Singhato et al. 2025 — Toxic heavy metals (As, Cd, Hg, Pb) in 20 seafood species from Thai markets and risk assessment

Singhato and colleagues measured total arsenic, cadmium, total mercury, and lead in 20 commonly consumed seafood species purchased from three markets in Chonburi Province, Thailand (eastern coastal region), and conducted dietary risk assessment using Thai per-capita and eater-only consumption data stratified by age group (3-5.9, 6-12.9, 13-17.9, 18-34.9, 35-64.9, ≥65 years) and three cooking methods (boiling, frying, grilling). Wedge shell (Mercenaria mercenaria) and musk crab (Charybdis feriata Linnaeus) exhibited the highest concentrations of arsenic (21.4 and 16.2 mg/kg FW respectively) and wedge shell also the highest cadmium (7.3 mg/kg FW, exceeding the Thai 2.0 mg/kg bivalve cadmium limit). Risk characterization (margin-of-exposure for As and Pb, hazard quotient against PTMI/PTWI for Cd and Hg) identified arsenic as the principal exposure-driven concern, with the youngest age group (3-5.9 y) most vulnerable; cadmium and mercury risks were minimal apart from discrete hotspots; lead margins of exposure fell below 100 (potential concern) across most eater-only seafood-consumer scenarios.

Key numbers

All concentrations reported as mean ± SD on fresh weight (FW) basis, n=3 replicate analytical measurements per species (Table 6, p. 9). Cooking yield factors from a companion study (Rueangsri et al. 2025, ref [16]) were applied during risk assessment but raw FW concentrations are the analytical results presented here.

Method-validation values

• Recoveries from spiked samples: As 90 ± 5% (range 85-95%), Cd 100 ± 5% (95-105%), Hg 111 ± 4% (107-115%), Pb 110 ± 4% (106-113%) (p. 5).
• LODs (3SD of ten replicate measurements at lowest detectable concentration): As 0.0007, Cd 0.0009, Hg 0.0011, Pb 0.0011 mg/kg FW (p. 5).
• LOQs (10SD): As 0.0063, Cd 0.0062, Hg 0.0124, Pb 0.0123 mg/kg FW (p. 5).
• Certified reference materials: NIST SRM 1566b (oyster tissue) and NMIJ 7402-a (codfish tissue) (p. 4).
• Rhodium used as internal standard at 50 mg/L in 5% (v/v) HNO₃ (p. 4).

Concentrations of toxic elements in 20 seafood species (Table 6, mg/kg FW; mean ± SD, n=3)

ND = not detected (below LOD of 0.001 mg/kg). <LOQ values for Hg specifically denote <0.012 mg/kg; for As and Cd, <0.006 mg/kg; for Pb, <0.012 mg/kg (p. 9 footnote).

Reported concentration ranges (p. 9)

• Total arsenic: 0.2-21.4 mg/kg FW across species.
• Cadmium: 0.001-7.39 mg/kg FW across species (with wedge shell 7.39 mg/kg exceeding the Thai 2.0 mg/kg cadmium limit for bivalves and squid).
• Mercury: 0.001-0.06 mg/kg FW across species (below the Thai 0.5 mg/kg limit in all samples).
• Lead: 0.01-0.22 mg/kg FW across species (below the Thai 0.3 mg/kg limit for fish and 1.0 mg/kg general limit).
• Three species with highest As: wedge shell (21.4 ± 2.0), musk crab (16.2 ± 1.8), Indo-Pacific horseshoe crab eggs (10.6 ± 1.7) mg/kg FW.
• Three species with highest Cd: wedge shell (7.3 ± 1.8), cockle (1.7 ± 0.3), clam (1.3 ± 0.6) mg/kg FW.

Regulatory comparison (Thai Ministry of Public Health Notification No. 414 B.E. 2563, 2020)

• Maximum allowable concentrations stated in the paper (p. 2 and p. 9): inorganic As 2.0 mg/kg, Cd in bivalves and squid 2.0 mg/kg, Cd in fish 1.0 mg/kg, Hg 0.5 mg/kg, Pb 1.0 mg/kg general (lower threshold of 0.3 mg/kg for fish).
• Total As concentrations exceeded the 2.0 mg/kg threshold in wedge shell (21.4) and musk crab (16.2); however, the Thai 2.0 mg/kg limit is specifically for inorganic arsenic, so the paper applies an estimated iAs fraction (5% for shrimp/prawn/crab/squid/shellfish; 2% for marine fish, citing Dwiyitno et al. 2024) for risk assessment rather than direct iAs comparison.
• Wedge shell Cd (7.39 mg/kg) exceeded the 2.0 mg/kg bivalve limit.
• All Hg and Pb concentrations remained below Thai limits.

Health-based guidance values applied in risk assessment (Table 2, p. 6)

• Arsenic: BMDL 0.06 µg/kg BW/day (skin, bladder, lung cancer endpoint).
• Cadmium: PTMI 25 µg/kg BW/month (renal tubular dysfunction endpoint).
• Mercury: PTWI 4 µg/kg BW/week (renal toxicity endpoint).
• Lead: BMDL 0.5 µg/kg BW/day (3-17.9 y developmental neurotoxicity); BMDL 0.63 µg/kg BW/day (≥18 y chronic kidney disease).

Body weights applied (Table 5, p. 8, kg)

• 3-5.9 y: 17.25; 6-12.9 y: 33.38; 13-17.9 y: 53.42; 18-34.9 y: 63.12; 35-64.9 y: 63.53; ≥65 y: 55.77.

Per-capita seafood consumption (Table 3, p. 7, g/person/day; selected averages with 97.5th percentiles)

• Shrimp and prawn: average 1.5-5.2 across ages; 97.5th percentile 13.71-41.15.
• Crabs: average 0.27-1.19; 97.5th 2.93-9.43.
• Squids: average 0.89-5.42; 97.5th 6.86-41.14.
• Razor clam, wedge shell: average 0.04-0.16; 97.5th 0.00-0.47.
• Oysters: average 0.02-0.63; 97.5th 0.00-6.60.
• Cockle: average 0.32-2.06; 97.5th 2.57-20.57.
• Clam: average 0.24-0.75; 97.5th 2.73-8.20.
• Mussels: average 0.27-1.46; 97.5th 2.80-12.00.
• Indo-Pacific horseshoe crab (eggs): average 0.06-0.24; 97.5th 0.73-3.14.
• Marine fish: average 0.69-1.70; 97.5th 7.14-24.00.

Arsenic margin-of-exposure (Table 7, per-capita basis, selected entries)

The paper uses the EFSA framework: MOE = BMDL_iAs (0.06 µg/kg BW/day) / TE; the EFSA health-concern threshold band for arsenic is MOE values of 2.0-0.4 (average consumers) and 0.9-0.2 (high consumers) — MOE values within or below this band indicate potential health risks (p. 6). This As-MOE threshold band is distinct from the Pb MOE<100 convention. Examples of per-capita As MOE entries (Table 7a-d) for the youngest (3-5.9 y) age group, boiled preparation:

• Pacific white shrimp 1.8; banana prawn 1.7; giant tiger prawn 2.4; ornate rock lobster 0.3; musk crab 1.1.
• Blue crab 7.3; serrated mud crab 8.3; red frog crab 12.0.
• Splendid squid 1.5; cuttlefish 1.7; bigfin reef squid 0.7.
• Razor clam 200.2; oysters 379.6; cockle 7.0; clam 207.1; mussels 12.2; wedge shell 14.5; Indo-Pacific horseshoe crab eggs 29.1.
• Northern whiting fish 17.7; silver pomfret 90.3.

Applying the EFSA arsenic health-concern band (MOE ≤2.0 indicating potential risk for average consumers), the 3-5.9 y age group shows MOE values within or below the at-risk band for nearly all 20 species at average consumption (only razor clam, oysters, clam, and silver pomfret remain above the 2.0 band at this age/cooking-method), with ornate rock lobster (0.3) and bigfin reef squid (0.7) presenting the highest concern at sub-unit MOE.

Cadmium hazard quotient findings (Section 3.2.3, pp. 13-14)

• Per-capita basis: no species exceeded HQ = 1 at mean consumption. At the 97.5th percentile, cockle exceeded unity across all cooking methods (boiled 6-34.9 y, fried 3-34.9 y, grilled 3-64.9 y). Additional method- and age-specific exceedances: musk crab (fried/grilled, 3-5.9 y), cuttlefish (fried, 3-5.9 y), ornate rock lobster (grilled, 3-12.9 y).
• Eater-only basis: consistent exceedances across cooking methods and age strata for musk crab, oysters, cockle, wedge shell, and red frog crab; mean-level exceedances added for blue crab, cuttlefish, mussels, and serrated mud crab in specific age groups.

Mercury hazard quotient findings (Section 3.2.4, p. 14-15)

• Per-capita: no significant risk at average or 97.5th percentile consumption.
• Eater-only: most samples posed no concern; the exception was grilled cockle, which at the 97.5th percentile in ages 3-12.9 y indicated potential risk.

Lead margin-of-exposure findings (Section 3.2.5, p. 15)

• Per-capita view: cooking method differentiated risk (boiled produced the largest share of values above 100, grilled intermediate, fried smallest); low-risk exemplars (mean and 97.5th percentile both above 100 across all age groups) were boiled razor clam/clam/wedge shell and boiled silver pomfret; grilled razor clam/clam/wedge shell; fried razor clam/wedge shell.
• Eater-only view: both statistics generally below 100 across many groups and age strata; the only age-specific exception exceeding 100 was boiled northern whiting in ≥65 y consumers (mean MOE = 102.09). No eater-only species-by-method combination met the low-risk two-tier threshold.

Key conclusions (p. 17)

• All elements within Thai legal standards except wedge shell and musk crab (As exceedances on a total-As basis) and wedge shell (Cd exceedance).
• No health risks from Cd across species except wedge shell.
• Lead MOE below 100 across eater-only seafood-consumer group, indicating significant lead exposure health risk for that subgroup.
• Wedge shell and musk crab presented a high level of concern for arsenic exposure, with the 3-5.9 y age group most vulnerable.

Methods (brief)

Twenty seafood species were selected based on Thai national aquaculture statistics and the Thai Food Composition Database. Samples (500 g per purchase) were collected from three to four vendors at three seafood markets in Chonburi Province in February-March 2025. Specimens were weighed before and after removal of non-edible portions. Thermal processing employed boiling (in deionized water using Milli-Q EQ 7000, Merck KGaA, Darmstadt, Germany), frying (in palm oil), and grilling (electric grill pan). Cooked samples were homogenized (Ultra-Turrax T25, IKA, Staufen, Germany), freeze-dried (Kinetic Engineering Co., Bangkok, Thailand) at -55 °C and 0.040 mbar for 24 h, milled (A11 Basic Analytical Mill, IKA, Staufen, Germany), and stored at -20 °C until analysis.

Sample digestion followed AOAC 2015.01: approximately 0.25 g per sample digested in microwave vessels with 4 mL concentrated HNO₃ (Suprapure 65%, Sigma-Aldrich, St. Louis, MO, USA), 1 mL 30% H₂O₂, and 0.1 mL of internal standard solution containing rhodium at 50 mg/L in 5% (v/v) HNO₃. Digests were transferred to acid-cleaned 50 mL HDPE centrifuge tubes and diluted to 20 mL with deionized water.

Toxic element concentrations were determined by inductively coupled plasma triple quadrupole mass spectrometry (ICP-MS/MS; Agilent 8800 ICP-QQQ, Agilent Technologies, Waldbronn, Germany), with results reported as mg/kg FW. Standard stock solutions for As, Cd, Hg, Pb, and Rh were obtained from NIST (Gaithersburg, MD, USA). Certified reference materials NIST SRM 1566b (oyster tissue) and NMIJ 7402-a (codfish tissue) were analyzed for QA. Moisture was determined gravimetrically per AOAC 925.23. Each measurement was carried out in triplicate (n=3); means ± SD were reported. One-way ANOVA with Duncan’s multiple range post-hoc test was used to compare species means (SPSS v24.0, SPSS Inc., Chicago, IL, USA); p < 0.05 was treated as significant.

The paper reports total arsenic and total mercury (not speciated). For arsenic risk assessment, an inorganic-arsenic fraction was estimated (not measured). The Discussion (Section 4.2, p. 16) cites a baseline assumption that iAs constitutes 10% of total arsenic in seafood (ref [22], EFSA Margin of Exposure resource), but the operative fractions actually applied to the MOE calculation were 5% of total As for shrimp/prawn/crab/squid/shellfish and 2% for marine fish, following Dwiyitno et al. 2024 (ref [25]). Margin-of-exposure was computed against the EFSA inorganic-As BMDL of 0.06 µg/kg BW/day. Cadmium risk was characterized as a hazard quotient against the JECFA PTMI of 25 µg/kg BW/month. Mercury risk was characterized as HQ against the JECFA PTWI of 4 µg/kg BW/week. Lead risk was characterized as MOE against the EFSA developmental-neurotoxicity BMDL of 0.5 µg/kg BW/day (3-17.9 y) and the chronic-kidney-disease BMDL of 0.63 µg/kg BW/day (≥18 y). Cooking yield factors from Rueangsri et al. 2025 (ref [16], Supplementary Table S8) were applied to convert raw-sample concentrations to as-consumed concentrations during the risk-assessment step.

Consumption data were taken from the National Bureau of Agricultural Commodity and Food Standards (Thailand), stratified into per-capita (full-population average) and eater-only (subgroup that consumes the species) intakes, by age group (3-5.9, 6-12.9, 13-17.9, 18-34.9, 35-64.9, ≥65 y), and reported as both average and 97.5th percentile intakes.

Limitations declared by the paper: bioaccessibility of arsenic from the seafood matrix under human digestive conditions was not measured, and the paper recommends future bioaccessibility work to refine the iAs-based MOE estimates (p. 17). Arsenic speciation was assumed (5% / 2% iAs fractions), not measured. Triplicate measurements (n=3) per species; sampling restricted to three markets in a single coastal province during a two-month window.

Implications

Certification: The paper provides Thai-market occurrence data for total As, Cd, total Hg, and Pb across 20 commonly consumed seafood species, on a fresh-weight basis, with triplicate-mean precision. The strongest occurrence signals for HMTc seafood-category threshold work are the As and Cd concentrations in wedge shell (Mercenaria mercenaria) and the As concentration in musk crab (Charybdis feriata) — both of which exceed Thai regulatory limits. The Hg and Pb concentrations are uniformly low (most ≤0.1 mg/kg FW) and do not contribute differentiating evidence at the species level within this dataset. For HMTc threshold work specifically, the iAs values must be interpreted with care: the paper does not measure inorganic arsenic; the 5%/2% estimation is a literature-borrowed conversion factor from Dwiyitno et al. 2024 and the actual iAs fraction in Thai shellfish may vary.

Courses: This paper is a useful teaching example for two distinct points. First, that high total-arsenic concentrations in marine shellfish and crustaceans (often dominated by arsenobetaine and other organic forms) do not necessarily indicate proportionate inorganic-arsenic exposure risk, and that the regulatory limit on total As is fundamentally a proxy for iAs that may be poorly calibrated for species with high organic-As content. Second, that population-average (per-capita) risk metrics can mask substantial risk in the eater-only subpopulation; the lead MOE values here are within the safe range for the Thai population on average but consistently below 100 for actual seafood consumers.

App: Route as Thai-market broad seafood/shellfish occurrence evidence for cadmium and total arsenic. The wedge shell and musk crab rows are the most informative individual data points; routine bivalve and crustacean rows reinforce the broader Thai-Southeast-Asia occurrence pattern. Do not use the iAs MOE figures as occurrence evidence for inorganic arsenic; the iAs values are model outputs not measurements.

Wiki pages this source may touch

• seafood
• shellfish
• fish
• bivalve-molluscs
• molluscs
• seafood
• shellfish
• fresh-fish
• fish-marine-non-predatory
• arsenic-total
• cadmium
• mercury-total
• lead
• jecfa-cadmium-ptmi
• jecfa-arsenic-bmdl
• efsa-arsenic-contam-2009
• efsa-lead-contam-2010

Verification notes

Identifying metadata: title, authors, year, journal, volume, article number, and DOI taken from the PDF first page (Foods 2025, 14, 3725; https://doi.org/10.3390/foods14213725). License is CC BY 4.0 as printed on the PDF first page (“This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license”). Article received 5 October 2025, revised 29 October 2025, accepted 29 October 2025, published 30 October 2025.

Filename note: this PDF was retrieved by the Kimi Agent under the filename “Comprehensive Risk Assessment of Metals and Minerals in Seafood Using Bioaccessibility Correction.pdf” (which does not match the paper’s actual title), and the same directory also contains a separately-named file “Risk Assessment of Toxic Heavy Metal Exposure in Selected Seafood Species from Thailand.pdf” — md5 verification confirmed both filenames refer to byte-identical files (md5 5e4283750def5ba8697aca4e97f3adf3). The KADC_ raw_handle preserves the Kimi-Agent-corrupted filename for downstream raw-file traceability. The paper’s actual content is the Singhato et al. 2025 Thailand seafood risk assessment, not a bioaccessibility-correction study.

Identity checks: zero existing wiki source page returns on the DOI grep (10.3390/foods14213725), the cite-key grep (singhato2025), and the KADC_ raw_handle. NEW source page.

Speciation discipline: the paper measures total arsenic and total mercury; metals frontmatter uses tAs and tHg accordingly. The 5% and 2% inorganic-arsenic fractions used in the paper’s risk assessment are model estimates (citing Dwiyitno et al. 2024), not direct iAs measurements; these are reported in the Key numbers section as model outputs, not as primary occurrence evidence for inorganic arsenic.

Brand firewall (Part 12): the paper does not name any commercial brands of the seafood sampled (samples were collected from vendors at unnamed local markets in Chonburi Province). Scientific-method vendor names are preserved per Part 12 Exception 2: Milli-Q EQ 7000 (Merck KGaA, Darmstadt, Germany), Ultra-Turrax T25 (IKA, Staufen, Germany), A11 Basic Analytical Mill (IKA, Staufen, Germany), Memmert ULE 400 (Memmert GmbH, Büchenbach, Germany), Mettler AT201 (Mettler Toledo, Hamilton, New Zealand), Agilent 8800 ICP-QQQ (Agilent Technologies, Waldbronn, Germany), Suprapure HNO₃ (Sigma-Aldrich, St. Louis, MO, USA), NIST SRM 1566b oyster tissue (NIST, Gaithersburg, MD, USA), NMIJ 7402-a codfish tissue (National Metrology Institute of Japan), SPSS v24.0 (SPSS Inc., Chicago, IL, USA). These are method-side instrument and reference-material identifiers, not brand-name attributions of contamination values.

Wiki/HMTc firewall (Part 2): the paper’s “Discussion” section contains several cross-literature comparison statements (Arampongpun 2017, Dokmaikaw and Suntaravitun 2019, Juwa et al., Arbsuwan 2021, Ritonga et al. 2022, Thongra-ar 2014); these have not been imported into this page as cross-literature synthesis claims. Only the paper’s own measurements, risk-assessment outputs, and stated conclusions are reproduced. No HMTc threshold proposals were derived from the paper’s findings.

Unit and basis discipline: all concentrations on this page are mg/kg fresh weight (FW) as the paper reports them. No silent conversion to ppb or to dry weight has been performed. The cooked-state adjusted concentrations used in risk assessment apply cooking-yield-factor corrections (from Rueangsri et al. 2025 ref [16] and Supplementary Table S8 of the present paper), but the raw FW concentrations in Table 6 are what this page reproduces.

2026-06-03 audit application (fresh-context Claude Opus 4.7 audit subagent, verdict REVISE):

• Finding 1 (⚠️ Check 1, As-MOE narrative incorrectly applied the Pb <100 cutoff): verified against PDF p. 6 — the EFSA-defined health-concern threshold for arsenic MOE is the 2.0-0.4 (average consumers) and 0.9-0.2 (high consumers) band; the <100 cutoff is the Pb convention, not As. Applied: the per-capita MOE preamble and the post-table narrative both now reference the correct EFSA As risk band (≤2.0) and explicitly note that this is distinct from the Pb <100 convention. The post-table sentence also names which four species (razor clam, oysters, clam, silver pomfret) remain above the 2.0 band at 3-5.9 y boiled.
• Finding 2 (⚠️ Check 3, omission of 10% iAs baseline reference): verified against PDF p. 16 Discussion 4.2 — the paper does cite a baseline assumption that iAs constitutes 10% of total As in seafood (ref [22], EFSA Margin of Exposure resource), alongside the operative 5%/2% figures from Dwiyitno et al. 2024 (ref [25]) that were actually applied to the risk calculation. Applied: Methods section now notes both the 10% baseline reference and the 5%/2% operative figures, with the distinction between baseline and applied figures made explicit.
• All Check 2/4/5 findings clean per subagent. Check 3 was otherwise clean. Routing audit re-run after fix: 0 blocking changes.

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.