UK COT 2013 — Potential risks from aluminium in the infant diet

Summary

The UK Committee on Toxicity of Chemicals in Food, Consumer Products and the Environment (COT) issued Statement 2013/01 at the request of the Scientific Advisory Committee on Nutrition (SACN), which was reviewing dietary recommendations for infants aged 0–12 months and children aged 1–5 years. The Statement collates UK exposure data for aluminium from drinking water (Drinking Water Inspectorate 2011 survey), breast milk (Baxter et al. 1991 UK data plus international comparators), infant formulae (Food Standards Agency 2006 survey of 32 formulae), and commercial infant foods (FSA 2006 survey of 153 samples across 11 food categories), and uses these data to estimate weekly aluminium exposure as a percentage of the JECFA 2011 Provisional Tolerable Weekly Intake (PTWI) of 2 mg/kg bw/week. The Committee concludes that estimated dietary exposures of UK infants — up to ≈39% of the PTWI from infant formula and complementary foods combined, plus a worst-case 12% additional contribution from high-aluminium drinking water used to reconstitute powdered formula — do not indicate a toxicological concern or warrant changes to Government advice, with the principal caveat that soya-based powdered formula produces the highest exposures (21% of the PTWI on its own).

Key numbers

Health-based guidance value applied throughout the Statement:

• JECFA 2011 PTWI: 2 mg Al/kg bw (2000 µg/kg bw) per week, applied to all aluminium compounds in food including food additives. Derived by JECFA from the Poirier et al. 2011 GLP twelve-month neurodevelopmental study in Sprague-Dawley rats (NOAEL 30 mg Al/kg bw/day; LOAEL 100 mg/kg bw/day for grip-strength and renal effects in offspring; uncertainty factor 100) (paragraphs 28–29, p.6). The COT concluded the JECFA PTWI derivation was sound and adopted it for this assessment.
• WHO 2009 caveat: health-based guidance values are not generally applicable to infants under 12 weeks, who may be at risk at lower exposures (paragraph 30, p.7).

Drinking water — Drinking Water Inspectorate (DWI) 2011 survey, England and Wales (n = 42,400), Northern Ireland (n = 1,730), Scotland (n = 5,020) (Table 1, p.8):

• 1st percentile: 0*–1** µg/L (England/Wales), 4 µg/L (Northern Ireland), 4 µg/L (Scotland).
• 25th percentile: 0*–6** µg/L (E/W), 19 µg/L (NI), 16 µg/L (Scotland).
• Mean: 9.0*–11** µg/L (E/W), 38 µg/L (NI), 31 µg/L (Scotland).
• Median: 6.6*–7.3** µg/L (E/W), 28 µg/L (NI), 25 µg/L (Scotland).
• 75th percentile: 13 µg/L (E/W), 45 µg/L (NI), 38 µg/L (Scotland).
• 99th percentile: 55 µg/L (E/W), 205 µg/L (NI), 134 µg/L (Scotland).
• EC Directive 98/83/EC indicator parameter value: 200 µg Al/L (paragraph 34, p.8).
• E/W values reported as ranges spanning lower- and upper-bound LOD treatments (* = LOD treated as zero; ** = LOD treated as LOD); NI uses half-LOD; samples below LOD are explicit in source-table notes.

Breast milk (Table 2, p.9):

• Australia (Weintraub et al. 1986): mean 30 µg Al/L (n and median not reported).
• UK (Baxter et al. 1991): n = 8; mean 27 µg/L; median 15 µg/L; range 3–79 µg/L.
• Spain (Fernandez-Lorenzo et al. 1999): n = 45; mean 23.9 ± 9.6 µg/L; median 25.0 µg/L; range 7–42 µg/L.
• Austria (Krachler et al. 2000): n = 27; median 67 µg/L; range <10–380 µg/L.
• Morocco (Zaida et al. 2007): n = 396; mean 17.3 ± 13.9 µg/L; range 1.3–62.2 µg/L.

Infant formulae — FSA 2006 UK survey, “as sold” basis (Table 3, p.10):

• All formulae (n = 32): mean 899 µg Al/kg as sold; range <100–2423 µg/kg.
• Cows’ milk-based (n = 27): mean 817 µg/kg; range <100–2423 µg/kg.
• Goats’ milk-based (n = 3): mean 878 µg/kg; range 235–1412 µg/kg.
• Soya-based (n = 2): mean 2027 µg/kg; range 1808–2246 µg/kg.
• Ready-to-consume cows’ milk-based (n = 14): mean 84 µg/kg; range <17–162 µg/kg.

Commercial infant foods — FSA 2006 UK survey, “as sold” basis (Table 4, p.10):

• Baby rice (n = 8): mean 780 µg/kg; range 203–2284.
• Biscuits (n = 8): mean 4571 µg/kg; range 1021–19715.
• Breakfast foods (n = 27): mean 1024 µg/kg; range 71–4288.
• Cereal bars / rice cakes (n = 9): mean 7205 µg/kg; range 834–25388.
• Desserts (n = 12): mean 1690 µg/kg; range 127–12744.
• Fish (n = 7): mean 470 µg/kg; range 129–1475.
• Fruit puree (n = 7): mean 1529 µg/kg; range 402–4406.
• Meat (n = 45): mean 1074 µg/kg; range 113–4097.
• Pasta / dairy (n = 16): mean 914 µg/kg; range 95–3928.
• Rusks (n = 7): mean 2612 µg/kg; range 763–4455.
• Vegetables (n = 7): mean 1419 µg/kg; range 111–4621.

Estimated dietary exposure — breast milk only (Table 5, p.11), µg/kg bw/week:

• Mean Al in breast milk (27 µg/L): 26 (0–4.0 mo, 800 mL/day), 38 (0–4.0 mo, 1200 mL/day), 19 (>4.0–6.0 mo, 800 mL/day), 29 (>4.0–6.0 mo, 1200 mL/day).
• Maximum Al in breast milk (79 µg/L): 75 / 112 / 57 / 85 µg/kg bw/week, same age × consumption combinations.
• Bodyweights used: 5.9 kg (0–4.0 mo, from DH 1994); 7.8 kg (>4.0–6.0 mo, from DNSIYC 2013).

Estimated dietary exposure — infant formula only (Table 6, p.12), µg/kg bw/week, excluding reconstitution water:

• Powdered cows’ milk-based (mean 106 µg/L as prepared): 101 / 151 / 76 / 114 across the four age × consumption combinations.
• Powdered goats’ milk-based (148 µg/L): 140 / 211 / 106 / 159.
• Powdered soya-based (293 µg/L): 278 / 417 / 210 / 316.
• Ready-to-consume cows’ milk-based (58 µg/L): 55 / 83 / 42 / 62.

Possible additional exposure from drinking water used to reconstitute powdered formula (Table 7, p.13), µg/kg bw/week (water accounts for ~85% of reconstituted formula volume; 680 and 1020 mL/day used):

• Median E/W water (7.3 µg/L): 5.9 / 8.8 / 4.5 / 6.7.
• 99th-percentile E/W water (55 µg/L): 44 / 67 / 34 / 50.
• 99th-percentile NI water (205 µg/L): 165 / 248 / 125 / 188.

Total dietary exposure from infant formula + commercial infant foods (paragraph 45, Table 8 p.14), µg/kg bw/week, from FSA 2006 data using manufacturer feeding guidelines (ages 0–3, 4–6, 7–9, 10–12 mo):

• 2003 survey normal diet: 98 / 994 / 1225 / 1239.
• 2003 survey soya diet: 574 / 1694 / 1554 / 1526.
• 2006 survey normal diet: 104 / 200 / 424 / 776.
• 97.5th-percentile exposure at age 7–12 mo (1986 consumption data + 1997–1999 FSA survey): 686 µg/kg bw/week; same age band using 2003 survey data: 532 µg/kg bw/week.

PTWI utilisation in the conclusions (paragraphs 49–51, 58–61, pp.15–17):

• Exclusively breastfed infants: up to 6% of PTWI (highest exposure in high-level consumers aged 0–3 months on maximum breast-milk Al concentration).
• Exclusively ready-to-consume formula-fed infants: ≈4% of PTWI.
• Exclusively powdered formula-fed infants: 8% (cows’ milk-based), 11% (goats’ milk-based), 21% (soya-based) of PTWI from formula alone.
• Water used in reconstitution: up to a further 12% of PTWI under worst-case 99th-percentile Northern Ireland water (205 µg/L).
• Combined formula + reconstitution-water worst case: up to 33% of PTWI for soya-based powdered formula reconstituted with 99th-percentile NI water in exclusively formula-fed infants aged 0–3 mo (≈665 µg/kg bw/week).
• Mean exposure from infant formula + commercial infant foods (2004–2005 FSA data): <776 µg/kg bw/week (≈39% of PTWI).
• Comparison reference (Bishop et al. 1997; Fewtrell et al. 2009): 315 µg/kg bw/week intravenous Al from standard parenteral solutions in neonates was associated with reduced lumbar-spine and hip bone mass in adolescence; oral worst-case dietary estimate (665 µg/kg bw/week) is ≈2× this parenteral dose, but with ≈0.5% oral absorption versus 100% parenteral, the margin between absorbed dietary Al and the parenteral dose linked to bone effects is >95-fold.

Other points of reference cited:

• Aluminium chlorohydrate dermal absorption from antiperspirants: 0.25 µg/day, ≈2.5% of a 10 mg dietary intake (Flarend et al. 2001) — not relevant to infants (paragraph 31, p.7).
• Council Directive 98/83/EC: 200 µg Al/L indicator parameter for drinking water (paragraph 34, p.8).
• Commission Regulation (EU) 10/2011: Al permitted as additive in plastic food contact materials (paragraph 33, p.8).
• Total of 0.04 mg/day from atmospheric inhalation in the general population (EFSA 2008; WHO 1997) (paragraph 31, p.7).
• Lowermoor 1988 water-pollution incident (Camelford, North Cornwall): 20 tonnes of aluminium sulphate accidentally introduced; ≈20,000 people exposed; COT subgroup (COT 2013) concluded delayed or persistent harm to health unlikely but residual uncertainty for those who were infants at the time (paragraphs 25–26, p.6).

Methods (brief)

Statement integrates four UK data streams and the international peer-reviewed and regulatory literature on aluminium toxicokinetics and toxicology. UK exposure inputs: (i) Drinking Water Inspectorate water-quality data for 2011, aggregating samples reported by 29 water companies serving England and Wales, plus Scottish Government data and Northern Ireland Water data, with company-by-company variation in limits of detection acknowledged in source-table footnotes; (ii) Food Standards Agency 2006 survey of metals in foods and formulae for infants, providing the n=32 infant-formula and n=153 commercial-infant-food concentration data; (iii) Baxter et al. 1991 UK breast-milk study (n=8) for UK-specific breast-milk concentrations, supplemented by published Australian, Spanish, Austrian, and Moroccan datasets; (iv) Department of Health DNSIYC 2013 (Diet and Nutrition Survey of Infants and Young Children, 2011) for infant bodyweights (>4.0–6.0 mo) and DH 1994 COMA report for younger infants (0–4.0 mo). Toxicological inputs: EFSA 2008 Scientific Opinion on Al safety; JECFA 2007 and 2011 evaluations (the 2011 evaluation establishing the 2 mg/kg bw/week PTWI from Poirier et al. 2011); ATSDR 2008 Toxicological Profile. Dietary exposure to Al was calculated as (mean Al concentration × daily consumption volume × 7) ÷ bodyweight, with exposure volumes adjusted to 680/1020 mL/day (85% of 800/1200 mL/day) for reconstitution-water contributions. The Statement makes explicit that absorption was not modelled — all exposures are external (intake) doses compared against the PTWI on an external-dose basis.

Implications

Certification. Provides the operative UK regulatory baseline for aluminium in infant formula and complementary foods through 2013 (FSA 2006 survey data). For aluminium contributions to HMTc threshold work on infant formula, the soya-based powdered formula data point (mean 2027 µg/kg as sold; n=2) is the highest-concentration product format documented in the UK survey, with the powdered cow’s milk-based mean at 817 µg/kg as sold and ready-to-consume cow’s milk-based at 84 µg/kg as sold representing the lower bound. The complementary-foods data identify cereal bars / rice cakes (mean 7205 µg/kg, n=9) and biscuits (mean 4571 µg/kg, n=8) as the highest-Al baby-food categories, with rusks (2612 µg/kg) and desserts (1690 µg/kg) following. Per the Statement’s own analysis (paragraph 50), even unusually high public-water aluminium concentrations are unlikely to push total infant Al intake above the PTWI; only the Northern Ireland 99th percentile (205 µg/L) approached the EC indicator parameter of 200 µg/L in the 2011 DWI data.

Courses. Worked example of regulatory exposure calculation: how a tolerable-weekly-intake (PTWI) reference value is used to characterise risk from a complex dietary mixture (formula + complementary foods + reconstitution water). The Statement illustrates the standard practice of comparing total estimated intake against a percentage of the PTWI rather than against a single binding source. Also useful for showing how UK regulatory bodies frame the residual uncertainty for the 0–12-week age band where health-based guidance values are formally not applicable.

App. Adds UK 2006 commercial-formula and infant-food concentration data for aluminium across multiple infant product categories. The soya-based-powder vs cow’s-milk-powder vs ready-to-consume gradient (2027 / 817 / 84 µg/kg as sold) is a clear product-format contamination signal for any aluminium occurrence pool. The complementary-foods rankings (cereal bars and biscuits highest; meat, pasta-dairy, vegetables mid-range; fish lowest at 470 µg/kg) extend the format gradient into solid-food categories.

Microbiome. No direct microbiome endpoints; Statement does not discuss Al–gut-microbiota interactions.

Provenance notes

Statement is published under the UK Open Government Licence by the UK Committee on Toxicity of Chemicals in Food, Consumer Products and the Environment, an independent scientific advisory committee. Cite-key follows the institutional-author convention used elsewhere in the corpus (e.g., efsa-aluminium-afc-2008). The Statement number is COT 2013/01; published June 2013. Original retrieval through Kimi agent corruption-folder workflow (May 2026); raw_path reflects the Kimi-fetch folder structure. Currently published-canonical access URL maintained by the UK Food Standards Agency (cot.food.gov.uk); when retrieving for verification, the COT subdomain may redirect to the FSA archive. The original FSA cot.food.gov.uk footnote URLs in the Statement are now archived to the National Archives webarchive in places.

Wiki pages this source may touch

• aluminum
• breastmilk
• soy-based-infant-formula
• non-soy-infant-formula
• infant-formula-ingredients
• water
• infant-rice-cereal
• fruit-purees
• meat
• fish
• infant-formula-powder-non-soy
• infant-formula-powder-soy-based
• infant-formula-rtf-liquid-non-soy
• baby-cereals-dry-rice-based
• breakfast-cereals
• biscuits
• fruit-purees
• meat-and-poultry-purees
• fish-containing-baby-foods
• efsa-aluminium-twi
• jecfa-aluminium-ptwi-2011 (missing-slug proposal — no regulation page yet for the JECFA 2011 PTWI revision; see scheer2017 verification notes)

Verification notes

• 2026-06-01: Fresh ingest via v2.0 manual-fetch autonomous skill (cot2013-uk-aluminium-infant-diet).
• Subagent audit (2026-06-01) verdict REVISE; two findings handled:
  • Check 2 ❌ flagged [[products/biscuits]] as missing from the 2026-05-18 taxonomy snapshot, suggesting retarget to snacks-crackers-biscuits. Verified false positive: wiki/products/biscuits.md exists in the current wiki as a provisional scaffold created 2026-05-19 (post-snapshot, via heal-gaps from original2020-original-a declaration). The routing audit routes this source to products/biscuits cleanly as locked_hmtc_row direct_evidence. No change applied to the wiki page; the taxonomy snapshot itself is stale and should be regenerated separately.
  • Check 5 ⚠️ flagged “Drinking water rarely binds for infant Al exposure” in the Certification block as an HMTc-mechanics paraphrase that overreaches paragraph 50. Verified concern correct: paragraph 50 states “even unusually high concentrations of aluminium in public water are unlikely to take total intake above the PTWI.” Implications/Certification paragraph revised to attribute the claim to paragraph 50 directly and to use the source’s own wording.
• Checks 1, 3, 4 returned ✅ clean; numerical fidelity verified across Tables 1–8 and conclusion paragraphs 49–51 / 58–61.

Page history

The five most recent substantive edits to this page. The full version history lives in git; when DOI minting comes online (see schema docs), each entry below will also link to a version-pinned DataCite DOI.