Brodziak-Dopierała and Fischer 2023 — total mercury in fish from Polish retail

Brodziak-Dopierała and Fischer measured total mercury in 68 fish purchased from points of sale across Poland in 2021-2022, covering 18 species split into 12 saltwater and 6 freshwater. Analysis used direct thermal-decomposition atomic absorption spectrometry (AMA 254). The range of Hg across all samples was 4-827 µg/kg wet weight; the arithmetic mean in marine fish (100 µg/kg) was higher than in freshwater fish (63 µg/kg), but the difference was not statistically significant (p = 0.097). Tuna species — yellowfin and Atlantic bluefin — carried the highest per-species means (265 and 278 µg/kg) and were the only two species in the dataset with Estimated Hazard Quotient values above 1 (1.22 and 1.28); for every other species tested, EHQ stayed below 1 and the authors classified consumption as low-risk. None of the per-sample values exceeded the EU regulatory ceilings under EC 1881/2006.

Key numbers

All concentrations are total mercury (tHg) measured by AMA 254 thermal-decomposition AAS on edible-part (muscle) tissue, reported in mg/kg wet weight in the source. Values below are restated in µg/kg to keep units consistent with other fish-mercury source pages on the wiki. The authors did not measure methylmercury directly; the EDI/EHQ calculations in the source use total Hg as the input concentration with a reference dose of 0.0001 mg/kg/day.

Whole-study statistics (Table 2, n=68)

The marine-vs-freshwater difference was not statistically significant (Mann-Whitney U, p = 0.097). Overall range across all 68 samples was 4-827 µg/kg, with a single tuna outlier at 827 µg/kg (the highest single value in the study, from a tuna in FAO area 34).

Per-species mean Hg (Table 3)

Marine species:

Freshwater species:

Within the marine group, the two tuna species were the only entries with EHQ > 1; halibut at EHQ = 0.71 was the next-highest. Within freshwater, the three predatory species — Nile perch, pike, pike perch — clustered at ~125-140 µg/kg, while rainbow trout, brown trout, and carp sat at 10-18 µg/kg. The authors note that all freshwater carp samples were farmed in Poland.

Wild-caught versus farmed

Farming roughly fivefold lowered mean Hg relative to wild-caught across the dataset taken as a whole.

Exposure assessment (Tables 3 and discussion)

The exposure model used a 70 kg adult, daily fish consumption of 32.14 g, exposure duration of 70 years (assuming reduced consumption during infancy and early childhood), and an oral reference dose of 0.0001 mg/kg/day for Hg per US EPA guidance. EDI values across all species ranged from 0.004 to 0.128 µg/kg b.w./day. Comparing against EFSA tolerable weekly intakes (1.3 µg/kg b.w. for MeHg and 4 µg/kg b.w. for inorganic Hg, equivalent to daily intakes of 0.186 and 0.571 µg/kg b.w. respectively), every species in the dataset sat well below both TWI-derived daily limits. EHQ values ranged from 0.04 (brown trout) to 1.28 (Atlantic bluefin tuna); only the two tuna species crossed the EHQ = 1 threshold the authors used to flag potential health concern.

EU regulatory comparison

EU Commission Regulation 1881/2006 (cited as ref. [22] in the paper) and successor Regulation 2023/915 set a maximum Hg of 1.0 mg/kg wet weight for the listed predatory species (including tuna) and 0.5 mg/kg for most other fishery products. The single highest individual sample in this study (827 µg/kg, a bluefin tuna) sits at roughly 83% of the 1.0 mg/kg predatory ceiling. None of the per-sample values exceeded the applicable ceiling.

Methods (brief)

Sixty-eight fish were purchased at retail in Poland during 2021-2022 across hypermarkets, fish-only shops, and aquaculture breeding tanks, with the goal of representing species “readily available at the point of sale and popular with Polish consumers.” Product origin and FAO catch-area data came from producer labelling. Marine species were predominantly sourced from FAO area 27 (NE Atlantic) with additional samples from FAO areas 21, 34, 41, 51, 61, and 67.

Whole small fish (≤2 kg) were eviscerated; for larger fish, ≥100 g fillet samples were taken from the central part. Edible-part flesh was homogenised on an Analytical Mill A 11 basic (IKA, Warsaw, Poland). Three test aliquots of approximately 100 mg were taken per fish using a RADWAG analytical balance (Radom, Poland).

Total Hg was determined by atomic absorption spectrometry on an AMA 254 analyser (Altec, Praha, Czech Republic) using direct thermal-decomposition with gold-amalgam pre-concentration at 253.65 nm with oxygen carrier gas (≥99.5% purity, inlet pressure 200-250 kPa). Method timings: drying 200 s, decomposition 250 s, measurement 90 s. Reported lower detection limit 0.01 ng Hg per sample. Each sample was measured in triplicate.

Quality control used two Polish MODAS certified reference materials: MODAS-3 HerTis (herring tissue, certified 227 ± 21 ng/g; analysed 221 ± 24 ng/g, recovery 97.4%) and MODAS-5 CodTis (cod tissue, certified 310 ± 22 ng/g; analysed 272 ± 19 ng/g, recovery 87.7%). Both CRMs were sourced from the MODAS consortium (Institute of Nuclear Chemistry and Technology, Warsaw, and the Technical University of Gdańsk).

Statistical analysis used Microsoft Excel and Statistica v13.3 PL (Statsoft, Kraków). Because the data failed Shapiro-Wilk normality testing, discussion of central tendency emphasised medians, and non-parametric tests were used for group comparisons: Mann-Whitney U for two groups and Kruskal-Wallis ANOVA for three or more groups, with significance at p ≤ 0.05.

Methylmercury was not measured directly. The EDI/EHQ exposure model uses total Hg with the US EPA oral reference dose of 0.0001 mg/kg/day; comparisons against EFSA TWIs in the text (MeHg 1.3 µg/kg b.w., inorganic Hg 4 µg/kg b.w.) treat the measured tHg as conservative bounding for both species.

Implications

Certification: This source provides usable as-purchased Polish-retail occurrence data for total mercury across twelve marine and six freshwater fish species, with per-species means and EDI/EHQ calculations. The dataset is broad in coverage but per-species n is small (1-9 individuals per species; eight of the eighteen species are represented by ≤3 samples), so the per-species means should be read as point estimates rather than robust distribution summaries. The wild-caught vs farmed contrast (71 vs 14 µg/kg mean across the full dataset) is a useful structural observation but is confounded with species composition (e.g., farmed carp is also a low-trophic-level species, so the species and provenance effects are not separable here). Tuna remains the standout high-Hg category, with both yellowfin (265 µg/kg) and Atlantic bluefin (278 µg/kg) crossing EHQ = 1 under the authors’ adult-consumer exposure model. Among non-tuna predators, Atlantic halibut at 154 µg/kg is the next-highest marine value, and Nile perch / pike / pike perch (125-138 µg/kg) lead the freshwater group. The dataset is total-Hg only; speciation must be sourced elsewhere for MeHg-specific values.

Courses: Compact case study for explaining (a) how predator status interacts with fishing area to produce the tuna outlier without lifting the whole-population mean above EU limits, (b) the EDI/EHQ exposure-modelling workflow using TWI-derived daily intakes and an oral reference dose, and (c) why “no sample exceeded the regulatory ceiling” and “tuna EHQ > 1 under adult-consumer modelling” are compatible findings on the same dataset.

App: Per-species tHg means feed marine-predatory, marine-non-predatory, and freshwater fish occurrence cells on a wet-weight as-consumed basis. Do not propagate to MeHg cells; methylmercury was not measured in this study, and the paper’s EDI/EHQ figures use total Hg as input.

Microbiome: Not addressed.

Verification notes

PDF read in full across all 15 pages of the published Toxics 11(8):717 article (DOI 10.3390/toxics11080717). All numerical tables (Table 1 species and origin; Table 2 study-wide statistics; Table 3 per-species concentrations, EDI, and EHQ) were transcribed against the PDF.

Paper-internal inconsistency noted: the Materials and Methods section states “samples of the edible parts (flesh) of 68 fish belonging to 18 different saltwater (N = 12) and 6 freshwater (N = 6) species,” and Table 2 reports freshwater n = 30, while Table 1’s freshwater species counts (Nile perch 5 + rainbow trout 1 + brown trout 1 + pike 3 + carp 8 + pike perch 5) sum to 23 samples — leaving a 7-sample gap between Table 1 and Table 2 on the freshwater side. The marine side reconciles (Table 1 sums to 38, matching Table 2). The wiki carries the Table 2 marine/freshwater totals (38/30, summing to 68 as stated) because these are the statistics the authors use for all downstream analysis; the gap appears to be either a typographic shortfall in Table 1 (e.g., a missing line) or transposed sample counts, but does not affect the per-species mean concentrations or EHQ values reported in Table 3.

Species classification note: Table 1 places wels catfish (Silurus glanis) under the “Marine” heading despite Silurus glanis being a freshwater species. The wiki preserves the authors’ table grouping but flags the discrepancy here.

Speciation: The paper measured total mercury by direct thermal-decomposition AAS only; no methylmercury speciation was performed. The metals: frontmatter carries tHg only. The paper compares its EDI values against EFSA’s MeHg TWI in the discussion (a conservative bound treating measured tHg as if it were entirely MeHg), but this is a comparison move, not a measurement of MeHg, so MeHg is not added to the metals array.

Brand-firewall handling: No brand names appear in the source. Sample identification used species names and FAO catch areas only. Scientific-equipment vendor names (AMA 254 / Altec; IKA Analytical Mill; RADWAG balance; Statistica / Statsoft; MODAS CRMs from the Institute of Nuclear Chemistry and Technology and Technical University of Gdańsk) are retained per the Part 12 Exception 2 scientific-method carve-out.

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.