Signes-Pastor et al. 2016 — Inorganic arsenic in rice-based products for infants and young children

UK-market survey speciating inorganic arsenic (Asi), DMA, and MMA in 179 commercial rice-based products sold for weaning and toddler-age consumption (baby rice, rice cereals, rice crackers), compared statistically against the US FDA 2014 published rice/rice-product survey for the same three product categories. Asi median across the UK baby-rice subset was 121 µg/kg dry weight, with 14% of samples exceeding the JECFA-proposed 200 µg/kg maximum for polished rice; rice crackers from the UK market showed a significantly higher Asi median (111 µg/kg) than the US-market rice cracker subset (79 µg/kg, p=0.011). Organically-produced samples, which were predominantly whole-grain, showed significantly higher Asi than non-organic milled samples within each product category. The paper closes by arguing that regulatory limits specifically for infant rice-based foods are urgently needed because the modelled per-serving Asi exposures fall within the EFSA BMDL01 risk range for the youngest infant population.

Key numbers

Present-study UK samples (Table 1, mg/kg dry weight):

Baby rice (n=29, 5 manufacturer codes M001–M005):

• ΣAs range across manufacturers: 0.063 – 0.334 mg/kg
• Asi percentage of ΣAs range: 51.4 – 84.6 %
• Asi concentration range: 0.056 – 0.268 mg/kg
• 14% of baby rice samples were above the JECFA-proposed Asi maximum for polished rice (0.200 mg/kg)
• Highest-Asi manufacturer (single brand) had median 0.190 mg/kg (p<0.001 vs other brands in the set)
• DMA range: 0.030 – 0.123 mg/kg; MMA range: 0.001 – 0.003 mg/kg

Rice cereals (n=53, 6 manufacturer codes M006–M011):

• ΣAs range: 0.042 – 0.396 mg/kg
• Asi percentage: 14.2 – 89.6 %
• Asi range: 0.008 – 0.323 mg/kg
• 2% of rice cereal samples above the JECFA Asi maximum
• One manufacturer (code M009) had a higher Asi median (0.234 mg/kg, p<0.001) — up to 1.6× the JECFA maximum
• DMA range: 0.005 – 0.082 mg/kg; MMA range: 0.001 – 0.003 mg/kg

Rice crackers (n=97, 15 manufacturer codes — M001, M005, M006, M009, M011, M012, M013, M014, M015, M016, M017, M018, M019, M020, M021):

• ΣAs range: 0.019 – 0.246 mg/kg
• Asi percentage: 71.1 – 100 % (rice crackers showed the highest Asi proportion of the three product types)
• Asi range: 0.019 – 0.212 mg/kg
• 1% of rice cracker samples above the JECFA Asi maximum
• Highest-Asi brand (single manufacturer) median 0.205 mg/kg (p<0.001)
• DMA range: 0.001 – 0.172 mg/kg

Organic vs non-organic stratification (Table 2, mg/kg dry weight, present study):

Within-category Asi differences for organic vs non-organic were significant for baby rice (p<0.001), rice cereals (p<0.001), and rice crackers (p<0.001). The organic samples were predominantly whole-grain rice; the elevated Asi is attributed to higher Asi at the pericarp/aleurone (bran) layer relative to milled (polished) rice.

EU (present study UK) vs US FDA 2014 published-survey comparison (Table 3, mg/kg dry weight):

UK rice crackers had significantly higher Asi than US rice crackers (median 111 vs 79 µg/kg, p=0.011); UK and US baby rice and rice cereals did not differ significantly in Asi. UK samples consistently had a higher Asi proportion of total speciated As, attributed by the authors to lower DMA in UK rice (geographic/cultivar effects on As speciation).

Per-serving exposure estimates (Table 4 and Fig. 2, present study; WHO body-weight standards):

Baby rice consumed by 4-month-old infants (recommended serving 15 g w.wt., range 15.0 – 15.0):

• Asi intake per serving (UK present study, n=29): median 0.7 µg, range 0.3 – 2.2 µg
• Asi intake per serving (US FDA 2014, n=85): median 1.7 µg, range 0.6 – 3.8 µg
• 3 servings/day at median body weight (6.70 kg) and median Asi/serving = 0.31 µg Asi/kg b.w./day
• This could be 1.67- to 3.37-fold higher with 4 servings/day and 3rd-percentile body weight (5.35 kg) or 97th-percentile Asi/serving

Rice cereals consumed by 12-month-old young children (recommended serving 30 g w.wt., range 22.0 – 35.0):

• Asi intake per serving (UK present study, n=53): median 2.2 µg, range 0.2 – 9.3 µg
• Asi intake per serving (US FDA 2014, n=105): median 2.6 µg, range 0.8 – 11.0 µg
• 2 servings/day at median body weight (9.30 kg) = 0.47 µg Asi/kg b.w./day
• This could be 2.48- to 4.58-fold higher with 4 servings/day and 3rd-percentile body weight (7.50 kg) or 97th-percentile Asi/serving

Rice crackers consumed by 12-month-old young children (recommended serving 30 g w.wt., range 8.0 – 50.0):

• Asi intake per serving (UK present study, n=97): median 3.5 µg, range 0.5 – 9.1 µg
• Asi intake per serving (US FDA 2014, n=199): median 2.4 µg, range 0.2 – 8.2 µg
• 1 serving/day at median body weight = 0.38 µg Asi/kg b.w./day
• This could be 4.95- to 8.93-fold higher with 4 servings/day at 3rd-percentile body weight or 97th-percentile Asi/serving

All three scenarios fall within the EFSA (2009) BMDL01 range of 0.3 – 8 µg Asi/kg b.w./day, which the authors interpret as evidence that infant rice-product Asi exposure cannot be excluded from the cancer-and-skin-lesion benchmark range.

Analytical performance:

• LOD for DMA and Asi: 0.003 mg/kg dry weight (n=6 blanks)
• Speciation CRM recoveries (mean ± SE, n=15): DMA 105 ± 3 %, MMA 93.4 ± 5 %, Asi 94.5 ± 2 %
• All samples were above LOD for Asi; only a few below LOD for DMA and MMA (½ LOD substituted)

Methods (brief)

UK-market sampling: 179 rice-based products purchased Feb 2014 – Mar 2016 from 36 UK food retailers (15 local shops + 21 supermarkets) — duplicate samples of the same product and brand were always purchased from different retailers. Manufacturers were assigned anonymous codes M001 – M021. Three product categories: baby rice (n=29), rice cereals (n=53), rice crackers (n=97).

Sample preparation: freeze-dried, then powdered to nominal homogeneity using a Retsch PM100 rotary ball-mill with zirconium-oxide-lined vessel and zirconium-oxide grinding balls. 0.1 g aliquots digested in 10 mL 1 % Aristar nitric acid (overnight rest, then microwave digestion on a CEM MARS 6 in batches of approximately 30 samples — three-stage slow heating program: ramp to 55 °C in 5 min and hold 10 min; ramp to 75 °C in 5 min and hold 10 min; ramp to 95 °C in 5 min and hold 30 min). 1 mL aliquot transferred to 2 mL polypropylene vial; 10 µL analytical-grade H₂O₂ added to convert arsenite to arsenate prior to chromatography.

Speciation analytical chain: Thermo Scientific IC5000 ion chromatography system with a Thermo AS7 2 × 250 mm column and Thermo AG7 2 × 50 mm guard column, ammonium-carbonate gradient mobile phase (A: 20 mM NH₄CO₃; B: 200 mM NH₄CO₃; gradient program detailed in §2.2), interfaced to a Thermo iCAP Q ICP-MS monitoring m/z 75 with He collision-cell gas to suppress argon-chloride interference. Authentic standards run for DMA, Asi (arsenate form), MMA, tetramethylarsonium (TETRA), and arsenobetaine (AB); presence under each chromatographic peak calibrated using a DMA concentration series.

Quality control: each digestion batch included a method blank and NIST 1568b rice flour CRM. Additionally, batches of 30 samples were prepared which include a blank and the rice-flour CRM. As-speciation certified reference material (DMA, Asi, MMA) used for speciation recovery; certified DMA 0.182, MMA 0.012, Asi 0.092 mg/kg; recoveries 105 ± 3 %, 93.4 ± 5 %, 94.5 ± 2 % respectively (n=15). LOD for DMA and Asi: 0.003 mg/kg d.w. (½ LOD substituted for the small number of <LOD DMA/MMA samples in statistical analysis).

Statistics: general linear model and Duncan’s multiple-range test for between-manufacturer differences; IBM SPSS Statistics 21.0. US comparison data: Asi, DMA, MMA, total As, and ΣAs values from the published US FDA 2014 rice-and-rice-products survey (US FDA 2014; 85 baby rice, 105 rice cereals, 199 rice crackers) — these are reanalysed in Table 3 alongside the UK present-study data, not independently sampled by this paper.

Speciation flag: the paper reports both inorganic arsenic (Asi) and total speciated arsenic (ΣAs = Asi + DMA + MMA) by direct chromatographic speciation. Total As (sum of all species including any unspeciated As) is reported only for the US FDA-derived rows in Table 3; the UK present-study rows report ΣAs (speciation sum) rather than acid-digested total As.

Authors disclose Marie Curie Intra-European Fellowship funding (PIEF-GA-2013-622096) within the EU 7th Framework Programme.

Implications

Certification (HMTc): contributes A-tier UK-market occurrence data on inorganic arsenic in three rice-based-product subcategories that are within the HMTc infant-and-child-foods category. The 29 baby-rice and 53 rice-cereal samples both contribute Asi data to the dry-as-sold baby-cereals-dry-rice-based row; the 97 rice-cracker samples contribute Asi data to the teething-and-snacks-rice-based row. Source-stated medians, ranges, and percentile-style information are available; source does not report sample-level Asi values that could be admitted one-by-one into a pooled percentile distribution. The dataset is also directly comparable to the US FDA 2014 rice/rice-products survey, which is on a separate FDA source page; the present paper’s Table 3 is the statistical bridge between the two.

Courses: useful case study for brand QA and regulatory-affairs courses on speciation discipline (Asi vs ΣAs vs total As), on the organic / whole-grain Asi elevation in rice products marketed as “organic” or “wholegrain” (an audience-counterintuitive finding that brands and retailers should be prepared to communicate), and on the per-serving exposure framing for infant populations.

App: contributes Asi occurrence data (UK market) for the rice → baby-cereals-dry-rice-based and rice → teething-and-snacks-rice-based cells of the contamination-likelihood matrix. The 14% exceedance over the JECFA 0.200 mg/kg proposed maximum for baby rice is a notable upper-tail anchor.

Wiki pages this source may touch

• arsenic-inorganic
• arsenic-total
• baby-cereals-dry-rice-based
• teething-and-snacks-rice-based
• rice
• rice-cereal
• rice-cakes
• efsa-arsenic-contam-2009
• jecfa-arsenic-bmdl
• fda-iAs-rice-cereal-100ppb

Verification notes

Enhanced 2026-05-17 from the prior 2026-05-07 revision under the v2.0 manual-fetch ingest workflow. Changes from the prior version:

1. Added raw_handle: MFK_inorganic-arsenic-in-rice-based-products-for-infan and corrected raw_path to the actual PDF location under raw/Manual Fetch Kimi /06_Infant_Foods_Formula/06_Infant_Foods_Formula/. The prior path raw/studies/... did not exist on the filesystem.

2. Frontmatter slug fields converted to quoted-wikilink form per current schema (Part 14 of CLAUDE.md and docs/gpt-collaboration/system-prompt.md): products and ingredients are now ["[[products/...]]", ...] instead of bare strings.

3. Added matrices: [rice, infant-cereal, baby-snack] (was absent). Values picked from the controlled vocabulary in docs/gpt-collaboration/system-prompt.md — rice for raw-ingredient context, infant-cereal for baby-rice and rice-cereal products, baby-snack for rice crackers.

4. jurisdictions corrected to [GB, EU] (was [EU, US]). The paper’s primary sampling is UK-only (36 UK retailers); the US data are a reanalysis of the published US FDA 2014 rice/rice-products survey, not independent sampling by this paper. Source-page jurisdictions should reflect where this paper sampled, not where comparison-context data originated. The GB code matches related Carey/Meharg-lab papers (e.g., carey2018-rice-dilution-infant-food-eu-arsenic.md).

5. Added sample_n: 179 and a sample_population description that explicitly notes the UK-vs-US distinction so downstream consumers do not double-count US FDA 2014 data already represented in the FDA source pages.

6. product_metal_scope widened from [iAs] only to [iAs, tAs] for both routed product pages. The paper reports both Asi and ΣAs (total speciated As) for UK samples and additionally cites US FDA total As; tAs is in-scope as a routed metal alongside iAs.

7. Body restructured to current schema:

   • Legacy ## Summary heading replaced with un-headed opening paragraph per current wiki/sources/ template.
   • ## Key Numbers (title-case) → ## Key numbers (sentence-case per schema).
   • ## Evidence Fitness, ## Routing Notes, ## Limitations, and ## Wiki Pages Updated On Ingest sections replaced by ## Methods (brief), ## Implications, ## Wiki pages this source may touch, and ## Verification notes. The “evidence fitness” content (medians/ranges only, not sample-level values) is preserved inside ## Implications under Certification.

8. Key numbers expanded substantially from the prior 6-row Table 3 summary. The prior version omitted: (a) the per-manufacturer baby-rice, rice-cereal, and rice-cracker breakdowns from Table 1; (b) the organic vs non-organic stratification (Table 2) — a major finding of the paper; (c) the full Table 3 comparison with DMA, MMA, ΣAs, total As, and Asi% columns; (d) the per-serving exposure estimates (Table 4) and the 0.31 / 0.47 / 0.38 µg Asi/kg b.w./day modelled exposures at median body weight; (e) the JECFA 0.200 mg/kg proposed-maximum exceedance percentages (14%, ~2%, ~1% for baby rice, rice cereals, rice crackers); (f) analytical performance figures (recoveries, LODs).

9. Methods section added (was absent). Documents Retsch PM100 ball mill, CEM MARS 6 microwave digester, Thermo IC5000 + Thermo iCAP Q ICP-MS speciation chain, NIST 1568b rice-flour CRM, IBM SPSS 21.0. All instrument-vendor names are scientific-method exception per Part 12 Exception 2 (locked 2026-05-17), not Part 12 violations.

10. Brand-firewall pass clean. The paper anonymises manufacturers using codes M001 – M021 — no brand names appear in the paper itself. The wiki page reports per-manufacturer ranges and the existence of significantly-higher-Asi single manufacturers without naming or numbering the codes, to avoid any wiki-side construction of a brand ranking. Aggregate / product-form descriptors used throughout.

11. Wiki/HMTc firewall pass clean. No threshold proposals, consumer-audience advisories, or cross-source synthesis claims in the body. The Implications section names what this paper contributes to threshold work without proposing values. The JECFA 0.200 mg/kg figure is reported as a paper-stated reference value, not as a proposed HMTc threshold.

12. Speciation flagging clarified. The paper directly speciates Asi, DMA, MMA via HPLC–ICP-MS. metals: correctly includes both iAs and tAs. Methods section explicitly notes that the UK rows report ΣAs (speciation sum) while the US FDA rows additionally have an acid-digested total-As column.

Preserved from prior revision: cite_key, doi, evidence_tier, license, access_url, pdf_url, raw_sha256.

Audit application (2026-05-17)

Fresh-context audit subagent verdict: REVISE. Each ❌/⚠️ finding independently verified by re-reading Table 1, Table 2, Table 3, and the relevant prose against the source PDF. Applied:

• Manufacturer-code listing for rice crackers corrected. Previously written as “spanning M001, M005–M021” which implied a contiguous range. Source Table 1 rice-cracker section lists 15 codes: M001, M005, M006, M009, M011, M012, M013, M014, M015, M016, M017, M018, M019, M020, M021 — non-contiguous. Replaced with the full list.
• “~2%” / “~1%” hedges removed. Source §3.2 and §3.3 state “2% of the rice cereal samples” and “1% of the rice crackers samples” without hedging. Wiki now matches exact source wording.
• Table 3 baby-rice US ΣAs range corrected. Wiki previously wrote “(0.111–0.262)” — this value does not appear in Table 3; it was a transposition from Table 2’s baby-rice organic-milled ΣAs cell. Source Table 3 prints “0.215 (0.550–0.341)” — itself an apparent source typo where the range upper bound (0.341) is less than the lower (0.550). Page now matches the source typo and flags it inline with [sic] and a likely-intended-value note keyed to the total-As column bounds.
• Table 3 rice crackers UK Asi range corrected. Source Table 3 prints “(0.018–0.201)”; wiki previously transcribed “(0.018–0.211)” — a one-digit transposition in the upper bound. Fixed.
• Methods microwave-heating program clarified. Source §2.1 describes three ramp/hold pairs: 55 °C ramp 5 min + 10 min hold; 75 °C ramp 5 min + 10 min hold; 95 °C ramp 5 min + 30 min hold. Wiki’s prior compressed notation incorrectly implied a 5-min hold at 55 °C. Rewritten as full ramp-and-hold sentence.

Audit findings reviewed and not applied (false positives — disproved by direct source re-read):

• Table 2 baby-rice organic-milled n=9 Asi range. Audit claimed the wiki’s “(0.024–0.120)” should be “(0.087–0.141)” and that (0.024–0.120) was the DMA range for that row. Source verification: the Table 2 DMA column for organic-milled baby rice (n=9) shows a dash ”—” (DMA not reported), and the Asi column directly reads “0.107 (0.024–0.120)“. The audit’s (0.087–0.141) value does not appear anywhere in Table 2; the audit appears to have hallucinated it. Wiki value is correct per the source table.
• Table 2 rice-crackers non-organic whole-grain n=2 Asi range. Audit claimed “(0.018–0.211)” should be “(0.018–0.245)“. Source Table 2 row for non-organic whole-grain rice crackers (n=2): Asi 0.081 (0.018–0.211), ΣAs 0.098 (0.018–0.245). The audit conflated the Asi and ΣAs column values. Wiki value matches the Asi column correctly.
• Table 2 rice-crackers organic whole-grain n=52 Asi median. Audit claimed wiki’s “0.121 (0.048–0.195)” should be “0.150 (0.06–0.328)” — same false-positive pattern. Source Table 2 row: Asi 0.121 (0.048–0.195), ΣAs 0.150 (0.06–0.328). Wiki value matches the Asi column correctly. (This is also the value the paper’s §3.3 prose cites for the organic-vs-non-organic Asi comparison in rice crackers, providing internal corroboration.)
• Baby rice “highest-Asi manufacturer (single brand) median 0.190 mg/kg” framing. Audit flagged this as an attribution concern but then confirmed both the 0.190 value and the source attribution. The wiki framing correctly distinguishes Asi-concentration ranking (where M005 alone is highest) from Asi-percentage ranking (where M001/M003/M005 are jointly elevated). Not a correction; kept as written.

Audit Checks 2, 3, 4, 5 returned ✅ or only minor ⚠️ already addressed above. Page is post-audit consistent with the source for all transcribed numerical values; the one remaining [sic] annotation surfaces a source-internal typo that Karen / downstream reviewers should be aware of.

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.