Late Lessons, Jensen Huang and AI
Markdown version

LL1-14 — Ch14 Hormones as growth promoters: the precautionary principle or a political risk assessment?#

Late lessons from early warnings: the precautionary principle 1896–2000 (EEA Environmental Issue Report No 22, 2001). Report pages 149–156; PDF pages 149–156 (printed and PDF page numbers coincide for this chapter). The chapter text runs pp. 149–154. Pages 155–156 are references.

Reading record. I read the whole text extract in order, through the last marker (PDF 156 / report p. 156). I rendered pp. 149–156 from the PDF and checked them visually against the extract. The extraction is faithful: two-column text, bulleted lists, and both tables on p. 154 read correctly. Spelling slips in the printed original include “trenbelone” (trenbolone), “bortholinic gland” (Bartholin’s gland), “Ralgo” (the zeranol implant Ralgro), “Peakal”/”Peakall”, “Bulkey” (Bulkley) and “Jakes” (the author is T. H. Jukes; see the bias check).

Sources outside the chapter. From the same PDF I used the author biographies (p. 195), the editors’ framing in Chapter 1 (pp. 11–12) and the editors’ use of this chapter in Chapter 16 (pp. 173–181). For “what happened next” and to check facts, I went to primary documents where I could: - WTO Appellate Body report WT/DS26/AB/R, WT/DS48/AB/R (16 January 1998), and the WTO one-page case summary. - WTO Article 22.6 arbitration decisions WT/DS26/ARB and WT/DS48/ARB (12 July 1999), and the WTO DS320 case page. - Directive 2003/74/EC (OJ L 262, 14.10.2003). - Hess & Clark v. FDA, 495 F.2d 975 (D.C. Cir. 1974), and Rhone-Poulenc v. FDA, 636 F.2d 750 (D.C. Cir. 1980). - JECFA 52nd meeting monograph on oestradiol, progesterone and testosterone (WHO Food Additives Series 43, 2000). - PubMed records of cited and later papers.

One secondary source fills gaps: the Congressional Research Service report R40449 (Johnson, January 2017, via everycrsreport.com). I used it for the 1996 lifting of US tariffs, the 2009 MoU terms, and its quotation from the 2007 EFSA opinion, which I could not reach directly. Each outside claim is flagged below.

Audit re-checks (see Audit log): the AB report text (paras 2–3, 124–125, 135 fn 107, 186–187, 194–196, 200–201, 206–208, 245), the Fara et al. 1979 and Aksglaede et al. 2006 abstracts (Europe PMC), the Jukes 1976 bibliographic record (Europe PMC), a profile of T. H. Jukes (R. Conniff, Alicia Patterson Foundation), and the report’s own pp. 11–12, 84–90, 173–181, 195 and index (p. 209) were re-read against these notes. The Hess & Clark and Rhone-Poulenc quotations could be re-verified only in part (search-result summaries), so treat their exact wording as provisional.

Authors and standpoint#

Authors. Jim W. Bridges and Olga Bridges, both at the University of Surrey (p. 149; biographies p. 195).

The authors were participants, not only analysts. The WTO record shows the Panel cited a paper by “J. Bridges and O. Bridges”, Hazards of Growth Promoting Agents and Strategies of Risk Assessment, presented at the 1995 EC Scientific Conference (AB report, footnote 107, in the section on MGA evidence at para. 135) [ext]. The Panel cited it alongside its point that JECFA, Codex and the EC had assessed each of the other five hormones individually, that is, in reasoning that went against the EC on MGA. That is the same conference the chapter describes on p. 152. Jim Bridges also sat on SCAN, which the chapter calls “the lead committee for issues concerning growth promoters” (p. 150), from 1991 to 1997. That was after the Lamming episode the chapter criticises. So the lead author advised the EU side in the trade dispute that the chapter analyses. The chapter text never says this. It is visible only in the biography at the back of the report.

Stance. The chapter’s stance is more even-handed than the lead author’s role might suggest. - Critical of the EU. It calls the original EU ban “in reality, a political risk assessment” (p. 154). It says the Commission decided “to overrule the views of both committees” (p. 151). It reports that the Lamming Committee “was disbanded” in 1987 and that “the EU did not publish its interim conclusions” (p. 150). The text does not say who disbanded it; Table 14.2 says “disbanded by EU”. It concedes there is “no good evidence” that the ban protected public health (p. 153). It also reports claims that the ban may have raised risk by driving illegal use (p. 151). - Critical of the US and JECFA. It criticises the narrow scope of risk assessments (p. 150), the FDA’s reasoning and cost estimates (pp. 149–150), the farming lobby (p. 150), and the neglect of the environment by drug agencies (p. 152). - Final judgement. Precaution is appropriate where the science base is inadequate to confirm safety, and more recent research “probably justifies” keeping the ban (p. 154).

Overall this is a pro-precaution chapter by a regulatory toxicologist. It uses the case to argue for better uncertainty characterisation and risk–benefit machinery, not simply to celebrate the ban.

Panels and commentaries. None. Panels are a feature of the 2013 volume. The chapter does have two tables with different provenance: - Table 14.1 (FDA approval dates) is sourced to “Henderson et al., 1988”. (The audit checked the positions of the two source lines on the PDF page: “Henderson” sits beside Table 14.1 and “EEA” beside Table 14.2.) - Table 14.2 (“early warnings and actions”) is sourced to “EEA”. It was presumably compiled by the editors, not the authors. Several of its entries differ from the chapter text; see the bias check.

How the editors frame and use the chapter elsewhere in the report. These are the editors’ views, not the chapter authors’. - p. 11: one of three chapters on issues “of some conflict between North America and Europe”. - p. 12, Table 1.1: the US DES growth-promoter ban of 1972–79 is an example of US “precautionary prevention”, “nearly 10 years before the EU ban in 1987”. - Chapter 16 draws on the chapter under these lessons: - “blind spots” (children as the at-risk group, p. 173) - interdisciplinary obstacles (wildlife concerns attracting little regulatory attention, p. 174) - real-world conditions (JECFA’s narrow scope; misuse; Puerto Rico and Italy; sensitive young boys, p. 175) - scrutinising pros and cons (“narrow and partial assessments”, p. 176) - regulatory independence (the 1974 US reinstatement “followed strong pressure from the farming lobby”, p. 179) - institutional obstacles (different national decisions on the same information, p. 181)

Some of these editorial uses go beyond what the chapter says; see the bias check.

Section-by-section notes#

My comments are in square brackets. Corrections checked against primary sources are collected in the bias check.

14.1 Introduction (pp. 149–151)#

Biology (p. 149). - Oestrogens act at very low serum levels (0.1–1 pg/ml). - Their effect on male reproductive development has been known since Zuckerman (1940). The mammalian default is female development when hormone signalling is absent. Both sexes need oestrogens for fertility. - Above physiological levels, oestradiol-17β given during foetal or perinatal life in four rodent and rabbit species can impair male pituitary–hypothalamic and later testicular function (four studies, 1975–83). - [This sets up the chapter’s later emphasis on early-life windows of sensitivity.]

Growth promotion and DES (p. 149). - Oestrogen–androgen combinations maximise growth mainly by improving nitrogen retention. - After 1945 such hormones were adopted to raise meat output. DES, a cheap, well-absorbed synthetic analogue of oestradiol, became the favoured promoter for cattle, sheep and poultry. It was given by implant or in feed. - Concern arose in the early 1970s when DES was “confirmed to be a human carcinogen”. Scientists nonetheless reached a consensus that residue risk was insignificant, because residues were below analytical detection and far below therapeutic doses.

DES: EU versus US (pp. 149–150). - The chapter says DES growth-promoter use “continued in some Member States … longer than in the United States”. It dates the EU-wide ban to 1987, “because of uncertainty as to whether there was a definable ‘no effect’ dose” for tumour induction, and notes that some Member States had banned it earlier. [The 1987 date is probably wrong: see the bias check.] - In the US, the chapter says DES was banned in 1972 as a carcinogen under the 1958 Delaney Clause. That clause was hard to apply because most foods contain trace carcinogens, but public opinion kept it in law. Regulators “had been forced to refer to ‘minimum detectable levels’” to justify their inaction under the clause in the case of DES. - The FDA put the consumer cost of a ban at about USD 500 million a year, on questionable assumptions. The residue cancer risk was estimated at one case in 133 years (Jukes 1976, printed as “Jakes”). - In 1974 DES use was “temporarily reinstated” because of “procedural deficiencies in the original bill”. The farm lobby warned of serious economic harm from “a further ban” despite existing alternatives (Table 14.1). The resulting “breathing space” let the pharmaceutical industry develop more hormonal promoters. [The chapter does not say the lobby caused the reinstatement; the editors do, p. 179. See the bias check.] - A scientific debate on what DES residue level carried no significant risk ran alongside. In 1976 the FDA set 2 ppb as the regulatory (minimum detectable) level. It estimated actual residues at about 0.5 ppb but could not verify this by measurement.

The FDA’s safety argument (p. 150). This is the chapter’s clearest example of a reference-population error. - The FDA cited the safety of oral contraceptive oestrogens and the high natural oestrogen levels of pregnancy as “crucial evidence” that residues were harmless. - The chapter’s objections: - Young children, with low natural oestrogen, were the likely at-risk group (McMartin 1978). - DES is structurally quite different from oestradiol and the pill’s oestrogens.

The 1979 ban and its costs (p. 150). - DES was banned in 1979 because no residue level could be identified below which carcinogenic effects would not occur. - The feared costs were “probably groundless”. There was “little evidence of a sustained increase” in meat production costs. For the US, the chapter says this “could be explained” by available substitutes and, “in part”, by wrong assumptions in the cost calculation. - The FDA kept supporting “other oestrogenic compounds” (oestradiol, trenbolone, zeranol) for their perceived importance to the economic efficiency of meat production. [Trenbolone is a synthetic androgen, not an oestrogen, so the chapter’s label is loose.]

EU expert committees (p. 150). - Lamming Committee. In 1982 this EU working group, drawn from the Scientific Committee for Food and SCAN (SCAN being the lead on growth promoters), reached an interim conclusion that oestradiol and several other hormonal promoters were safe in cattle. - The chapter asserts this was “clearly unpopular” with officials. - Later work did not change the committee’s view. It “was disbanded in 1987 and the EU did not publish its interim conclusions”. Members published independently (Vet Record 1987). - JECFA (WHO/FAO, 1988) agreed with Lamming, using the standard method “still employed by scientific advisory committees today”. Its scope had three limits: 1. risk under authorised use only, although there were already “indications of significant accidental or deliberate misuse” that might be expected to raise residues 2. single promoters, not combinations 3. manufacturers’ data only

[The chapter does not explicitly tie these scope limits to its p. 154 diagnosis (committees not asked to characterise uncertainty), but they are the most concrete evidence the chapter offers for it. The chapter calls JECFA the “Joint Expert Committee on Food” and, on p. 154, the committee “of WHO”; it is the Joint FAO/WHO Expert Committee on Food Additives.]

The EU ban (p. 150). - Shortly after JECFA reported, the EU banned oestradiol and other natural and synthetic steroid hormones for growth promotion. - The ban was first adopted in 1985, annulled by the European Court on procedural grounds after a UK challenge, and re-agreed in 1988. - It covered oestradiol-17β, testosterone, progesterone, zeranol, trenbolone acetate and MGA. - In 1989 it was applied to imports, with exemptions for exporters that already banned the promoters or ran hormone-free programmes. [The chapter says “third world countries”; the correct term is third countries, meaning all non-EU states.] - The chapter says this action “could be regarded as an application of the precautionary principle, although the principle was not formalised at the time”. - [Internal chronology problem. The chapter says the EU ban came “shortly after the publication of JECFA’s conclusions” (1988), but also that it was “first adopted in 1985”, before JECFA had met (32nd meeting, 1987) or reported. So the 1985 decision cannot have overruled JECFA. The p. 154 statement that the action was “not at the time supported by” JECFA fits the 1988 re-adoption only.]

Why the Commission overruled both committees (p. 151). The chapter says three factors “appear to have had a particular influence”, confusingly described as bearing on “DES’s use as a growth promoter”: 1. DES had been shown to cause “vaginal clear-cell adenoma” in young women (Herbst and Bern). [Wording issues: it is adenocarcinoma, and the cause was prenatal pharmaceutical exposure, not residues.] 2. Growing public worry about hormones, including the first-generation pill (breast cancer, thrombosis). 3. Reports of oestrogenic contamination causing anomalies of growth, sexual development and puberty: - Puerto Rico: over 10,000 cases (Perez-Comas 1988). Linked to high serum oestrogens, but the source was not clearly identified. - Italy: similar effects, believed due to accidental food contamination by DES (Fara et al. 1979). - Italian baby food (1980): made with homogenised veal, it contained significant DES, attributed to implants not removed at slaughter.

Risk is multi-factorial (p. 151). Human risk depends on: - the product, dose, site and time to slaughter - lifetime consumption of treated meat - background exposure to environmental oestrogens - individual susceptibility

Misuse (p. 151). Six forms: - excess dose - complex mixtures - wrong injection site - unremoved implants or injection sites, where concentrations are much higher - shortened withdrawal periods - illegal substances

Illegal use and perverse effects (p. 151). - Illegal use in Member States was the subject of the 1989 Pimenta Report (a European Parliament inquiry [ext]; the chapter does not say who produced it and does not list it in the references). The report “found no evidence in the use of oestradiol-17β” [ambiguous wording]. It still endorsed the ban because it “facilitated controls and consumer confidence in meat”. - The chapter reports “claims” that the ban led to illegal use, “not only of the ‘safer’ steroids but also the more toxic ones, such as DES” (citing Loizzo et al. 1984), so it “may have led to an increased risk to the consumer”. The chapter reports these claims without endorsing them. - Without “a substantial regular monitoring programme”, the extent of misuse and any added risk “is hard to gauge”. - DES was reportedly found “last year” in US meat imported into Switzerland. [I have not verified this.]

Reversibility (p. 151). Most accidental-contamination effects were transient and reversible. But long-term effects of exposure before puberty are unknown, and the lack of a demonstrated no-effect threshold adds to the uncertainty.

14.2 Wildlife (pp. 151–152)#

14.3 Uncertainties for human health (pp. 152–153)#

14.4 Has the Commission’s approach proved sound? (p. 153)#

14.5 Overall conclusions (p. 154)#

Table 14.1 — US FDA approval dates for anabolic agents (p. 154; source Henderson et al. 1988)#

Trade name Agent Approved
Synovex-S 200 mg progesterone / 20 mg oestradiol benzoate 20/2/1956
Synovex-H 200 mg testosterone propionate / 20 mg oestradiol benzoate 16/7/1958
“Ralgo” [Ralgro] 36 mg or 12 mg zeranol 5/11/1969
MGA 0.25–0.50 mg/day 3/6/1977

[This supports the claim that alternatives existed before 1972–74. MGA’s approval falls within the post-1974 “breathing space”.]

Table 14.2 — Early warnings and actions (p. 154; source EEA)#

The table lists events from the 1970s to 2001, summarised in my timeline below. It is sourced to “EEA”, so it was presumably compiled by the editors. Several entries go beyond, shift or contradict the text: - 1985: says Lamming was ignored “because the scope of their assessments had not been broad enough”. The text gives no such contemporaneous reason. - 1988: says the ban was “based on uncertainty of their effects on humans”. The text says the action was “principally in response to public concern” (p. 154). - 1989 extension: “EU ban extended to other growth promoters and to imports”. The text says the 1989 extension was to imports only (p. 150). It also repeats “third world countries”. - 1989 Pimenta: “finds illegal use of growth promoters in some Member States”. The text says illegal use was the report’s “subject” and that it found “no evidence in the use of oestradiol-17β” (p. 151). These are not strictly contradictory, but the text never says the report found illegal use. - 1995 conference: credited with leaving uncertainties on immune, endocrine and cancer effects. The text attributes those concerns to the later SCVPH opinions and says only that the conference drew “no definitive conclusions” (p. 152). - 1999 SCVPH: “no threshold levels can be defined for six growth promoters”. The chapter text does not report this; it says only that both SCVPH opinions called for attention to sensitive populations (p. 152). The table entry matches the external record (see timeline), but it is not in the text. - 2000 workshop: “confirms impacts” of veterinary drugs on wildlife. The text says “confirmed concerns”, and p. 153 says there is no conclusive environmental evidence.

References (pp. 155–156)#

28 items, mostly primary papers from 1940–1999, plus EC 1996 proceedings, JECFA TRS 763 (1988), SCVPH 1999 and the 1997 WTO panel reports. The pairing of citations with claims is unreliable in several places; see the bias check.

Case timeline#

Chapter claims carry page numbers. Items marked [ext] come from outside sources and are flagged.

Date Event Source / notes
1940 Oestrogens known to affect male reproductive development Zuckerman 1940 (p. 149). Background knowledge, not a warning about growth promoters.
Post-1945 Oestrogens, alone or with androgens, introduced to raise meat production; DES becomes the favoured promoter p. 149
1956, 1958, 1969 FDA approves Synovex-S, Synovex-H, Ralgro (zeranol) Table 14.1, p. 154. Alternatives to DES exist early.
1962 US “DES proviso”: animal-drug exception to Delaney, conditional on no residue found by approved methods [ext] Hess & Clark 1974. The chapter mentions only the “1958 Delaney Clause” (p. 149).
1970–73 Wildlife and endocrine warnings: DDT–oestradiol 1970; catfish/DES 1972; bird feminisation; later trout, alkylphenols, seals and mink p. 152. “Ignored until the late 1980s”.
1971 New gas–liquid chromatography (GLC) testing by USDA finds small DES residues in beef livers [ext] Hess & Clark 1974
Early 1970s DES “confirmed to be a human carcinogen”; scientific consensus nonetheless that residue risk was “insignificant” p. 149
1972 (4 Aug) FDA withdraws DES feed-premix approvals, calling it “a regulatory, not a public health problem” with “no evidence of a public health hazard” [ext] Hess & Clark (the feed-premix withdrawal was litigated in the companion case Chemetron v. HEW, 495 F.2d 995, decided the same day; exact wording only partly re-verified in the audit). The chapter says banned “on the grounds that it was a carcinogen” (p. 149).
1973 (27 Apr) FDA withdraws DES implant approvals after radiotracer studies show residues at 120 days [ext] Hess & Clark
1974 (24 Jan) Court sets aside FDA orders because no evidentiary hearing was held; DES use resumes [ext] Hess & Clark. The chapter says “procedural deficiencies in the original bill” (p. 150).
1974–79 “Breathing space”: farm lobby cost claims; pharmaceutical industry develops more promoters p. 150
1976 FDA sets 2 ppb as the minimum detectable/regulatory level; estimates residues at about 0.5 ppb, unverified p. 150. The 1980 court says the 2 ppb standard predated 1971 evidence, so the chapter’s date may be wrong.
1977 MGA approved in US Table 14.1
1977–78 Breast enlargement outbreak at a Milan school (first noted November 1977, followed up to end of 1978) p. 151 (Fara 1979). [ext] Fara abstract: no oestrogen contamination detected in tested school meals; an “uncontrolled supply of poultry and beef was suspected”; enlargement “disappeared within 8 months”.
1979 FDA bans DES as an animal drug after a hearing (44 Fed. Reg. 54,851): no demonstrable no-effect level; manufacturers failed to show DES similar to natural oestrogens; benefits not shown to outweigh risks p. 150. [ext] Rhone-Poulenc 1980 affirms.
1980 DES found in Italian veal baby food p. 151
1981 (31 Jul) Directive 81/602/EEC prohibits “substances having a hormonal action” and thyrostatics in farm animals, except therapeutic/zootechnical use and five hormones pending review [ext] AB report paras 2–3 (re-checked). DES was not among the exempted hormones. This conflicts with the chapter’s “1987” EU DES ban (p. 149).
1982 Lamming Committee interim conclusion: oestradiol and other hormonal promoters safe in cattle p. 150
1980s Puerto Rico: over 10,000 cases of premature sexual development; source unidentified p. 151
1985 (31 Dec) Directive 85/649/EEC bans the hormones for growth promotion p. 150. [ext] AB report fn 6 gives the date. Adopted before JECFA met (1987) or reported (1988), although the chapter places the ban “shortly after” JECFA’s conclusions.
1987 Lamming Committee disbanded; conclusions unpublished by EU; members publish in Vet Record p. 150
1987/88 JECFA (32nd meeting; TRS 763, 1988) concludes natural hormones safe under authorised use p. 150
1988 (23 Feb; 7 Mar) ECJ annuls 85/649 on procedure (UK challenge); substantively identical Directive 88/146/EEC adopted, including an import ban on meat from treated animals from third countries p. 150. [ext] AB fn 6 and para 4.
1989 Import prohibition takes effect; US imposes 100% duties on about $93m of EU goods (from 1 January 1989); Pimenta Report; Collins Report pp. 151–153. [ext] CRS for the US tariff date.
1995 EC Scientific Conference on Growth Promotion in Meat Production, Brussels (29 Nov–1 Dec); the Bridgeses present a paper p. 152. [ext] AB fn 107.
1996 US and Canada start WTO proceedings (US consultations 26 Jan; panels established 20 May and 16 Oct); US 1989 tariffs lifted (May, per CRS); Directive 96/22/EC consolidates the ban [ext] WTO; CRS; AB para 5. The chapter says 1997 (p. 153).
1997 (18 Aug) Panel reports: EC violated SPS Arts 3.1, 5.1, 5.5 p. 153. [ext] WTO.
1998 (16 Jan; 13 Feb) Appellate Body upholds the Art 5.1 (and 3.3) violation; reverses on 3.1, 5.5 and burden of proof; reports adopted p. 153 (partly). [ext] AB report.
1999 (30 Apr) SCVPH: risk identified; no threshold and no ADI for any of the six hormones; prepubertal children of greatest concern; oestradiol a “complete carcinogen” Table 14.2, pp. 152–154. [ext] Dir. 2003/74 recitals 5–7.
1999 (13 May) End of reasonable period for EC compliance [ext] WT/DS26/ARB para 38
1999 (12 Jul) WTO arbitrators set retaliation at US$116.8m a year (US) and C$11.3m a year (Canada) [ext]. The chapter gives “GBP 100 million (ca. EUR 160 million)” (p. 153).
1999–2000 UK Veterinary Products Committee (Oct 1999), EC CVMP (Dec 1999) and JECFA (Feb 2000) supply information contrary to the SCVPH view; SCVPH (3 May 2000) sees no reason to revise [ext] Dir. 2003/74 recital 8
2000 Copenhagen workshop on hormones and endocrine disrupters “confirmed concerns” p. 152
2001 EU still under sanctions p. 154

Later developments (after the chapter; all [ext]): - 2002: SCVPH reconfirms (Dir. 2003/74 recital 8). - 2003: Directive 2003/74/EC makes the oestradiol-17β ban permanent and the ban on the other five hormones provisional (recital 10). - 2007: EFSA opinion, quoted via CRS. - 2008: DS320/321 Appellate Body (16 October) reverses panel findings against the EC’s new measure but cannot complete the analysis. - 2009 US MoU and 2011 Canada MoU: market access through a duty-free quota for beef from untreated animals (20,000 t, then 45,000 t, per CRS); retaliation reduced. - 2013: revised MoU. - 2016: USTR moves to reinstate tariffs (CRS).

Lags. - Wildlife signal: first published warnings 1970–73; first attention in late 1980s (Collins Report 1989); still “remains to be established” in 2001 (p. 152). A lag of roughly 15–20 years before even assessment, with no regulatory action on environmental grounds by 2001. - US DES: carcinogenicity “confirmed” in the “early 1970s” (p. 149); the report’s own DES chapter dates the link between prenatal DES and vaginal cancer to April 1971 (p. 90). First withdrawal 1972–73; reversed on procedure 1974; final ban 1979. About eight years. The chapter frames the delay mainly through economic lobbying (p. 150). The primary record shows a procedural-law cause for the 1974 reversal. - EU DES: the chapter says 1987 (p. 149). The directive record indicates 1981. - Natural and synthetic hormones in the EU: the ban came against expert advice, not after a lag. This case is the inverse of most Late Lessons cases, with action ahead of scientific consensus rather than behind it. - Harms: unlike most chapters, no human harm from residues of legally used promoters is established in the chapter (p. 153). The harms documented are: - DES harm from pharmaceutical use (a different exposure route) - episodes attributed, with varying confidence, to illegal or accidental contamination (p. 151) - suggestive wildlife effects (p. 152) - economic costs of the trade dispute (p. 153)

The authors’ own lessons and conclusions#

Lessons derived from their evidence: 1. The original EU ban was not grounded in a characterised assessment of uncertainty, because the committees were never asked for one. It was therefore “in reality, a political risk assessment” (p. 154). Evidence: committee terms and conduct (p. 150); drivers of public concern (p. 151). 2. Standard risk assessment as practised by JECFA was too narrow. It covered authorised use only, single substances and manufacturer data (p. 150). It left out misuse (pp. 150–151), combinations (p. 150) and wildlife (p. 151). Related gaps the chapter attributes more generally to the residue studies or the regulatory process are unassayed metabolites (p. 152), sensitive subgroups (pp. 150, 152–153), environmental exposure pathways (p. 153) and animal welfare, which was “not given significant attention” when the promoters were introduced (p. 152) and was absent from the trade dispute (p. 153). 3. The safety case rested on inappropriate comparators: pill users and pregnant women instead of children, and “natural” oestradiol instead of structurally different DES (p. 150). It also rested on unreliable measurement of children’s own hormone levels (pp. 152–153). 4. Economic arguments against restriction were overstated (“probably groundless”). There was “little evidence of a sustained increase” in costs, which “could be explained” by substitutes and, in part, by wrong cost assumptions (p. 150). 5. Drug regulators neglected environmental effects through institutional disinterest and comforting assumptions about dilution and degradation (p. 152). 6. Outcome evidence cuts both ways. No good evidence the ban protected health; no reliable safety evidence accumulated where use continued; no conclusive environmental evidence (p. 153). 7. Bans may have perverse effects by encouraging illegal use, possibly of more toxic substances; the chapter reports this as “claims”, and says the effect “is hard to gauge” without monitoring (p. 151). 8. With hindsight, “more recent scientific research … probably justifies” continuing the ban (p. 154). [p. 154 does not name the research. The likeliest referents in the chapter are the finding that oestradiol-17β is a genotoxic carcinogen (p. 153) and the absence of a demonstrable threshold (p. 151); the link is my inference.]

Recommendations and advocacy: - Precaution is appropriate where the science base cannot confirm safety (p. 154). This is a normative premise, stated rather than argued. - It is “very important” that scientific committees are requested to identify the uncertainties in their assessments (p. 154). - Rigorous and transparent mechanisms are needed to weigh risks against benefits (pp. 153–154), including formal mechanisms “for trading risks and benefits for the public” (p. 153). - Implied rather than stated: environmental impacts and animal welfare should enter both assessments and trade disputes (pp. 152–153). - A framing question rather than a recommendation: “where the benefit of doubt should lie” (p. 153).

What the chapter does not recommend. It does not say committees should have been overruled, or that bans should follow public concern. Its argument is procedural: ask the right question, and the decision stops being purely “political”. It also does not claim that the ban has been vindicated by evidence of harm.

Mechanisms and dynamics#

1. The question asked shapes the answer#

The chapter’s core mechanism is that expert bodies answer the question they are given, within conventional boundaries. - JECFA assessed authorised use, single agents and manufacturer-supplied data (p. 150). - Wildlife impacts were “not considered by either the Lamming Committee or JECFA” (p. 151). - Neither was “required to properly characterise the degree of uncertainty” (p. 154). - Their “safe” verdicts were conditional on good practice and on a limited set of exposures. When the conditions were relaxed, the verdicts no longer spoke to the policy question.

The WTO Appellate Body later made the same observation in legal terms. Most studies found use “safe” if hormones are given under good veterinary practice; where that condition is not met, “the logical inference is that the use … may or may not be ‘safe’” (AB para 206) [ext].

The chapter implies, rather than argues, that a narrow framing is not neutral. It tends to produce reassuring conclusions, because excluded pathways count as zero.

2. Reference populations, baselines and the “denominator” problem#

Safety arguments compared residues with doses in high-exposure groups: pill users and pregnant women (p. 150). The FDA’s 1% criterion compared residue intake with endogenous production (p. 153; the chapter does not date it).

Both depend on assumed baselines. For prepubertal children, baselines were low and poorly measured, near detection limits (pp. 152–153). So small absolute additions could be large relative ones.

This is a structural blind spot. A risk metric expressed as a proportion of a natural background is only as good as the lowest, least-measured background it assumes. The editors elevate this to a general lesson about sensitive subgroups versus “average” responses (p. 175).

3. Detection limits as de facto safety standards#

“Minimum detectable levels” served as regulatory levels (pp. 149–150). “Below the limit of analytical detection” was treated as evidence of negligible risk (p. 149). The FDA’s residue estimate could not be checked by measurement (p. 150), and metabolites were not assayed (p. 152).

The US court record sharpens this [ext]. The 1962 DES proviso made legality depend on “no residue … found” by approved methods. - When the USDA’s gas–liquid chromatography (1971) and radiotracer studies (1973) found residues the older mouse-uterine test could not, approval collapsed. - This happened even though the FDA said it had “no reason to believe” there was a public health hazard (Hess & Clark, 495 F.2d 975).

Regulatory status tracked measurement capability, not new evidence of harm. The chapter captures the “detection = safety” frame but not this reversal. It is a general dynamic: where zero-tolerance rules are pegged to detection, better instruments create regulatory shocks, and “not detected” is mistaken for “absent”.

4. “Natural”, “similar” and “physiological”: framing as argument#

The safety case leaned on: - naturalness (“natural oestrogens”, the “physiological range”, p. 152) - analogy with pharmaceutical oestrogens (p. 150)

The chapter’s counters: - DES is structurally different from oestradiol and the pill’s oestrogens (p. 150). - A natural hormone can itself be a genotoxic carcinogen (p. 153). - “Physiological range” is contested (p. 152).

In 1979 the FDA found that manufacturers “had not sustained their burden of proving that DES is similar to natural estrogens” (Rhone-Poulenc 1980) [ext]. On the public side, the chapter says concern was about “involuntary and unnatural hormone exposure” (p. 154). So “natural/unnatural” did rhetorical work on both sides. For proponents it signalled safety; for publics it signalled contamination.

5. Economic claims, lobbying and substitution#

The chapter presents a pattern in which predicted costs of restriction were overstated and delay gave time to prepare substitutes. It gives no cost data, so this rests on one case. The substitutes here were other hormones, so the chemistry and the controversy continued. The chapter calls the case for DES specifically “a purely economic one”, with no benefit to the animals or the environment (p. 153). [My inference, not the chapter’s: benefits that accrue only to producers bear on how much uncertainty is tolerable.]

The FDA’s 1979 environmental analysis did credit DES with some environmental benefits: less animal waste and less feed to grow (Rhone-Poulenc) [ext]. The chapter’s “no benefit … for the environment” (p. 153) is therefore contestable.

6. The gap between authorised use and real use#

Risk was assessed under authorised use. Real use included higher doses, mixtures, wrong injection sites, unremoved implants, shortened withdrawal periods and illegal substances (p. 151). The Italian baby-food case shows the unremoved-implant pathway (p. 151).

The Pimenta Report endorsed the ban partly because it “facilitated controls” (p. 151). This matters: a total ban is easier to enforce, because any detection is a violation, than a conditional permission that needs dose and withdrawal compliance. Enforceability is thus a hidden criterion in the choice between a ban and conditional approval.

The Appellate Body later confirmed that control and enforcement difficulties may legitimately be part of a risk assessment. But it found the EC had not actually assessed them (AB paras 205–208) [ext].

7. Risk–risk trade-offs and displacement#

The chapter reports “claims” that the ban drove illegal use, including DES, so that it “may have led to an increased risk to the consumer” (p. 151).

This is the classic displacement dynamic: prohibition pushes activity into unmonitored channels. The chapter pairs it with a monitoring gap. Without “a substantial regular monitoring programme” the net effect “is hard to gauge” (p. 151). The evidence cited is weak for this ban: Loizzo 1984 predates the 1985/88 ban, though it does document illegal DES use in Italian baby food after DES had already been prohibited. The only other evidence is DES detected “last year” in US meat imported into Switzerland (p. 151).

8. Science–policy interface and institutional behaviour#

The Commission overruled its committee. The chapter says the committee’s conclusion was “clearly unpopular” with officials. The committee was disbanded and its conclusions were not published, so members published independently (p. 150).

[My interpretation; the chapter reports the facts but does not draw this link.] This is a legitimacy dynamic. Not publishing inconvenient advice leaves the decision open to the charge of being “political”. It also lets opponents present the expert record as one-sided in their favour. In the WTO case the EC itself listed the 1982 “Lamming Report” among the scientific material it relied on. The Panel treated it, with the 1988 and 1989 JECFA reports, as meeting “the minimum requirements of a risk assessment”, and found these concluded that growth-promotion use of the hormones (except MGA) was “safe” (AB paras 195–196, re-checked) [ext].

The chapter locates the root failure upstream, in how committees were commissioned, not in the Commission’s willingness to act (p. 154).

9. How public concern formed#

According to the chapter, EU concern was driven by: - DES harm from a different use (pharmaceutical, in pregnancy) - worries about the pill - epidemiological episodes of uncertain cause: Puerto Rico’s unidentified source, and Italy’s suspected contamination (pp. 151, 153)

This is risk perception by class membership. Harm from one member and one route (DES in pregnancy) raised concern about the class (hormones) and a different route (residues in meat). The chapter treats this concern as a legitimate trigger for precaution (p. 154) but does not present it as evidence of harm. Its “political risk assessment” label rests explicitly on a different point: the committees were never asked to characterise uncertainty (p. 154). The editors later write that the Puerto Rico and Italy effects “were blamed on the illegal or incompetent use of animal growth promoters” (p. 175). That wording is reportorial, but it ties Puerto Rico to growth promoters, whose source the chapter says “was not clearly identified” (p. 151).

10. Institutional remits and environmental blind spots#

The chapter tentatively (“may be ascribed to”) explains why 1970s wildlife warnings “were ignored until the late 1980s” by the “lack of interest of drug agencies in the environment” and the assumption that excreted drugs dilute and degrade quickly (p. 152).

The mechanism is that an institution assesses harms within its remit (animal and human health) and treats other pathways as someone else’s problem. The editors generalise this under “interdisciplinary obstacles” (p. 174).

11. Divergent jurisdictions, trade law and the burden of proof#

The same evidence produced different decisions: - The US banned DES earlier but kept other hormones for economic reasons (p. 150). - The EU banned all six hormones (p. 150).

The divergence became a trade conflict with unilateral retaliation, then WTO litigation and authorised sanctions (p. 153). The trade regime then became the arbiter of what counts as adequate evidence.

The chapter frames the dispute as “where the benefit of doubt should lie” (p. 153). It notes an institutional gap: there is no formal mechanism for trading off public risks and benefits, and environment and welfare are absent from the dispute (p. 153).

On the burden of proof, primary sources show three different regimes [ext]: - US animal-drug law: once the FDA brought new evidence, manufacturers had to show safety (Rhone-Poulenc). - WTO: the Appellate Body reversed the panel’s placing of the burden on the regulating member. But it required measures to be “based on” a risk assessment specific to the risk at stake (AB paras 200, 208). - Precaution: the Appellate Body held that the precautionary principle does not override SPS Arts 5.1 and 5.2. It is “reflected” in Art 5.7, the preamble and Art 3.3, and governments “commonly act from perspectives of prudence and precaution” where risks are irreversible (AB paras 124–125).

The chapter’s reading, that the appeal clarified risk assessment need not be quantitative and may rest on minority “qualified and respected sources” (p. 153), is accurate as to AB paras 186–187 and 194. The chapter does acknowledge “the success of the United States” (p. 153), but it presents the upheld finding as narrow when it was dispositive: the ban was not “based on” a risk assessment, so the EU lost.

The costs of divergence were borne by third parties: sanctions hit a “variety of EU exports” unrelated to beef (p. 153). This is a distributional effect of trade enforcement.

12. Symmetric ignorance: nobody generates the decisive evidence#

The chapter’s most balanced observation is on p. 153. After the ban there was no good evidence of health protection. Where use continued, no reliable safety evidence had accumulated.

[The rest of this section is my inference, not the chapter’s argument.] Once positions harden, neither side may have strong incentives to generate the long-term, low-dose, sensitive-subgroup outcome evidence that could settle the question. The ban removes the exposed population in one jurisdiction; continued use removes the regulatory motive in the other. The dispute then persists on inference and extrapolation. One qualification from the chapter itself: the EU “has continued to encourage scientific evaluation” and was “sponsoring a number of ongoing research projects” (p. 152). So the gap is in decisive outcome evidence, not in research funding as such. The chapter’s two recommendations respond partly to this, but do not address who should fund resolution.

13. Distribution of benefits, risks and costs#

14. Complexity, mixtures and timing#

Together these undermine a single-substance, average-adult, threshold-based framework. The chapter does not develop a formal alternative.

Transferable insights (technology-neutral)#

  1. Expert “safe” verdicts are conditional on the question and scope set for the assessors. Assessments limited to intended use, single agents and sponsor-supplied data can certify safety without addressing real-world exposure, combinations or misuse. - Evidence: JECFA scope (p. 150); misuse forms (p. 151); p. 154 diagnosis. Independently echoed by the WTO Appellate Body (AB para 206) [ext]. - Strength: strong. Documented in the chapter and corroborated by a primary legal source.

  2. When assessors are not asked to characterise uncertainty, decision-makers must choose between a confident “safe” and public concern, and whatever they decide looks political. Requiring explicit uncertainty statements is a structural remedy. - Evidence: p. 154; committee history, p. 150. - Strength: moderate. The logic is sound and central to the chapter, but it shows little about the committees’ actual terms of reference.

  3. Safety arguments built on the average or on high-exposure comparison groups can miss the most sensitive subgroup. Where the baseline is low, small absolute additions are large relative ones. - Evidence: pill and pregnancy comparisons versus children (p. 150); prepubertal boys and the 1% criterion (pp. 152–153). Later: Aksglaede et al. 2006 found children’s oestradiol levels “lower than originally claimed” and FDA production-rate estimates “highly overestimated” [ext]. JECFA 1999 added a tenfold factor for sensitive populations [ext]. - Strength: strong.

  4. When legal thresholds are defined by what can be detected, “not detected” gets read as “safe”, and regulatory status can flip with advances in measurement rather than new evidence of harm. - Evidence: pp. 149–150, 152. The primary record shows withdrawals in 1972–73 triggered by more sensitive analytical methods, with the FDA disclaiming a health hazard [ext: Hess & Clark]. - Strength: strong for the first half (“not detected” read as “safe”), which the chapter documents. The second half (status flipping with better measurement) rests on the external court record, not on the chapter, whose account of the 1972 ban does not mention it.

  5. Appeals to “natural”, “physiological” or “similar to an accepted product” work as safety arguments without being evidence of safety. Structural and biological differences can be decisive, and a natural agent can itself be hazardous. - Evidence: pp. 150, 152, 153. In 1979 the FDA rejected the similarity argument [ext: Rhone-Poulenc]. - Strength: moderate. Well illustrated, but the “genotoxic carcinogen” claim for oestradiol was contested.

  6. Predicted economic costs of restriction can be overstated; substitutes may already exist, and delays give industry time to develop more. - Evidence: cost estimate with “questionable assumptions” (p. 149); lobby claims despite alternatives (p. 150); “little evidence of a sustained increase” in costs (p. 150); Table 14.1. - Strength: moderate. The claim is plausible and consistent with the dates, but the chapter gives no cost data. The 1980 court found the manufacturers had not proved economic benefits [ext].

  7. Prohibition can displace activity into illegal or unmonitored channels. Without monitoring, the net effect of a ban on risk is unknowable. - Evidence: p. 151. - Strength: suggestive. Presented as “claims”; the cited source predates the ban; the mechanism is generic and plausible.

  8. Institutions assess harms within their remit and discount pathways outside it, especially when comforting assumptions (dilution, degradation) fill the gap. Early warnings in the neglected domain are ignored for long periods. - Evidence: p. 152 (1970s warnings ignored until late 1980s); p. 153. Later field studies of feedlot effluent (Orlando et al. 2004) and the finding that trenbolone metabolites regenerate in surface waters (Qu et al. 2013) undercut the degradation assumption [ext]. - Strength: moderate. The neglect is documented. The institutional explanation is offered tentatively (“may be ascribed to”, p. 152). Later studies show the degradation assumption was wrong, but they do not test the remit explanation.

  9. Overruling, disbanding or not publishing inconvenient expert advice undermines the legitimacy of the decision and of the advisory system, and leaves the decision exposed when challenged. - Evidence: p. 150; p. 154. In the WTO case the Panel treated the EC’s own 1982 Lamming Report (with JECFA) as a risk assessment concluding the hormones were safe (AB para 196) [ext]. - Strength: moderate. The facts are reported; the motive (“unpopular”) is asserted. The legitimacy link is my inference; the chapter does not draw it explicitly.

  10. A total ban can be chosen partly because it is easier to enforce and more trusted than conditional permission. That is a legitimate criterion but seldom made explicit.

    • Evidence: Pimenta Report endorsed the ban for “controls and consumer confidence” (p. 151).
    • Strength: moderate. A single source, but it is the stated rationale of an official inquiry. The AB accepted control difficulties as relevant in principle [ext].

  11. Where two jurisdictions decide differently on the same evidence, the disagreement can be settled by an international body that sets its own evidentiary standards. That body may accept minority views and non-quantitative, real-world risk factors, yet still require assessments specific to the risk in question.

    • Evidence: p. 153 [ext: AB paras 124–125, 186–187, 194, 200, 205–208].
    • Strength: strong on the legal content. The chapter itself understates the EU’s loss.

  12. After divergent decisions, the outcome evidence that could settle the question may not accumulate on either side. The ban removes exposure; continued use removes the motive to study. Disputes then run on for decades.

    • Evidence: p. 153 (“no good evidence” of protection; no reliable safety evidence). Later: the EU’s 2003 reassessment still could not give “a quantitative estimate of the risk” (Dir. 2003/74 recitals 6–7) [ext]. EFSA 2007 said whether residues contribute to cancer risk “is currently unknown” (via CRS) [ext]. Counter-evidence: the EU kept commissioning evaluations and sponsoring research (p. 152).
    • Strength: moderate. The evidential stalemate is documented. The incentive explanation is my inference, not the chapter’s.

  13. Who benefits shapes how much uncertainty is tolerable. Where benefits go only to producers and uncertain risks fall on consumers, bystanders or ecosystems, the case for requiring positive evidence of safety is stronger. That needs explicit risk–benefit machinery.

    • Evidence: pp. 150, 153–154.
    • Strength: asserted. Normative, and largely my extension. The chapter’s “purely economic” judgement is made of DES specifically (p. 153), and its “no environmental benefit” claim is contestable [ext: Rhone-Poulenc].

  14. Public concern can generalise from harm caused by one use or route to a whole class of agents or uses, and drive action ahead of expert consensus.

    • Evidence: p. 151; p. 153; p. 154.
    • Strength: moderate. The chapter’s account of the drivers is plausible but not documented (no surveys or policy records).

  15. Adverse effects that are reversible in individual episodes do not settle long-term or developmental risk. With no demonstrated threshold, uncertainty about early-life exposure persists.

    • Evidence: p. 151; pp. 149, 152.
    • Strength: moderate.

Limitations, contestation and bias check#

Standpoint and possible interest#

Factual errors and inaccuracies (checked against primary sources)#

  1. EU DES ban date. The chapter says 1987 (p. 149); Chapter 1’s Table 1.1 repeats this (p. 12). Directive 81/602/EEC (31 July 1981) already banned hormonal substances in farm animals, exempting only therapeutic/zootechnical use and five named hormones pending review; DES was not exempted (AB paras 2–3) [ext]. The 1987 date appears wrong unless it refers to something I have not found.
  2. “Third world countries” (p. 150; Table 14.2). The EU import ban applied to all “third countries” (non-EU states), including the US and Canada (AB para 4; Dir. 96/22) [ext]. This matters, because the import ban is what triggered the transatlantic dispute.
  3. DES “vaginal clear-cell adenoma” (pp. 151, 153). The lesion is clear-cell adenocarcinoma, and it followed prenatal pharmaceutical exposure. The report’s own DES chapter (pp. 84–90) and index (p. 209) use “adenocarcinoma”. The p. 151 phrasing (“DES, which had been extensively used as a growth promoter, caused …”) implies harm from DES used as a growth promoter.
  4. US 1972 ban “on the grounds that it was a carcinogen, violating … 1958 Delaney Clause” (p. 149). The operative rule was the 1962 DES proviso: no residue by approved methods. The FDA framed the 1972 withdrawal as “a regulatory, not a public health problem”, with “no evidence of a public health hazard” (Hess & Clark) [ext].
  5. 1974 reinstatement “because of procedural deficiencies in the original bill” (p. 150). A court set aside FDA orders for failing to give an evidentiary hearing; there was no bill [ext]. The editors turn this into “followed strong pressure from the farming lobby” (p. 179), which the chapter does not say and the court record does not support. The chapter only says the lobby made claims during the period.
  6. 2 ppb “set in 1976” (p. 150). The 1980 court says the 2 ppb standard was set before 1971 [ext]. Possible date error.
  7. 1979 ban grounds cited to “Jakes, 1976” (p. 150). A 1976 source cannot document a 1979 decision. The substance, no demonstrable no-effect level, is confirmed by Rhone-Poulenc [ext]. The author is Thomas H. Jukes (“Diethylstilbestrol in beef production: what is the risk to consumers?”, Prev Med 5:438–453, 1976; PubMed 967836, confirmed via Europe PMC). Jukes publicly argued for continued DES use in cattle, saying the risk to consumers was minuscule [ext: R. Conniff profile, Alicia Patterson Foundation; see also Jukes, “Estrogens in beefsteaks”, JAMA 1974]. So the “1 cancer in 133 years” figure (p. 149) is cited to a prominent defender of DES use. The chapter does not say this.
  8. WTO chronology and outcome (p. 153). - Proceedings began in 1996 (US consultations 26 January 1996; panel established 20 May 1996), and the panel reported in August 1997 [ext]. The chapter’s “in 1997 … challenged … in front of a WTO panel” is therefore imprecise rather than simply wrong. - The chapter says the only point upheld on appeal was lack of residue-specific focus. That is the AB’s central ground (para 200), but it was dispositive, not minor. The AB upheld the finding that the ban was not based on a risk assessment (Art 5.1, hence 3.3) on three grounds:
    • general studies were not specific to residues in meat (para 200)
    • there was no assessment of MGA (para 201)
    • there was no assessment of misuse and control risks (paras 207–208)
    • The EU lost and had to comply by 13 May 1999 [ext].
    • The AB reversed the panel on Art 3.1, on Art 5.5 (discrimination or disguised restriction) and on the burden of proof (WTO summary) [ext].
  9. 1996 withdrawal of US tariffs “as not compatible with WTO law” (p. 153). CRS says the tariffs “remained in effect until May 1996, when the EU sought a WTO panel against the U.S action” [ext, secondary]. I found no WTO ruling on them. The chapter’s causal gloss is unsupported.
  10. Sanctions “GBP 100 million (ca. EUR 160 million)” (p. 153; Table 14.2). Authorised retaliation was US$116.8 million a year (US) and C$11.3 million a year (Canada), set on 12 July 1999 [ext: WT/DS26/ARB, WT/DS48/ARB]. That is roughly US$124–125 million combined, below the chapter’s figure.
  11. Italian school outbreak attributed to DES (p. 151). The Fara et al. abstract (re-checked via Europe PMC) reports no oestrogen contamination in tested school meals; an “uncontrolled supply of poultry and beef was suspected”. It does not name DES [ext]. The enlargement “disappeared within 8 months”, which is consistent with the chapter’s point about transient effects.
  12. Framing of the Commission’s decision as about “DES’s use” (p. 151). The overruled committees and the 1985/88 decision concerned the six hormones. Internally inconsistent.
  13. Ban chronology (pp. 150, 154). The chapter places the EU ban “shortly after” JECFA’s 1988 conclusions and says it was not supported by JECFA “at the time”. But it also says the ban was first adopted in 1985, before JECFA had assessed the hormones. Internally inconsistent for the 1985 decision.
  14. Minor labels (pp. 150, 154). Trenbolone is listed among “other oestrogenic compounds”, but it is an androgen. JECFA is called the “Joint Expert Committee on Food” and the committee “of WHO”; it is the Joint FAO/WHO Expert Committee on Food Additives.

Citation problems#

In-text citations often do not match the claims: - Peakall 1970 (a DDT paper) cited for the DES catfish study. - Bulkey [Bulkley] 1972 (catfish) cited for the trout study. - Delong et al. 1973 (sea lions) cited for bird feminisation. - Adkins 1975 (quail) cited for oestrogen and breast cancer in women. - Henderson et al. 1988 (cancer lecture) cited for the trade dispute and as the source of Table 14.1. - Johnstone et al. 1978 (“Sex reversal in salmonid cultures”) cited, first in the list, for DDT, PCBs and alkylphenols disrupting oestrogen receptor function in wildlife. - SCVPH 1999 cited for a 1970 DDT finding. This is a secondary citation. The primary paper, Peakall 1970 on DDT and blood oestradiol, is in the list but attached to the catfish claim. The citations in this paragraph appear shifted by one. - Loizzo et al. 1984 (the Italian baby-food DES follow-up; PubMed 6143928) cited for claims that the 1985/88 ban caused illegal use, which it predates. It does document illegal DES use after DES was already prohibited, so it bears on the general displacement point, but not on this ban. - The Collins Report and the Pimenta Report are not referenced.

These do not necessarily invalidate the claims, several of which are real findings. But the chapter’s evidentiary scaffolding cannot be taken at face value.

Thin evidence and overstatement#

Counter-arguments acknowledged and ignored#

Hindsight and case selection#

What later evidence says (post-2001; all [ext], selected primary sources)#

Regulation and trade: - The EU made the oestradiol-17β ban permanent and kept the other five provisionally (Directive 2003/74/EC). It relied on SCVPH opinions of 1999, 2000 and 2002 finding no threshold and no ADI for any of the six hormones, prepubertal children of greatest concern, and oestradiol a “complete carcinogen”. - The EU challenged continued US and Canadian sanctions (DS320/321). The AB (16 October 2008) reversed key panel findings against the EU’s new measure but could not complete the analysis. - The dispute was managed rather than resolved: US–EU MoU 2009, revised 2013; Canada–EU MoU 2011. The EU expanded a duty-free quota for beef from untreated cattle (20,000 t, then 45,000 t, per CRS), and retaliation was reduced. USTR moved to reinstate tariffs in 2016 (CRS). - A 2019 US–EU arrangement on the quota is reported but I have not verified it from a primary source. - US finishing-cattle production still routinely uses “growth enhancing technologies” (Duckett 2026, Vet Clin North Am Food Anim Pract).

Science: - Children’s sensitivity: Aksglaede et al. 2006 (Hum Reprod Update) found prepubertal oestradiol levels “lower than originally claimed”, children “extremely sensitive”, no established threshold, and FDA 1999 production-rate estimates “highly overestimated and should be revised”. This is strong later support for the chapter’s p. 152–153 argument. - Human health outcomes: EFSA 2007, as quoted by CRS: epidemiology convincingly associates red meat consumption with some hormone-dependent cancers, but whether hormone residues contribute “is currently unknown”. Swan et al. 2007 (Hum Reprod) found lower sperm concentration in sons of mothers who ate more beef in pregnancy (observational; the authors suggest a possible role for xenobiotics in beef; the paper drew published comment). Suggestive, not decisive. - Endogenous oestrogen and breast cancer: Key et al. 2002 (JNCI), pooling nine prospective studies, found endogenous sex hormone levels “strongly associated” with postmenopausal breast cancer risk. This supports the general premise on p. 153 but not a residue-specific effect. - Environment: Orlando et al. 2004 (EHP) found demasculinised male and defeminised female wild fathead minnows exposed to feedlot effluent, with potent androgenic activity. Qu et al. 2013 (Science) found trenbolone metabolites regenerate in surface water (“product-to-parent reversion”), undermining degradation-based fate assumptions. Both support the chapter’s environmental concern (p. 152) and specifically its criticism of the dilution/degradation assumption. - Puerto Rico: Colón et al. 2000 (EHP) found phthalates in the serum of girls with premature breast development. This is an alternative, non-meat hypothesis for part of the epidemic the chapter cites (p. 151). The contemporaneous Sáenz de Rodríguez et al. 1985 (J Pediatr) had suspected meat, having found oestradiol-equivalent activity in some meat samples. The source remains unresolved in the sources I checked.

Notable quotes#

Kept short (under 15 words) and few, given the report’s copyright.

  1. “the original EU ban was, in reality, a political risk assessment.” (p. 154)
  2. “Neither committee was required to properly characterise the degree of uncertainty” (p. 154)
  3. “young children with low natural levels of oestrogens were the likely ‘at risk’ group” (p. 150)
  4. “The ‘breathing space’ enabled the pharmaceutical industry to develop additional hormonal growth promoters.” (p. 150)
  5. “the ban may have led to an increased risk to the consumer” (p. 151)
  6. “no good evidence that the EU ban … has protected the health of the public.” (p. 153)
  7. “The case for DES use as a growth promoter was a purely economic one.” (p. 153)
  8. “The dispute raises the issue of where the benefit of doubt should lie.” (p. 153)

Open questions#

  1. What were the committees actually asked? What were the Lamming group’s and JECFA’s terms of reference, and did any EU mandate ask for uncertainty characterisation? The chapter’s diagnosis depends on this but does not document it.
  2. Why was the Lamming group disbanded? Is there documentary evidence for official displeasure (p. 150), or was it routine?
  3. When was DES banned EU-wide? Is there any reading under which “1987” (p. 149) is right, or is 81/602 (1981) definitive?
  4. Did the ban increase illegal use? What do EU residue-monitoring data since Directive 96/23/EC show for illegal hormone and stilbene use? Is there evidence the ban raised or lowered net exposure (p. 151)?
  5. Was the chapter’s vindication borne out? Has post-2007 science (EFSA or later EU reviews, JECFA re-evaluations, IARC) moved on oestradiol genotoxicity at dietary levels and on children’s exposure (the p. 154 “probably justifies” claim)?
  6. Were the economic predictions right? What happened to meat production costs after the US 1979 and EU 1988 bans (p. 150)? What have hormone-free supply chains cost?
  7. Has the environmental concern been acted on? Has regulation anywhere acted on feedlot hormone discharges since 2001, and did environmental assessment enter veterinary drug approval as the chapter implies it should (pp. 152–153)?
  8. How did the trade experience affect later EU precaution? Did the hormones case shape the EU’s 2000 Communication on the precautionary principle and later framings of risk assessment and risk management? This is outside the chapter; it needs checking.
  9. Were the recommendations adopted? Did requirements for committees to state uncertainty follow (for example, later EU guidance on uncertainty analysis in scientific assessments), and did they change outcomes?
  10. Who funds the decisive evidence? The EU funded research (p. 152), yet neither side had generated decisive outcome evidence (p. 153). Who should fund long-term, low-dose, sensitive-subgroup studies, and how?

Audit log#

Independent audit against the chapter text (pp. 149–156, all pages read), the PDF (table source positions on p. 154; pp. 11–12, 84–90, 173–181, 195, 209) and selected outside sources (the AB report text; Europe PMC records).

  1. Added a note on audit re-checks under “Sources outside the chapter”, flagging that the Hess & Clark/Rhone-Poulenc wording was only partly re-verified.
  2. Added from p. 195 that Olga Bridges “has participated in a number of public enquiries”.
  3. Specified that the Panel cited the Bridges paper in its MGA reasoning, beside the point about substance-by-substance assessments (AB fn 107, re-checked).
  4. Stance: corrected “Commission … disbanded one and did not publish its conclusions” to the chapter’s passive “was disbanded” and “the EU did not publish its interim conclusions”; noted Table 14.2’s “by EU”.
  5. Recorded that the Table 14.1/14.2 source attributions were confirmed from the positions of the source lines on the PDF page.
  6. DES EU date: replaced “no ‘no effect’ dose could be defined” with the chapter’s “uncertainty as to whether there was a definable ‘no effect’ dose”; added that DES use continued in some Member States longer than in the US.
  7. Delaney: replaced “to avoid applying it” with the chapter’s “forced to refer to ‘minimum detectable levels’” to justify inaction.
  8. 1974: added “temporarily” and “further ban”; flagged that the chapter does not blame the lobby for the reinstatement (the editors do, p. 179).
  9. Added the omitted point that a scientific debate on safe DES residue levels ran alongside the 1976 FDA level.
  10. Costs: softened “showed no sustained rise” to “little evidence of a sustained increase”; kept the chapter’s hedge (“could be explained”) and its second reason (wrong assumptions).
  11. Flagged that trenbolone is an androgen, although the chapter lists it among “other oestrogenic compounds”.
  12. Lamming: “never published its conclusions” corrected to “did not publish its interim conclusions”.
  13. JECFA: quoted “indications of significant accidental or deliberate misuse”; softened the “evidential core” bracket to an interpretive note; flagged JECFA’s name and the “of WHO” slips.
  14. EU ban: replaced “de facto application” with the chapter’s hedged “could be regarded as an application”.
  15. Added the omitted internal-chronology problem: the 1985 ban predates JECFA’s assessment, although the chapter places the ban “shortly after” it.
  16. Three factors: added the chapter’s hedge that they “appear” to have influenced the Commission.
  17. Pimenta: marked “European Parliament inquiry” as [ext]; quoted the chapter exactly; noted the report is unreferenced.
  18. Illegal use: made clear these are unendorsed “claims”; “cannot be gauged” corrected to “is hard to gauge”.
  19. Wildlife: removed the unsupported “in birds” (the chapter gives no species); added “likely to affect a wide range of species”; added the tentative “may be ascribed to”.
  20. Prepubertal boys: “undermines the FDA criterion” softened to the chapter’s “particularly relevant”; marked the denominator point as an implication.
  21. 14.4: noted that “purely economic” is said of DES only; added the omitted “very considerable … yet to be fully evaluated” economic consequences; noted Henderson 1988 cited for the trade dispute; added the chapter’s own acknowledgement of US success.
  22. Appeal clarification: “is not quantitative” corrected to “is not necessarily quantitative in nature”.
  23. 14.5: recommendations quoted exactly (“very important … are requested”); scope “acceptability of the use of chemical substances” added; noted p. 154 does not name the “more recent research”.
  24. Table 14.2: “compiled by the editors” hedged to “presumably”. Pimenta discrepancy reframed as unsupported rather than contradictory. Added three missed discrepancies: 1988 “based on uncertainty”, 1989 “extended to other growth promoters”, and 1999 SCVPH “no threshold” not in the text.
  25. References: “About 30” corrected to 28.
  26. Timeline: flagged Chemetron v. HEW as the feed-premix companion case; corrected the Milan outbreak dates (1977–78) and added “disappeared within 8 months”; quoted 81/602 wording (re-checked, AB para 3); added a 1985-before-JECFA note.
  27. Lags: replaced the unsourced “carcinogenicity evidence 1971” with the chapter’s “early 1970s” plus the report’s Chapter 8 date (April 1971, p. 90).
  28. Authors’ lessons #2: separated what JECFA’s scope omitted from gaps the chapter attributes to residue studies or the wider process (animal welfare, metabolites).
  29. Authors’ lessons #4 and #7: softened to the chapter’s hedged wording.
  30. Authors’ lessons #8: marked the referents of the “more recent research” as my inference.
  31. Mechanism 1: “ignored wildlife” corrected to “not considered”.
  32. Mechanism 2: removed the unsupported “Later” dating of the FDA 1% criterion.
  33. Mechanism 5: softened “typical dynamics” to a single-case pattern; restricted “purely economic” to DES; marked the tolerable-uncertainty point as my inference.
  34. Mechanism 7: noted that Loizzo documents illegal DES use under an earlier prohibition; added the Swiss detection as the only other evidence.
  35. Mechanism 8: marked the legitimacy framing as interpretation; “suppressing” became “not publishing”; AB paras 195–196 re-checked, with “(except MGA)” added.
  36. Mechanism 9: corrected the claim that concern-not-evidence is why the ban was “political” (the chapter’s stated ground is the unasked uncertainty question); made the editors’ p. 175 wording precise.
  37. Mechanism 10: added the chapter’s tentative “may be ascribed to”.
  38. Mechanism 11: replaced “understates that the EU still lost” (the chapter acknowledges US success) with “presents the dispositive finding as narrow”.
  39. Mechanism 12: marked the incentive argument as my inference and added counter-evidence that the EU kept sponsoring research (p. 152).
  40. Insight 4: the strength now separates the chapter-supported half from the half resting on the external court record.
  41. Insight 6: softened (“can be”, “may already exist”) and quoted the chapter’s cost wording.
  42. Insight 8: downgraded from moderate-to-strong to moderate, because later studies do not test the remit explanation.
  43. Insight 9: noted that the legitimacy link is inferred.
  44. Insight 12: reworded (“outcome evidence may not accumulate”), added the p. 152 counter-evidence and downgraded to moderate.
  45. Insight 13: noted that “purely economic” is said of DES only and that the insight is largely my extension.
  46. Bias check, standpoint: specified the Panel’s use of the Bridges paper.
  47. Error #3: cited the report’s DES chapter (pp. 84–90) and index (p. 209) for “adenocarcinoma”.
  48. Error #7: confirmed the author as T. H. Jukes (Europe PMC) and his public defence of DES use in cattle (Conniff profile), replacing “should be checked”.
  49. Error #8: WTO “1997” reframed as imprecise (proceedings 1996, panel report 1997); the upheld finding described as the AB’s central and dispositive ground.
  50. Error #11: Fara abstract re-verified with an exact quote; added transience.
  51. Added error #13 (internal ban chronology) and error #14 (trenbolone and JECFA-name slips).
  52. Citation problems: added the Johnstone 1978 mismatch; noted that the wildlife-paragraph citations appear shifted (Peakall 1970 is the DDT paper); nuanced Loizzo; added the unreferenced Pimenta Report.
  53. Counter-arguments: replaced the unverified “genuine anxieties” with the AB para 245 wording; tidied the CAP parenthesis; added the omitted harmonisation driver (AB para 245); renumbered the list.
  54. Framing: made the editors’ p. 175 and p. 179 amplifications precise.
  55. Open question 10: reworded in light of the p. 152 EU research funding.
  56. Digest: corrected the US chronology (chapter’s 1972 ban vs external 1972–73 withdrawals; “temporarily”, “procedural deficiencies”).
  57. Digest: corrected “The Commission did not publish … and disbanded it” to the chapter’s wording; added the 1985-before-JECFA note.
  58. Digest: replaced “drivers were public concern” with the chapter’s three hedged factors plus p. 154’s “principally”.
  59. Digest: noted that p. 154 does not specify which “more recent research” it relies on.
  60. Digest: made the wildlife explanation tentative; completed the cost explanation; replaced “Misuse was widespread” with the chapter’s hedged wording.
  61. Digest: softened “undermines the FDA’s rule” to “particularly relevant”.
  62. Digest insights: #4 split for strength; #5 and #6 softened; #7 downgraded to moderate; #8 “hiding” became “not publishing”; #10 reworded and downgraded; #11 DES caveat added.
  63. Digest caveats: 1981 claim hedged; WTO date softened; AB outcome reframed; Jukes confirmed; internal slips, Table 14.2 and the harmonisation omission added; the editorial-amplification line made specific.
  64. Scanned both files for references to specific contemporary technologies or companies not in the source: none found; nothing to remove.
  65. Digest: “JECFA ignored misuse” corrected to “considered only authorised use despite indications of … misuse”; “The EU’s scientific committees” corrected, since JECFA is WHO/FAO, not EU.