Brodziak-Dopierała et al. 2023 — Mercury in oil-based dietary supplements on the Polish market

This peer-reviewed study measures total mercury (tHg) in 36 oil-based dietary supplements purchased from Polish pharmacies, drugstores, and herbal shops, using atomic absorption spectrometry via the amalgamation technique (AMA 254, Altec; LOD 0.01 ng Hg). The 36 preparations span vegetable oils (n = 18, all cold-pressed and unrefined), cod liver oils (n = 12), and shark liver oils (n = 6). All samples fell below the EU maximum level for dietary supplements of 0.10 mg/kg (100 µg/kg). Counterintuitively, vegetable oil supplements contained significantly higher tHg concentrations than fish oil supplements. No preparation approached a health-relevant dose at standard recommended intakes.

This paper is distinct from Brodziak-Dopierała et al. 2024 (which covers a broader set of medicines and dietary supplements, n = 139, 2022–2023 purchase dates) and from Torović et al. 2024 (which covers fish oil supplements only in Serbia/Srpska, n = 42).

Key numbers

All concentrations tHg, µg/kg (= ppb), wet weight. Measurements by AAS amalgamation, AMA 254. Table 2 reports both arithmetic mean ± SD and median for each group; non-parametric tests were used because Shapiro–Wilk rejected normality, so the median is the appropriate measure of central tendency.

Overall (n = 36):

• Range: 0.023–0.427 µg/kg
• Arithmetic mean ± SD: 0.165 ± 0.092 µg/kg
• Median: 0.165 µg/kg
• Q1, Q3: 0.088, 0.220 µg/kg
• Coefficient of variation: 55.8%
• Differences across the three subgroups statistically significant (Kruskal–Wallis p < 0.001)

By supplement type (Table 2):

• Vegetable oils (n = 18): mean 0.227 ± 0.077 µg/kg; median 0.218 µg/kg; range 0.131–0.427
• Cod liver oils (n = 12): mean 0.126 ± 0.063 µg/kg; median 0.106 µg/kg; range 0.030–0.207
• Shark liver oils (n = 6): mean 0.061 ± 0.023 µg/kg; median 0.065 µg/kg; range 0.023–0.087
• Fish oils combined (cod + shark liver, n = 18): median 0.088 µg/kg
• Vegetable > fish oils difference statistically significant (p < 0.001); the paper’s prose uses median values when comparing groups

Vegetable oil–specific individual preparations (Figure 5, Table 3 — botanical class only, no brand identification):

• Cedar nut oil: 0.427 µg/kg (highest in the study)
• Evening primrose oil: 0.320 µg/kg
• Camelina sativa oil (one of two preparations sampled): 0.203 µg/kg; the other: 0.173 µg/kg
• Safflower oil: 0.173 µg/kg
• Grape seed oil: 0.131 µg/kg (lowest vegetable oil)

Cod liver oil preparations (Figure 4; values by within-study preparation number, no brand identification):

• Highest: preparation No. 7 at 0.207 µg/kg; preparation No. 3 at 0.205 µg/kg
• Mid-range (preparations Nos. 1, 4, 5): approximately 0.180 µg/kg per paper text
• Lowest: preparation No. 12 at 0.030 µg/kg

Shark liver oil preparations (Figure 3; values by within-study preparation number, no brand identification):

• Highest: preparation No. 1 at 0.087 µg/kg
• preparation No. 6 at 0.077 µg/kg; No. 2 at 0.073 µg/kg
• Lowest: preparation No. 4 at 0.023 µg/kg

Dietary intake estimates at manufacturer-recommended doses (Table 3):

• Highest daily Hg intake among all preparations: 0.0045 µg/day (Safflower oil, 20 mL/day, and Organic Soybean oil, 25 mL/day)
• Highest annual Hg intake: 1.64 µg/year (Safflower oil, prep No. 1)
• Other vegetable oils with annual intake > 1 µg/year: Organic Soybean 1.46, Wheat Germ 1.02, Evening Primrose 1.02
• Lowest daily Hg intake: 0.00001 µg/day (Cod-Liver Oil, prep No. 12)
• Maximum TWI contribution among all preparations: 0.030% of EFSA TWI for methylmercury (Safflower and Organic Soybean, prep No. 1 and 7); all other preparations ≤ 0.022%
• EFSA TWI references used in paper: methylmercury 1.3 µg/kg bw/week → 91 µg/week for 70 kg adult; inorganic mercury 4 µg/kg bw/week → 280 µg/week
• All preparations more than 200-fold below the 0.10 mg/kg (100 µg/kg) EU maximum level for mercury in dietary supplements (Commission Regulation (EU) 629/2008)

Reference comparison from discussion:

• Augustsson et al. 2021: Hg in fish oil supplements 1.1 µg/kg; vegetable-oil supplements 2.1 µg/kg — approximately 5–10× higher than the present Polish study
• Smutna et al. 2009: fish oil total Hg 0.013–2.03 µg/kg (upper range much wider)
• Foran et al. 2003 (US): fish oil supplements 6 to 12 µg/kg (notably higher, earlier-generation products)
• Rucklidge & Shaw 2020 (New Zealand): all fish oil supplements below LOD = 10 µg/kg

Methods

AAS amalgamation technique; AMA 254 spectrometer (Altec, Praha, Czech Republic); wavelength 253.65 nm; inlet pressure 200–250 kPa, oxygen carrier gas (purity ≥ 99.5%); drying 200 s, decomposition 250 s, measurement 90 s. LOD = 0.01 ng Hg per sample. No sample preparation required before AMA 254 analysis. Recovery from certified reference material INCT-MPH-2 (Mixed Polish Herbs, Institute of Nuclear Chemistry and Technology, Warsaw): measured 0.0184 ± 0.0003 mg/kg, recovery 102.3%. Three 100 µL sub-samples per preparation; result = arithmetic mean of three measurements. Non-normal distribution confirmed by Shapiro–Wilk; non-parametric tests used (Mann–Whitney U for two-group comparisons, Kruskal–Wallis ANOVA on ranks for three-group comparisons). Significance level p ≤ 0.05. Sample collection via DISCOVERY Comfort DV200 automatic single-channel pipette with a 20–200 µL ejector (HTL LAB SOLUTIONS).

Speciation: This study measures total mercury (tHg) only. No speciation into methylmercury (MeHg) vs. inorganic Hg was performed. For vegetable oils, the predominant form is likely inorganic Hg; for fish-derived oils (cod liver, shark liver), MeHg may contribute but its proportion is not established by this study. The finding that vegetable oil supplements contain higher tHg than fish oils runs counter to the bioaccumulation narrative for MeHg, consistent with vegetable oils accumulating inorganic rather than methylmercury (the paper does not test this hypothesis directly but notes the counterintuitive direction).

Implications

Certification: Contributes a 36-sample tHg occurrence dataset for Polish oil-based dietary supplements (HMTc Category 16, Row 21) measured in a single survey window. Highest observed concentration 0.427 µg/kg (cedar nut vegetable oil) is more than 200-fold below the EU dietary-supplement maximum of 0.10 mg/kg (100 µg/kg). Vegetable-oil supplements measured significantly higher tHg than fish-oil supplements in this dataset, counter to the bioaccumulation narrative. Augustsson 2021 and Foran 2003 measured notably higher values in other markets and eras; any cross-market threshold-setting must triangulate across these datasets rather than relying on this study alone.

Courses: Useful teaching case for the distinction between expected contamination narratives (fish = mercury bioaccumulation) and observed data (vegetable oil supplements measured higher tHg in this study), reinforcing the importance of direct measurement over assumption. Also a teaching case for honest mean-vs-median reporting under non-normal distributions.

App: tHg medians for cod liver oil (0.106 µg/kg) and shark liver oil (0.065 µg/kg) provide reference data for fish-oil supplement ingredient profiles; vegetable oil median 0.218 µg/kg provides reference for vegetable-oil supplement category.

Microbiome: Not applicable.

Verification notes

2026-05-26: Page enhanced from earlier ingest. Defects corrected:

• raw_path updated to the actual PDF filename (previous value was truncated to a non-existent file path).
• access_url added (DOI URL).
• products enhanced to include [[products/supplements-oil-based]] (HMTc Category 16 Row 21, the specific locked routing target) in addition to broader [[products/dietary-supplements]].
• Subgroup central-tendency values previously labeled “mean” were the medians per Table 2. Page now reports both arithmetic mean ± SD and median for each subgroup, matching Table 2 verbatim, and notes that the paper’s prose uses medians when comparing groups because the data are non-normal.
• Added per-preparation values for the highest/lowest cod liver and shark liver oils (Figures 3–4, Table 3) and additional annual-intake figures from Table 3.
• Added EFSA TWI numerical anchors (1.3 µg/kg bw/week for MeHg; 4 µg/kg bw/week for inorganic) from page 10, and the maximum 0.030% TWI contribution.
• Added Rucklidge & Shaw 2020 New Zealand reference comparison (page 10).

2026-05-26 audit pass: Auto-audit subagent verdict REVISE. Findings applied:

• Stripped all commercial brand names from the Key numbers section (Part 12 brand firewall). The earlier draft listed Iskial, Iskial Plus, Rekinol, Life Shark Liver Oil, Tran Moller’s, Life Tran z Wątroby Dorsza, and “Tran z Witaminami” attached to specific contamination values; these are now reported by within-study preparation number only. The same fix simultaneously resolved an audit-flagged transposition (prep No. 7 in the paper’s Table 1 is “Tran Vitamex” / “Tran w Kapsułkach Vitamex”, not “Tran z Witaminami” which is prep No. 11) — the brand label is no longer present in either spot.
• Removed an eyeballed Figure-5 estimate (“wheat germ oil ~0.280 µg/kg”) that was not a verbatim Table value; Table 3 lists wheat germ in-sample at 0.00023 µg/g (= 0.23 µg/kg), so the estimate was also numerically high.
• Tightened Implications → Certification framing from “supports a low-priority risk classification” to occurrence-data reporting per Part 2.

Slug-vocabulary gaps flagged but not corrected this session (hard constraint: do not create new ingredient pages):

• ingredients/cod-liver-oil is referenced in this source’s frontmatter and Wiki-pages-touched section but does not exist as a wiki ingredient page (freq-1 across the corpus as of 2026-05-26). Below the freq≥2 auto-stub threshold; should be re-evaluated when additional cod-liver-oil sources land.
• ingredients/shark-liver-oil same as above (freq-1).
• The current routing audit does not surface either as unresolved, so routing passes; the gap is taxonomy-completeness, not routing-blocking.

Wiki pages updated on ingest

• cod-liver-oil
• shark-liver-oil
• vegetable-oil
• supplements-oil-based
• dietary-supplements
• mercury

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.