Hindsight check: LL2-13 (Ch 13 Ethinyl oestradiol in the aquatic environment)#

Source section: EEA, Late lessons from early warnings: science, precaution, innovation (EEA Report No 1/2013), Ch 13, by Susan Jobling and Richard Owen (pp. 279–307; text pp. 279–297, Table 13.1 p. 298). There are no panels. Check window: publication (2013) to late September 2026. Checked: 25 September 2026.

Method note. - General web search was unavailable for this pass because the session’s search budget was exhausted. I retrieved sources by fetching primary repositories directly: - the EU Publications Office (the Cellar document store and its SPARQL endpoint), for the full texts of EU directives, Commission decisions, Commission communications and impact assessments, and Court of Justice notices and orders; - the European Parliament Legislative Observatory (OEIL), for procedure timelines; - the Joint Research Centre (JRC) publications repository (abstract pages only; the PDFs load by script and could not be fetched); - Europe PMC, for peer-reviewed abstracts and one open-access full text; - GOV.UK (its search API, Environment Agency reports and attachments); - the UK Water Industry Research (UKWIR) website; - the NHS Business Services Authority (NHSBSA) open-data portal, queried directly for prescription items; - the Swiss Water Association (VSA) micropollutant platform (micropoll.ch); - the European Medicines Agency (EMA) public assessment reports. - Some sources refused automated retrieval: - EUR-Lex (HTTP 202, empty body). I read the same legal texts through the Publications Office Cellar, which serves the Official Journal versions. I cite the EUR-Lex/ELI address for readers. - The OECD’s 2019 report on pharmaceutical residues (403) and NHS Digital’s 2003–2013 prescription statistics (403). I therefore have no official 2013 baseline for English prescribing and use NHSBSA open data from 2014 onward instead. - The JRC’s 2018 review of the first watch list. I rely on its abstract and on a later peer-reviewed JRC summary (Loos et al. 2024). Where a point rests on an abstract only, I say so. - Annex 3 does not apply. EE2 is a new case in the 2013 volume, so there is no Annex 3 update. Annex 3 covers only the 2001 cases. - I read only the section digest, the section’s source extract and web sources. Page numbers are report pages of the 2013 volume. - Conflict of roles. Jobling co-authored several of the later studies I use: the Environment Agency’s 2017 roach resurvey (published 2020) [S24] and the population-genetics papers [S25, S26]. Two of these (Hamilton et al. 2014 and 2020) cut against the chapter’s population-level fears. That is a point in favour of the group’s even-handedness, but these are not independent checks. I flag independent evidence where it exists: the Canadian whole-lake experiments [S27, S28], the EU regulatory record [S1–S9] and the UK water industry’s own monitoring [S19–S22].

Overview#

The chapter makes four kinds of claim. Thirteen years on, they have fared differently.

1. The regulatory story: delay was worse than the chapter feared, but regulation arrived and is stricter than proposed. - In 2013 the EU did not list EE2 and E2 as priority substances. Directive 2013/39/EU instead put them, with diclofenac, on a new “watch list” for monitoring [S1]. They were monitored from 2015 until the four-year limit expired in 2019 [S2–S4]. - The Commission’s proposal to list them came in October 2022, six years after the 2016 review the July 2012 amendment had envisaged [S6, S7]. - Directive (EU) 2026/805 (30 March 2026) finally lists EE2, E2 and oestrone (E1) as priority substances [S8]: - The EE2 annual-average standard is 0.017 ng/L in inland waters, half the 0.035 ng/L proposed in 2012, and 0.0016 ng/L in other surface waters. - It takes effect from 22 December 2027. The target for good chemical status is 22 December 2039, twelve years after the 2027 date in the “delaying” 2012 amendment and eighteen years after the 2021 date in the original proposal (pp. 293–294). - In parallel, the 2024 recast of the Urban Wastewater Treatment Directive requires “quaternary” treatment against micropollutants at large plants and in at-risk areas by 2045. At least 80% of its cost falls on producers of human medicines and cosmetics through extended producer responsibility (EPR) [S9]. - The chapter never discussed polluter-pays (see digest caveats). Cost allocation therefore moved in a direction it did not anticipate, and cost is now fought over in court rather than only in cost–benefit analysis [S12].

2. The technical claims have largely held. - Conventional sewage treatment removes EE2 poorly, and effluents exceed its no-effect level. The UK water industry’s own £190m monitoring programme (2010–2022) confirmed this [S19–S21]. - Measuring EE2 at the proposed standard remains extremely hard. Watch-list monitoring showed some member states could not reach the required detection limits [S14]. The 2026 directive acknowledges that methods “are not always sensitive enough” [S8, recital 29]. - EU rules require a limit of quantification at or below 30% of the standard [S13]. For EE2 that means about 5 pg/L in rivers and about 0.5 pg/L in other waters. - The EU’s answer is mandatory effect-based (bioassay) monitoring of total oestrogenic activity in 2030–2031, a partial vindication of the chapter’s point about mixtures [S8, Art. 8a(3)–(5)].

3. The projections of cost and implementation were too optimistic in the early-mover case, and untested in the UK. - Switzerland did build full-scale micropollutant treatment, but more slowly and at higher cost than footnote 10 projects [S16, S17]: - the first ozonation plant opened in 2014, not spring 2013; - 39 plants were running by December 2025, against about 120 targeted by 2040; - the target covers about 70% of domestic wastewater, not 80%; - realised annual costs for the first 13 plants were CHF 12–55 per connected person including capital, against the chapter’s EUR 5. - England never built the national granular activated carbon (GAC) programme whose EUR 32–37 billion estimate the chapter cites, so that figure has never been tested. England’s current chemicals programme is far smaller: £387m for more than 700 measures across all chemicals, 2025–2030 [S23]. - EU-wide, the JRC’s updated estimate for quaternary treatment is EUR 1.48–1.8 billion a year at full implementation [S10]. Industry and Poland contest this as an underestimate [S12].

4. The ecological and human-health science has become more nuanced, and the chapter’s hedges look well judged. - Feminisation persists. The Environment Agency’s 2017 resurvey found intersex at 60% of sites, with no significant change in frequency at 80% of them. Vitellogenin had fallen and no feminised ducts were seen, which suggests lower exposure [S24]. - English roach populations appear self-sustaining despite feminisation [S25]. There is no clear genetic adaptation to oestrogen [S26]. This is the chapter’s “maybe they will not” (p. 296). - The Canadian lake population recovered within about three to four years once EE2 additions stopped [S27]. The same experiment showed knock-on food-web effects, including a 23–42% fall in lake trout biomass [S28]. - Human health. Meta-analyses report continuing, even accelerating, declines in sperm counts [S37, S38], though this remains contested [S39]. The chapter’s scepticism about an oestrogen cause and its pointer to anti-androgens have aged well [S40–S42]. - Substitution has not happened. In English community prescribing, EE2-free combined pills were under 1% of combined-pill items in 2025 [S33]. EE2 items did fall by about half between 2014 and 2025, however, as prescribing shifted towards progestogen-only methods for reasons unrelated to the environment [S33].

Net weight. - The chapter’s diagnosis holds up well: potency, poor removal, measurement limits, and cost as the axis of contestation. - So does its central governance lesson: delay under a cost-conditioned precautionary framing. - Its specific numbers (UK and Swiss costs, timelines) should be treated as indicative only. - Its population-level concern should be read in the calibrated way the authors themselves wrote it.

Claim-by-claim#

Claim 1: EU regulatory trajectory. The January 2012 proposal listed EE2 and E2 as priority substances with an EE2 annual-average standard of 0.035 ng/L and possible enforcement by 2021; a July 2012 amendment would defer the standard to the 2016 review, with compliance by 2027 (pp. 293–294)#

Original claim (pp. 293–294; also p. 282 and p. 296). - The Commission’s January 2012 proposal listed E2 and EE2 with an EE2 annual-average environmental quality standard (EQS) of 0.035 ng/L for inland surface waters. It “could be taken into account in the 2015 … River Basin Management Plans … with enforcement required by 2021” (p. 293). - The July 2012 draft report (rapporteur Richard Seeber) proposed not specifying an EQS for “substances of pharmaceutical relevance” until the 2016 review, “with the aim of meeting the EQS by 2027” (p. 294, fn 9). The chapter warned that the decision “may be stalled by recent representations … from both the water and pharmaceutical industries” (p. 294), and that even if agreed it “will not come into force until at least 2015” (p. 296).

Subsequent developments - 2013: watch list, not priority listing. - Directive 2013/39/EU (12 August 2013) added other priority substances but not the oestrogens. Article 8b required that “Diclofenac …, 17-beta-estradiol (E2) … and 17-alpha-ethinylestradiol (EE2) … shall be included in the first watch list, in order to gather monitoring data for the purpose of facilitating the determination of appropriate measures to address the risk posed by those substances” [S1]. - Article 8c required a “strategic approach to pollution of water by pharmaceutical substances” within two years, with proposals for measures by 14 September 2017 [S1]. Recital 15 called pharmaceutical residues “an emerging environmental concern” [S1]. - 2015–2019: monitoring. The first watch list (Decision (EU) 2015/495, 20 March 2015) added oestrone (E1). It set a maximum acceptable method detection limit for EE2 of 0.035 ng/L [S2]. The 2018 list kept E1, E2 and EE2 [S3]. The 2020 list removed them because the maximum four-year monitoring period had expired in 2019 [S4]. - 2019: the strategic approach arrived three and a half years late (11 March 2019). It cited the feminisation of male fish exposed to “the main ingredient in the contraceptive pill” [S35]. - 2022: the Commission’s proposal (COM(2022) 540, 26 October 2022) proposed listing EE2, E2 and E1 individually [S6, S7]. - The impact assessment found a “relatively large distance to target” for EE2, meaning widespread and large exceedances. - It judged that individual listing would “place onus on source control” [S6]. - It recorded “possible societal impacts from loss of use (contraceptive pill, HRT, hormone treatments) if Ethinyl estradiol (EE2) is restricted/banned” [S6]. - 2026: adoption. The procedure ran: - Parliament first reading, April 2024; - interinstitutional agreement, 21 October 2025; - Council position, 5 March 2026; - Parliament second reading, 26 March 2026; - publication as Directive (EU) 2026/805 on 20 April 2026 [S7, S8]. The directive: - lists EE2 (No 46), E2 (No 47) and E1 (No 59) as priority substances, category “Pharmaceuticals – estrogenic hormone”; - sets annual-average EQS of 1.7 × 10⁻⁵ μg/L (0.017 ng/L) for EE2 in inland waters and 1.6 × 10⁻⁶ μg/L (0.0016 ng/L) in other surface waters; - sets 0.00018 μg/L (0.18 ng/L) for E2 and 3.6 × 10⁻⁴ μg/L (0.36 ng/L) for E1 in inland waters; - derives no short-term maximum (MAC-EQS) [S8]. For newly identified substances, the standards take effect from 22 December 2027, “with the aim of achieving good surface water chemical status … by 22 December 2039”. A final programme of measures is due in the 2033 river basin management plans, and transposition by 21 December 2027 [S8]. - Related measures. - Directive (EU) 2024/3019 (the urban wastewater recast, 27 November 2024, in force 1 January 2025) requires quaternary treatment at all plants of at least 150,000 population equivalent (p.e.): 20% of them by 2033, 60% by 2039 and all by 2045. It also applies to agglomerations of at least 10,000 p.e. discharging into listed at-risk areas, in steps from 10% by 2033 to 100% by 2045 [S9, Art. 8]. - At-risk areas include rivers with a dilution ratio below 10 and areas where extra treatment is needed to meet water-quality standards [S9, Art. 8(2)]. - Producers of human medicines and cosmetics must cover “at least 80 % of the full costs” of quaternary treatment and micropollutant monitoring from 31 December 2028 [S9, Art. 9, Annex III]. - The UK after EU exit. I found no EE2 standard in England. The Environment Agency’s 2025 statement of chemical challenges lists the 2015 “specific pollutants” and discusses oestrogenic effects in roach, but does not mention an EE2 standard or an EE2 permit limit [S23]. (This is absence of evidence in the documents I retrieved, not proof that none exists.)

Complications - The 2013 watch-list wording (“to address the risk posed”) shows the EU accepted the risk in 2013 and treated monitoring as a precursor to measures, not as a test of whether a risk existed [S1]. - The eventual standard is stricter than the 2012 proposal (0.017 against 0.035 ng/L). The long delay therefore did not weaken the scientific target. It deferred the compliance date and changed who pays.

Verdict: strengthened. The chapter’s worry about delay proved, if anything, too mild. Even the “delaying” amendment’s timetable (review 2016, compliance 2027) was overtaken by a 2022 proposal and a 2039 target. But the decision was not stalled for good, and the standard adopted is tighter than the one proposed.

Implications for weight. The chapter’s account of how a cost-conditioned precautionary process defers action (p. 296) is well supported by what followed. The sequence was a monitoring instrument, then a missed review date, then a stricter standard with a distant compliance date. Readers should not infer that delay means dilution. In this case delay coexisted with a tightening of the science-based target, and the burden moved from standard-setting to financing and implementation.

Claim 2: Conventional sewage treatment removes only about 54% of EE2 at best, so the 0.1 ng/L no-effect concentration is exceeded in many UK effluents whatever the treatment; GAC can bring effluent below it with no significant fish vitellogenin or intersex response (p. 292)#

Original claim (p. 292). - In the UK National Demonstration Programme, nitrifying activated sludge “removed some 54 %”, with tertiary treatment removing a further 0–38% of the remainder. - “the proposed EE2 Predicted No Effect Concentration (PNEC) of 0.1 ng/L is exceeded in many UK final effluents … irrespective of conventional treatment type”. - GAC produced “final effluents below the EE2 PNEC and with no significant induction in fish VTG or intersex”. The chapter notes that “in one study reproduction in fish was slightly, but significantly impacted” (p. 292).

Subsequent developments - The UK Chemicals Investigation Programme (CIP). The ten large water and sewerage companies of England and Wales funded: - CIP1 (£25m, 2010–2015); - CIP2 (£140m, 2015–2020), which prioritised 600 works and “carried out technology trials”; - CIP3 (£25m, 2020–2022) [S19]. CIP4 runs over 2025–2027 [S23]. The effluent database holds about 3 million determinand readings and is open to anyone who registers [S19]. - EE2 is among the most poorly removed pharmaceuticals. An analysis of the two UK-wide programmes (45 works sampled on 20 occasions) found “poorer removal … for ethinyloestradiol, diclofenac, propranolol, the macrolide antibiotics, fluoxetine, tamoxifen and carbamazepine”. All but the last two were in effluents at concentrations above their estimated PNEC. “as many as 890 WwTW in the UK (approximately 13% of all WwTW) may cause exceedances of estimated riverine PNECs” for the set of pharmaceuticals studied [S21, abstract]. The Environment Agency repeats this conclusion in its 2025 statement [S23]. - Ten-year trends. A summary of 2010–2020 CIP effluent data found significant falls for nickel, DEHP, nonylphenol, tributyltin, brominated diphenyl ethers and triclosan, attributed to “tighter regulatory controls and/or improved wastewater treatment”. EE2 is not among the substances named as falling [S20, abstract]. - Advanced treatment works at full scale. - Switzerland’s first long-term full-scale ozonation plant (Neugut, 105,000 p.e.) removed “a broad range of micropollutants by >79% on average” across 550 substances at 0.55 g O₃/g dissolved organic carbon. A biological or GAC post-treatment was needed to remove ozonation by-products [S18, abstract]. - Swiss and EU law now define performance by 80% removal of a basket of indicator substances. Pharmaceuticals such as carbamazepine and diclofenac are used, not EE2, because EE2 concentrations are too low to measure reliably [S9, Annex I Table 3; S16]. - Lab-scale and review evidence. A 2022 review of more than 70 post-2014 studies reported chemical, biological, adsorptive and ion-exchange methods with “up to 100% removal efficiency”. Most were tested at μg/L concentrations far above environmental levels, the same limitation the chapter flags (p. 292) [S31].

Complications - I found no later UK publication that re-estimates the 54% figure itself. - CIP2’s technology-trial reports are held in the UKWIR library and were not retrievable here. I cannot say what they concluded about EE2 specifically.

Verdict: held up. Independent, industry-funded, national-scale monitoring confirmed that conventional treatment leaves EE2 above effect thresholds in many effluents. Advanced treatment (ozone, activated carbon) is now proven at full scale for broad micropollutant removal.

Implications for weight. This is one of the chapter’s firmest empirical foundations. One caution: regulators now judge treatment by removal of proxy substances rather than EE2 itself. Whether EE2 targets are actually met downstream will have to be checked by the new EQS monitoring and effect-based methods (Claim 6).

Claim 3: The cost of compliance. GAC at about 1,360 works in England and Wales would cost EUR 32–37 billion (provisional UK estimate) plus about 14 kg CO₂ per person per year; EU-wide estimates range from EUR 11–18 per person per year (EC) to 25–50% of sewerage charges (EUREAU); sand filtration is almost as effective as GAC against feminisation at two-thirds the cost (pp. 292–293)#

Original claim (pp. 292–293, 295). - The UK figure is a “provisional” government estimate cited via Owen and Jobling (2012), together with capital costs of over EUR 3m for a 50,000 p.e. works and over EUR 8m for 250,000 p.e. - “There is clearly considerable uncertainty here”, and national costs “could be considerably lower” elsewhere (p. 293). - Sand filtration after activated sludge was “almost as effective as GAC at preventing the feminisation of male fish albeit it was two thirds cheaper” (Baynes et al. 2012; p. 293). - The chapter asks whether it is “scientifically incorrect to blame just EE2 for the costs” when GAC would remove other substances too (p. 293).

Subsequent developments - England did not build the programme. Nothing I retrieved shows a national GAC roll-out for EE2. - England’s current Water Industry National Environment Programme (WINEP, 2025–2030) includes “over 700 separate measures of source control or end of pipe control for chemicals … costing the water industry £387 million, which includes £70 million of chemical investigations” [S23]. That covers all chemicals, not EE2 specifically. - The Environment Agency’s 2020 fish survey found that “an absence of chemical monitoring data prevented further exploration” of whether treatment upgrades explained lower exposure [S24]. - The EUR 32–37 billion estimate has therefore never been tested against outturn. - The EU reframed the problem, which answered the chapter’s question about blaming EE2. The EU chose broad-spectrum quaternary treatment for a class of micropollutants, not EE2-specific treatment [S9, Art. 8]. - The 2022 impact assessment put costs at about EUR 1.2 billion a year by 2040 (EUR 1.56 billion in 2025 prices). - The JRC’s December 2025 update found “higher per-capita treatment expenses, yet total costs remain within the original uncertainty margins”. It projects EUR 1.48–1.8 billion a year at full implementation in 2045, or −5% to +15% against the inflation-adjusted estimate [S10, abstract]. - Spread across the whole EU population (about 449 million; my calculation, for scale only), that is roughly EUR 3–4 per resident per year. The costs are in fact concentrated on connected populations at the plants that must upgrade. - Who pays, and the dispute over cost. - EPR places at least 80% on producers of medicines and cosmetics [S9, Art. 9]. - Recital 23 sets contributions “proportionate to the quantities … and the hazardousness”, in order “to favour the substitution of substances and products generating micropollutants residues” [S9]. - EFPIA (the European pharmaceutical industry federation) sued on 7 March 2025. Its first plea is that the impact assessment and supporting study “overestimate the contribution of human medicines to micropollution and underestimate the costs of the Directive’s required quaternary treatment” (T-158/25) [S12]. - Poland (C-193/25) argues the costs are “disproportionate to the expected results” and that the rules breach polluter-pays by ignoring other emitters [S12]. - The Commission undertook in June 2025 to “conduct an updated study of costs and its potential impacts on concerned sectors” [S11]. The JRC update above appears to be the cost part of that commitment. - Swiss outturn is covered under Claim 4. - Carbon. The Swiss experience confirms that micropollutant stages “increase the greenhouse gas emissions of a WWTP”. Electricity use is 2–9 kWh per p.e. per year, higher for ozone than activated carbon [S16]. I found no like-for-like check of the chapter’s 14 kg CO₂ per person per year for GAC, which is mainly embodied carbon in the activated carbon itself. - Sand filtration. Regulators did not adopt it as a compliance route. In Switzerland it appears only as a post-treatment after ozone or powdered activated carbon (“O3 + SF”, “PAK vor SF”) [S17]. The EU’s 80% indicator-removal standard is framed around broad-spectrum processes [S9, Annex I Table 3]. I found no later replication of the Baynes et al. result in the retrieved literature.

Complications - The chapter’s UK figure and the EU figures measure different things: GAC at about 1,360 works to meet a stringent EE2 EQS, against quaternary treatment at larger plants to meet a removal-percentage standard. They cannot be compared directly. - The EU’s lower aggregate figure does not show that the UK estimate was wrong. It shows that the framing of the compliance measure drives the cost, as the chapter argued (pp. 293, 295).

Verdict: partly held up. - Held up: the claims that costs would be large, contested, dependent on context and a brake on decisions. Cost is now litigated. - Untested: the specific UK figure, because the programme was never built. - Not taken up: the sand-filtration option.

Implications for weight. The lesson that cost estimates are contested evidence, varying with who produces them and with the chosen compliance measure (digest, “Strong”), is reinforced. The chapter’s numbers should not be used as realised costs. The chapter’s question about attributing the cost to EE2 alone was answered in practice by pooling EE2 with other micropollutants and charging producers, a route it did not discuss.

Claim 4: Switzerland will fit full-scale micropollutant treatment at more than 100 plants treating about 80% of municipal wastewater; the first full-scale ozonation plant will open in spring 2013 at about EUR 5 per person per year (p. 295, fn 10)#

Original claim (p. 295, fn 10). “Full-scale treatments will be installed on more than 100 Swiss WWTP treating about 80 % of the Swiss municipal wastewater. The first full-scale ozonation plant will go in operation in Spring 2013 and will cost only 5 EUR/person/year including both capital and operational costs, because sand filtration already exists on this plant.” The chapter presents Switzerland as acting “even in the absence of regulation of EE2 and E2”.

Subsequent developments - Legal basis and funding (VSA, April 2025) [S16]: - Selection criteria have applied since 2016: plants with more than 80,000 people connected; more than 24,000 in lake catchments; more than 8,000 where the receiving water has a high share of wastewater. - All plants pay a national wastewater fee of CHF 9 per person per year into a fund “limited until 2040”. - The fund pays 75% of eligible initial investment. Plants already operating a micropollutant stage are exempt from the fee. - Scale and coverage. “a total of around 120 WWTPs are to be equipped … by 2040”, treating “around 70% of domestic wastewater”. A further 15 are considering merging with other plants [S16]. - Pace. - The VSA list dated December 2025 shows 39 plants in operation and 46 in planning or construction, some of which will divert flows rather than build new stages [S17]. - The first was Neugut (ozone plus sand filtration), in operation since 2014 [S17]; Bourgin et al. describe it as “the first WWTP in Switzerland to implement a long-term full-scale ozonation” [S18]. - Three plants were running by the end of 2016, eight by the end of 2018, 33 by spring 2025 and 39 by December 2025 [S16, S17]. - Realised costs. For the first 13 micropollutant stages, annual operating costs were CHF 2.5–17 per connected person. Including capital (interest and depreciation), annual costs were “between CHF 12 and 55 per connected person and year”, or CHF 7–32 per p.e. [S16]. - Scope expanded. - Since 2020, Swiss law sets ecotoxicological limits in surface waters for three drugs: azithromycin, clarithromycin and diclofenac. - The current programme halves the length of river exceeding them, but “measures are required at a further 300 or so WWTPs”, which Motion 20.4262 instructs the government to require [S16]. - The VSA reports “first positive changes visible in the waters”, including sharp falls in diclofenac in the Glatt and improved brown-trout gene-expression markers [S16].

Complications - The chapter’s EUR 5 figure was for one large plant that already had sand filtration. The VSA’s range covers mixed plant sizes and technologies, so the comparison is not like-for-like. - On any reading, though, the typical realised cost is several times the EUR 5 figure. - The VSA is the sector’s own association, and its report is an implementation review, not an independent evaluation.

Verdict: partly held up. - Direction correct. Switzerland did build a national programme without an EE2-specific rule, funded by a dedicated fee. - Specifics optimistic. The first plant opened a year later than projected. Coverage will be about 70%, not 80%, although more plants (about 120) are involved than the “more than 100” projected. By the end of 2025 only a third of the targeted plants were running, and costs per person are well above EUR 5.

Implications for weight. - The Swiss case supports the claim that a policy-led, broad micropollutant approach is feasible and can command acceptance. The VSA credits “early involvement of the relevant stakeholders” [S16]. - It also shows the usual pattern of early-mover projections: slower roll-out, higher unit costs and later expansion of scope. - Treat the EUR 5 figure as a best-case plant, not a national average.

Claim 5: There is “reasonable certainty” that EE2 plays a significant role in fish reproductive harm and “extremely high” certainty about the risks of oestrogen mixtures; population-level effects remain uncertain, since wild populations “might collapse in the future … but maybe they will not” (pp. 294–296)#

Original claim (pp. 294–296). - “There is reasonable certainty, based on sufficient scientific evidence, that EE2 plays a significant role in causing the reproductive health impacts observed in fish” (p. 294). - “the level of scientific certainty concerning the risks … posed by oestrogens as mixtures … is extremely high” (p. 295). - But: “Uncertainties remain, particularly around fish population level effects” and “These populations might collapse in the future, as indicated in the Great Lakes study of EE2 (Kidd et al., 2007), but maybe they will not” (pp. 295–296). The Kidd study was in fact in Canada’s Experimental Lakes Area, not the Great Lakes (digest).

Subsequent developments - Institutional acceptance of the hazard and mixture risk. - The EU listed all three steroid oestrogens after scientific-committee review of the EQS dossiers. It found a “relatively large distance to target” for EE2 [S6, S8]. - It mandated effect-based monitoring because “the cumulative risk from estrogenic pharmaceuticals should be addressed by effect-based monitoring” [S8, recital 15]. - The Commission’s 2019 strategic approach cites fish feminisation as the leading example of pharmaceutical harm [S35]. - Feminisation persists, at apparently lower exposure. The Environment Agency revisited 10 historical sites in 2017 and sampled 466 roach (report published 30 October 2020; authors Baynes, Jobling, Lange, Tyler) [S24]: - “feminisation in wild male roach is still a widespread phenomenon, present at 60% of the sites sampled”; - “For the majority (80%) of sites, the frequency of intersex has not significantly changed”; - male plasma vitellogenin fell significantly at 7 of 10 sites, though it stayed above baseline at most; - no males had feminised ducts, against 94% of sites historically. The authors read this as “suggestive of a reduction in environmental exposure … but with continued impacts”. The survey was smaller than earlier ones and had no chemistry. The Environment Agency’s 2025 statement repeats the finding and announces a wider two-year CIP4 resurvey for 2025–2026 [S23]. - English roach populations appear self-sustaining. - Hamilton et al. (2014; co-authors include Jobling, Sumpter and Tyler) estimated effective population size (Ne) at 28 sites. They found “no significant negative correlation between N(e) and the predicted estrogen exposure”, and populations confined to effluent-dominated stretches “for multiple generations” that “have survived”. They caution that “a reduction in N(e) of up to 65% is still possible for the most contaminated sites” [S25]. - Hamilton et al. (2020) found “no strong evidence” of genetic selection driven by oestrogen exposure [S26]. - The lake experiment: collapse, recovery and food-web effects (independent of the chapter authors). - After EE2 additions at 5–6 ng/L stopped, the fathead minnow population recovered. Vitellogenin returned to baseline by year 3, and abundance and size structure recovered by spring of year 4. The recovered fish were descendants of the original population [S27, 2015]. - The same experiment found no direct toxicity to algae, microbes or invertebrates. But lake trout biomass fell 23–42%, “most probably an indirect effect from the loss of its prey species” [S28, 2014]. - Wider context. A 2025 review of chemical pressures on British river invertebrates ranked metals and pesticides as greatest concern. It judged pharmaceuticals “relatively low risk” to invertebrates on available data, with some exceedances “requir[ing] further investigation” [S30]. That review concerns invertebrates, not fish, so it does not test the chapter’s claims directly.

Complications - The English field evidence is dominated by one research group that includes the chapter’s first author. Its population-genetic findings cut against alarm, which argues against motivated reasoning, but they are not independent replication. - The Environment Agency survey could not attribute effects to EE2 specifically because there was no chemistry [S24]. - Rivers typically carry EE2 below the 5–6 ng/L of the lake experiment (digest caveat).

Verdict: held up. The chapter’s calibrated statement fits the later evidence well: - Hazard: EE2 and oestrogen mixtures cause reproductive harm in fish, now formally accepted by the EU. - Persistence: widespread feminisation continues in English rivers. - Uncertainty: populations studied in English rivers appear to persist, while controlled high-exposure lakes show collapse and, once exposure stops, recovery. The chapter’s “but maybe they will not” has so far been borne out for English roach.

Implications for weight. The lesson that contestation moves from “is it happening?” to “so what?” (digest, “Strong”) is reinforced. The “so what” question about populations remains the hinge and is now partly answered in the less alarming direction. Anyone using this case as a lens should keep two findings apart: individual-level harm, which is persistent, widespread and irreversible in the individual, and population-level harm, which has not been shown in rivers but is plausible at higher exposure and has indirect food-web effects. The chapter itself keeps them apart.

Claim 6: Measuring compliance at the proposed EQS is “hardly routine”, immunoassays have selectivity and sensitivity problems, and so an EE2 EQS “may be very hard to implement” (p. 295)#

Original claim (p. 295). “The techniques available to measure such very low levels of EE2 in natural waters and effluents are hardly routine.” Immunoassay methods have been “fraught with issues around selectivity and sensitivity”. So “while it is technically possible to develop an EQS, in practice it may be very hard to implement”. It will also be “costly to monitor and demonstrate legal compliance itself”.

Subsequent developments - Watch-list experience. - The EU set a maximum acceptable method detection limit of 0.035 ng/L for EE2 in 2015 and 2018 [S2, S3]. - JRC scientists later reported that first watch-list results “showed that some countries had difficulties to reach an analytical Limit of Quantification (LOQ) below or equal to the Predicted No-Effect Concentrations (PNEC) or EQS”. The JRC ran workshops and a literature review on EE2 analysis [S14, abstract; see also S15]. - A 2022 review found EE2 “well detectable in many countries, both above and below” 0.035 ng/L, “although analytical methods used for the quantification often are unsatisfactory regarding their limit of detection” [S31]. - The new standard is lower still. - Directive 2009/90/EC requires a limit of quantification “equal or below a value of 30 % of the relevant environmental quality standards”. Where no method meets this, monitoring uses “best available techniques not entailing excessive costs” [S13, Art. 4]. - At 0.017 ng/L inland and 0.0016 ng/L in other waters [S8], that implies LOQs of about 0.005 ng/L and 0.0005 ng/L (my calculation). - The 2026 directive itself states that “the analytical methods available on the market are not always sensitive enough to achieve the proposed quality standards”. It asks the Commission to assess a voluntary joint EU monitoring facility [S8, recital 29]. - The 2022 impact assessment warned that a “lack of granular data for E1, E2, EE2 … could lead to less effective management” [S6]. - Shift to effect-based methods. - Member states must monitor oestrogenic activity with effect-based methods over two years from 1 January 2030, alongside chemical monitoring of E1, E2 and EE2. The Commission must adopt technical specifications by 1 December 2027, then report on whether an “effect-based trigger value for estrogens” can be used for screening and chemical status [S8, Art. 8a(3)–(5) of Directive 2008/105/EC as amended]. - The definition of good chemical status now anticipates effect-based trigger values [S8]. - The directive states that “sufficiently robust effect-based monitoring methods already exist for estrogenic substances” [S8, recital 23]. - These are bioassays measuring total oestrogenic activity, not the single-substance immunoassays the chapter criticised.

Verdict: strengthened. The implementation problem was real, persisted through the watch-list period, and is acknowledged in the text of the final law. Because the EQS is now stricter, the analytical challenge is harder than in 2012. The regulatory response has been to add a mixture-level biological metric rather than rely on substance-by-substance chemistry alone.

Implications for weight. The lesson that measurement capability limits both knowledge and enforceability (digest, “Strong”) is among the best supported in the chapter. One consequence the chapter did not foresee: measurement limits pushed regulators towards effect-based, mixture-level compliance metrics, which align better with the chapter’s own claim about mixture risk (p. 295).

Claim 7: Replacing EE2 as the pill’s active ingredient, or redesigning pharmaceuticals to protect the environment, would be “a constructive precautionary approach” (p. 295)#

Original claim (p. 295, fn 11; p. 296). The chapter asks whether EE2 could be replaced “leaving the remaining oestrogens to be removed conventionally at less cost”. It suggests industry could keep contraceptive choice “while substantively altering the design of pharmaceuticals to protect the environment”. It poses this as a question (“Is this option preferable?”) and notes research on progestogen-only pills (fn 11). Elsewhere it notes that progestins also affect fish (p. 290; digest).

Subsequent developments - No environmental substitution. - I queried NHSBSA open data for items dispensed in the community in England [S33]. Combined oral contraceptives without EE2 were about 0.7% of combined-contraceptive items in calendar 2025: estradiol valerate with dienogest (8,894 items), estradiol with nomegestrol (8,217) and estetrol with drospirenone (1,220), against about 2.48 million EE2-containing items. - The European Commission’s 2022 impact assessment treated restriction of EE2 as a potential “societal impact” to avoid, not as an option [S6]. - EE2 use has nonetheless fallen, for non-environmental reasons. - In the same data, all EE2-containing items fell from 4.86 million (2014) to 2.52 million (2025), down 48%. Total prescription items rose 21% over the same period (1.065bn to 1.288bn) [S33]. - Progestogen-only contraceptive items (4.11m) now exceed combined hormonal items (2.49m); in 2014 the figures were 3.84m and 4.67m [S33]. - The UK consulted in 2021 on pharmacy (non-prescription) availability of desogestrel progestogen-only pills (GOV.UK consultation pages, July 2021), which moves some supply outside these data. - Caveat: these are community prescription items only. They exclude sexual-health clinic supply and over-the-counter sales, and items are not doses or kilograms. They show a direction of travel, not an environmental load. - The shift towards progestins moves, rather than removes, the environmental question (p. 290). No progestin was on any EU watch list through 2025 [S2–S5]. - A less potent oestrogen exists, with trade-offs. - Estetrol/drospirenone (Drovelis) was authorised in the EU on 19 May 2021 [S32]. Its EMA assessment gives estetrol (E4) a lowest fish life-cycle no-observed-effect concentration (NOEC) of 0.69 μg/L (690 ng/L), several hundred times the EE2 level at which the chapter reports breeding inhibition (above 2 ng/L, p. 290). - But the E4 dose is 15 mg against EE2’s 20–35 μg. A modelled 99.3% of E4 passes through treatment to surface water. The refined predicted concentration (0.0312 μg/L) gives a risk quotient “< 1” but awaited an updated assessment. - For the partner progestin, drospirenone, the assessment concluded “potential surface water environmental risk” [S32]. - Policy instruments that could reward redesign. - The 2019 EU strategic approach promised to “support the development of pharmaceuticals intrinsically less harmful for the environment”, subject to funding, and to consider “greener” procurement [S35]. - The Commission’s 2023 pharmaceutical proposal would make an incomplete environmental risk assessment grounds for refusal. It would also create a programme of assessments for medicines “authorised before 30 October 2005 that have not been subject to any ERA” [S36]. EE2 products belong to that legacy group, although the proposal does not name them. - The directive’s interinstitutional agreement was reached by March 2026, but it had not been adopted by September 2026 (plenary forecast 19 October 2026) [S36]. - The urban wastewater EPR sets fees by quantity and hazardousness “to favour the substitution of substances” [S9, recital 23].

Verdict: unclear. The substitution route the chapter floated has not been taken for environmental reasons, and policy momentum went to end-of-pipe treatment paid for by producers. Yet EE2 exposure from contraception may be falling in England anyway, because of clinical and consumer shifts. A lower-potency oestrogen exists but is marginal in use and brings its own environmental questions. Hazard-weighted EPR fees and legacy environmental assessments could still create incentives, but have not yet been tested.

Implications for weight. The later evidence supports the chapter’s implicit point (digest: “High-benefit products get end-of-pipe fixes”) more than its hopeful suggestion. Its own caveat that progestins also affect fish (p. 290) turns out to be important: substitution within a therapeutic class can shift, rather than remove, the environmental burden.

Claim 8: EE2 decisions were made in a “poorly understood, closed process” with the public as “silent witnesses”; without public input, decisions will default to costs alone (p. 297)#

Original claim (p. 297; also pp. 283–284). - “the public has been and continues to be silent witnesses”. Applying precaution “must encourage public participation” so that costs of action and inaction are debated openly. - Deciding “based largely on costs alone … loads the dice before they are thrown” (p. 297). - Earlier, a contract kept national survey results confidential until 1994 (p. 284).

Subsequent developments - Transparency of evidence improved in England. The CIP data portal (about 3 million readings from more than 600 works) is open to anyone who creates an account [S19]. Summary results appear in peer-reviewed papers [S20–S22]. The Environment Agency published its 2017 resurvey [S24]. This contrasts with the 1990s confidentiality the chapter describes (p. 284). - Consultation, not deliberation, at EU level. - The EU reviews used stakeholder consultation and scientific-committee review. The 2026 directive records “an extensive consultation with experts from the Commission services, Member States, stakeholder groups and” the scientific committee SCHEER [S8, recital 13]. The impact assessment also ran an online open public consultation. For surface water, respondents were “most concerned about pharmaceuticals (average score of 4.2; rated ≥4 by 72% of respondents)” among emerging pollutants [S6]. That is a public input the chapter found missing, though it is a self-selected survey, not a deliberation over acceptable harm. - I found no evidence of dedicated public deliberation on the acceptability of fish feminisation in the EU or the UK. - The Swiss VSA attributes the acceptance of its programme to “early involvement of the relevant stakeholders” [S16]. That is stakeholder, not general-public, involvement. - Cost allocation changed. The biggest governance shift is financial, not deliberative: - EPR puts at least 80% of quaternary-treatment costs on medicine and cosmetics producers [S9]. - The 2026 directive also requires a Commission report by 11 May 2029 on extending EPR to monitoring costs under the Water Framework Directive [S8, Art. 19a]. - Cost moved into the courts. - On 7–10 March 2025, sixteen actions were brought at the General Court (T-156/25 to T-171/25), by EFPIA, Cosmetics Europe and generic and originator firms including Teva, Zentiva, Fresenius Kabi and Polpharma. Poland brought a parallel action (C-193/25) [S12]. - On 18 February 2026 the General Court dismissed the company and association actions as inadmissible, for “Lack of individual concern”. An appeal (C-411/26 P) was lodged on 27 April 2026 [S12]. - On 5 June 2026 the Irish High Court referred validity questions (C-614/26). These include “the alleged significant underestimation of the annual costs” and the impracticality of substituting product constituents by 31 December 2028 [S12]. - Poland’s action had no judgment in the metadata I could query on 25 September 2026. - Delay has been partly procedural. The Commission delivered its pharmaceutical strategic approach in 2019 against a 2015 deadline [S1, S35], and its priority-substances proposal in 2022 against the 2016 review date [S6].

Complications - The chapter treats public participation as the remedy for cost-dominated decisions. The later record shows a different corrective, a polluter-pays reallocation of costs, which the chapter did not discuss (digest caveat). - That reallocation has not ended cost-centred argument. It has turned the argument into a legal dispute over proportionality and attribution.

Verdict: partly held up. - Held up: the observation that decisions ran through expert and stakeholder channels with little public deliberation. Cost remained the central axis of contest. - Weakened: the prediction that decisions would default to cost alone. The EU adopted a stricter standard in 2026 despite cost objections and shifted who pays. - Improved since the period the chapter describes: transparency of evidence.

Implications for weight. The “silent witnesses” diagnosis remains a fair description of process, but the chapter’s causal claim (no public input, so cost decides) is only partly supported. The later record points to another lever the chapter omitted: redesigning who pays can unlock action without resolving the public-values question. That lever creates its own contest, over the evidence used to apportion costs.

Claim 9: Human health. Sperm counts appear to have fallen (113 to 66 million/ml, 1940–1990); the oestrogen theory of testicular dysgenesis syndrome is “not entirely convincing”, and anti-androgens are more likely culprits (Box 13.1, pp. 285–288)#

Original claim (Box 13.1 and pp. 286–288). - Carlsen et al. (1992) found a decline from 113 to 66 million/ml between 1940 and 1990. European studies of young men were “consistent with sperm counts having fallen”. - The WHO “redefine[d] the ill people as ‘normal’” by lowering reference values (p. 287). - The oestrogen theory is “not entirely convincing”. Anti-androgens “are more likely culprits”, though “there is no human or experimental animal data to support this” (quoting Sharpe 2009; p. 288).

Subsequent developments - The decline, re-estimated. - Levine et al. (2017; 185 studies, 42,935 men, samples 1973–2011) reported “a significant decline in sperm counts … driven by a 50-60% decline among men unselected by fertility from North America, Europe, Australia and New Zealand” [S37]. - Their 2023 update (223 studies, samples 1973–2018) reported a global decline, including for the first time among unselected men in South/Central America, Asia and Africa. They suggest “this world-wide decline is continuing in the 21st century at an accelerated pace” [S38]. - Still contested. Boulicault et al. (2021) argue that the “Sperm Count Decline hypothesis” rests on contestable assumptions. They propose a “biovariability” framework in which much variation is “non-pathological and species-typical” [S39]. A 2024 Danish sperm-bank study found a recent decline in motility (title only) (Lassen et al., Hum Reprod 2024; not otherwise assessed here). - Causes. - Skakkebaek et al. (2016) conclude that “environmental exposures arising from modern lifestyle, rather than genetics, are the most important factors”. They regard several adult disorders as signs of testicular dysgenesis arising in the womb [S40]. - A systematic review of phthalates (anti-androgenic plasticisers) found “robust evidence of an association between DEHP and DBP exposure and male reproductive outcomes” [S41]. - The EU restricted four such phthalates in consumer articles from 7 July 2020 on reproductive-toxicity grounds [S42]. - I found no later evidence implicating contraceptive-derived EE2 in drinking water as a cause. - The WHO reference values. The 6th edition of the WHO semen manual (2021) says its ranges “have been misinterpreted as distinct limits between fertility and infertility”. It frames them as distributions from fertile men [S43]. This partly answers Box 13.1’s charge but does not resolve the public-health question.

Verdict: partly held up. - Decline: strengthened by larger meta-analyses but still disputed. - Scepticism about an oestrogen cause: held up. - Pointer towards anti-androgens: strengthened, both epidemiologically and in regulation. - The box’s rhetoric (WHO redefining “the ill” as “normal”) remains advocacy, not a finding.

Implications for weight. Treat Box 13.1 as context. Its most durable content is the authors’ own caution against the oestrogen hypothesis. The human-health material should not be used to strengthen the EE2 case itself, and the chapter does not try to.

Claim 10: Scope. EE2 is a “test case for many thousands of low-level pollutants”, and UK medicine use is predicted to more than double by 2050 (pp. 282, 284)#

Original claim (pp. 282, 284, 297). - EE2 “is in many ways a test case for many thousands of low-level pollutants that infiltrate our environment ubiquitously” (p. 282). - “the UK Office of National Statistics predicts that the country’s medicine usage will more than double by 2050” (p. 284, citing Nature 2011). - “EE2 is a perfect case study of how we are responding as a society” (p. 297).

Subsequent developments - The template spread. - The watch-list mechanism created for EE2, E2 and diclofenac became a standing EU instrument, now in its fifth list [S2–S5]. - The 2026 revision adds nine pharmaceuticals as surface-water priority substances: EE2, E2, E1, azithromycin, carbamazepine, clarithromycin, diclofenac, erythromycin and ibuprofen. It adds groundwater standards for carbamazepine, sulfamethoxazole and primidone. It anticipates standards for “sum(s) of selected pharmaceuticals by mode of action” [S8]. - The 2024 recast of the Urban Wastewater Treatment Directive treats micropollutants as a class requiring quaternary treatment [S9]. - The Environment Agency now cites a cross-government “Pharmaceuticals in the Environment (PiE) Group” [S23]. - Pharmaceutical pollution is recognised worldwide. A survey of 1,052 sites in 104 countries found concentrations of at least one pharmaceutical above safe levels for aquatic organisms, or above levels of concern for antimicrobial resistance, at 25.7% of sites [S44]. A companion analysis found potential ecotoxicological concern at about 43.5% of sites, including from progestins [S45]. - Medicine use. - English community prescription items rose from 1.065bn (2014) to 1.288bn (2025) [S33]. That is about 1.75% a year compound (my calculation), a rate that would roughly double use in about 40 years. - NHSBSA reports 1.26bn items in 2024/25, up 4% on 2023/24 [S34]. - The Environment Agency cites ageing as “a principal cause in the increase in pharmaceutical consumption” [S23]. - Items are not a measure of active-ingredient mass, and I could not retrieve the 2011 projection itself.

Verdict: strengthened. EE2 did become the test case the chapter described: the instruments invented for it were generalised to pharmaceuticals and micropollutants as a class. The medicine-use trend in England is broadly consistent with the projection on an items basis.

Implications for weight. The chapter’s framing claim is one of its most durable. Lessons drawn from EE2 were in practice applied to a wider class of pollutants. That supports using this case as a lens on how institutions handle diffuse, low-dose, high-benefit exposures.

Summary of verdicts#

Technology-neutral lessons this check supports (for later use as a lens)#

Each is tied to the section’s pages and to the later evidence above.

1. Deferral instruments can stretch a proposed limit into decades of delay without softening its scientific basis. A “monitor first” mechanism and missed review dates moved a 2021 compliance horizon to 2039, while the adopted standard became stricter than the one deferred. Delay and dilution are separable, and should be tracked separately (pp. 293–294, 296; [S1–S8]).
2. When a single-agent fix looks unaffordable, institutions may reframe the problem around a class of agents and a broad technology. That changes both the cost–benefit calculus and the unit of accountability. The chapter’s question “is it scientifically incorrect to blame just EE2 for the costs?” was answered by pooling (p. 293; [S9, S10, S16]).
3. Moving the payer can unlock action where public deliberation did not, but the contest then shifts to attributing costs. Producer-pays rules triggered litigation over proportionality and over the data used to apportion contributions (pp. 290–291, 296–297; [S9, S11, S12]).
4. Measurement limits outlast standard-setting and push regulators towards effect-based, mixture-level metrics. Standards can be set below what routine chemistry can verify. Proxy-based compliance (indicator substances, bioassays) then becomes the practical currency (pp. 284, 292, 295; [S2, S8, S13, S14]).
5. Individual-level harm and population-level harm can diverge for long periods. Persistent, widespread individual effects can coexist with self-sustaining populations. Controlled whole-system experiments show both collapse and recovery, plus indirect effects that single-species tests miss (pp. 290, 295–296; [S24–S28]).
6. Early-mover projections tend to understate time and unit cost, and scope tends to grow after roll-out. The first plant was later than projected, coverage lower, per-person cost higher, and hundreds more plants were added later (p. 295 fn 10; [S16, S17]).
7. High-benefit products tend to get end-of-pipe and financing fixes rather than substitution. Where product mix changes, it is usually for reasons other than the environment. Substitutes can relocate rather than remove the burden, as with progestins and higher-dose, lower-potency alternatives (pp. 289–290, 295; [S6, S32, S33]).
8. Opening evidence up is an institutional lesson that can be learned. A shift from confidential surveys to publicly accessible national monitoring data followed this case. Gaps in chemical monitoring still limit evaluation of whether upgrades worked (pp. 283–284; [S19, S24]).
9. A case framed as a test case does become a template. Instruments designed for one agent (watch lists, pharmaceutical priority listing, class-wide treatment) were generalised. Early institutional choices in a flagship case therefore carry disproportionate weight (pp. 282, 297; [S5, S8, S9]).

Sources#

All retrieved 25 September 2026 unless noted. EUR-Lex refused automated retrieval (HTTP 202). EU legal texts were read in their Official Journal form through the Publications Office Cellar (https://publications.europa.eu/resource/celex/<CELEX>); ELI/EUR-Lex addresses are given for readers.

EU water legislation and implementing decisions - [S1] Directive 2013/39/EU of the European Parliament and of the Council of 12 August 2013 amending Directives 2000/60/EC and 2008/105/EC as regards priority substances in the field of water policy. OJ L 226, 24.8.2013, p. 1 (recital 15; Articles 8b, 8c). https://eur-lex.europa.eu/eli/dir/2013/39/oj (read via https://publications.europa.eu/resource/celex/32013L0039) - [S2] Commission Implementing Decision (EU) 2015/495 of 20 March 2015 establishing a watch list of substances for Union-wide monitoring. OJ L 78, 24.3.2015, p. 40 (EE2 maximum method detection limit 0.035 ng/L; E1 added). https://eur-lex.europa.eu/eli/dec_impl/2015/495/oj (via CELEX 32015D0495) - [S3] Commission Implementing Decision (EU) 2018/840 of 5 June 2018 establishing a watch list … and repealing Decision (EU) 2015/495. OJ L 141, 7.6.2018, p. 9. https://eur-lex.europa.eu/eli/dec_impl/2018/840/oj (via CELEX 32018D0840) - [S4] Commission Implementing Decision (EU) 2020/1161 of 4 August 2020 establishing a watch list … (recital: watch-list obligation for E1, E2, EE2 “ceased in 2019”). https://eur-lex.europa.eu/eli/dec_impl/2020/1161/oj (via CELEX 32020D1161) - [S5] Commission Implementing Decisions (EU) 2022/1307 of 22 July 2022 and (EU) 2025/439 of 28 February 2025 establishing watch lists (checked for hormones and progestins; none listed). https://eur-lex.europa.eu/eli/dec_impl/2022/1307/oj ; https://eur-lex.europa.eu/eli/dec_impl/2025/439/oj (via CELEX 32022D1307, 32025D0439) - [S6] European Commission. Impact Assessment Report accompanying the proposal … amending Directive 2000/60/EC, Directive 2006/118/EC and Directive 2008/105/EC. SWD(2022) 540 final, 26 October 2022 (distance-to-target; option assessment for oestrogenic hormones; “possible societal impacts from loss of use”). Via https://publications.europa.eu/resource/celex/52022SC0540 ; proposal COM(2022) 540 final. - [S7] European Parliament Legislative Observatory. Procedure file 2022/0344(COD) (timeline to adoption as Directive (EU) 2026/805). https://oeil.europarl.europa.eu/oeil/en/procedure-file?reference=2022/0344(COD) - [S8] Directive (EU) 2026/805 of the European Parliament and of the Council of 30 March 2026 amending Directives 2000/60/EC, 2006/118/EC and 2008/105/EC. OJ L, 2026/805, 20.4.2026 (recitals 13, 15, 23, 29, 56; Art. 3(1a) deadlines; Art. 8a(3)–(5) effect-based monitoring; Art. 19a EPR report; Annex I Part A entries 46, 47, 59; transposition by 21 December 2027). http://data.europa.eu/eli/dir/2026/805/oj (via CELEX 32026L0805) - [S9] Directive (EU) 2024/3019 of the European Parliament and of the Council of 27 November 2024 concerning urban wastewater treatment (recast). OJ L, 2024/3019, 12.12.2024 (recitals 18, 20, 21, 23, 47; Art. 8 quaternary treatment; Art. 9 EPR; Annex I Table 3; Annex III). http://data.europa.eu/eli/dir/2024/3019/oj (via CELEX 32024L3019) - [S10] Pistocchi, A. Updated estimation of the costs of quaternary wastewater treatment in the EU: A comparison of cost models. JRC144745, European Commission Joint Research Centre, 10 December 2025 (abstract page only). https://publications.jrc.ec.europa.eu/repository/handle/JRC144745 - [S11] European Commission. European Water Resilience Strategy. COM(2025) 280 final, 4 June 2025 (commitment to “an updated study of costs” for UWWTD EPR). Via https://publications.europa.eu/resource/celex/52025DC0280 - [S12] Court of Justice of the European Union, notices and orders (OJ C series; metadata via the Publications Office SPARQL endpoint, https://publications.europa.eu/webapi/rdf/sparql): - Case T-158/25, EFPIA v Parliament and Council, action brought 7 March 2025, OJ C/2025/2674, 19.5.2025. http://data.europa.eu/eli/C/2025/2674/oj - Cases T-156/25 to T-171/25 (actions of 7 and 10 March 2025 by Accord Healthcare France, Dermapharm, EFPIA, Adamed, Fresenius Kabi, hameln pharma, Hexal, Exeltis, Polpharma, Puren, EG Labo, Teva Nederland, Zentiva, Cosmetics Europe, BGP Products, Laboratorios Normon). - General Court orders of 18 February 2026 in Joined Cases T-156/25 et al. (Urban wastewater I), T-158/25 (Urban wastewater II) and T-169/25 (Urban wastewater III): “Standing to bring proceedings – Lack of individual concern”. CELEX 62025TO0156, 62025TO0158, 62025TO0169. https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:62025TO0156 - Case C-193/25, Poland v Parliament and Council, action brought 10 March 2025, OJ C/2025/2189, 22.4.2025. http://data.europa.eu/eli/C/2025/2189/oj - Case C-411/26 P, appeal brought 27 April 2026 by Dermapharm and Others (CELEX 62026CN0411). - Case C-614/26, Irish Pharmaceutical Healthcare Association and Medicines for Ireland, reference from the High Court (Ireland) of 5 June 2026, OJ C/2026/4575, 7.9.2026. http://data.europa.eu/eli/C/2026/4575/oj - [S13] Commission Directive 2009/90/EC of 31 July 2009 laying down technical specifications for chemical analysis and monitoring of water status (Art. 4: LOQ at or below 30% of EQS; otherwise “best available techniques not entailing excessive costs”). https://eur-lex.europa.eu/eli/dir/2009/90/oj (via CELEX 32009L0090)

Analytical methods and watch-list reviews - [S14] Loos, R. et al. “Summary recommendations on ‘Analytical methods for substances in the Watch List under the Water Framework Directive’.” Science of the Total Environment (2024; online 20 November 2023). doi:10.1016/j.scitotenv.2023.168707 (abstract). https://doi.org/10.1016/j.scitotenv.2023.168707 - [S15] Loos, R., Marinov, D., Sanseverino, I., Napierska, D., Lettieri, T. Review of the 1st Watch List under the Water Framework Directive and recommendations for the 2nd Watch List. JRC111198, 25 June 2018 (abstract page only; PDF not retrievable). https://publications.jrc.ec.europa.eu/repository/handle/JRC111198

Switzerland - [S16] Wunderlin, P., Bitterwolf, S. Micropollutant elimination in Swiss wastewater treatment plants: Experiences and developments. VSA (Swiss Water Association), April 2025 (file dated 2 May 2025). https://micropoll.ch/wp-content/uploads/2025/05/20250502_Micropollutant-Elimination-in-Switzerland.pdf - [S17] VSA Platform “Verfahrenstechnik Mikroverunreinigungen”. Liste der ARA mit MV-Stufe (list of plants with micropollutant stages), status December 2025. https://micropoll.ch/wp-content/uploads/2025/12/20251223-Liste-MV.pdf (index page: https://micropoll.ch/Mediathek/liste-der-ara-mit-mv-stufe/) - [S18] Bourgin, M. et al. “Evaluation of a full-scale wastewater treatment plant upgraded with ozonation and biological post-treatments.” Water Research 129 (2018), online 20 October 2017. doi:10.1016/j.watres.2017.10.036 (abstract). https://doi.org/10.1016/j.watres.2017.10.036

United Kingdom: water industry and Environment Agency - [S19] UK Water Industry Research. Water Chemicals Investigation Programme (CIP1 £25m 2010–2015; CIP2 £140m 2015–2020; CIP3 £25m 2020–2022; data portal). https://ukwir.org/water-chemicals-investigation-programme - [S20] Gardner, M.J., Comber, S.D.W., Ellor, B. “Summary of data from the UKWIR chemical investigations programme and a comparison of data from the past ten years’ monitoring of effluent quality.” Science of the Total Environment (2022; online 4 April 2022). doi:10.1016/j.scitotenv.2022.155041 (abstract). https://doi.org/10.1016/j.scitotenv.2022.155041 - [S21] Comber, S., Gardner, M., Sörme, P., Leverett, D., Ellor, B. “Active pharmaceutical ingredients entering the aquatic environment from wastewater treatment works: A cause for concern?” Science of the Total Environment (2018; online 26 September 2017). doi:10.1016/j.scitotenv.2017.09.101 (abstract). https://doi.org/10.1016/j.scitotenv.2017.09.101 - [S22] Comber, S.D.W., Gardner, M.J., Ansell, L., Ellor, B. “Assessing the impact of wastewater treatment works effluent on downstream water quality.” Science of the Total Environment (2022; online 11 July 2022). doi:10.1016/j.scitotenv.2022.157284 (abstract). https://doi.org/10.1016/j.scitotenv.2022.157284 - [S23] Environment Agency. Chemicals: challenges for the water environment. First published 22 October 2021; updated 20 November 2025 (sections 4.3.6, 4.3.7, 4.4.1.2; WINEP 2025–2030 £387m). https://www.gov.uk/government/publications/chemicals-challenges-for-the-water-environment ; document: https://assets.publishing.service.gov.uk/media/691ef316c62d57eb3a7107d8/Chemicals_challenges_in_the_water_environment.odt - [S24] Baynes, A., Jobling, S., Lange, A., Tyler, C.R. Oestrogenic effects in wild fish: comparing contemporary and historical impacts. Environment Agency report SC170001/R, 30 October 2020. https://www.gov.uk/government/publications/oestrogenic-effects-in-wild-fish-comparing-contemporary-and-historical-impacts ; PDF: https://assets.publishing.service.gov.uk/media/61bb4ae8d3bf7f055fce755c/Oestrogenic_effects_in_wild_fish_comparing_contemporary_and_historical_impacts_-_report.pdf

Fish populations and ecosystems - [S25] Hamilton, P.B., Nicol, E., De-Bastos, E.S., Williams, R.J., Sumpter, J.P., Jobling, S., Stevens, J.R., Tyler, C.R. “Populations of a cyprinid fish are self-sustaining despite widespread feminization of males.” BMC Biology 12:1 (13 January 2014). doi:10.1186/1741-7007-12-1. https://doi.org/10.1186/1741-7007-12-1 - [S26] Hamilton, P.B. et al. (incl. Jobling, Tyler). “Investigation into adaptation in genes associated with response to estrogenic pollution in populations of roach (Rutilus rutilus) living in English rivers.” Environmental Science & Technology (online 23 November 2020). doi:10.1021/acs.est.0c00957 (abstract). https://doi.org/10.1021/acs.est.0c00957 - [S27] Blanchfield, P.J., Kidd, K.A., Docker, M.F., Palace, V.P., Park, B.J., Postma, L.D. “Recovery of a wild fish population from whole-lake additions of a synthetic estrogen.” Environmental Science & Technology 49 (2015), online 17 February 2015. doi:10.1021/es5060513 (abstract). https://doi.org/10.1021/es5060513 - [S28] Kidd, K.A. et al. “Direct and indirect responses of a freshwater food web to a potent synthetic oestrogen.” Philosophical Transactions of the Royal Society B 369 (1 November 2014). doi:10.1098/rstb.2013.0578 (abstract). https://doi.org/10.1098/rstb.2013.0578 - [S29] Wedekind, C. “Demographic and genetic consequences of disturbed sex determination.” Philosophical Transactions of the Royal Society B 372 (1 September 2017). doi:10.1098/rstb.2016.0326 (abstract; background only). https://doi.org/10.1098/rstb.2016.0326 - [S30] Poyntz-Wright, I.P., Harrison, X.A., Tyler, C.R. “Evaluation of the pollution pressures posed by groups of chemicals on British riverine invertebrate populations.” Biological Reviews (online 21 September 2025; 2026 issue). doi:10.1111/brv.70075 (abstract). https://doi.org/10.1111/brv.70075 - [S31] Klaic, M., Jirsa, F. “17α-Ethinylestradiol (EE2): concentrations in the environment and methods for wastewater treatment – an update.” RSC Advances 12 (27 April 2022). doi:10.1039/d2ra00915c (open-access full text, PMC9044539). https://doi.org/10.1039/d2ra00915c

Medicines, prescribing and pharmaceutical policy - [S32] European Medicines Agency. Drovelis (drospirenone/estetrol): EPAR – public assessment report, EMA/212533/2021; marketing authorisation 19 May 2021 (section 2.3.5, environmental risk assessment). https://www.ema.europa.eu/en/medicines/human/EPAR/drovelis ; PDF: https://www.ema.europa.eu/en/documents/assessment-report/drovelis-epar-public-assessment-report_en.pdf - [S33] NHS Business Services Authority. Prescription Cost Analysis (PCA) annual statistics, open data, calendar-year tables PCA_STP_SNOMED_2014, PCA_STP_SNOMED_2019 and PCA_ICB_SNOMED_2025, queried by SQL on 25 September 2026 (total items; items for BNF chemical substances containing “ethinylestradiol”; BNF paragraphs 7.3.1 and 7.3.2). https://opendata.nhsbsa.net/dataset/prescription-cost-analysis-pca-annual-statistics - [S34] NHS Business Services Authority. Prescription Cost Analysis – England 2024/25, 5 June 2025. https://www.nhsbsa.nhs.uk/statistical-collections/prescription-cost-analysis-england/prescription-cost-analysis-england-202425 - [S35] European Commission. European Union Strategic Approach to Pharmaceuticals in the Environment. COM(2019) 128 final, 11 March 2019. Via https://publications.europa.eu/resource/celex/52019DC0128 - [S36] European Commission. Proposal for a Directive … on the Union code relating to medicinal products for human use. COM(2023) 192 final, 26 April 2023 (recitals 70–73; environmental risk assessment provisions, including the programme for products authorised before 30 October 2005 and ERA monographs). Via https://publications.europa.eu/resource/celex/52023PC0192 ; procedure file 2023/0132(COD): https://oeil.europarl.europa.eu/oeil/en/procedure-file?reference=2023/0132(COD)

Human reproductive health - [S37] Levine, H. et al. “Temporal trends in sperm count: a systematic review and meta-regression analysis.” Human Reproduction Update 23(6) (1 November 2017). doi:10.1093/humupd/dmx022 (abstract). https://doi.org/10.1093/humupd/dmx022 - [S38] Levine, H. et al. “Temporal trends in sperm count: a systematic review and meta-regression analysis of samples collected globally in the 20th and 21st centuries.” Human Reproduction Update 29(2) (1 March 2023). doi:10.1093/humupd/dmac035 (abstract). https://doi.org/10.1093/humupd/dmac035 - [S39] Boulicault, M. et al. “The future of sperm: a biovariability framework for understanding global sperm count trends.” Human Fertility (online 10 May 2021; 2022 issue). doi:10.1080/14647273.2021.1917778 (abstract). https://doi.org/10.1080/14647273.2021.1917778 - [S40] Skakkebaek, N.E. et al. “Male reproductive disorders and fertility trends: influences of environment and genetic susceptibility.” Physiological Reviews 96(1) (1 January 2016). doi:10.1152/physrev.00017.2015 (abstract). https://doi.org/10.1152/physrev.00017.2015 - [S41] Radke, E.G., Braun, J.M., Meeker, J.D., Cooper, G.S. “Phthalate exposure and male reproductive outcomes: A systematic review of the human epidemiological evidence.” Environment International (online 16 October 2018). doi:10.1016/j.envint.2018.07.029 (abstract). https://doi.org/10.1016/j.envint.2018.07.029 - [S42] Commission Regulation (EU) 2018/2005 of 17 December 2018 amending Annex XVII to REACH as regards DEHP, DBP, BBP and DIBP. OJ L 322, 18.12.2018, p. 14 (restriction in articles from 7 July 2020). https://eur-lex.europa.eu/eli/reg/2018/2005/oj (via CELEX 32018R2005) - [S43] Björndahl, L. “A paradigmatic shift in the care of male factor infertility: how can the recommendations for basic semen examination in the sixth edition of the WHO manual and the ISO 23162:2021 standard help?” Reproductive BioMedicine Online (online 14 March 2022). doi:10.1016/j.rbmo.2022.03.011 (abstract). https://doi.org/10.1016/j.rbmo.2022.03.011

Global pharmaceutical pollution - [S44] Wilkinson, J.L., Boxall, A.B.A., Kolpin, D.W. et al. “Pharmaceutical pollution of the world’s rivers.” PNAS 119(8) (online 14 February 2022; issue 22 February 2022). doi:10.1073/pnas.2113947119 (abstract). https://doi.org/10.1073/pnas.2113947119 - [S45] Bouzas-Monroy, A., Wilkinson, J.L., Melling, M., Boxall, A.B.A. “Assessment of the potential ecotoxicological effects of pharmaceuticals in the world’s rivers.” Environmental Toxicology and Chemistry 41 (2022), online 22 June 2022. doi:10.1002/etc.5355 (abstract). https://doi.org/10.1002/etc.5355