HBBF 2025 — Total and inorganic arsenic, cadmium, lead, and mercury in 145 US retail rice samples and 66 alternative grain samples

Healthy Babies Bright Futures commissioned Brooks Applied Labs to test 211 containers of rice and alternative grains purchased from retailers in 20 US metro areas during 2024-2025, measuring total arsenic, inorganic arsenic, cadmium, lead, and total mercury. Inorganic arsenic was detected in 100% of 145 rice samples, with more than one in four (>25%) exceeding the FDA action level of 100 ppb iAs for infant rice cereal. Brown rice grown in the Southeastern US and arborio risotto rice from Italy showed the highest average total heavy metal burdens (151 ppb and 142 ppb respectively); California-grown rice averaged 65 ppb total heavy metals, the lowest of the rice categories tested. Alternative grains averaged 33 ppb total heavy metals versus 108 ppb for rice, a roughly 69% reduction.

Key numbers

• Sample design: 211 total grain containers from 20 metro areas; 145 rice samples across 105 brands; 66 alternative-grain samples spanning amaranth, barley, buckwheat, bulgur, couscous, farro, millet, quinoa, and spelt (p. 5, p. 29).
• Inorganic arsenic detection: 100% of 145 rice samples had detectable iAs; 28% (more than one in four) exceeded the FDA 100 ppb iAs action level for infant rice cereal (p. 1, Executive Summary; p. 15, Insight #4).
• Average total heavy metal levels by rice category (sum of iAs + Cd + Pb + tHg, ppb): California rice (sushi, Calrose) 65 ppb; Thai jasmine 86 ppb; Indian basmati 100 ppb; Southeast US and “USA” white rice 118 ppb; Italian arborio (risotto) 142 ppb; Southeast US and “USA” brown rice 151 ppb (Figure 1, p. 7).
• All-rice vs alternative grains: rice average 108 ppb total heavy metals; non-rice grains 33 ppb (69% lower) (Figure 2, p. 8). Rice contained 28 times more arsenic and 1.5 times more cadmium than alternative grains; average iAs was 84.8 ppb in rice vs 3.1 ppb in alternative grains (footnote 9, p. 8).
• Average cadmium: 18.8 ppb in rice vs 27.5 ppb in alternative grains (alternatives higher for Cd specifically) (footnote 9, p. 8).
• Lead: parboiled saffron-seasoned yellow rice was the highest-Pb rice subtype tested; the single highest sample reached 36 µg/kg Pb — 32 times the all-rice average lead and 5 times the next-highest sample from any other brand. Two lots from the same producer (best-by dates 3 months apart) measured 29.4 and 36 µg/kg Pb respectively, indicating a persistent rather than isolated finding (Appendix A pp. 34-35; Insight #5, p. 16).
• Cooking effect: “cooking rice like pasta” - boiling in 6-10 cups water per 1 cup rice and draining - reduces arsenic by up to 60% (cited body of prior literature; HBBF presents as no-cost mitigation; p. 3).
• Alternative-grain cost: alternative grains averaged ~$0.53/servingvs$0.10/serving for rice (~5x cost premium) (p. 8, citing Statista 2024).
• Sample-level Appendix A inorganic-arsenic ranges by rice subtype (selected, ug/kg dry-weight unless noted RTH):
  • Arborio rice (n=10): iAs 72.8-156 ug/kg; tAs 77.3-213 ug/kg; Cd 5.1-138 ug/kg
  • Basmati rice (n=19): iAs 16.2-123 ug/kg; tAs 24.6-522 ug/kg; Cd 6.8-108 ug/kg
  • Brown rice (n=15): iAs 63.4-201 ug/kg; tAs 76.7-317 ug/kg; Cd 2.9-31 ug/kg
  • Calrose rice (n=5): iAs 39.9-79.6 ug/kg; tAs 65.5-136 ug/kg; Cd 2.7-5.1 ug/kg
  • Jasmine rice (n=22): iAs 15.2-119 ug/kg (the low end is a ready-to-heat brown jasmine sample); tAs 30-167 ug/kg; Cd 1.5*-23.1 ug/kg
  • Sushi rice (n=6): iAs 37.5-80.9 ug/kg; tAs 41.3-169 ug/kg; Cd 2.6*-10.3 ug/kg
  • White rice long-grain (n=~50): iAs 20.6-130 ug/kg; tAs 29.3-342 ug/kg; Cd 2.5*-66.8 ug/kg; many Cd >20 ug/kg
  • Parboiled saffron-seasoned yellow rice (n=5 across the white-rice and Dixie Lily yellow-rice listings): Pb 4.9-36 ug/kg, consistently higher than non-saffron rice; the two highest samples (29.4, 36 ug/kg Pb) came from the same producer, lots 3 months apart.
• Alternative-grain summaries (Appendix A): amaranth/barley/buckwheat/bulgur/couscous/farro/millet/quinoa/spelt; iAs not measured for most (NA in column); tAs typically <12 ug/kg (mostly estimated values flagged ”*”); Cd often higher than rice (e.g., quinoa Cd 15.9-59.9 ug/kg; spelt 32.4-67.3 ug/kg; couscous Cd 9.4-43.7 ug/kg; farro 8.8-44.8 ug/kg).

Methods (brief)

Testing performed by Brooks Applied Laboratories (BAL, near Seattle, Washington) on 211 retail-purchased grain containers shipped from 20 US metro areas between May 2024 and March 2025 (Appendix B). BAL is NELAP-accredited and ISO 17025 certified for the elemental analyses and arsenic-speciation analysis used in this study. Sample digestion used H₂O₂ and concentrated HNO₃ in a controlled microwave digestion program (AOAC 2015.01, modified). Total recoverable arsenic, cadmium, lead, and mercury were quantified by triple-quadrupole ICP-MS (ICP-QQQ-MS); the QQQ stage suppresses polyatomic and doubly-charged interferences, which matters for monoisotopic As. Arsenic speciation used a separate dilute-acid digest analyzed by ion chromatography followed by collision-reaction-cell ICP-MS (IC-ICP-CRC-MS) to resolve As(III), As(V), MMA, and DMA; reported “inorganic arsenic” is the sum of As(III) + As(V), consistent with the FDA approach. Each digestion batch carried four method blanks, a certified reference material (SRM), a laboratory duplicate, and a matrix-spike / matrix-spike-duplicate pair. Units: ug/kg (ppb) on dry-grain basis, except samples noted as “ready-to-heat” reported on wet-weight basis. The public report uses “ND (<x.x)” to indicate non-detect below the method detection limit and an asterisk (*) to denote estimated values falling between LOD and LOQ; per-sample LODs/LOQs are not in the public PDF but are noted as available in BAL’s detailed reports. The report further estimates inorganic arsenic in alternative grains as 70% of total arsenic (footnote 9, p. 8). Author-stated limitation: instant rice may form an additional toxic arsenic species (DMMTA, dimethylmonothioarsenate) during high-heat manufacturing that is not captured in standard iAs methods (Carrijo 2022; Colina Blanco 2024; Yadav 2024 cited).

Implications

• Certification (HMTc): Contributes Pb, Cd, tAs, iAs, and tHg occurrence rows for retail US rice across the 16 named rice subtypes shown in Figure 3 (arborio, basmati, brown, Calrose, Carolina Gold, Charleston Gold, glutinous, Japonica, jasmine, sushi, Thai long-grain, white long-grain, mixed rice, Morelos, Kalijira, wild blend) and 9 alternative-grain categories. Sample size and geographic spread make this a useful pooled source for rice (ingredient) and infant-rice-cereal-adjacent (product) standards work; format axis is dry-grain-as-sold for most samples and ready-to-heat (wet-weight) for select RTH products.
• Courses: Teachable case for brand QA and supply-chain audiences on how rice variety, growing region, and processing format (e.g., RTH, parboiled, seasoned blends like saffron yellow rice) drive heavy-metal load across an otherwise commodity ingredient.
• App: Adds n=145 rice samples plus 66 alternative-grain samples to Rice for Pb, Cd, iAs, tAs, tHg. Also touches Quinoa (n=9 samples) and the broader Non Rice Grains umbrella.

Wiki pages this source may touch

• Arsenic, Inorganic
• Arsenic, Total
• Cadmium
• Lead
• Mercury, Total
• Rice
• Rice Flour
• Quinoa
• Non Rice Grains
• Infant Cereal
• Baby Cereals / Grain Products, Dry (Rice-Based)
• FDA Closer to Zero — 100 ppb Inorganic Arsenic Action Level for Infant Rice Cereal
• FDA Closer to Zero — Program Overview

Verification notes

• This is a B-tier NGO report (HBBF + GreenLatinos + Virginia Women’s Residency), not peer-reviewed. Methodology summary (Appendix B) is high-level; per-sample LOD/LOQ values are not in the public PDF but are referenced as available in the lab’s detailed reports. Treated as B-tier per Part 13.
• Brand firewall (Part 12, strict 2026-05-17 reading): Appendix A names 105 brands across 211 samples. This source page aggregates findings at rice-subtype and grain-category level only. Brand names are stripped from Key numbers; outlier findings that previously named producers (parboiled saffron-seasoned yellow rice Pb cluster; ready-to-heat brown jasmine low-iAs sample) are now described by product-form descriptor alone. Routine brand-by-brand rankings are not reproduced; the appendix tables remain in raw/ for reference.
• 2026-05-29 merge-enhance: populated products: with rice-bulk-grain (direct evidence, 145 rice samples), baby-cereals-dry-rice-based (regulatory context: report frames findings against FDA’s 100 ppb iAs action level for infant rice cereal and discusses homemade rice cereal made from these grains), and other-grain-products (direct evidence, 66 alternative-grain samples). Cleared the routing-malformed advisory. Stripped two brand attributions from the saffron-seasoned yellow rice and Vigo-line yellow-rice descriptors in Key numbers, and corrected a misattributed sample on the jasmine sub-block (the 15.2 ug/kg iAs jasmine sample is the ready-to-heat brown-jasmine sample, not a California-grown white-jasmine sample). Numerical fidelity spot-checks against Figures 1-2, footnote 9, Insight #5, and Appendix A pp. 29-37 all matched.
• 2026-05-29 audit application (fresh-context subagent, REVISE verdict): corrected three Appendix A sample-count typos carried over from the 2026-05-17 version (Calrose n=6 → 5; Brown rice n=16 → 15; Quinoa n=8 → 9); reconciled an inconsistency in the Implications opener (said “10 named rice subtypes” but enumerated 16 from Figure 3); tightened the FDA exceedance figure from “>25% (more than one in four)” to the source’s precise “28%”; corrected the saffron-rice Pb comparison wording from “5 times the next-highest non-saffron sample” to the source-verbatim “5 times the next-highest sample from any other brand” (p. 16); expanded Methods with Appendix B detail (NELAP/ISO 17025 accreditation, AOAC 2015.01 modified microwave digestion with H₂O₂+HNO₃, ICP-QQQ-MS for total metals, IC-ICP-CRC-MS for As speciation, QA/QC batch composition). One ⚠️ left unaddressed: the audit flagged matrices: [rice, cereal-grain] as possibly-novel matrices vocabulary; left as-is because cereal-grain was inherited from the 2026-05-17 version and the matrices snapshot is open. The audit also briefly named brands inside the audit’s own change-log when describing the brand-stripping edits; rewritten here to avoid re-introducing brand names in the wiki body.

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.