Assessment of the Risk of Contamination of Food for Infants and Toddlers

This narrative literature review surveys 83 publications on contaminants in foods for infants and toddlers aged 0.5–3 years, covering toxic elements (Pb, Cd, As, Hg), acrylamide, bisphenol A, dioxins, furan, mycotoxins, nitrates/nitrites, pesticide residues, polycyclic aromatic hydrocarbons, 3-MCPD/glycidyl esters, and mineral oil hydrocarbons. The toxic-elements section (3.1) compiles findings from infant-formula and complementary-food studies across Norway, the USA (Miami), Poland, Brazil, Nigeria, Spain (Menorca), France, Canada, Spain, Columbia, Australia, Sweden, Korea, Tanzania, and Portugal. For HMI purposes, this page reports only the heavy-metals (toxic-elements) content; the non-metal sections are out of scope. The review reproduces values from constituent papers without re-normalizing units, so unit basis (wet versus dry weight, µg/g versus mg/kg versus µg/kg) appears as the originating studies reported it.

Key numbers

Toxic-elements review summary (Section 3.1 and review Table 2). Numbers are reproduced as the review reports them; the cited primary papers are the authoritative source for each value.

Pb and Cd in infant formulas, infant foods, and complementary foods:

• Iwegbue 2010 (Nigeria, n=26 infant food/commercial food, Table 2 ref [14]): 96% of samples exceeded the FDA Pb allowable concentration (0.4 mg/kg) and 53% exceeded the maximum FDA Cd concentration (0.1 mg/kg). Note: the review’s prose paragraph attributes this finding to Igweze 2020 [9], but Table 2 attributes the n=26 / 96%-Pb / 53%-Cd row to ref [14] = Iwegbue 2010; this page follows Table 2’s attribution as the more structured representation and flags the source-internal inconsistency.
• Gardener 2019 (Miami, USA, n=564 infant food, Table 2 ref [10]): Pb contamination in 37% of samples and Cd contamination in 57% of samples; RfD exceeded by 0–3%, Pb and Cd by 6–23%. Review reports no significant difference between organic and conventional cultivation; rice-based products contained significantly more Pb and Cd.
• Winiarska-Mieczan 2013 (Poland, n=114 baby desserts, juices, dinners, Table 2 ref [12]): no sample exceeded the Maximum Consented Limit (MCL) for Pb and Cd.
• De Castro 2010 (Brazil, n=55 infant formula and milk, Table 2 ref [13]): median Pb concentration 0.109 mg/kg, exceeding the FAO/WHO maximum standard (0.02 mg/kg); Pb in 7 infant formulas and 22 milk samples exceeded the Brazilian standard (0.2 mg/kg in infant formula; 0.05 mg/kg in milk).
• Igweze 2020 (Nigeria, n=190 infant formula, Table 2 ref [9]): 1 sample exceeded the PTWI (Pb and Cd).
• Skrbic 2016 (Spain, n=90 infant food, Table 2 ref [11]): Pb detected in 2 samples, As in 3 samples; RfD exceeded for Cd and Pb.

Cd specifically:

• Iwegbue 2010 (Nigeria, ref [14], reported by the review as “In another study”): Cd was detected in 65% of samples; the highest amounts were found in potatoes, cookie cereal bars, vegetables, and bread; 2.5% of children aged 3 years and under exceeded the TWI for Cd. (Note: the food-category list — potatoes, cookie cereal bars, vegetables, bread — is broader than Iwegbue 2010’s commercial-infant-formula scope and may indicate a further source-internal reference error; reproduced as the review attributes it.)
• Gao 2020 (snacks consumed by children 1–6 years): no exceedance of Cd intake standards; EDI did not exceed the MTDI (maximum tolerable daily intake = 46 µg/d); flour products contributed 37% of total Cd delivery from snacks.
• Sweden, n=18 infant food: Cd exceeded in rice-based products.

Hg in infant and toddler foods:

• Guérin/Hulin 2018 (France, n=291 children’s products): no sample exceeded the maximum allowable concentration; Hg detected in prepared fish- and meat-based products (maximum 7.4 µg/kg), in dairy desserts, cereal products, and in one infant formula.
• Junqué 2017 (Menorca, Spain, n=45 commercial food for toddlers, fish products): fish products exceeded the MRL for Hg (0.5–1.0 mg/kg) in 66% of samples.
• Dabeka 2012 (Canada, n=150 infant formula): no sample exceeded the standards.
• Martins 2013 (Portugal, n=87 infant food): no sample exceeded the PTWI.

Long-term exposure (children under 6 years) — Kim 2014 (Korea, n=119 commercial food for toddlers, Pb/Cd/Hg):

• Pb exposure exceeded the reference value (RfT = 0.5 µg/kg bw) among 35% of children.
• 42% of children exceeded the TWI of Cd; Hg exposure was below the TWI.
• Main sources of Pb exposure: dairy products and vegetables.
• Main sources of Cd and Hg exposure: cereal products, fish, and shellfish.
• Review notes (citing Dabeka 2012 and Martins 2013) that two analyzed studies suggest infant formulas are safe in terms of Pb content.

Arsenic in infant foods:

• Gu 2020 (Australia, n=39 commercial food for toddlers): 75% of samples contaminated with inorganic As; exceeded EU maximum levels. Highest As concentrations in rice noodles, whole grain rice, and crackers.
• Columbia, n=48 commercial food for toddlers: 91% of samples contained inorganic As; 100% exceeded total As reference level (review wording).
• Rotenberg 2017 and Ljung 2011 also reported high As contamination (review reference [22] and [23]).
• Melø 2008 (Norway, n=76 infant formula and milk): no sample was contaminated for Pb, Cd, As, or Hg. Oatmeal had the highest concentration of Cd (1.02 µg/g); fish had the highest concentrations of As (1.02 µg/g) and Hg (6 µg/g). None of the products contained amounts of As that would pose a health risk per the review.
• Tanzania, n=43 commercial food for toddlers (Cd and Hg): no sample exceeded the PTWI.

Safe-level reference values cited by the review (Table 1, attributed to references [3,4,5,6,7,8]):

• Arsenic: PTWI = 15 µg/kg/bw.
• Cadmium: MTD = 4.1 µg/day; PTWI = 7 µg/kg/bw.
• Lead: MTD = 0.5 µg/day; PTWI = 25 µg/kg/bw.
• Mercury: PTWI = 1.6 µg/kg/bw.

(Table 1 also lists safe levels for acrylamide, bisphenol A, dioxins, furan, mycotoxins, nitrates/nitrites, and pesticide residues; those are non-metal contaminants and out of HMI scope.)

Methods (brief)

Narrative literature review. PRISMA-style flow with the authors’ own modifications. Search databases: PubMed (primary), Google Scholar, and Scopus. Search window: October 2020 – March 2021. Date range of included publications: 2004–2021. Search terms combined “contamination”, “baby food”, “infant formula”, “toxic elements”, “dioxins”, “acrylamide”, “bisphenol”, “furan”, “mycotoxins”, “nitrates”, “polycyclic aromatic hydrocarbon”, “pesticides”, “3-MCPD”, “glycidyl esters”, and “mineral oil hydrocarbons” under the headings “food safety”, “food contamination”, and “environmental contaminants”. Identification: n=5,922 from databases + n=3,236 from other sources. After deduplication: 6,278 records. Screening: 4,298 titles; 4,152 excluded. Eligibility: 146 full-text articles assessed. Inclusion: 83 studies with full access. Exclusion reasons reported: focused not on the purpose (n=19), too small a sample size/study group (n=19), unsuitable age group (n=4), test material being human breast milk (n=3), animal studies (n=3), other (n=15). Studies were excluded if they involved animal measurements, prenatal exposure, adult exposure, contaminant-exposed sick children, ages under 6 months or over 3 years, insufficient sample/study size, lack of full access, were themselves literature reviews, failed to specify products for children, or used inappropriate study material (e.g., human milk). Non-English-language publications were excluded. No formal meta-analysis. No quantitative synthesis of the constituent studies. No instrument-level or analytical-method reporting (the review summarises primary-study findings without describing the methods used in each primary study).

Implications

Certification: The review reproduces occurrence and intake summaries for Pb, Cd, As, and Hg in infant formulas, complementary foods, and snacks across 14 jurisdictions. As a narrative review without re-analysis or unit normalization, it is most useful as a discovery surface pointing to the constituent primary studies for any HMTc threshold-setting work in Category 6 (Infant Foods & Formula). Primary-study attributions (Igweze 2020, Gardener 2019, De Castro 2010, Iwegbue 2010, Gao 2020, Guérin/Hulin 2018, Junqué 2017, Kim 2014, Gu 2020, Rotenberg 2017, Ljung 2011, Melø 2008, Dabeka 2012, Martins 2013, Bielecka 2020) should be ingested individually for occurrence-data use; this page should not be pooled into the Cat 6 standards workbench as a primary source.

Courses: Useful as a single-paper survey of the heavy-metals literature on infant/toddler foods circa 2021, particularly for the Polish-authored framing of regulatory reference values (FDA RfD, FAO/WHO MCL, EFSA TWI/PTWI) applied to baby-food categories. The discussion section’s framing of why infants and toddlers are particularly vulnerable (higher intake per kg body weight, immature urinary/biliary systems, faster gastric emptying, higher absorption) is teachable as background.

App: Not directly applicable — the review does not provide product-level concentration distributions usable by the ingredient-list inference model.

Microbiome: Not applicable — the review does not address microbiome outcomes.

Wiki pages this source may touch

• baby-food
• infant-formula-powder
• infant-rice-cereal
• rice
• fish
• infant-formula-powder
• infant-food-general
• baby-cereals-dry-rice-based
• baby-cereals-dry-non-rice
• teething-and-snacks-rice-based
• fish-containing-baby-foods
• lead
• cadmium
• arsenic-total
• arsenic-inorganic
• mercury-total

Verification notes

• 2026-06-09 (Manual Fetch Discovery ingest, Claude Opus 4.7 v2 skill, fresh path): wrote source page from the PDF in raw/Manual Fetch Discovery/. Three identity checks (DOI 10.3390/nu13072358, raw_handle MFD_mielech2021, cite-key author/year) all returned no hit; this is a fresh ingest.
• 2026-06-09 (auto-audit application, fresh-context subagent verdict QUARANTINE → revised here): the audit subagent caught three numerical-fidelity errors in the initial ingest. Each was independently verified against the source and corrected:
  • Igweze ↔ Iwegbue Nigeria attribution swap: the initial ingest assigned the 96%/53% Pb/Cd finding to Igweze 2020 [9] and the “1 sample exceeded PTWI” finding to Iwegbue 2010 [14], following the review’s prose paragraph in Section 3.1. Re-reading the source: Table 2 (pp. 5–6) shows the n=26 row with 96%/53% pointing to ref [14] = Iwegbue and the n=190 row with “1 sample exceeded PTWI” pointing to ref [9] = Igweze. The review’s own prose and Table 2 disagree on which reference the 96%/53% finding belongs to. The page now follows Table 2’s structured attribution and flags the source-internal contradiction transparently in the Key numbers bullet.
  • Melø 2008 As/Hg transposition: the initial ingest reported As=6 µg/g and Hg=1.02 µg/g for fish. The review states “fish had the highest concentrations of As and Hg (1.02 µg/g, 6 µg/g, respectively) [24]”, where “respectively” maps to the “As, Hg” order: As=1.02 µg/g, Hg=6 µg/g. Corrected. (The 6 µg/g Hg figure is very high for fish and likely reflects either a unit issue in the original Melø 2008 paper or a transcription decision by the review authors; this page reports the value as the review states it without re-interpretation, per the review’s own no-re-normalization posture.)
  • Cd 65% misattribution to Gardener 2019: the initial ingest attributed the “Cd was detected in 65% of samples; potatoes, cookie cereal bars, vegetables, bread; 2.5% > TWI” finding to Gardener 2019. Source p. 4 attributes this to ref [14] = Iwegbue 2010, not Gardener (ref [10]). Re-attributed to Iwegbue 2010 and flagged that the food-category list does not fit Iwegbue 2010’s commercial-infant-formula scope, suggesting a further source-internal reference error that we reproduce faithfully rather than silently correct.
  • Pushback on Check 1 ⚠️ items (author attributions for Winiarska-Mieczan, Skrbic, Dabeka, Martins): the audit flagged these as “added by wiki without source basis,” but each attribution is verifiable through the review’s reference list (Section “Reference”, pp. 21–22) cross-referenced with the bracketed citation numbers in Section 3.1 prose and Table 2: ref [11] = Skrbic 2016 (Table 2 Spain n=90 row), ref [12] = Winiarska-Mieczan 2013 (Table 2 Poland n=114 row), ref [19] = Dabeka 2012 (Table 2 Canada n=150 row), ref [20] = Martins 2013 (Table 2 Portugal n=87 row). These attributions stand.
  • Verification-note Rotenberg framing correction: the initial ingest’s verification note called Rotenberg 2017 part of “the non-metal section context”; Rotenberg 2017 (ref [22]) is actually in the heavy-metals section of the review as a study of co-exposure to MeHg and iAs in rice-based infant cereals. The speciation-flagging verification note has been corrected.
• Scope decision: this is a narrative review covering both heavy metals (in scope for HMI) and non-metal contaminants (acrylamide, bisphenol, dioxins, furan, mycotoxins, nitrates/nitrites, pesticides, PAHs, 3-MCPD, MOH — out of scope). Key numbers extracts only the toxic-elements content (Section 3.1, Table 2, and the relevant Discussion paragraphs). The non-metal sections (3.2–3.11) are noted in the lead paragraph but not extracted to Key numbers.
• Speciation flagging: the review uses “As” and “Hg” mostly without speciation labels. Where the constituent study measured inorganic arsenic (Gu 2020 Australia, Columbia n=48 study, Rotenberg 2017 rice-cereal MeHg+iAs co-exposure) iAs is recorded in metals:. Where the cited study measured total or unspeciated arsenic (Melø 2008 Norway, Skrbic 2016 Spain), tAs is recorded. For mercury, the review and its constituent studies are not consistent in distinguishing MeHg from total Hg; conservative tHg is used because the toxic-elements section’s quantitative findings are reported as unspeciated Hg in Guérin/Hulin 2018, Junqué 2017, Kim 2014, Dabeka 2012, and Martins 2013. Rotenberg 2017 is named in the review’s heavy-metals section (ref [22]) as a study of MeHg-and-iAs co-exposure in rice-based infant cereals; only its iAs co-exposure context is summarized here, not the MeHg side.
• Unit basis: the review carries values in mixed bases (mg/kg, µg/kg, µg/g, µg/L, µg/day, µg/kg/bw) as the originating studies reported them. This page does not re-normalize; the constituent primary studies are authoritative for unit basis on each value.
• Brand firewall (Part 12): the source does not name commercial brands in the toxic-elements section; no brand-firewall action required.
• Wiki/HMTc firewall (Part 2): no synthesis claims, no HMTc threshold proposals, no consumer-audience risk advisories. Safe-level reference values (PTWI, RfD, TWI, MTD) are reported as the review’s Table 1 lists them; not converted into wiki-side recommendations.
• The review’s mortality/toxicology Discussion paragraphs (e.g., Pb in infancy disrupts the nervous system; Cd is an IARC carcinogen; inorganic As is a carcinogen) are not extracted to Key numbers because they are background framing rather than primary-data findings; they are summarized briefly in the lead paragraph only.
• Routing input: products: is set to the umbrella infant-formula-powder slug plus broad complementary-food slugs (infant-food-general, baby-cereals dry rice-based/non-rice, teething-and-snacks rice-based, fish-containing-baby-foods). The routing layer will fan these out to sibling pages. The review is broad and does not commit to soy vs non-soy or powder vs RTF for infant formula across its cited studies, so the umbrella slug is correct per Part 5b routing input rules.

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.