Huff et al. 2025 — Heavy metals in spices from Lancaster, PA

This study measured arsenic, cadmium, and lead in 116 spice samples (82 store-purchased, 34 home-donated) from Lancaster, Pennsylvania, a city with a notably high refugee resettlement rate and elevated childhood blood lead levels relative to the state average. As, Cd, and Pb were detected in 90.37%, 93.1%, and 98.3% of samples respectively, using FDA EAM 4.7 ICP-MS method at an ISO 17025:2017 accredited laboratory. Lead was the dominant concern: 40.5% of all samples exceeded the New York State Department of Health (NYSDOH) 2023 guidance of 0.21 ppm, with a single curry powder sample reaching 31.9 ppm, the only sample whose lead concentration would yield an EDI exceeding the FDA reference level. The three highest single-analyte concentrations (As 0.497 ppm in turmeric from Thailand, Cd 0.294 ppm in paprika from the USA, Pb 31.9 ppm in curry powder from the USA) were all purchased from the same store with an internationally-implying name, pointing at single-supplier concentration risk. The study adds US commercial market data for spice blends (masala, curry powder) to a growing US-focused literature documenting spices as an underrecognized lead-exposure pathway, particularly for immigrant and refugee communities.

Key numbers

Overall median concentrations (n = 116):

• tAs: 0.048 ppm (IQR 0.028–0.075), range 0.007–0.497 ppm
• Cd: 0.056 ppm (IQR 0.034–0.083), range 0.007–0.333 ppm
• Pb: 0.177 ppm (IQR 0.086–0.339), range 0.007–31.9 ppm

Detection frequency: tAs 90.37%, Cd 93.1%, Pb 98.3% of samples.

Store-purchased (n = 82) vs. home-donated (n = 34) medians (Wilcoxon rank sum):

• tAs: 0.055 (IQR 0.035–0.080) vs. 0.034 (IQR 0.019–0.049) ppm; p = 0.001 (significant)
• Cd: 0.055 (IQR 0.037–0.081) vs. 0.058 (IQR 0.029–0.092) ppm; p = 0.98
• Pb: 0.182 (IQR 0.105–0.352) vs. 0.119 (IQR 0.054–0.279) ppm; p = 0.06

Proportion exceeding NYSDOH 2023 reference levels:

• tAs (limit 0.21 ppm, iAs basis): 6.03% all; 8.54% store-purchased; 0.00% home-donated
• Cd (limit 0.26 ppm): 3.45% all; 2.44% store-purchased; 5.88% home-donated
• Pb (limit 0.21 ppm): 40.5% all; 43.9% store-purchased; 32.4% home-donated

Proportion exceeding WHO 2007 reference levels (herbal medicines):

• tAs (5 ppm, mode of national limits from Canada, China, Malaysia, Singapore, Thailand): 0.00%
• Cd (0.3 ppm): 1.72% (all); 0.00% store; 5.88% home
• Pb (10 ppm): 0.09% (all); 1.22% store; 0.00% home

Highest single-analyte concentrations (all from the same store with an internationally-implying name):

• tAs 0.497 ppm in turmeric from Thailand
• Cd 0.294 ppm (text reports the highest Cd as 0.294 ppm; Table 1 IQR range goes to 0.333 ppm) in paprika from the USA
• Pb 31.9 ppm in curry powder from the USA

USA-origin (n = 19) vs other-country-origin (n = 55) store-purchased subsample, median (IQR):

• tAs: USA 0.060 (0.035–0.111) vs other 0.051 (0.037–0.077); p = 0.38
• Cd: USA 0.058 (0.026–0.099) vs other 0.054 (0.037–0.069); p = 0.84
• Pb: USA 0.182 (0.137–0.336) vs other 0.187 (0.109–0.406); p = 0.82
• NYSDOH exceedance: tAs 1.11% USA vs 0.00% other; Cd 1.11% USA vs 0.00% other; Pb 42.1% USA vs 47.3% other

Spearman rank correlations between metals (in samples):

• tAs–Cd: rs = 0.33, p = 0.0003 (weak but significant)
• Cd–Pb: rs = 0.34, p = 0.002 (weak but significant)
• tAs–Pb: rs = 0.58, p < 0.001 (moderate, significant)

Estimated daily intake (EDI) for children aged 1–5 (IR = 0.35 g/day, BW = 15 kg):

• Median: tAs 0.016 µg/day (0.0011 µg/kg/day); Cd 0.020 µg/day (0.0012 µg/kg/day); Pb 0.062 µg/day (0.0041 µg/kg/day)
• Highest: tAs 0.174 µg/day (0.0114 µg/kg/day); Cd 0.117 µg/day (0.00766 µg/kg/day); Pb 11.2 µg/day (0.734 µg/kg/day)
• Reference doses: EPA RfD As 0.6 µg/kg/day, Cd 0.54 µg/kg/day; FDA Pb reference level 2.2 µg/day. The highest Pb EDI (11.2 µg/day, curry powder) is the only sample whose intake exceeds the FDA reference level. Two other samples had EDIs above 1 µg/day but below the FDA action level: a store-purchased paprika at 1.77 µg/day and a home-donated spice at 1.41 µg/day.

LOD for all three metals: 0.01 ppm; non-detects replaced with LOD/√2 per Croghan & Egeghy (2003).

Methods (brief)

ICP-MS (Agilent 7900) using FDA EAM 4.7 (FDA, 2015) for As, Cd, and Pb in spices. 2 g of homogenised spice digested on a hot block with 4 mL HNO₃ (Fischer Chemical), 1 mL H₂O₂ (Sigma-Aldrich), and 10 mL water; 0.25 mL HCl (Fischer Chemical) added post-digestion and brought to 50 mL with water. Analyses run at AGQ Labs USA (Oxnard, CA), ISO 17025:2017 accredited. Quality control: custom certified reference material from Inorganic Ventures (composite of AGQ 22, AGQ 23, AGQ 29, ACG 23Q, MSHG-10PPM, ICP STD Hg 1000) for both initial and continuing calibration verification; calibration via Inorganic Ventures multi-elemental standard; trace metallic impurities in all reagents pre-screened by ICP-MS or ICP-OES (reported NA); Milli-Q water throughout. LOD 0.01 ppm for all three metals; non-detects replaced with LOD/√2 (Croghan & Egeghy, 2003). Internal SOP PE-2118 governed linearity, repeatability, spike recovery, LOD/LOQ, and uncertainty calculations.

Speciation: total arsenic was reported, not inorganic arsenic. The NYSDOH 0.21 ppm reference standard for As is specifically an iAs limit; authors note that because total As did not exceed the iAs limit in the affected samples, actual iAs exceedance rates would be lower than the 6.03% reported here. The wiki uses tAs accordingly. No Hg or Cr was measured.

Statistical analysis: Stata 14. Medians and IQRs because data were non-normal. Wilcoxon rank sum test for store-vs-home and USA-vs-other-country comparisons (α = 0.05). Spearman correlation for between-metal relationships.

Exposure assessment: intake estimates from USEPA-WWEIA Food Commodity Intake Database (FCID 2005–2010), derived from NHANES dietary interview data, restricted to cinnamon, ginger (dried), pepper (black and white), turmeric, and “spices (other).” Children aged 1–5 selected because of vulnerability to heavy-metal effects and the 1–5 age band for routine childhood lead screening (Wengrovitz & Brown, 2009). IR 0.35 g/day; BW 15 kg per EPA Exposure Factors Handbook (2011). EDI = (C_metal × IR) / BW.

Implications

Certification: The 40.5% exceedance rate against NYSDOH 2023 Pb guidance in commercial spice samples from a single US city is among the highest published US market-basket rates for any spice study, and the 31.9 ppm Pb concentration in a curry powder sample is in the range associated historically with lead-chromate adulteration. The authors connect the observed As–Pb correlation (rs = 0.58, p < 0.001) and the elevated single-analyte maxima to the broader literature on intentional colourant adulteration, citing Forsyth et al. 2024 on turmeric lead-chromate adulteration in South Asia. Three of the highest single-analyte concentrations originating from the same retail store points at clustered single-supplier concentration risk that supplier-audit programs would want to surface. The dataset supports inclusion of spice blends (masala, curry powder) as a distinct subcategory of higher-risk spices for routing into the spices product page.

Courses: Documents the immigration and refugee health angle: Lancaster has disproportionately high refugee resettlement and elevated childhood blood lead levels (in 2018 Pennsylvania was among the three states with the highest prevalence of childhood BLL ≥5 µg/dL nationally), and spices are both culturally important and a documented lead source in this community. Illustrates how standard childhood lead risk assessments (focused on paint and water) can miss spice exposure, especially in immigrant households. Strong case study for QA training on underregulated food categories: there is no US federal limit for heavy metals in spices, New York State guidance is currently the only US state-level reference, and the WHO 2007 guidance was developed for herbal medicines rather than culinary spices, so applying it to spices forces an assumption of methodological transferability the authors flag.

App: Provides US commercial market data for mixed spice products (masala, curry powder) showing median Pb 0.177 ppm and maximum Pb 31.9 ppm across 116 samples; useful for flagging spice blends and curry powders as higher-risk subcategories relative to single-ingredient spices. Total As reaching 0.497 ppm in turmeric from Thailand supports flagging turmeric as a tAs-elevated spice ingredient. Per-spice composition of the sampled set is documented (masala 28%, curry powder 17%, paprika 7%, turmeric 6%, annatto 5%, other 37%); per-spice means are not reported in the source.

Microbiome: not addressed.

Verification notes

• 2026-05-29 (Claude Opus 4.7, ingest-next-manual-fetch-pdf v2.0 — Phase 1): EXISTING path merge-enhance. Pre-existing page (last updated: 2026-05-14) carried the legacy raw_handle: manual-fetch-kimi, no access_url, no raw_sha256, no tier_rationale, no sampling_locations, and used the legacy ## Wiki pages updated on ingest heading. Merge-enhance preserves cite-key huff2025-spices-lancaster-pa and the existing Key numbers (verified against the PDF and confirmed accurate). Additions: USA-origin vs other-country sub-analysis (Table 3 of source); Spearman correlations between metals; full EDI table for both median and highest concentrations including the two paprika and home-donated samples that exceeded 1 µg/day Pb; CRM composition (Inorganic Ventures AGQ 22/23/29, ACG 23Q, MSHG-10PPM, ICP STD Hg 1000); detection frequencies (As 90.37%, Cd 93.1%, Pb 98.3%) at the level reported in Results rather than the rounded abstract values; explicit speciation note that tAs not iAs is reported. Updated raw_handle to the MFK_ form; added the SHA-256 (computed via shasum -a 256); added the canonical DOI URL as access_url.
• Three identity checks: DOI grep on 10.1007/s10661-025-14064-9 hit only this page; raw_handle grep on the MFK_ form was empty (the prior page used the generic manual-fetch-kimi handle); cite-key grep on huff2025 matched only this page. No true-duplicate conflict.
• Cd numerical note: the abstract reports the Cd “highest concentration” anecdote as 0.294 ppm in paprika from the USA, while Table 1 reports the Cd column range as 0.007–0.333 ppm. The text on page 6 (“Highest concentration … 0.294 ppm”) and the abstract use 0.294 ppm; Table 1 shows a column maximum of 0.333 ppm. The wiki Key numbers reproduce both values verbatim where the source uses each. This is a paper-internal mismatch the authors did not explicitly reconcile and is preserved as such rather than reconciled wiki-side.
• WHO Cd exceedance: Table 2 of the source reports 1.72% (all samples); the Conclusion text reports 1.77% on page 7. The Table 2 value (1.72%) is treated as the source of truth for Key numbers because Table 2 is the structured presentation; the Conclusion’s 1.77% appears to be a typesetting mismatch. No correction applied to the source; the wiki uses the Table value.
• WHO Pb column ordering anomaly in Table 2: the printed values are all 0.09 | stores 1.22 | homes 0.00, which is internally inconsistent because the all-sample exceedance proportion (0.09%) cannot mathematically be lower than the store-sample subset (1.22%) when the home subset is 0.00%. A coherent reading would put the values in all 1.22 | stores 0.09 | homes 0.00 order or treat the all-column as a typographical error. The wiki reproduces Table 2 verbatim and flags the anomaly here rather than reconciling wiki-side. Audit subagent (2026-05-29) raised this as a ⚠️ presentation concern.
• Table 3 “1.11% USA” denominator anomaly: Table 3 reports US-origin exceedance proportions of 1.11% for both tAs and Cd against N=19 store-purchased USA-origin samples. Arithmetically 1/19 = 5.26%, 1/82 = 1.22%, 1/90 ≈ 1.11%; the 1.11% figure does not match the stated N=19 denominator. The most parsimonious reading is that the percentages in Table 3 were computed against a different denominator (possibly the full N≈90 store-and-home set, or another subgroup) than the stated row N. The wiki reproduces Table 3 verbatim and flags the anomaly here rather than recomputing. Audit subagent (2026-05-29) raised this as a ⚠️ concern; the values themselves are faithfully copied from the source table.
• Speciation: paper measures total As by ICP-MS; iAs not separately quantified. NYSDOH 0.21 ppm reference is an iAs limit, so the 6.03% tAs exceedance rate is an upper bound on iAs exceedance. Recorded as tAs in metals: per CLAUDE.md Part 14.
• Brand firewall (CLAUDE.md Part 12, strict reading locked 2026-05-17): source body and Supplemental Table 1 contain brand names and store names for the store-purchased samples. The wiki page describes the three single-analyte maxima as “from turmeric from Thailand,” “from paprika from the USA,” and “from curry powder from the USA,” all “from the same store with an internationally-implying name,” without naming the brand or store. Scientific-method vendor names (Agilent 7900 ICP-MS, AGQ Labs USA, Fischer Chemical, Sigma-Aldrich, Inorganic Ventures CRMs and standards, Milli-Q water, Stata 14, ISO 17025:2017) are preserved per the 2026-05-17 Exception 2 (scientific reproducibility).
• HMTc firewall (CLAUDE.md Part 2): Implications section reports what this paper contributes to threshold work for the spices product row without proposing a numerical HMTc threshold. The phrase “would want to surface” describing audit programs is paper-level guidance for QA workflows, not a threshold proposal.
• Frontmatter slug check: metals: Pb, Cd, tAs ✓ (per Part 14 vocabulary); ingredients: herbs-and-spices ✓, turmeric ✓, cinnamon ✓ (verified by ls wiki/ingredients/); products: spices ✓ (wiki/products/spices.md exists); matrices: spices, dried-herbs, curry-powder, masala — all four used by sibling spice/herb sources (verified by grep -rh '^matrices:' wiki/sources/); jurisdictions: US ✓.
• Regulations cited in the source (NYSDOH 2023 Subpart 5-2 metals-in-spices guidance, WHO 2007 Guidelines for Assessing Quality of Herbal Medicines, JECFA 2002 food-additive heavy-metal limits, EU Regulation 2023/915, US FDA EAM 4.7) are referenced contextually in body text and Implications but no separate [[regulations/...]] wikilinks are added in this pass per the CLAUDE.md Part 10 rule that regulation pages are created in a separate workflow on first encounter with the correct rule ID/citation.
• Sampling year range not stated in the source; left null. Sampling location captured as Lancaster-PA (city plus surrounding areas per the Materials and methods section).
• 2026-05-29 (Claude Opus 4.7, ingest-next-manual-fetch-pdf v2.0 — Phase 3 audit application): fresh-context audit subagent (Phase 2) returned REVISE with three ⚠️ presentation findings — (1) WHO Pb column ordering in Table 2 (now flagged in this section above); (2) Table 3 “1.11% USA” denominator question (now flagged above); (3) Forsyth-2024 framing too synthesis-style in Implications. Five ✅ clean checks for slug vocabulary, brand firewall, speciation/methods discipline, and numerical fidelity on the overall medians/IQRs/ranges, store-vs-home comparisons, NYSDOH exceedances, Spearman correlations, EDI table, and reference doses. Findings applied: (a) two new verification notes documenting the Table 2 WHO Pb and Table 3 1.11% source-internal anomalies; (b) Implications/Certification rephrased to attribute the Forsyth-2024/lead-chromate-adulteration framing explicitly to the authors (“The authors connect…”) rather than leaving the phrasing as wiki-side cross-source synthesis. No numerical values changed in the page; the values themselves were verified correct against the PDF in the audit. Findings rejected as false positives: none — though the Check 1 ”❌” flag on the 1.11% figure was downgraded to a ⚠️ because the audit’s own bullet text confirmed the values are verbatim from Table 3 (the ❌ pointed at a source-internal arithmetic inconsistency, not at a wiki transcription error).

Wiki pages this source may touch

• herbs-and-spices
• turmeric
• cinnamon
• spices
• lead
• cadmium
• arsenic

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.