The contribution of infant formula to the food survey-based dietary exposure of nine selected elements

Overview

This peer-reviewed German total diet study (TDS) quantifies both the occurrence of nine contaminants and essential elements in powdered infant formula — iAs, Cd, Cr, Pb, Mn, Hg, Ni, Se, and Zn — and the dietary exposure of formula-consuming infants and toddlers aged 0.5 to 3 years, using measured consumption data from the KiESEL survey (n=114 children) combined with occurrence data from the BfR MEAL Study. The study compares exposure against health-based guidance values (HBGVs) and uses a margin of exposure (MoE) approach for Pb and iAs, finding that Cd exposure exceeded the TWI for approximately 30% of infants and toddlers, and that iAs MoEs fell below one for all formula-consuming children. Infant formula contributed up to 64% of reported Zn exposure in infants; reported formula-contribution proportions for iAs, Cr, and Ni were 18%, 19%, and 20%, while Mn was 7%.

Key numbers

Occurrence data (BfR MEAL Study, powdered infant formula, all values upper bound, µg/kg dry weight unless noted):

EU maximum level for Cd in powdered infant formula (cow’s milk): 10 µg/kg; BfR MEAL mean of 4.00 µg/kg is well below this limit.

Dietary exposure from infant formula and total diet (µg/kg body weight/day unless noted, upper bound; medians shown where stated; parenthetical formula-contribution percentages are the source’s Table 5 proportions):

Infants (0.5 to <1 year): iAs from formula median 0.03; reported formula-contribution proportion 18%; total iAs median 0.22 µg/kg bw/day; MoE for total iAs median 0.28 (all 51 infants below MoE of 1, using BMDL05 0.06 µg/kg bw/day for skin cancer).

Toddlers (1 to <3 years): iAs from formula median 0.02; reported formula-contribution proportion 10%; total iAs median 0.21 µg/kg bw/day; MoE median 0.29 (all 63 toddlers below MoE of 1).

Pb: Infants median total 0.26 µg/kg bw/day (P95: 0.33); formula median 0.03; reported formula-contribution proportion 12%; MoE median 1.92 using BMDL01 0.50 µg/kg bw/day for neurotoxicity. No child had MoE below 1.

Cd: Infants median total 0.32 µg/kg bw/day (P95: 0.48); formula median 0.03; reported formula-contribution proportion 10%; 84% of EFSA TWI (2.5 µg/kg bw/week) at mean; 16 of 51 infants (31%) exceeded TWI. Toddlers: median total 0.30 µg/kg bw/day; 18 of 63 (29%) exceeded TWI.

Ni: Infants median total Ni 2.96 µg/kg bw/day; well below EFSA TDI of 13 µg/kg bw/day.

Inorganic Hg (derived from total Hg assuming 100% inorganic for non-fish/seafood): infants median 0.11 µg/kg bw/day; well below EFSA TWI of 4 µg/kg bw/week; reported formula-contribution proportion 18% for infants.

Cr: Total Cr exposure (assuming 100% Cr-III) reached a maximum of 1% of the EFSA TDI (300 µg/kg bw/day) at the 95th percentile (1.93 µg/kg bw/day infants, 2.71 µg/kg bw/day toddlers). Reported formula-contribution proportions were 19% for infants and 8% for toddlers.

Mn: Median total Mn exposure 0.83 mg/day infants (P95 1.51), 1.07 mg/day toddlers (P95 2.35); safe levels of intake 2 mg/day infants and 4 mg/day toddlers. Infants reached 76% of the safe level at the 95th percentile; one infant (2%) exceeded the safe level. No toddler exceeded. Reported formula-contribution proportions were 7% for infants and 4% for toddlers.

Se: Median total Se exposure 15.7 µg/day infants (P95 27.9), 18.9 µg/day toddlers (P95 29.2); UL 55 µg/day infants and 70 µg/day toddlers. Median Se exposure was approximately 30% of UL and the P95 reached 51% (infants). No child exceeded the UL. Reported formula-contribution proportions were 53% for infants and 31% for toddlers.

Zn: Reported formula-contribution proportions were 64% for infants and 43% for toddlers. Median total Zn exposure was 4.09 mg/day in infants and 4.70 mg/day in toddlers. The source states that no UL has been established for infants under one year; toddler Zn exposure was over 60% of the 7 mg/day UL at the median and 93% at the 95th percentile, with 2 toddlers exceeding the UL.

Consumption of infant formula (dry weight): infants mean 8.01 g/kg bw/day (median 6.61; P95 15.8 g/kg bw/day); toddlers mean 4.46 g/kg bw/day (median 3.86; P95 8.38 g/kg bw/day).

Methods (brief)

The BfR MEAL Study (Germany’s first TDS, 2017–2019) provided occurrence data. A single pooled “infant formula” sample combined 15 powdered subsamples: 8 infant formula (pre and stage 1), 4 follow-on formula (stages 2–3), 2 hypoallergenic, and 1 special-medical-needs formula that the source describes as based on cow milk. Elemental determination by ICP-MS (microwave digestion DIN EN 13805:2014); iAs was measured directly by HPLC-ICP-MS/MS for 52 foods where iAs could be expected, while infant formula used the EFSA conversion factor assuming 70% of total As as iAs. Total Hg by direct Hg analyzer (DMA-80). LOD/LOQ for infant formula (dry matrix): iAs LOD 0.001 mg/kg, LOQ 0.003 mg/kg; Cd LOD 0.002 mg/kg, LOQ 0.005 mg/kg; Pb LOD 0.001 mg/kg, LOQ 0.004 mg/kg; Cr LOD 0.02 mg/kg, LOQ 0.05 mg/kg; Ni LOD 0.03 mg/kg, LOQ 0.1 mg/kg; Hg LOD 0.003 mg/kg, LOQ 0.005 mg/kg.

Dietary exposure calculated deterministically: mean concentration multiplied by individual reported consumption from KiESEL (n=114 formula-consuming, non-breastfed children aged 0.5–<3 years), divided by individual body weight, summed across 356 MEAL foods. Upper bound approach: values below LOD set to LOD; values below LOQ and above LOD set to LOQ. Breastfed children (n=13) excluded due to absence of breast milk consumption data.

Limitation note: TDS pooled samples preclude characterizing upper-tail distribution of individual products. The 50–100% left-censoring for iAs, Cd, Pb, Hg, and Ni in infant formula means the true exposure from formula is likely lower than the upper bound estimates reported. iAs conversion factor (70%) may overestimate iAs; direct speciation data suggest iAs is actually the predominant As form in infant formula, meaning true iAs could be higher than 70% of total As. Assuming 100% Cr content corresponds to Cr(III) and 100% Hg in infant formula corresponds to inorganic Hg introduces additional uncertainty.

Implications

This TDS-based study contributes infant-formula exposure evidence for Cd, iAs, Pb, Mn, Zn, and inorganic Hg in German infants and toddlers. Its strongest source-page value is the combination of measured formula consumption from KiESEL with BfR MEAL occurrence data, which makes it a population-exposure profile rather than a product-by-product contamination survey.

Cd exceeded the EFSA TWI for approximately 30% of infants and toddlers, and iAs MoEs were below 1 for all formula-consuming children. The low Pb MoEs (approximately 2 at the median) also support retaining Pb in infant-formula evidence summaries, while the Mn and Zn results show why essential elements may still matter in exposure review when formula is a major dietary contributor.

The study does not provide per-product concentration distributions usable as a Path A occurrence dataset. TDS pooling and left-censoring make its values most appropriate for documenting exposure context and identifying analytes that merit continued monitoring, with explicit caveats that the reported formula concentrations are pooled-sample upper-bound means.

Wiki pages this source may touch

• Infant Formula Powder
• Infant Formula Ingredients
• Milk and Dairy
• Lead
• Cadmium
• Arsenic, Inorganic
• Nickel
• Chromium
• Manganese
• Zinc
• Mercury
• Commission Regulation (EU) 2023/915 cadmium maximum levels

Verification notes

• Cross-vendor audit (Codex, 2026-05-17) corrected the truncated PDF raw_path, replaced invalid matrix powdered-formula with infant-formula-powder, corrected the Zn/UL statement to reflect that no UL is established for infants under one year and that the 93% UL comparison applies to toddlers, clarified the iAs conversion-factor method for infant formula, replaced the invalid EU regulation link with eu-2023-915-cadmium, and tightened HMTc implication language away from certification-analyte justification.
• Merge-enhance (Claude, 2026-05-18, v2.0 manual-fetch skill): added Mn and Zn to metals: to route this source to the manganese and zinc pages (paper provides exposure-vs-HBGV findings for both that were previously missing from Key numbers); added Cr, Mn, and Se exposure paragraphs to Key numbers covering the EFSA TDI / safe-level-of-intake / UL comparisons reported in Tables 5–6; verified per-table that all numbers match the source. Brand-firewall recheck under the strict 2026-05-17 reading: page body and Methods name no brands attached to contamination values; the only corporate string is the “DMA-80” Hg analyzer model, which is the Exception 2 scientific-method carve-out.
• Audit subagent (Claude general-purpose, 2026-05-18) flagged toddler Mn P95 as 2.68 mg/day when Table 5 reports 2,348 µg/day = 2.35 mg/day; verified against source — corrected to 2.35 (the 2.68 figure was the European-average upper-bound from the paper’s Discussion, not the present study’s own P95). Same audit flagged the “Wiki pages this source may touch” bulleted list as inconsistent with the post-2026-05-18 metals: frontmatter (missing chromium, manganese, zinc, mercury); propagated the four routing bullets to match. iAs attribution-slip and dietary-intake matrix concerns noted but left as-is (cosmetic pattern consistent across other analyte bullets; matrix string benign).
• Manual-fetch recheck (Codex, 2026-05-18) updated the legacy raw_handle, added DOI access URL and SHA-256 provenance, added Se/iHg to the analyte list because the paper reports selenium exposure and converts total Hg to inorganic-Hg exposure, and rewrote the legacy Certification/Courses/App implications into source-evidence context only. No metals/selenium link was added because the current wiki has no selenium metal page.
• Fresh-context audit (Codex subagent Hume, 2026-05-18) returned REVISE. Applied the verified findings: corrected the intro sentence that wrongly grouped Mn with the ~20% formula-contribution analytes (source Table 5: iAs 18%, Cr 19%, Ni 20%, Mn 7%, Zn 64% for infants), clarified that the Key numbers parenthetical contribution percentages are source-reported Table 5 proportions, and softened the Methods description so only the special-medical-needs subsample is described as cow-milk-based per source p. 3. Also changed routing from the non-soy powder row to generic infant-formula-powder because the source does not clearly resolve the soy/non-soy split.

Update history

The five most recent substantive edits to this page, classified major (evidence or structure moved), correction (a published value or statement was wrong and has been fixed), or minor (narrative rewritten without changing the underlying evidence). Each description is derived from what the edit did to this page; the linked commit is the authoritative record, routine regeneration passes are excluded, and 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.