for an overview of the descriptors in the inventory). It is possible that these groups may still evolve before the implementation of the case studies, as explained in section 4.2.2. The following figures summarize the types of proposed case studies: the largest circles represent the target populations, the intermediate circles represent the target substances, and the smallest circles represent the partners ADDITIONAL DELIVERABLE AD6.3 66 interested in the case study. Substances marked with a star are present in several types of case studies and reflections are under way to see which mergers are possible. Figure 9. Representation of the case studies currently envisaged for general environment. Figure 10. Representation of the case studies currently envisaged for occupational exposure. ADDITIONAL DELIVERABLE AD6.3 67 Figure 11. Representation of the case studies currently envisaged for the combination of general and occupational exposures. 4.2.4. Prioritization of case studies and partner organization in working groups From the proposed list of case studies, chemical families are being prioritized for a start at the beginning of year 2 or later during the project. The prioritization choice is taking into account several elements. Firstly, priority is given to chemicals families that are part of the list prioritized in task T6.2 (section 1.2): PFAS, heavy metals, pesticides, and mycotoxins. Other criteria are also taken into account: the actual interest of the family in the other projects of T6.2; the actual interest of numerous partners in P6.2.1.b; the availability of external exposure data and human biomonitoring studies; the anticipation of interest in combining the two living environments; and the level of risk related to the magnitude of the internal exposure levels exceeding the human biomonitoring guidance values from previous projects such as HBM4EU. During, April 2023, working groups per chemical family will be set up and meetings will be organized. These working groups will bring together the partners oriented a priori in the three different types of case studies to make it possible to go as far as it is possible to combine general life and occupational exposures. Each group will be coordinated by two partner institutes, in order to represent the two environments of life: general and occupational. The objective of these working groups is first to refine the description of the case studies according to the available data and possible collaborations. More specifically, the availability of monitoring data, biomonitoring and exposure factors will be investigated. Moreover, definition of the substance(s) of interest, delimitation of the population(s) of interest, and the type of living environment (general ADDITIONAL DELIVERABLE AD6.3 68 environment, occupational environment and combining both) will be specified in each situation. Also, the strategy for combining living environments will be studied and may differ in regards of the nature of the data available. As explained above, different modelling methods are needed to reflect an association of living environments starting from characteristics measured on the whole population, and an association of living environments starting from a population characterized by an occupational activity. These elements will be set up to start the first case studies at the beginning of year 2 of PARC. 5. Conclusion This deliverable shows the first lines of the strategy that is being developed under the A6.2.1 to propose a more integrative risk assessment and management across regulatory silos. We presented main preliminary findings regarding model, data and guidance inventories, the development of aggregation methods and of the case studies. The developed strategy and its application will help to move away from a compartmentalized view of risk assessment and will provide tools to facilitate comparisons between entry routes and exposure situations and, ultimately, prioritize areas for action and prevention. This work based on concrete results will provide a better understanding of the relative contributions of sources and routes of exposure and thus will support effective risk management measures for part of PARC prioritized chemical families. ADDITIONAL DELIVERABLE AD6.3 69 6. References Ali, Nadeem, Muhammad Imtiaz Rashid, Nabil A. Alhakamy, Sultan Hassan Alamri, and Syed Ali Musstjab Akber Shah Eqani. 2023. ‘Profiling of Phthalates, Brominated, and Organophosphate Flame Retardants in COVID-19 Lockdown House Dust; Implication on the Human Health’. Science of The Total Environment 856:158779. doi: 10.1016/j.scitotenv.2022.158779. Axel Clausen, Per, Rikke L. Lindeberg Bille, Tobias Nilsson, Vivi Hansen, Bo Svensmark, and Søren Bøwadt. 2003. ‘Simultaneous Extraction of Di(2-Ethylhexyl) Phthalate and Nonionic Surfactants from House Dust: Concentrations in Floor Dust from 15 Danish Schools’. Journal of Chromatography A 986(2):179–90. doi: 10.1016/S00219673(02)02007-1. Bennett, Deborah H., Manuele D. Margni, Thomas E. McKone, and Olivier Jolliet. 2002. ‘Intake Fraction for Multimedia Pollutants: A Tool for Life Cycle Analysis and Comparative Risk Assessment’. Risk Analysis 22(5):905–18. doi: 10.1111/1539-6924.00260. Bruinen de Bruin, Yuri, Antonio Franco, Andreas Ahrens, Alick Morris, Hans Verhagen, Stylianos Kephalopoulos, Valeria Dulio, Jaroslav Slobodnik, Dick T. H. M. Sijm, Theo Vermeire, Takaaki Ito, Koki Takaki, Jonathas De Mello, Jos Bessems, Maryam Zare Jeddi, Celia Tanarro Gozalo, Kevin Pollard, Josephine McCourt, and Peter Fantke. 2022. ‘Enhancing the Use of Exposure Science across EU Chemical Policies as Part of the European Exposure Science Strategy 2020–2030’. Journal of Exposure Science & Environmental Epidemiology 32(4):513–25. doi: 10.1038/s41370-021-00388-4. Carne, G., S. Leconte, V. Sirot, N. Breysse, P. M. Badot, A. Bispo, I. Z. Deportes, C. Dumat, G. Rivière, and A. Crépet. 2021. ‘Mass Balance Approach to Assess the Impact of Cadmium Decrease in Mineral Phosphate Fertilizers on Health Risk: The Case-Study of French Agricultural Soils’. Science of the Total Environment 760. doi: 10.1016/j.scitotenv.2020.143374. Cherrie, John W., Sean Semple, Yvette Christopher, Ahsan Saleem, Graeme W. Hughson, and Andrew Philips. 2006. ‘How Important Is Inadvertent Ingestion of Hazardous Substances at Work?’ Annals of Occupational Hygiene 50(7):693–704. doi: 10.1093/annhyg/mel035. Cornelis, Christa. 2012. Case Study: Brominated Flame Retardants. ID N97E8 TAGS BDE_209 final.docx. VITO. De Brouwere, Katleen, Jurgen Buekers, Christa Cornelis, Christian E. Schlekat, and Adriana R. Oller. 2012. ‘Assessment of Indirect Human Exposure to Environmental Sources of Nickel: Oral Exposure and Risk Characterization for Systemic Effects’. Science of The Total Environment 419:25–36. doi: 10.1016/j.scitotenv.2011.12.049. De Oliveira Fernandes, Eduardo, Joana Madureira, Josefa Barrero, Otmar Geiss, and Stylianos Kephalopoulos. 2016. Monitoring and Auditing of Indoor Air Quality in European Buildings: Status and Perspectives. ADDITIONAL DELIVERABLE AD6.3 70 Delmaar, Christiaan J. E., Roel Schreurs, Martine I. Bakker, Jordi Minnema, and Bas G. H. Bokkers. 2022. ‘PACEMweb: A Tool for Aggregate Consumer Exposure Assessment’. Journal of Exposure Science & Environmental Epidemiology 1–9. doi: 10.1038/s41370022-00509-7. Descatha, Alexis, Marc Fadel, Grace Sembajwe, Susan Peters, and Bradley A. Evanoff. 2022. ‘Job-Exposure Matrix: A Useful Tool for Incorporating Workplace Exposure Data Into Population Health Research and Practice’. Frontiers in Epidemiology 2. Directorate-General for Environment (European Commission). 2020. Update on Progress and Implementation: European Union Strategic Approach to Pharmaceuticals in the Environment. LU: Publications Office of the European Union. Duffus, John H., Monica Nordberg, and Douglas M. Templeton. 2007. ‘Glossary of Terms Used in Toxicology, 2nd Edition (IUPAC Recommendations 2007)’. Pure and Applied Chemistry 79(7):1153–1344. doi: 10.1351/pac200779071153. ECHA. 2011. Guidance on Information Requirements and Chemical Safety Assessment. Part A: Introduction to the Guidance Document. Version 1.1. ECHA-2011-G-15-EN. ECHA. ECHA. 2012. REACH—Guidance on Information Requirements and Chemical Safety Assessment. Helsinki: ECHA. ECHA. 2013a. Evaluation of New Scientific Evidence Concerning DINP and DIDP in Relation to Entry 52 of Annex XVII to REACH Regulation (EC) No 1907/2006. ECHA. ECHA. 2013b. Guidance on Information Requirements and Chemical Safety Assessment. Chapter R.20: Table of Terms and Abbreviations. ECHA-12-G-26.1-EN. ECHA. ECHA. 2015. Biocides Human Health Exposure Methodology. Helsinki: ECHA. ECHA. 2016a. Guidance on Information Requirements and Chemical Safety Assessment. Chapter R.14: Occupational Exposure Assessment. Version 3.0. ECHA-2016-G-04-EN. ECHA. ECHA. 2016b. Guidance on Information Requirements and Chemical Safety Assessment. Chapter R.15: Consumer Exposure Assessment. Version 3.0. ECHA-2016-G-04-EN. ECHA. ECHA. 2016c. Guidance on Information Requirements and Chemical Safety Assessment. Chapter R.16: Environmental Exposure Assessment. Version 3.0. ECHA-16-G-03-EN. Version 3.0. ECHA-16-G-03-EN. ECHA. ECHA. 2016d. Guidance on Information Requirements and Chemical Safety Assessment. Chapter R.19: Uncertainty Analysis. Version 3.0. ECHA-16-G-03-EN. ECHA. ECHA. 2016e. Guidance on Information Requirements and Chemical Safety Assessment. Part D: Framework for Exposure Assessment. Version 2.0. ECHA-16-G-08-EN. ECHA. ADDITIONAL DELIVERABLE AD6.3 71 ECHA. 2016f. Guidance on Information Requirements and Chemical Safety Assessment. Part E: Risk Characterisation. Version 3.0. Version 3.0. ECHA-2016-G-04-EN. ECHA. ECHA. 2023a. ‘Guidance on Information Requirements and Chemical Safety Assessment’. Retrieved 17 October 2022 (https://echa.europa.eu/fr/guidancedocuments/guidance-on-information-requirements-and-chemical-safetyassessment). ECHA. 2023b. ‘Understanding REACH - ECHA’. Retrieved 23 February 2023 (https://echa.europa.eu/regulations/reach/understanding-reach). EEA and EC. 2023. ‘About HBM4EU’. HBM4EU. Retrieved 2 May 2023 (https://www.hbm4eu.eu/about-us/about-hbm4eu/). EFSA. 2012. ‘Scientific Opinion on Risk Assessment Terminology’. EFSA Journal 10(5):2664. doi: 10.2903/j.efsa.2012.2664. EFSA. 2014. ‘Scientific Opinion on the Safety Assessment of Carvone, Considering All Sources of Exposure’. EFSA Journal 12(7):3806. doi: 10.2903/j.efsa.2014.3806. EFSA. 2016. ‘Overview of Existing Methodologies for the Estimation of Non-Dietary Exposure to Chemicals from the Use of Consumer Products and via the Environment’. EFSA Journal 14(7):e04525. doi: 10.2903/j.efsa.2016.4525. EFSA, Chris Anagnostopoulos, Maria Anastassiadou, Anna Federica Castoldi, Adeline Cavelier, Tamara Coja, Federica Crivellente, Bruno Dujardin, Andy Hart, Wim Hooghe, Samira Jarrah, Kyriaki Machera, Elena Menegola, Francesca Metruccio, Christian Sieke, and Luc Mohimont. 2022. ‘Retrospective Cumulative Dietary Risk Assessment of Craniofacial Alterations by Residues of Pesticides’. EFSA Journal 20(10). doi: 10.2903/j.efsa.2022.7550. EFSA, Maria Anastassiadou, Alba Brancato, Luis Carrasco Cabrera, Lucien Ferreira, Luna Greco, Samira Jarrah, Aija Kazocina, Renata Leuschner, Jose Oriol Magrans, Ileana Miron, Ragnor Pedersen, Marianna Raczyk, Hermine Reich, Silvia Ruocco, Angela Sacchi, Miguel Santos, Alois Stanek, Jose Tarazona, Anne Theobald, and Alessia Verani. 2019. ‘Pesticide Residue Intake Model‐ EFSA PRIMo Revision 3.1’. EFSA Supporting Publications 16(3). doi: 10.2903/sp.efsa.2019.EN-1605. EFSA, Claudia Cascio, Jean Lou Dorne, Georges Kass, Davide Arcella, Marco Binaglia, Bruno Dujardin, Julia Fabrega, Claudia Heppner, and Djien Liem. 2022. ‘Theme (Concept) Paper ‐ Advancing Aggregate Exposure to Chemicals in EU (ExpoAdvance)’. EFSA Supporting Publications 19(10). doi: 10.2903/sp.efsa.2022.e201001. EFSA, Peter S. Craig, Bruno Dujardin, Andy Hart, Antonio F. Hernández‐Jerez, Susanne Hougaard Bennekou, Carsten Kneuer, Bernadette Ossendorp, Ragnor Pedersen, Gerrit Wolterink, and Luc Mohimont. 2020. ‘Cumulative Dietary Risk Characterisation of Pesticides That Have Acute Effects on the Nervous System’. EFSA Journal 18(4). doi: 10.2903/j.efsa.2020.6087. ADDITIONAL DELIVERABLE AD6.3 72 EFSA, Peter S. Craig, Bruno Dujardin, Andy Hart, Antonio F. Hernandez‐Jerez, Susanne Hougaard Bennekou, Carsten Kneuer, Bernadette Ossendorp, Ragnor Pedersen, Gerrit Wolterink, and Luc Mohimont. 2020. ‘Cumulative Dietary Risk Characterisation of Pesticides That Have Chronic Effects on the Thyroid’. EFSA Journal 18(4). doi: 10.2903/j.efsa.2020.6088. EFSA, Francesca Romana Mancini, Maria de Fátima Tavares Poças, Evelin Fabjan, Stefano Frattini, Niko Hellsten, Evgenia Stojanova, Katleen Baert, Claudia Cascio, Marios Georgiadis, Irene Munoz Guajardo, Katharina Volk, and Laurence Castle. 2022. ‘Protocol for the Exposure Assessment as Part of the Risk Assessment of Phthalates, Structurally Similar Substances and Replacement Substances Potentially Used as Plasticisers in Materials and Articles Intended to Come into Contact with Food’. EFSA Supporting Publications 19(5):7288E. doi: 10.2903/sp.efsa.2022.EN-7288. EFSA Panel on Food Additives and Flavourings (FAF), Maged Younes, Gabriele Aquilina, Laurence Castle, Gisela Degen, Paul J. Fowler, Maria Jose Frutos Fernandez, Peter Fürst, Ursula Gundert-Remy, Rainer Gürtler, Trine Husøy, Melania Manco, Wim Mennes, Peter Moldeus, Sabina Passamonti, Romina Shah, Ine Waalkens-Berendsen, Detlef Wölfle, Matthew Wright, Romualdo Benigni, Claudia Bolognesi, Polly Boon, Kevin Chipman, Joop De Knecht, Karin Nørby, Davide Arcella, Stefania Barmaz, Maria Carfì, Marcello Laganaro, Carla Martino, Alexandra Tard, Giorgia Vianello, and KarlHeinz Engel. 2022. ‘Scientific Guidance on the Data Required for the Risk Assessment of Flavourings to Be Used in or on Foods’. EFSA Journal 20(12):e07673. doi: 10.2903/j.efsa.2022.7673. EFSA Panel on Food Contact Materials, Enzymes and Processing Aids (CEP), Vittorio Silano, José Manuel Barat Baviera, Claudia Bolognesi, Andrew Chesson, Pier Sandro Cocconcelli, Riccardo Crebelli, David Michael Gott, Konrad Grob, Evgenia Lampi, Alicja Mortensen, Gilles Rivière, Inger-Lise Steffensen, Christina Tlustos, Henk Van Loveren, Laurence Vernis, Holger Zorn, Jean-Pierre Cravedi, Cristina Fortes, Maria de Fatima Tavares Poças, Ine Waalkens-Berendsen, Detlef Wölfle, Davide Arcella, Claudia Cascio, Anna F. Castoldi, Katharina Volk, and Laurence Castle. 2019. ‘Update of the Risk Assessment of Di-Butylphthalate (DBP), Butyl-Benzyl-Phthalate (BBP), Bis(2Ethylhexyl)Phthalate (DEHP), Di-Isononylphthalate (DINP) and Di-Isodecylphthalate (DIDP) for Use in Food Contact Materials’. EFSA Journal 17(12):e05838. doi: 10.2903/j.efsa.2019.5838. EFSA Panel on Food Contact Materials, Enzymes, Flavourings and Processing Aids (CEF). 2015. ‘Scientific Opinion on the Risks to Public Health Related to the Presence of Bisphenol A (BPA) in Foodstuffs’. EFSA Journal 13(1):3978. doi: 10.2903/j.efsa.2015.3978. Eichler, Clara M. A., Elaine A. Cohen Hubal, Ying Xu, Jianping Cao, Chenyang Bi, Charles J. Weschler, Tunga Salthammer, Glenn C. Morrison, Antti Joonas Koivisto, Yinping Zhang, Corinne Mandin, Wenjuan Wei, Patrice Blondeau, Dustin Poppendieck, Xiaoyu Liu, Christiaan J. E. Delmaar, Peter Fantke, Olivier Jolliet, Hyeong-Moo Shin, Miriam L. Diamond, Manabu Shiraiwa, Andreas Zuend, Philip K. Hopke, Natalie von Goetz, Markku Kulmala, and John C. Little. 2021. ‘Assessing Human Exposure to SVOCs in ADDITIONAL DELIVERABLE AD6.3 73 Materials, Products, and Articles: A Modular Mechanistic Framework’. Environmental Science & Technology 55(1):25–43. doi: 10.1021/acs.est.0c02329. EMA. 2006. ‘Environmental Risk Assessment of Medicinal Products for Human Use - Scientific Guideline’. European Medicines Agency. Retrieved 23 February 2023 (https://www.ema.europa.eu/en/documents/scientific-guideline/guidelineenvironmental-risk-assessment-medicinal-products-human-use-first-version_en.pdf). EMA. 2018a. ‘Data Submission on Authorised Medicines (Article 57)’. European Medicines Agency. Retrieved 23 February 2023 (https://www.ema.europa.eu/en/humanregulatory/post-authorisation/data-medicines-iso-idmp-standards/data-submissionauthorised-medicines-article-57). EMA. 2018b. ‘Download Medicine Data’. European Medicines Agency. Retrieved 23 February 2023 (https://www.ema.europa.eu/en/medicines/download-medicine-data). EMA. 2018c. ‘EudraVigilance’. European Medicines Agency. Retrieved 1 March 2023 (https://www.ema.europa.eu/en/human-regulatory/researchdevelopment/pharmacovigilance/eudravigilance). EMA. 2018d. ‘ICH Guideline E17 on General Principles for Planning and Design of MultiRegional Clinical Trials - Scientific’. European Medicines Agency. Retrieved 23 February 2023 (https://www.ema.europa.eu/en/ich-guideline-e17-general-principles-planningdesign-multi-regional-clinical-trials-scientific). EMA. 2018e. ‘ICH Guidelines’. European Medicines Agency. Retrieved 23 February 2023 (https://www.ema.europa.eu/en/human-regulatory/researchdevelopment/scientific-guidelines/ich-guidelines). EMA. 2018f. ‘Partners & Networks’. European Medicines Agency. Retrieved 23 February 2023 (https://www.ema.europa.eu/en/partners-networks). EMA. 2018g. ‘Pre-Authorisation Guidance’. European Medicines Agency. Retrieved 23 February 2023 (https://www.ema.europa.eu/en/human-regulatory/marketingauthorisation/pre-authorisation-guidance). EMA. 2018h. ‘Reporting Requirements for Marketing-Authorisation Holders’. European Medicines Agency. Retrieved 23 February 2023 (https://www.ema.europa.eu/en/human-regulatory/post-authorisation/datamedicines-iso-idmp-standards/reporting-requirements-marketing-authorisationholders). EMA. 2018i. ‘Research and Development’. European Medicines Agency. Retrieved 23 February 2023 (https://www.ema.europa.eu/en/human-regulatory/research-development). EMA. 2018j. ‘Revised Guideline to Assess Risk of Human Medicines for the Environment’. European Medicines Agency. Retrieved 23 February 2023 (https://www.ema.europa.eu/en/news/revised-guideline-assess-risk-humanmedicines-environment). ADDITIONAL DELIVERABLE AD6.3 80 SCCS. 2023b. Scientific Advice – Children Exposure on Methyl Salicylate (Methyl 2Hydroxybenzoate). SCCS. SCCS. 2023c. ‘Scientific Committee on Consumer Safety (SCCS)’. Health.Ec.Europa.Eu. Retrieved 23 February 2023 (https://health.ec.europa.eu/scientificcommittees/scientific-committee-consumer-safety-sccs_en). SCCS and SCHEER. 2016. Rules of Procedure. The Scientific Committees on Consumer Safety (SCCS) and Health, Environmental and Emerging Risks (SCHEER). SCCS, SCHEER. Schlüter, Urs, Jessica Meyer, Andreas Ahrens, Francesca Borghi, Frédéric Clerc, Christiaan Delmaar, Antonio Di Guardo, Tatsiana Dudzina, Peter Fantke, Wouter Fransman, Stefan Hahn, Henri Heussen, Christian Jung, Joonas Koivisto, Dorothea Koppisch, Alicia Paini, Nenad Savic, Andrea Spinazzè, Maryam Zare Jeddi, and Natalie von Goetz. 2022. ‘Exposure Modelling in Europe: How to Pave the Road for the Future as Part of the European Exposure Science Strategy 2020–2030’. Journal of Exposure Science & Environmental Epidemiology 32(4):499–512. doi: 10.1038/s41370-022-00455-4. Schossler, Patrícia, Tobias Schripp, Tunga Salthammer, and Müfit Bahadir. 2011. ‘Beyond Phthalates: Gas Phase Concentrations and Modeled Gas/Particle Distribution of Modern Plasticizers’. Science of The Total Environment 409(19):4031–38. doi: 10.1016/j.scitotenv.2011.06.012. Spankie, Sally, Anne Sleeuwenhoek, Martie van Tongeren, and Denis Sarigiannis. 2012. TAGS. Realistic Estimation of Exposure to Substances from Multiple Sources. The Tiered Aggregate Exposure Assessment. B5 - CEFIC Long-Range Research Initiative Funded Project. Chlorpyrifos Case Study. CEFIC, IOM, CERTH, VITO. Swartjes, Frank A. 2007. ‘Insight into the Variation in Calculated Human Exposure to Soil Contaminants Using Seven Different European Models’. Integrated Environmental Assessment and Management 3(3):322–32. doi: 10.1002/ieam.5630030303. Tebby, Cleo, Hilko van der Voet, Georges de Sousa, Emiel Rorije, Vikas Kumar, Waldo de Boer, Johannes W. Kruisselbrink, Frédéric Y. Bois, Moosa Faniband, Angelo Moretto, and Céline Brochot. 2020. ‘A Generic PBTK Model Implemented in the MCRA Platform: Predictive Performance and Uses in Risk Assessment of Chemicals’. Food and Chemical Toxicology 142:111440. doi: 10.1016/j.fct.2020.111440. van Tongeren, Martie, John Cherrie, Sally Spankie, Christa Cornelis, Katleen de Brouwere, Rudi Torfs, Anastasios Karabelas, and Spyros Karakitsios. 2011. Structured Criteria and Process for Launching an Aggregate Exposure Assessment. project deliverable. IOM. US EPA. 1989. Report to Congress on Indoor Air Quality - Volume Ill: Indoor Air Pollution. Research Needs Statement. EPA/400/1-89/001 D. US EPA, Graham Glen, Valerie Zartarian, Luther Smith, and Jiangping Xue. 2012. The Stochastic Human Exposure and Dose Simulation Model for Multimedia, Multipathway Chemicals (SHEDS-Multimedia): Residential Module. SHEDS-Residential Version 4. Technical Manual. US EPA. ADDITIONAL DELIVERABLE AD6.3 81 US EPA, ORD. 2011. ‘Exposure Factors Handbook Glossary’. Retrieved 2 May 2023 (https://www.epa.gov/sites/default/files/2015-09/documents/efh-glossary.pdf). US EPA, ORD. 2015. ‘About the Exposure Factors Handbook’. Retrieved 23 February 2023 (https://www.epa.gov/expobox/about-exposure-factors-handbook). US EPA, ORD, and CPHEA. 2003. Framework for Cumulative Risk Assessment. EPA/600/P02/001F. Washington Office, Washington, DC: US EPA. US EPA, Jianping Xue, Valerie Zartarian, and Steve Nako. 2012. The Stochastic Human Exposure and Dose Simulation Model for Multimedia, Multipathway Chemicals (SHEDSMultimedia): Dietary Module. SHEDS-Dietary Version 1. Technical Manual. US EPA. US EPA, Valerie Zartarian, Graham Glen, Luther Smith, and Jianping Xue. 2008. Stochastic Human Exposure and Dose Simulation Model for Multimedia, Multipathway, Chemicals. SHEDS-Multimedia Model. Version 3. Technical Manual. US EPA. Van Holderbeke, Mirja, Tine Fierens, Arnout Standaert, Christa Cornelis, Céline Brochot, Philippe Ciffroy, Erik Johansson, and Johan Bierkens. 2016. ‘Assessing Multimedia/Multipathway Exposures to Inorganic Arsenic at Population and Individual Level Using MERLIN-Expo’. Science of The Total Environment 568:794–802. doi: 10.1016/j.scitotenv.2016.04.038. Vanacker, Marie, Paul Quindroit, Karine Angeli, Corinne Mandin, Philippe Glorennec, Céline Brochot, and Amélie Crépet. 2020. ‘Aggregate and Cumulative Chronic Risk Assessment for Pyrethroids in the French Adult Population’. Food and Chemical Toxicology 143:111519. doi: 10.1016/j.fct.2020.111519. Vanacker, Marie, Jessica Tressou, Guillaume Perouel, Philippe Glorennec, and Amélie Crépet. 2020. ‘Combining Data from Heterogeneous Surveys for Aggregate Exposure: Application to Children Exposure to Lead in France’. Environmental Research 182:109069. doi: 10.1016/j.envres.2019.109069. Vermeire, T. G., D. T. Jager, B. Bussian, J. Devillers, K. den Haan, B. Hansen, I. Lundberg, H. Niessen, S. Robertson, H. Tyle, and P. T. J. van der Zandt. 1997. ‘European Union System for the Evaluation of Substances (EUSES). Principles and Structure’. Chemosphere 34(8):1823–36. doi: 10.1016/S0045-6535(97)00017-9. van der Voet, Hilko, Johannes W. Kruisselbrink, Waldo J. de Boer, Marco S. van Lenthe, J. J. B. (Hans) van den Heuvel, Amélie Crépet, Marc C. Kennedy, Johanna Zilliacus, Anna Beronius, Emiel Rorije, Corinne Sprong, and Jacob D. van Klaveren. 2019. ‘The EuroMix Model Toolbox MCRA 9’. doi: 10.5281/ZENODO.3462182. van der Voet, Hilko, Johannes W. Kruisselbrink, Waldo J. de Boer, Marco S. van Lenthe, J. J. B. (Hans) van den Heuvel, Amélie Crépet, Marc C. Kennedy, Johanna Zilliacus, Anna Beronius, Cleo Tebby, Céline Brochot, Claudia Luckert, Alfonso Lampen, Emiel Rorije, Corinne Sprong, and Jacob D. van Klaveren. 2020. ‘The MCRA Toolbox of Models and Data to Support Chemical Mixture Risk Assessment’. Food and Chemical Toxicology 138:111185. doi: 10.1016/j.fct.2020.111185. ADDITIONAL DELIVERABLE AD6.3 82 Wei, Wenjuan, Olivier Ramalho, and Corinne Mandin. 2019. ‘A Long-Term Dynamic Model for Predicting the Concentration of Semivolatile Organic Compounds in Indoor Environments: Application to Phthalates’. Building and Environment 148:11–19. doi: 10.1016/j.buildenv.2018.10.044. WHO, IPCS. 2008. Uncertainty and Data Quality in Exposure Assessment. World Health Organization. WHO, IPCS. 2009. Assessment of Combined Exposures to Multiple Chemicals: Report of a WHO/IPCS International Workshop on Aggregate/Cumulative Risk Assessment. IPCS Harmonization Project Document No. 7. ISBN: 9789241563833. WHO, IPCS. 2011. ‘Risk Assessment of Combined Exposure to Multiple Chemicals: A WHO/IPCS Framework’. Regulatory Toxicology and Pharmacology 60(2, Supplement):S1–14. doi: 10.1016/j.yrtph.2011.03.010. Wild, Christopher Paul. 2005. ‘Complementing the Genome with an “Exposome”: The Outstanding Challenge of Environmental Exposure Measurement in Molecular Epidemiology’. Cancer Epidemiology, Biomarkers & Prevention 14(8):1847–50. doi: 10.1158/1055-9965.EPI-05-0456. Zartarian, Valerie, Jianping Xue, Rogelio Tornero-Velez, and James Brown. 2017. ‘Children’s Lead Exposure: A Multimedia Modeling Analysis to Guide Public Health DecisionMaking’. Environmental Health Perspectives 125(9):097009. doi: 10.1289/EHP1605. Zenié, Alexandre, and Vittorio Reina. 2007. ‘ExpoFacts Database’. ADDITIONAL DELIVERABLE AD6.3 83 Table S1. Short name and extended name of the tasks/activities/projects cited in the document Short name of the PARC task/activity/project Extended name of the PARC task/activity/project T4.1 Task 4.1: Human biomonitoring T4.2 Task 4.2: Environmental and multisources monitoring T6.2 Task 6.2: Integrative exposure and risk assessment A6.2.1 Activity 6.2.1: Aggregated exposure assessment from multiple sources and routes for general population and workers P6.2.1.a Project 6.2.1.a: Developing, performing and validation of source to dose modelling including selected case studies P6.2.1.b Project 6.2.1.b: Strategy for aggregate exposure from general life and occupational exposures A6.2.2 Activity 6.2.2: Modelling exposure through life P6.2.2.a Project 6.2.2.a: Refinement and development of PBPK models for human risk assessment A6.2.3 Activity 6.2.3: Mixture exposure and risk assessment A6.2.4 Activity 6.2.4: Human health impact assessment and risk indicators T7.3 Task 7.3: Innovative analyses (in particular innovative, methods for combined analyses of FAIR data, for uncertainty management, and for extracting knowledge from non-structured data) T8.3 Task 8.3: Integrative models Appendix Appendix 1: Short name and extended name of the tasks/activities/projects cited in the document ADDITIONAL DELIVERABLE AD6.3 84 These inventories are available on request to the responsible authors. Table S2 summarises the descriptors, and description of descriptors used in model inventories. Figure S1 is a simplified representation of this inventory. Figure S2 is a simplified representation of the inventory of aggregate exposure models. Table S2. Descriptors, and description of descriptors used in model inventories. Descriptor Explanation of the descriptor Who completed this line? Names and affiliations of the people who completed the row for a specific model. Short name of the model Short name of the model, often an acronym. Full name of the model Full name of the model. URL link or useful reference URL link or useful reference. Owner (country, organisation) Owner (country, organisation). Source-to-dose modelling? Enter if it is source-to-dose modelling. If so, specify which type of modelling: rather indoor, outdoor, or consumer product related. General description A general description of the model, to get an overview of its capabilities. Target substance(s) Substances within the scope of the model. Mixtures? If several substances are considered, indicate 'yes', otherwise indicate 'no'. Multiple exposure sources? If multiple sources of exposure, indicate 'yes', otherwise indicate 'no'. Exposure sources Macro-criterion subdivided into several criteria corresponding to different possible exposure sources: water, soil, food, dust, etc. Main emission sources Description of the main emission sources. Relevant for source-to-dose models. Multiple routes? If multiple routes of exposure, indicate 'yes', otherwise indicate 'no'. Exposure routes Macro-criterion subdivided into several criteria corresponding to different possible exposure routes: ingestion, dermal, inhalation. Targeted population groups Macro-criterion subdivided into several criteria corresponding to different possible target populations: adults, adolescents, children. Activity-related exposure Macro-criterion subdivided into several criteria corresponding to different possible activity-related exposure: workers, bystanders/residents, consumers. Application of model Macro-criterion subdivided into several criteria corresponding to different possible applications of the model/tool. Exposure duration Macro-criterion subdivided into several criteria corresponding to different possible exposure duration: acute, subchronic, chronic. Modelling approach Description of the modelling approach (probabilistic or deterministic modelling?, capturing uncertainty and variability?, managing censorship?, consideration of time?, etc). To be further detailed for inventories modelling methods with an explicit source-aggregation-by-route strategy. Description of the inputs Description of the inputs. Description of the outputs Description of the outputs. Appendix 2: Overview of the descriptors and the structure of the model inventory ADDITIONAL DELIVERABLE AD6.3 85 Internal consistency: model calibration, validation or evidence-based parameters Are the parameters interpretable when relevant, or are they derived from a calibration process when necessary. Applicability domain Applicability domain: a target population? spatial coverage? temporal coverage? Target substances? Availability Availability of the model: free access, free license, not free, etc. Usability of input data Input data file encoded in a transparent way?, can be created, written and edited by all software intended to process the type of file? Usability of output data Output data file encoded in a transparent way?, can be created, written and edited by all software intended to process the type of file? Accessibility to the model (insert URL if available) Accessibility to the model: where to download, or where to use it? User friendliness User friendliness: is it easy to handle? Disclosure of the model description (equation/formula, parameter values, etc.) Is the model well described? Critical aspects regarding the link to the MCRA toolbox Is the model included in MCRA? What are the opportunities if not? Programming environments Programming environments. Level of maintenance Level of maintenance. Last update Last update. Language Language used in the program and in the documentation. Data input and Confidential Business Information Does the information inputted into the model remain confidential to the user (i.e. stays protected and is not made available to others)? Robustness Robustness: Are there examples of the use of the model in the literature? Has the model undergone a verification process, a peer-review? Remarks Additional information. Secondary source Secondary sources of information: other inventories, other SharePoint sheets. ADDITIONAL DELIVERABLE AD6.3 86 Descriptor 1 Descriptor 2 Descriptor 3 … Descriptor n Summary of the content of the inventories: 44 outdoors source-to-dose models. ← description of the first model 37 indoor source-to-dose models. … 19 models without a source-to-dose module. ← description of the last model 45 models consider the dietary ingestion route, 44 the non dietary ingestion routes, 51 the dermal route, and 59 consider the inhalation route. Sheet 1 Sheet 2 Sheet 3 Descriptor 1 Descriptor 2 Descriptor 3 … Descriptor n 63 models and tools. ← description of the first model 26 models out of 76, partly or totally concern the dermal route, and 27 out of 76 partly or totally concern the inhalation route. … ← description of the last model Sheet 1 Sheet 2 Sheet 3 Descriptor 1 Descriptor 2 Descriptor 3 … Descriptor n 66 outdoor source-to-dose models. ← description of the first model 25 indoor source-to-dose models. … ← description of the last model Sheet 1 Sheet 2 Sheet 3 Figure S1. Simplified representation of the three sub-inventories of the exposure model inventory. Sheet 1' refers to human exposure models, 'Sheet 2' refers to occupational exposure models, and 'Sheet 3' refers to 'environmental' models, not having human exposure as an output. Model name 1rst model … last model Model name 1rst model … last model Descriptor 1 Descriptor 1 … … Descriptor n Descriptor n table 1 table 2 Figure S2. Simplified representation of the structure of two inventories aggregate exposure models inventories. ‘Table 1’ refers to general life exposure models, and ‘Table 2’ refers to occupational exposure models. ADDITIONAL DELIVERABLE AD6.3 87 This inventory is available on request to the responsible authors. Table S3 summarises the descriptors, and description of descriptors used in the data inventory. Figure S3 is a simplified representation of this inventory. Table S3. Descriptors, and description of descriptors used in data inventory. Descriptor Explanation of the descriptor Who completed this line? This descriptor must be completed. Names and affiliations of the people who completed the line. This makes it possible to trace changes, and to know who to attribute them to. Quick insight of the data type This descriptor must be completed. One of the following types is possible: * Contamination data * Internal human exposure data * External human exposure data (mainly relevant for JEMs, in occupational exposure assessment) * Exposure factors Short title of the database This descriptor must be complete if available. It is a short title of the database, often an acronym. Full title of the database This descriptor must be completed. It is the complete title of the database. URL link to any useful information about the database This descriptor must be completed if available. URL link or useful reference. Database producer This descriptor must be completed. The producer of the database: country, organisation. General description This descriptor must be completed. A general description of the database, to get an overview of the data and the quality of the metadata. Description of the spatial coverage and granularity This descriptor must be completed. Description of the spatial coverage and granularity: country, spatial scale, etc. Description of the temporal coverage and granularity This descriptor must be completed. Description of the temporal coverage and granularity: sampling times, data reporting dates, etc. Individual or aggregated data? This descriptor must be completed. Are these statistics summarising the data, or are they raw data, or other types of data transformation? Number of individuals or records in the database This descriptor must be completed. Number of statistical units in the database, measured quantities, and finally, number of records. Target substances This descriptor must be completed if relevant. Specify the targeted substances, where relevant. Sub-descriptors per substance are present in the inventory, and not reported in this table. Contamination data, by source of exposure This descriptor must be completed if relevant. This descriptor is detailed in sub-descriptors, not reported in this table. This descriptor concerns the sources of exposure. If relevant, indicate useful information about the specific origin of these data, if not redundant with the general description of the data: study, places, number of analyses, duration of the study. Appendix 3: Overview of the descriptors and the structure of the data inventory ADDITIONAL DELIVERABLE AD6.3 88 Biomonitoring data (internal exposure) This descriptor must be completed if relevant. This descriptor is detailed in sub-descriptors, not reported in this table. This descriptor concerns the biomonitoring. If relevant, indicate useful information about the specific origin of these data, if not redundant with the general description of the data: study, places, number of analyses, duration of the study. These data are already inventoried in PARC/A6.2.3. The present inventory refers to this inventory. If partners wish to share data not present in the A6.2.3. inventory, then they can put it in P6.2.1.b inventory. External human exposure data (mainly relevant for JEMs, in occupational exposure assessment) This descriptor must be completed if relevant. This descriptor is detailed in sub-descriptors, not reported in this table. This descriptor concerns the external exposure data gathered in job-exposure matrices. If relevant, indicate useful information about the specific origin of these data, if not redundant with the general description of the data: study, places, number of analyses, duration of the study. These data are already inventoried by the OMEGA-NET project. The present inventory refers to this inventory. If partners wish to share data not present in the OMEGA-NET inventory, then they can put it in P6.2.1.b inventory. Exposure factors This descriptor must be completed if relevant. This descriptor is detailed in sub-descriptors, not reported in this table. This descriptor concerns exposure factors: generic physiological parameters, activity-related parameters, and consumer products usages factors. Targeted population groups This descriptor must be completed if relevant. This descriptor is detailed in sub-descriptors, not reported in this table. This descriptor concerns the targeted population groups. Activity-related exposure This descriptor must be completed if relevant. This descriptor is detailed in sub-descriptors, not reported in this table. This descriptor concerns the activity-related exposure: worker, bystanders/residents, and consumers. Exposure duration This descriptor must be completed if relevant. This descriptor is detailed in sub-descriptors, not reported in this table. This descriptor concerns the exposure duration: acute, subchronic, and chronic. Availability This descriptor must be completed. This descriptor is detailed in subdescriptors, not reported in this table. This descriptor concerns the availability of the database: freely available, fee-based licence, free licence, etc. Data accessibility (insert URL if available) This descriptor must be completed. This descriptor is detailed in subdescriptors, not reported in this table. This descriptor concerns the data accessibility: data are web-based? Are downloadable? Etc. Disclosure of the dataset description This descriptor must be completed. This descriptor is detailed in subdescriptors, not reported in this table. This descriptor concerns the disclosure of the data description. Are they described in a published document? Is it confidential? Etc. ADDITIONAL DELIVERABLE AD6.3 89 Quality of the metadata This descriptor must be completed. This descriptor is detailed in subdescriptors, not reported in this table. This descriptor concerns the quality of the metadata: Is the description of the data of good quality? Is the frequency of updates filled in and respected? Is the license filled in? Is there at least one resource with a declared open format? Is the spatial coverage filled in? Is the spatial granularity filled in? Is the temporal coverage filled in? Is the temporal granulometry filled in? Level of maintenance This descriptor must be completed. This descriptor concerns the level of maintenance. Last update This descriptor must be completed. This descriptor concerns the last update. Language This descriptor must be completed. This descriptor concerns the language of the database, the metadata, and the possible documentation. Data input and Confidential Business Information (CBI) must be completed. Does the information inputted into the model remain confidential to the user (i.e. stays protected and is not made available to others)? Remarks This descriptor could be completed. It concerns additional remarks. Descriptor 1 Descriptor 2 Descriptor 3 … Descriptor n Constraints Constraints Constraints Constraints Constraints ← description of the first dataset … ← description of the last dataset Figure S3. Simplified representation of the structure of the data inventory