Heavy Metal Index

Coordinated Sampling Project 41: Microbial and Heavy Metal Detections in Herbs and Spices

Source access

Publication
Local Health Authorities Analytical Committee, Edith Cowan University
Year
2025
Authors
LHAAC
Source type
government report
License
unknown

LHAAC 2025 — CSP41: Microbial and heavy metal detections in herbs and spices, Western Australia

This government surveillance project (Coordinated Sampling Project 41, conducted 18 November 2024 to 31 January 2025) assessed the microbiological safety and heavy metal contamination of 380 herb and spice products purchased from retail businesses across 23 local government areas of Western Australia. ICP-MS analysis screened for nine metals (Al, As, Cd, Cr, Cu, Pb, Hg, Ni, Zn) following NATA-accredited methods; results were benchmarked against Codex Alimentarius standards (CXS 193-1995), the Australia New Zealand Food Standards Code, and EU MLs (Regulation 1881/2006 as amended by 2021/1317). The project was conducted in direct response to FSANZ recommendations following the 2023 US cinnamon applesauce recall and global concerns about lead in cinnamon products; Environmental Health Officers were specifically instructed to prioritise cinnamon-containing products. Key finding: As, Cd, and Hg remained within conservative safety thresholds at the population level; Cr exceeded 2 mg/kg in 5.8% of samples (with extreme dried-herb outliers in the hundreds of mg/kg); and lead exceeded the Codex/EU ML of 0.8 mg/kg for dried spices in 29 spice-powder samples and the EU ML of 10 mg/kg for dried herbs in 6 dry-herb samples, with ground cinnamon the dominant offending commodity.

Key numbers

Sample breakdown (n=380):

  • Spice powder or mix: 203 samples
  • Fresh spice: 77 samples
  • Fresh herbs: 42 samples
  • Other (whole seeds, dried flowers, etc.): 32 samples
  • Dry herbs: 26 samples

Arsenic (As, total): 380/380 samples within the conservative 2 mg/kg nominal threshold per the Figure 9 summary; the Figure 10 caption notes one Fresh spice sample at or above 2 mg/kg (paper-internal phrasing inconsistency between the Figure 9 summary chart and the Figure 10 caption). JECFA BMDL for inorganic As: 7.9 µg/kg bw/day; a 5 g spice intake at 2 mg/kg = ~10 µg As/day, well below the BMDL on this exposure assumption.

Cadmium (Cd): All 380 samples < 1 mg/kg. Samples above the 0.7 mg/kg nominal threshold were predominantly cinnamon-containing products in Fresh spice and Spice powder categories. JECFA PTMI for Cd: ~7 µg/kg bw/week; >200 g/week of 2 mg/kg Cd spice would be needed to approach the limit.

Chromium (Cr, total): Table 6 of the source states 22 samples (5.8%) exceeded 2 mg/kg; the Figure 12A descriptive text states 27. The detailed by-category breakdown (Figure 12A) sums to 27 (6 dry herb, 0 fresh herb, 3 fresh spice, 1 other, 17 spice powder/mix) and 27/380 = 7.1%; the summary chart (Figure 9) and Table 6 use 22 (5.8%). Both numbers are reported above per the source as authored. The “majority were Spice powder or mix (n=17)” attribution is consistent across both expressions. Extreme outliers (from Figure 12A and Discussion §5.2):

  • Fresh chilli: 40 mg/kg
  • Undescribed fresh herb: 61 mg/kg
  • Dried bay leaves: 460 mg/kg
  • Dried curry leaves: 970 mg/kg

The ICP-MS method reports total Cr; Cr-VI speciation was not performed.

Lead (Pb): Above the 2 mg/kg nominal threshold in 13 samples total — 10 spice powder, 2 dry herb, 1 fresh spice (Figure 13). Above the Codex/EU ML of 0.8 mg/kg for dried spices in 29 spice powder samples; above the EU ML of 10 mg/kg for dried herbs (EU 1881/2006 as amended by 2021/1317) in 6 dry herb samples (Table 6, Discussion §5.2). Ground cinnamon was the dominant commodity exceeding the spice-side thresholds.

Mercury (Hg, total): All 380 samples < 1 mg/kg; analyst LOR set at 0.02 mg/kg. Above-LOR detections by category (Figure 14): Dry Herb 9, Fresh Herb 13, Fresh Spice 8, Other 4, Spice powder or mix 44 — total 78 samples above LOR. Of the 44 spice-powder above-LOR detections, 25 were from cinnamon products. JECFA PTWI for methylmercury: 1.6 µg/kg bw/week; 10 g/day spice intake at 0.02 mg/kg total Hg = 0.2 µg/day, far below the methylmercury PTWI on this exposure assumption (note: the source reports total Hg, not MeHg; the MeHg comparison is a conservative upper bound).

Aluminium (Al): Total of 278 samples analysed (one analyst — Agrifood Technology — was unable to perform Al, Cu, Ni, or Zn). Detected across all categories, range 20–5,100 mg/kg (Figure 15). Spice powder or mix had the highest count (159 samples between 20 and 1,400 mg/kg, 0 above 1,400 mg/kg). Dried parsley accounted for the >1,400 mg/kg cohort (1 Dry Herb sample above 1,400 mg/kg, reaching the upper end of the 5,100 mg/kg range).

Nickel (Ni): Majority of samples < 2 mg/kg; 32 samples above 2 mg/kg distributed across categories (Spice powder or mix: 16, Dry Herb: 11, Other: 2, Fresh Spice: 2, Fresh Herb: 1).

Copper (Cu) and Zinc (Zn): Measured by one analyst on 278 samples. Spice powder or mix had the most detections above the 5 mg/kg indicative threshold for both (Cu: 111 samples; Zn: 136 samples). Not in the HMT&C primary analyte set; reported here for completeness.

Lead chromate (PbCrO4) adulteration check: The Discussion §5.2 Table 6 addresses the possibility of illicit lead-chromate colouring of turmeric, paprika, and cinnamon (per ASTA 2024). The CSP 41 samples did not show the co-elevated Pb-and-Cr signature expected from lead chromate adulteration, so no evidence of this adulteration mode was found.

Regulatory references used: Codex CXS 193-1995 (Contaminants and Toxins in Foods); Australia New Zealand Food Standards Code Schedule 19; EU Regulation 1881/2006 as amended by 2021/1317 (Pb in dried spices 0.8 mg/kg, dried herbs 10 mg/kg); FSANZ Compendium of Microbiological Criteria for Food (2022); JECFA evaluations (As BMDL, Cd PTMI, MeHg PTWI). The 2 mg/kg “nominal threshold” used in the report is a conservative screening value, not a regulatory limit, used where no formal ML applies.

Methods (brief)

ICP-MS following acidic digestion with an automated hot block; both NATA-accredited. Four sample categories (fresh herbs, fresh spice, dry herb, spice powder/mix) plus an “Other” category were collected by Environmental Health Officers from 23 WA local government areas during 18 November 2024 – 31 January 2025. Two participating analysts: Agrifood Technology (AT) and Eurofins Analytical Reference Laboratory (EARL). AT was unable to perform Cu, Ni, Zn, and Al analyses, so those four metals were quantified on the 278-sample EARL subset; the other five metals (Al excluded here — see prior sentence; As, Cd, Cr, Pb, Hg) were run on all 380 samples by both analysts. Results compared against Codex, FSANZ, and EU MLs where applicable; 2 mg/kg used as a conservative nominal threshold for metals lacking a formal limit. Microbial methods (NATA-accredited): TPC by pour plate; E. coli, coliforms, and faecal coliforms by MPN; Salmonella and Listeria monocytogenes by MDS molecular detection; Bacillus spp. by spread plate; Clostridium perfringens by pour plate. Four Bacillus cereus-group isolates were referred to PathWest for whole-genome sequencing and toxin-gene profiling.

Limitations

  • No formal Australian or Codex ML exists for most metals in herbs and spices; the 2 mg/kg “nominal threshold” used in the report is a conservative screening value rather than a regulatory limit.
  • Composite blended products (e.g., spice mixes) mean detections may not be attributable to a single ingredient — “lead in cinnamon vs turmeric in a mixed powder” is the example given in the source’s Assumptions section.
  • Single point-in-time sampling; results do not capture seasonal, supplier, or lot-to-lot variation.
  • ICP-MS gives total elemental concentrations; no speciation for As (iAs vs total) or Cr (Cr-VI vs Cr-III). The Discussion notes that the extreme Cr values in dried bay leaves and curry leaves “may reflect total chromium including Cr-III; Cr-VI speciation would be needed for a full toxicological assessment.”
  • Paper-internal inconsistency on the Cr exceedance count: Table 6 and the Figure 9 summary chart give 22 samples (5.8%) above 2 mg/kg; the Figure 12A descriptive text gives 27 (and the per-category breakdown sums to 27). Both are reported here as authored.
  • A second paper-internal inconsistency on Cr in spice powders: Table 6 (Discussion §5.2 Cr row) states “chiefly in spice powders (14 detections; Figure 13)” — but Figure 12B (the Cr-in-spice-powders breakdown) shows 17 samples in the >2.0 mg/kg band, and Table 6 also mis-cites the figure as 13 (the Pb figure) rather than 12B. The 17 figure from Figure 12B is used in this page’s Key numbers section.
  • A paper-internal inconsistency on the Pb exceedance count above the 2 mg/kg nominal threshold: the Figure 9 summary chart on page 16 shows 6 samples above; the Figure 13 by-category chart and the page 18 narrative text both give 13 (10 spice powder + 2 dry herb + 1 fresh spice). This page’s Key numbers section uses 13, matching Figure 13 and the page 18 narrative.
  • The Aluminium 278-sample subset (single-analyst) is not a random sub-sample of the 380 — it is the EARL-tested cohort. Population-level Al inferences therefore carry an analyst-selection caveat the source does not quantify.
  • The mercury exposure comparison uses total Hg measurements against a methylmercury PTWI; this is a conservative upper bound on methylmercury exposure, not a direct measurement.
  • Results were not assessed for consumption-frequency or portion-weighted cumulative dietary intake; the report explicitly states it does not model cumulative dietary intake.

Implications

Certification: Lead contamination in cinnamon products is a confirmed and commercially significant risk, aligning with the US cinnamon applesauce recall narrative and the FSANZ-driven design of this project. The Pb exceedance pattern — ground cinnamon as the dominant offender, 29 of 203 spice powder samples (14%) above the Codex/EU ML of 0.8 mg/kg, and 6 of 26 dry herb samples (23%) above the EU 10 mg/kg dried-herb ML — should inform standards work for the spices category. The extreme Cr values in dried bay leaves (460 mg/kg) and curry leaves (970 mg/kg) are striking but cannot be interpreted as Cr-VI without speciation; recommendation for downstream synthesis is to flag speciation as a gap rather than read the totals as toxicologically determinative.

Courses: Useful real-world case study for the supply-chain surveillance module. Demonstrates how a consumer-safety event (the US cinnamon recall) propagated into regulatory monitoring in a different jurisdiction, and provides current data on the magnitude of microbiological versus chemical contamination risks in herbs and spices, plus the methodological detail (NATA accreditation, two-analyst design, ICP-MS digestion protocol) needed to teach what a credible surveillance study looks like.

App: Confirmed contamination signals usable for ingredient contamination_profile updates by the synthesis workflow: Pb in ground cinnamon (29/203 spice powder samples above the 0.8 mg/kg Codex/EU ML; ground cinnamon dominant); Cr outliers in dried curry leaves (970 mg/kg) and bay leaves (460 mg/kg); Al elevation in dried parsley (>1,400 mg/kg in one sample, range extending to 5,100 mg/kg). As, Cd, and Hg are within conservative thresholds at the population level (with the cinnamon-subset Cd note that the 0.7 mg/kg nominal-threshold cohort is predominantly cinnamon).

Verification notes

  • 2026-05-28 (Claude Opus 4.7, v2.0 manual-fetch merge-enhance — Phase 3 audit application): fresh-context audit subagent (Phase 2) returned REVISE with three actionable items: (1) ❌ Check 2 flagged [[metals/aluminium]] as outside the taxonomy snapshot — verified against wiki/metals/ and the snapshot’s metals list, the correct slug is aluminum (US spelling); fixed. (2) ⚠️ Check 2 flagged [[regulations/eu2023-contaminants-maximum-levels]] as misattributing the source’s regulatory anchor — verified: source Table 6 Pb row explicitly cites “EU 1881/2006; 2021/1317”, which is the superseded predecessor; added [[regulations/eu-1881-2006-contaminants-superseded]] (the predecessor reg the paper actually names) while keeping the current [[regulations/eu2023-contaminants-maximum-levels]] reference since the data was collected after 2023/915 took effect (May 2023) and both are relevant. (3) ⚠️ Check 1 flagged two additional paper-internal inconsistencies not previously surfaced: the Pb Figure 9 (6 above 2 mg/kg) vs Figure 13 (13 above) contradiction, and the Cr Table 6 “14 detections in spice powders” vs Figure 12B (17 in the >2.0 mg/kg band) contradiction — verified both against the PDF and added to the Limitations section. Subagent flagged Check 1 ⚠️ on the Cu/Zn framing as borderline but the wiki language is correctly category-attributed to spice powder or mix, not all-category totals — false positive, not changed. Subagent flagged Check 2 ⚠️ on matrices vocabulary as outside the snapshot’s authoritative list — false positive, matrices vocabulary is an open routing-input vocabulary and the strings used (herbs, spices, spice-powder, fresh-herbs, dried-herbs) are conventional. Subagent flagged Check 5 ⚠️ on the Certification line as “skirting close” but not crossing — borderline, kept as written (no specific HMTc threshold proposed; “should inform standards work” stays on the allowed side per Part 2).
  • 2026-05-28 (Claude Opus 4.7, v2.0 manual-fetch merge-enhance — Phase 1 prior to audit): page was originally created 2026-05-14 in the kimi-cond05 batch ingest with the generic raw_handle: manual-fetch-kimi, a truncated raw_path, a duplicated products entry, and no Limitations section. Phase 1 fixes against re-read of the source PDF (28 pages):
    • Set per-PDF raw_handle: MFK_coordinated-sampling-project-41-microbial-and-heav.
    • Restored full raw_path with the trailing-space folder name (“Manual Fetch Kimi ”) and the un-truncated filename matching the on-disk file.
    • Added raw_sha256: 8682a66d58047a1afae89ffa162ac5f8bb1844e0df74f7747425994ecdb9f899.
    • Added tier_rationale documenting the A-tier rationale (government-of-record provenance, NATA-accredited methods, n=380, FSANZ-recommendation-driven design) given the absence of a peer-reviewed publication venue.
    • De-duplicated products: ["[[products/spices]]"] (was listed twice in the prior frontmatter).
    • Removed the “Wiki pages updated on ingest” reference to [[regulations/codex-cxs-193-1995-contaminants]], which is not an existing wiki page; the only Codex CXS-193 page in this wiki is [[regulations/codex-cxs-193-1995-tin-canned-foods]] and that does not match the contaminants standard the source cites. Replaced with the existing [[regulations/au-nz-food-standards-code-schedule-19-contaminants]] and [[regulations/eu2023-contaminants-maximum-levels]] references for the AU and EU MLs the source actually leans on.
    • Expanded Key numbers to add the missing 6-dry-herb-samples-above-EU-10-mg/kg-ML finding (Table 6, Discussion §5.2), to surface the paper-internal Cr-count inconsistency between Table 6 (22, 5.8%) and Figure 12A descriptive text (27), and to add the lead-chromate adulteration-screen finding (no co-elevated Pb-Cr signature, so no evidence of PbCrO4 adulteration in this sample set).
    • Reframed the Mercury exposure comparison to make explicit that total Hg was measured but the source compares to the methylmercury PTWI; this is a conservative upper bound, not a direct MeHg measurement.
    • Added explicit ## Limitations section documenting the paper-internal Cr inconsistency, the 278-sample Al sub-cohort analyst-selection caveat, the no-speciation caveat for As and Cr, the composite-product attribution caveat, and the single-time-point sampling caveat.
  • Author list ([LHAAC]), year (2025), publication, source_type (government-report), evidence_tier (A), license (unknown), no_doi_assigned (true), sample_n (380), and sample_population framing were unchanged or only lightly reworded for precision.

Wiki pages updated on ingest

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.

CommitDateDescription
b0f3d382026-06-12batch | corpus rescreen b04 old terminal skips