LL2-A3 digest — Annex 3: An update of some case studies from vol. 1#

EEA Report No 1/2013, pp. 717–737 (PDF 719–739). Full notes: notes/LL2-A3.md. [External — verify] marks material not from the annex.

What it is#

Nine short updates to nine of the fourteen 2001 case studies, with an unsigned editorial introduction (p. 717). There are no panels and no industry or dissenting voices, and no declarations of interest. Most authors revisit their own cases or promote their own lab’s work. The hormones piece is the European Commission’s own lawyer summarising a WTO ruling (pp. 736–737). The ozone piece is an anonymous summary of a 2006 WMO assessment (p. 728).

Editors’ thesis: harm “expands over time” to new harms, often at exposures “previously considered ‘safe’” (p. 717).

Core content#

• Fish (Ruzzin, p. 718). A new warning. In the author’s rodent studies, background POPs in farmed salmon “can fully abolish” the benefits of fish. A switch to vegetable feed halved the omega-3 ratio (2005–09), with the health effects unstudied. Evidence is suggestive.
• Benzene (Infante, pp. 720–721). Blood cancers at exposures below 1 ppm; childhood leukaemia “appears” raised with prenatal and childhood exposure (exposure inferred from proxies). Risk assessments are “likely” to underestimate risk 3–9-fold (nonlinear metabolism). The Pliofilm cohort may omit 4–9 deaths, a claim that rests on unpublished work by the author (p. 722).
• Asbestos (Castleman and Gee, pp. 724–726). The richest account of political economy in the annex.
• At least 107,000 occupational deaths a year (WHO, no reference given); bans in over 50 countries “offset” by growth in China and India.
• Canada blocked listing under the Rotterdam Convention until 2012, and Russia pressured Thailand.
• Substitutes cost little: USD 65 extra for an asbestos-free Thai town-house roof (0.5% of overall cost; unsourced); fibre-cement 10–15% more to make, partly offset downstream (World Bank).
• Cheap optical microscopy was used for cost reasons, and so shaped the evidence base and the metric of exposure limits (p. 725).
• PCBs (Koppe, p. 727). Links to diabetes via epigenetics; the same genes altered by PCBs and by DDT-type pesticides, read as additive; “denial” said to have hampered development of cures. Thinly sourced.
• Ozone (p. 728). Montreal Protocol working, with recovery by mid-century “assuming global compliance”. Substitute gases also contribute to climate change. The annex’s clearest success story, which the editors leave unreconciled with their thesis.
• DES (Swan, pp. 729–730). The strongest update. NCI cohorts show 12 outcomes in daughters (HR 1.4–8.1) with dose gradients, harm in sons, and tentative third-generation signals. It ends with a warning about BPA.
• Antimicrobial growth promoters (Silbergeld, pp. 731–733).
• Early bans (per the author) cost no productivity. An industry-run trial showed small or no benefit, and the author’s own reanalysis found a “net loss for producers” even assuming a significant performance gain.
• The practice outlived its 1940s rationale as a vitamin B12 substitute (she concedes it may once have had real benefits).
• Early-1970s warnings about horizontal gene transfer (the shared “reservoir”) were “apparently ignored”.
• Verdict: “intransigence in the face of overwhelming evidence”.
• MTBE (Belpoggi, p. 735). Calls MTBE “probably carcinogenic” on the strength of her lab’s uncited rodent studies, apparently recasting a case originally about groundwater [External — verify].
• Hormones (Christoforou, pp. 736–737). The WTO Appellate Body faulted the Panel on several grounds, including:
• compromised expert independence;
• a misallocated burden of proof;
• a “restrictive notion of risk assessment” in relation to misuse and abuse of hormones;
• deciding “the best science” itself rather than the “more limited exercise” its mandate required.

International standards are not decisive, and a rigid “critical mass” test for insufficient evidence was rejected. But the AB “could not complete the analysis”, so the ban’s WTO-consistency was left unconfirmed, and the AB held the US and Canada had not acted unilaterally; lifting sanctions needed agreement or further litigation.

Main mechanisms#

• Harm widens as detection improves.
• Risk assessment underestimates through nonlinear biology, one canonical dataset, cheap instruments, one-agent-at-a-time assessment and assumed ideal use.
• Legacy stocks and long latency delay relief.
• Hazardous uses grow in less-regulated markets as others restrict them; exporters block information treaties.
• Sponsoring ministries side with industry.
• Benefit claims go untested.
• Systems optimised for producer metrics degrade properties that matter to users, and the consequences go unstudied.
• Change comes from coalitions, media, elections and courts.

Transferable insights (the full list of 23 is in the notes)#

1. Harm profiles widen and effect thresholds fall with continued study (pp. 717, 720, 729). Moderate: the sample was chosen because harm occurred.
2. Linear extrapolation is not automatically conservative (p. 721). Moderate: the finding is contested.
3. One canonical dataset can anchor regulation for decades (p. 721). Suggestive.
4. Cheap instruments shape the evidence and the metric of limits (p. 725). Suggestive to moderate: mechanism stated, effect on limits not shown.
5. Benefit claims need the same scrutiny as risk claims; when benefits fail, the burden shifts (pp. 732–733). Moderate.
6. Practices can outlive their rationale (p. 733). Suggestive.
7. Safer alternatives often cost little at system level (pp. 724–725). Moderate to strong.
8. Local restriction coincides with growth of the hazard elsewhere, and exporters may block information-sharing (pp. 724–726). Strong for exporter obstruction (asbestos); moderate for displacement, since causation is not shown.
9. Harms that pool in a shared resource make each local use a system-wide cost (pp. 731–732). Moderate to strong.
10. Risk assessment should include real-world misuse (p. 737). Moderate.
11. Whether precaution survives adjudication turns on who carries the burden of proof, how much deference reviewers give the regulator, and whether experts are independent (pp. 736–737). Moderate: partisan source.
12. Relief from legacy harm takes decades and depends on continued compliance (pp. 724–725, 728). Strong.

Caveats#

• The annex is selective, confirmatory advocacy, not a review.
• Key evidence is self-cited, unpublished or uncited (benzene, MTBE, PCBs).
• There are internal slips: the editors’ DES example is loosely worded about who is at risk (p. 717 vs. p. 729); AB report numbers are inconsistent (p. 736); the MRSA claim is unsourced and the “80%” use figure is loosely framed (p. 731); Silbergeld’s “two recent industry studies” include a USDA survey (p. 732).
• [External — verify] Later events strengthened several warnings: BPA restrictions, IARC Group 1 for PCBs, asbestos bans in Canada and the US, action on growth promoters, lower EU benzene limits. MTBE carcinogenicity, the breadth of benzene’s cancer spectrum, and the fish–diabetes link remain contested.