Late Lessons, Jensen Huang and AI
Markdown version

LL1-14 hindsight check: Hormones as growth promoters (Bridges and Bridges), Late lessons from early warnings (EEA, 2001), Ch. 14, pp. 149–156#

Checked 25 September 2026. The check covers what happened between 2001 and September 2026. Page numbers refer to the 2001 report. “My calculation” marks arithmetic I did myself from the cited figures.

Overview#

The chapter makes three kinds of claim: - A science verdict. Oestradiol-17β is a “genotoxic carcinogen”, and newer research “probably justifies” the EU ban (pp. 153–154). - Observations about how the evidence was produced. - What assessors were asked to look at. - Which subgroup was at risk. - What the drug agencies ignored. - That nobody generated the decisive evidence. - A political-economy account. Costs, illegal use, the WTO and sanctions.

A quarter-century later, the observations about evidence have held up best, the science verdict is still contested, and some of the numbers do not survive checking.

Vindicated or strengthened - Low-baseline children (Claim 3). - Each generation of more sensitive assays has confirmed the chapter’s point. - Aksglaede et al. (2006) concluded that prepubertal children’s oestradiol had been overestimated. So had the daily production rates in the FDA’s 1999 estimates, which regulators were still using. - A GC-MS/MS study found oestradiol in prepubertal boys “undetectable or extremely low”, with a median below 3.7 pmol/L, that is under about 1 pg/ml (Courant et al., 2010). - The EU wrote this group into law. Directive 2003/74/EC names prepubertal children as “the group of greatest concern”. - The WTO Appellate Body (2008, para. 725) faulted a panel for dismissing this evidence. - Outside the EU, regulators did not act on it. In 2022 the FDA revised its permitted oestradiol increments using new food-consumption values. The permitted extra intake stayed at about 60 ng a day (my calculation). That implies the old endogenous-production baseline was kept. - Environmental blind spot (Claim 5). - Field and laboratory work confirmed that feedlot effluent is hormonally active and alters fish reproduction (Orlando et al., 2004; Soto et al., 2004; Durhan et al., 2006). - Metabolites of the synthetic androgen trenbolone can re-form in surface water after apparently breaking down (Qu et al., 2013; Ward et al., 2015). That directly undercuts the “dilution and degradation” assumption the chapter criticised. - Two qualifications: - The main culprit in the field studies was an androgenic promoter, not the oestrogenic ones the chapter focused on. - Population-level effects remain unestablished (Ankley et al., 2018). - Evidence stalemate (Claim 4). - No post-2001 study I found resolves whether the ban protected health or whether continued use is safe. - EFSA (2007) found the contribution of residues to hormone-dependent cancers “unknown”. - A 2015 review concluded that risks “cannot” yet be quantified (Nachman and Smith, 2015). - The FDA still requires no official analytical method to monitor oestradiol or trenbolone residues (FOI summary, May 2026).

Contested, overstated or wrong - “Demonstrated genotoxic carcinogen” (Claim 2). - What is broadly accepted is narrower: - Oestradiol has “genotoxic potential” (JECFA 1999). - Its carcinogenicity probably involves receptor-mediated proliferation and possibly genotoxic metabolites (NTP, 15th Report on Carcinogens, 2021). - Who disagreed: - Non-EU assessors and the EU’s own veterinary medicines committee (CVMP) did not accept that genotoxicity is relevant at residue doses. - The 2008 WTO panel’s experts found the EU’s evidence did not show genotoxicity had been “demonstrated”. The Appellate Body reversed that panel on procedural grounds but made no scientific finding. - “Probably justifies” the ban (Claim 1). The EU reaffirmed its position (SCVPH 2000 and 2002, Directive 2003/74/EC, EFSA 2007). No other major assessor has moved towards it. JECFA has not re-evaluated oestradiol since its 52nd meeting. The US (2022, 2026 approvals) and Canada (2022) maintain that residues are safe. - Sanctions “of the order of GBP 100 million (EUR 160 million)” (Claim 9). - The authorised level was US$116.8m plus C$11.3m a year: about EUR 117m at 1999 rates, or EUR 139m (about GBP 86m) at 2001 rates (my calculation). - That figure is a ceiling on the value of trade subject to punitive duties, not a measured loss. - The duties were cut to US$38m in 2009 and ended in 2011. Canada’s were suspended the same year. - Costs of a DES ban “probably groundless”; the 1974–79 “breathing space” produced more promoters (Claim 7). I found no ex post economic evaluation. The chapter’s own Table 14.1 shows only one new approval (MGA, 1977) inside the 1974–79 window. - The ban may have raised consumer risk through illegal use, including DES (Claim 6). - The EU monitoring premise was already out of date: Directive 96/23/EC required residue monitoring. - Monitoring since then shows: - low official non-compliance; - rare stilbene findings EU-wide (three in 2013, including DES in cattle; one DES finding in 2014); - continued illegal use, which researchers say official figures understate. - Whether the ban raised net consumer risk remains unknown.

What the chapter could not see - Management instead of resolution. - The dispute was never adjudicated to a final result. The 2008 Appellate Body could not complete its analysis. - The trade conflict was settled by market access, not science: - a 2009 memorandum of understanding (MoU) with the US; - a 2011 MoU with Canada; - a 2013 revision; - a 2019 agreement that gave the US 35,000 of the EU’s 45,000 tonnes of hormone-free beef quota. - Cost of the settlement: - Other exporters lost access. - The EU kept the ban. - The US kept its WTO authorisation to retaliate. - Reassessment stopped. The EU’s last reassessment is EFSA 2007. JECFA’s last evaluation of the natural hormones is from 1999/2000. Neither side has produced the “rigorous and transparent” risk–benefit weighing the chapter called for. - Water regulation. Directive (EU) 2026/805 made natural oestradiol and oestrone EU priority substances in surface water. It targets oestrogens from all sources and is not specific to livestock.

Weight for the lens. The chapter’s durable lessons are about how evidence is framed and produced: - the scope that assessors are given; - low-baseline subgroups; - detection limits standing in for safety; - remit blind spots; - the evidence stalemate that follows a precautionary decision.

All five are reinforced by later events, several of them from primary WTO and regulatory records. The chapter’s direct scientific verdict (a “demonstrated” non-threshold carcinogen that “probably justifies” the ban) should be cited only as one side of a live dispute. Its quantitative and economic claims should not be cited without correction.

2013 update. - Annex 3 of Late lessons II (EEA, 2013, pp. 736–737) does contain an “update” for this case. It is not a scientific update. It is “a personal summary” of the 2008 Appellate Body report by Theofanis Christoforou, the European Commission’s Principal Legal Adviser, who was a party representative. It calls the ruling “a considerable victory to the Community”. - Annex 2, Table A2.1 (p. 702) recasts the case: - The first early warning becomes “1972/1973 oestrogen effects on wildlife”. - “Years of substantial inaction” becomes “16+”. - Human health risk is left out entirely. - Both should be read as advocacy by interested parties, not as independent hindsight.

Claim 1. More recent research “probably justifies” the EU continuing its ban on precautionary grounds (p. 154)#

Original claim. - The original ban was “in reality, a political risk assessment”, because neither the Lamming Committee nor JECFA was asked to characterise uncertainty. - Nevertheless, “more recent scientific research, however, probably justifies the application by the EU of the precautionary principle to continue the ban” (p. 154). - The research meant is chiefly the claim that oestradiol is a genotoxic carcinogen (p. 153) and the SCVPH 1999 opinion that “no threshold levels can be defined” (Table 14.2, p. 154).

What happened since. - EU reaffirmation. - The SCVPH confirmed its 1999 opinion in May 2000 and again on 10 April 2002. It drew on 17 EU-commissioned studies (1998–2001). - Directive 2003/74/EC (22 September 2003) did three things: - It made the oestradiol ban “permanent”. - It continued the ban on the other five hormones “provisionally” under Article 7 of the General Food Law (Regulation 178/2002). - It recorded the SCVPH findings: “no threshold levels and, therefore, no acceptable daily intake (ADI) can be established” and oestradiol “has to be considered as a complete carcinogen” (recitals 5–6, 10). - The Directive also records dissent (recital 8), from the UK Veterinary Products Committee (October 1999), JECFA (February 2000) and the EU’s own CVMP (December 1999). The CVMP “noted in particular that oestradiol 17ß has a carcinogenic effect only after prolonged exposure and at levels which are considerably higher than those needed for a physiological (oestrogenic) response”. - EFSA 2007. EFSA’s CONTAM Panel reviewed evidence published after 2002. Its conclusions: - Epidemiology gives “convincing evidence for an association between the amount of red meat consumed and certain forms of hormone-dependent cancers. Whether hormone residues in meat contribute to this risk is currently unknown.” - The new data “do not call for a revision” of the SCVPH assessments (quoted in CRS, 2017, p. 8). - EFSA’s press release adds that “there is a lack of data on the type and amount of GPH residues in meat on which to make a quantitative exposure assessment”. - I found no later EFSA opinion on these residues (Europe PMC and Crossref searches of EFSA Journal, 2008–2026). - JECFA and Codex. - JECFA’s database lists no evaluation of oestradiol after its 52nd meeting (held 1999, report 2000). That meeting set an ADI of 0–50 ng/kg body weight and left MRLs “not specified”. - The WTO Appellate Body records that “Codex has adopted an international standard for oestradiol-17β, based on evaluations carried out by JECFA” (AB 2008, para. 536). - Other national regulators. - The US position is that “there is a clear world-wide scientific consensus supporting the safety of these approved and licensed hormones when used according to good veterinary practice”. It cites the UK VPC (1999, 2006) and Australia’s APVMA (2003) reviews (CRS, 2017, pp. 9–10). - The FDA continued approving new hormone-implant indications through 2026, including SYNOVEX ONE Grower (20 May 2026). - Health Canada (page modified 12 December 2022): “Scientific research to date has not demonstrated that food products from animals treated with these hormonal growth promoters pose a threat to human health.” - The WTO test (2005–2008). - In US/Canada – Continued Suspension (DS320/321), the 2008 panel relied on its experts. It found that the SCVPH evidence “did not support the conclusion that the genotoxicity of oestradiol-17β has been demonstrated and that residues … lead to increased risk of cancer” (AB 2008, para. 510). - The Appellate Body reversed that finding. Its grounds were procedural: - The panel had consulted two experts whose “institutional affiliation with JECFA … was likely to affect or give rise to justifiable doubts as to their independence or impartiality” (para. 481). - The panel had acted as risk assessor instead of reviewer (paras. 590–591). - It “was unable to complete the analysis and therefore made no findings” on whether the bans were consistent with the SPS Agreement (WTO case summary).

Verdict: contested. - “Probably justifies” remains the EU’s position, formally restated in 2003 and 2007. - It is not the position of JECFA/Codex, the US, Canada or the other national reviewers cited in the dispute. - No new decisive evidence has emerged on either side since 2007. - The WTO process tested the question twice and resolved it neither way.

Weight. Treat “probably justifies” as a judgement about which level of uncertainty is acceptable, and whose uncertainty counts. It is not an evidential finding. In the EU’s own formulation at the WTO, its level of protection is “no (avoidable) risk” from “genotoxic chemical substances that are intended to be added deliberately to food” (AB 2008, para. 536). With that premise the ban follows even if the risk is tiny. Without it, the ban does not. That supports the chapter’s framing lesson (the question asked shapes the answer) more than its substantive verdict.

Claim 2. “Recent research has demonstrated that oestradiol-17β (although a natural hormone), is a genotoxic carcinogen” (p. 153)#

Original claim. Stated as settled fact (p. 153). The overall conclusions repeat it as the basis for the ban (p. 154).

What happened since. - What JECFA actually concluded (52nd meeting, WHO Food Additives Series 43, 2000). - Oestradiol “did not cause gene mutations in vitro”, but there was “more consistent evidence for the induction of micronuclei in vitro, aneuploidy in vitro, cell transformation in vitro, oxidative damage to DNA in vivo, and DNA single-strand breakage in vivo”. - “The Committee concluded that estradiol has genotoxic potential.” - But because tumours in long-term studies occurred “only in hormone-dependent tissues”, “the carcinogenicity of estradiol is most probably a result of its interaction with hormonal receptors”. - It set a threshold-based ADI of 0–50 ng/kg bw. The ADI was derived from a NOEL of 0.3 mg/day in postmenopausal women, with a factor of 10 for individual variation and “an additional factor of 10 … to protect sensitive populations”. - Mechanistic research since 2000 has built the case that oestrogens act through two pathways: - receptor-driven proliferation; - catechol-oestrogen quinone metabolites that form depurinating DNA adducts (Liehr, 2000; Yue et al., 2003; Santen et al., 2009). - Mammary tumours in oestrogen-receptor-α knockout mice are cited as evidence of a receptor-independent genotoxic route (Yue et al., 2003). This strengthens the “genotoxic potential” side. - Weighing the pathways. - The US National Toxicology Program (15th Report on Carcinogens, 2021) lists steroidal oestrogens as “known to be human carcinogens” (first listed 2002). It states: “The evidence is strong that estrogen carcinogenesis is mediated through activation of the estrogen receptor”, with “possibly direct and indirect genotoxic effects”. The “relative importance of each mechanism is likely a function of the specific estrogen and of the exposed tissue”. - A 2004 review of steroid genotoxicity concluded that natural oestrogens “are generally negative in the ICH core battery of tests”. It also held that their genotoxic potential “may play no role under normal physiological and therapeutic conditions” (Joosten et al., 2004). The first author’s listed affiliation is the toxicology department of Organon, a maker of hormonal medicines (PubMed record). - In the dispute. - Canada argued that the SCVPH’s evidence showed genotoxicity “in vitro, which does not indicate that it is genotoxic in vivo” (Canada’s submission as summarised in AB 2008, para. 170). - The 2008 panel sided with this view on the advice of its experts. The Appellate Body set the finding aside for procedural reasons, without deciding the science (see Claim 1). - Standing of the dissenters. The dissent recorded in Directive 2003/74 (recital 8) came from the EU’s own CVMP. So the disagreement was not simply the EU against the US.

Verdict: contested. The chapter overstated a live dispute as settled. - “Genotoxic potential” is now broadly accepted, and mechanistic work since 2001 has strengthened it. - “Demonstrated genotoxic carcinogen”, in the no-threshold, residue-relevant sense the chapter needs, is not the mainstream assessment. It is the EU’s minority position, which the WTO Appellate Body later said a member may legitimately rely on if it comes “from qualified and respected sources” (AB 2008, para. 591).

Weight. Do not cite p. 153 as fact. The accurate formulation: - Oestradiol has genotoxic potential, and genotoxic metabolites may contribute to oestrogen carcinogenesis. - Whether this matters at residue-level oral exposures is disputed. - The EU treats it as a non-threshold hazard; JECFA, Codex, the US, Canada and the CVMP do not.

The episode is a good example of an assessment moving from “hazard shown in some assays” to “demonstrated carcinogen” in a policy text. The 2001 editors went further again when they reused it (digest, “Editorial amplification”).

Claim 3. Prepubertal boys are the at-risk group; their endogenous oestrogen is measured near detection limits, undermining the FDA’s 1% criterion (pp. 150, 152–153)#

Original claim. - The FDA’s safety argument compared residues with oestrogen levels in pill users and pregnant women. It ignored young children “with low natural levels of oestrogens” (p. 150). - Prepubertal boys’ levels “were close to the limits of detection”. So any exogenous oestrogen “represents a relatively high percentage of the total body oestrogen”. That is “particularly relevant” to the FDA criterion that intake “should constitute less than 1% of the individual’s daily endogenous production” (pp. 152–153).

What happened since. - Assays. - JECFA’s 1999 table put prepubertal male serum oestradiol at “< 10” pg/ml and daily production at “< 0.014” mg/day (FAS 43, Table 1). - Aksglaede et al. (2006), reviewing newer assays, concluded: - “circulating levels of estradiol in prepubertal children are lower than originally claimed”; - children “may respond … even at serum levels below the current detection limits”; - “no threshold has been established”; - the FDA’s 1999 daily production estimates for children, “still used in risk assessments are highly overestimated and should be revised”. - Courant et al. (2010), using GC-MS/MS, found prepubertal boys’ oestradiol “undetectable or extremely low (median < 3.7 pmol/liter)”. That is below about 1 pg/ml, an order of magnitude below the JECFA table’s upper bound. - A 2020 LC-MS/MS method (detection limit 4 pmol/L) could finally tell boys’ levels from girls’ in childhood (Frederiksen et al., 2020). - These papers come largely from one Copenhagen group (Skakkebaek, Juul and Andersson) and from the French laboratory that serves as EU reference laboratory (Le Bizec). The measurement findings are not contested. How far they bear on risk is contested. - EU uptake. Directive 2003/74/EC (recital 5): “of the various susceptible risk groups, prepubertal children constitute the group of greatest concern”. EFSA 2007 kept prepubertal children as the key susceptible group (CRS, 2017, p. 8). - WTO. - Before the 2008 panel, the EU argued that more sensitive methods “had identified lower endogenous levels of oestradiol in pre-pubertal children than previously assumed by the detection method referred to in JECFA’s risk assessments” (AB 2008, para. 722). - One panel expert (Sippell) called the ultrasensitive methods a “quantum leap in [oestrogen] assay methodology” (para. 724). - The panel dismissed the point. The Appellate Body held that it had applied “an excessively high threshold”. “It suffices that new scientific developments call into question whether the body of scientific evidence still permits of a sufficiently objective assessment of risk” (para. 725). - US regulation. - Until 2022, 21 CFR 556.240 allowed oestradiol increments above natural levels of 120 ppt (muscle), 240 (liver), 360 (kidney) and 480 (fat). - On 29 March 2022 the FDA raised these to 0.2, 0.6, 1.2 and 1.2 ppb “to reflect the use of revised food consumption values” (87 FR 17947). The new values are 300 g muscle, 100 g liver, 50 g kidney and 50 g fat (FOI summary, NADA 141-348, July 2022). - Multiplying out, both the old and new increments allow about 60 ng of extra oestradiol per day for each tissue (for example 120 ppt × 500 g = 200 ppt × 300 g = 60 ng) (my calculation; the 500 g figure is the FDA’s former standard muscle-consumption value, which I did not re-verify). - So the underlying safe-intake figure was carried over unchanged, and with it, apparently, the endogenous-production baseline that Aksglaede et al. said should be revised. I could not retrieve the October 2021 FOI summary that explains the change. - Balance. - JECFA’s ADI approach does not rest on the 1% endogenous criterion. It uses a NOEL in women with a tenfold factor for sensitive groups. - JECFA noted that oral oestradiol has low bioavailability (5% for fine-particle oestradiol in women). - JECFA estimated excess intake of total oestrogens from treated beef at 30–50 ng per person per day. - So lower endogenous levels increase the relative share of dietary oestrogen. They do not by themselves show an effect at those intakes. No study I found tests effects of residue-level dietary oestradiol in children.

Verdict: strengthened on the measurement point. Later assays confirmed that children’s levels were overestimated and still sit at the edge of detection. - The EU adopted the point. - The WTO Appellate Body said a panel must take it seriously. - The US and JECFA frameworks were not revised on this basis. - Whether it translates into risk is unresolved.

Weight. This is one of the chapter’s most robust lessons (digest insight 3). Safety margins expressed relative to a baseline are only as good as the measurement of the lowest baseline in the population, and those numbers can persist for decades after being shown wrong. It also supports insight 4 (detection limits standing in for safety). Here the detection limit applied to the reference population rather than to the residue.

Claim 4. “No good evidence” the EU ban protected health; but where the hormones are used, “reliable evidence has not been accumulated on their safety” (p. 153)#

Original claim. Both halves of the stalemate (p. 153). The chapter adds that high oestrogen exposure is increasingly linked to breast and prostate cancer, and that geographical and migrant patterns point to environmental causes (p. 153).

What happened since. - The one direct human study. - Swan et al. (2007) studied 387 partners of pregnant women in five US cities (1999–2005). In sons of women who reported more than seven beef meals a week in pregnancy, sperm concentration was 24.3% lower. The proportion below 20 million/ml was three times higher (17.7% v. 5.7%). - Other meat, and the men’s own meat intake, showed no association. - The authors suggested “anabolic steroids and other xenobiotics in beef” as a possible cause. - Limitations: - mothers self-reported their pregnancy diet decades later; - exposure to hormone-treated beef was not measured; - pesticides and other contaminants are alternative explanations. - I found no replication in Europe PMC searches of 2008–2026 literature. This is an absence in my searches, not proof that none exists. - Regulatory assessments. - EFSA (2007): the contribution of residues to hormone-dependent cancers is “currently unknown” (Claim 1). - A US review by public-health researchers concluded that, “given the lack of chronic bioassays of oral toxicity of the seven hormone compounds in the public literature and the limitations of existing residue surveillance programs, it is not currently possible to provide a quantitative characterization of risks” (Nachman and Smith, 2015). - Surveillance. FDA FOI summaries for recent implant approvals state: “Tolerances for trenbolone and estradiol are not required; therefore, an official analytical method for monitoring their residues in cattle is not required” (NADA 141-348, approved 20 May 2026). The system produces no routine residue data on the natural hormone and one synthetic. - Red meat and cancer. - The association EFSA pointed to is real. IARC classified processed meat as carcinogenic (Group 1) and red meat as probably carcinogenic (Group 2A) in 2015 (Bouvard et al., 2015). - IARC’s evaluation attributed the hazard chiefly to compounds formed in processing and high-temperature cooking. It did not implicate hormone residues (as summarised by Domingo and Nadal, 2016). - Puerto Rico. One of the chapter’s drivers of public concern was the Puerto Rico premature thelarche outbreak (p. 151). I did not re-check later work on its cause.

Verdict: held up. Twenty-five years on, the stalemate the chapter described persists: - No epidemiology shows harm from, or protection against, residues at authorised use. - The jurisdiction using the hormones does not routinely generate residue data that could settle it.

Weight. Strong support for digest insight 10: after a precautionary decision, neither side funds the decisive studies. Two refinements from hindsight: - The EU did fund 17 studies (1998–2001). They addressed hazard and control, not population outcomes. - The approving jurisdiction structurally removed the need for residue data by declaring tolerances unnecessary.

The absence of evidence is thus produced by the regulatory design on both sides, not just by neglect.

Claim 5. The environmental impact of oestrogenic growth promoters “remains to be established”; 1970s wildlife warnings were ignored because agencies assumed dilution and rapid degradation (p. 152)#

Original claim. - Wildlife evidence from the 1970s was “ignored until the late 1980s”. - The chapter attributes this to “the lack of interest of drug agencies in the environment”. It adds that it was “widely held” that excreted growth promoters would be “very diluted” and “quickly degrade” (p. 152). - “The extent of the environmental impact … remains to be established” (p. 152), and no “conclusive evidence” of significant risk had been produced (p. 153).

What happened since. - Field evidence. Three studies published in 2004–2006, two in Nebraska and one by US EPA scientists, documented hormonal activity downstream of feedlots: - Orlando et al. (2004): wild fathead minnows below a feedlot showed male “demasculinization” and female “defeminization”, and the effluent was potently androgenic. - Soto et al. (2004): androgenic and oestrogenic activity in water receiving feedlot effluent. - Durhan et al. (2006): the trenbolone metabolites 17α- and 17β-trenbolone in a feedlot discharge. - Fate: the degradation assumption fails. - Qu et al. (2013, Science) showed that trenbolone metabolites’ rapid photodegradation “is reversible under conditions representative of those in surface waters”. This produces “diurnal cycling and substantial regeneration”. The authors concluded that “regulatory risk assessment paradigms must account for transformation products”. - Modelling then showed that reversion combined with stream–sediment exchange “increase[s] environmental persistence” (Ward et al., 2015). The mechanism was confirmed by Baltrusaitis et al. (2016). - Runoff studies show hormones and metabolites leaving manure-amended fields (Jones et al., 2014; Biswas et al., 2017). - Effects. - A critical review with regulatory and industry co-authors concluded that nanogram-per-litre 17β-trenbolone “can cause changes in endocrine function in the short term, and adverse apical effects in longer exposures”. It can also “greatly skew sex ratios” (Ankley et al., 2018). - The same review says “additional research is warranted to address uncertainties as to the degree/breadth of environmental exposures and potential population-level effects”. - Behavioural effects in fish were reported at field-realistic concentrations (Bertram et al., 2018). - Regulatory responses. - EFSA (2007) noted “associations between hormone-treated beef cattle production and adverse effects in wild fish species” (press release). - US. - FDA environmental assessments for recent trenbolone/oestradiol implants are prepared by the sponsor and approved by the FDA. They model watershed concentrations over 30 years and compare them with fish-reproduction no-effect levels. - The 2024 assessment gives 90th-percentile risk quotients of 0.04 (oestradiol) and 0.32 (trenbolone), concluding that “significant environmental effects are highly unlikely” (EA for SYNOVEX ONE Grower in dry-lot cattle, prepared for the May 2026 approval; the date is not shown in my text extract). - A text search of the three assessments I retrieved found no mention of product-to-parent reversion. - US EPA’s drinking-water Contaminant Candidate List 4 (final 2016) included oestradiol, oestrone, oestriol, equilin and other steroid oestrogens. CCL 5 (November 2022) kept only 17α-ethinylestradiol. Trenbolone appears on neither. - EU. Directive (EU) 2026/805 (30 March 2026) lists 17β-oestradiol, oestrone and 17α-ethinylestradiol as priority substances. It sets annual-average environmental quality standards of 0.00018 µg/L for oestradiol in inland surface waters. It covers oestrogens from all sources, chiefly human excretion as well as manure. Trenbolone is not listed. That is consistent with its non-use in the EU. - The 2013 EEA re-telling. Late lessons II now dates this case’s first early warning to “1972/1973 oestrogen effects on wildlife” (Annex 2, Table A2.1, p. 702).

Verdict: strengthened. Two caveats go with it. - The dilution-and-degradation assumption the chapter criticised has been directly undermined for at least one promoter. - Field effects near feedlots are documented. - Caveat 1: the strongest evidence concerns an androgenic synthetic promoter (trenbolone), not the oestrogenic ones the chapter emphasised. - Caveat 2: “the extent of the environmental impact” at population level still “remains to be established” (Ankley et al., 2018). So the chapter’s own hedge also stands.

Weight. Strong support for digest insight 7: remit-bound institutions fill gaps outside their domain with comforting assumptions. It adds a sharper mechanism: - Standard fate assessment treats a transformation as removal. A reversible transformation makes that bookkeeping wrong. - The sponsor-prepared risk quotients show the gap persisting in approval practice as late as 2026.

Claim 6. The EU ban may have increased consumer risk by encouraging illegal use, including of DES, but “in the absence of a substantial regular monitoring programme” this is hard to gauge (p. 151)#

Original claim. - Misuse takes several forms: overdosing, cocktails, wrong injection sites, retained implants, shortened withdrawal, illegal substances. - There were “claims” that the ban “has led to illegal use, not only of the ‘safer’ steroids but also the more toxic ones, such as DES”. The chapter cites Loizzo et al. (1984). - “The ban may have led to an increased risk to the consumer.” - The chapter also notes DES found in US meat imported into Switzerland (p. 151).

What happened since. - Monitoring existed and expanded. - EU residue monitoring was already mandatory under Directive 96/23/EC (Annex I Group A: stilbenes, steroids, resorcylic acid lactones, beta-agonists). The chapter’s premise was out of date for the EU. - The framework has since moved to Regulation (EU) 2017/625 and Implementing Regulation 2022/1646. In 2024 the EU member states, Iceland and Norway reported 493,664 samples (EFSA, 2026). - What it finds. - Official monitoring in 2007 found “<0.2% non-compliance for the use of illegal growth-promoters” in EU cattle (Nebbia et al., 2011). - A Polish summary of EFSA’s EU-wide data for 2011–2014 reports: - 116 steroid (Group A3) non-compliances, mostly nortestosterone and boldenone, with “individual cases” of oestradiol, testosterone, trenbolone, methyltestosterone and stanozolol; - three stilbene cases in 2013 (DES in cattle, hexoestrol in pigs); - DES in pigs, one case in 2014; - none in 2007–2012 (Matraszek-Żuchowska et al., 2017). - Zeranol-group findings rose to 108 in 2014. But zeranol residues can also derive from contamination with the mould toxin zearalenone. A ratio test is used to separate the two, so raw counts overstate abuse (same source). - I could not retrieve the per-group tables of the 2022–2024 EFSA reports; the Wiley host blocked access. - Underestimation. - Researchers developing biomarker screens argue that official figures “may underestimate the real incidence of GP abuse in meat cattle breeding” (Nebbia et al., 2011). - Illegal use “remains a frequent practice” in parts of the EU (Pegolo et al., 2012). This work concerns mainly low-dose corticosteroid and beta-agonist regimes and hormone “cocktails”, which are hard to detect (Riedmaier et al., 2014). - Incidents. In 2002, pharmaceutical waste containing medroxyprogesterone acetate entered pig feed via glucose syrup, and exposed farms were placed under official control (van Leengoed et al., 2002). This was a contamination incident, not a growth-promotion scheme. It showed that the control system could detect a hormonal contaminant in the food chain. - The US argument at the WTO. - The US and Canada argued that there are “no economic incentives to fail to observe good veterinary practices”. - The Appellate Body held that the 2008 panel had wrongly excluded misuse and abuse from the risk assessment (AB 2008, paras. 553–555). This is the chapter’s point, now accepted in WTO law as a relevant risk-assessment consideration.

Verdict: partly held up. - Illegal use did persist after the ban. DES is still occasionally found, and official monitoring probably understates abuse. - The chapter’s claim that monitoring was absent was out of date for the EU. - I found no evidence that the ban increased net consumer risk. - The net effect remains, as the chapter said, unmeasured.

Weight. Digest insight 6 (bans can displace activity into unmonitored channels) stays suggestive. The better-supported lesson is narrower: - Enforcement of a ban depends on detection technology. - Official non-compliance rates reflect what the tests can see, not the true incidence. - Misuse risk is a legitimate part of risk assessment (AB 2008, para. 555).

Claim 7. The predicted consumer cost of a US DES ban (about USD 500m a year) was “probably groundless”; the 1974–79 “breathing space” let industry develop more hormonal promoters (pp. 149–150)#

Original claim. - The FDA’s estimate “involved a number of questionable assumptions” (p. 149). - When the ban came in 1979 “there was little evidence of a sustained increase in the costs of meat production”, because alternatives existed and the cost calculation was wrong (p. 150). - The 1974–79 reinstatement gave “the pharmaceutical industry” time “to develop additional hormonal growth promoters” (p. 150).

What happened since. - No ex post evaluation found. Europe PMC and Crossref searches did not locate an ex post economic evaluation of the 1979 DES withdrawal, and no source I found quantifies meat-price effects. The chapter cites none either. - The chapter’s own evidence. - Table 14.1 (p. 154) shows three of the four listed alternatives approved before the 1972 withdrawal: Synovex-S (1956), Synovex-H (1958) and Ralgro/zeranol (1969). - Only MGA (3 June 1977) falls within the 1974–79 “breathing space”. - The substitution argument is therefore well supported by the table. The “breathing space” argument rests on one product. - The logic has been reinforced ex ante. - Modelling of the removal of all growth-enhancing technologies from US beef estimated effects equivalent to “an 8.2% tax on beef production”, plus higher land, feed and emissions per unit of beef (Capper and Hayes, 2012). The technologies are implants, in-feed hormones, ionophores and beta-agonists. - A Brazilian study reached similar conclusions; two of its co-authors work for Zoetis, which makes implants (Capper et al., 2021). - Neither study tests the DES case. Together they show that the cost of a restriction depends mainly on whether substitutes remain, which is the chapter’s own explanation. - Continued innovation. The FDA kept approving new implant products and indications through 2026, for example extended-release trenbolone/oestradiol implants (NADA 141-348, 2014–2026). Development continued long after the “breathing space”, so it cannot be credited to it specifically.

Verdict: unclear. - The claim is plausible and consistent with the substitution logic. - The quantitative judgement (“probably groundless”) remains unverified. - The “breathing space” claim is only weakly supported by the chapter’s own table.

Weight. Digest insight 5 (cost forecasts are inflated because substitutes exist or appear) should stay moderate and be cited with its condition: - Restricting one agent within a class of substitutes is cheap. - Restricting the whole class is not.

Do not cite the USD 500m figure or the “breathing space” story as established.

Claim 8. US success at the WTO “has encouraged further actions” against EU precautionary health measures; whether health-based precaution is acceptable under WTO rules is uncertain (p. 153)#

Original claim. - US success “has encouraged further actions on other products where the EU has adopted a precautionary approach on health grounds”. “The question is whether … it would prove acceptable to the WTO” (p. 153). - The appeal clarified that a ban can rest on a risk assessment that is not quantitative, takes account of control difficulties, and relies on “qualified and respected sources … even if these are in the minority” (p. 153).

What happened since. - Further challenges followed. - In EC – Biotech (DS291–293), the US, Canada and Argentina challenged the EU’s approvals process. The panel was established in August 2003; its report was adopted in November 2006 without appeal. - It found a general de facto moratorium causing “undue delay” in 24 of 27 product procedures. None of the member-state safeguard bans “were based on a risk assessment”. - It did not accept that evidence was “insufficient” for Article 5.7 purposes (WTO case summary). - The US also requested a panel in 2009 on the EU ban on poultry treated with pathogen-reduction washes (DS389). The panel was established on 19 November 2009, and the WTO record shows nothing after that. - I have not verified that the hormones outcome caused these cases. The timing is consistent with the chapter’s claim. - The Appellate Body clarified the law in the EU’s favour on key points (2008, DS320/321). - A member “may properly base an SPS measure on divergent or minority views, as long as these views are from qualified and respected sources” (para. 591). - A panel reviews whether a risk assessment “is supported by coherent reasoning and respectable scientific evidence”, not whether it “is correct” (para. 590). - A member choosing higher protection need not “frame the scope and methods of its risk assessment … in the same manner as the international body” (para. 685). - The panel’s “critical mass”/”paradigm shift” test for “insufficient” evidence under Article 5.7 was “too inflexible” (paras. 705, 721, 725). - Risks of misuse must be considered (para. 555). - Experts affiliated with the body whose assessment is in dispute compromise a panel’s impartiality (para. 481). - But it “made no findings as to the consistency or inconsistency” of the bans (WTO case summary). - Note: the Congressional Research Service (CRS, 2017, p. 6) says the Appellate Body stated “that the EU’s ban is not incompatible with WTO law”. That overstates the ruling. - Later case law. In Korea – Radionuclides (DS495, 2019), the Appellate Body reversed a panel’s findings against Korea’s import bans on Japanese fishery products. The grounds were that the panel mis-specified Korea’s level of protection and the “similar conditions” test. So an importing member’s precautionary bans survived appeal. - Enforcement machinery. “Currently, the Appellate Body is unable to review appeals given its ongoing vacancies”; the last member’s term ended on 30 November 2020 (WTO). Any new SPS dispute can now be appealed into a void, which makes the acceptability question practically unresolvable in litigation. - How the dispute was actually managed. - US, 2009. In the US–EU MoU (13 May 2009), the EU opened a duty-free 20,000-tonne quota for “High Quality Beef” from untreated cattle. The US cut sanctions from US$116m to US$38m and dropped “carousel” rotation, which would have affected over US$200m of EU exports (Commission MEMO/09/239). - US, 2011. The US terminated the remaining duties early on 27 May 2011, after a ruling by the Federal Circuit court of appeals. The termination was retroactive to entries after 29 July 2007 (76 FR 30987–30989). - Canada, 2011. A Canada–EU MoU was notified on 17 March 2011 (WTO DS48). - 2012–2013. The quota rose to 45,000 tonnes in 2012. A revised MoU was dated 21 October 2013. - 2019. Because the quota was open to all qualifying suppliers, much of it went to Australia, Uruguay and Argentina (CRS, 2017, p. 15). The 2019 US–EU agreement (signed 2 August 2019) allocated “35,000 tonnes of this quota … to the U.S., phased over a 7-year period, with the remaining amount left available for all other exporters”. It was negotiated with “other substantial supplying countries” (IP/19/3012; IP/19/5010). - 2020. The allocation became operational from 1 January 2020 (SPEECH/20/2553). - Later trade deals. The EU said CETA “will not change … the ban on hormone-treated beef” (IP/17/3121, 2017). Of the 2025 US–EU political agreement, the Commission said it “fully respects the EU’s regulatory sovereignty and protects sensitive areas of EU agriculture, such as beef” (QANDA/25/1930, 29 July 2025).

Verdict: partly held up. - Further challenges to EU precautionary measures did follow, and the EU lost the biotech case. - The uncertainty about precaution under WTO rules was partly clarified in the EU’s favour in 2008, but never resolved for this measure. - The dispute ended as a managed truce: market access for untreated beef, a retained ban, and a retained US right to retaliate. It did not end in a legal or scientific judgment.

Weight. For digest insight 8 (overruling or hiding expert advice undermines legitimacy), hindsight adds a mirror image. The WTO found that relying on experts tied to the assessment in dispute also undermines legitimacy (AB 2008, para. 481). The case also yields a lesson the chapter did not draw (see “New lessons”): an unresolved science dispute can be settled by trading market access. The cost falls partly on third parties, in this case other exporters to the EU.

Claim 9. The EU “is currently experiencing sanctions against its exports of the order of GBP 100 million (ca. EUR 160 million) per annum” (p. 153; Table 14.2, p. 154)#

Original claim. Stated in the text (p. 153) and repeated in Table 14.2: “2001 EU still suffers from sanction to its exports of around EUR 160 million per year” (p. 154).

What happened since. - The authorised amounts. - The WTO arbitrators set the US level at US$116.8m (12 July 1999). Concessions were suspended “at a level of US$ 116,8 million (USA) and CN$ 11.3 million (Canada)” from 26 July 1999 (EU DG Trade case page; USTR case page). - At ECB annual average rates, the combined total is: - about EUR 117m at 1999 rates (USD 1.0658 and CAD 1.5840 per euro); - about EUR 139m at 2001 rates (USD 0.8956 and CAD 1.3864), which is about GBP 86m at the 2001 rate of 0.6219 GBP per euro (my calculation). - The chapter’s GBP 100m (EUR 160m at 2001 rates) therefore overstates the authorised level by roughly 15–40%, depending on the year. - What the figure means. The authorised figure is the annual value of EU exports subject to 100% duties. It is not a measured loss. - What was hit. The 1999 US list included beef and pork products, goose pâté, Roquefort, truffles, Dijon mustard, chocolate and other foods. It left out the UK, “because it has indicated support for lifting the ban” (CRS, 2017, p. 11). - Effectiveness. Commentators cited by CRS judged the duties “mostly ineffective since they do not provide any direct benefit to the U.S. beef industry”. On that view it was “U.S. and EU consumers who lose by paying higher prices” (CRS, 2017, p. 11). - Trajectory. - Reduced to US$38m under the 2009 MoU. - US duties terminated in May 2011, retroactive to mid-2007 entries after the Federal Circuit ruling. - Canadian sanctions suspended under the 2011 MoU. - What the sanctions secured. Under the quota regime, EU-reported imports of US beef reached nearly US$260m in 2015 (fresh/chilled), up from US$5–6m a year in 2000–2006 (CRS, 2017, pp. 18–19). - Not found. I found no study that measures the realised loss to EU exporters.

Verdict: partly held up. Substantial sanctions were in force in 2001, so the claim is right in kind. But the chapter’s figure overstates the authorised level and presents a ceiling as a cost. Within a decade the sanctions were first reduced and then removed.

Weight. Cite the primary figures (US$116.8m plus C$11.3m a year from July 1999), not “EUR 160 million”. The trajectory is more instructive than the number. Retaliation mainly penalised third-party importers and consumers. It was ended by market access plus a domestic court ruling, not by either side conceding on the science.

Claim 10. Scientific committees should be required to identify the uncertainties in their assessments; rigorous and transparent mechanisms are needed for weighing risks against benefits (p. 154)#

Original claim. - “It is very important … that scientific committees are requested to identify the uncertainties in their assessments” - “rigorous and transparent mechanisms must be developed for evaluating risks against benefits” (p. 154). - The chapter also notes the absence of “any formal mechanisms for trading risks and benefits for the public” (p. 153).

What happened since. - Uncertainty: adopted in EU and international guidance. - EFSA’s Scientific Committee (2018): “all EFSA scientific assessments must include consideration of uncertainties. Therefore, the application of this Guidance is unconditional for EFSA. Assessments must say what sources of uncertainty have been identified and characterise their overall impact on the assessment conclusion.” - One carve-out echoes the chapter’s critique of JECFA’s “standard approach” (p. 150). In “assessments that follow standardised procedures, it is only necessary to identify non-standard uncertainties”. - WHO/IPCS issued guidance on “evaluating and expressing uncertainty in hazard characterization” (2014; 2nd edition 2018). - EFSA’s 2009 guidance on transparency preceded both. - Risk–benefit: adopted only for health. - EFSA’s 2010 risk–benefit guidance compares health risks with health benefits. It “does not address social, economic and other considerations such as ‘cost-effectiveness’”. - The weighing the chapter had in mind is health risks to consumers against economic benefits that go only to producers (p. 153). That remains outside formal assessment and inside risk management (for example “other legitimate factors” under Regulation 178/2002). - Did it change outcomes here? - The hormones case has not been reassessed under the new uncertainty frameworks. There has been no EFSA opinion since 2007 and no JECFA evaluation of oestradiol since 1999/2000. - The outcome since 2009 has been set by trade agreements. - So the case offers no test of whether explicit uncertainty analysis would have changed the result. - The EU’s 2003 legal response did, however, treat uncertainty explicitly: - It separated a “permanent” ban (oestradiol) from a “provisional” one (the other five) pending “more complete scientific information”. - It invoked the precautionary principle in Article 7 of Regulation 178/2002 (Directive 2003/74/EC, recital 10). - The Appellate Body treated this split as legally meaningful (paras. 703–725).

Verdict: partly held up. - The uncertainty recommendation was adopted in principle, and in EFSA’s case as a binding internal rule. - The risk–benefit recommendation was adopted only in a health-versus-health form. - Neither has been applied to this case, so any effect on outcomes is untested.

Weight. Digest insight 2 (assessors not asked to characterise uncertainty leave decisions looking political) is reinforced as a norm, not as a demonstrated effect. The hindsight adds a caution. Formal uncertainty guidance can exempt exactly the “standardised” assessments the chapter criticised. Risk–benefit machinery can be built so as to exclude the economic benefits that drove the dispute.

Minor factual checks#

Implications for the section’s transferable insights#

Stated in technology-neutral terms and keyed to the digest’s list.

Insight (digest #, pages) Effect of the post-2001 record
1. “Safe” verdicts are conditional on the scope set for assessors (p. 150) Reinforced. WTO law now says a member seeking higher protection need not frame “scope and methods” as the international body did (AB 2008, para. 685). Misuse risk belongs in scope (para. 555). Strong.
2. Assessors not asked to characterise uncertainty leave decisions looking political (p. 154) Reinforced as a norm. EFSA made uncertainty analysis “unconditional” (2018), but the case was never reassessed under it. Moderate.
3. Averages and high-exposure comparators hide sensitive, low-baseline subgroups (pp. 150, 152–153) Strengthened. Successive assays show prepubertal boys’ oestradiol at or below about 1 pg/ml (Courant et al., 2010). The EU codified the point (2003), and the WTO Appellate Body required panels to engage with it (2008). US increments still imply an unchanged ~60 ng/day allowance (my calculation). Strong.
4. Detection-based thresholds make “not detected” read as “safe” (pp. 149–150, 152) Reinforced and extended. Detection limits also governed the reference population’s baseline, not only residues. In US regulation, “no tolerance required” means “no official method required” (FDA FOI, 2026). Strong.
5. Cost forecasts of restriction tend to be inflated; substitutes exist or are developed (p. 150) Unverified for DES. Ex ante models suggest restricting a whole class is costly (8.2% tax-equivalent, Capper and Hayes, 2012). Restricting one member of a class with substitutes is not. Moderate, with that condition.
6. Bans can displace activity into unmonitored channels (p. 151) Partly supported. Illegal use persists and is probably under-detected. Stilbenes, including DES, have been found occasionally (2013–14). No evidence that net consumer risk rose. Suggestive.
7. Remit-bound institutions neglect harms outside their domain (p. 152) Strengthened. Feedlot effluent alters fish reproduction, and trenbolone metabolites re-form in water (Qu et al., 2013). Approval-stage assessments still use risk quotients that do not mention reversion (FDA EAs, 2021–2026). Strong for the mechanism.
8. Overruling or hiding expert advice undermines legitimacy (p. 150) Reinforced, symmetrically. Relying on experts affiliated with the disputed assessment also compromised a WTO panel’s impartiality (AB 2008, para. 481). Moderate.
9. Ban versus conditional permission turns partly on enforceability (p. 151) Supported. The Appellate Body accepted control difficulties as relevant to risk assessment (paras. 553–555). The US and Canada argued the opposite: there is no incentive to misuse. Moderate.
10. After a precautionary decision, nobody funds the decisive studies (p. 153) Strongly reinforced. EFSA 2007 says “unknown”; Nachman and Smith (2015) say risk is not quantifiable. The only direct human study (Swan et al., 2007) is unreplicated in my searches. Strong.
11. Benefits flowing only to producers lower tolerable uncertainty; needs explicit risk–benefit machinery (pp. 153–154) Not taken up. EFSA’s risk–benefit guidance explicitly excludes economic considerations (2010). The dispute was settled by trade quotas instead. Asserted; no test.

New lessons from hindsight, not drawn in the chapter (technology-neutral): - An unresolved evidence dispute can be settled by trading market access, with costs pushed onto third parties. The 2009–2019 arrangements left the science where it was. They converted the conflict into a quota, 78% of which went to the complainant, at the expense of other exporters (IP/19/5010; CRS, 2017). - Adjudication can stall on scientific process rather than substance. The key 2008 ruling turned on expert independence and the standard of review. Then the appellate machinery itself stopped functioning (2019–2020). Formal dispute resolution may never deliver a verdict on the underlying risk. - A regulatory baseline can outlive the evidence that undermines it. The US safe-intake figure appears unchanged 20 years after the children’s production estimates were said to be “highly overestimated” (Aksglaede et al., 2006). A later technical revision (2022) changed the consumption values but not, on the arithmetic, the baseline. - Advocacy re-tellings harden a case. Both the EU and the US sides later presented the record selectively: - The EEA’s 2013 update was written by the EU’s own litigator. - The CRS summary overstated the 2008 ruling. - The original chapter’s authors included an EU adviser (digest). Hindsight on contested cases needs primary legal and regulatory texts, not either side’s summary. - Transformation is not removal. A contaminant that degrades and then re-forms defeats assessments that count breakdown as disappearance (Qu et al., 2013; Ward et al., 2015).

Method and access notes#

Sources#

EU law and EU scientific bodies

  1. Directive 2003/74/EC of the European Parliament and of the Council of 22 September 2003 amending Council Directive 96/22/EC (OJ L 262, 14.10.2003), recitals 5–10. https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:32003L0074 (retrieved via http://publications.europa.eu/resource/celex/32003L0074, 25 Sep 2026)
  2. EFSA CONTAM Panel. Opinion … related to hormone residues in bovine meat and meat products. EFSA Journal 2007;5(7):510 (18 July 2007). https://doi.org/10.2903/j.efsa.2007.510 (not read in full; quoted via source 30 and 3)
  3. EFSA. “EFSA concludes review of new scientific data on potential risks to human health from certain hormone residues in beef”, press release, 18 July 2007. https://www.efsa.europa.eu/en/press/news/070718
  4. EFSA. Report for 2024 on the results from the monitoring of residues of veterinary medicinal products in live animals and animal products. EFSA Supporting Publications 2026:EN-9913 (February 2026; abstract only). https://doi.org/10.2903/sp.efsa.2026.EN-9913
  5. EFSA Scientific Committee. Guidance on Uncertainty Analysis in Scientific Assessments. EFSA Journal 2018;16(1):5123. https://doi.org/10.2903/j.efsa.2018.5123 ; and The principles and methods behind EFSA’s Guidance on Uncertainty Analysis. EFSA Journal 2018;16(1):5122. https://doi.org/10.2903/j.efsa.2018.5122
  6. EFSA Scientific Committee. Guidance on human health risk-benefit assessment of foods. EFSA Journal 2010;8(7):1673. https://doi.org/10.2903/j.efsa.2010.1673 ; and Guidance … on Transparency in the Scientific Aspects of Risk Assessments carried out by EFSA. Part 2: General Principles. EFSA Journal 2009;7(5):1051 (title only). https://doi.org/10.2903/j.efsa.2009.1051
  7. 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, 20.4.2026), Annex I entries 46, 47, 59. http://publications.europa.eu/resource/celex/32026L0805 ; procedure file 2022/0344(COD): https://oeil.europarl.europa.eu/oeil/en/procedure-file?reference=2022/0344(COD)

International assessments and standards

  1. JECFA, 52nd meeting. Estradiol-17β, progesterone and testosterone. WHO Food Additives Series 43 (2000). https://www.inchem.org/documents/jecfa/jecmono/v43jec05.htm
  2. JECFA database entry: estradiol-17beta (evaluations 1987, 2000). https://apps.who.int/food-additives-contaminants-jecfa-database/Home/Chemical/1835 (accessed 25 Sep 2026)
  3. National Toxicology Program. Report on Carcinogens, 15th ed. (2021): Estrogens, Steroidal. https://ntp.niehs.nih.gov/sites/default/files/ntp/roc/content/profiles/estrogenssteroidal.pdf
  4. WHO/IPCS. Guidance document on evaluating and expressing uncertainty in hazard characterization, 2nd ed. (2018). https://www.who.int/publications/i/item/9789241513548
  5. Bouvard, V. et al. (IARC Monograph Working Group). Carcinogenicity of consumption of red and processed meat. Lancet Oncology 2015;16:1599–1600. https://doi.org/10.1016/s1470-2045(15)00444-1

WTO records

  1. WTO. DS26, EC – Hormones: case page. https://www.wto.org/english/tratop_e/dispu_e/cases_e/ds26_e.htm ; one-page summary (2026 edition). https://www.wto.org/english/tratop_e/dispu_e/cases_e/1pagesum_e/ds26sum_e.pdf
  2. WTO. DS48 (Canada) case page. https://www.wto.org/english/tratop_e/dispu_e/cases_e/ds48_e.htm
  3. WTO Appellate Body. US – Continued Suspension of Obligations in the EC – Hormones Dispute, WT/DS320/AB/R, 16 October 2008 (paras. 481–482, 510, 536, 553–555, 590–591, 685, 705, 721–725). https://docs.wto.org/dol2fe/Pages/SS/directdoc.aspx?filename=Q:/WT/DS/320ABR.pdf ; case page https://www.wto.org/english/tratop_e/dispu_e/cases_e/ds320_e.htm ; one-page summary https://www.wto.org/english/tratop_e/dispu_e/cases_e/1pagesum_e/ds320sum_e.pdf
  4. WTO. One-page summary, EC – Approval and Marketing of Biotech Products (DS291–293). https://www.wto.org/english/tratop_e/dispu_e/cases_e/1pagesum_e/ds291sum_e.pdf
  5. WTO. One-page summary, Korea – Radionuclides (DS495). https://www.wto.org/english/tratop_e/dispu_e/cases_e/1pagesum_e/ds495sum_e.pdf
  6. WTO. DS389, EC – Poultry meat and poultry meat products (US). https://www.wto.org/english/tratop_e/dispu_e/cases_e/ds389_e.htm
  7. WTO. Appellate Body page (status). https://www.wto.org/english/tratop_e/dispu_e/appellate_body_e.htm (accessed 25 Sep 2026)
  8. European Commission, DG Trade. WT/DS26 case page. https://policy.trade.ec.europa.eu/enforcement-and-protection/dispute-settlement/wto-dispute-settlement/wto-disputes-cases-involving-eu/wtds26-european-communities-measures-affecting-meat-and-meat-products-hormones_en
  9. USTR. EU – Measures concerning meat and meat products (hormones). https://ustr.gov/issue-areas/enforcement/dispute-settlement-proceedings/european-union-%E2%80%93-measures-concerning-meat-an

Trade settlement records

  1. European Commission. MEMO/09/239, “Memorandum on Beef Hormones dispute signed with the United States”, 13 May 2009. https://ec.europa.eu/commission/presscorner/detail/en/MEMO_09_239
  2. USTR. “Termination of Action and Further Monitoring in Connection With the EC-Beef Hormones Dispute”, 76 FR 30987–30989, 27 May 2011. https://www.govinfo.gov/content/pkg/FR-2011-05-27/pdf/2011-13282.pdf
  3. European Commission. IP/19/3012, “The European Union and the United States reach an agreement on imports of hormone-free beef”, 14 June 2019. https://ec.europa.eu/commission/presscorner/detail/en/ip_19_3012
  4. European Commission. IP/19/5010, “The European Union and the United States sign an agreement on imports of hormone-free beef”, 2 August 2019. https://ec.europa.eu/commission/presscorner/detail/en/ip_19_5010 ; EEAS version: https://www.eeas.europa.eu/node/66164_en
  5. European Commission. SPEECH/20/2553, Commissioner Hogan at CSIS, 16 January 2020. https://ec.europa.eu/commission/presscorner/detail/en/speech_20_2553
  6. European Commission. IP/17/3121, “EU-Canada trade agreement enters into force”, 20 September 2017. https://ec.europa.eu/commission/presscorner/detail/en/ip_17_3121
  7. European Commission. QANDA/25/1930, “EU-US trade deal explained”, 29 July 2025. https://ec.europa.eu/commission/presscorner/detail/en/qanda_25_1930
  8. European Central Bank. Annual average reference rates, USD, CAD and GBP per EUR, 1999–2001. https://data-api.ecb.europa.eu/service/data/EXR/A.USD.EUR.SP00.A (and .CAD., .GBP.)
  9. Congressional Research Service. The U.S.-EU Beef Hormone Dispute, R40449, version 21, updated 9 January 2017 (secondary; used for positions, trade data and quotations of EFSA 2007). https://www.congress.gov/crs_external_products/R/PDF/R40449/R40449.21.pdf

US and Canadian regulators

  1. 21 CFR 556.240, Estradiol and related esters (current text; amended 87 FR 17947, 29 March 2022). https://www.law.cornell.edu/cfr/text/21/556.240 ; earlier text (CFR 2012 edition) https://www.govinfo.gov/content/pkg/CFR-2012-title21-vol6/pdf/CFR-2012-title21-vol6-sec556-240.pdf ; 87 FR 17947 https://www.govinfo.gov/content/pkg/FR-2022-03-29/pdf/2022-06395.pdf
  2. FDA. FOI Summary, supplemental NADA 141-348, SYNOVEX ONE FEEDLOT (approved 18 July 2022). https://animaldrugsatfda.fda.gov/adafda/app/search/public/document/downloadFoi/12669
  3. FDA. FOI Summary, supplemental NADA 141-348, SYNOVEX ONE GROWER (approved 20 May 2026). https://animaldrugsatfda.fda.gov/adafda/app/search/public/document/downloadFoi/18506
  4. FDA. Environmental assessments for SYNOVEX implants: ONE Grower (21 July 2021) https://animaldrugsatfda.fda.gov/adafda/app/search/public/document/downloadEA/3383 ; Primer and Choice in dry lot (10 November 2023) https://animaldrugsatfda.fda.gov/adafda/app/search/public/document/downloadEA/4522 ; ONE Grower in dry lot (prepared for the May 2026 approval, citing the 2024 EA; date not shown in my extract) https://animaldrugsatfda.fda.gov/adafda/app/search/public/document/downloadEA/5570
  5. FDA. “Steroid Hormone Implants Used for Growth in Food-Producing Animals” (content current as of 24 October 2024). https://www.fda.gov/animal-veterinary/product-safety-information/steroid-hormone-implants-used-growth-food-producing-animals
  6. Health Canada. “Hormonal growth promoters” (modified 12 December 2022). https://www.canada.ca/en/health-canada/services/drugs-health-products/veterinary-drugs/factsheets-faq/hormonal-growth-promoters.html
  7. US EPA. Contaminant Candidate List 4 chemical contaminants https://www.epa.gov/ccl/chemical-contaminants-ccl-4 ; CCL 5 chemical contaminants (final 14 November 2022) https://www.epa.gov/ccl/ccl-5-chemical-contaminants

Oestradiol toxicology and children’s hormone levels

  1. Liehr, J. G. Is estradiol a genotoxic mutagenic carcinogen? Endocrine Reviews 2000;21:40–54. https://doi.org/10.1210/edrv.21.1.0386
  2. Yue, W. et al. Genotoxic metabolites of estradiol in breast: potential mechanism of estradiol induced carcinogenesis. J Steroid Biochem Mol Biol 2003;86:477–86. https://doi.org/10.1016/s0960-0760(03)00377-7
  3. Joosten, H. F. et al. Genotoxicity of hormonal steroids. Toxicology Letters 2004;151:113–34. https://doi.org/10.1016/j.toxlet.2004.01.018
  4. Santen, R. et al. Estrogen mediation of breast tumor formation involves estrogen receptor-dependent, as well as independent, genotoxic effects. Ann NY Acad Sci 2009;1155:132–40. https://doi.org/10.1111/j.1749-6632.2008.03685.x
  5. Aksglaede, L. et al. The sensitivity of the child to sex steroids: possible impact of exogenous estrogens. Human Reproduction Update 2006;12:341–9. https://doi.org/10.1093/humupd/dml018
  6. Courant, F. et al. Assessment of circulating sex steroid levels in prepubertal and pubertal boys and girls by a novel ultrasensitive gas chromatography-tandem mass spectrometry method. J Clin Endocrinol Metab 2010;95:82–92. https://doi.org/10.1210/jc.2009-1140
  7. Frederiksen, H. et al. Sex-specific estrogen levels and reference intervals from infancy to late adulthood determined by LC-MS/MS. J Clin Endocrinol Metab 2020;105:754–68. https://doi.org/10.1210/clinem/dgz196

Human epidemiology and exposure

  1. Swan, S. H. et al. Semen quality of fertile US males in relation to their mothers’ beef consumption during pregnancy. Human Reproduction 2007;22:1497–502. https://doi.org/10.1093/humrep/dem068
  2. Nachman, K. E., Smith, T. J. Hormone use in food animal production: assessing potential dietary exposures and breast cancer risk. Current Environmental Health Reports 2015;2:1–14. https://doi.org/10.1007/s40572-014-0042-8
  3. Domingo, J. L., Nadal, M. Carcinogenicity of consumption of red and processed meat: what about environmental contaminants? Environmental Research 2016;145:109–15. https://doi.org/10.1016/j.envres.2015.11.031
  4. Jukes, T. H. Diethylstilbestrol in beef production: what is the risk to consumers? Preventive Medicine 1976;5:438–53 (pre-2001; spelling check only). https://doi.org/10.1016/0091-7435(76)90060-8

Environmental fate and effects

  1. Orlando, E. F. et al. Endocrine-disrupting effects of cattle feedlot effluent on an aquatic sentinel species, the fathead minnow. Environ Health Perspect 2004;112:353–8. https://doi.org/10.1289/ehp.6591
  2. Soto, A. M. et al. Androgenic and estrogenic activity in water bodies receiving cattle feedlot effluent in Eastern Nebraska, USA. Environ Health Perspect 2004;112:346–52. https://doi.org/10.1289/ehp.6590
  3. Durhan, E. J. et al. Identification of metabolites of trenbolone acetate in androgenic runoff from a beef feedlot. Environ Health Perspect 2006;114(Suppl 1):65–8. https://doi.org/10.1289/ehp.8055
  4. Qu, S. et al. Product-to-parent reversion of trenbolone: unrecognized risks for endocrine disruption. Science 2013;342:347–51. https://doi.org/10.1126/science.1243192
  5. Jones, G. D. et al. Surface and subsurface attenuation of trenbolone acetate metabolites and manure-derived constituents in irrigation runoff on agro-ecosystems. Environ Sci Process Impacts 2014;16:2507–16. https://doi.org/10.1039/c4em00385c
  6. Ward, A. S. et al. Coupled reversion and stream-hyporheic exchange processes increase environmental persistence of trenbolone metabolites. Nature Communications 2015;6:7067. https://doi.org/10.1038/ncomms8067
  7. Baltrusaitis, J. et al. Reversible photohydration of trenbolone acetate metabolites. Environ Sci Technol 2016;50:6753–61. https://doi.org/10.1021/acs.est.5b03905
  8. Biswas, S. et al. Effect of rainfall timing and tillage on the transport of steroid hormones in runoff from manure amended row crop fields. J Hazard Mater 2017;324:436–47. https://doi.org/10.1016/j.jhazmat.2016.11.009
  9. Ankley, G. T. et al. A critical review of the environmental occurrence and potential effects in aquatic vertebrates of the potent androgen receptor agonist 17β-trenbolone. Environ Toxicol Chem 2018;37:2064–78. https://doi.org/10.1002/etc.4163
  10. Bertram, M. G. et al. Field-realistic exposure to the androgenic endocrine disruptor 17β-trenbolone alters ecologically important behaviours in female fish across multiple contexts. Environmental Pollution 2018;243:900–11. https://doi.org/10.1016/j.envpol.2018.09.044

Residue control and illegal use

  1. Matraszek-Żuchowska, I. et al. Control of anabolic hormone residues in tissues of slaughter animals in Poland during the period of 2011–2015. J Vet Res 2017;61:69–79 (includes a summary of EU-wide EFSA data, 2011–2014). https://doi.org/10.1515/jvetres-2017-0009
  2. Nebbia, C. et al. Novel strategies for tracing the exposure of meat cattle to illegal growth-promoters. Veterinary Journal 2011;189:34–42. https://doi.org/10.1016/j.tvjl.2010.06.016
  3. Pegolo, S. et al. Transcriptomic markers meet the real world: finding diagnostic signatures of corticosteroid treatment in commercial beef samples. BMC Vet Res 2012;8:205. https://doi.org/10.1186/1746-6148-8-205
  4. Riedmaier, I. et al. Identification of a potential gene expression biomarker signature in bovine liver to detect the abuse of growth promoters. Food Addit Contam A 2014;31:641–9. https://doi.org/10.1080/19440049.2014.886341
  5. van Leengoed, L. et al. [The weakest link: medroxyprogesterone acetate in pig feed]. Tijdschrift voor Diergeneeskunde 2002 (abstract only). PMID 12244854. https://pubmed.ncbi.nlm.nih.gov/12244854/

Economics

  1. Capper, J. L., Hayes, D. J. The environmental and economic impact of removing growth-enhancing technologies from U.S. beef production. J Anim Sci 2012;90:3527–37. https://doi.org/10.2527/jas.2011-4870
  2. Capper, J. L. et al. Modeling the effects of steroid implant use on the environmental and economic sustainability of Brazilian beef production. Transl Anim Sci 2021;5:txab144 (co-authors from Zoetis). https://doi.org/10.1093/tas/txab144

Late lessons follow-up

  1. EEA. Late lessons from early warnings: science, precaution, innovation, EEA Report 1/2013, Annexes: Annex 2, Table A2.1 (p. 702); Annex 3, “Hormones as growth promoters” by T. Christoforou (pp. 736–737). https://www.eea.europa.eu/en/analysis/publications/late-lessons-2/late-lessons-chapters/late-lessons-ii-annex-1