Garofalo et al. 2025 — Ten-year Italian seafood Cd/Pb/Hg compliance monitoring

A ten-year (2014-2023) regulatory compliance monitoring programme run by the Italian Istituto Zooprofilattico Sperimentale del Lazio e della Toscana (IZSLT) analysed 5,854 seafood samples / 9,809 chemical analyses for cadmium, lead, and total mercury against the EU 2023/915 maximum levels (which superseded EC 1881/2006). One hundred forty-two samples (2.43%) were non-compliant. Cadmium exceedances concentrated in cephalopod molluscs (17 of 22 Cd exceedances were cephalopods, mainly squids); mercury exceedances concentrated in marine apex predators (118 of 119 Hg exceedances were marine fish, with swordfish at 11.30%, shark species at 6.48%, and tuna species at 3.11% non-compliance); lead exceedances were rare (one bivalve sample, a truncate donax). A preliminary estimated-weekly-intake calculation for swordfish raised concerns for frequent consumers in the Italian population.

Key numbers

Sample distribution across taxonomic groups (n = 5,854) (Section 2.1, p. 3):

Analyses per metal (n = 9,809 total): THg 4,300; Cd 3,338; Pb 2,171.

Cadmium concentrations by taxonomic group, mg/kg wet weight (Table 2, p. 7):

The 5.60 mg/kg Cd maximum was found in an oceanic squid.

Lead concentrations by taxonomic group, mg/kg wet weight (Table 3, p. 7):

The 3.20 mg/kg Pb maximum was found in a truncate donax (Donax trunculus) from Lazio, Italy.

Total mercury concentrations by taxonomic group, mg/kg wet weight (Table 4, p. 8):

Mercury in freshwater fish was significantly lower than in marine fish (Mann-Whitney U = 255,206.5; Z = 9.348; p < 0.001). The 13.00 mg/kg THg maximum was found in a Pacific marlin.

Total mercury in marine fish by species group, mg/kg wet weight (Table 7, p. 10):

Non-compliance summary (Tables 5-6, pp. 8-9):

• Total non-compliant: 142 of 5,854 samples (2.43%).
• Cadmium: 22 non-compliant (17 cephalopods, range 1.40-5.60 mg/kg; 3 bivalves, range 1.30-1.60 mg/kg; 2 crustaceans, range 0.77-0.85 mg/kg; 0 marine fish).
• Lead: 1 non-compliant (1 bivalve at 3.20 mg/kg).
• Mercury: 119 non-compliant (118 marine fish, range 0.74-13.00 mg/kg; 1 cephalopod, a Peruvian squid at 1.90 mg/kg).

Non-compliance by sampling-request type (Table 5, p. 8):

• HACCP self-control by food companies: 3 of 48 analysed (6.25%).
• Official import control: 36 of 2,528 (1.42%).
• Official control — processing and trading: 102 of 2,634 (3.87%).
• Official control — Italian primary production: 1 of 496 (0.20%).

Specific non-compliant cases highlighted (Section 3.2, p. 9): three Chilean blue mussels (Mytilus chilensis) exceeded EU Cd limits; two American lobsters (Homarus americanus) exceeded EU Cd limits; one Peruvian squid exceeded EU Hg limits (1.90 mg/kg); one Donax trunculus from Lazio exceeded EU Pb limits (3.20 mg/kg) with co-occurring Arsenic 2.5 mg/kg, ndl-PCB sum 1.53 ng/g, and WHO-PCDD/PCDF-TEQ/g (upper bound) 0.0574 pg/g and 0.175 pg/g wet weight.

EU maximum limits applied (Table 1, p. 5; mg/kg wet weight):

Predatory marine fish (swordfish, sharks, tuna species, marlin) carry the 1.00 mg/kg THg limit; most other marine fish carry 0.50 mg/kg; non-predatory species in the lower tier carry 0.30 mg/kg.

Estimated Weekly Intake of THg through swordfish, Italian population (Table 8, p. 11; µg/kg b.w. per week; reference TWI for MeHg = 1.3 µg/kg b.w., for iHg = 4 µg/kg b.w., EFSA CONTAM 2012):

For the general population, EWI of THg via swordfish alone falls within the MeHg TWI (1.3 µg/kg b.w./week). For frequent swordfish consumers — particularly the 3.9% of adults in the 2018 survey averaging 10.64 g/kg b.w./week — EWI ranges 5.03-7.98 µg/kg b.w./week, exceeding the MeHg TWI by 3.9-6.1× under EFSA’s conservative assumption that 100% of THg in predatory fish occurs as MeHg.

Chemical detection distribution overall (Section 3.1, p. 6): of the 9,809 analyses, 55% were measurable (n=5,385), 33% were <LOQ (n=3,277), 12% were <LOD (n=1,147). Pb had the lowest measurable fraction (44%) and Cd the highest (61%).

Methods (brief)

Samples were collected over 2014-2023 under four sampling-plan types: HACCP self-control by food companies, official import control, official control during processing and trading, and official control of Italian primary production. Samples were placed in sterile bags, transported to the IZSLT chemical laboratories in Rome and Florence, and stored at -20 °C until analysis.

A 1.0 g aliquot of homogenised sample was digested in a Teflon tube with 8.0 mL HNO3 (67-69%) and 1.0 mL H2O2 (30-32%) in a MILESTONE ETHOS ONE SK10T microwave oven at 1500 W and 190 °C for 20 min, then diluted to 100 mL with Milli-Q water. Reagent blanks were prepared under the same conditions. An internal standard mixture of Ge, In, and Y was added at 1 µg/mL final.

Cd and Pb were determined by ICP-MS on a Thermo Fisher iCap Q with CETAC 500 Series ASX-520 autosampler (RF 1550 W; auxiliary gas 0.8 L/min; nebuliser 1.06 L/min; plasma 14.0 L/min). Isotopes monitored: 111Cd, 206Pb, 207Pb, 208Pb, 200Hg. Total mercury was determined by both ICP-MS (200Hg) and cold vapour atomic absorption spectrometry (CVAAS) on a PerkinElmer FIMS-100 Flow Injection Mercury System with AS-90 autosampler, using 0.2% (w/v) sodium borohydride in 0.05% (w/v) sodium hydroxide as reducing agent and 3% (v/v) HCl as carrier. Cd/Pb calibration standards: 0.1, 0.5, 1.0, 5.0, 20.0 ng/mL; Hg ICP-MS calibration: 0.05, 0.1, 0.5, 1.0, 2.0 ng/mL; Hg CVAAS calibration: 2.0, 10.0, 20.0, 40.0, 50.0 ng/mL.

LODs were 0.006 mg/kg for Cd and Pb and 0.010 mg/kg for Hg; LOQs were 0.020 mg/kg for Cd and Pb and 0.030 mg/kg for Hg. Linearity R² > 0.999. Measurements were performed in triplicate, in duplicate when results approached or exceeded 80% of the EU maximum limit. Quality control was ensured using DORM-5 reference material (National Research Council Canada), analysed in duplicate during each analytical session; recovery acceptance window 80-120%.

Method validation followed Commission Regulation (EC) No 333/2007 (specificity, linearity, LOD, LOQ, repeatability, reproducibility, recovery, precision). Compliance was assessed accounting for measurement uncertainty.

Statistical analyses used Stata/SE 16.1 (StataCorp LP, College Station, TX). Per taxonomic group and metal: mean, SD, median, maximum, p95, p99. Values of 0.001 and 0 were substituted for <LOQ and <LOD results respectively when computing descriptive statistics.

Speciation note: total mercury (THg) was the only measured Hg species. The EWI calculation assumed 100% of THg in predatory fish occurs as MeHg, per EFSA’s conservative convention; this is an exposure-modelling assumption, not a measured value. No iAs/tAs speciation was performed (As was reported only as a co-occurring element in one specific exceedance sample). Cd and Pb are speciation-irrelevant in food matrices.

Limitation: the paper describes the methodology currently in use; earlier years of the ten-year window used different but equivalent methods, which the authors do not detail.

Implications

Certification: The study contributes ten-year regulatory-compliance occurrence data for Cd, Pb, and THg across seven seafood taxonomic groups, with the largest Italian-monitoring corpus published to date for these analytes. Direct evidence for HMTc seafood-row threshold work includes the per-taxonomic-group p95/p99 occurrence distributions and the per-species-group THg non-compliance fractions (swordfish 11.30%, sharks 6.48%, tuna 3.11%). The Cd cephalopod signal (mean 0.15 mg/kg, p99 1.45 mg/kg, n=1,500) is a comparably substantial occurrence contribution. The dataset is at the EU 2023/915 regulatory floor (compliance/non-compliance framing), not below it, so it bounds the upper tail of the Italian/European seafood marketed product.

Courses: Useful as a case study of multi-stakeholder ten-year surveillance design — HACCP, import, processing, and primary-production sampling plans aggregated under a single regional public-health authority — and of EU 2023/915 compliance assessment in a high-volume official-control setting. The HACCP-self-control non-compliance rate (6.25%, n=48) being substantially higher than the official-control rates illustrates the value of independent monitoring versus self-attestation.

App: Direct occurrence data the consumer app can consume for marine apex predators (swordfish, shark, tuna), bivalve molluscs, cephalopods, crustaceans, and freshwater fish. The species-group-resolved THg distributions in Table 7 are the highest-resolution input for per-species Hg risk weighting.

Microbiome: Not addressed.

Wiki pages updated on ingest

• mercury-total
• cadmium
• lead
• fish
• seafood
• shellfish
• bivalve-molluscs
• molluscs
• freshwater-fish
• swordfish
• shark
• seafood
• fresh-fish
• fish-marine-predatory
• fish-marine-non-predatory
• fish-freshwater
• eu2023-contaminants-maximum-levels

Verification notes

2026-05-29 merge-enhance pass (Claude Opus 4.7, manual-fetch-PDF skill v2.0): the prior 2026-05-15 page version was re-ingested against the full PDF and corrected on the following points.

• metals: removed MeHg. The paper measured total Hg only (by both ICP-MS at 200Hg and by CVAAS on the PerkinElmer FIMS-100). The MeHg/THg ratio of 100% used in the EWI section is an EFSA modelling convention, not a measured speciation. Per CLAUDE.md Part 14, metals frontmatter must reflect what was measured.
• products: removed canned-fish and fish-containing-baby-foods (the paper studies raw whole-seafood regulatory monitoring and does not address canned product or baby-food formulations). Added fish-marine-predatory, fish-marine-non-predatory, and fish-freshwater (the paper’s Table 7 reports species-group-resolved data directly relevant to these product rows).
• ingredients: added swordfish and shark (direct evidence per Table 7).
• matrices: replaced the generic [whole-food, regulatory-compliance-monitoring] with the actual seafood-matrix vocabulary present in the dataset.
• Methods section: corrected the prior page’s claim that THg was measured “by ICP-MS (note: not CVAAS)“. The paper used both ICP-MS and CVAAS for Hg. Added the DORM-5 reference material (NRC Canada), the PerkinElmer FIMS-100 system, calibration ranges, LODs/LOQs, and the EC 333/2007 validation reference.
• Key numbers section: added the full Table 2, 3, 4, 7, and 8 contents, which were absent from the prior version.
• Cd exceedance count corrected from “17” to 22 total (17 cephalopods + 3 bivalves + 2 crustaceans per Table 6). The prior page conflated the cephalopod-only count with the total.
• Hg exceedance count corrected from “118, all marine fish” to 119 total (118 marine fish + 1 Peruvian squid per Sections 3.2 and 4).
• Implications section toned down per CLAUDE.md Part 2: removed prior page’s “HMTc certification can either (a) preclude these species … or (b) require species-specific testing at every lot” framing, which proposed certification policy. The current text describes what the paper contributes to threshold work, not what HMTc should do.
• Section heading renamed ## Methods → ## Methods (brief) per the source-page template.
• sampling_year_range populated (2014-2023); sampling_locations populated.
• Brand-firewall (Part 12): no food brands appear in this paper. Instrument/reference-material vendor names (Thermo Fisher iCap Q, PerkinElmer FIMS-100, MILESTONE ETHOS ONE, CETAC ASX-520, NRC Canada DORM-5, StataCorp Stata/SE 16.1) are scientific-method names and are retained per the 2026-05-17 strict-reading Exception 2.

Preserved from prior version: cite_key, raw_handle, raw_path, license, near_duplicates, ingest_method, created_by, page filename, and the `

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.