LL2-03 hindsight check: Lead in petrol ‘makes the mind give way’ (Needleman and Gee), Late lessons from early warnings: science, precaution, innovation (EEA Report No 1/2013), Ch. 3, pp. 46–75#

Checked 25 September 2026. This check covers what happened between 2013 and September 2026 that bears on the chapter’s claims, evidence, predictions and recommendations. Page numbers are report pages. Herbert Needleman, one of the chapter’s two authors, is also a protagonist in the story. The other, David Gee, edited the Late lessons reports. Leaded petrol was not a case in the 2001 report, so Annex 3 of the 2013 report has no update for it. Every development below is from outside the report.

Overview#

The chapter’s core science has held up well: lead is neurotoxic at low doses, no threshold has been found, and removing lead from petrol brought very large benefits. Some of that science has been strengthened and written into law. The weaker parts are specific figures, dates and illustrations that were wrong or already out of date in 2013. The chapter also treats the story as more finished than it was.

Vindicated or strengthened - No threshold, and “safe” levels keep falling (claim 1). The trend the chapter describes in Box 3.11 continued: - CDC replaced its 10 µg/dL “level of concern” with a population reference value of 5 µg/dL in 2012, and lowered it to 3.5 µg/dL in 2021. It now states there is “no known safe blood lead level”. - England halved its public health intervention level from 10 to 5 µg/dL in July 2021. - WHO (2026) says “no level of exposure to lead … is known to be without harmful effects”. - In 2024 the EU cut the binding occupational blood lead limit from 70 to 30 µg/100 ml, falling to 15 µg/100 ml from 2029. It also labelled lead a “non-threshold reprotoxic substance” (Directive (EU) 2024/869). - A 2013 reanalysis of the key pooled study by consultants, one of whom had co-authored an earlier critique, “tended to support” its conclusions (Crump et al., 2013). - The size of the benefits (claim 2). US child blood lead fell 94.5% between 1976–80 and 2011–16 (Egan et al., 2021). A 2022 study estimated that early-life lead exposure cost the US population 824 million IQ points (McFarland et al., 2022). Newer studies that use natural experiments link leaded petrol to crime, schooling, earnings and elderly mortality. These include Sweden (Grönqvist et al., 2020), Rhode Island (Aizer and Currie, 2019) and US motor racing (Hollingsworth and Rudik, 2021). - The legacy and global burden (claim 5). Current estimates are far larger than the 2004 figure the chapter quotes. Up to 800 million children (1 in 3) have blood lead at or above 5 µg/dL (UNICEF and Pure Earth, 2020). WHO now attributes more than 3.5 million deaths in 2023 to lead. E-waste did grow, from 53.6 Mt in 2019 to 62 Mt in 2022, with only 22.3% formally recycled. But the main sources in poorer countries turned out to be battery recycling, mining and smelting, adulterated spices and paint. - Corporate conduct (claim 8). The bribery happened as the chapter says, and the primary record is worse than the chapter reports. In 2006–07 Octel/Innospec paid USD 155,000 to make sure an Iraqi field trial of a competing octane additive (MMT) failed. It also funded “Lead Defense activities” to stop Indonesia switching to unleaded fuel (SEC complaint, 2010). - Early warners (lesson 5, p. 71). In the Flint drinking-water crisis (2014–15), the state regulators’ response to independent testers and a paediatrician was “aggressive dismissal, belittlement, and attempts to discredit” them. That is the official task force’s finding (Flint Water Advisory Task Force, 2016).

Overstated, wrong or dated - “The level that remains in place today” (Box 3.11, p. 62). This was out of date when the report was published: CDC had replaced the 10 µg/dL level of concern in 2012. - “5–6 points” of IQ gain (Box 3.10). The box copies WHO (2010) almost word for word. Grosse et al. (2002), the source it cites, estimate 2.2–4.7 points, and that gain comes from all sources of the decline, not petrol alone. - The Octel convictions (p. 68). In 2010 it was the company that pleaded guilty, in the UK and the US. Former executives pleaded guilty in 2012 or were convicted by a jury in June 2014, and all were sentenced in August 2014. - Lead and antisocial behaviour (claim 4). The link is supported: a 2025 systematic review by EPA scientists finds a “likely causal” association. But the size of the crime effect has been inflated by publication bias (Higney et al., 2022). A New Zealand cohort in which blood lead did not track social class found no dose-response with criminal convictions (Beckley et al., 2018). - The alcohol alternative (claim 7). Ethanol did become the main octane source in US petrol a century later. But ethanol at scale needed mandates and has large land-use and carbon costs. The cellulosic alcohol Midgley described in 1921 is still commercially marginal in 2026. The post-lead substitutes carried hazards of their own: in 2025 IARC classified MTBE as possibly carcinogenic and petrol itself as carcinogenic. - The mobile-phone example (claim 10). The footnote says “there are no studies in children”. The CEFALO study of 7–19-year-olds was published in July 2011, so this was already inaccurate. Later evidence has weakened the example: MOBI-Kids (2022), COSMOS (2024) and a WHO-commissioned systematic review (2024) all found no clear association.

What the chapter could not see or did not say - Leaded fuel is not finished. The global phase-out for road vehicles was completed only in July 2021 (Algeria). Leaded aviation fuel remains the largest single source of lead to US air: about 470 tons, or 70% of emissions, in 2017. EPA made its endangerment finding in October 2023. The FAA aims to end leaded aviation fuel by the end of 2030. In the EU, REACH authorisations allow continued use of tetraethyl lead (TEL) until April 2032. Innospec, formerly Octel, describes itself as “the world’s only producer” of TEL for aviation fuel. - UK monitoring came back only in part. England rebuilt case-based surveillance (SLiC 2010–12, a pilot in 2014, LEICSS from 2016). But it still has no population blood lead survey. In 2018 the UK National Screening Committee declined population screening, citing, among other things, a lack of prevalence data. UKHSA says “there is no data available” on how many children exceed the intervention level. - The recommendations were taken up unevenly, and gains can be reversed. - The EU adopted whistleblower protection (2019), study-notification and verification-study rules for food-chain risk assessment (2019), and in 2025 an EEA-run “early warning and action system for emerging chemical risks”, due by 2 January 2027. - The US adopted scientific-integrity policies in 2021–24, but they were rescinded or rewritten under Executive Order 14303 (May 2025). - The whistleblower directive protects only reports of breaches of EU law. An early warning about a lawful but hazardous product is not clearly covered.

Weight for the lens. Mechanism-level lessons from this chapter can carry substantial weight. These include: - no threshold, so “safe” levels fall over time; - small individual effects that add up across a population; - decades-long legacy harm that falls on the poor; - sponsors dominating the evidence base; - dismissal of early warners; - the importance of an unrelated technological co-driver.

Each has been reinforced since 2013. Some within lead itself, such as early warners at Flint and evidence-rigging in the MMT trial, are fresh repetitions of the pattern. Do not quote the chapter’s specific numbers without correction: the CDC level, the 5–6 IQ points, the Octel dates and the “no studies in children” footnote. Its claims on crime and the alcohol alternative need qualification. The strongest new lesson the chapter points towards but never states is this: “temporary” exemptions last. The aviation exemption written into the 1996 US ban is still in place 30 years later, now backed by an EU authorisation to 2032.

Claim 1. There appears to be no threshold for lead neurotoxicity in children; harm occurs at 5 µg/dL and lower; the IQ loss per µg/dL is steeper below 10 µg/dL than above it; and the CDC level of concern of 10 µg/dL “remains in place today” (Box 3.1, p. 47; Box 3.11, p. 62; p. 61)#

Original claim. - Box 3.1 (after WHO, 2010): “lead is associated with neurobehavioural damage at concentrations in the blood of 5 μg/dl and even lower. There appears to be no threshold level below which lead causes no injury to the developing human brain” (p. 47). - Page 61: “most of the cognitive impairment seeming to occur at blood lead levels as low as 5 μg/dl (Lanphear et al., 2000)”. - Box 3.11 traces the CDC level from 60 µg/dL (1960s) down to 10 µg/dL (early 1990s), “the level that remains in place today”. It then gives the pooled-analysis finding: a rise from below 1 to 10 µg/dL “was associated with a six IQ point decrement, which is considerably greater than the decrement associated with an increase … from 10 μg/dl to 20 μg/dl” (Lanphear et al., 2005). - It says EFSA withdrew support for the provisional tolerable weekly intake in 2010 (p. 62).

What happened since. - CDC abandoned the 10 µg/dL “level of concern” in 2012, before the report was published. An MMWR update records that “In 2012, CDC introduced the population-based blood lead reference value (BLRV)”. That value was 5 µg/dL, the 97.5th percentile of US children aged 1–5. CDC lowered it to 3.5 µg/dL in October 2021, stating that “there is no known safe blood lead level” (Ruckart et al., MMWR 70(43), 29 October 2021). - The conceptual change matters. The new value is a statistical benchmark (“more lead than most other children”), not a health-based threshold. That is the move from “normal” to “typical” that the chapter credits to Patterson (pp. 57–58). - EFSA (April 2010) set a BMDL01 for developmental neurotoxicity of 12 µg/L (1.2 µg/dL) blood lead. It concluded that the PTWI “is no longer appropriate as there is no evidence for a threshold for critical lead-induced effects” (EFSA CONTAM Panel, EFSA Journal 8(4):1570). - Box 3.11 calls the body the “European Food Standards Agency”; it is the European Food Safety Authority. - The chapter’s own source (WHO, 2010) credits the formal withdrawal of the PTWI to JECFA in June 2010. - I found no later EFSA re-evaluation of lead hazard. - Reanalysis of the pooled data. Crump et al. (2013), a team from Louisiana Tech and consultancies (ENVIRON, Gradient) that included a co-author of a 2006 critique of the non-linear curve, reanalysed Lanphear et al. (2005). They found “some small errors and questionable decisions” but concluded that “our reanalysis tended to support their conclusions”: - there was “statistical evidence that the exposure-response is non-linear over the full range … for a given increase in blood lead, the associated IQ decrement is greater at lower BPb levels”; - “at BPb below 10 µg/dL, the exposure-response is adequately modeled as linear”; - there was an association at maximal blood lead of 7 µg/dL or less and at concurrent levels as low as 5 µg/dL or less. - Lanphear et al. corrected the 2005 analysis in 2019 (EHP erratum). Two errors in the Boston cohort data “had only minor effects on the estimates and did not change the overall results”. The corrected decrements are: - 3.8 IQ points (95% CI 2.3–5.3) for 2.4 to 10 µg/dL; - 1.8 points for 10 to 20 µg/dL; - 1.1 points for 20 to 30 µg/dL. - Regulators and health agencies have moved towards “no known safe level”. - WHO. The June 2026 fact sheet says “There is no level of exposure to lead that is known to be without harmful effects” and that levels “as low as 3.5 µg/dL may be associated with decreased intelligence”. WHO’s 2021 clinical guideline recommends intervention at 5 µg/dL or above. - England. UKHSA halved its case definition from 10 to 5 µg/dL on 5 July 2021 after an evidence review. Its 2025 report states “There is no defined safe threshold”. - EU occupational law. Directive (EU) 2024/869 (13 March 2024) cuts the binding biological limit value to 30 µg Pb/100 ml until 31 December 2028 and 15 µg/100 ml from 1 January 2029, down from 70. It adds the notation “non-threshold reprotoxic substance”, stating: “It is not scientifically possible to identify a level below which exposure to lead … would be safe for the development of the offspring”. The Risk Assessment Committee (RAC) recommended, non-bindingly, that women of childbearing age stay below 4.5 µg/100 ml. - US EPA. EPA issued a new Integrated Science Assessment for Lead in January 2024 (EPA/600/R-23/375). I did not read its causality appendices. - IARC. IARC has announced a new evaluation of inorganic lead compounds for 19–26 October 2027 (Meeting 144). That meeting concerns cancer, not neurotoxicity. - Wider endpoints. Heart and blood-vessel deaths now dominate burden estimates: - Lanphear et al. (2018) followed NHANES III adults and linked blood lead in the 1–6.7 µg/dL range to all-cause and cardiovascular mortality. - Larsen and Sánchez-Triana (2023) estimate 5.5 million cardiovascular deaths worldwide in 2019, “six times higher than the GBD 2019 estimate”. - This supports the chapter’s lesson 2, that mortality and acute endpoints mislead. But these figures are model-dependent (see claim 5).

Verdict: strengthened for the no-threshold and low-dose claims. The specific statement that 10 µg/dL “remains in place today” was out of date at publication. Lead has joined a small group of agents that regulators in several jurisdictions formally treat as having no identified threshold. What the evidence shows is “no threshold has been identified”, and the chapter’s “appears” is the right hedge: residual confounding at the lowest exposures cannot be ruled out statistically. The supralinear curve (steeper at low doses) survived an adversarial reanalysis. Below 10 µg/dL, though, a linear model fits as well.

Weight. Strong support for digest insights 5 (the chosen endpoint defines the hazard; assumed thresholds delay prevention) and 7 (“safe” levels fall as measurement improves and exposure drops). Correct the CDC statement before quoting. The replacement of a “level of concern” by a statistical reference value is a useful mechanism in its own right: an institution that cannot find a safe dose switches from a safety threshold to a benchmark of what is typical.

Claim 2. Phasing out leaded petrol was associated with a >90% fall in mean US blood lead and a 5–6 point gain in population IQ, worth USD 100–300 billion per US birth cohort (Grosse et al., 2002); similar declines followed in Europe, China, India, Mexico and elsewhere (Box 3.10, p. 62)#

Original claim. Box 3.10 (source: WHO, 2010) says the 1976–1995 phase-out “was associated with a more than 90% reduction in the mean blood lead concentration”. It also says that in the same period lead was removed from food-can solder and new paint. “An estimated gain of 5–6 points in mean population IQ score was associated with the decline … calculated to yield an annual economic benefit of between USD 100 billion and USD 300 billion in each birth cohort” (Grosse et al., 2002). Declines are also reported in Europe and in “China, El Salvador, India, Mexico and Thailand”.

What happened since. - The source misquoted. Grosse et al. (2002) estimate that US preschool children in the late 1990s “had IQs that were, on average, 2.2–4.7 points higher” than with the late-1970s blood lead distribution. The benefit for “each year’s cohort of 3.8 million 2-year-old children ranges from $110 billion to $319 billion”, as present-value lifetime earnings. - The “5–6 points” figure appears in WHO (2010), which Box 3.10 reproduces nearly word for word. Grosse’s abstract does not contain it. - Grosse’s estimate covers the fall in blood lead from all sources: petrol, solder and paint. It is not a petrol-only attribution. - The US decline continued. NHANES data show the geometric mean blood lead in children aged 1–5 fell from 15.2 µg/dL (1976–80) to 0.83 µg/dL (2011–16), a 94.5% decrease. Levels remain higher among non-Hispanic Black children, low-income families and those in older housing (Egan et al., EHP, March 2021). - The historical scale of harm. McFarland, Hauer and Reuben (PNAS, March 2022) back-cast blood lead from 1940 using petrol-lead data: - more than 170 million Americans alive in 2015 (53%) had early-life blood lead above 5 µg/dL; - exposure above that level “was nearly universal (>90%) among those born 1951 to 1980”; - total losses came to 824 million IQ points. - Causal designs attributing effects specifically to leaded petrol (economics literature after 2013): - Grönqvist, Nilsson and Robling (JPE, 2020) followed 800,000 Swedish children “differentially exposed to the phaseout of leaded gasoline”. Even low exposure affected long-run outcomes including crime and human capital. Boys were more affected. Reductions on the scale of the recent redefinitions of elevated blood lead “can increase earnings by 4%”. - Aizer and Currie (REStat, 2019) used road proximity and the de-leading of petrol as instruments for 125,000 Rhode Island children. For boys, a 1 µg/dL rise in blood lead increased the probability of school suspension by 6% and of juvenile detention by 57%. - Hollingsworth and Rudik (AEJ: Policy, 2021) used motor-racing exemptions. Leaded fuel raised ambient lead, elevated blood lead rates and elderly mortality: “each gram of lead added to gasoline exceeds $1,100 in damages”. - The global picture. UNEP (30 August 2021) put the benefits of the worldwide phase-out at more than 1.2 million premature deaths prevented per year and USD 2.45 trillion saved per year. That estimate comes from a UNEP-sponsored analysis (Tsai and Hatfield, 2011) and is best treated as an order of magnitude. - Other countries, with qualifications. - China. Median blood lead in rural children fell from 6.3 to 3.1 µg/dL between 2002 and 2012, and the share above 5 µg/dL fell from 63.6% to 14.2% (Liu et al., 2023). - India. A meta-analysis of 2010–18 studies found a mean child blood lead of 6.86 µg/dL. Levels “remain elevated despite regulatory action to eliminate leaded petrol” (Ericson et al., Environ Int, 2018). In some countries petrol was not the dominant source, so phasing it out was necessary but not sufficient.

Verdict: partly held up. The size and direction of the benefit are well supported, and later causal studies strengthen the attribution to petrol specifically. The chapter’s IQ figure misreports its own cited source, and its framing leaves out co-occurring reductions from other sources. “Similar declines” did occur in China, but in India levels stayed high.

Weight. Strong support for digest insight 15 (small individual effects add up across a population). Use Grosse’s 2.2–4.7 points, or McFarland et al., rather than “5–6”. Present the benefit as the result of a combined programme against lead: petrol, solder and paint. The newer natural-experiment studies are better evidence for petrol-specific harm than the chapter’s own sources.

Claim 3. Nearly all countries had phased out leaded petrol by 2012–13, and unleaded petrol accounted for about 99% of world sales (pp. 46, 62, 71); leaded petrol was “finally completely phased out in the US in 1995” (p. 62)#

Original claim. “nearly all countries worldwide had phased out leaded petrol by 2012” (p. 46). “Worldwide, unleaded petrol now accounts for an estimated 99% of total sales” (Box 3.10, p. 62, from WHO 2010). Table 3.1 for 2013 reads “Nearly all countries worldwide have phased out leaded petrol” (p. 71). “Leaded petrol was finally completely phased out in the US in 1995” (p. 62).

What happened since. - The global phase-out for road vehicles was completed in July 2021. UNEP announced on 30 August 2021 that Algeria, the last country selling leaded petrol, stopped in July 2021. That ended a campaign by UNEP’s Partnership for Clean Fuels and Vehicles, which began in 2002. WHO (2010), the chapter’s source, had listed nine countries still selling leaded petrol in late 2010. - The US on-road ban took effect on 1 January 1996. It exempted “aircraft, racing cars, farm equipment, and marine engines” (US EPA press release, 29 January 1996). “Completely phased out in … 1995” is therefore inaccurate twice over: on the date, and because uses remained. - Aviation fuel remains the main use. - EPA’s final endangerment finding (20 October 2023) states that lead from piston-engine aircraft is “the largest single source of lead to air in the U.S.” In 2017 it was about 470 tons, 70% of annual lead emissions to air, emitted at and near thousands of the roughly 20,000 US airports. - The finding obliges EPA to set emission standards and the FAA to set fuel standards. EPA’s page (updated 8 May 2026) gives no rulemaking timeline. - The FAA-industry EAGLE initiative aims “to eliminate leaded aviation fuels in piston-engine aircraft safely by the end of 2030”. The FAA approved an unleaded 100-octane fuel for general aviation by supplemental type certificate (G100UL, September 2022) and a second, narrower approval (Swift 100R, September 2024). - Section 827 of the 2024 FAA Reauthorization Act required a transition plan. A draft went out for comment from January to March 2026, and V1.0 has been published, with V2.0 expected “tentatively in late 2027” (FAA page, updated 30 July 2026). The FAA counts about 220,000 piston aircraft. - TEL is still made in the UK and authorised in the EU. Innospec’s FY2025 10-K (filed 18 February 2026) states: “We believe we remain the world’s only producer of tetra ethyl lead for use in aviation gasoline”. It also states that under EU REACH “three users successfully appl[ied] for EU REACH Authorization permitting continued TEL use until end of April 2032”. It notes an Alaska derogation to the end of 2032 under the US 2030 goal. The company’s own view is that “there is no currently available alternative”, which the FAA approvals above contest in part. Its principal TEL site, Ellesmere Port, carries USD 65.1 million in plant-closure provisions. - Racing. Hollingsworth and Rudik (2021) found that leaded fuel in motor racing measurably raised local elderly mortality until racing series switched fuels.

Verdict: held up for road transport. The phase-out was indeed nearly complete in 2012–13 and was completed in 2021. The chapter does not mention aviation fuel, the remaining and now largest single source of airborne lead in the US. Its “completely phased out” in the US is inaccurate.

Weight. The phase-out is a genuine global success. Its speed among late adopters is consistent with Hilton’s (2001) finding that “later abatement implies faster abatement”. The aviation case adds a lesson the chapter does not state: exemptions granted as temporary become durable. The 1996 exemption has lasted three decades. The sole producer defends it by arguing that no alternative exists. EU authorisations now extend it to 2032.

Claim 4. Childhood lead exposure is linked to antisocial and delinquent behaviour; “studies have further confirmed the link” (Byers 1943; Needleman 1996; WHO 2010) (p. 61)#

Original claim. Needleman “followed up the 1943 Byers discovery of the chronic anti-social behaviour of children who had ‘recovered’ from acute lead poisoning, confirming the association between childhood lead poisoning and anti-social adolescent behaviour (Needleman, 1996). Studies have further confirmed the link between lead and anti-social behaviour (WHO, 2010)” (p. 61). WHO (2010) lists links to “conduct disorder, juvenile delinquency, drug use and incarceration”. It also says crime reductions after the petrol phase-out are benefits that “have yet to be monetized”.

What happened since. - Available before 2013 but not cited. - Marcus et al. (2010) meta-analysed 19 studies (8,561 children). They found an average r of .19, or .15 excluding hair-lead studies, “strikingly similar in magnitude to the relation between lead exposure and decreased IQ”. - The Cincinnati prospective cohort linked prenatal and childhood blood lead to adult arrests: an arrest rate ratio of 1.40 per 5 µg/dL prenatal blood lead, and 1.48 for violent arrests per 5 µg/dL at age 6 (Wright et al., PLoS Med, 2008). - A null result where lead did not track social class. Beckley et al. (JAMA Pediatrics, 2018) followed 553 members of the Dunedin cohort (New Zealand, born 1972–73). Blood lead at age 11 averaged 11.0 µg/dL and was not associated with low socioeconomic status. Blood lead “was a poor discriminator between no conviction and conviction (area under the curve, 0.58)”. The authors concluded that “Findings failed to support a dose-response association between BLL and consequential criminal offending”. - Quasi-experimental evidence for an effect. Aizer and Currie (2019), in Rhode Island, and Grönqvist et al. (2020), in Sweden, both find effects on detention and crime through changes in non-cognitive skills (see claim 2). - A meta-analysis of the lead–crime literature. Higney, Hanley and Moro (Regional Science and Urban Economics, 2022) pooled 542 estimates from 24 studies. They found that “The effect of lead is overstated in the literature due to publication bias”. The corrected mean effect is a partial correlation of 0.16. Lead abatement “may be responsible for 7–28% of the fall in homicide in the US”. “Lead increases crime, but does not explain the majority of the fall in crime”. - Systematic reviews. - Talayero et al. (PLOS Global Public Health, 2023) reviewed 17 individual-level studies. They found “limited data” but concluded that “an excess risk for criminal behavior in adulthood exists” with early exposure. - Shaffer et al. (Environment International, September 2025), led by scientists in the US EPA Office of Research and Development, reviewed 43 epidemiological and 37 animal studies using EPA evidence-integration methods. The evidence was “slight” for aggression and “moderate” for antisocial-related disorders and for violating social norms. Overall, “evidence indicates a likely causal association between Pb and antisocial behavior”.

Verdict: partly held up. A causal link between childhood lead and antisocial behaviour is now the mainstream reading, and a 2025 EPA-authored review rates it “likely causal”. “Confirmed” overstates it. The effect on crime is smaller than much of the earlier literature suggested, publication bias is documented, and one well-designed cohort in which lead did not track social class found no association with convictions. The larger claim that lead explains most of the 1990s crime drop, which the chapter does not make, is not supported.

Weight. Treat this as moderate evidence for behavioural harm, not as a settled fact about crime. The episode carries its own caution: dramatic findings on social outcomes are especially prone to publication bias, even when the underlying toxicology is sound.

Claim 5. Legacy lead in soils and dust, and informal e-waste recycling in low-income countries, are continuing or emerging child lead hazards; in 2004, 16% of children worldwide were above 10 µg/dL (pp. 46–47, 62, 69; Box 3.12, p. 69)#

Original claim. “lead concentrations in soils and sediments remain high. Meanwhile, electronic wastes … also cause elevated blood lead levels” (p. 46). “lead in electronic waste is an emerging hazard for children in poor countries in Asia and Africa, where waste from rich countries is dumped” (p. 69). “In 2004, 16% of all children worldwide were estimated to have levels above 10 μg/dl (WHO, 2010)” (Boxes 3.11 and 3.12, printed twice).

What happened since. - The global exposure estimates are much larger, though measured at a lower threshold. - UNICEF and Pure Earth (July 2020) estimated that “1 in 3 children globally—up to 800 million” have blood lead at or above 5 µg/dL, using IHME data. - Ericson et al. (Lancet Planetary Health, March 2021) systematically reviewed low- and middle-income countries. They estimated that 632 million children (48.5%) in the 34 countries with poolable data exceed 5 µg/dL. “Major sources … were informal lead acid battery recycling and manufacture, metal mining and processing, electronic waste, and the use of lead as a food adulterant, primarily in spices.” - The 2004 figure of 16% above 10 µg/dL and these above-5 figures are not directly comparable. Improved measurement in poorer countries also accounts for some of the difference. - The burden of disease has been revised sharply upward, and is model-dependent. - WHO’s fact sheet (updated 10 June 2026): “Lead exposure was attributed to more than 3.5 million deaths globally in 2023”, mainly cardiovascular, and more than 71.5 million disability-adjusted life years (DALYs). - Larsen and Sánchez-Triana (Lancet Planetary Health, September 2023) estimated 765 million IQ points lost among children under 5 and 5.5 million adult cardiovascular deaths in 2019. That is six times the GBD 2019 estimate, and they put the global cost at USD 6.0 trillion (6.9% of GDP). - The spread between estimates is itself a caution. These are modelled from limited blood lead data. - E-waste grew as predicted, but is one source among several. - WHO (15 June 2021): more than 18 million children and adolescents work in the informal industrial sector, which includes waste processing. E-waste reached 53.6 Mt in 2019, of which only 17.4% was formally recycled. - The Global E-waste Monitor 2024 (November 2024) reports 62 Mt in 2022, 82% more than in 2010, with 22.3% documented as formally recycled. The total is projected to reach 82 Mt by 2030. - A 2023 meta-analysis of blood lead in children exposed to e-waste exists (Huang et al., Environ Sci Pollut Res). I did not read beyond its title. - Sources the chapter underplays. - WHO names lead-acid batteries, paint, e-waste and consumer products. As of January 2024, “48% of countries have legally-binding controls on lead paint”. - Turmeric adulterated with lead chromate was a major source in Bangladesh. An enforcement and media intervention cut market samples with detectable lead from 47% (2019) to 0% (2021) (Forsyth et al., Environ Res, 2023). This shows fast gains are possible when a source is identified. - Legacy sources in rich countries. - In England in 2024 the most commonly reported exposures were soil (24%), paint (17%) and drinking-water pipes (12%). 48% of cases lived in the most deprived areas (UKHSA LEICSS report, December 2025). - In the US, Flint’s 2014 water-source switch doubled the incidence of elevated blood lead in young children, from 2.4% to 4.9% (Hanna-Attisha et al., AJPH, 2016). - US EPA finalised the Lead and Copper Rule Improvements, requiring most lead service lines to be replaced (30 October 2024). It also lowered the dust-lead hazard standards (12 November 2024), with a correction in July 2025. I did not check how these have been implemented since 2025.

Verdict: strengthened. The chapter was right that the lead problem outlasts petrol and falls on poor children and poor countries. Later estimates make the problem larger than the chapter implied. E-waste proved a real and growing hazard. Battery recycling, mining and smelting, adulterated spices and paint, however, are now seen as larger sources in many low- and middle-income countries.

Weight. Strong support for digest insight 14 (persistent, diffuse harm outlives the product’s use and falls on the vulnerable). Quote the newer figures with their thresholds stated. Flag that burden-of-disease numbers, especially cardiovascular deaths, rest on models, with estimates for 2019–2023 ranging roughly from about 0.9 million to 5.5 million deaths.

Claim 6. The macroeconomic costs of lead regulation were “insignificant”, and the costs claimed at the time were “significantly biased, due to the vested interests that supported the analyses”; industry claimed “one million barrels of oil a day” against the EPA’s estimate of 30,000 barrels a day (Panel 3.3, p. 65; p. 60)#

Original claim. Von Storch et al., on Germany: “The macroeconomic costs of the regulation seem to have been insignificant in spite of concerns that they would be substantial (Hagner, 2000) … the costs claimed at the time of the regulations turned out to be significantly biased, due to the vested interests that supported the analyses” (p. 65). US: “The EPA estimated that the oil penalty from phasing out lead was 30 000 barrels per day”. A New York Times advertisement of 2 December 1973 claimed removal “would have the effect of dumping one million barrels of oil a day” (p. 60).

What happened since. - No post-2013 ex post study of leaded-petrol phase-out costs. I found none for Germany, the EU or the US that re-examines these claims directly. The German evidence is still von Storch et al. (2003) and Hagner (2000). I also found no ex post estimate of the US “oil penalty” to compare with the 30,000 versus one million barrels a day figures. That comparison sets an EPA estimate against an advertisement; neither is an ex post measurement. - General evidence on whether regulators overestimate costs is mixed. - Harrington, Morgenstern and Nelson (JPAM, 2000) found total costs overestimated for 14 of 28 rules and underestimated for 3. Unit costs, though, were “often accurate”, except for economic-incentive rules, where they were “consistently overestimated”, often because of “unanticipated technological innovation”. The US lead phasedown was a tradable-credit programme of this kind. - Simpson (J Benefit-Cost Analysis, 2014) argued that the common finding of overestimation “does not necessarily demonstrate that ex ante estimates are biased”, and could not reject unbiasedness. - Morgenstern (J Benefit-Cost Analysis, 2018), with 34 new ex ante/ex post comparisons, found “a slight tendency to overestimate both costs and benefits”. - The phase-down’s design lowered compliance costs. Kerr and Newell (J Industrial Economics, 2003) found that the tradable-permit system “provided incentives for more efficient technology adoption decisions” at US refineries. This predates 2013 but is not cited by the chapter. - The benefit side has grown. Per-gram damages exceed USD 1,100 (Hollingsworth and Rudik, 2021). Global cost-of-exposure estimates run to trillions (Larsen and Sánchez-Triana, 2023). Whatever the true costs, net benefits look larger in 2026 than in 2013.

Verdict: partly held up. That phasing out lead had large net benefits is stronger now than in 2013. The specific claims have not been re-examined since: that macroeconomic costs were insignificant, and that ex ante costs were “significantly biased” because of vested interests. The broader literature finds a modest tendency to overestimate costs, often because of unanticipated innovation, rather than systematic bias.

Weight. Moderate. Digest insight 11 (industry cost projections exceed the costs actually incurred) is consistent with the evidence but should not be stated as a law. The best-supported version is narrower. When a rule leaves room for innovation, and especially when it uses tradable credits, ex ante cost estimates tend to be too high. The one-million-barrels advertisement illustrates advocacy, not a measured forecasting error.

Claim 7. Counterfactual: a 10-year phase-out “at any time since 1925” would have spurred less hazardous and perhaps more efficient additives and engine designs; an “equally effective” and cleaner alcohol alternative existed in 1925 (pp. 46, 69; Panel 3.1, pp. 54–55)#

Original claim. “despite the availability of an equally effective alcohol additive which was assessed by experts to be cleaner, the leaded route to fuel efficiency was chosen” (p. 46). “a 10-year phase out of leaded petrol at any time since 1925 would have encouraged innovators to develop less hazardous and perhaps more efficient fuel additives and engine designs” (p. 69). Panel 3.1 (Kovarik) cites 1933 Navy tests in which 20% ethanol blends and leaded petrol were “almost exactly equivalent”. It also records Midgley’s 1921 view that alcohol from farm cellulose could fuel US vehicles, but “The catch was that it would cost two dollars per gallon” (p. 54).

What happened since. - Ethanol did become the octane source, a century late. Since at least 2016, E10 has accounted for more than 95% of US motor gasoline (US EIA, 4 May 2016). An ORNL history of octane and compression ratio since 1925 traces how fuel antiknock quality and engine design evolved together (Splitter et al., 2016). This goes some way to vindicating Kovarik’s point that alcohol blends could do the antiknock job. - Ethanol at scale needed mandates and has large side effects. Lark et al. (PNAS, 2022) found that the US Renewable Fuel Standard: - raised corn prices by 30%; - expanded corn cultivation by 2.8 million hectares; - increased fertiliser use and water-quality degradation; - left the carbon intensity of corn ethanol “no less than gasoline and likely at least 24% higher”. “Cleaner” held for tailpipe smoke and antiknock performance. At the scale of a whole fuel supply, the environmental balance is contested. - Midgley’s cellulosic alcohol remains marginal in 2026. EPA’s final Renewable Fuel Standard rule for 2026–27 (1 April 2026) states that cellulosic biofuel growth “has primarily been driven by renewable CNG/LNG, although small volumes of liquid cellulosic biofuels, particularly ethanol produced from [corn kernel fiber], have also played a contributing role”. It also partially waived the 2025 cellulosic requirement because of a shortfall. The “two dollars per gallon” problem Midgley identified has not been solved a century later, despite mandates. - Post-lead substitutes carried their own hazards. - MTBE, the main oxygenate and octane substitute in the US and Europe, caused groundwater contamination (the subject of a separate Late lessons case). In March 2025 IARC classified MTBE and ETBE as Group 2B (possibly carcinogenic) and automotive gasoline as Group 1 (carcinogenic to humans), on evidence for bladder cancer and acute myeloid leukaemia (Turner et al., Lancet Oncol, 2025; IARC Volume 138, published July 2026). - MMT (manganese) was the substitute Octel paid to sabotage in Iraq (claim 8). - The “less hazardous” substitutes a phase-out would have “spurred” were not automatically benign. - Induced innovation: evidence under modern conditions. The US phasedown induced refineries to adopt octane-raising technology in response to market incentives (Kerr and Newell, 2003). Late-adopting countries phased out quickly (Hilton, 2001). Both support the feasibility of a phase-out on a ten-year scale once refining technology exists. Neither can test the 1925 counterfactual.

Verdict: contested. A ten-year phase-out “at any time since 1925” would have produced safer innovation: that is an untestable counterfactual. The modern evidence shows that phase-outs can induce technology adoption, and that ethanol can supply octane. “Equally effective” is defensible for engine performance. “Cleaner” and practical at scale in 1925 are not established. The chapter underplays alcohol’s cost (which its own panel reports), its supply limits, and the hazards of the substitutes that were actually adopted.

Weight. Low as a factual claim; moderate as a mechanism. The durable lessons are that alternatives are “denied and then forgotten” once capital is committed (Panel 3.1; digest insight 10), and that regulation can induce innovation. Pair them with a lesson the chapter omits: substitutes need their own hazard assessment. The MTBE sequel and the 2025 IARC classifications make this concrete.

Claim 8. In summer 2010 two former Octel executives were convicted of bribing Indonesian and Iraqi officials to keep TEL in use; UK monitoring of children’s blood lead ended after 1987; the UK was one of the last industrialised countries to embrace unleaded petrol (Panel 3.4, p. 68)#

Original claim. “in the summer of 2010 two former senior executives of Octel were convicted of having bribed government officials in Indonesia and Iraq to continue allowing the use of tetra ethyl lead as a fuel additive in those countries (Leigh et al., 2010)” (source: a Guardian article of 30 June 2010). “official efforts to monitor childhood blood lead levels in British children then came to an end … It remains difficult … to estimate the proportion of children in the United Kingdom with elevated blood lead levels” (p. 68).

What happened since (and just before publication). - The company’s pleas (March 2010). SEC litigation release 21454 (18 March 2010) and the accompanying complaint set out the following. - Scale. Innospec, formerly Octel, made illicit payments of about USD 6.35 million and promised a further USD 2.87 million, for contracts worth USD 176.7 million in revenue and USD 60.1 million in profit. - Iraq. Kickbacks under the UN Oil-for-Food Programme (2001–03), then bribes to Iraqi officials until late 2007 to secure TEL contracts. - Indonesia. Bribes from 2000 until 2005, “when Indonesia’s need for TEL ended”. - Sabotaging a competitor. Innospec paid its Iraqi agent USD 155,000 “so that he could ensure the failure of a 2006 field trial test of MMT, a fuel product manufactured by a competitor … that competed with TEL”. The agent reported success “against all odds”. - Delaying unleaded fuel. In Indonesia the agent “had been working hard to try and stop Indonesian officials from switching to unleaded fuel in January 2005” and “might need some extra money to support the Lead Defense activities”. - Penalties. USD 40.2 million in total, including a USD 12.7 million criminal fine imposed in England. The SEC called it the first corruption settlement coordinated between the SEC, DOJ and SFO. - Individual convictions (2012–2014). The UK Serious Fraud Office (4 August 2014) announced the sentences: - Dennis Kerrison (former CEO), 4 years; - Paul Jennings (former CEO), 2 years; - Miltiades Papachristos, 18 months; - David Turner, 16 months suspended. Turner pleaded guilty in January 2012 and Jennings in June and July 2012. Kerrison and Papachristos were convicted by a jury in June 2014 “in relation to Indonesia only”. The judge called the corruption “endemic, institutionalised and ingrained”, with harm including “its effect on the environment”. “Innospec itself pleaded guilty in March 2010.” - The chapter’s “summer 2010 … two former senior executives … convicted” does not match these records. My inference, not verified: it may conflate the company’s 2010 pleas, or a related US plea by the Iraqi agent, with the later individual cases. I could not access the Guardian source. - UK child blood lead surveillance: case-based surveillance resumed; population monitoring did not. UKHSA’s LEICSS annual report 2025 (18 December 2025) sets out the history: - Rebuilding. A study of elevated blood lead in children (SLiC, 2010–12) recommended laboratory-based surveillance. A pilot began in 2014. It became permanent as LEICSS in 2016 and was integrated into a UKHSA prevention and surveillance group in December 2021. - Lower threshold. The case definition was halved to 5 µg/dL from 5 July 2021. - Few cases found. 2024 saw 247 cases (23 per million children), which UKHSA believes is “significantly lower than the estimated incidence”. - No population data. “there is no data available regarding the number of children in England who are exposed to lead resulting in concentrations above this threshold”. - Screening declined. In 2018 the UK National Screening Committee “did not recommend a systematic population screening program due to concerns about testing and treatment, as well as the lack of current population prevalence data”. UKHSA instead extrapolates US prevalence, giving about 34,500 children aged 1–5 above 5 µg/dL. IHME estimates as many as 213,702 children aged 0–19 above 5 µg/dL in the UK in 2019. - Octel’s successor still makes TEL for aviation fuel at Ellesmere Port (claim 3).

Verdict: partly held up. - Bribery: the substance is confirmed and in one respect worse. Paying to make a competing additive’s field trial fail is a direct instance of the “evidence manipulation” pattern. - Dates and descriptions of the convictions: wrong. - UK monitoring: the chapter’s point holds for population monitoring, which still does not exist in 2026. Case surveillance has been rebuilt since 2014. - The UK as a late adopter: not re-examined.

Weight. The Innospec record is strong primary evidence for digest insights 2 (sponsor control of evidence) and 8 (concentrated benefits against diffuse harms). It also illustrates export of harm to other countries (digest, main mechanisms). The UK surveillance story adds a circular mechanism, which is my reading rather than UKHSA’s. Screening was declined partly for lack of prevalence data, and prevalence data are missing because there is no population survey. This matches the chapter’s point about how the question put to experts sets the evidence bar (Millstone, p. 68).

Claim 9. Recommendations: independent, publicly funded long-term prospective monitoring when new technologies are introduced; legal protection and support for early-warning scientists; public release of industry and government decision records (pp. 70–71)#

Original claim. “There is a need for sufficient incentives and funds for independent long-term prospective monitoring of potential health hazards when new technologies are introduced” (p. 70). “Such scientists need more support from society via recognition for their work, help with their defence and legal protection against discrimination” (p. 71). “The concrete record of decision-making by industries, scientists and governments need to be made publicly available … This usually only occurs many years after the relevant events and then only via legal cases for compensation” (p. 71).

What happened since.

EU: substantial uptake in law, with limits - Whistleblower protection. - Directive (EU) 2019/1937 (23 October 2019) protects people who report breaches of Union law. Its scope includes product safety, protection of the environment, food and feed safety, and public health, and transposition was due by 17 December 2021. - The Commission’s implementation report (COM(2024) 269, 3 July 2024) found that only 3 Member States transposed on time. The Commission opened infringement proceedings against 24 in January 2022 and referred 6 to the Court of Justice in March 2023. - Key limit: “breaches” means acts that “are unlawful” or “defeat the object or the purpose” of EU rules. A scientist warning about a lawful but possibly hazardous product is not clearly protected. The chapter’s early warners (Henderson, Patterson, Needleman) were largely in that position. - Transparency of evidence in the food chain. Regulation (EU) 2019/1381 (20 June 2019; applying to applications from 27 March 2021) does three things. Business operators must notify EFSA of the studies they commission (Art. 32b). The public is consulted on the studies planned for renewals (Art. 32c). The Commission can commission EU-funded “verification studies” in “cases of high societal importance where there are serious controversies or conflicting results”. This speaks directly to the industry monopoly on data (lesson 1, p. 70). - Chemicals data and early warning. Regulation (EU) 2025/2455 (26 November 2025): - establishes a common data platform on chemicals, including human biomonitoring data held by the EEA; - requires the EEA to establish “a Union early warning system for emerging chemical risks” by 2 January 2027 (Art. 22); - lets ECHA “commission studies” to fill knowledge gaps; - creates a Database of Study Notifications by 2 November 2027 (Art. 9). Its contents “shall be considered confidential and shall not be made public”. This is close to the chapter’s call for independent monitoring. It is aimed at chemicals in general and is not yet operating. - Publicly funded monitoring. The European human biomonitoring initiative (HBM4EU, 2017–22) was followed in 2022 by PARC, the Partnership for the Assessment of Risks from Chemicals, a seven-year EU research partnership (Ramhøj et al., 2024).

US: gains, then reversal - Scientific integrity policies. Agencies adopted scientific-integrity policies under the January 2021 presidential memorandum (for example, final HHS and NIH policies in October–December 2024). Executive Order 14303, “Restoring Gold Standard Science” (signed 23 May 2025), then: - reverted agencies to the policies in force on 19 January 2021; - ordered agencies to “revise or rescind” policies issued between 2021 and 2025; - placed enforcement under “a senior appointee designated by the agency head”, who may “correct scientific information” and refer staff for discipline. - Transparency under the order. The order also requires agencies to publish “the data, analyses, and conclusions” behind scientific information, and to protect employees “from efforts to prevent or deter consideration of alternative scientific opinions”. - How to read it. In the terms of the chapter’s lessons, the order both expands disclosure and moves control of scientific-integrity decisions to political appointees. It cuts both ways against the chapter’s concern about protecting dissenting early warners.

Decision records - ToxicDocs. Historians Gerald Markowitz and David Rosner, whose Deceit and denial the chapter acknowledges, and Merlin Chowkwanyun launched ToxicDocs (version 1.0, 2018; Columbia University/CUNY). It is a searchable archive of “millions of pages of once-secret corporate documents about asbestos, polyvinyl chloride, benzene, silica, and lead”, largely from litigation. - Enforcement records. The SEC complaint and SFO records (claim 8) are another route by which records became public. Both confirm the chapter’s observation that records emerge mainly through legal proceedings.

Recurrence of the problem - Flint. The Flint Water Advisory Task Force (21 March 2016) found that Michigan’s environment department “failed in its fundamental responsibility” and that both state agencies “stubbornly worked to discredit and dismiss others’ attempts to bring the issues of unsafe water, lead contamination … to light”. Its response was “often one of aggressive dismissal, belittlement, and attempts to discredit these efforts and the individuals involved”. The independent testers and the paediatrician whose blood lead analysis showed the problem were vindicated (Hanna-Attisha et al., 2016).

Verdict: partly held up (read as: the recommendations remain valid, have been partly adopted, and the gains are reversible). - The EU has legislated versions of all three recommendations. Each has gaps: whistleblower protection is limited to breaches of law, the study database is confidential, and the early-warning system is not yet operating. - US scientific-integrity protections were strengthened and then rewritten in 2025. - Flint showed the dismissal of early warners recurring within a decade, in a lead case.

Weight. Strong support for digest insight 13 (early warners face retaliation; vindication comes late) and for the value of independent monitoring. For the lens, the most useful hindsight point is fragility. Protections created by executive policy were reversed within four years. Protections in legislation, as in the EU, have lasted but are narrow.

Claim 10. The EEA’s 2007, 2009 and 2011 early warnings on possible head-cancer risk to children from mobile phones as a current example of the “no evidence of harm” error; “there are no studies in children” (p. 50, footnote 3)#

Original claim. “there are no studies in children of the potential head cancer hazard of using mobile phones … Yet it is widely asserted that there are no risks to children … In 2007, 2009 and 2011 the EEA issued ‘early warnings’ about the potential hazard of head cancers from mobile phones, particularly in younger people. See Chapter 21” (p. 50, footnote 3). The footnote attaches to the chapter’s lesson that “no evidence of harm” is not “evidence of no harm”.

What happened since (and just before publication). - A study in children already existed. CEFALO (Aydin et al., JNCI, 27 July 2011) was a case-control study of 7–19-year-olds in four countries. Regular users were “not statistically significantly more likely” to have brain tumours (OR 1.36, 95% CI 0.92–2.02). In a subset with operator records, risk was related to time since first subscription, which the authors did not interpret as causal. The footnote’s “no studies in children” was inaccurate when the report was published in 2013. - MOBI-Kids (Castaño-Vinyals et al., Environ Int, 2022) covered 14 countries, 899 cases aged 10–24 and 1,910 controls. It “provides no evidence of a causal association between wireless phone use and brain tumours in young people”. Many odds ratios were below 1, attributed partly to recall bias, so “we cannot rule out … a small increased risk”. Critics dispute the design (Hardell and Moskowitz, 2023). - COSMOS, a prospective cohort (Feychting et al., Environ Int, 2024), followed 264,574 adults for a median 7.1 years. The hazard ratio per 100 cumulative call-hours was 1.00 (0.98–1.02) for glioma. Critics call it “methodologically-flawed” (Moskowitz et al., 2024). - The WHO-commissioned systematic review (Karipidis et al., Environ Int, 2024) rated, with GRADE, “moderate certainty evidence” that head exposure from mobile-phone use “likely does not increase the risk of glioma, meningioma, acoustic neuroma, pituitary tumours, and salivary gland tumours in adults, or of paediatric brain tumours”. - IARC’s position is open. The 2011 Group 2B classification stands. IARC’s 2024 Advisory Group gave radiofrequency radiation high priority for re-evaluation in 2025–29, citing “New human cancer and animal cancer evidence”. I found no scheduled meeting as of September 2026. - The mobile-phone case is covered in its own chapter (Ch. 21). This is only a check of the footnote’s use of it.

Verdict: weakened. As an illustration of the “no evidence of harm” error, the example has aged badly. The factual premise was wrong at publication, and the large studies since, including one in young people, have found no clear association. The general lesson it was meant to illustrate is not affected. IARC’s pending re-evaluation means the question is not formally closed.

Weight. Low for this example. It is a caution about the chapter’s method: lessons from a well-evidenced case (lead) were extended to a live controversy without weighing the evidence, a risk the digest already flagged. When “absence of evidence is not evidence of absence” is invoked, the lens should ask whether studies designed to find the harm have since been done, and what they found.

Other factual checks#

• US phase-out date (p. 62): the on-road ban took effect on 1 January 1996, not 1995, with exemptions for aircraft, racing, farm and marine engines (US EPA, 1996).
• “European Food Standards Agency” (Boxes 3.11 and 3.12): should be the European Food Safety Authority. The PTWI withdrawal is attributed to JECFA in the chapter’s own source (WHO, 2010).
• Box 3.10 is taken nearly verbatim from WHO (2010), including the “5–6 points” and “99%” figures. The two sentences on 16% and on EFSA are printed in both Box 3.11 and Box 3.12.
• Octel conviction date (p. 68): see claim 8. The digest had flagged it; it is confirmed wrong.
• Lanphear et al. (2005): errors in the Boston cohort data were corrected in 2019 without changing the conclusions (claim 1).
• The pre-2013 behaviour literature (Marcus et al., 2010; Wright et al., 2008) was stronger support than the WHO (2010) citation the chapter used.

Implications for the section’s transferable insights#

Additional lessons hindsight suggests. Each is framed neutrally and tied to this case. - Temporary exemptions last. A carve-out granted while a use is phased out can outlast the phase-out by decades. The durability comes from a sole supplier, a small but organised user base and certification hurdles (pp. 62, 68; claim 3). - Substitutes need their own assessment. Removing one hazard can install another if the replacements are adopted without comparable scrutiny (Panel 3.1; claim 7). - Protections set by executive policy are reversible. Science-integrity protections created by executive policy were reversed within four years. Legislated protections have lasted but are narrowly drawn (claim 9). - Dramatic social-outcome findings are prone to publication bias. This holds even when the underlying biology is sound (claim 4).

Method and access notes#

• Tools. The web-search budget for this session was exhausted before this check began. Sources were found through the Europe PMC, OpenAlex and Crossref APIs, the EU Publications Office (Cellar and SPARQL), the Federal Register API, the GOV.UK content API and SEC EDGAR, and read by direct fetches of primary pages and PDFs.
• Unreachable sources. I could not open the Guardian article cited by the chapter, the DOJ press release on Innospec (bot-protected), or EUR-Lex HTML pages. EU acts were read in full from the Publications Office.
• SFO press release. The 2014 release was read from an Internet Archive snapshot of the SFO page.
• Not read beyond abstract, title or summary. EPA’s 2024 lead ISA appendices; the 2024 IARC priorities paper (read via a summary of its table); Huang et al. (2023); Splitter et al. (2016); Tsai and Hatfield (2011, cited via UNEP).
• Not checked. Implementation after 2025 of US drinking-water and dust-lead rules; any ex post study of the “oil penalty”.
• Strand boundary. No material outside Strand A was consulted.

Sources#

Dose-response, thresholds and reference values - Ruckart PZ, Jones RL, Courtney JG, et al. (29 Oct 2021). Update of the blood lead reference value — United States, 2021. MMWR 70(43): 1509–1512. https://doi.org/10.15585/mmwr.mm7043a4 - EFSA Panel on Contaminants in the Food Chain (Apr 2010). Scientific opinion on lead in food. EFSA Journal 8(4): 1570. https://doi.org/10.2903/j.efsa.2010.1570 - WHO (2010). Childhood lead poisoning. Geneva. https://iris.who.int/handle/10665/136571 - Lanphear BP, Hornung R, Khoury J, et al. (2005). Low-level environmental lead exposure and children’s intellectual function: an international pooled analysis. Environ Health Perspect 113: 894–899. https://doi.org/10.1289/ehp.7688 - Lanphear BP, et al. (Sep 2019). Erratum: “Low-level environmental lead exposure and children’s intellectual function”. Environ Health Perspect 127(9): 099001. https://doi.org/10.1289/EHP5685 - Crump KS, Van Landingham C, Bowers TS, Cahoy D, Chandalia JK (Sep 2013). A statistical reevaluation of the data used in the Lanphear et al. (2005) pooled-analysis. Crit Rev Toxicol 43(9): 785–799. https://doi.org/10.3109/10408444.2013.832726 - WHO (updated 10 Jun 2026). Lead poisoning (fact sheet). https://www.who.int/news-room/fact-sheets/detail/lead-poisoning-and-health - Public Health England (5 Jul 2021). Evaluation of whether to lower the public health intervention concentration for lead exposure in children. https://assets.publishing.service.gov.uk/media/60df32488fa8f50ab4035d29/Evaluation_of_whether_to_lower_the_public_health_intervention_concentration_for_lead_exposure_in_children.pdf - Directive (EU) 2024/869 of 13 March 2024 amending Directives 2004/37/EC and 98/24/EC as regards the limit values for lead and its inorganic compounds and for diisocyanates. OJ L, 19.3.2024. https://eur-lex.europa.eu/eli/dir/2024/869/oj - US EPA (Jan 2024). Integrated Science Assessment for Lead (EPA/600/R-23/375). https://www.epa.gov/isa/integrated-science-assessment-isa-lead - IARC Monographs (23 Jul 2026). Meeting 144 “Metallic lead and inorganic lead compounds; cadmium and cadmium compounds” is announced (meeting 19–26 Oct 2027). https://monographs.iarc.who.int/news-events/meeting-144-metallic-lead-and-inorganic-lead-compounds-cadmium-and-cadmium-compounds-is-announced/ - Lanphear BP, Rauch S, Auinger P, Allen RW, Hornung RW (Mar 2018). Low-level lead exposure and mortality in US adults: a population-based cohort study. Lancet Public Health 3: e177–e184. https://doi.org/10.1016/S2468-2667(18)30025-2

Benefits, attribution and blood lead trends - Grosse SD, Matte TD, Schwartz J, Jackson RJ (Jun 2002). Economic gains resulting from the reduction in children’s exposure to lead in the United States. Environ Health Perspect 110: 563–569. https://doi.org/10.1289/ehp.02110563 - Egan KB, Cornwell CR, Courtney JG, Ettinger AS (Mar 2021). Blood lead levels in U.S. children ages 1–11 years, 1976–2016. Environ Health Perspect 129(3): 037003. https://doi.org/10.1289/EHP7932 - McFarland MJ, Hauer ME, Reuben A (Mar 2022). Half of US population exposed to adverse lead levels in early childhood. PNAS 119(11): e2118631119. https://doi.org/10.1073/pnas.2118631119 - Grönqvist H, Nilsson JP, Robling P-O (2020). Understanding how low levels of early lead exposure affect children’s life trajectories. J Political Economy 128(9): 3376–3433. https://doi.org/10.1086/708725 - Aizer A, Currie J (2019). Lead and juvenile delinquency: new evidence from linked birth, school, and juvenile detention records. Rev Econ Stat 101(4): 575–587. https://doi.org/10.1162/rest_a_00814 - Hollingsworth A, Rudik I (Aug 2021). The effect of leaded gasoline on elderly mortality: evidence from regulatory exemptions. AEJ: Economic Policy 13(3): 345–373. https://doi.org/10.1257/pol.20190654 - Ericson B, Dowling R, Dey S, et al. (2018). A meta-analysis of blood lead levels in India and the attributable burden of disease. Environ Int 121: 461–470. https://doi.org/10.1016/j.envint.2018.08.047 - Liu X, Zhang H, Zhang Y, et al. (2023; online Jun 2022). The time trend of blood lead and cadmium levels in rural Chinese children: China Nutrition and Health Survey 2002 and 2012. Biol Trace Elem Res 201: 2162–2169. https://doi.org/10.1007/s12011-022-03333-0

Global phase-out and remaining leaded fuel - UNEP (30 Aug 2021). Era of leaded petrol over, eliminating a major threat to human and planetary health. https://www.unep.org/news-and-stories/press-release/era-leaded-petrol-over-eliminating-major-threat-human-and-planetary - US EPA (29 Jan 1996). EPA takes final step in phaseout of leaded gasoline (press release). https://www.epa.gov/archive/epa/aboutepa/epa-takes-final-step-phaseout-leaded-gasoline.html - US EPA (20 Oct 2023). Finding that lead emissions from aircraft engines that operate on leaded fuel cause or contribute to air pollution that may reasonably be anticipated to endanger public health and welfare. Federal Register. https://www.federalregister.gov/documents/2023/10/20/2023-23247/finding-that-lead-emissions-from-aircraft-engines-that-operate-on-leaded-fuel-cause-or-contribute-to - US EPA. Regulations for lead emissions from aircraft (updated 8 May 2026). https://www.epa.gov/regulations-emissions-vehicles-and-engines/regulations-lead-emissions-aircraft - Federal Aviation Administration. Building an unleaded future by 2030 (last updated 30 Jul 2026). https://www.faa.gov/unleaded - Innospec Inc. Form 10-K for fiscal year 2025 (filed 18 Feb 2026). https://www.sec.gov/Archives/edgar/data/1054905/000119312526056502/iosp-20251231.htm - Hilton FGH (2001). Later abatement, faster abatement: evidence and explanations from the global phaseout of leaded gasoline. J Environ Dev 10(3): 246–265. https://doi.org/10.1177/10704965-0101003-03

Lead and antisocial behaviour or crime - Marcus DK, Fulton JJ, Clarke EJ (2010). Lead and conduct problems: a meta-analysis. J Clin Child Adolesc Psychol 39(2): 234–241. https://doi.org/10.1080/15374411003591455 - Wright JP, Dietrich KN, Ris MD, et al. (27 May 2008). Association of prenatal and childhood blood lead concentrations with criminal arrests in early adulthood. PLoS Med 5(5): e101. https://doi.org/10.1371/journal.pmed.0050101 - Beckley AL, Caspi A, Broadbent J, et al. (26 Dec 2017; print 2018). Association of childhood blood lead levels with criminal offending. JAMA Pediatrics 172(2): 166. https://doi.org/10.1001/jamapediatrics.2017.4005 - Higney A, Hanley N, Moro M (Nov 2022). The lead-crime hypothesis: a meta-analysis. Regional Science and Urban Economics 97: 103826. https://doi.org/10.1016/j.regsciurbeco.2022.103826 - Talayero MJ, Robbins CR, Smith ER, Santos-Burgoa C (Aug 2023). The association between lead exposure and crime: a systematic review. PLOS Glob Public Health 3(8): e0002177. https://doi.org/10.1371/journal.pgph.0002177 - Shaffer RM, Carlson LM, Hester K, et al. (online Sep 2025; print Nov 2025). Lead exposure and antisocial behavior: a systematic review of human and animal evidence. Environ Int 205: 109786. https://doi.org/10.1016/j.envint.2025.109786

Global exposure, legacy sources, e-waste - UNICEF and Pure Earth (Jul 2020). The toxic truth: children’s exposure to lead pollution undermines a generation of future potential. https://www.unicef.org/reports/toxic-truth-childrens-exposure-to-lead-pollution-2020 - Ericson B, Hu H, Nash E, et al. (Mar 2021). Blood lead levels in low-income and middle-income countries: a systematic review. Lancet Planet Health 5: e145–e153. https://doi.org/10.1016/S2542-5196(20)30278-3 - Larsen B, Sánchez-Triana E (online Sep 2023; print Oct 2023). Global health burden and cost of lead exposure in children and adults: a health impact and economic modelling analysis. Lancet Planet Health 7: e831–e840. https://doi.org/10.1016/S2542-5196(23)00166-3 - WHO (15 Jun 2021). Soaring e-waste affects the health of millions of children, WHO warns. https://www.who.int/news/item/15-06-2021-soaring-e-waste-affects-the-health-of-millions-of-children-who-warns - ITU/UNITAR (Nov 2024). The Global E-waste Monitor 2024. https://ewastemonitor.info/the-global-e-waste-monitor-2024/ - Huang J, Zeng Z, Xu X, et al. (2023). Blood lead levels of children exposed to e-waste: a systematic review and meta-analysis. Environ Sci Pollut Res. https://doi.org/10.1007/s11356-023-27114-x (title only consulted) - Forsyth JE, Baker M, Nurunnahar S, et al. (Jun 2023). Food safety policy enforcement and associated actions reduce lead chromate adulteration in turmeric across Bangladesh. Environ Res 232: 116328. https://doi.org/10.1016/j.envres.2023.116328 - UK Health Security Agency (18 Dec 2025). Lead Exposure in Children Surveillance System (LEICSS) annual report, 2025. https://www.gov.uk/government/publications/lead-exposure-in-children-surveillance-reports-from-2021/lead-exposure-in-children-surveillance-system-leicss-annual-report-2025 - Public Health England / UKHSA. Lead exposure in children: surveillance reports (2018–2021 series, including the SLiC study). https://www.gov.uk/government/publications/lead-exposure-in-children-surveillance-reports - Hanna-Attisha M, LaChance J, Sadler RC, Champney Schnepp A (Feb 2016; online Dec 2015). Elevated blood lead levels in children associated with the Flint drinking water crisis. Am J Public Health 106: 283–290. https://doi.org/10.2105/AJPH.2015.303003 - Flint Water Advisory Task Force (21 Mar 2016). Final report. https://www.michigan.gov/-/media/Project/Websites/formergovernors/Folder6/FWATF_FINAL_REPORT_21March2016.pdf - US EPA (30 Oct 2024). National Primary Drinking Water Regulations for Lead and Copper: Improvements (LCRI). Federal Register. https://www.federalregister.gov/documents/2024/10/30/2024-23549/national-primary-drinking-water-regulations-for-lead-and-copper-improvements-lcri - US EPA (12 Nov 2024). Reconsideration of the dust-lead hazard standards and dust-lead post-abatement clearance levels. Federal Register. https://www.federalregister.gov/documents/2024/11/12/2024-25070/reconsideration-of-the-dust-lead-hazard-standards-and-dust-lead-post-abatement-clearance-levels

Costs of regulation - von Storch H, Costa-Cabral M, Hagner C, et al. (2003). Four decades of gasoline lead emissions and control policies in Europe: a retrospective assessment. Sci Total Environ 311: 151–176. https://doi.org/10.1016/S0048-9697(03)00051-2 - Harrington W, Morgenstern RD, Nelson P (2000). On the accuracy of regulatory cost estimates. J Policy Anal Manage 19(2): 297–322. https://doi.org/10.1002/(SICI)1520-6688(200021)19:2<297::AID-PAM7>3.0.CO;2-X - Simpson RD (Jun 2014). Do regulators overestimate the costs of regulation? J Benefit-Cost Anal 5(2): 315–332. https://doi.org/10.1515/jbca-2014-0027 - Morgenstern RD (2018; online Nov 2017). Retrospective analysis of U.S. federal environmental regulation. J Benefit-Cost Anal 9(2): 285–304. https://doi.org/10.1017/bca.2017.17 - Kerr S, Newell RG (Sep 2003). Policy-induced technology adoption: evidence from the U.S. lead phasedown. J Industrial Economics 51(3): 317–343. https://doi.org/10.1111/1467-6451.00203

Alternatives and substitutes - US Energy Information Administration (4 May 2016). Almost all U.S. gasoline is blended with 10% ethanol. https://www.eia.gov/todayinenergy/detail.php?id=26092 - Splitter D, Pawlowski A, Wagner R (Jan 2016). A historical analysis of the co-evolution of gasoline octane number and spark-ignition engines. Front Mech Eng 1: 16. https://doi.org/10.3389/fmech.2015.00016 (abstract only) - Lark TJ, Hendricks NP, Smith A, et al. (Feb 2022). Environmental outcomes of the US Renewable Fuel Standard. PNAS 119(9): e2101084119. https://doi.org/10.1073/pnas.2101084119 - US EPA (1 Apr 2026). Renewable Fuel Standard (RFS) program: standards for 2026 and 2027, partial waiver of 2025 cellulosic biofuel volume requirement, and other changes. Federal Register. https://www.federalregister.gov/documents/2026/04/01/2026-06275/renewable-fuel-standard-rfs-program-standards-for-2026-and-2027-partial-waiver-of-2025-cellulosic - Turner MC, Godderis L, Guénel P, et al. (online 20 Mar 2025; print May 2025). Carcinogenicity of automotive gasoline and some oxygenated gasoline additives. Lancet Oncol 26(5): 548–549. https://doi.org/10.1016/S1470-2045(25)00165-2 - IARC Monographs (23 Jul 2026). Volume 138: Automotive gasoline and some oxygenated additives (news item). https://monographs.iarc.who.int/news-events/volume-138-automotive-gasoline-and-some-oxygenated-additives/

Octel/Innospec - US Securities and Exchange Commission (18 Mar 2010). Litigation Release No. 21454: SEC files settled FCPA charges against Innospec, Inc. https://www.sec.gov/litigation/litreleases/lr-21454 - US Securities and Exchange Commission (18 Mar 2010). Complaint, SEC v. Innospec, Inc. https://www.sec.gov/litigation/complaints/2010/comp21454.pdf - US Department of Justice (18 Mar 2010). Innospec Inc. pleads guilty to FCPA charges and defrauding the United Nations (press release; not accessible, cited from secondary reference). https://www.justice.gov/archives/opa/pr/innospec-inc-pleads-guilty-fcpa-charges-and-defrauding-united-nations-admits-violating-us - UK Serious Fraud Office (4 Aug 2014). Four sentenced for role in Innospec corruption. https://www.sfo.gov.uk/2014/08/04/four-sentenced-role-innospec-corruption/ (read via Internet Archive snapshot of 28 Nov 2022: http://web.archive.org/web/20221128081659/https://www.sfo.gov.uk/2014/08/04/four-sentenced-role-innospec-corruption/)

Recommendations: whistleblowers, transparency, monitoring - Directive (EU) 2019/1937 of 23 October 2019 on the protection of persons who report breaches of Union law. OJ L 305, 26.11.2019. https://eur-lex.europa.eu/eli/dir/2019/1937/oj - European Commission (3 Jul 2024). Report on the implementation and application of Directive (EU) 2019/1937, COM(2024) 269. https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:52024DC0269 - Regulation (EU) 2019/1381 of 20 June 2019 on the transparency and sustainability of the EU risk assessment in the food chain. OJ L 231, 6.9.2019. https://eur-lex.europa.eu/eli/reg/2019/1381/oj - Regulation (EU) 2025/2455 of 26 November 2025 establishing a common data platform on chemicals … and establishing a monitoring and outlook framework for chemicals. https://eur-lex.europa.eu/eli/reg/2025/2455/oj - Ramhøj L, Svingen T, Vanhaecke T (2024). Editorial: European partnership on the assessment of risks from chemicals (PARC). Front Toxicol 6. https://doi.org/10.3389/ftox.2024.1461967 - Executive Order 14303, Restoring Gold Standard Science (signed 23 May 2025). Federal Register, 29 May 2025. https://www.federalregister.gov/documents/2025/05/29/2025-09802/restoring-gold-standard-science - Toxic Docs (Columbia University / CUNY). About. https://www.toxicdocs.org/about (accessed 25 Sep 2026)

Mobile phones (footnote 3) - Aydin D, Feychting M, Schüz J, et al. (27 Jul 2011). Mobile phone use and brain tumors in children and adolescents: a multicenter case-control study (CEFALO). J Natl Cancer Inst 103: 1264–1276. https://doi.org/10.1093/jnci/djr244 - Castaño-Vinyals G, Sadetzki S, Vermeulen R, et al. (2022; online Dec 2021). Wireless phone use in childhood and adolescence and neuroepithelial brain tumours: results from the international MOBI-Kids study. Environ Int 160: 107069. https://doi.org/10.1016/j.envint.2021.107069 - Hardell L, Moskowitz JM (online 2022; print 2023). A critical analysis of the MOBI-Kids study of wireless phone use in childhood and adolescence and brain tumor risk. Rev Environ Health 38: 409–421. https://doi.org/10.1515/reveh-2022-0040 - Feychting M, Schüz J, Toledano MB, et al. (Mar 2024). Mobile phone use and brain tumour risk — COSMOS, a prospective cohort study. Environ Int 185: 108552. https://doi.org/10.1016/j.envint.2024.108552 - Moskowitz JM, Frank JW, Melnick RL, et al. (Aug 2024). COSMOS: a methodologically-flawed cohort study of the health effects from exposure to radiofrequency radiation from mobile phone use (letter). Environ Int 190: 108807. https://doi.org/10.1016/j.envint.2024.108807 - Karipidis K, Baaken D, Loney T, et al. (online Aug 2024; print Sep 2024). The effect of exposure to radiofrequency fields on cancer risk in the general and working population: a systematic review of human observational studies – Part I. Environ Int 191: 108983. https://doi.org/10.1016/j.envint.2024.108983 - Berrington de González A, Masten SA, Bhatti P, et al. (2024). Advisory Group recommendations on priorities for the IARC Monographs. Lancet Oncol 25(5): 546–548. https://doi.org/10.1016/S1470-2045(24)00208-0 ; full text https://pmc.ncbi.nlm.nih.gov/articles/PMC12640648/