Zmudzinska et al. 2022 — Heavy Metals in Ready-to-Eat Baby Foods, Poland

Polish survey assessing As, Cd, Hg, and Pb in 397 ready-to-eat products marketed for children aged 0.5–3 years, sampled across six product categories from Polish brick-and-mortar and online retail between December 2020 and September 2021. The study reports per-category means, maxima, exceedance counts against EU baby-food limits, and exposure estimates (EDI, EWI, PTWI%, BMDL%, THQ, CR) by age subgroup. Pb exceeded the applicable EU limit in 4.53% of samples and Hg in 1.5%; As and Cd showed no exceedances.

Key numbers

Whole-sample means (n=397, µg/kg wet weight):

Category-level means (µg/kg wet weight; X ± SD):

EU maximum-limit exceedances (Table 3 in source):

As and Cd: zero exceedances reported across all categories. <LOD counts: As 69, Cd 47, Pb 10, Hg “the same samples” (Hg <LOD count not numerically reported).

Risk assessment (Tables 5–6 in source; total cohort n=397):

• Pooled EDI (µg/day): tAs 0.162 ± 0.598; Cd 0.224 ± 0.439; tHg 0.314 ± 0.438; Pb 1.006 ± 1.192
• Pooled EWI (µg/week): tAs 1.139 ± 4.196; Cd 1.571 ± 3.076; tHg 2.204 ± 3.070; Pb 7.046 ± 8.348
• Pooled %PTWI: Cd 2.25%; tHg 13.84%
• Pooled %BMDL: tAs 0.54%; Pb 20.23%
• Highest category-level Pb %BMDL: dairy 23.47%, baby dinners 33.2% (Table 5 in source)
• THQ values for all elements remained <1 (max pooled THQ ≈ 5.44×10⁻⁸ for As, 2.61×10⁻⁸ for Cd, 1.65×10⁻⁷ for Hg, 2.82×10⁻⁷ for Pb)
• CR values below the 10⁻⁴ carcinogenic-risk threshold (As 2.44×10⁻⁷, Cd 1.41×10⁻⁶, Pb 8.56×10⁻⁹)

Age-subgroup means (Table 7 in source; µg/kg wet weight):

Statistically significant differences: tAs and Pb both differ across age groups (Pb most strongly; products without age declaration had the highest mean Pb).

Methods (brief)

Sample size n=397, distributed across baby dinners (102; poultry 24, beef 16, pork 13, fish 18, rabbit 11, vegetarian 20), porridges (50; with milk 8, milk+fruit 15, cereal-gluten 12, gluten-free cereal 15), fruit/vegetable mousses (58; fruit+veg 9, fruit 33, fruit+cereal 6, fruit+dairy 6, vegetables 4), baby drinks (64; fruit drinks & water 22, fruit juices 42), snacks “for the hand” (62; waffles/crisps 30, biscuits/cookies 17, fruit bars 15), and dairy (60; yellow cheese 28, yoghurt 32). Sampling Dec 2020 – Sep 2021 in Białystok (northeast Poland) brick-and-mortar (hypermarkets, discount stores, children’s-food stores) and Polish online retail.

For As, Cd, Pb: closed-loop microwave digestion of 0.25–0.35 g sample in 4 mL concentrated HNO₃ (Berghof Speedwave, Eningen, Germany) in four phases (10 min/170 °C/20 atm, 10 min/190 °C/30 atm, 40 min/210 °C/40 atm, 18 min cooling); diluted 10×. Analysis by ICP-MS (PerkinElmer NexION 300D, Waltham, MA, USA) with kinetic-discrimination chamber (KED) for As (collision/KED mode) and standard mode for Cd and Pb. LODs: As 0.019, Cd 0.017, Pb 0.16 µg/kg (LOD = 3 × SD of 10 blanks).

For Hg: no preliminary digestion; direct AAS amalgamation on 0.018–0.023 g sample using AMA-254 (Leco Corp./Altec Ltd., Prague, Czech Republic). Sample dried and ashed with oxygen at 600 °C; Hg vapor amalgamator-trapped, released, measured at 254 nm. LOD 0.003 ng/kg.

Reference materials: Corn Flour INCT-CF-3 (Institute of Nuclear Chemistry and Technology, Warsaw) for porridges and snacks; Skim Milk Powder CRM 63R (Community Bureau of Reference BCR) for dairy; Simulated Diet D (Swedish National Food Administration, Uppsala) for baby dinners. Recoveries 98.3–102.0 %; precision 2.2–4.6 % (six independent runs per CRM).

Speciation: As reported as total As only (no inorganic-As speciation); Hg reported as total Hg (no MeHg/inorganic-Hg speciation). Discussion notes that the highest-Hg dinner contained hake, consistent with fish-driven Hg load.

Statistics: Shapiro–Wilk non-normality; non-parametric Mann–Whitney U and Kruskal–Wallis ANOVA; results summarised as median + Q1/Q3 with mean ± SD also reported. Significance thresholds p<0.05, p<0.01, p<0.001. Statistica software (TIBCO Software Inc., Palo Alto, CA, USA).

Exposure indicators used: EDI = C × Cons (consumption from Pitnuts 2016 Polish nutrition survey for children aged 0.5–3 years, mean age 1 y 9 mo, BW 10.5 kg); EWI = EDI × 7; PTWI = EDI × 7 / BW (PTWI_Cd 7 µg/kg BW/day, PTWI_Hg 1.6 µg/kg BW/day); BMDL_0.5 for As (3 µg/kg BW/day), BMDL for Pb (0.5–6 µg/kg BW/day); THQ with RfD_As,Hg 0.3 and RfD_Cd,Pb 1.0 µg/kg BW/day (US EPA); CR with SF_As 1.5, SF_Cd 6.3, SF_Pb 0.0085 mg/kg/day (US EPA); THQ and CR not computed for Hg (US EPA non-carcinogen designation).

Implications

Certification (HMTc): contributes Polish-market wet-weight occurrence data across six baby-food categories (dinners, porridges, mousses, drinks, snacks, dairy) with category-level means and per-category exceedance counts versus EU baby-food maxima. Particularly informative for Pb across baby drinks (1.76% exceedance, the highest single-category rate in the study) and snacks (1.50%), and for Hg in fish-containing dinners (3 of 18 fish dinners drove 3/6 total-cohort Hg exceedances). The 138.99 µg/kg Pb maximum in a fruit/cottage-cheese/grape-juice mousse and the 37.25 µg/kg Hg maximum in a hake-based dinner are useful upper-tail anchors. Single-paper data; routing layer fans out to per-category product pages.

Courses: useful case study for brand QA/regulatory-affairs courses on RTE baby-food multi-category surveillance — illustrates how Pb signal concentrates in drinks and snacks while Hg signal concentrates in fish-containing dinners, with implications for category-specific testing prioritisation. The within-category subgroup data (e.g., gluten-free porridges driving the porridge tAs mean; wafers/crisps driving snack tAs) is the kind of granular pattern brands should expect in their own supply-chain monitoring.

App: contributes wet-weight occurrence data for several ingredient/product cells — rice and rice-containing snacks for tAs, fish and fish-containing dinners for tHg and Cd, fruit-based mousses and fruit juices for Pb. Synthesis pass will resolve the ingredient/metal cells.

Wiki pages this source may touch

• arsenic-total
• cadmium
• mercury-total
• lead
• fish-containing-baby-foods
• fruit-purees
• non-root-vegetable-purees
• meat-and-poultry-purees
• mixed-meals-non-rice
• mixed-meals-rice-containing
• infant-cereal
• teething-and-snacks-non-rice
• teething-and-snacks-rice-based
• fruit-juices-apple-containing
• fruit-juices-non-apple
• flavored-waters
• fish
• fruit
• rice
• vegetables
• meat-and-poultry
• milk-and-dairy

Verification notes

Enhanced 2026-05-17 from the prior 2026-05-12 revision under the v2.0 manual-fetch ingest workflow. Changes from the prior version:

1. Corrected As-max attribution. Prior version stated “wafer/crisps (84.71 µg/kg)” for the As maximum. Source body (Results §3.1) and Discussion both state the 84.71 µg/kg As maximum was in a fruit bar with added rice (“rice bar 84.71 µg/kg”). The abstract’s “wafer/crisps” framing is inconsistent with the body and was the source of the prior error. Corrected to “fruit bar with added rice.”

2. Corrected Hg-max attribution. Prior version stated “poultry dinners (37.25 µg/kg)” for the Hg maximum, copied from the abstract. Source body (Results §3.1, line “The highest Hg value was 37.25 µg/kg in a lunch based on vegetables, rice, and hake”) and Discussion (“those containing hake had the highest levels of Hg (37.25 µg/kg)”) both state the highest-Hg product was a hake (fish) baby dinner, not poultry. The abstract’s “dinners with poultry” wording is a paper-internal abstract/body discrepancy resolvable in favour of the body; the prior session flagged this as anomalous in a footnote but did not correct the attribution. Corrected to “hake/rice/vegetable baby dinner.” The earlier defensive note about “fish-containing components in multi-ingredient poultry dinners” is removed as no longer needed.

3. Methods section expanded with instrument vendors (PerkinElmer NexION 300D for ICP-MS; Leco AMA-254 for Hg amalgamation; Berghof Speedwave microwave digester), reference materials (Corn Flour INCT-CF-3, Skim Milk Powder CRM 63R, Simulated Diet D), LODs (As 0.019, Cd 0.017, Pb 0.16 µg/kg; Hg 0.003 ng/kg), full digestion protocol, statistical software, and the explicit US EPA / JECFA constants used for the EDI/PTWI/BMDL/THQ/CR calculations. Vendor/material names are scientific-method exception per the locked 2026-05-17 brand-firewall reading (CLAUDE.md Part 12 Exception 2), not brand-attribution-to-contamination.

4. Key numbers expanded with the full Table 3 (per-category exceedance counts), Table 5 (EDI/EWI/PTWI/BMDL by category), Table 6 (THQ/CR by category), and Table 7 (age-subgroup means). Prior version had only the abstract-level summary.

5. Speciation flagging clarified. Page now states explicitly: As reported as total As (no inorganic-As speciation); Hg reported as total Hg (no MeHg speciation). metals: frontmatter correctly uses tAs and tHg.

6. Products field expanded from 4 slugs (fish-containing-baby-foods, fruit-purees, meat-and-poultry-purees, baby-cereals-dry-non-rice) to 12, reflecting the six product categories the paper actually distinguishes (drinks, snacks, mousses, porridges, dinners, dairy partial). The prior baby-cereals-dry-non-rice slug was incorrect — the porridges are wet/ready-mixed, not dry-as-sold; replaced with infant-cereal (broad slug). Added: mixed-meals-non-rice, mixed-meals-rice-containing, non-root-vegetable-purees, infant-cereal, teething-and-snacks-non-rice, teething-and-snacks-rice-based, fruit-juices-apple-containing, fruit-juices-non-apple, flavored-waters. All slugs verified against docs/gpt-collaboration/taxonomy-snapshot.md (2026-05-17 snapshot).

7. Ingredients field expanded from 2 slugs (fish, fruit) to 6 (added vegetables, rice, milk-and-dairy, meat-and-poultry) to reflect the multi-ingredient compositional analysis. The rice slug is significant because the paper explicitly identifies rice as the carrier of As across snacks (fruit-bar-with-rice) and the hake/rice/vegetable Hg-max dinner.

8. Matrices field corrected. Prior version had [ready-to-eat-baby-foods, wet-weight]; “ready-to-eat-baby-foods” is not in the matrices vocabulary (docs/gpt-collaboration/system-prompt.md) and “wet-weight” is a basis indicator, not a matrix. Replaced with valid matrix terms: baby-food-puree, baby-snack, fruit-puree, vegetable-puree, infant-cereal, fruit-juice. Wet-weight basis is now documented in Methods.

9. Legacy heading replaced. ”## Wiki pages updated on ingest” → ”## Wiki pages this source may touch” per current schema. No content inside the section was generator-owned (no marker blocks), so direct rewrite is safe.

10. Brand-firewall pass clean. Source names six baby-food producers (Nestle, Nutricia, Hipp, Humana, Holle, Helpa) in §2.1 as supply-source context, not as contamination attributions. These are NOT named in the wiki page; only product-form descriptors are used (per Part 12 strict reading locked 2026-05-17). Method-vendor names ARE retained per Exception 2.

11. Wiki/HMTc firewall pass clean. No threshold proposals, consumer-audience advisories, or cross-source synthesis claims in the page body. The Implications section names what the paper contributes to threshold work without proposing values.

Preserved from prior revision: cite_key, raw_handle, raw_path, license, access_url, evidence_tier, doi.

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.