2. Water quality monitoring for endocrine disrupting chemicals: from traditional chemical analysis to effect-based monitoring

To manage endocrine disruptors in water, well-designed monitoring programmes can support policy action. While water quality monitoring usually focusses on detecting a shortlist of substances (substance-by-substance monitoring), this approach is insufficient to address endocrine disruption. As described in Chapter 1, endocrine disrupting chemicals (EDCs) are found in various classes of chemicals (e.g., pesticides, pharmaceuticals, packaging, steroids) and it is impossible to monitor each and every potential EDC. Moreover, only few chemicals are currently identified or suspected as EDCs, even though effects may be observed in freshwater organisms, posing a challenge for the selection of chemicals to monitor on a substance-by-substance basis. The problematics of EDCs call for additional monitoring approaches.

One emerging solution are effect-based methods (EBM). EBMs are increasingly applied for water monitoring in research since the 2000s (Escher, Neale and Leusch, 2021[1]; Di Paolo et al., 2016[2]; Fairbrother et al., 2019[3]; Robitaille et al., 2022[4]; Wernersson et al., 2015[5]). EBM is achieved through bioanalytical assays or bioassays that detect the activity – or effect – of water samples in organisms, embryos, tissues, or cells. If a change occurs in the bioassay, it indicates the presence of chemical(s) which can generate that change. Bioassays exist to detect various types of endocrine activity.

In 2018, California has formalised the use of cell bioassays as a water quality policy tool, including bioassays which test for estrogenicity (California State Water Board, 2018[6]). Moreover, bioassays are used by utilities, water authorities and industries across the world, usually as a screening tool to detect endocrine disruptive effects. The European Commission is also considering extending the Water Framework Directive (EU, 2000[7]) to include the regulation of endocrine effects (European Commission, 2022[8]).

This chapter inventories prevailing and promising techniques to monitor and assess endocrine activity and endocrine disruption through chemical and biological analysis. An overview of these methods is given in Figure 2.1. The chapter also addresses ways to validate the results obtained in monitoring, either to confirm the hazard, to identify the culprit substance, or to identify pollution hotspots. The chapter also highlights the enabling factors, including threshold values, funding, laboratory access and sampling, to advance effect-based monitoring. The last section of this chapter gives a brief overview of the barriers and uncertainties that may challenge the wide use of effect-based monitoring for policy development. It also argues that decision-makers must accept a certain level of uncertainty when developing policy responses.

This section defines and assesses targeted and non-targeted chemical analyses to monitor endocrine-disrupting compounds in water.

Water quality management is typically done by the development of threshold values for concentrations of single chemicals found in freshwater, such as environmental quality standards, water quality criteria, or environmental norms. Targeted chemistry gives a direct conclusion on compliance with regulation: the chemical concentration is either below or above the threshold level. Countries develop threshold levels based on the available knowledge of the toxicity and the exposure levels of a chemical with the objective of protecting human and/or ecosystem health. The enforcement of those standards is done via classical single-chemical monitoring. In this type of monitoring, targeted chemical analysis is used to determine the concentration of an individual chemical of interest in a water sample. The concentration is then compared to the associated standard.

Targeted analysis can also be useful to monitor known highly active EDCs for which no quality standard exists. For example, EPA Victoria in Australia has conducted two monitoring campaigns on emerging contaminants in wastewater using targeted chemistry (Box 2.1). Data of such targeted analysis can be instrumental in linking the effects observed in bioassays to the culprit compounds (Section 2.3.1).

While targeted chemistry is widely used for various chemicals for water management, it is currently a rather limited approach within the EDC context for several reasons:

  1. 1. Currently only a few EDCs are covered in regulatory monitoring programmes. For example, the European Union Water Framework Directive (EU, 2000[7]) has Environmental Quality Standards for several suspected or conformed EDCs, such as Di(2-ethylhexyl)- phthalate (DEHP), nonylphenols, octylphenols, tributyltin compounds, perfluorooctane sulfonic acid and its derivatives (PFOS), brominated diphenyl ethers and hexabromo cyclododecane (HBCDD). However, this is only a small fraction of the more than 100 EDCs listed as identified, under evaluation, or considered as EDC on the platform Endocrine Disruptors Lists (edlists.org, n.d.[18])1. Moreover, endocrine disruptive substances monitored may not necessarily be the highest-potency substances. The scarcity of available standards can be linked to a lack of formal identification of EDCs, to insufficient data for their development and to inadequate methods for including endocrine endpoints (Chapter 3, Section 3.3.1). Despite this limitation in the regulatory context, targeted chemical analysis is still useful in monitoring known substances for which no quality standard exists.

  2. 2. EDCs can cause effects at very low concentrations (below ng/L). However, current chemistry analyses for common EDCs have limits of detection higher than the range required to evaluate their risk (see the example in Box 2.1). Hence, the available methods are ill-suited for the required risk assessment. Efforts are made to decrease the limits of detection, increase accuracy, and streamline sample processing (Metcalfe et al., 2022[19]). One example of such progress is the development of a method to detect steroids and bisphenols at levels as low as 0.1-0.5 ng/L (Goeury et al., 2022[20]). However, it will take time for those methods to be standardised and made accessible globally.

  3. 3. Chemical monitoring is a top-down approach that only scratches the surface of the problem (WHO-UNEP, 2013[21]) as an analysis of the wide array of chemicals present in an environmental sample is expensive and fundamentally impossible. This is due to limits in our knowledge of all existing chemicals (“unknown unknowns”) - including breakdown and transformation products (Hecker and Hollert, 2009[22]).

  4. 4. Targeted chemical analyses do not address mixture effects (Brack et al., 2019[23]). Chemistry data is compared to individual standards and overlooks the risks posed by chemical mixtures. Bioassays can capture mixtures (Wernersson et al., 2015[5]) (Section 2.3.1).

As mentioned above, only a few chemicals are assessed in routine water monitoring programmes. It is estimated that only 5% of all known chemicals are monitored using targeted analyses (McCord, Groff and Sobus, 2022[24]). To address this issue, non-targeted analyses (NTAs) are increasingly used. Like the name indicates, NTAs do not have necessarily pre-defined target chemicals. Rather they aim to identify all chemicals present in an environmental sample, without quantifying the concentration of each chemical detected.

NTAs can analyse “known unknowns” and “unknown unknowns”. Most NTAs analyse “known unknowns”, which are chemicals of which at least the structure is classified in databases and of which some toxicity data is available. NTAs aim to include chemicals that are not yet regulated or routinely monitored. Those analyses are often referred to as suspect screening analyses (SSA) (Paszkiewicz et al., 2022[25]). SSA can also include the quantification of selected chemicals. NTAs can also look at “unknown unknowns” for which not even the molecular structure is clearly defined or registered in databases (Paszkiewicz et al., 2022[25]). High-resolution mass spectrometry (HRMS) is the typical method of choice for any type of NTA (McCord, Groff and Sobus, 2022[24]; Paszkiewicz et al., 2022[25]).

NTA is a useful screening tool to map EDCs and other chemicals present in water (McCord, Groff and Sobus, 2022[24]; Hollender et al., 2019[26]). Such methods are useful in developing a baseline or archive of the chemical composition of a water sample, in detecting accidental spills, in capturing (synthetic) EDCs that cannot yet be detected by bioassays, and in analysing mixtures of chemicals.

NTAs could help track the impact of pollution sources by looking at their specific fingerprint instead of by surveying specific chemicals (Brack et al., 2019[27]).For example, samples were analysed with NTAs at multiple sites of River Holtemme in Germany. By clustering the acquired data, researchers were able to identify patterns of chemicals specific to their sources of contamination, such as wastewater treatment plant (WWTP) effluents. The research even identified the contribution of each WWTP to the pollution in a section of the river (Beckers et al., 2020[28]).

Furthermore, NTA can provide a good digital record of chemical pollution over time (Alygizakis et al., 2019[29]; Hollender et al., 2019[26]). This can be used for retrospective analysis even for contaminants which were not of concern as endocrine active at the time of the measurement. Keeping records of NTAs can also help evaluate the evolution of pollution through time to see for example if a contaminant is ubiquitous (i.e., present all the time), if new contaminants were introduced, or if contaminants detected in prior studies disappeared. This information could be used in the long term to prioritise action for new and ubiquitous contaminants, as well as assessing the impact of remediation action (Brack et al., 2019[27]; Hollender et al., 2019[26]). The Norman Network has kept NTA records in a Digital Sample Freezing Platform (Norman Network, n.d.[30]).

NTA technologies are evolving into automated routine monitoring systems. This is exemplified by the case study of the International Rhine Monitoring station in Switzerland (Box 2.2). The automation of the workflow enables the station to monitor water quality daily with NTA. NTA has identified accidental spills and alerted drinking water treatment stations downstream. NTA has also led to mitigation action in a manufacturing company after the detection of a continuously released hazardous compound.

Except when standards are used for SSA, most NTAs cannot be used in risk-based regulation as quantification remains a challenge. Still, NTAs could be useful in hazard-based regulation as only the presence of the chemical is sufficient to justify action (McCord, Groff and Sobus, 2022[24]). Hence, if an authority decides to adopt a hazard-based approach to EDCs, with a zero tolerance to EDCs present in a water sample, NTAs could be applied to detect the presence of substances. However, depending on the cost per sample, targeted chemical analysis may be more cost-effective.

While NTAs might not be readily useful for regulation, NTAs can be used to prioritise EDCs and other chemicals. For example, prioritisation of site-specific contamination can be done by looking at the rarity of a chemical in water, such as demonstrated in a German study (Krauss et al., 2019[32]). Moreover, the Environmental Agency (EA) of England, United Kingdom, is investigating how to integrate NTAs in their Prioritisation and Early Warning System (PEWS) for chemicals (Sims, 2022[33]).

While NTA, combined with other methods, has a strong potential in the future monitoring of EDCs, there are still a lot of limitations for their use for regulatory purposes around the world.

  1. 1. NTAs are still mainly qualitative (McCord, Groff and Sobus, 2022[24]; Hollender et al., 2019[26]), as the concentration of each chemical cannot be determined, with the exception of standards used for SSA. Otherwise, only relative quantification can be done. Research efforts are being done to allow quantification for the purpose of risk assessment using surrogate standards or modelling responses based on chemical structure (McCord, Groff and Sobus, 2022[24]). Until those methods are mainstreamed, non-targeted chemistry can be used for pre-screening, setting a water quality baseline of known and unknown substances present in water, and prioritisation. NTAs could also be useful in the context of hazard-based approaches that do not tolerate any presence of certain substances, though chemical analysis might be more cost-effective for these purposes.

  2. 2. NTA is not standardised, time-consuming and requires analytical expertise (McCord, Groff and Sobus, 2022[24]; Paszkiewicz et al., 2022[25]), which makes those methods more difficult to apply on a regular basis. To use NTAs for regulatory purposes, there is first a need for standardisation and harmonisation of methods to ensure the quality of data (McCord, Groff and Sobus, 2022[24]; Luo et al., 2022[34]; Hollender et al., 2019[26]). Efforts are also made to make the technology quicker and more accessible (e.g., price and expertise requirement) (Hollender et al., 2019[26]). There is a need for automation of the data processing for high-throughput analysis (McCord, Groff and Sobus, 2022[24]).

  3. 3. There is a growing need to develop databases for sharing NTA data to enable their comparison, retrospective analysis, facilitate technical support by experts and increase international collaboration (Hollender et al., 2019[26]). Some databases already exist, such as the Global Natural Products Social Molecular Networking (GNPS) (Wang et al., 2016[35]) or the Digital Sample Freezing Platform (DSFP) from the Norman Network (Alygizakis et al., 2019[29]). Data acquisition needs to be harmonised to facilitate data submission and data comparison. Organisations such as the International Commission for the Protection of the River Rhine (ICPR) are working towards that goal (Hollender et al., 2019[26]).

  4. 4. Most NTAs concentrate on known chemicals for which the chemical structure has at least been identified. There is a need to increase spectra identification to increase the information available in databases such as the NORMAN MassBank (NORMAN Network, n.d.[36]). However, some chemicals might not be detected and efforts need to be put in place to improve the method to enable the discovery of new chemicals (Escher, Stapleton and Schymanski, 2020[37]).

This section presents and discusses the advantages and disadvantages of bioassays and in situ wildlife monitoring, two biological approaches that can be used to monitor the adverse effects of EDCs in water.

A promising approach to solve the issues linked to chemical monitoring for EDC risk assessment in freshwater is effect-based monitoring or effect-based methods (EBM). Like the name suggests, this monitoring approach is based on the detection and quantification of effects caused by chemicals found in a sample (Brack et al., 2019[23]). This type of monitoring uses bioanalytical methods, or bioassays. Bioassays are biological test methods performed using in vitro (cell-based or cell-free) or in vivo (whole organism) models to detect effects in a concentration-dependent manner on toxicological endpoints of concern (Brack et al., 2016[38]; Robitaille et al., 2022[4]). They consist of testing the biological activity of a sample using responses of (sub)cellular systems or whole organisms (Brack et al., 2016[38]). Box 2.3 contains a simple explainer of bioassays.

If a bioassay (that measures an endocrine mode of action) responds to a water sample, it indicates potential endocrine activity in the water sample. Bioassays do not directly identify the chemical triggering the activity, but they provide signal that there is a potential concern. Bioassays are often, but not always, more sensitive than chemical analysis. There is a high correlation between results found in bioassays and chemical measurements, indicating that both methods agree on the overall endocrine potential of samples (Könemann et al., 2018[39]; Escher, Neale and Leusch, 2021[40]). However, bioassays and chemistry do not correlate well at low concentrations because, first of all, bioassays can detect activity below the limit of detection (LOD) of chemical methods, and second, bioassays can detect mixtures from chemicals that are individually below their LOD.

Bioassays can be used regardless of any prior knowledge on the chemical composition of the water sample. Any chemical, known and unknown EDCs, that triggers an activity in a bioassay could be detected. Furthermore, bioassays will inform on the activity of chemical mixtures found in the sample. Since mixtures are still characterised by uncertainty, their identification is a critical added value of bioassays (Box 2.4).

For endocrine activity and endocrine disruption, bioassays are designed to detect endocrine-specific endpoints (Table 2.1). The most studied endpoints are the EATS modalities: Estrogen, Androgen, Thyroid and Steroidogenesis. Estrogen modalities are well studied. Modalities for invertebrates are also gaining traction: Juvenile Hormones (Jh) and ecdysteroids (Ec) (OECD, 2018[50]). Thyroid disruption is notably known for disrupting metamorphosis in amphibians. Effects on the glucocorticoid receptor and transthyretin (TTR) displacement have also been observed in freshwater, but these effects are less well studied (OECD, 2022[51]). For water testing, the most common endpoint evaluated involves the interaction of chemicals with hormone receptors, especially nuclear receptors for:

  • Estrogen (ER),

  • Androgen (AR),

  • Thyroid hormones (TR),

  • Progesterone (PR),

  • Glucocorticoids (GR).

Other bioassays look at the synthesis of hormones (steroidogenesis assays) or at the hormone transport in blood (transthyretin binding assay) (Robitaille et al., 2022[4]). In whole organisms, endpoints such as fecundity, growth, metamorphosis for amphibians and biomarkers (e.g. vitellogenin, female egg yolk precursor) can be measured (Table 2.1).

Bioassays can also be informative in risk assessment and thresholds similar to chemical standards can be developed. These types of thresholds are generally referred to as effect-based trigger values (EBT) (Escher et al., 2018[56]). Effect-based trigger values are the threshold values, or water quality indicators, for bioassays. EBTs help interpret whether the effects detected in a bioassay are acceptable or not (Neale et al., 2023[57]). More information on setting EBTs for bioassays is given in Section 2.6.1.

For water quality monitoring, it is recommended to use a set of different bioassays to obtain a complete picture of the different effects present in a water sample (Neale, Leusch and Escher, 2020[58]). After all, a single bioassay can only detect one or a few modalities, whereas a set of bioassays - applied at the same time, covering multiple modalities or endocrine endpoints - make the water quality assessment more comprehensive. A set of bioassays is referred to as a “battery of bioassays”. There is no standard recommendation for a battery of bioassays, and different methods are used by various countries. It should be noted that, generally, batteries of bioassays comprise more effects than endocrine activity, depending on the monitoring purpose (Escher, Neale and Leusch, 2021[40]). Some suggest that a minimal battery of bioassays should include testing for ER, AhR and oxidative stress, adding genotoxicity in drinking water research (Escher et al., 2014[59]; Neale et al., 2022[60]; Rosenmai et al., 2018[61]).

While the interest in bioassays for water quality monitoring is growing, bioassays largely remain non-standardised tools, except for several whole organism tests that are not favoured for routine water quality monitoring due to concerns related to animal testing. This situation hinders their widespread adoption for water quality regulation and policy. Gaps that hinder the mainstreaming of effect-based monitoring approaches are the following:

  1. 1. Effect-based trigger values (EBT) need to be in place to determine the level of risk of each observed effect. However, most bioassays do not have a harmonised or internationally agreed standard or trigger value that determines to what extent the observed effect is (potentially) harmful (Escher et al., 2018[56]). This remains up to the discretion of individual water authorities, academia, industries, and bioassay developers. This gives rise to a patchwork of trigger values and diagnostic tools. Moreover, it is currently dependent on the formal identification of EDCs which can be a long and tedious task. Sections 2.6.1 discusses effect-based trigger values in more detail.

  2. 2. There is a lack of standardisation for bioassay methods, sample collection and preparation, result analysis, and the calculation of biological equivalent concentrations (BEQ). Such standardisation methods are available for chemical assessments, but the options are limited when it comes to water quality assessment. For water monitoring, standardised ISO methods are only available for specific estrogenic bioassays (ISO 19040 series), the calculation of BEQ (ISO 23196:2022) and water sampling (ISO 5667 series) (Table 2.1). Developing performance standards for bioassays can level the playing field for vendors wishing to enter the bioassay market and accelerates the validation of methods (see also the case study of California, Box 2.5). Finally, there is a need for technical guidance for regulators and utilities on how to apply bioassays (Neale et al., 2022[60]). Platforms, such as the Water Safety Portal (WHO and IWA, n.d.[62]), could host case studies and guidance documents. Section 3.5.2, Chapter 3, discusses standardisation in more detail.

  3. 3. Countries have different levels of bioanalytical capacity. Laboratories with the capacity to process and analyse (water quality-related) bioassays are scarce in many countries. This also includes the infrastructure for animal facilities for in vivo bioassays or cell culture laboratories for in vitro bioassays. Laboratory infrastructure is discussed in Section 2.6.3.

  4. 4. There is still a lack of specific, validated bioassays for several modes of action (European Environment Agency, 2020[63]; Brack et al., 2018[64]; Robitaille et al., 2022[4]). For example, estrogenic bioassays have more methods than any other endpoints (Table 2.1). In contrast, the thyroid modality has no test guidelines for in vitro bioassays. There is a need to invest in method validation for other endocrine endpoints to consider a broad range of effects related to endocrine disruption (Martyniuk et al., 2022[65]) (see also Box 3.8 on the Pepper platform). The EURION initiative (European Cluster on Identification of Endocrine disruptors) aims to bridge the gaps for non-EATS pathways, such as for metabolic disease, thyroid, neuroendocrine hormones, and for the female reproductive system (Martyniuk et al., 2022[65]; EURION, n.d.[66]). Method validation is a long, costly, and tedious process. Section 3.5, Chapter 3, discusses validation and makes recommendations for improvement of the validation process.

  5. 5. There is still a lack of confidence in the ability to extrapolate the results from in vitro bioassays to their outcomes in humans or ecosystems (see also Box 2.6 on Adverse outcome pathways). More work needs to be done on quantitative in vitro to in vivo extrapolation. This could also help decrease animal use in the long. This aligns with the objective of programmes for the evaluation of single-chemicals such as the ToxCast/Tox21 of the US (Dix et al., 2007[67]; Krewski et al., 2010[68]), the EU-ToxRisk and ONTOX in the EU (Daneshian et al., 2016[69]; Vinken et al., 2021[70]), and the OECD guidelines for the evaluation of EDCs (OECD, 2018[50]). Moreover, in vitro bioassays do not mimic the exact effects happening in a whole organism. This includes, for example, the bioavailability of compounds, the uptake, metabolism, distribution and excretion (ADME) of substances, the impact of chronic exposure or even the sensitivity. Research is ongoing to increase the realism of in vitro bioassays (Robitaille et al., 2022[4]). It should be noted that, for the purposes of water quality monitoring, a bioassay does not need to represent the exact impacts on whole organisms, just like targeted chemical analysis does not represent the exact impact on whole organisms. The purpose is to get an indication of potential risks present in a water sample.

  6. 6. For ecosystem protection, bioassays need to be developed to include a diverse range of species. Most bioassays are designed for human receptors (Robitaille et al., 2022[4]). While hormones are generally conserved across species, proteins such as hormone receptors have evolved independently, which could lead to some differences in sensitivity. For ambient water quality monitoring aiming to protect aquatic ecosystems, it would be ideal to have access to in vitro bioassays representing a higher diversity of species.

  7. 7. Current test guidelines for individual bioassays do not give a full picture of water quality as this would require a battery of bioassays (Di Paolo et al., 2016[2]; Brack et al., 2019[23]). Developing a battery of bioassays requires specialised expertise.

While bioassays are interesting for routine risk management, they might not completely capture the ecological consequences of endocrine disruption (Windsor, Ormerod and Tyler, 2018[74]). In situ wildlife monitoring methods survey species in the wild for any significant physical, molecular or behavioural changes, which could indicate changes in the Predicted No-Effect Concentration (PNEC).

By analysing water samples only in a laboratory setting, water regulators may overlook impacts that are happening in the wild. For example, fish surveys helped identify reproduction issues in various water bodies across the world close to wastewater treatment plants and industries (Jobling et al., 1998[75]; Marlatt et al., 2022[76]; Sumpter, 2005[77]; Hewitt et al., 2008[78]). Those studies led to the identification of compounds found in wastewater, such as EE2, which could lead to endocrine disruption. Another example of the necessity of in situ wildlife monitoring is the observation of the development of male sex organs, known as imposex, in sea snails (Ellis and Agan Pattisina, 1990[79]; Smith, 1981[80]; Beyer et al., 2022[81]). Imposex was later linked to tributyltin (TBT), a biocidal agent in boat paint, which led to its ban (Beyer et al., 2022[81]). Increased wildlife monitoring would benefit research both into bioaccumulation/bioconcentration and into the differences between species, especially invertebrates, in which data are scarce (Fernandez, 2019[82]). Moreover, currently available bioassays would have overlooked the activity of TBT, as its main mechanism of action (via the retinoid X-receptor) is not assessed in most bioassays (Beyer et al., 2022[81]).

In situ surveys rely on the study of indicator species. Those indicator species are used to assess the changing quality of an environment in relation to pollution (Siddig et al., 2016[83]). Species selected as indicators are ideally sensitive to changes in their environment, are local and commonly distributed on the territory of interest, representative of their ecosystem, and well documented. Species can also be selected based on their cultural or economic importance (Hutchinson et al., 2006[84]). Moreover, wildlife monitoring programmes ideally evaluate more than one species to have a better representation of an ecosystem. An example of a such programme that was able to assess endocrine disruption is the Environmental Effects Monitoring (EEM) programme in Canada (Box 2.7). EEM surveys fish to ensure the protection of fish health and their habitat under the Fisheries Act regulations (Environment Canada, 1998[85]) in part to protect the fishing industry and for conservation.

For the selected indicator species, specific biomarkers are measured. Biomarkers act as indicators of a change in a biological organism. In the study of contaminants, biomarkers aim to evaluate either exposure (i.e. evaluate if the organism was in contact with a contaminant) or effect (i.e. evaluate if the organism was affected negatively by its environment) in a given organism (Hutchinson et al., 2006[84]). Any measurable change can be called a biomarker, ranging from physiological (e.g. body and organ mass, tissue histology, sexual secondary characteristics) to molecular change (e.g. protein production and gene expression) (Hutchinson et al., 2006[84]). It should be noted that biomarkers may have different meanings depending on the context and species at hand (Dang and Kienzler, 2019[86]). One of the most widely used biomarkers in associated with endocrine activity is the presence of vitellogenin (VTG) in the blood or liver of organisms. VTG is a precursor of the egg yolk, making it a biomarker for females as males do not produce eggs. It can be used, for example, to detect if a male fish was exposed to estrogenic substances as the production of VTG will have increased (Hutchinson et al., 2006[84]). The EEM programme in Canada studied biomarkers comprising age, weight-at-age, condition factor (weight/length3), and relative weight of the liver and gonads (Box 2.7).

The data collected during in situ wildlife monitoring can be used to assess the health of selected species or the ecosystem in general. This risk assessment will normally involve the comparison of the site of interest to a reference site (e.g. upstream of discharge) which is considered not polluted. If a significant change is detected in the health of selected species between both sites, it can be necessary to take action. For example, in the EEM programme, trigger values were established over time for specific fish biomarkers. When the values are exceeded, this triggers an investigation procedure by industry, which can lead to actions to mitigate the problem (see details in Box 2.7).

As with all monitoring methods, in situ wildlife monitoring has limitations:

  1. 1. There is a need to develop more biomarkers for all modes of action of EDCs. For example, VTG, one of the most used biomarkers for endocrine disruption, is not adapted for all species, such as invertebrates (Windsor, Ormerod and Tyler, 2018[74]) and can present problems of variability (Hutchinson et al., 2006[84]). Biomarkers need to include more mechanisms of action for endocrine disruption, covering all key characteristics of EDCs (La Merrill et al., 2020[114]) (see also Box 1.2, ‘Ten key characteristics of endocrine disrupting chemicals’). Omics (transcriptomics, proteomics and metabolomics) can help in the discovery of new biomarkers and could eventually help risk assessment in the future (Martyniuk, 2018[115]).

  2. 2. In situ wildlife monitoring often looks at one or a small subset of indicator species which can mischaracterise the impact on the whole ecosystem. There is a need to include more species in those surveys to increase the understanding of the food-web and cascade of consequences of EDCs on the trophic system, as well as to take into account biodiversity in groups of species such as fish and invertebrates (Fernandez, 2019[82]; Saaristo et al., 2018[116]; Windsor, Ormerod and Tyler, 2018[74]). Moreover, whilst hormones are generally conserved among most species, the effects of EDCs can differ among species (Hutchinson et al., 2006[84]). Hence, looking at only a few selected species might bias risk assessment.

  3. 3. Wildlife surveys are generally field intensive, expensive, time consuming and involve mostly lethal or invasive sampling for species. New technology like environmental DNA (eDNA) (Box 2.8) could help survey the presence of species by reducing the burden of field work as well as reducing lethal and invasive methods. The former point can be of high importance when dealing with endangered species.

  4. 4. The data developed by wildlife monitoring programmes can be difficult to link to EDCs or pollution. The EEM programme in Canada illustrates this challenge (Box 2.7). It took several years to gather the evidence on pulp mill effluent effects on fish health and to develop a fish bioassay before being able to mitigate the cause. Moreover, data interpretation within species may require additional evidence. For instance, non-EDCs could trigger a change in fish and changes in fish species can be linked to pathways other than estrogen, androgen and steroidogenesis (Dang, 2014[117]).

  5. 5. There is a need to develop tools that assess the risks of pollution at the ecosystem level rather than at the species level. Biological indices are a common tool to indicate impacts at the ecosystem level. For microbial ecosystems, the Pollution-Induced Community Tolerance (PICT, (Tlili et al., 2016[118])) helps risk assessment by predicting the effect of a chemical or mixture based on the tolerance of the community in comparison to reference site. For invertebrates, the Species at Risk (SPEAR) index predicts the impact of pesticides on invertebrate communities based on species sensitivity to pesticides (Schäfer et al., 2007[119]; Hunt et al., 2017[120]). The improvement of such tools and the inclusion of other species such as vertebrates could help accelerate and facilitate risk assessment of pollution in ecosystems.

When EBMs, such as bioassays, have detected endocrine activity in a water sample, the source of this activity is often unknown. An additional step of analysis is needed to identify the chemical(s) causing the activity. This can be done through effect-directed analysis (EDA).

EDA is a method in which a sample is first separated into multiple fractions. Those fractions are then analysed in parallel by both non-targeted chemical analysis and bioassays. The results for each method are then put together to identify culprit chemicals found in those fractions where biological activity is detected (Brack, 2003[121]). EDA can be used to detect a range of EDCs, including new and emerging hormone-like contaminants (Houtman et al., 2004[122]; Simon et al., 2013[123]; Muschket et al., 2018[124]; Hashmi et al., 2018[125]; Gwak et al., 2022[126]; Houtman et al., 2020[127]; Zwart et al., 2018[128]).

Several case studies demonstrate the usefulness of EDA in identifying the culprit chemicals. A study in Korea (Gwak et al., 2022[126]) looked at the efficiency of different steps of treatment in a WWTP, applying bioassays for ER, AR, GR and AhR. The treatment removed all activity except for estrogenicity. After further investigation with EDA, the researchers found that the activity was caused by the pharmaceuticals arenobufagin and loratadine. The activity was confirmed by exposing the same in vitro bioassay to the pure molecule. Another case study is the Holtemme River in Germany, where anti-androgenic activity was suspected to cause decreased reproduction in fish (Muschket et al., 2018[124]). With EDA, fluorescent dye (4-methyl-7-diethylaminocoumarin) was identified as the source of the activity. The activity of the dye was further confirmed in vivo. Both cases demonstrate that the identification of chemicals is an important tool for risk management and abatement actions, as illustrated in Box 2.9.

Currently, EDA is relatively costly and laborious to be used for routine monitoring (Brack et al., 2018[64]). However, advancements have been made in this regard with novel high-throughput techniques (Houtman et al., 2020[127]; Zwart et al., 2018[128]), which should make it more available to other users in the years to come. Another remaining challenge will be to increase the chemical analytical capacity as some of the activity detected is not always followed by chemical detection (Hashmi et al., 2020[131]; Houtman et al., 2020[127]; Zwart et al., 2018[128]). As an example, EDA was used to explain endocrine activity (ER, AR, GR, PR) in the Danube river (Hashmi et al., 2018[125]; Hashmi et al., 2020[131]). In general, EDA was able to explain the activity detected by bioassays, however part of the GR activity was not explained (Hashmi et al., 2020[131]). The authors hypothesised that it could be a method artefact or that the chemicals causing the effects are in very low concentration, but their additive effect can still be seen.

The previous sections described existing and upcoming methods for monitoring EDCs in freshwater. Each method has its strengths and limitations (Table 2.2). As there are probably infinite options of monitoring programmes, this section proposes a set of questions that should be asked during the process of designing a monitoring programme. While these questions do not necessarily provide definite guidance on the monitoring programme design, they can inform on avenues to explore. The ideal environmental monitoring system combines multiple methods of monitoring to strengthen and exploit synergies as they provide important complementary information (Brunner et al., 2020[132]; Hollender et al., 2019[26]). Some countries, therefore, apply a combination of methods. The second part of this section provides country cases of combinations of monitoring methods.

Before being able to monitor EDCs, it is important that the monitoring strategy and programme is designed for the intended purpose. In a perfect world, every type of water and source should be monitored for all chemicals and effects with the best available techniques. However, choices need to be made based on multiple factors such as cost, time, available expertise, equipment and environmental conditions such as temperature, weather and geography. Hence, it is important to first confirm the intent of the programme. This section proposes a set of questions that that guide the process of designing a monitoring programme.

As mentioned throughout this report, there are many types of water to monitor, such as wastewater, recycled water, surface water, groundwater, and drinking water. For human health concerns, it is relevant to look at source waters (particularly when a region relies on a single source), drinking water, recycled water used for irrigation, fish products, or recreational water. Australia and California, United States, set up specific monitoring programmes to ensure safety of recycled water (Escher, Neale and Leusch, 2015[133]; California State Water Board, 2018[6]). Monitoring recycled water is ever more relevant, as certain regions are increasingly using recycled wastewater due to the droughts associated with climate change. Even if there is no immediate risk for human health, monitoring can be a powerful communication tool to inform policy makers on water quality (OECD, 2022[51]).

For wastewater, programmes can be designed to survey and regulate the release of pollution from municipal wastewater treatment plants, but also effluents of specific types of industry (e.g., pulp & paper mills, pharmaceutical manufacturing, mining, see for example the EEM Programme in Canada, Box 2.7).

For all water types, it is also important to consider the limit of quantification required for the choice of methods. For example, drinking water, which is generally obtained from a cleaner source and highly treated, will have low to undetectable levels of contaminants in comparison to wastewater. Hence, some methods might not be sensitive enough to capture contaminants found in drinking water. To not waste resources, it should be ensured that the limit of quantification (LOQ) of the selected method is relevant for the type of water to guarantee the usefulness of the results. Selecting the most sensitive method - with the lowest LOQ – is not necessarily the best option, as sometimes very low levels of endocrine activity do not pose a risk to humans or ecosystems.

This question relates to the protection goal of the monitoring programme: human health or ecosystem health. It can inform on the prioritisation of water type as seen in the previous question. More importantly, this choice will impact the calculation of threshold levels or trigger values. Threshold values are derived based on toxicological risk data, either considering the risk to human health or to ecosystem health (such as benthic organisms, freshwater biota, or critical species). Different species have different tolerance levels to contaminants. For exposure assessment it is important to realise that aquatic organisms are exposed 24/7 to surface water, while human drinking water uptake is estimated to be approximately two litres per day. A threshold level is therefore heavily influenced by the underlying toxicological risk data and protection goal. When both human and ecosystem health are prioritised, the lowest Predicted No-Effect Concentration (PNEC) value can be useful.

It is important to define the purpose and the level of ambition of the monitoring programme. When limited prior knowledge is available, a programme could aim to collect baseline data and identify potential hotspots, such as through targeted chemical analysis (Box 2.1), non-target screening (Box 2.2), or bioassays combined with effect-directed analysis. Other monitoring strategies can be applied to identify hotspots, such as the SIMONI strategy in Box 2.10 (van der Oost et al., 2017[134]). Monitoring initiatives can react to acute situations, such as observed abnormalities in fish physiology or behaviour or concerns raised by the public (Sanchez et al., 2011[129]) (Box 2.9). In such situations a more extensive programme, combining different methods, may be more appropriate to establish a cause and effect relationship, to generate trust in the results and to justify follow-up action. Other monitoring programmes assess if water is fit for purpose (recycled water, drinking water, recreation). In such cases a routine method that embeds an early warning system may be appropriate (Box 2.10). Lastly, monitoring programmes can be used to enforce regulation or permits by setting threshold levels, such as trigger values, quality standards, or concentration levels. In such cases, regulatory “lock-ins” are important to consider, such as unintentional government-required animal testing (Section 2.6.4) or discriminating between methods by preselecting one or a few methods in regulatory standards (Table 2.3).

Water quality assessment is predominantly based on risk-based approaches (see also Chapter 3). As a consequence, the need to develop a threshold or trigger value that defines the acceptable level of risk will arise (Section 2.6.1). However, it can be plausible to adopt a hazard-based approach where EDCs are considered a hazard at any concentration. The threshold level will correspond to zero, i.e. no concentration is allowed in water. The choice between risk-based or hazard-based approaches impacts the selection of methods (e.g. highly sensitive methods for hazard-based approaches), analysis of results and the prioritisation of sampling method.

There is a need to define who is doing what and who bears the cost of the monitoring programme. For example, who is doing the analysis and the design of the study? Who is paying for the analysis? Who is reviewing the results? What in-house capacity is available? While this might be less consequential for small research-based programmes with their own specific research fund, this can play an important role for routine monitoring. The EEM programme in Canada is an example of a monitoring programme where the role of each stakeholder is well defined in Box 2.7. The industry is responsible for monitoring and covers the cost for the conduct of the study, while the government provides guidance documents and assesses the design and the results of the study.

For human health, it is important to consider populations that are particularly vulnerable to EDCs (Section 3.4.4, Chapter 3). For ecosystem health, there might be a need to prioritise the protection of endangered species or species of cultural or economic importance. This could impact the choice of species to be studied in a in situ wildlife monitoring campaign, the selection of threshold or trigger value, and site selection.

Determining the desired type of monitoring can help define the frequency of measures and the feasibility based on available resources. Currently, there are four main types of monitoring (Neale et al., 2022[60]). The first type of monitoring is a ‘system assessment’ which aims to determine the baseline of contamination of the selected water. This type of monitoring can be done as a first screen or repeated over long periods of time (month or years). The second one is ‘validation monitoring’ which evaluates the efficacy of a measure to reduce pollution, like a wastewater treatment plant. This monitoring might be done once to a few times. The third type is ‘operational monitoring’, used to evaluate if water treatment infrastructure is operating well to ensure constant quality of the treatment. However, this might be more difficult for chemicals such as EDCs since, in general, the methods described in section 2.2 and 2.3 require analysis that take more than a day. Finally, ‘verification monitoring’ verifies the compliance of treatment plants. This is often done on quarterly or biannual basis for various parameters and could include monitoring methods for EDCs for all the methods described.

As seen in previous sections, various types of monitoring approaches exist, and while each has its advantages and disadvantages, together they make a very strong monitoring toolbox (Table 2.2). Even though one monitoring approach might be selected over another (e.g. for reasons of cost, time, effectiveness), it is ultimately recommended to combine methods as each provides important complementary information (Brunner et al., 2020[132]; Hollender et al., 2019[26]). Since the information given by each method is of a different nature, one might need tools to integrate all the different datasets. Moreover, methods do not have to be used all at the same time but can be integrated in different stages. For example, one method might be used for pre-screening and follow-up methods can be used to further investigate the issue. Examples of ways to integrate monitoring methods are given in this section.

Bioassays and NTA are increasingly used as a pre-screening or early warning tool to detect endocrine activity in water. Neither method, however, reveals the culprit chemical. When the potential culprit chemicals are known, targeted analysis (Section 2.2.1) can help identify potential chemicals that trigger the detected activity. In some cases, the activity might not be explained by known chemicals and more investigation is needed. This can be done with the help of effect-directed analysis (EDA) (Section 2.4).

The Smart Integrated Monitoring (SIMONI) approach of Waternet, the water authority of Amsterdam, the Netherlands, applies bioassays as an early warning system for surface water quality (Box 2.10). The monitoring programme revealed that the main sources of contamination were landfills, sewage overflow, sewage water effluents and agriculture. SIMONI comprises two Tiers of monitoring. Tier 1 is a routine risk identification by applying two methods: relatively simple bioassays performed on passive samples, and chemical analysis of grab samples is conducted for metals, ammonium, and other substances. The results of Tier 1 are analysed against effect-based trigger values and threshold values. If these values indicate an increased risk, targeted research is prompted in Tier 2. Tier 2 combines broad spectrum chemistry, in vivo bioassays, and effect-directed analysis. When there are concerns for human health, non-targeted analysis and advanced bioassays may be applied.

Switzerland developed an online toolbox of monitoring methods to support cantons in selecting the appropriate combination of methods for surface water quality monitoring (Box 2.11).

Abnormalities in wildlife can be observed by routine wildlife monitoring, or even from observations by local communities. In situ wildlife analysis of specific physical endpoints is generally the first step. This analysis is typically conducted in the potentially contaminated site and a reference site. Bioassays can then be applied to confirm if effects are caused by chemical pollution. Mapping pressures (such as municipal or industrial effluents, landfills, agricultural activities) can guide on the selection of relevant substances for targeted analysis to identify the culprit. Laboratories carrying out the chemical analysis should be sufficiently equipped to report back on low detection limits, i.e. nanogram/litre concentrations. Effect-directed analysis is another tool to identify the culprit. A workflow to respond to observed abnormalities is well described by Sanchez et al. and Creusot et al. (2014[130]; 2011[129]) (Box 2.9), following a case of observed adverse effects in wild fish living near pharmaceutical manufacture discharges in France.

Abnormalities can arise from unregulated substances and regulators may have limited powers when guidelines do not exist. High confidence in the assessment results is paramount for industry and regulators to recognise the problem and to justify follow-up action. Thorough research, however, can increase the lag time between observation and action. Pre-defined protocols and methods could reduce this lag time.

This section covers success factors related to implementing bioassays as monitoring method. It discusses setting water quality standards and trigger values; options to minimise the costs; access to laboratories; considerations in relation to animal testing; and water sampling. These success factors can facilitate cost-effective deployment of bioassays for policy purposes in a range of contexts.

This paragraph discusses the options for setting threshold levels for concentrations of endocrine disrupting chemicals or endocrine disrupting effects. Threshold values are commonly used in setting environmental quality standards or as a condition in a discharge permit. There are roughly three types of thresholds: 1) water quality criteria for chemical analysis, 2) effect-based trigger values for bioassays; and 3) trigger values for in situ monitoring of wild species; (Been et al., 2021[138]; Neale, Leusch and Escher, 2020[139]; Escher et al., 2018[56]; van der Oost et al., 2017[134]; James, Kroll and Minier, 2023[140]). The three types of standards discussed in this section are complementary to one another, and a mix of standards can be appropriate.

This publication does not provide any definite guidance on the appropriate threshold values. Rather, it discusses the considerations when setting water quality standards for endocrine activity or disruption. Ultimately, determining the acceptable level of risk is a complex decision, usually made by government regulators (in consultation with scientists, stakeholders, industry, and other groups).

Typically, substances are regulated on a substance-by-substance basis. However, current substance-by-substance regulation does not always capture the endocrine properties of chemicals in water quality criteria or standards. Regulators normally work with a calculation method for deriving water quality criteria or environmental quality standards, considering many environmentally harmful properties of substances, such as acute toxicity. In practice, the calculation methods do not fully consider the endocrine disrupting properties of substances (James, Kroll and Minier, 2023[140]) (see also Box 2.12).

France is exploring how to adjust existing water quality standards considering the endocrine properties of substances that are already prioritised on the Environmental Quality Standards list. The French National Institute for Industrial Environment and Risks (Ineris) found that 70% of the Environmental Quality Standards of the substances that are potentially endocrine active or disruptive did not consider endocrine activity as part of the method, although there are substance-specific data suggesting or evidencing such activities (James, Kroll and Minier, 2023[140]) (see also Box 2.12). France therefore developed a method to derive Environmental Quality Standards that better reflect the endocrine disruptive properties of substances, which is further described in Box 2.12.

Effect-based trigger values (EBTs) are the threshold values, or water quality indicators, for bioassays. EBTs help interpret whether the effects detected in a bioassay are acceptable or not. While bioassay results provide a lot of information already, as the detected responses can be compared in time and space, they do not assess the potential risk as such, because not all levels of activity are a risk to humans or aquatic species, particularly given that some bioassays could be very sensitive to low doses of contamination (De Baat et al., 2020[145]). “Exceedance of an effect-based trigger value signals the presence of a hazard, induced by one or more potentially harmful compounds. However, this does not necessarily mean there is a risk” (Been et al., 2021[138]; van der Oost et al., 2017[134]). It rather means that below the trigger value, the chance of adverse effects on humans or environment is low (Been et al., 2021[138]; Neale et al., 2023[57]). Trigger values are most used as a pre-screening value for further analysis (van der Oost et al., 2017[134]), but can also be used as a regulatory standard for water quality (California State Water Board, 2018[6]).

Trigger values are essential to determine whether there is a (potential) risk and whether follow-up action is required. Trigger values should therefore be established at a realistic level, as low trigger values can lead to many “hits” or unnecessary follow-up actions (i.e. the trigger value was too rigid) and high trigger values may overlook issues of concern and not lead to appropriate follow-up actions (i.e. the trigger value was too tolerant) (Been et al., 2021[138]; Dingemans et al., 2018[146]). Establishing trigger values at just the right level has been subject of a long-standing scientific and policy debate and it appears to be one of the major barriers towards implementing EBMs (OECD, 2022[51]).

Some jurisdictions, such as California, apply effect-based threshold values instead of effect-based trigger values for bioassays. Though there is some nuance between the two, this report uses these terms interchangeably. One difference is that California’s threshold values for bioassays are tiered, meaning that if the bioassay result is five times above the threshold value, it triggers different actions than when it is ten times above the threshold value, and so on.

Effect-based trigger values can be established for specific types of bioassays, “assay-specific trigger values”, or for specific endpoints regardless of the brand of bioassay, “generic trigger values” (Neale, Leusch and Escher, 2020[139]; De Baat et al., 2020[145]). There are advantages and disadvantages to each approach (Table 2.3). De Baat et al. (2020[145]) recommend using assay-specific trigger values, as these are more accurate because each bioassay is. However, from a public policy perspective, generic trigger values are more appropriate for regulatory purposes such as setting environmental quality norms or discharge permits. Generic trigger values, combined with bioassay performance standards, allow any bioassay provider to enter the market and makes it easier to substitute bioassays with an equivalent (due to costs, shortages, laboratory capacity and other reasons). For example, California favoured an endpoint-specific approach to be able to easily replace bioassays with alternatives that are, for instance, more economical or more easily deployable by laboratories.

Trigger values are commonly expressed in terms of concentration levels (nanogramme per litre) of the biological equivalent concentration (BEQ) of a reference chemical (e.g. estrogen equivalents for estrogenic bioassays). The BEQ allows comparing the activity of a chemical or mixture by comparing it to a reference chemical. For example, the BEQ for estrogenic activity translates the levels of activity caused by a mixture of chemicals, such as of contraceptive pills and sex hormones, in the concentration level of estrogens. The concentration levels refer to a concentration of a reference chemical for the effect-based trigger value, but in fact, it expresses the cumulative effect of all chemicals present in a sample, including unknown chemicals. Box 2.13 contains a simple explainer of BEQs.

Effect-based trigger values can be set for the protection of human health or ecosystem health, each of which yields different trigger values (Been et al., 2021[138]). In many cases, the trigger values for the protection of ecosystems can be more stringent, as most aquatic organisms are physically smaller than humans, which can make them more susceptible to pollutants, and humans naturally have higher concentrations of hormones in their bodies. In addition, an aquatic organism is continuously exposed to freshwater, or at least for a large portion of its life.

There are roughly four ways of deriving an effect-based trigger value:

  1. 1. Using available toxicological data on safe levels (for humans or wildlife) of a reference chemical relevant to the bioassay (Been et al., 2021[138]). This yields a threshold value that is similar to concentration levels applied for chemical analysis, e.g. E2 for estrogenic activity. The next step is to transform this threshold into a BEQ to be able to use the thresholds in the data analysis of bioassays. For many substances, a water quality threshold level already exists, for example in drinking water guidelines or environmental regulation. In such cases, existing guideline values can easily be adapted for a selected bioassay, this is called “read-across” (Escher et al., 2018[56])(see Table 2.4). If multiple chemicals have been identified as highly potent substances for one type of activity, the integration of all their existing guidelines in the calculation of one EBT should be considered. If there is no regulatory value available, a value can still be derived by looking at available BEQ data on known potent chemicals, and then estimate the BEQ level that is hazardous to no more than, for example, 5% of aquatic organisms (using the Species Sensitivity Distribution) (van der Oost et al., 2017[134]). Another example of such an approach is a recent study that used PNEC for the risk assessment of 56 WWTPs across 15 European countries (Finckh et al., 2022[148]).

  2. 2. Comparing in vivo and in vitro bioassay responses for a selected chemical and determining at which moment an effect can be detected in vivo. The in vitro equivalent of the in vivo tipping point is then selected as the trigger value. This was done for estrogenic activity by comparing effects in fish embryo and in vitro bioassays of multiple cell-lines (Brion et al., 2019[149]).

  3. 3. When no toxicological data is available, a simple method was recently proposed to calculate an EBT. This method consists of setting the threshold at the concentration generating 10% of effect for the reference compound of the selected bioassay (Neale et al., 2023[57]). This method enables to differentiate activity from the noise of the method and the resulting EBTs are at worst at one order of magnitude of the ones designed using previously mentioned methods. Hence, they can provide a good first approximation when no data is available.

  4. 4. If there is no toxicological data or possibility to derive a trigger value in the laboratory, EBTs can be determined in the field. This can be done by acquiring data on various sites and water types for which the expected water quality can be classified. Based on the level of activity observed for each water type or site, a threshold can be established at the level that allows us to distinguish between expected water qualities.

Trigger values are unavoidably associated with uncertainties due to incomplete knowledge about the composition of a sample, the quality of underlying data and the dynamics of mixture effects (Working Group Chemicals, 2021[147]). The Working Group Chemicals under the European Water Framework Directive developed a decision framework to derive effect-based trigger values based on the breadth and quality of knowledge and data available on the risks associated with the relevant effects and chemicals (Working Group Chemicals, 2021[147]). It prioritises four methods of deriving a trigger value. The trigger values derived in Tier 4 are the most robust and based on complex methods; the ones in Tier 1 are the least robust and based on simple methods.

  • Tier 4: derived based on in vivo and in vitro studies that have been calibrated against one another. In addition, chemical-mixture effects and risks have been quantified based on monitoring data.

  • Tier 3: derived based on in vitro studies, and data from chemical monitoring and mixture risk assessments.

  • Tier 2: derived based on data for existing environmental quality standards for single compounds, enriched with data on other compounds that trigger the bioassay.

  • Tier 1: derived based on data of a reference compound that has the most potent effect in a bioassay, ideally based on an existing environmental quality standard.

Programmes that monitor fish and other species in the wild require methods that measure a (statistically significant) change in fish health. The trigger values adopted by the Canadian Environmental Effects Monitoring (EEM) programme are called “Critical Effect Size Triggers”. The EEM programme is a comparative method based in part on five “core fish endpoints”, namely age, weight-at-age (growth rate), relative gonad size, relative liver size and condition (weight/length3). To assess the effects of pollution, a comparison is made between fish living in habitats exposed to effluent pollution and the reference fish that live in reference or unexposed habitats. If a statistically significant difference is detected on one of the five endpoints, this gives lead to further investigation on the potential impact of effluent pollution. By means of illustration, the Critical Effect Size Triggers are a ≥ 25% change in relative gonad or liver size, or a 10% change in condition factor (for more on Canada’s EEM programme, see Box 2.7).

Analysing the costs of new water quality monitoring methods is not straightforward and depends on several factors. The cost components of water quality monitoring, excluding method development costs, comprise (Kienle et al., 2015[152]; Drewes et al., 2018[150]):

  • Sampling and pre-treatment of samples.

  • Capital expenditure on laboratory equipment.

  • Laboratory consumables and test products, such as kits and/or cell lines. Note that the costs of bioassays that require a license are, at the moment, relatively higher than license-free bioassays.

  • Labour required to maintain, prepare and operate the samples and analysis.

The outsourcing of services can affect the costs of a monitoring programme. Distance to the nearest qualified laboratory is an issue in some countries where samples need to be shipped domestically or abroad for analysis.

Comparing the cost-effectiveness of methods is ambiguous, though some general statements can be made. Targeted chemical analysis of well-regulated or routinely-monitored chemicals benefits from economies of scale which reduces analytical costs (Working Group Chemicals, 2021[147]). Such economies of scale have not yet been reached with bioanalytical methods and non-targeted chemical analysis. In a way, bioassays can be more cost-efficient than targeted chemical analysis as they respond to a group of substances, which is inherently impossible with substance-by-substance methods. Moreover, anecdotal evidence suggests that the required infrastructure and equipment for EBMs (e.g. incubators, sterile hood and plate-readers) have lower cost than for analytical chemistry (e.g. mass spectrometer). However, EBMs may need additional methods, such as effect-directed analysis, to identify the culprit chemical with certainty. Moreover, comparing the different bioassay approaches, in vitro methods are generally more cost-effective than in vivo methods, as they can be more easily scaled up by automation and high-throughput technologies (Drewes et al., 2018[150]; Working Group Chemicals, 2021[147]). Lastly, non-target screening is a relatively expensive method due to the need for specialised experts and the capital cost of equipment.

The costs of bioassays differ per cell line provider, laboratory, type of services outsourced (depending on in-house capacity), and country (Kienle et al., 2015[152]; Drewes et al., 2018[150]; Working Group Chemicals, 2021[147]). In the Netherlands, the implementation of a complete set of bioassays (including, but not limited to, assays detecting endocrine activity) costs about EUR 800-1100, which comes down to around EUR 100 per bioassay (De Baat, Van Den Berg and Pronk, 2022[151]). The cost of estrogenic effect monitoring has been estimated at approximately EUR 140-200 per sample within the European Union (Working Group Chemicals, 2021[147]). It is generally expected by the experts who contributed to this publication that costs associated with bioassays will diminish as demand increases.

A smart monitoring programme design can potentially reduce costs. For example, the same samples can be used for multiple purposes, such as chemical analyses and effect-based analyses (Wernersson et al., 2015[5]). Moreover, it could be worthwhile researching if some routine chemical analysis can be partially replaced with bioassays that capture the same chemicals. However, this has not been widely explored in the literature and requires further research. Lastly, cost-effective choices can be made in determining the comprehensiveness of a battery of bioassays. Knowledge about the environmental pressures can direct the selection of a battery of methods. For instance, estrogenic effect assays could be prioritised if a water body is primarily exposed to sewage effluents. The Water Quality Guidelines in the Netherlands distinguish between a “basic battery” of six bioassays and an “additional battery” (De Baat, Van Den Berg and Pronk, 2022[151]). In Canada, the frequency of Environmental Effects Monitoring is reduced if there are no effects observed over two consecutive monitoring cycles (Box 2.7).

In some countries, very few to no laboratories have the expertise or the infrastructure to perform and analyse bioassays. To make bioassays more widely available for regulators, various types of laboratories could be considered, including research laboratories, contract laboratories, water utility/authority laboratories, and medical laboratories (OECD, 2022[51]). Medical laboratories often have long-standing experience with bioanalytical methods but may need additional guidance on water sample preparation and treatment. Outsourcing bioanalytical analysis to university laboratories may not be appropriate for water safety analysis, as it requires specialised expertise and certification. Various steps of the analytical process, from sample preparation to analysis, can be outsourced to the test method developer or cell line supplier. Interlaboratory comparison should be performed to ensure the robustness of methods across laboratories (industry, academia, government facilities), platforms/vendors and relevant sample matrices (OECD, 2022[51]).

Non-targeted screening and effect-directed analysis also require highly specialised equipment and experts. These methods may not be available to every country and at every budget. International collaborative research projects and outsourcing analysis to laboratories abroad are common practice to overcome the barrier of limited laboratory access.

In many cases, in vivo bioassay methods are a form of animal testing. Fish species are commonly used in freshwater and effluent testing (Robitaille et al., 2022[4]). Designing a monitoring programme or regulatory standard that unintentionally stimulates animal testing should be avoided, particularly if non-animal methods are available. In some countries, invertebrates and fish and frog embryos are accepted methods that reduce animal suffering. By regulating or integrating bioassays into regulatory practices or test guidelines, countries run the risk to lock in practices of animal testing for regulatory compliance.

There are many reasons why animal testing is used in water monitoring. For some endpoints, in vivo methods may be the only method sufficiently sensitive or reliable to make statements on toxicity of a water or effluent sample. In vivo methods can also be used as a second-step test to confirm effects found in vitro settings. Regulatory standards can also be a driver for animal testing. Governments sometimes require that companies use in vivo methods to monitor compliance with regulatory standards, such as in the oil and gas industry (Hughes, Maloney and Bejarano, 2021[153]).

It is worth considering that in vivo tests are more expensive and time-consuming. Mittal et al. (2022[154]) estimate that traditional, animal-based, ecotoxicity tests for a single chemical “cost USD $118,000, require 135 animals, and take 8 weeks”, while New Approach Methods cost “USD $2,600, require 20 animals (or none), and take up to 4 weeks to test 16 (to potentially hundreds of) chemicals” (Mittal et al., 2022[154]).

Moreover, bioassays should not be considered as a tool that exactly represents what is happening in the water sample. Rather, bioassays should be valued on par with targeted chemistry: as a proxy for the state of our water quality. It cannot be expected that bioassays be closer to reality than chemical analysis. Bioassays simply provide additional pieces of information that inform on potential risks. Using in vivo bioassays for compliance monitoring therefore most likely overshoots the purpose of a routine water quality monitoring programme, particularly when in vitro assays are available. In this context, there is a concern that the international definition of endocrine disruptors may become a driver for regulatory animal testing, as it implies that an adverse health effect needs to occur in an intact organism: “An endocrine disruptor is an exogenous substance or mixture that alters function(s) of the endocrine system and consequently causes adverse health effects in an intact organism, or its progeny, or (sub)populations” (IPCS, 2002[155]). For the purposes of (routine) water quality monitoring, permanent compliance with the definition may be unnecessary as in vitro tests, combined with in silico methods, can provide valuable information on the effect on the mixture effects of all EDC present in a sample (Escher, Neale and Leusch, 2021[156]).

Some recommendations can be made to avoid unnecessary animal testing for freshwater and effluent quality testing:

  • Avoid developing regulations or standards that lock in government-required animal testing, and instead design flexible standards that allow for alternative methods in the future. This also includes the development of effect-based trigger values. If an in vivo-based-effect-based trigger value is embedded in regulation, it may lead to government-required animal testing.

  • In most OECD countries, the use of animals for scientific or regulatory testing is regulated and reported to the public. However, testing for water quality regulation is sometimes beyond the scope of animal testing statistics. Sharing data of animal testing for water quality regulation can help avoid unnecessary animal testing and ensure humane treatment of animals in unavoidable cases.

  • Embed the 3R principles of Replacement (avoiding animal testing), Reduction (limit the number of animals exposed to animal testing), and Refinement (limit the suffering and distress of animals) in water monitoring and regulation (Russel and Burch, 1960[157]). Concrete ways of embedding the 3R principles in water practices is adding an article on “Choice of methods” in regulation or guidelines, prioritising non-animal methods in validating test guidelines. A lot can be learnt from chemicals regulation and practices (Scholz et al., 2013[158]).

For assessing risk in freshwater, the sampling strategies and the sample preparations matter. The development of guidelines and standard operating procedures would facilitate the use of bioassays (Neale et al., 2022[60]). The sampling strategy is first designed depending on the objective of the sampling campaign (Escher, Neale and Leusch, 2021[159]). This objective will depend on the water to test (e.g. surface water, drinking water, wastewater) and the information sought (e.g. assess efficiency of treatment, find hotspots in surface water). Clearly identifying the objective of the sampling will help determine what is necessary for the rest of the sampling strategy. International or national guidelines exist for sampling water from different sources (e.g. ISO 5667 series, (European Commission, 2009[160])) to help determine how to perform the sampling (e.g. number of samples, type of bottle, conservation of samples). While they might not be specific to endocrine disruptors, they can still guide on the strategies to be used.

The method and timing of water collection also matters. The traditional method of collecting water samples is referred to as grab sampling, i.e., taking a sample of water directly at the site (Escher, Neale and Leusch, 2021[159]). However, those samples represent only one moment in time for the selected site and might not be representative of the contaminants that can be found generally at the site. To mitigate this issue, composite samples are often done for water treatment plants (Escher, Neale and Leusch, 2021[159]). For that, water samples will be collected throughout 24 hours and mixed to form one composite sample. Composite samples take into account the variation of contaminants during the day. In research, there is a growing interest in passive sampling (Luo et al., 2022[34]; Escher, Neale and Leusch, 2021[159]) to increase the representativeness of a site over time. Passive sampling uses a device that contains a sorbent which collects chemicals over a chosen period at a given site.

After the collection, samples will need a pre-treatment before being able to use them for chemical analysis and in vitro bioassays. This pre-treatment or sample preparation is necessary to concentrate the sample for the analysis, but also to remove components that might interfere with the analysis (Luo et al., 2022[34]; Robitaille et al., 2022[4]). As the sample is modified in this process, some chemicals can be lost (see Annex 2.A). Hence, methods are often judged on their capacity to retain chemicals of interest, called ‘recovery’.

For sample preparation, there is a clear need for standardisation, as well as a need for increasing the capacity of processing samples (Paszkiewicz et al., 2022[25]; Luo et al., 2022[34]; Metcalfe et al., 2022[19]; Robitaille et al., 2022[4]). Embedding guidelines for water sample preparation within existing international test methods or guidelines, such as ISO Standards or OECD Test Guidelines, is worth considering.

This chapter described the available methods for monitoring EDCs and endocrine activity in water, based on case studies from across OECD countries. It also discussed potential barriers and uncertainties in monitoring EDCs and endocrine activity. Figure 2.2 presents a conceptual framework summarising the monitoring possibilities and follow-up actions in four Levels. Level 0 guides through the design of the monitoring programme with the help of questions as described in Section 2.5.1. Level 1 addresses the choice of methods, which can a single method or a combination of methods (described in Sections 2.2 and 2.3). Level 2 provides an overview of all the validation methods to confirm the hazard, identify the culprit chemical or map the sources (described in Section 2.5.2). Finally, Level 3 describes potential action that can be taken after either a threshold was exceeded (Level 1) or a risk was confirmed (Level 2), which will be discussed in the next Chapter.

It is important to note that monitoring is not a pollution reduction measure in itself (OECD, 2019[161]). Monitoring can support in prioritising or justifying action, but uncertainties will persist – particularly given continued international manufacturing and trade of existing and new substances, the further release of EDCs and other CECs into the environment, and challenges arising from environmental change and degradation - such as climate change, biodiversity decline, land degradation and desertification. These pressures only increase the imperative for governments to avoid “decision paralysis” and identify options for near-term preventive action for the safety of humans and the environment. The next Chapter sets out such instruments to manage EDCs in freshwater.


[43] Altenburger, R. et al. (2015), Future water quality monitoring - Adapting tools to deal with mixtures of pollutants in water resource management, https://doi.org/10.1016/j.scitotenv.2014.12.057.

[29] Alygizakis, N. et al. (2019), “NORMAN digital sample freezing platform: A European virtual platform to exchange liquid chromatography high resolution-mass spectrometry data and screen suspects in “digitally frozen” environmental samples”, TrAC Trends in Analytical Chemistry, Vol. 115, pp. 129-137, https://doi.org/10.1016/j.trac.2019.04.008.

[11] ANZECC and ARMCANZ (2000), Australian and New Zealand Guidelines for Fresh and Marine Water Quality. Volume 3: Primary industries – Rationale and background information (Chapter 9), The Guidelines., Australian and New Zealand Environment and Conservation Council and Agriculture and Resource Management Council of Australia and New Zealand.

[10] ANZG (2021), Toxicant default guidelines for protecting aquatic ecosystems., Australian and New Zealand Governments and Australian state and territory governments, https://www.waterquality.gov.au/anz-guidelines (accessed on 20 January 2023).

[28] Beckers, L. et al. (2020), “Unraveling longitudinal pollution patterns of organic micropollutants in a river by non-target screening and cluster analysis”, Science of The Total Environment, Vol. 727, p. 138388, https://doi.org/10.1016/j.scitotenv.2020.138388.

[138] Been, F. et al. (2021), “Development of a framework to derive effect-based trigger values to interpret CALUX data for drinking water quality”, Water Research, Vol. 193, p. 116859, https://doi.org/10.1016/j.watres.2021.116859.

[100] Belknap, A. et al. (2006), “Identification of compounds associated with testosterone depressions in fish exposed to bleached kraft pulp and paper mill chemical recovery condensates”, Environmental Toxicology and Chemistry, Vol. 25/9, p. 2322, https://doi.org/10.1897/05-638R.1.

[81] Beyer, J. et al. (2022), “The ecotoxicology of marine tributyltin (TBT) hotspots: A review”, Marine Environmental Research, Vol. 179, p. 105689, https://doi.org/10.1016/j.marenvres.2022.105689.

[41] Bopp, S. et al. (2019), “Regulatory assessment and risk management of chemical mixtures: challenges and ways forward”, Critical Reviews in Toxicology, Vol. 49/2, pp. 174-189, https://doi.org/10.1080/10408444.2019.1579169.

[47] Bopp, S. et al. (2019), “Regulatory assessment and risk management of chemical mixtures: challenges and ways forward”, Critical Reviews in Toxicology, Vol. 49/2, pp. 174-189, https://doi.org/10.1080/10408444.2019.1579169.

[121] Brack, W. (2003), “Effect-directed analysis: a promising tool for the identification of organic toxicants in complex mixtures?”, Analytical and Bioanalytical Chemistry, Vol. 377/3, pp. 397-407, https://doi.org/10.1007/s00216-003-2139-z.

[23] Brack, W. et al. (2019), “Effect-based methods are key. The European Collaborative Project SOLUTIONS recommends integrating effect-based methods for diagnosis and monitoring of water quality”, Environmental Sciences Europe, Vol. 31/1, https://doi.org/10.1186/s12302-019-0192-2.

[38] Brack, W. et al. (2016), “Effect-directed analysis supporting monitoring of aquatic environments — An in-depth overview”, Science of The Total Environment, Vol. 544, pp. 1073-1118, https://doi.org/10.1016/j.scitotenv.2015.11.102.

[44] Brack, W. et al. (2015), “The SOLUTIONS project: Challenges and responses for present and future emerging pollutants in land and water resources management”, Science of The Total Environment, Vol. 503-504, pp. 22-31, https://doi.org/10.1016/j.scitotenv.2014.05.143.

[64] Brack, W. et al. (2018), “Towards a holistic and solution-oriented monitoring of chemical status of European water bodies: how to support the EU strategy for a non-toxic environment?”, Environmental Sciences Europe, Vol. 30/1, https://doi.org/10.1186/s12302-018-0161-1.

[27] Brack, W. et al. (2019), “High-resolution mass spectrometry to complement monitoring and track emerging chemicals and pollution trends in European water resources”, Environmental Sciences Europe, Vol. 31/1, p. 62, https://doi.org/10.1186/s12302-019-0230-0.

[149] Brion, F. et al. (2019), “Monitoring estrogenic activities of waste and surface waters using a novel in vivo zebrafish embryonic (EASZY) assay: Comparison with in vitro cell-based assays and determination of effect-based trigger values”, Environment International, Vol. 130, p. 104896, https://doi.org/10.1016/j.envint.2019.06.006.

[132] Brunner, A. et al. (2020), “Integration of target analyses, non-target screening and effect-based monitoring to assess OMP related water quality changes in drinking water treatment”, Science of The Total Environment, Vol. 705, p. 135779, https://doi.org/10.1016/J.SCITOTENV.2019.135779.

[6] California State Water Board (2018), Water quality control policy for recycled water, State Water Resources Control Board, California Environmental Protection Agency.

[12] CCME (2021), Summary table. Canadian Environmental Quality Guidelines., Canadian Council of Ministers of the Environment (CCME), https://ccme.ca/en/current-activities/canadian-environmental-quality-guidelines (accessed on 20 January 2023).

[130] Creusot, N. et al. (2014), “Identification of Synthetic Steroids in River Water Downstream from Pharmaceutical Manufacture Discharges Based on a Bioanalytical Approach and Passive Sampling”, Environmental Science & Technology, Vol. 48/7, pp. 3649-3657, https://doi.org/10.1021/es405313r.

[69] Daneshian, M. et al. (2016), “Highlight report: Launch of a large integrated European in vitro toxicology project: EU-ToxRisk”, Archives of Toxicology, Vol. 90/5, pp. 1021-1024, https://doi.org/10.1007/s00204-016-1698-7.

[117] Dang, Z. (2014), “Fish biomarkers for regulatory identification of endocrine disrupting chemicals”, Environmental Pollution, Vol. 185, pp. 266-270, https://doi.org/10.1016/j.envpol.2013.11.006.

[86] Dang, Z. and A. Kienzler (2019), “Changes in fish sex ratio as a basis for regulating endocrine disruptors”, Environment International, Vol. 130, p. 104928, https://doi.org/10.1016/j.envint.2019.104928.

[151] De Baat, M., S. Van Den Berg and T. Pronk (2022), Deltafact: Het toepassen van bioassays binnen Nederland, Foundation for Applied Water Research STOWA, Amersfoort, https://www.datocms-assets.com/38996/1644586511-deltafact-het-nieuwe-bioassay-spoor_v5.pdf (accessed on 7 February 2023).

[145] De Baat, M. et al. (2020), “Advancements in effect-based surface water quality assessment”, Water Research, Vol. 183, p. 116017, https://doi.org/10.1016/j.watres.2020.116017.

[2] Di Paolo, C. et al. (2016), “Bioassay battery interlaboratory investigation of emerging contaminants in spiked water extracts – Towards the implementation of bioanalytical monitoring tools in water quality assessment and monitoring”, Water Research, Vol. 104, https://doi.org/10.1016/j.watres.2016.08.018.

[146] Dingemans, M. et al. (2018), “Risk‐based approach in the revised European Union drinking water legislation: Opportunities for bioanalytical tools”, Integrated Environmental Assessment and Management, Vol. 15/1, pp. 126-134, https://doi.org/10.1002/ieam.4096.

[67] Dix, D. et al. (2007), “The ToxCast Program for Prioritizing Toxicity Testing of Environmental Chemicals”, Toxicological Sciences, Vol. 95/1, pp. 5-12, https://doi.org/10.1093/toxsci/kfl103.

[52] Dodder, N., A. Mehinto and K. Maruya (2015), SCCWRP technical report 854 - Monitoring of constituents of emerging concern (CECs) in California’s aquatic ecosystems - pilot study design and QA/QC guidance, SCCWRP.

[150] Drewes, J. et al. (2018), Monitoring Strategies for Constituents of Emerging Concern (CECs) in Recycled Water: Recommendations of a Science Advisory Panel Convened by the State Water Resources Control Board, State Water Resources Control Board, https://ftp.sccwrp.org/pub/download/DOCUMENTS/TechnicalReports/1032_CECMonitoringInRecycledWater.pdf (accessed on 7 February 2023).

[97] Dubé, M. and D. MacLatchy (2001), “Identification and treatment of a waste stream at a bleached-kraft pulp mill that depresses a sex steroid in the mummichog (Fundulus heteroclitus)”, Environmental Toxicology and Chemistry, Vol. 20/5, pp. 985-995, https://doi.org/10.1002/etc.5620200508.

[96] Dubé, M. and D. MacLatchy (2000), “Endocrine responses of Fundulus heteroclitus to effluent from a bleached-kraft pulp mill before and after installation of reverse osmosis treatment of a waste stream”, Environmental Toxicology and Chemistry, Vol. 19/11, pp. 2788-2796, https://doi.org/10.1002/etc.5620191125.

[18] edlists.org (n.d.), Endocrine Disruptor Lists, https://edlists.org/ (accessed on 15 March 2023).

[79] Ellis, D. and L. Agan Pattisina (1990), “Widespread neogastropod imposex: A biological indicator of global TBT contamination?”, Marine Pollution Bulletin, Vol. 21/5, pp. 248-253, https://doi.org/10.1016/0025-326X(90)90344-8.

[113] Environment and Climate Change Canada (2020), Pulp & Paper Effluent Regulations - ANNUAL REPORT 2020, Environment and Climate Change Canada, Gatineau, QC, Canada, https://publications.gc.ca/collections/collection_2022/eccc/En1-87-2020-eng.pdf (accessed on 18 January 2023).

[112] Environment and Climate Change Canada (2019), Pulp and Paper Effluent Regulations - ANNUAL REPORT 2019, Environment and Climate Change Canada, Gatineau, QC, Canada, https://publications.gc.ca/collections/collection_2021/eccc/En1-87-2019-eng.pdf (accessed on 18 January 2023).

[107] Environment Canada (2014), Sixth National Assessment of Environmental Effects Monitoring Data from Pulp and Paper Mills Subject to the Pulp and Paper Effluent Regulation, Environment Canada, Ottawa, ON, Canada, https://publications.gc.ca/collections/collection_2014/ec/En14-84-2014-eng.pdf (accessed on 18 January 2023).

[95] Environment Canada (2012), Status report on the Pulp and Paper Effluent Regulations, Environment Canada, Ottawa, ON, Canada, https://publications.gc.ca/collections/collection_2012/ec/En14-66-2012-eng.pdf (accessed on 18 January 2023).

[91] Environment Canada (2010), 2010 Pulp and Paper Environmental Effects Monitoring (EEM) Technical Guidance Document, Environment Canada, Ottawa, ON, Canada, https://www.ec.gc.ca/esee-eem/3E389BD4-E48E-4301-A740-171C7A887EE9/PP_full_versionENGLISH%5B1%5D-FINAL-2.0.pdf (accessed on 18 January 2023).

[92] Environment Canada (2005), Pulp and paper environmental effects monitoring guidance document. Eem/2005/1., National EEM Office, Environment Canada.

[85] Environment Canada (1998), Pulp and paper technical guidance for aquatic environmental effects monitoring., National EEM Office, Environment Canada.

[59] Escher, B. et al. (2014), “Benchmarking Organic Micropollutants in Wastewater, Recycled Water and Drinking Water with In Vitro Bioassays”, Environmental Science & Technology, Vol. 48/3, pp. 1940-1956, https://doi.org/10.1021/es403899t.

[56] Escher, B. et al. (2018), “Effect-based trigger values for in vitro and in vivo bioassays performed on surface water extracts supporting the environmental quality standards (EQS) of the European Water Framework Directive”, Science of the Total Environment, Vol. 628-629, https://doi.org/10.1016/j.scitotenv.2018.01.340.

[156] Escher, B., P. Neale and F. Leusch (2021), Bioanalytical Tools in Water Quality Assessment, IWA Publishing, https://doi.org/10.2166/9781789061987.

[40] Escher, B., P. Neale and F. Leusch (2021), “Design of test batteries and interpretation of bioassay results”, in Bioanalytical Tools in Water Quality Assessment, IWA Publishing, https://doi.org/10.2166/9781789061987_0265.

[1] Escher, B., P. Neale and F. Leusch (2021), “Introduction to bioanalytical tools in water quality assessment”, in Bioanalytical Tools in Water Quality Assessment, IWA Publishing, https://doi.org/10.2166/9781789061987_0001.

[159] Escher, B., P. Neale and F. Leusch (2021), “Sampling, sample preparation and dosing”, in Bioanalytical Tools in Water Quality Assessment, IWA Publishing, https://doi.org/10.2166/9781789061987_0245.

[133] Escher, B., P. Neale and F. Leusch (2015), “Effect-based trigger values for in vitro bioassays: Reading across from existing water quality guideline values”, Water Research, Vol. 81, pp. 137-148, https://doi.org/10.1016/j.watres.2015.05.049.

[37] Escher, B., H. Stapleton and E. Schymanski (2020), “Tracking complex mixtures of chemicals in our changing environment”, Science, Vol. 367/6476, pp. 388-392, https://doi.org/10.1126/science.aay6636.

[7] EU (2000), “Directive 2000/60/EC”, Official Journal of the European Communities, Vol. L 269/September 2000.

[66] EURION (n.d.), European Cluster To Improve Identification Of Endocrine Disruptors, https://eurion-cluster.eu/ (accessed on 25 February 2023).

[8] European Commission (2022), Proposal for a Directive of the European Parliament and of the Council amending Directive 2000/60/EC establishing a framework for Community action in the field of water policy, Directive 2006/118/EC on the protection of groundwater against pollution and deterioration and Directive 2008/105/EC on environmental quality standards in the field of water policy, https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A52022PC0540 (accessed on 20 December 2022).

[141] European Commission (2018), “Technical Guidance For Deriving Environmental Quality Standards (Guidance Document No. 27)”, European Communities, Vol. 11-12 June/27.

[160] European Commission (2009), Guidance Document No. 19 Guidance on Surface Water Chemical Monitoring Under the Water Framework Directive - Common Implementation Strategy for the Water Framework Directive (2000/60/EC), WFD.

[63] European Environment Agency (2020), The European environment —state and outlook 2020. Knowledge for transition to a sustainable Europe, European Environment Agency, https://www.eea.europa.eu/soer/publications/soer-2020 (accessed on 27 June 2022).

[3] Fairbrother, A. et al. (2019), “Toward Sustainable Environmental Quality: Priority Research Questions for North America”, Environmental Toxicology and Chemistry, Vol. 38/8, pp. 1606-1624, https://doi.org/10.1002/etc.4502.

[82] Fernandez, M. (2019), “Populations Collapses in Marine Invertebrates Due to Endocrine Disruption: A Cause for Concern?”, Frontiers in Endocrinology, Vol. 10, https://doi.org/10.3389/fendo.2019.00721.

[148] Finckh, S. et al. (2022), “Endocrine disrupting chemicals entering European rivers: Occurrence and adverse mixture effects in treated wastewater”, Environment International, Vol. 170, p. 107608, https://doi.org/10.1016/j.envint.2022.107608.

[20] Goeury, K. et al. (2022), “Occurrence and seasonal distribution of steroid hormones and bisphenol A in surface waters and suspended sediments of Quebec, Canada”, Environmental Advances, Vol. 8, p. 100199, https://doi.org/10.1016/j.envadv.2022.100199.

[126] Gwak, J. et al. (2022), “Molecular Characterization of Estrogen Receptor Agonists during Sewage Treatment Processes Using Effect-Directed Analysis Combined with High-Resolution Full-Scan Screening”, Environmental Science & Technology, Vol. 56/18, pp. 13085-13095, https://doi.org/10.1021/acs.est.2c03428.

[125] Hashmi, M. et al. (2018), “Effect-directed analysis (EDA) of Danube River water sample receiving untreated municipal wastewater from Novi Sad, Serbia”, Science of The Total Environment, Vol. 624, pp. 1072-1081, https://doi.org/10.1016/j.scitotenv.2017.12.187.

[131] Hashmi, M. et al. (2020), “Effect‐Directed Analysis of Progestogens and Glucocorticoids at Trace Concentrations in River Water”, Environmental Toxicology and Chemistry, Vol. 39/1, pp. 189-199, https://doi.org/10.1002/etc.4609.

[22] Hecker, M. and H. Hollert (2009), Effect-directed analysis (EDA) in aquatic ecotoxicology: State of the art and future challenges, https://doi.org/10.1007/s11356-009-0229-y.

[78] Hewitt, L. et al. (2008), “Altered reproduction in fish exposed to pulp and paper mill effluents: roles of individual compounds and mill operating conditions”, Environmental Toxicology and Chemistry, Vol. 27/3, p. 682, https://doi.org/10.1897/07-195.1.

[31] Hollender, J. et al. (2017), “Nontarget Screening with High Resolution Mass Spectrometry in the Environment: Ready to Go?”, Environmental Science & Technology, Vol. 51/20, pp. 11505-11512, https://doi.org/10.1021/acs.est.7b02184.

[26] Hollender, J. et al. (2019), “High resolution mass spectrometry-based non-target screening can support regulatory environmental monitoring and chemicals management”, Environmental Sciences Europe, Vol. 31/1, https://doi.org/10.1186/s12302-019-0225-x.

[127] Houtman, C. et al. (2020), “High resolution effect-directed analysis of steroid hormone (ant)agonists in surface and wastewater quality monitoring”, Environmental Toxicology and Pharmacology, Vol. 80, p. 103460, https://doi.org/10.1016/j.etap.2020.103460.

[122] Houtman, C. et al. (2004), “Identification of estrogenic compounds in fish bile using bioassay-directed fractionation”, Environmental Science and Technology, Vol. 38/23, https://doi.org/10.1021/es049750p.

[153] Hughes, S., E. Maloney and A. Bejarano (2021), “Are Vertebrates Still Needed in Routine Whole Effluent Toxicity Testing for Oil and Gas Discharges?”, Environmental Toxicology and Chemistry, Vol. 40/5, pp. 1255-1265, https://doi.org/10.1002/etc.4963.

[120] Hunt, L. et al. (2017), “Species at Risk (SPEAR) index indicates effects of insecticides on stream invertebrate communities in soy production regions of the Argentine Pampas”, Science of The Total Environment, Vol. 580, pp. 699-709, https://doi.org/10.1016/j.scitotenv.2016.12.016.

[84] Hutchinson, T. et al. (2006), “Screening and Testing for Endocrine Disruption in Fish—Biomarkers As “Signposts,” Not “Traffic Lights,” in Risk Assessment”, Environmental Health Perspectives, Vol. 114/Suppl 1, pp. 106-114, https://doi.org/10.1289/ehp.8062.

[144] Ineris (2023), VGE. Valeur Guide Environnementale : Carbendazime (n°CAS 10605-21-7), French National Institute for Industrial Environment and Risks (Ineris) , Verneuil-en-Halatte.

[143] Ineris (2023), VGE. Valeur Guide Environnementale : S-Métolachlore – n° CAS : 87392-12-9, Métolachlore – n° CAS : 51218-45-2, et leurs produits de dégradation., French National Institute for Industrial Environment and Risks (Ineris), Verneuil-en-Halatte.

[155] IPCS (2002), Global Assessment of the state-of-the-science of Endocrine Disruptors, International Programme on Chemical Safety, World Health Organization, https://apps.who.int/iris/handle/10665/67357.

[140] James, A., A. Kroll and C. Minier (2023), “Towards a better consideration of endocrine disruption within the technical guidance for deriving environmental quality standards”, Regulatory Toxicology and Pharmacology, Vol. 143, p. 105457, https://doi.org/10.1016/j.yrtph.2023.105457.

[142] James, A., A. Kroll and C. Minier (2023), “Towards a better consideration of endocrine disruption within the technical guidance for deriving environmental quality standards”, Regulatory Toxicology and Pharmacology, Vol. 143, p. 105457, https://doi.org/10.1016/j.yrtph.2023.105457.

[75] Jobling, S. et al. (1998), “Widespread Sexual Disruption in Wild Fish”, Environmental Science & Technology, Vol. 32/17, pp. 2498-2506, https://doi.org/10.1021/es9710870.

[152] Kienle, C. et al. (2015), Methoden zur Beurteilung der Wasserqualität anhand von ökotoxikologischen Biotests. Ergebnisse einer Literaturrecherche und einer Expertenbefragung, Ecotox Centre, Dübendorf, https://www.centreecotox.ch/media/90556/2015_kienle_biotests_methodenempfehlung.pdf (accessed on 7 February 2023).

[17] King, O. et al. (2017), Proposed aquatic ecosystem protection guideline values for pesticides commonly used in the Great Barrier Reef catchment area: Part 1 (amended) - 2,4-D, Ametryn, Diuron, Glyphosate, Hexazinone, Imazapic, Imidacloprid, Isoxaflutole, Metolachlor, Metribuzin, Metsulfuron-methyl, Simazine, Tebuthiuron. (amended March 2018), Department of Environment and Science, Brisbane, Queensland, Australia.

[39] Könemann, S. et al. (2018), “Effect-based and chemical analytical methods to monitor estrogens under the European Water Framework Directive”, TrAC Trends in Analytical Chemistry, Vol. 102, pp. 225-235, https://doi.org/10.1016/j.trac.2018.02.008.

[42] Kortenkamp, A. and M. Faust (2018), “Regulate to reduce chemical mixture risk”, Science, Vol. 361/6399, pp. 224-226, https://doi.org/10.1126/science.aat9219.

[102] Kovacs, T. et al. (2007), Cycle 4 national investigation of cause project : final report, Environment Canada, Burlington, ON, Canada.

[108] Kovacs, T. et al. (2013), “A survey of Canadian mechanical pulp and paper mill effluents: Insights concerning the potential to affect fish reproduction”, Journal of Environmental Science and Health, Part A, Vol. 48/10, pp. 1178-1189, https://doi.org/10.1080/10934529.2013.776440.

[110] Kovacs, T. et al. (2011), “Kraft mill effluent survey: Progress toward best management practices for reducing effects on fish reproduction”, Environmental Toxicology and Chemistry, Vol. 30/6, pp. 1421-1429, https://doi.org/10.1002/etc.526.

[106] Kovacs, T., P. Martel and M. Ricci (2007), “A Shortened Adult Fathead Minnow Reproduction Test Developed for Investigation of Cause and Investigation of Solution Work Involving Pulp and Paper Mill Effluents”, Water Quality Research Journal, Vol. 42/2, pp. 91-100, https://doi.org/10.2166/wqrj.2007.012.

[32] Krauss, M. et al. (2019), “Prioritising site-specific micropollutants in surface water from LC-HRMS non-target screening data using a rarity score”, Environmental Sciences Europe, Vol. 31/1, p. 45, https://doi.org/10.1186/s12302-019-0231-z.

[68] Krewski, D. et al. (2010), “Toxicity Testing in the 21st Century: A Vision and a Strategy”, Journal of Toxicology and Environmental Health, Part B, Vol. 13/2-4, pp. 51-138, https://doi.org/10.1080/10937404.2010.483176.

[114] La Merrill, M. et al. (2020), “Consensus on the key characteristics of endocrine-disrupting chemicals as a basis for hazard identification”, Nature Reviews Endocrinology, Vol. 16/1, pp. 45-57, https://doi.org/10.1038/s41574-019-0273-8.

[93] Lowell, R. et al. (2005), Nwri scientific assessment report series no. 5. National assessment of pulp and paper environmental effects monitoring data: Findings from cycles 1 through 3., National water research institute, Burlington, ON, Canada.

[34] Luo, Y. et al. (2022), “Chemical and biological assessments of environmental mixtures: A review of current trends, advances, and future perspectives”, Journal of Hazardous Materials, Vol. 432, p. 128658, https://doi.org/10.1016/j.jhazmat.2022.128658.

[98] MacLatchy, D. et al. (2010), “Reproductive Steroid Responses in Fish Exposed to Pulp Mill Condensates: An Investigation of Cause Case Study”, Water Quality Research Journal, Vol. 45/2, pp. 163-173, https://doi.org/10.2166/wqrj.2010.020.

[76] Marlatt, V. et al. (2022), “Impacts of endocrine disrupting chemicals on reproduction in wildlife and humans”, Environmental Research, Vol. 208, p. 112584, https://doi.org/10.1016/j.envres.2021.112584.

[103] Martel, P. et al. (2010), Cycle 5 national investigation cause project : final report, Environment Canada, Ottawa, ON, Canada.

[111] Martel, P. et al. (2011), “Effluent monitoring at a bleached kraft mill: Directions for best management practices for eliminating effects on fish reproduction”, Journal of Environmental Science and Health, Part A, Vol. 46/8, pp. 833-843, https://doi.org/10.1080/10934529.2011.579850.

[109] Martel, P. et al. (2017), “The Relationship between Organic Loading and Effects on Fish Reproduction for Pulp Mill Effluents across Canada”, Environmental Science & Technology, Vol. 51/6, pp. 3499-3507, https://doi.org/10.1021/acs.est.6b05572.

[115] Martyniuk, C. (2018), “Are we closer to the vision? A proposed framework for incorporating omics into environmental assessments”, Environmental Toxicology and Pharmacology, Vol. 59, pp. 87-93, https://doi.org/10.1016/j.etap.2018.03.005.

[65] Martyniuk, C. et al. (2022), “Emerging concepts and opportunities for endocrine disruptor screening of the non-EATS modalities”, Environmental Research, Vol. 204, p. 111904, https://doi.org/10.1016/j.envres.2021.111904.

[24] McCord, J., L. Groff and J. Sobus (2022), “Quantitative non-targeted analysis: Bridging the gap between contaminant discovery and risk characterization”, Environment International, Vol. 158, p. 107011, https://doi.org/10.1016/j.envint.2021.107011.

[87] McMaster, M. et al. (1992), “Milt characteristics, reproductive performance, and larval survival and development of white sucker exposed to bleached kraft mill effluent”, Ecotoxicology and Environmental Safety, Vol. 23/1, pp. 103-117,https://doi.org/10.1016/0147-6513(92)90025-X.

[88] McMaster, M. et al. (1991), “Changes in hepatic mixed-function oxygenase (MFO) activity, plasma steroid levels and age at maturity of a white sucker (Catostomus commersoni) population exposed to bleached kraft pulp mill effluent”, Aquatic Toxicology, Vol. 21/3-4, pp. 199-217, https://doi.org/10.1016/0166-445X(91)90073-I.

[54] Mehinto, A. et al. (2015), “Interlaboratory comparison of in vitro bioassays for screening of endocrine active chemicals in recycled water”, Water Research, Vol. 83, pp. 303-309, https://doi.org/10.1016/j.watres.2015.06.050.

[19] Metcalfe, C. et al. (2022), “Methods for the analysis of endocrine disrupting chemicals in selected environmental matrixes”, Environmental Research, Vol. 206, p. 112616, https://doi.org/10.1016/j.envres.2021.112616.

[154] Mittal, K. et al. (2022), “Resource requirements for ecotoxicity testing: A comparison of traditional and new approach methods”, bioRxiv, https://doi.org/10.1101/2022.02.24.481630.

[94] Munkittrick, K. et al. (2009), “A REVIEW OF POTENTIAL METHODS OF DETERMINING CRITICAL EFFECT SIZE FOR DESIGNING ENVIRONMENTAL MONITORING PROGRAMS”, Environmental Toxicology and Chemistry, Vol. 28/7, p. 1361, https://doi.org/10.1897/08-376.1.

[101] Munkittrick, K. et al. (2002), “Overview of Freshwater Fish Studies from the Pulp and Paper Environmental Effects Monitoring Program”, Water Quality Research Journal, Vol. 37/1, pp. 49-77, https://doi.org/10.2166/wqrj.2002.005.

[89] Munkittrick, K. et al. (1992), “Reproductive Dysfunction and MFO Activity in Three Species of Fish Exposed to Bleached Kraft Mill Effluent at Jackfish Bay, Lake Superior”, Water Quality Research Journal, Vol. 27/3, pp. 439-446, https://doi.org/10.2166/wqrj.1992.030.

[124] Muschket, M. et al. (2018), “Identification of Unknown Antiandrogenic Compounds in Surface Waters by Effect-Directed Analysis (EDA) Using a Parallel Fractionation Approach”, Environmental Science & Technology, Vol. 52/1, pp. 288-297, https://doi.org/10.1021/acs.est.7b04994.

[60] Neale, P. et al. (2022), “Effect-based monitoring to integrate the mixture hazards of chemicals into water safety plans”, Journal of Water and Health, Vol. 20/12, pp. 1721-1732, https://doi.org/10.2166/wh.2022.165.

[57] Neale, P. et al. (2023), “Effect‐Based Trigger Values Are Essential for the Uptake of Effect‐Based Methods in Water Safety Planning”, Environmental Toxicology and Chemistry, https://doi.org/10.1002/etc.5544.

[139] Neale, P., F. Leusch and B. Escher (2020), Effect Based Monitoring in Water Safety Planning, Global Water Research Coalition, London.

[58] Neale, P., F. Leusch and B. Escher (2020), “Use of effect-based monitoring for the assessment of risks of low-level mixtures of chemicals in water on man and the environment”, Global Water Research Coalition, http://www.globalwaterresearchcoalition.net/_r4284/media/system/attrib/file/825/GWRC_EBM%20in%20WST_Factsheet%203_1%206%20April%202020.pdf (accessed on 23 December 2022).

[13] NHMRC and NRMMC (2011), Australian Drinking Water Guidelines Paper 6 National Water Quality Management Strategy., National Health and Medical Research Council and National Resource Management Ministerial Council, Canberra, Commonwealth of Australia.

[163] Niziński, P. et al. (2021), “Perchlorate – properties, toxicity and human health effects: an updated review”, Reviews on Environmental Health, Vol. 36/2, pp. 199-222, https://doi.org/10.1515/reveh-2020-0006.

[36] NORMAN Network (n.d.), NORMAN MassBank, https://massbank.eu/MassBank/.

[30] Norman Network (n.d.), Digital Sample Freezing Platform, https://dsfp.norman-data.eu/ (accessed on 15 March 2023).

[55] NWRI (2020), Bioanalytical tools for detection and quantification of estrogenic and dioxin-like chemicals in water recycling and reuse, National Water Research Institute.

[9] O’Connor, N. and D. Stevens (2019), Environmental risk assessment for emerging contaminants in sewage in Victoria, VicWater publication.

[51] OECD (2022), Summary Note of the OECD Workshop on Developing Science-Informed Policy Responses to Curb Endocrine Disruption in Freshwater, 18-19 October 2022, OECD, Paris, https://www.oecd.org/water/Summary-note-endocrine-disruption-in-freshwater.pdf (accessed on 6 February 2023).

[161] OECD (2019), Pharmaceutical Residues in Freshwater: Hazards and Policy Responses, OECD Studies on Water, OECD Publishing, Paris, https://doi.org/10.1787/c936f42d-en.

[46] OECD (2018), Considerations for Assessing the Risks of Combined Exposure to Multiple Chemicals, Series on Testing and Assessment No. 296, OECD, Environment, Health and Safety Division, Environment Directorate.

[49] OECD (2018), Considerations for Assessing the Risks of Combined Exposure to Multiple Chemicals, Series on Testing and Assessment No. 296, OECD Environment, Health and Safety Division, Environment Directorate.

[50] OECD (2018), Revised Guidance Document 150 on Standardised Test Guidelines for Evaluating Chemicals for Endocrine Disruption, OECD Series on Testing and Assessment, No. 150, OECD Publishing, Paris, https://doi.org/10.1787/9789264304741-en.

[73] OECD (2017), “Revised Guidance Document on developing and assessing Adverse Outcome Pathways (ENV/JM/MONO(2013)6)”, in OECD Series on Testing and Assessment Assessment.

[71] OECD (n.d.), Adverse Outcome Pathway Knowledge Base (AOP-KB), https://aopkb.oecd.org/index.html.

[72] OECD (n.d.), OECD Series on Adverse Outcome Pathways, OECD Publishing, Paris, https://doi.org/10.1787/2415170X.

[104] Parrott, J. et al. (2010), “Responses in a Fathead Minnow (Pimephales promelas) Lifecycle Test and in Wild White Sucker (Catostomus commersoni) Exposed to a Canadian Bleached Kraft Mill Effluent”, Water Quality Research Journal, Vol. 45/2, pp. 187-200, https://doi.org/10.2166/wqrj.2010.022.

[25] Paszkiewicz, M. et al. (2022), “Advances in suspect screening and non-target analysis of polar emerging contaminants in the environmental monitoring”, TrAC Trends in Analytical Chemistry, Vol. 154, p. 116671, https://doi.org/10.1016/j.trac.2022.116671.

[162] Pleus, R. and L. Corey (2018), “Environmental exposure to perchlorate: A review of toxicology and human health”, Toxicology and Applied Pharmacology, Vol. 358, pp. 102-109, https://doi.org/10.1016/j.taap.2018.09.001.

[16] Rehberger, K. et al. (2020), “Long-term exposure to low 17α-ethinylestradiol (EE2) concentrations disrupts both the reproductive and the immune system of juvenile rainbow trout, Oncorhynchus mykiss”, Environment International, Vol. 142, p. 105836, https://doi.org/10.1016/j.envint.2020.105836.

[4] Robitaille, J. et al. (2022), “Towards regulation of Endocrine Disrupting chemicals (EDCs) in water resources using bioassays – A guide to developing a testing strategy”, Environmental Research, Vol. 205, p. 112483, https://doi.org/10.1016/j.envres.2021.112483.

[61] Rosenmai, A. et al. (2018), “In vitro bioanalysis of drinking water from source to tap”, Water Research, Vol. 139, pp. 272-280, https://doi.org/10.1016/j.watres.2018.04.009.

[157] Russel, W. and R. Burch (1960), The Principles of Humane Experimental Technique, Methuen & Co. Limited, London, https://doi.org/10.5694/j.1326-5377.1960.tb73127.x.

[116] Saaristo, M. et al. (2018), “Direct and indirect effects of chemical contaminants on the behaviour, ecology and evolution of wildlife”, Proceedings of the Royal Society B, Vol. 285/1885, https://doi.org/10.1098/RSPB.2018.1297.

[129] Sanchez, W. et al. (2011), “Adverse effects in wild fish living downstream from pharmaceutical manufacture discharges”, Environment International, Vol. 37/8, https://doi.org/10.1016/j.envint.2011.06.002.

[90] Sandström, O., E. Neuman and P. Karås (1988), “Effects of a Bleached Pulp Mill Effluent on Growth and Gonad Function in Baltic Coastal Fish”, Water Science and Technology, Vol. 20/2, pp. 107-118, https://doi.org/10.2166/wst.1988.0051.

[14] Sardiña, P. et al. (2019), “Emerging and legacy contaminants across land-use gradients and the risk to aquatic ecosystems”, Science of The Total Environment, Vol. 695, p. 133842, https://doi.org/10.1016/j.scitotenv.2019.133842.

[53] SCCRWP (2014), Development of bioanalytical techniques for monitoring constituents/chemicals of emerging concern (CECs) in recycled water applications for the state of California, California Water Boards.

[119] Schäfer, R. et al. (2007), “Effects of pesticides on community structure and ecosystem functions in agricultural streams of three biogeographical regions in Europe”, Science of The Total Environment, Vol. 382/2-3, pp. 272-285, https://doi.org/10.1016/j.scitotenv.2007.04.040.

[158] Scholz, S. et al. (2013), “A European perspective on alternatives to animal testing for environmental hazard identification and risk assessment”, Regulatory Toxicology and Pharmacology, Vol. 67/3, pp. 506-530, https://doi.org/10.1016/j.yrtph.2013.10.003.

[99] Shaughnessy, K. et al. (2007), “Effects of kraft pulp mill condensates on plasma testosterone levels in mummichog (Fundulus heteroclitus)”, Ecotoxicology and Environmental Safety, Vol. 67/1, pp. 140-148, https://doi.org/10.1016/j.ecoenv.2006.04.003.

[83] Siddig, A. et al. (2016), “How do ecologists select and use indicator species to monitor ecological change? Insights from 14 years of publication in Ecological Indicators”, Ecological Indicators, Vol. 60, pp. 223-230, https://doi.org/10.1016/j.ecolind.2015.06.036.

[123] Simon, E. et al. (2013), “Effect-directed analysis to explore the polar bear exposome: Identification of thyroid hormone disrupting compounds in plasma”, Environmental Science and Technology, Vol. 47/15, https://doi.org/10.1021/es401696u.

[33] Sims, K. (2022), ECG Environmental Brief No. 35 - Chemicals of concern: a prioritisation and early warning system for England, Bulleting from the Royal Society of Chemistry and Environmental Chemistry Group.

[80] Smith, B. (1981), “Male characteristics on female mud snails caused by antifouling bottom paints”, Journal of Applied Toxicology, Vol. 1/1, pp. 22-25, https://doi.org/10.1002/jat.2550010106.

[77] Sumpter, J. (2005), “Endocrine Disrupters in the Aquatic Environment: An Overview”, Acta hydrochimica et hydrobiologica, Vol. 33/1, pp. 9-16, https://doi.org/10.1002/aheh.200400555.

[136] Swiss Federal Council (1998), Waters Protection Ordinance of 28 October 1998, Status as of 1 February 2023.

[118] Tlili, A. et al. (2016), “Pollution-induced community tolerance (PICT): towards an ecologically relevant risk assessment of chemicals in aquatic systems”, Freshwater Biology, Vol. 61/12, pp. 2141-2151, https://doi.org/10.1111/fwb.12558.

[105] van den Heuvel, M. et al. (2010), “Evaluation of Short-Term Fish Reproductive Bioassays for Predicting Effects of a Canadian Bleached Kraft Mill Effluent”, Water Quality Research Journal, Vol. 45/2, pp. 175-186, https://doi.org/10.2166/wqrj.2010.021.

[135] van der Oost, R. et al. (2017), “SIMONI (Smart Integrated Monitoring) as a novel bioanalytical strategy for water quality assessment: Part II–field feasibility survey”, Environmental Toxicology and Chemistry, Vol. 36/9, https://doi.org/10.1002/etc.3837.

[134] van der Oost, R. et al. (2017), “SIMONI (smart integrated monitoring) as a novel bioanalytical strategy for water quality assessment: Part I – model design and effect-based trigger values”, Environmental Toxicology and Chemistry, Vol. 36/9, https://doi.org/10.1002/etc.3836.

[15] Victorian Government (2017), Environment Protection Act 2017, No. 51 of 2017, Authorised Version No. 011, https://content.legislation.vic.gov.au/sites/default/files/2023-06/17-51aa011-authorised.pdf.

[45] Vinggaard, A. et al. (2022), “P15-08 PANORAMIX: Providing risk assessments of complex real-life mixtures for the protection of Europe’s citizens and the environment”, Toxicology Letters, Vol. 368, p. S217, https://doi.org/10.1016/j.toxlet.2022.07.589.

[48] Vinggaard, A. et al. (2022), “P15-08 PANORAMIX: Providing risk assessments of complex real-life mixtures for the protection of Europe’s citizens and the environment”, Toxicology Letters, Vol. 368, p. S217, https://doi.org/10.1016/j.toxlet.2022.07.589.

[70] Vinken, M. et al. (2021), “Safer chemicals using less animals: kick-off of the European ONTOX project”, Toxicology, Vol. 458, p. 152846, https://doi.org/10.1016/j.tox.2021.152846.

[137] VSA Platform for Water Quality (n.d.), About the Modular Stepwise Procedure, https://modul-stufen-konzept.ch/en/about-the-msp/ (accessed on 25 February 2023).

[35] Wang, M. et al. (2016), “Sharing and community curation of mass spectrometry data with Global Natural Products Social Molecular Networking”, Nature Biotechnology, Vol. 34/8, pp. 828-837, https://doi.org/10.1038/nbt.3597.

[5] Wernersson, A. et al. (2015), “The European technical report on aquatic effect-based monitoring tools under the water framework directive”, Environmental Sciences Europe, Vol. 27/7, https://doi.org/10.1186/s12302-015-0039-4.

[62] WHO and IWA (n.d.), Water Safety Portal, https://wsportal.org/ (accessed on 3 April 2023).

[21] WHO-UNEP (2013), State of the Science of Endocrine Disrupting Chemicals 2012, Inter-Organisation Programme for the Sound Management of Chemicals, United Nations Environment Programme and the World Health Organization, https://apps.who.int/iris/handle/10665/78101.

[74] Windsor, F., S. Ormerod and C. Tyler (2018), “Endocrine disruption in aquatic systems: up-scaling research to address ecological consequences”, Biological Reviews, Vol. 93/1, https://doi.org/10.1111/brv.12360.

[147] Working Group Chemicals (2021), Technical Proposal for Effect-Based Monitoring and Assessment under the Water Framework Directive: Report to the Common Implementation Strategy (CIS) Working Group Chemicals on the outcome of the work performed in the subgroup on Effect-Based Methods (EBM), Sub-group for effect-based methods of the CIS Working Group Chemicals, Directorate-General Environment, https://www.normandata.eu/sites/default/files/files/Highlights/211013_EBM%20report_FINAL_WG_Chem_Oct_2021%20%281%29.pdf (accessed on 7 February 2023).

[128] Zwart, N. et al. (2018), “High-Throughput Effect-Directed Analysis Using Downscaled in Vitro Reporter Gene Assays to Identify Endocrine Disruptors in Surface Water”, Environmental Science and Technology, Vol. 52/7, https://doi.org/10.1021/acs.est.7b06604.

As water samples are modified in the sample preparation process, some chemicals may be lost. For targeted chemistry, the sample preparation will be selective to the chemicals desired and will have high recovery (Metcalfe et al., 2022[19]). However, for bioassays and non-targeted chemistry, the preparation process aims to keep as many chemicals, while preventing the matrix interference during the analysis (Paszkiewicz et al., 2022[25]; Luo et al., 2022[34]; Robitaille et al., 2022[4]). This means that, even with a lot of effort, some chemicals will inevitably be lost. For example, for in vitro bioassays, the most used method is solid-phase extraction (SPE) (Luo et al., 2022[34]; Robitaille et al., 2022[4]) which are columns containing sorbents similar to the one used in passive sampling. The water will be passed through the sorbent which will trap certain chemicals. One important notion to understand is that while sorbents (e.g., HLB) are designed to catch as many chemicals as possible, it is not possible to retain all, though multiple solid-phase extraction methods with different sorbents can be used for a given sample. Most methods used currently (Luo et al., 2022[34]; Robitaille et al., 2022[4]) will use sorbents that keep mostly hydrophobic molecules, i.e. molecules that do not like water which encompass a majority of EDCs (Escher, Neale and Leusch, 2021[159]; Robitaille et al., 2022[4]). However, some chemicals, such as metals, will be lost in the process (De Baat et al., 2020[145]). It is important to take this limitation into account as some EDCs will be removed in the sampling process. Perchlorate, which can disrupt the thyroid axis, is a case in point (Pleus and Corey, 2018[162]; Niziński et al., 2021[163]).


← 1. Joint ED-list by Belgium, Denmark, France, Netherlands, Spain and Sweden.

Metadata, Legal and Rights

This document, as well as any data and map included herein, are without prejudice to the status of or sovereignty over any territory, to the delimitation of international frontiers and boundaries and to the name of any territory, city or area. Extracts from publications may be subject to additional disclaimers, which are set out in the complete version of the publication, available at the link provided.

© OECD 2023

The use of this work, whether digital or print, is governed by the Terms and Conditions to be found at https://www.oecd.org/termsandconditions.