2. Opportunities to build a policy-relevant knowledge base

Improving knowledge on pharmaceuticals in water, and their effects and risks on human health and the environment, is an important foundation on which pharmaceutical authorisation, environmental risk assessments and water quality policies can be built. This chapter inventories and assesses the strengths and weaknesses of various innovative monitoring and modelling approaches to assess the impacts and risks of pharmaceuticals in water. The chapter takes stock of country and international initiatives to improve the knowledge base and highlights the need for data sharing and institutional coordination.

    

2.1. Key messages

Numerous active pharmaceutical ingredients (APIs) have been detected in surface waters as a result of advancements in analytical technologies. Certain pharmaceuticals are being monitored in surface waters of OECD countries according to watch-lists, although the vast majority of them remain unmonitored. Furthermore, many pharmaceuticals in use have not been assessed for their environmental impact, and the environment is not considered in the risk-benefit analysis when authorising new medicines for human use.

Knowledge gaps persist on the sources, transport and fate of APIs (including transformation products and metabolites) in water bodies, and their toxicological and chronic impacts (including additive effects of mixtures) to a multitude of targets in organisms, populations, and communities, over short and long-terms. Member countries are looking for innovative ways to maximise the benefits of existing knowledge, and develop new tools and decision frameworks that are cost- and time-effective.

Without adequate knowledge of the potential hazards APIs may pose, assessing their environmental risk is challenging. As it would be impractical (technically and economically) to commence analytical monitoring en masse for all APIs in the environment, more integrated and holistic approaches to environmental monitoring are needed, in combination with modelling, prioritisation methods and data sharing initiatives.

In order to address knowledge gaps, perform robust environmental risk assessments and design informed policy responses, it is necessary to establish standardisation and guidance on the best practices in analytical methods and risk assessment, data quality and exchange, and prioritisation of pharmaceuticals and susceptible water bodies. Country and international initiatives are crucial to improve the knowledge base, and the exchange and review of data.

Lastly, it is important to note that improving knowledge is not a pollution reduction measure in itself. Governments should take advantage of alternative innovative monitoring and modelling technologies to simplify the prioritisation and identification of APIs of concern, and improve the speed and quality of risk assessments and cost benefit analysis, but they should not wait for exact science before taking action.

2.2. Environmental risk assessment and authorisation of pharmaceuticals

2.2.1. Introduction

One of the key legislative factors influencing the presence of the pharmaceuticals in the environment is the current framework for Environmental Risk Assessment (ERA), which is a part of the Market Authorisation process of new pharmaceuticals. The objective of an ERA is to determine the potential adverse effects that pharmaceuticals pose to ecological health, and is a combined evaluation of hazards and exposure. Toxicity is determined through toxicity testing (in vivo), which involves an assessment of the harm to aquatic organisms. To assess whether exposure levels are safe, scientists calculate the ratio of the predicted environmental concentration (PEC; based on distributions) to the predicted no-effect concentration (PNEC; the predicted level below which there are no negative consequences for the ecosystem, or an ecological safety threshold). Risk is identified when substances are present in water at higher levels than would be safe for the environment (PEC:PNEC ratios greater than one) (EMA, 2006[1]). Of specific concern are substances which are persistent, bioaccumulative and toxic (PBT), or very persistent and very bioaccumulative (vPvB) (see section 1.4.2), and substances with endocrine disruptive (where not intended), mutagenic or carcinogenic properties.

Most (88%) of the pharmaceuticals targeting human proteins do not have comprehensive environmental toxicity data, or they are not publically available. This highlights the need for both intelligent approaches to prioritise legacy human drugs for a tailored environmental risk assessment and a transparent database that captures environmental data (Gunnarsson et al., 2019[2]).

Due to increased awareness regarding pharmaceuticals in the environment, pharmaceutical producers and regulatory bodies have started to assess potential environmental risks when designing and authorising new pharmaceuticals. The following sections outline current ERA practice in countries regarding human and veterinary pharmaceuticals, and how they are used in decision-making when new pharmaceuticals are being authorised.

2.2.2. Environmental risk assessment of new pharmaceuticals in practice

Within the EU, since 2006, an ERA is required for all new marketing authorisations of human pharmaceuticals according to Article 8(3) of Directive 2001/83/EC (Box 2.1). Pharmaceuticals entered on the market before 2006 do not require an ERA in a retrospective manner, unless a line extension (new version or an enhancement of an existing product) is requested. An important point to note is that even if a risk is identified in the ERA, it is not considered in the benefit-risk assessment of authorisation, and therefore cannot be used as grounds for refusal in marketing. However, on a case-by-case basis, specific arrangements to limit impact on the environment can be considered including, for example, specific labelling (EMA, 2006[1]).

Box 2.1. Environmental Risk Assessment in the authorisation process of new human pharmaceuticals (post 2006), EU

Realising that pharmaceuticals could pose an environmental risk, the European Medicines Agency developed guidance for ERAs in 2006. During the authorisation process, the applicant is required to evaluate environmental impact, in terms of exposure and effects, and submit the assessment of the potential risk. The assessment is a two-phase procedure (see figure below) initiated with Phase I, where the predicted environmental concentration (PEC) for surface water is calculated and the distribution is measured. If the PEC is equal to, or above, a trigger value of 0.01 μg/L, a second phase of analysis is carried out. Substances that are known to affect the reproduction of organisms at low concentrations, or have the potential to bioaccumulate, should also enter Phase II.

Phase II is a two-tiered (A and B) environmental fate and effect analysis. In this phase, the properties of the substance (persistence, bioaccumulation and toxicity) are investigated. In Phase II A, PNECs are calculated for surface water, groundwater and microorganisms based on a standard long-term toxicity test on fish, daphnia and algae (e.g. according to OECD Test Guidelines 201, 211 and 210). When the results suggest potential risk, Phase II is extended for further evaluation on the fate of the substance and/or its metabolites in the aquatic environment (Phase II B) (EMA, 2006[1]).

Figure 2.1. Outline of the three steps of the Environmental Risk Assessment for medicinal products for human use, EU

Sources: (Caneva et al., 2014[3]) (EMA, 2006[1]).

Similar to the EU, ERAs are required by the U.S. Food and Drug Administration (FDA) for new (post 2008) drug applications (with some exceptions). New FDA (2016[4]) guidance supplements the Environmental Assessment Guidance issued in (1998[5]) by addressing specific considerations for drugs that have potential oestrogenic, androgenic, or thyroid hormone pathway activity in the environment, and the conditions under which the sponsor should submit an environmental assessment or may apply for a claim of categorical exclusion. However, also like the EU, the results are not considered as part of the benefit-risk assessment of authorisation.

Canada’s New Substances Program is responsible for administering the New Substances Notification Regulations (Chemicals and Polymers) and the New Substances Notification Regulations (Organisms) of the Canadian Environmental Protection Act, 1999 (CEPA). Collectively known as the ‘Regulations’, they are an integral part of the federal government's national pollution prevention strategy. As part of the "cradle to grave" management approach for toxic substances laid out in CEPA, the Regulations were created to ensure that no new substances (chemicals, polymers or animate products of biotechnology) are introduced into the Canadian marketplace before an assessment of whether they are potentially toxic has been completed, and any appropriate or required control measures have been taken. In September 2001, substances in products regulated by the Food & Drugs Act became subject to the Regulations. This includes substances used in pharmaceuticals, veterinary drugs, biologics (including genetic therapies), medical devices, cosmetics and personal care products, food additives, novel foods and natural health products. Under the Regulations, a New Substances Notification package, containing all information prescribed in the Regulations, may be required before a new substance can be imported into or manufactured in Canada. The type of information required and the timing of the notification will depend on such factors as the type of substance, the quantity that will be imported or manufactured, the intended use of the substance and the circumstances associated with its introduction. When a potential risk to human health or the environment is identified for a new substance, CEPA empowers the Government of Canada to develop risk management measures prior to or during the earliest stages of its introduction into Canada. This ability to act early makes the New Substances Program a unique and essential component of the federal management of toxic substances.

The International Cooperation on Harmonisation of Technical Requirements for Registration of Veterinary Medicinal Products (VICH), launched in 1996, provides a basis for harmonising technical requirements for veterinary product registration in the EU, U.S. and Japan. Observers of the VICH also include Australia, New Zealand, Canada and South Africa. The VICH platform stipulates guidelines on how environmental impact assessments (EIA) should be performed (Box 2.2), but does not constitute a role to provide guidance to establish regulatory systems or regulations for marketing authorisation of veterinary pharmaceuticals (VICH, 2018[6]).

Box 2.2. The International Cooperation on Harmonisation of Technical Requirements for Registration of Veterinary Medicinal Products: Guidance on EIA

The International Cooperation on Harmonisation of Technical Requirements for Registration of Veterinary Medicinal Products (VICH) provides guidance on carrying out environmental impact assessments of veterinary pharmaceuticals. The guidance builds on a two-phase approach, similar to the one used for human pharmaceuticals in the EU (see Box 2.1):

• Phase I builds on a series of questions that the applicant must answer and predicted environmental concentration (PEC) calculations, which decides if the pharmaceutical in question should enter Phase II investigation. The trigger values (i.e. calculated PEC) for Phase II are 100 µg/kg and 1 µg/L for soil and indirect aquatic exposure, respectively. Certain therapeutic groups automatically enter Phase II regardless of the PEC calculation, e.g. ecto-and endoparasiticides intended for use in livestock (VICH, 2000[7]).

• Phase II requires an assessment of the environmental fate and effects of the pharmaceutical, and includes studies of: (1) physico-chemical properties (e.g. water solubility, octanol/water partitioning); (2) binding to soils; (3) biodegradation in soil or aquatic test systems; and (4) acute effects of the drug residue on select aquatic and terrestrial species (Tier A). The approach is not fully harmonised in the VICH regions due to differences in animal husbandry and land-use practices. Results indicating a risk can require further study (Tier B), for example, when substances have a logKow distribution greater than 4, which signals potential for bioaccumulation (EMA, 2004[8]).

Sources: (EMA, 2004[8]) (VICH, 2000[7]).

Within the EU, in contrast to human pharmaceuticals, the marketing authorisation of veterinary medicines does require the environmental risks to be included in the benefit-risk assessment (Directive 2001/82/EC, as amended by Directive 2004/28/EC), i.e. the benefits are weighed up against the environmental risks before putting them on the market (EMA, 2009[9]). However, the current guidelines on benefit-risk are not clear on how trade-offs should be assessed (Chapman et al., 2017[10]). Generally, veterinary pharmaceuticals are authorised if the benefits are thought to outweigh the environmental risks or because no alternative treatment is available (UBA, 2018[11]). In 2016, a guideline was adopted within the EU for hazard-based assessment of PBT and vPvB substances used in veterinary medicinal products. So far, all potential PBTs identified belong to the therapeutic group of parasiticides (EMA, 2017[12]).

Since the introduction of pharmaceutical regulations in EU, the German Environment Agency (UBA) has been evaluating ERAs for human and veterinary medicines before they are marketed. They estimate that 10% of pharmaceutical products indicate a potential environmental risk (Küster and Adler, 2014[13]). Of greatest concern are hormones, antibiotics, analgesics, antidepressants and anticancer pharmaceuticals used for human health, and hormones, antibiotics and parasiticides used as veterinary pharmaceuticals (Küster and Adler, 2014[13]).

2.2.3. Concerns with current environment risk assessment of pharmaceuticals

Several researchers and OECD member countries stress the limitations of current practices for ERAs for the authorisation of human and veterinary pharmaceuticals:

• ERAs are not considered in the risk-benefit analysis of marketing authorisation for human pharmaceuticals; current approval for human pharmaceuticals is based on safety, efficacy and quality. This issue has been raised by several scientists (e.g. (Ågerstrand et al., 2015[14]) (Küster and Adler, 2014[13])) as well as within government national action plans related to pharmaceuticals in the environment. Germany for instance, stresses the need to incorporate the environment as a factor in the benefit-risk balance (similar to the framework for veterinary pharmaceuticals). Sweden has also pushed for the inclusion of environmental risk in the benefit-risk analysis of human pharmaceuticals, as a driver for pharmaceutical companies to take greater responsibility and implement environmental risk mitigation measures when necessary.

• ERAs are not required retrospectively for already-authorised pharmaceuticals (pre 2006 in the EU and pre-2008 in the US) that may be of high environmental risk, and there is no formal mechanism to review assessments as information on the risks improve. Potential environmental risks of human and veterinary pharmaceuticals have been identified for different pharmaceutical classes such as parasiticides, analgesics, antidepressants, antibiotics, hormones and anticancer medicines (see section 1.4). Within the EU, there are about 2000 veterinary pharmaceutical products on the market, most of which have not been fully tested for their toxicity (Kools et al., 2008[15]). Likewise, most (88%) human pharmaceuticals do not have comprehensive environmental toxicity data, or they are not publically available (Gunnarsson et al., 2019[2]).

• In the current ERA process, some pharmaceuticals do not reach the trigger value for a thorough assessment (i.e. may stop after the first phase of an ERA). For example, endocrine active compounds have to be assessed, but others, such as cytotoxics, are not assessed despite their potential high hazard (Kümmerer et al., 2016[16]). ERAs also lack consideration of antimicrobial resistance; a more diverse selection of bacteria in ERA would increase protectiveness of ERA (Ågerstrand et al., 2015[14]; Le Page et al., 2017[17]; Le Page et al., 2019[18]).

• ERAs for pharmaceuticals are product-based, and not assessed per API used within the product. This may lead to different conclusions about the risk for the same API used in different pharmaceutical products. It also duplicates the laboratory work, and the animals being used for testing (Ågerstrand et al., 2015[14]). Moreover, this complicates the search for information in a systematic way.

• There is no uniform regulation for ERA of mixture and additive effects. There is a lack of peer-reviewed toxicity data for commonly identified chemical mixtures which would enable assessment of the hazard as a whole rather than being based on individual components (WHO, 2017[19]). For example, Cleuvers (2008[20]) found that toxicity of a mixture of non-steroidal anti-inflammatory drugs against Daphnia was considerably higher even at concentrations in which the single substances showed no, or only very slight, effects. Reproduction was decreased by 100% at concentrations where no effects on survival could be observed, which means that this destructive effect on the Daphnia population would be totally overlooked by an acute test using the same concentrations. Section 1.4.2 outlines how mixtures of pharmaceuticals, and with other contaminants in the environment, can possess a joint toxicity greater than the toxicity of individual substances. Some efforts have been made within the EU to establish guidance for chemical mixtures, for example, the State of Art Report on Mixture Toxicity (Kortenkamp Andreas and Faust, 2009[21]) and the Communication from the European Commission on Combination effects of Chemicals1 (Godoy and Kummrow, 2017[22]).

• There is no uniform regulation for ERA of metabolites and transformation products, and multiple routes of exposure in OECD countries. The ecotoxicity of some metabolites and transformation products can have higher additive and toxicity effects than that of their parent pharmaceutical compounds (Godoy and Kummrow, 2017[22]). Furthermore ecosystems and humans may be continuously exposed via a number of pathways of low-dose mixtures that can have additive effects (Backhaus, 2014[23]).

• When, as identified by ERA, human or veterinary pharmaceuticals pose a risk to the environment, risk mitigation measures can be recommended. However, compliance with risk mitigation measures is not enforced and is essentially voluntary (BIO Intelligence Service, 2013[24]).

• ERAs and pharmaceutical authorisation is resource-intensive, although it is less costly than pesticide or biocide registration. In 2016, the average number of new active pharmaceutical ingredients registered by OECD governments was 34, and the cost of non-clinical testing of such substances is likely to be several million euros (OECD, 2019[25]).

• Lack of involvement of different stakeholders and data sharing. The framework for ERAs for pharmaceuticals is unlike other regulations of chemicals (e.g. the EU Registration, Evaluation, Authorisation and Restriction of Chemicals (REACH) regulation) where data sharing is encouraged and stakeholders are invited to comment on draft opinions of the authorisation process and ERAs (Ågerstrand et al., 2017[26]). Since ERA information cannot be cross-referenced, data cannot be reused from one dossier to another, even if the concerned medicinal products contain the same API.

Further thinking on the ERA framework is needed to address the above challenges, including the potential for grouping risks based on the properties of pharmaceuticals (e.g. toxicity, mobility, persistence) and the receiving water (e.g. surface water, groundwater, drinking water) in order to predict, identify and mitigate future emerging environmentally persistent pharmaceutical pollutants. There is a need to better understand human and environmental exposures, through the use of both investigative monitoring and modelling.

Although the efficiency, efficacy and cost of undertaking risk assessments, as well as cost of the control of potential mixtures in relation to the health benefits of pharmaceuticals need to be carefully considered, a number of useful tools and models have been, and continue to be, developed to carry out ERA of chemical mixtures and overcome some of the challenges listed above. For example, OECD guidance on Considerations for Assessing the Risks of Combined Exposure to Multiple Chemicals (OECD, 2018[27]) can help. The World Health Organisation (2017[19]) also provides guidance on the risk assessment and management of chemical mixtures (Box 2.3).

More systematic studies will help to further the understanding of the transport, occurrence, exposure pathways and fate of pharmaceuticals in the environment, as well as susceptible species and relevant effects endpoints. Standardisation of protocols for sampling and analysing pharmaceuticals would help to facilitate the comparison of data (WHO, 2012[28]; WHO, 2017[19]). The following sections of this chapter look at existing frameworks for monitoring pharmaceuticals in the environment, and documents new research developments in monitoring technologies and approaches.

Box 2.3. UN World Health Organisation guidance on managing the risks of chemical mixtures

The WHO proposes a number of questions that can assist in making decisions about whether to treat a group of substances as a mixture for risk assessment, and what management approaches may be considered.

• Do the chemicals always occur as a mixture and, if not, how frequently do they occur together and under what circumstances?

• Does the proportion of substances vary and is there a small number that usually dominate?

• Are the substances of similar water solubility?

• Can one or two substances act as a surrogate for the others (for both risk assessment and management)?

• How stable is the mixture (i.e. is it always similar)?

• Can the components of the mixture be measured by the same method?

• How readily will components of the mixture be removed in the available wastewater and drinking-water treatment?

• Are there other upstream interventions that can be applied?

Source: (WHO, 2017[19]).

2.3. Existing frameworks for monitoring pharmaceuticals in water

Monitoring refers to repeated measurements of predetermined, specific endpoints over broad spatial and temporal scales aiming to assess the status and trends of water bodies (Ekman et al., 2013[29]). Many OECD countries have monitoring programmes to measure the concentrations of selected pharmaceuticals in surface water in order to determine appropriate measures to address the risk posed by these substances in the future (e.g. inclusion in systematic monitoring or the development of Environmental Quality Norms). However, they monitor only a selected number of pharmaceuticals. Groundwater and drinking water sources are less frequently monitored. In developing economies, monitoring of pharmaceuticals is often not a priority. Monitoring data is particularly under-represented in Asia, Africa and South America - the very regions of the world that are likely to have the highest consumption and release of pharmaceuticals due to high population density, limited wastewater treatment (aus der Beek et al., 2016[30]), and in some locations, pharmaceutical manufacturing. This section provides a brief summary of selected pharmaceutical monitoring programmes in OECD countries.

The EU Watch List under the Water Framework Directive requires monitoring of several pharmaceuticals (hormones and antibiotics) in surface water (Box 2.4). Research is underway to develop a groundwater watch list (Box 2.5). Korea created a candidate list of contaminants of emerging concern (CECs) for surface water monitoring in 2016 that includes eight pharmaceuticals (Acetylsalicylic acid, Sulfamethoxazole, Sulfamethazine, Sulfathiazole, Naproxen, Clarithromycin, Trimethoprim and Carbamazepine). In the U.S., the 1996 Safe Drinking Water Act (SDWA requires) that once every five years, the EPA publish a list of currently unregulated contaminants, known as the Contaminant Candidate List (CCL). In developing the CCL, the SDWA directs the EPA to consider health effects and occurrence information for unregulated contaminants and further specifies that the Agency place those contaminants on the list that present the greatest public health concern related to exposure from drinking water. As part of the CCL process, the EPA evaluates contaminants that are generally considered to be contaminants of emerging concern in drinking water, including pharmaceuticals. The SDWA also requires the EPA to monitor public water systems, every five years, for no more than 30 unregulated contaminants under the Unregulated Contaminant Monitoring Rule (UCMR). The EPA’s selection of contaminants for a particular UCMR cycle is largely based on a review of the CCL (US EPA, 2016[31]).

Box 2.4. Monitoring of pharmaceuticals in surface water, as required under the EU Water Framework Directive

Under the Water Framework Directive (WFD), the surface water Watch List is a list of potential water pollutants that are required to be monitored and reported by EU Member States to determine the risk they pose to the aquatic environment and whether EU Environmental Quality Norms (EQN) should be set for them. The list is reviewed every 2 years.

The first Watch List was published in 2015. It included 10 substances or groups of substances which included the hormones 17-Alpha-ethinylestradiol (EE2), 17-Beta-estradiol (E2) and oestrone (E1), and the pain killer diclofenac.

In 2018, the Watch List was updated to remove diclofenac as sufficiently high-quality monitoring data had been collected. Five antibiotics were added to the list: 17-alpha-ethinylestradiol (EE2), 17-beta-estradiol (E2), oestrone (E1), macrolide antibiotics (erythromycin, clarithromycin, azithromycin), amoxicillin and ciprofloxacin. The inclusion of antibiotics is consistent with the European One Health Action Plan against Antimicrobial Resistance (AMR), which supports the use of the Watch List to improve knowledge of the occurrence and spread of antimicrobials in the environment.

Source: (EC, 2018[32])

Box 2.5. EU voluntary watch list for groundwater monitoring of pharmaceuticals

A Watch List for substances in groundwater is currently being developed in accordance with EU Directive 2014/80/EU (amending the Groundwater Directive, 2006/118/EC, under the WFD umbrella) to increase the availability of monitoring data on substances posing a risk or potential risk to bodies of groundwater, for which groundwater quality standards or threshold values should be set. Contrary to the Watch List for surface water, the monitoring will be voluntary (instead of mandatory).

Twelve EU member states submitted pharmaceutical datasets for a pilot-study aiming to gather and identify evidence to derive a watch list for groundwater. In these 12 member states, a wide range of pharmaceuticals (in total, approximately 300 different substances) were monitored and detected in groundwater. The table below presents the 17 substances most frequently detected above 0.1 µg/L in all 12 member states. Carbamazepine was most widely analysed, followed by diclofenac. Paracetamol was the most frequently detected substance as a percentage of sites monitored (24% of the sites). In addition to these substances, diatrizoic acid, primidon, ibuprofen and clofibrate are prominent in the summary data.

Table 2.1. Substances being detected above 0.1 µg/L more than twice in groundwater of 12 EU Member States
Substance	Total number of sites	>LoQa	>0.1 µg/l
ethylenediaminetetraacetic acid (EDTA)	105	61	61
carbamazepine	3692	445	57
ibuprofen	1865	49	15
diatrizoic acid	1394	66	15
diclofenac	3060	34	13
gabapentin	167	19	13
metformin	787	56	12
paracetamol (acetaminophen)	1036	248	11
clofibrate	2294	46	10
primidon	628	31	10
pentoxifylline	859	9	9
acetylsalicylic acid	600	7	7
phenazone	1477	40	5
sulfamethoxazole	2153	105	4
ioxithalamic acid	329	12	4
erythromycin	1255	25	3
iopamidol	1462	37	3
a. Number of samples above limit of quantification (LoQ)
Source: (Marsland and Roy, 2016[33]).

Taking monitoring results one step further, environmental quality norms (EQNs) can be developed to define the maximum allowable concentration of a single substance in water to protect ecosystems, and drinking water and bathing resources. The EU Water Framework Directive (WFD) (2000/60/EC), among others, builds on EQNs to secure water quality.

The European Commission suggested EQNs for each of the pharmaceuticals on the surface water Watch List (Box 2.4), but they are not incorporated in EU-legislation. Under the WFD, member states are required to select additional substances of national or local concern (so called river basin-specific substances), and to define corresponding EQNs (WFD Annex VIII) (European Commission, 2017[34]). Sweden has proposed 17 pharmaceuticals for monitoring in addition to the WFD Watch List, based on PBT properties, large usage, and/or detection in fish, surface water, drinking water and sludge (MPA, 2015[35]). In addition, Sweden has incorporated EQNs for 4 pharmaceuticals (Ciprofloxacin, Diclofenac, E2 and EE2) as river basin-specific substances according to the Swedish Agency for Marine and Water Management statues HVMFS 2018:17. The U.K has transposed obligation to monitor substances on the watch-list into national legislation. The Netherlands has recently proposed EQNs for Carbamazepine, Metoprolol and Metformin (RIVM, 2014[36]). Australia has developed regulatory guidelines for water recycling including drinking water-based guidelines (ranging from 0.35 to 1050 μg/L) for antibiotics, non-steroidal anti-inflammatories, β-adrenergic blockers, estrogenic hormones, and other general pharmaceuticals (Australian Government, 2008[37]).

It has been acknowledged that the number of pharmaceuticals measured by target analysis is not sufficient to provide an exhaustive overview of water quality; target-based environmental monitoring neglects unknown but potentially substantial portions of constituents in the aquatic environment (Daughton, 2004[38]). Hence, substances that are often detected can lead to the assumption that they are the most common, when in reality they are simply the most studied (aus der Beek et al., 2016[30]). Conversely, the workload and cost of monitoring, assessing and reviewing data for increasingly lengthy lists of substances is not sustainable, nor realistic. Screening is one way of identifying new candidates for future inclusion in monitoring programmes. Box 2.6 describes the process of the Danish approach to selecting pharmaceuticals for screenings. Switzerland has prioritised five indicator substances to reduce analytical costs of monitoring for an extensive list of CECs (Box 2.7).

Box 2.6. Monitoring and screening programme for pharmaceuticals, Denmark

In Demark, a national monitoring programme (NOVANA) governed by the Danish Environmental Protection Agency monitors EU priority substances and national river basin-specific substances in water. In addition to the NOVANA programme, Denmark also performs annual screening projects with the aim of identifying new candidates for future inclusion in the national monitoring programme. This approach is considered cost effective in order to collect knowledge of the presence of specific substances in the aquatic environment. Two pharmaceutical screenings have been performed, in 2008 and 2015. The following assumptions and requirements were used in order to select pharmaceuticals for the screenings:

• Pharmaceuticals specifically mentioned in the scientific literature as a potential high environmental risk

• Pharmaceuticals representing various applications

• Pharmaceuticals in high use in primary (hospital) and secondary (outpatient) care

• Pharmaceuticals that could be reliably analysed with available monitoring technologies.

The 2015 study selected and screened for 27 human pharmaceuticals that had not previously been included in the national monitoring programme (see Table below). Samples were taken from WWTPs (inlet, outlet and sludge) and freshwater and marine water bodies receiving wastewater discharges. A total of 66 samples were analysed within the project.

Table 2.2. Pharmaceuticals selected for screening in 2015, Denmark Pharmaceuticals selected for screening in 2015, Denmark
Pharmaceuticals in the screening	Application (therapeutic use)
Capecitabine, Tamoxifen	Anti-cancer
Carbamazepine, Citalopram, Codeine, Fen-tanyl, Propofol, Tramadol	Central nervous system agents
Miconazol	Antifungal
Amiloride, Atenolol, Bisoprolol, Losartan, Metoprolol, Propranolol	Cardiovascular agents
Amoxicillin, Azithromycin, Ciprofloxacin, Clarithromycin, Erythromycin, Fluconazol, Roxythromycin	Antibiotics
Diclofenac, Ibuprofen, Ketoprofen, Naproxen	NSAIDs
Norethisterone	Synthetic progestin

Based on the screening results, the following pharmaceuticals were included in the NOVANA programme: Azithromycin, Clarithromycin, Carbamazepin, Citalopram, Ibuprofen, Naproxen, Tramadol, Propranolol and Diclofenac. No legislation changes or determination of EQN for these substances has yet been carried out. The need for such actions will be further evaluated when the programme generates more data for these pharmaceuticals. The 2015 screening, funded by the Danish Environment Agency, cost € 95 000. This case study highlights the importance of having knowledge from drug databases about which pharmaceutical substances are being used, where they are used (outpatient or hospital), and in what quantities in order to rank and select pharmaceuticals for screenings.

Source: Summary of case study provided by Henrik Søren Larsen, Ministry of Environment and Food, Environmental Protection Agency, Denmark

Box 2.7. Identifying and prioritising indicator substances for CECs monitoring, Switzerland

Switzerland has prioritised five indicator substances to reduce analytical costs of monitoring for an extensive list of CECs. Out of a total of 250 substances (pharmaceuticals, pesticide and transformation products) identified in Swiss rivers, 47 indicator substances were identified through a selection process based on five criteria: i) partitioning of substances between water and solid phase; ii) persistence in the aquatic environment; iii) toxicity; iv) concentration patterns (continuous, periodic or intermittent); and v) probability of detecting a substance in surface waters.

To reduce the analytical costs for monitoring all 47 compounds, a subgroup of five indicator compounds was identified to be included in sampling programmes: carbamazepine (anticonvulsant or anti-epileptic drug), diclofenac (nonsteroidal anti-inflammatory drug), sulfamethoxazole (antibiotic), mecoprop (herbicide) and benzotriazole (anticorrosive agent). All of these substances can be measured with the same analytical method and are detectable in more than 90 % of all domestic WWTP effluents in Switzerland.

Source: (Götz, Kase and Hollender, 2011[39])

In order to overcome the almost infinite number of combination of pharmaceuticals (and their metabolites and transformation products) in the environment, novel methods and alternative holistic approaches are being developed. These emerging monitoring technologies may be overtaking the capacity of governments to react and put adequate responses in place. Some countries are revising monitoring programmes to include the effects of mixtures. The following section takes stock of recent advances in monitoring (including integrated approaches) and modelling that can provide an alternative to assessing APIs in the environment individually.

2.4. Advances in water quality monitoring and potential benefits for risk assessments and water quality policy making

The sheer number of potentially harmful pharmaceuticals and other emerging pollutants challenges traditional chemical monitoring efforts, and consequently there is a good chance that adverse impacts from unknown or unexpected chemicals and mixtures on aquatic communities and human health remain unrecognised. The problem is aggravated by analytical detection limits that may be too high for detecting chemicals at or below their predicted no-effect-concentrations (PNEC). Finally, a better understanding is required of how to link early biological responses to chemical exposure detectable in bioassays and biomarkers to ecological responses at the population and community level (Brack et al., 2015[40]).

A number of advances in water quality monitoring and modelling techniques can assist with some of these challenges, providing alternatives to the costly traditional spot (grab) sampling and chemical analysis approach. There is value in moving towards approaches which aim at identifying and anticipating new environmental pollutants as early as possible, i.e. before they become pervasive in the aquatic environment and before major health or economic consequences are felt (Daughton, 2004[38]). It is particularly important to discover impacts related to environmental contaminants before they occur on a population level, because damage at the population and ecosystem level can take a long time to repair.

New methods including non-target screening and suspected-target screening can be utilised to gain knowledge about the occurrence of so-called “known unknowns” and “unknown unknowns” of pollutants in water bodies2. Box 2.8 illustrates an example of non-target and suspected-target screening in Korea.

Box 2.8. Prioritisation of pharmaceuticals via suspect and non-target screening, Korea

The Yeongsan River is one of four major river basins in Korea. It is the most water scarce basin and has suffered from declining water quality from an increase in diffuse urban and agricultural pollution and toxic point source discharges (OECD, 2018[41]).

In a study by Park et al. (2018[42]), pharmaceuticals and personal care products (PPCPs) in the Yeongsan River, Korea were prioritised using suspect and non-target analysis by Liquid chromatography–high resolution mass spectrometry (LC-HSMS) (QExactive plus Orbitrap) followed by semi-quantitative analysis to confirm the priority of PPCPs.

The screening identified more than 50 PPCPs, of which 26 could be confirmed with reference standards. The confirmed substances were prioritised based on a scoring and ranking system. Twelve additional substances not included in the first ranking were semi-quantitatively analysed. In the final prioritisation list, carbamazepine, metformin and paraxanthine shared first-ranking place, followed by caffeine, cimetidine, lidocaine, naproxen, cetirizine, climbazole, fexofenadine, tramadol, and fluconazole. The authors suggest that these 12 PPCPs are the most highly exposable substances, and should be considered in future water monitoring of the Yeongsan River (Park et al., 2018[42]).

Sources: (Park et al., 2018[42]; OECD, 2018[41])

Effect-based monitoring can be a valuable approach to detect effects caused by contaminants at an earlier stage than a substance-by substance chemical monitoring approach (Wernersson et al., 2015[43]). They can also take into account the overall response from co-exposure to multiple, bioavailable pharmaceuticals and chemicals in the environment, including on different levels of biological organisation, such as community, population, individual and/or suborganism level (EC, 2014[44]). The European Chemical Monitoring and Emerging Pollutants sub-working group recommends the continuation of such monitoring in relation to: i) the detection and evaluation of effects caused by mixtures of pollutants, ii) as screening tools as part of ERAs to aid in the prioritisation of water bodies, iii) to establish early-warning systems, and iv) to provide additional support in water and sediment quality assessment, complementary to conventional chemical monitoring under the WFD (EC, 2014[44]) (Wernersson et al., 2015[43]). For more information on the state of the art of aquatic effect-based monitoring tools, refer to (EC, 2014[44]).

An overview of the various monitoring approaches, differentiated by chemical monitoring and effect-based monitoring, and their advantages and disadvantages are presented in Table 2.3.

Table 2.3. An overview of the different monitoring approaches for pharmaceuticals in water, and their advantages and disadvantages
Monitoring approach	Description	Advantages	Disadvantages	Indicative, relative cost
Chemical monitoring
Spot-sampling and target chemical Analysis	A discrete sample taken at one point in time and location, followed by laboratory analysis for a target chemical.	Appropriate for pharmaceuticals with known hazard.	Costly and labour-intensive. Episodic pollution events can be missed Fails to account for the bioavailability.	➚
Biosensors	An analytical device, used for the detection of a chemical substance that combines a biological component with a physicochemical detector.	Enables real-time in situ monitoring. Rapid, simple, low-cost techniques. Low number of samples required for quantification Valuable as an early warning signal.	Difficult to assess the impacts of exposure and the mixture effects of multiple contaminants on the ecosystem. Requires prior knowledge about the type of substance to be monitored for. Low selectivity (in most cases, exception being antibody-based sensors), low detection limits, risk of contamination with other microorganisms. Often requires skilled operators at the research stage, or more expensive user-friendly commercialised biosensors.	➘
Non-target and suspected screenings	Advanced techniques that often employ high resolution mass spectrometry and liquid chromatography to either match unknown sample features to compounds within spectral and/or spectra-less databases (suspect screening), or elucidate structures of unknowns that may not be contained in a database (non-target screening).	Measures a large amount of chemical features Useful to identify chemical structures of transformation products Does not require priori information on the compounds to be detected.	Non-target identification is a very time-intense process. Unidentified substances remain “unknowns”. Requires highly-trained personnel and expensive equipment. Can be difficult to distinguish between those signals worthwhile following up, and those that are just ‘noise’.	➚
Passive sampling	Technique involving the use of a collecting medium, such as a man-made device or biological organism, to accumulate chemical pollutants in the environment over time. Also known as diffuse sampling.	Can effectively concentrate pollutants compared to spot sampling. Gives information about source, sink and patterns of pollution at different locations. Provides time-weighted-average and equilibrium concentrations over the deployment time, rather than a snap shot at one moment. Accounts for bioavailable fractions. Low cost, non-mechanical, easy to deploy, little/no maintenance. Low number of samples required for quantification.	Does not identify unknowns. Requires careful pre-calibration. No standardised guidance. Performance can be affected by environmental variables, such as flow rate, temperature, pH, salinity and biofouling.	➘
Effect-based (ecological) monitoring
Ecological indicators	Biological assemblages or taxa that by their presence, condition or numbers indicate something about the state of the environment. Salmon and mayflies are well-known ecological indicators of the health of rivers.	Assesses the overall response from co-exposure to multiple, bioavailable chemicals (including transformation products), at different biological levels. Provides a highly integrated and relevant response.	Difficult to identify underlying causes and key events of the adverse effect of the ecosystem. Difficult to find reference conditions. Can be time-consuming and difficult to implement in a cost-effective manner without prior knowledge of what pharmaceuticals and effects (and in what biota) should be looked for.	➚
Bioassays (in vitro)	Test that measures the joint biological effect of all active chemicals in an environmental sample under defined laboratory conditions at the subcellular level, such as receptor activation and DNA damage. The OECD Test Guidelines Programme identifies new in vitro test methods that are candidates to become part of OECD Test Guidelines (OECD, 2018[45]).	Monitors the overall biological activity and the potential mechanism of toxicity. Highly sensitive. Low cost, appropriate as rapid screening tool, can be used to identify and track pollution sources and water bodies that need require further investigations. Provides benchmarking of mixture effects Can be used to identify toxic fractions and provide guidance for the identification of causative agents.	Limited to the targeted biological system and endpoints. No harmonised effect-based trigger values. Difficult to interpret results when they do not necessarily imply an adverse effect in exposed whole organisms (i.e. difficult to translate in vitro responses to in vivo effects).	➘
Bioassays (in vivo)	Test that measures the joint biological effect of all active chemicals in an environmental sample under defined laboratory conditions at the individual level. An example is the fish embryo acute toxicity test, adopted by the OECD as Test Guideline N.236, which is based on individual exposure of eggs to evaluate the embryotoxicity of samples with the aim to detect contaminants, including pharmaceuticals. Sometimes tests are performed in the field (in situ bioassays).	Provides broad spectrum analysis of the effects of a variety of substances (both known and unknown) and different types of toxicity to whole living organisms.	Costly and time-intensive. Often limited to the targeted biological endpoints.	➚
Biomarkers	Measure of biological alterations (e.g. biochemical, physiological, histological, or morphological changes) at cellular or individual levels, measured in organisms sampled in the field in a specific location (e.g. by collecting blood samples from mussels or fish). There are biomarkers of exposure (e.g. measured concentration of a target chemical in blood or the liver), which inform about the quality and/or quantity of exposure, and biomarkers of effects, which allow statements about effects and the health status of exposed organisms (e.g. genotoxicity, effects on the immune system or reproduction). The OECD Test Guidelines Programme identifies new biomarker endpoints (OECD, 2018[45]).	Measures biological alterations at cellular or individual levels in organisms at specific field locations. Identifies the impact from substances or mixtures of substances not previously identified to be of concern, and identify regions of decreased environmental quality. Provides benchmarking of mixture effects. Valuable as an early warning signal.	Difficult to establish linkage between alterations and adverse effects at community or population level. Interpreting the data can be challenging. Very few biomarkers can be linked directly to exposures to specific classes of chemicals.	➚
Omics	High-throughput molecular profiling technologies, such as genomics, metagenomics, transcriptomics, proteomics, metabolomics and metabonomics. Used to explore the roles, relationships, and actions of the various types of molecules that make up the cells of an organism.	Provides unique opportunities for the discovery of a new generation of biomarkers of exposure and disease risk. Can identify unexpected effects, monitor biological effects at the genome scale. Can be used to develop molecular biomarkers of exposure as early signals to predict effects. Provides information about the mechanism of toxicity and a basis for extrapolation of the effects across species, and effects on the whole ecosystem. Provides benchmarking of mixture effects.	Early stage of development. Lack of standardisation and guidance on best practices. Difficulties in interpreting data. Requires highly-trained personnel and expensive equipment.	➚
Notes: ‘➚’ denotes relatively costly; ‘➘’ denotes relatively inexpensive.
Sources: (Ejeian et al., 2018[46]) (Hernandez-Vargas et al., 2018[47]) (Chouler et al., 2015[48]) (Letzel, 2014[49]) (Novák et al., 2018[50]) (Brack et al., 2017[51]) (Anderson et al., 2012[52]) (EC, 2014[44]) (Wernersson et al., 2015[43]) (Escher et al., 2018[53]) (van der Oost et al., 2017[54]) (Könemann et al., 2018[55]) (Kase et al., 2018[56]) (Ekman et al., 2013[29]) (Leung, 2018[57]).

2.4.1. The value of integrating monitoring approaches

No single method or combination of methods is able to meet all divergent monitoring purposes (Altenburger et al., 2019[58]). However, employing various monitoring approaches (Table 2.3) together, utilising their various strengths, can provide a more holistic understanding of pharmaceuticals in the environment and their environmental effects (Figure 2.2) (Ekman et al., 2013[29]; Petrie, Barden and Kasprzyk-Hordern, 2015[59]) (Altenburger et al., 2019[58]).

The use of effect-based tools, passive sampling for bioaccumulative chemicals and an integrated strategy for prioritisation of contaminants, accounting for knowledge gaps, is advocated to improve and advance monitoring (Ekman et al., 2018[60]) (Brack et al., 2015[40]). Biomarkers can be effectively incorporated with other diagnostic markers of fish health, and also with analytical chemical monitoring approaches to provide evidence for the contributions of chemical exposures (Hook, Gallagher and Batley, 2014[61]). Integrating Adverse Outcome Pathways and omics would be a valuable practical tool to discover low-dose effects of substances, either individually or as a mixture (Leung, 2018[57]). In the OECD Conceptual Framework for Testing and Assessment of Endocrine Disruptors, in vitro assays are recommended to provide data on selected endocrine mechanisms and pathways before the substance eventually undergoes further investigations in vivo (OECD, 2018[62]; OECD, 2018[63]).

Figure 2.2. The components in integrated monitoring approaches and the key strengths and limitations of each approach

Source: (Ekman et al., 2013[29])

There are several integrated monitoring programmes in OECD countries. For example, as part of the U.S. Great Lakes Restoration Initiative, an integrated approach including chemical analysis, bioassays (in vitro, in vivo or in situ) and ecosystem monitoring is used to monitor pharmaceuticals and other emerging pollutants, which are identified as one of the highest priority stressors in the lakes (Ekman et al., 2013[29]). Also in the US, effect-based testing, combined with analytical chemistry data, is used to determine discharge permitting requirements (Box 2.9). In the Netherlands, passive sampling is used in combination with in situ, in vivo and in vitro bioassays to assess the impacts of wastewater discharge on water quality using the SIMONI (smart integrated monitoring) approach (van der Oost et al., 2017[64]). A Swiss study reveals the benefits of using chemical analysis, bioassays and mixture toxicity modelling to determine the impact of mixture effects and multiple sources (both point and diffuse) on instream water quality (Box 2.10).

Box 2.9. Whole Effluent Toxicity testing, US: A combination of testing methods to evaluate toxicity of wastewater

In the US, Whole Effluent Toxicity (WET) testing is used to determine the aggregate acute and short-term chronic toxicity of wastewater effluent on aquatic organisms. It is one way the EPA Clean Water Act's prohibition of the discharge of toxic pollutants in toxic amounts is implemented. WET tests measure the effects of wastewater on specific test organisms' ability to survive, grow and reproduce. The WET test methods consist of exposing living aquatic organisms (plants, vertebrates and invertebrates) to various concentrations of a sample of wastewater, usually from a facility's effluent stream. WET tests are used by the National Pollutant Discharge Elimination System permitting authority to determine whether a facility's permit will need to include WET requirements.

Two WET test manuals describe test procedures for effluents and receiving waters and include guidelines on test species selection and mobile toxicity test laboratory design. When toxicity is measured, the next step is to use the Toxicity Identification Evaluation (TIE) process to identify and reduce the toxicity. The TIE is a three-phase process that characterises, identifies, and confirms the cause or causes of toxicity. Once the identification/isolation process has confirmed the potential cause of toxicity, the next step is to determine what needs to be done to reduce or treat the chemical or chemicals causing toxicity in the effluent. In vitro assays are being used to screen water samples for the presence of compounds that activate specific cellular pathways (e.g. activating an estrogen-dependent pathway) or that target a particular cell type (e.g. bladder). Omics endpoints, measured in exposed organisms, represent the immediate and early response that underlie the development of adverse outcomes. Signatures from these new assays can be combined with traditional apical endpoints, and targeted and non-targeted analytical chemistry data, to strengthen the weight of evidence of biologically meaningful exposure and to identify causative agents.

Source: (US EPA, 2017[65])

Box 2.10. Integrated monitoring to evaluate the contribution of wastewater effluent on CECs burden in small streams, Switzerland

In a study by Neale et al. (2017[66]), chemical analysis and bioassays were combined to assess the CEC burden in small streams upstream and downstream of three WWTPs in the Swiss Plateau. The chemical analysis included 405 chemicals and in vitro bioassays assessed the following endpoints: activation of the aryl hydrocarbon receptor, androgen receptor and oestrogen receptor; photosystem II inhibition; acetylcholinesterase inhibition; adaptive stress responses for oxidative stress, genotoxicity and inflammation; and estrogenic activity and developmental toxicity in zebrafish embryos. Mixture toxicity modelling was applied to assess the contribution of detected chemicals to the observed effect.

The results from this study highlight the importance of combining bioassays with chemical analysis to provide the whole picture of the CEC burden to water bodies. The in vitro assay provided information about the mixture effects, while the chemical analysis showed differences in the chemical pollution profiles along different sampling stations. For most bioassays, very little of the observed effects could be explained by the detected chemicals. While higher concentrations and effects were observed in samples downstream of effluent discharge compared to upstream samples, both chemical analysis and bioanalysis showed that diffuse sources upstream during low flow conditions significantly affect the water quality and aquatic ecosystem. Consequently, upgrading WWTPs alone will not completely reduce the concentration and effects of CECs in the Swiss Plateau.

Source: (Neale et al., 2017[66]).

Passive sampling, together with effect-based approaches, are being considered as potentially suitable tools that could be employed for monitoring of European water bodies in the implementation strategy of the EU Water Framework Directive (European Commission, 2015). While both approaches are often employed independently, their use in combination has been demonstrated previously in studies focusing on WWTP effluents and affected rivers (Creusot et al., 2013; Jalova et al., 2013; Jarosova et al., 2012).

While the ideal environmental monitoring system would routinely include all three components (chemical analysis, bioassays, ecosystem/effect-based monitoring), resource limitations and other practical constraints generally dictate where they are employed on a case-by case basis (Ekman et al., 2013[29]). For example, some sites have limited prior knowledge and information about the pollution burden and possible effects, other sites have low ecological health status, and others have known effects and exposure where the major sources needs to be identified. Ekman et al. (2013[29]) suggest a stepwise process to design and implement a strategic and integrated monitoring approach. The first step is to start with a problem formulation considering the existing information about the site, management goals and particular regulatory motivators. Once these basics are understood, strategic decisions about which specific monitoring tools should be employed are possible, which then allows for informed decision-making on what actions may (or may not) be required.

One of the challenges using alternative tools and data types is the interpretation of endpoints in the context of biological effects. Biological activity (e.g. measured in vitro or in biomarkers) does not necessarily constitute a hazard. The Adverse Outcome Pathways (AOP) concept (OECD, 2018[45]) was developed to address this issue and to make it possible to relate endpoints from in vitro bioassays and biomarkers to endpoints useful for risk assessment (e.g. of survival, reproduction, development, and growth) (Ankley et al., 2010[67]) (Ankley and Edwards, 2018[68]). It is a conceptual chain that links the exposure of contaminants to their cellular concentrations and molecular initiating events, via pathway disturbance and key events, to response at the cellular, organism and population or community levels. The framework can be used for different purposes, including to predict the effects of mixtures (Carusi et al., 2018[69]). AOPs increase the efficiency of chemical safety assessments, reduce the need for animal testing, and has received significant attention and use in the regulatory toxicology community (Carusi et al., 2018[69]). Recently, the AOP has been moving from a linear concept to a pathway network considering the idea of multiple causes for adverse effects (Escher et al., 2017[70]) (Knapen et al., 2015[71]).

Another alternative method to predict the toxicity of pharmaceuticals, is to group APIs that are structurally similar and may therefore cause similar adverse environmental effects. The OECD has developed the Integrated Approaches to Testing and Assessment (IATA, which can include AOPs), providing a framework and tools for data gathering to maximise the amount of information about risks of chemicals. For example, the OECD Quantitative Structure-Activity Relationships (QSAR) Toolbox (OECD, 2018[72]) is a software application intended to be used by governments, chemical industry and other stakeholders to fill gaps in (eco)toxicity data needed for assessing the hazards of chemicals. The Toolbox incorporates information and tools from various sources into a logical workflow. It enables the identification of new methods/profilers for grouping chemicals through in vitro bioassays.

2.5. The added value of system modelling

Because of the lack of systematic monitoring programmes for pharmaceuticals, modelling serves as a valuable and cost-effective basis for prioritisation, risk assessments, and to address data and knowledge gaps. Modelling the source-to-effect chain can be an effective tool to identify and target sources of pollution for monitoring. They can also be useful for predicting the exposure and impacts of pollution sources, pharmaceutical types and mixture toxicity, and the effectiveness of technical and policy solutions. Modelling can be a constructive starting point to understand and discuss the source and effects of pharmaceuticals with stakeholders, from which cost-effective solutions can be developed in cooperation. Ultimately, it would be of value to move towards developing modelling approaches of real world exposure where organisms continually face mixtures of multitudes of stressors, which vary over time in composition and concentrations (Daughton, 2004[38]).

Computational tools, like QSAR (Quantitative Structure Activity Relationship) models, can be used to screen large sets of chemicals in a short time, with the aim of ranking, highlighting and prioritising the most environmentally hazardous for focusing further experimental studies (Sangion and Gramatica, 2016[73]). The QSAR approach has been used to model the toxicity of pharmaceuticals, both regarding mixtures and in the assessment of unknown substances such as transformation products (Escher et al., 2006[74]) (Rastogi, Leder and Kümmerer, 2014[75]) (Mahmoud et al., 2014[76]). The OECD has recognised the potential for QSARs to reduce the costs of testing, reduce the need for animal testing and to strengthen chemical regulation (Fjodorova et al., 2008[77]).

Modelling sources of pharmaceutical pollution can be used in order to investigate to what degree individual sources (such as WWTPs) impact water quality. For example, in the Netherlands, nation-wide consumption-based hydrological modelling has given spatial insight on the impact of WWTP discharge on concentrations of pharmaceuticals in surface water bodies (Box 2.11). In Sweden, modelling has determined the exposure potential of some of the nation’s top-used pharmaceuticals to the Baltic Sea (Box 2.12). Oldenkamp et al. (2018[78]) have developed a spatially explicit model (ePiE) which calculates concentrations of APIs in surface waters across Europe at a resolution of approximately 1 km. Such models can be helpful in prioritising APIs and spatially assessing the risks.

Box 2.11. Defining impact of wastewater treatment plants on susceptible functions, Netherlands

A nation-wide modelling and ranking exercise was undertaken in the Netherlands to investigate and prioritise which of their 345 WWTPs should be upgraded to reduce the impact of pharmaceuticals on receiving water bodies (in particular to EU nature protection areas) and the risk to raw drinking water sources. The model was based on two components: i) a water quality model representing the Dutch surface water network and its key hydrological features; and ii) a consumption-based emission model to project the loads from WWTPs to receiving rivers during both low and high discharge conditions. Two pharmaceuticals with different characteristics (carbamazepine and ibuprofen) underwent a detailed spatial analysis.

The vast majority of the total impact of all Dutch WWTPs, during both high and low discharge conditions, was attributed to 19% of the WWTPs with regard to the drinking water function, and to 39% of the WWTPs with regard to the nature protected areas function. The model thus provides a spatially smart and cost-effective way to identify and prioritise WWTP upgrades to improve water quality and reduce adverse environment effects.

Source: (Coppens et al., 2015[79]).

Box 2.12. Modelling to predict occurrence of multiple pharmaceuticals in Swedish surface waters and their release to the Baltic Sea

A Swedish study provides assessments of the emissions and occurrence of 54 pharmaceuticals in surface water and their releases to the Baltic Sea. Lindim et al. (2017[80]) used the STREAM-EU model to predict the exposure, fate and long-range transport of 54 human pharmaceuticals (selected from the 200 most consumed pharmaceuticals in Sweden in 2011) in all Swedish basins draining to the Baltic Sea and Danish Strait. The model considered point sources (urban WWTPs) and diffuse sources (through WWTP sludge application to soils) as human pharmaceutical entry-pathways to surface water. The model was parametrised with the knowledge that 25.4% of the sludge generated from WWTPs in Sweden is used on agriculture soils, according to national statistics. The model was constructed by using data on emissions, hydrology, pH, air and water temperatures, and physico-chemical properties and computationally-estimated partitioning and degradation properties of the substances.

The results showed a total flushing flow rate to the Baltic Sea of 42 ton/yr for the 54 pharmaceuticals combined. Of the 54 studied pharmaceuticals, 35 (65%) had predicted annual flushes to the sea higher that 1 kg/yr; the highest of which were for metformin (27 ton/yr), paracetamol (6.9 ton/yr) and ibuprofen (2.33 ton/yr). In the Stockholm urban area, 25 drugs had predicted concentrations higher than 1 ng/L, with 17 above 10 ng/L. Modelled results were in good agreement with monitored/measured values, with agreements of r² = 0.62 and r² = 0.95 for concentrations and for disposed amounts to sea, respectively. Persistence and hydrophobicity of the pharmaceuticals determined the long-range transport and exposure potential. The model indicated that piperacillin, lorazepam, metformin, hydroxycarbamide, hydrochlorothiazide, furosemide and cetirizine have high potential to reach the sea.

Source: (Lindim et al., 2016[81]) (Lindim et al., 2017[80])

Modelling of pharmaceuticals at the global scale has also been attempted. Oldenkamp et al. (2019[82]) modelled the aquatic risks (expressed as the PEC/PNEC ratios) of antibiotics carbamazepine and ciprofloxacin in 449 aquatic ecoregions worldwide over the period 2005 to 2015. The study combined spatially explicit chemical fate and effect modelling with predictions of pharmaceutical consumption. The results showed exposure of antibiotics to freshwater ecosystems increased 10-20 fold over the last 20 years. Aquatic risks due to carbamazepine exposure were typically low, although more densely populated (e.g. western and central Europe) and/or arid ecoregions (e.g. Arabian and Californian peninsulas) showed higher risk. Risks for ciprofloxacin were found to be much higher and more widespread (Oldenkamp, Beusen and Huijbregts, 2019[82]).

Despite the advantages of modelling, it is pertinent to note that models only provide an estimation of reality and therefore there are always uncertainties. Source and environmental fate models require detailed information regarding pharmaceutical inputs on emission rates, partitioning properties, as well as degradation rates to accurately predict environmental partitioning and transport (Lindim et al., 2017[80]). Measured partitioning and degradation rates are still scarce for the majority of pharmaceuticals. Furthermore, many pharmaceuticals are polar and mobile, meaning that they are difficult to model as they behave differently from, for example, PBT-substances (Reemtsma et al., 2016[83]).

Finally, it is important to note that improving knowledge is not a pollution reduction measure in itself. Governments should take advantage of alternative innovative monitoring and modelling technologies to minimise costs, and prioritise substances and water bodies of highest concern, but they should not wait for exact science before taking action. Since pharmaceuticals are ubiquitous, the goal is to strive towards a non-toxic environment - not a non-chemical environment. Chapter 3 looks at opportunities to design and implement innovative and effective policies to reduce pharmaceuticals in the environment.

References

[14] Ågerstrand, M. et al. (2015), “Improving environmental risk assessment of human pharmaceuticals”, Environmental Science and Technology, Vol. 49/9, pp. 5336-5345, https://doi.org/10.1021/acs.est.5b00302.

[26] Ågerstrand, M. et al. (2017), “An academic researcher’s guide to increased impact on regulatory assessment of chemicals”, Environmental Science: Processes and Impacts, Vol. 19/5, pp. 644-655, https://doi.org/10.1039/c7em00075h.

[58] Altenburger, R. et al. (2019), “Future water quality monitoring: improving the balance between exposure and toxicity assessments of real-world pollutant mixtures”, Environmental Sciences Europe, Vol. 31/1, p. 12, https://doi.org/10.1186/s12302-019-0193-1.

[52] Anderson, P. et al. (2012), Monitoring Strategies for Chemicals of Emerging Concern (CECs) in California’s Aquatic Ecosystems Recommendations of a Science Advisory Panel.

[67] Ankley, G. et al. (2010), “Adverse outcome pathways: A conceptual framework to support ecotoxicology research and risk assessment”, Environmental Toxicology and Chemistry, Vol. 29/3, pp. 730-741, https://doi.org/10.1002/etc.34.

[68] Ankley, G. and S. Edwards (2018), “The adverse outcome pathway: A multifaceted framework supporting 21st century toxicology”, Current Opinion in Toxicology, Vol. 9, pp. 1-7, https://doi.org/10.1016/J.COTOX.2018.03.004.

[30] aus der Beek, T. et al. (2016), “Pharmaceuticals in the environment-Global occurrences and perspectives”, Environmental Toxicology and Chemistry, Vol. 35/4, pp. 823-835, https://doi.org/10.1002/etc.3339.

[37] Australian Government (2008), Australian guidelines for water recycling: managing health and environmental risks (phase 2). Augmentation of drinking water supplies., National Health and Medical Research Council, Canberra.

[23] Backhaus, T. (2014), “Medicines, shaken and stirred: A critical review on the ecotoxicology of pharmaceutical mixtures”, Philosophical Transactions of the Royal Society B: Biological Sciences, Vol. 369/1656, https://doi.org/10.1098/rstb.2013.0585.

[24] BIO Intelligence Service (2013), Study on the environmental risks of medicinal products.

[40] 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.

[51] Brack, W. et al. (2017), “Towards the review of the European Union Water Framework management of chemical contamination in European surface water resources”, Science of the Total Environment, Vol. 576, pp. 720-737, https://doi.org/10.1016/j.scitotenv.2016.10.104.

[3] Caneva, L. et al. (2014), “Critical review on the Environmental Risk Assessment of medicinal products for human use in the centralised procedure”, Regulatory Toxicology and Pharmacology, Vol. 68/3, pp. 312-316, https://doi.org/10.1016/J.YRTPH.2014.01.002.

[69] Carusi, A. et al. (2018), “Harvesting the promise of AOPs: An assessment and recommendations”, Science of the Total Environment, Vol. 628-629, pp. 1542-1556, https://doi.org/10.1016/j.scitotenv.2018.02.015.

[10] Chapman, J. et al. (2017), “Three methods for integration of environmental risk into the benefit-risk assessment of veterinary medicinal products”, Science of the Total Environment, Vol. 605-606, pp. 692-701, https://doi.org/10.1016/j.scitotenv.2017.06.205.

[48] Chouler, J. et al. (2015), “Water Quality Monitoring in Developing Countries; Can Microbial Fuel Cells be the Answer?”, Biosensors, Vol. 5/3, pp. 450-470, https://doi.org/10.3390/bios5030450.

[20] Cleuvers, M. (2008), “Chronic Mixture Toxicity of Pharmaceuticals to Daphnia – The Example of Nonsteroidal Anti-Inflammatory Drugs”, in Pharmaceuticals in the Environment, Springer Berlin Heidelberg, Berlin, Heidelberg, https://doi.org/10.1007/978-3-540-74664-5_17.

[79] Coppens, L. et al. (2015), “Towards spatially smart abatement of human pharmaceuticals in surface waters: Defining impact of sewage treatment plants on susceptible functions”, Water Research, Vol. 81, pp. 356-365, https://doi.org/10.1016/j.watres.2015.05.061.

[38] Daughton, C. (2004), “Non-regulated water contaminants: Emerging research”, Environmental Impact Assessment Review, Vol. 24/7-8, pp. 711-732, https://doi.org/10.1016/j.eiar.2004.06.003.

[32] EC (2018), Updated surface water Watch List adopted by the Commission, https://ec.europa.eu/jrc/en/science-update/updated-surface-water-watch-list-adopted-commission (accessed on 25 January 2019).

[44] EC (2014), Technical report on aquatic effect-based monitoring tools, European Commission, https://doi.org/10.2779/7260.

[46] Ejeian, F. et al. (2018), “Biosensors for wastewater monitoring: A review”, Biosensors and Bioelectronics, Vol. 118, pp. 66-79, https://doi.org/10.1016/J.BIOS.2018.07.019.

[29] Ekman, D. et al. (2013), “Biological Effects–Based Tools for Monitoring Impacted Surface Waters in the Great Lakes: A Multiagency Program in Support of the Great Lakes Restoration Initiative”, Environmental Practice, Vol. 15/4, pp. 409-426, https://doi.org/10.1017/S1466046613000458.

[60] Ekman, D. et al. (2018), “Evaluation of targeted and untargeted effects-based monitoring tools to assess impacts of contaminants of emerging concern on fish in the South Platte River, CO”, Environmental Pollution, Vol. 239, pp. 706-713, https://doi.org/10.1016/J.ENVPOL.2018.04.054.

[12] EMA (2017), Reflection paper on the authorisation of veterinary medicinal products containing (potential) persistent, bioaccumulative and toxic (PBT) or very persistent and very bioaccumulative (vPvB) substances, European Medicines Agency , London, http://www.ema.europa.eu/docs/en_GB/document_library/Scientific_guideline/2017/05/WC500228196.pdf.

[9] EMA (2009), Recommendation on the Evaluation of the Benefit-Risk Balance of Veterinary Medicinal Products. EMEA/CVMP/248499/2007, European Medicines Agency Veterinary Medicines and Inspections, London, http://www.ema.europa.eu/docs/en_GB/document_library/Other/2009/10/WC500005264.pdf.

[1] EMA (2006), Guideline on the environmental risk assessment of medicinal products for human use. Doc. Ref. EMEA/CHMP/SWP/4447/00 corr 2, European Medicines Agency Pre-Authorisation Evaluation of Medicines for Human Use, London.

[8] EMA (2004), Guideline on Environmental Impact Assessment for Veterinary Medicinal Products Phase II. CVMP/VICH/790/03¬FINAL, European Medicines Agency Veterinary Medicines and Inspections, London, http://www.ema.europa.eu/docs/en_GB/document_library/Scientific_guideline/2009/10/WC500004393.pdf.

[53] 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, pp. 748-765, https://doi.org/10.1016/J.SCITOTENV.2018.01.340.

[74] Escher, B. et al. (2006), “Comparative ecotoxicological hazard assessment of beta-blockers and their human metabolites using a mode-of-action-based test battery and a QSAR approach”, Environmental Science and Technology, Vol. 40/23, pp. 7402-7408, https://doi.org/10.1021/es052572v.

[70] Escher, B. et al. (2017), “From the exposome to mechanistic understanding of chemical-induced adverse effects”, Environment International, Vol. 99, pp. 97-106, https://doi.org/10.1016/j.envint.2016.11.029.

[34] European Commission (2017), Strategies against chemical pollution of surface waters, http://ec.europa.eu/environment/water/water-dangersub/ (accessed on 11 September 2018).

[4] FDA (2016), Environmental Assessment: Questions and Answers Regarding Drugs With Estrogenic, Androgenic, or Thyroid Activity Guidance for Industry, U.S. Department of Health and Human Services Food and Drug Administration Center for Drug Evaluation and Research (CDER), Silver Spring, MD, http://www.fda.gov/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/default.htm.

[5] FDA (1998), Guidance for Industry Environmental Assessment of Human Drug and Biologics Applications, Food and Drug Administration, https://www.fda.gov/downloads/Drugs/Guidances/ucm070561.pdf.

[77] Fjodorova, N. et al. (2008), “Directions in QSAR modeling for regulatory uses in OECD member countries, EU and in Russia”, Journal of Environmental Science and Health - Part C Environmental Carcinogenesis and Ecotoxicology Reviews, Vol. 26/2, pp. 201-236, https://doi.org/10.1080/10590500802135578.

[22] Godoy, A. and F. Kummrow (2017), “What do we know about the ecotoxicology of pharmaceutical and personal care product mixtures? A critical review”, Critical Reviews in Environmental Science and Technology, Vol. 47/16, pp. 1453-1496, https://doi.org/10.1080/10643389.2017.1370991.

[39] Götz, C., R. Kase and J. Hollender (2011), Micropollutants - Assessment concept for organic trace substances from municipal sewage [In German], Swiss Federal Institute of Aquatic Science and Technology (Eawag), Dübendorf.

[2] Gunnarsson, L. et al. (2019), “Pharmacology beyond the patient – The environmental risks of human drugs”, Environment International, Vol. 129, pp. 320-332, https://doi.org/10.1016/J.ENVINT.2019.04.075.

[47] Hernandez-Vargas, G. et al. (2018), “Electrochemical biosensors: A solution to pollution detection with reference to environmental contaminants”, Biosensors, Vol. 8/2, https://doi.org/10.3390/bios8020029.

[61] Hook, S., E. Gallagher and G. Batley (2014), “The role of biomarkers in the assessment of aquatic ecosystem health.”, Integrated environmental assessment and management, Vol. 10/3, pp. 327-41, https://doi.org/10.1002/ieam.1530.

[56] Kase, R. et al. (2018), “Screening and risk management solutions for steroidal estrogens in surface and wastewater”, TrAC Trends in Analytical Chemistry, Vol. 102, pp. 343-358, https://doi.org/10.1016/J.TRAC.2018.02.013.

[71] Knapen, D. et al. (2015), “The potential of AOP networks for reproductive and developmental toxicity assay development”, Reproductive Toxicology, Vol. 56, pp. 52-55, https://doi.org/10.1016/j.reprotox.2015.04.003.

[55] 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.

[15] Kools, S. et al. (2008), “A ranking of European veterinary medicines based on environmental risks”, Integrated Environmental Assessment and Management, Vol. 4/4, pp. 399-408, https://doi.org/10.1897/IEAM_2008-002.1.

[21] Kortenkamp Andreas, T. and M. Faust (2009), State of the Art Report on Mixture Toxicity, Directorate General for the Environment: Report to the EU Commission, http://ec.europa.eu/environment/chemicals/effects/pdf/report_mixture_toxicity.pdf.

[16] Kümmerer, K. et al. (2016), “Antineoplastic compounds in the environment—substances of special concern”, Environmental Science and Pollution Research, Vol. 23/15, pp. 14791-14804, https://doi.org/10.1007/s11356-014-3902-8.

[13] Küster, A. and N. Adler (2014), “Pharmaceuticals in the environment: Scientific evidence of risks and its regulation”, Philosophical Transactions of the Royal Society B: Biological Sciences, Vol. 369/1656, https://doi.org/10.1098/rstb.2013.0587.

[17] Le Page, G. et al. (2017), “Integrating human and environmental health in antibiotic risk assessment: A critical analysis of protection goals, species sensitivity and antimicrobial resistance”, Environment International, Vol. 109, pp. 155-169, https://doi.org/10.1016/J.ENVINT.2017.09.013.

[18] Le Page, G. et al. (2019), “Variability in cyanobacteria sensitivity to antibiotics and implications for environmental risk assessment”, Science of The Total Environment, Vol. 695, p. 133804, https://doi.org/10.1016/J.SCITOTENV.2019.133804.

[49] Letzel, T. (2014), Non-target screening, suspected-target screening and target screening – of technologies and philosophies, databases and crafts, http://www.int.laborundmore.com/archive/555305/Non-target-screening,-suspected-target-screening-and-target-screening-%E2%80%93-of-technologies-and-philosophies,-databases-and-crafts.html (accessed on 11 September 2018).

[57] Leung, K. (2018), “Joining the dots between omics and environmental management”, Integrated Environmental Assessment and Management, Vol. 14/2, pp. 169-173, https://doi.org/10.1002/ieam.2007.

[80] Lindim, C. et al. (2017), “Model-predicted occurrence of multiple pharmaceuticals in Swedish surface waters and their flushing to the Baltic Sea”, Environmental Pollution, Vol. 223, pp. 595-604, https://doi.org/10.1016/J.ENVPOL.2017.01.062.

[81] Lindim, C. et al. (2016), “Evaluation of human pharmaceutical emissions and concentrations in Swedish river basins”, Science of the Total Environment, Vol. 572, pp. 508-519, https://doi.org/10.1016/j.scitotenv.2016.08.074.

[84] Little, J., C. Cleven and S. Brown (2011), “Identification of “Known Unknowns” Utilizing Accurate Mass Data and Chemical Abstracts Service Databases”, Journal of The American Society for Mass Spectrometry, Vol. 22/2, pp. 348-359, https://doi.org/10.1007/s13361-010-0034-3.

[76] Mahmoud, W. et al. (2014), “Identification of phototransformation products of thalidomide and mixture toxicity assessment: An experimental and quantitative structural activity relationships (QSAR) approach”, Water Research, Vol. 49/1, pp. 11-22, https://doi.org/10.1016/j.watres.2013.11.014.

[33] Marsland, T. and S. Roy (2016), “Groundwater Watch List: Pharmaceuticals Pilot Study. Monitoring Data Collection and Initial Analysis. For European Commission – and CIS Groundwater Working Group (WGGW)”, Amec Foster Wheeler Associate, Shrewsbury, http://www.vlakwa.be/fileadmin/user_upload/20160629-File3-1600204_Pharm_Pilot_Study_GRW.doc.

[35] MPA (2015), Miljöindikatorer inom ramen för nationella läkemedelsstrategin (NLS). [Environmental indicators under the national drugs strategy (NLS)] Rapport från CBL-kansliet, Läkemedelsverket, Swedish Medicinal Product Agency, Uppsala, https://lakemedelsverket.se/upload/nyheter/2015/miljoindikatorer-rapport-NLS_2015-09-07.pdf.

[66] Neale, P. et al. (2017), “Integrating chemical analysis and bioanalysis to evaluate the contribution of wastewater effluent on the micropollutant burden in small streams”, Science of the Total Environment, Vol. 576, pp. 785-795, https://doi.org/10.1016/j.scitotenv.2016.10.141.

[50] Novák, J. et al. (2018), “Effect-based monitoring of the Danube River using mobile passive sampling”, Science of The Total Environment, Vol. 636, pp. 1608-1619, https://doi.org/10.1016/J.SCITOTENV.2018.02.201.

[25] OECD (2019), Saving Costs in Chemicals Management: How the OECD Ensures Benefits to Society, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264311718-en.

[45] OECD (2018), Adverse Outcome Pathways, Molecular Screening and Toxicogenomics, http://www.oecd.org/chemicalsafety/testing/adverse-outcome-pathways-molecular-screening-and-toxicogenomics.htm (accessed on 27 January 2019).

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

[41] OECD (2018), Managing the Water-Energy-Land-Food Nexus in Korea: Policies and Governance Options, OECD Studies on Water, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264306523-en.

[63] OECD (2018), OECD Work Related to Endocrine Disrupters, http://www.oecd.org/env/ehs/testing/oecdworkrelatedtoendocrinedisrupters.htm (accessed on 27 January 2019).

[62] 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://dx.doi.org/10.1787/9789264304741-en.

[72] OECD (2018), The OECD QSAR Toolbox, http://www.oecd.org/chemicalsafety/risk-assessment/oecd-qsar-toolbox.htm (accessed on 27 January 2019).

[82] Oldenkamp, R., A. Beusen and M. Huijbregts (2019), “Aquatic risks from human pharmaceuticals—modelling temporal trends of carbamazepine and ciprofloxacin at the global scale”, Environmental Research Letters, Vol. 14/3, p. 034003, https://doi.org/10.1088/1748-9326/ab0071.

[78] Oldenkamp, R. et al. (2018), “A High-Resolution Spatial Model to Predict Exposure to Pharmaceuticals in European Surface Waters: ePiE”, Environmental science & technology, Vol. 52/21, pp. 12494-12503, https://doi.org/10.1021/acs.est.8b03862.

[42] Park, N. et al. (2018), “Prioritization of highly exposable pharmaceuticals via a suspect/non-target screening approach: A case study for Yeongsan River, Korea”, Science of the Total Environment, Vol. 639, pp. 570-579, https://doi.org/10.1016/j.scitotenv.2018.05.081.

[59] Petrie, B., R. Barden and B. Kasprzyk-Hordern (2015), “A review on emerging contaminants in wastewaters and the environment: Current knowledge, understudied areas and recommendations for future monitoring”, Water Research, Vol. 72, pp. 3-27, https://doi.org/10.1016/J.WATRES.2014.08.053.

[75] Rastogi, T., C. Leder and K. Kümmerer (2014), “Qualitative environmental risk assessment of photolytic transformation products of iodinated X-ray contrast agent diatrizoic acid”, Science of the Total Environment, Vol. 482-483/1, pp. 378-388, https://doi.org/10.1016/j.scitotenv.2014.02.139.

[83] Reemtsma, T. et al. (2016), “Mind the Gap: Persistent and Mobile Organic Compounds - Water Contaminants That Slip Through”, Environmental Science and Technology, Vol. 50/19, pp. 10308-10315, https://doi.org/10.1021/acs.est.6b03338.

[36] RIVM (2014), Environmental risk limits for pharmaceuticals Derivation of WFD water quality standards for carbamazepine, metoprolol, metformin and amidotrizoic acid, National Institute for Public Health and the Environment , Bilthoven.

[73] Sangion, A. and P. Gramatica (2016), “PBT assessment and prioritization of contaminants of emerging concern: Pharmaceuticals”, Environmental Research, Vol. 147, pp. 297-306, https://doi.org/10.1016/j.envres.2016.02.021.

[11] UBA (2018), Recommendations for reducing micropollutants in waters. Background- April 2018, German Environment Agency, Dessau-Roßlau.

[65] US EPA (2017), Whole Effluent Toxicity Methods, Clean Water Act Analytical Methods, United States Environmental Protection Agency, https://www.epa.gov/cwa-methods/whole-effluent-toxicity-methods (accessed on 25 June 2019).

[31] US EPA (2016), Fourth Unregulated Contaminant Monitoring Rule, Monitoring unregulated drinking water contaminants, https://www.epa.gov/dwucmr/fourth-unregulated-contaminant-monitoring-rule (accessed on 25 June 2019).

[64] 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, pp. 2400-2416, https://doi.org/10.1002/etc.3837.

[54] 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, pp. 2385-2399, https://doi.org/10.1002/etc.3836.

[6] VICH (2018), What is the role of VICH, https://vichsec.org/what-is-the-role-of-vich.html (accessed on 11 September 2018).

[7] VICH (2000), Environmental Impact Assessment (EIAs) for veterinary medicinal products (VMPs)- Phase I. VICH GL6, The International Cooperation on Harmonisation of Technical Requirements for Registration of Veterinary Medicinal Products , Bruxelles.

[43] Wernersson, A. et al. (2015), “The European technical report on aquatic effect-based monitoring tools under the water framework directive”, Environmental Sciences Europe, Vol. 27/1, pp. 1-11, https://doi.org/10.1186/s12302-015-0039-4.

[19] WHO (2017), Chemical mixtures in source water and drinking-water, World Health Organisation, Geneva.

[28] WHO (2012), Pharmaceuticals in Drinking Water, World Health Organisation, Geneva.

Notes

← 1. Communication from the Commission to the council: http://ec.europa.eu/transparency/regdoc/rep/1/2012/EN/1-2012-252-EN-F1-1.Pdf.

← 2. Unknowns” refer to a class of substances that cannot be categorised into known molecules or identified by standard evaluation methods (Little, Cleven and Brown, 2011[84]).

2. Opportunities to build a policy-relevant knowledge base