Improving nutritional intakes and reducing metal(loid) exposures from wild fish broth among Inuit pregnant women

Groleau et al. (2025) measured nutrient and metal(loid) partitioning between Inuit country-food ingredients (lake trout, Arctic char, mussel, clam, seaweed) and the broths cooked from them, to identify recipes that maximize nutrient intake while minimizing toxic-metal exposure for pregnant and breastfeeding Inuit women in Nunavik. Most fish, seaweeds, and bivalves were important nutrient sources (>20% of pregnancy daily intake for K, Ca, Mg, Fe, Zn, or Se), but only fish broth was itself a meaningful nutrient source; bivalve and seaweed broths contributed <5% of daily intakes for most elements. Large lake trout (>0.5 kg, n=43) carried elevated total mercury — >67% of individuals exceeded the Canadian RMRV of 0.5 µg/g ww in muscle, with a maximum of 4.5 µg/g ww raw and 5.3 µg/g ww cooked, and >93% of fish-muscle Hg was methylmercury (81% in raw Arctic char). Few metal(loid)s transferred to broth, with arsenic the exception: total-As concentrations in some broths exceeded the Canadian potable-water RMRV of 10 µg/L, though the dominant forms were AsB in fish broth (85–88%) and arsenosugars in seaweed and bivalve broths; the one exception was clam broth, in which inorganic As(V) accounted for 29% of total As.

Key numbers

Sample structure

Broth volumes: 5 L pots; daily-intake calculations assumed a 250 mL cup. Fish broths cooked ~1 h; bivalve broths cooked 5 min per cooking guidelines.

Total mercury (THg) and methylmercury (MeHg) in tissues (µg/g ww)

• Lake trout muscle, large individuals (>0.5 kg, n=43): >67% exceeded the Canadian RMRV of 0.5 µg/g for THg; maximum 4.5 µg/g ww raw and 5.3 µg/g ww cooked for the same individual.
• Arctic char muscle (n=4): well below the RMRV in raw and cooked tissue.
• MeHg as % of THg in fish muscle: >93% across all fish species, except raw Arctic char at 81%.
• Hg speciation by tissue: lower MeHg proportions were observed in mussels (<14%), clams (<56%), and seaweeds (<16%).
• Lake trout cheek MeHg: >1.0 µg/g ww — double the RMRV for commercial fish (cheeks consumed as a delicacy).
• Lake trout tongue and brain THg (combined range, paper does not separate tongue from brain): 0.154 to 0.640 µg/g ww — significantly less than muscle and cheek (attributed to high lipid content / lower protein content; MeHg binds protein-rich tissues).
• Arctic char cheek MeHg: 0.03 to 0.3 µg/g ww.
• Arctic char tongue and brain THg (combined range): <0.002 µg/g ww.

Arsenic — total (TAs) in tissues (µg/g ww)

• Raw seaweed: elevated; several samples exceeded the Canadian RMRV of 3.5 µg/g ww; cooked seaweed lower (water loss reduces TAs in seaweeds).
• Cooked mussel and clam: higher TAs than raw (increase during cooking).
• Lake trout muscle: low; one specimen reached 2 µg/g ww.
• Arctic char tissues (cheek, tongue, brain): at least 1.5× more TAs than lake trout cheeks; six times more AsL in Arctic char tissues than in lake trout.

Arsenic speciation in tissues (% of total As)

• Overall in ingredients: <6% of total As was in inorganic toxic forms.

Cadmium and lead in tissues (µg/g ww)

• All fish, mussel, clam, and seaweed samples remained below the relevant RMRVs for Cd and Pb, except: one seaweed sample at 1.6 µg Cd/g ww (above the EU 1 µg/g bivalve RMRV) and another seaweed sample at 0.57 µg Pb/g ww (above the EU 0.5 µg/g fish RMRV).
• Cooked-mussel mean Cd was higher than raw-mussel mean Cd (increase during cooking).
• HQ values for Cd and Pb in bivalves and seaweeds were below 1 for all ingredients except three seaweed samples for Pb (assuming daily consumption, which is not the dietary pattern in Nunavik).

Metal(loid)s in broths (µg/L)

• THg in broths: most below the Canadian potable-water RMRV of 1 µg/L; one lake trout broth exceeded this.
• Cd in broths: all below the potable-water RMRV of 7 µg/L.
• Pb in broths: all below the potable-water RMRV of 5 µg/L, except a few lake trout broths.
• TAs in broths: fish and seaweed broths exceeded the Canadian potable-water RMRV of 10 µg/L for TAs (Fig. S9), with some approaching 100 µg/L; HQ values for TAs in fish broth were 1.6–2.3 assuming daily consumption, but speciation must be considered.

Arsenic speciation in broths (% of total As in broth)

• Most broths contained <10% As in toxic inorganic forms, except clam broth.

Nutrient concentrations in raw ingredients (µg/g ww)

• K: 1000–5300 across ingredients; highest in fish (lake trout, Arctic char) — Arctic char muscle raw ~4900 µg/g.
• Ca: highest in seaweeds (1000–3500); bivalves 250–500; fish low (<100) except brook trout up to 6800 (likely bone fragments during analysis).
• Mg: highest in seaweeds (350–1850); fish lower; some seaweeds above 550.
• Fe: highest in seaweeds (50–250+); bivalves moderate; fish low.
• Zn: slightly higher in bivalves (8–23) than seaweeds (4–14) and fish (3–10).
• Se: at least 20× higher in bivalves and fish (0.20–1.39) than seaweeds (0.01–0.03).

Nutrient contributions to pregnancy daily intake (% of RDA per portion: 100 g ingredient or 250 mL broth)

From Table 1:

From Table 1 (broths, 250 mL cup, unfiltered):

• Fish broths were good sources of Ca (4.5–12%), Mg (15–20%), K (26–53%), and Se (83–89%); seaweed and bivalve broths were poor sources (<5% for most elements).
• Cooked bivalves were excellent sources of Fe (48–78%), Zn (78–82%), Se (95–117%).
• Cooked seaweed was an excellent source of Ca (32%), Fe (34%), Zn (81%).

PUFAs (mg/g ww in fish muscle; mg/L in broth)

• Fish muscle PUFAs: ≤2.5 mg/g ww on average, with a few DHA concentrations up to 5 mg/g ww.
• Lake trout DHA 1.1–3.2 mg/g ww; EPA 0.9–2.0 mg/g ww.
• Fish broth PUFAs typically above 20 mg/L; some lake trout broths exceeded 500 mg/L.
• Fish broth contribution per cup: <200 mg PUFAs; cumulative across multiple cups for habitual broth drinkers.

Iron-releasing nutritional device (LIF)

• LIF release in 250 mL cup of broth-relevant water over 2 h:
  • pH 3.5 (lemon-juice-acidified water): ~5 mg Fe at 120 min — approaches a commercial 10 mg supplement only when LIF is pre-conditioned in acidic water at pH 3.5.
  • pH 4.5–6.5 (broth-realistic pH range): 2–4 mg Fe over 2 h.
  • Tap water (pH 8.80): minimal Fe release.
• To match a 10 mg daily Fe supplement at pH 4.5–6.5 would require ~20 cups (~5 L) of broth containing LIF, versus ~1.5 L of unfiltered lake trout broth (which itself releases enough Fe to be a meaningful source).
• Marketed claim (per cited social-enterprise literature): 6–8 mg/L Fe release after 10 min boiling at pH ~6.5; the experimental conditions in this study did not reproduce that level.

Methods (brief)

Ingredients (lake trout, Arctic char, brook trout, lake whitefish, mussel, clam, seaweed) were sampled with Inuit communities in Nunavik (Inukjuak, Kangiqsualujjuaq, Puvirnituq, Salluit). Cooking experiments matched traditional preparation: fish boiled ~1 h in 5 L pots; mussels and clams boiled 5 min; seaweed boiled 1 h under four water conditions (MilliQ, tap water, tap water + lemon juice, tap water + chicken broth). A separate experiment tested LIF iron release in 1 L glass beakers at varying pH (3.5, 4.5, 5.5, 6.5, 7.5) and tap water alone.

Tissue samples (4 ± 2 g muscle, brain, cheek, tongue, seaweed) were lyophilized and homogenized. Tap water and broth samples were preserved with 2% concentrated HNO₃ (trace-metal grade) or 0.4% concentrated HCl. Total mercury (THg) was measured by direct mercury analyzer (DMA-80, Milestone). MeHg was measured by cold-vapor atomic fluorescence spectrometry (CVAFS; Tekran 2600) after digestion with hydroxylamine, isopropanol, and stannous chloride. Other metals (K, Ca, Mg, Fe, Zn, Se, total As, Cd, Pb) were measured by ICP-MS/MS (Agilent 8900 Triple Quadrupole). Three digestion routes were used depending on matrix: hotplate HNO₃/HCl for water and broths; pressure-steam-sterilizer HNO₃ + HCl/H₂O₂ for 2019 fish muscle; microwave digestion (Multiwave 5000, Anton Paar) for other tissues. Arsenic speciation (As(III), As(V), AsB, DMA, MMA, As-Sug, AsL) was performed by HPLC-ICP-MS/MS.

Quality control used the following certified reference materials: DORM-2, DORM-4 (NRCC; fish protein), TORT-3 (NRCC; lobster hepatopancreas), SO-2 (CCMET; soil), SRM-3232 (NIST; kelp powder), and PTC (C02A metals in water; Proficiency Testing Canada) for tap water and broths. Recoveries: tissues 99–134% (mean), tap water and broths 78–136%, 86–142%, 85–107%, 99–114% by element. Fatty-acid analysis followed a chloroform:methanol extraction and gas chromatography with flame-ionization detection (TransBIOTech, Lévis, Canada). Statistical analysis used Shapiro-Wilk normality, Levene’s homogeneity-of-variance, ANOVA with Tukey HSD post-hoc, or Kruskal-Wallis with Mann-Whitney U as appropriate; α = 0.05; RStudio v4.2.2. Daily-intake percentages were calculated against Health Canada DRIs for pregnant women, and hazard quotients (HQ) were calculated from EDI and RMRV / pTDI references (Health Canada, JECFA, WHO).

Implications

Certification (HMT&C): This paper contributes occurrence data for freshwater fish (lake trout, Arctic char), bivalves (mussel, clam), and seaweed under traditional Inuit preparation, including cooked-tissue values and broth-transfer fractions. The arsenic-speciation data (iAs <6% of total As in fish tissue; AsB 85–88% in fish broth; iAs(V) 29% in clam broth as a notable exception) reinforces the existing wiki position that tAs is not a substitute for iAs in fish/shellfish matrices. The mercury data — >67% of large lake trout (>0.5 kg) exceeding the 0.5 µg/g RMRV with maxima above 4 µg/g — is relevant to any country-food-adjacent product category that draws on large freshwater predatory fish. The cooked-tissue Cd increase in bivalves and the seaweed Cd outlier (1.6 µg/g ww) are also category-level occurrence data points for shellfish and seaweed product threshold work.

Courses: A useful teaching case for risk-benefit trade-offs in traditional-food consumption: nutrient density (Se, K, Mg, Fe) from country-food broths is genuinely substantial for vulnerable populations, but size-selective harvest of fish (smaller lake trout, or non-lake-trout species) materially reduces MeHg exposure without sacrificing the nutrient benefits. The LIF finding (released iron 2–4 mg/L at broth-realistic pH, well below the marketed 6–8 mg/L) is a cautionary note on extrapolating manufacturer claims to real food matrices.

App: This paper provides ingredient-level concentration ranges and speciation splits suitable for fish-broth and country-food risk modeling: tHg-to-MeHg conversion factors (~0.93 for fish muscle; ~0.81 for raw Arctic char muscle; much lower for bivalves and seaweeds), iAs-to-tAs splits in fish vs clam vs seaweed broths, and broth-transfer fractions for K, Ca, Mg, Fe, Zn, Se.

Microbiome: Not directly addressed in this paper.

Verification notes

Enhancement 2026-05-18 (Claude Code, manual-fetch ingest): The previously committed page (2026-05-14) had multiple defects fixed here:

• Wrong title — the previous version’s title (“Heavy metals, trace elements, and nutrients in traditional Inuit country foods…”) does not match the paper. The actual title is “Improving nutritional intakes and reducing metal(loid) exposures from wild fish broth among Inuit pregnant women”.
• Wrong authors — the previous version listed Dallaire R, Bherer L, Dewailly E, Ayotte P among the authors; the actual authors are Tania Groleau, Mélanie Lemire, Dominic E. Ponton, and Marc Amyot.
• Invented sample_population — the previous version included walrus, seal, beluga, ringed seal, and char eggs; this paper does not measure marine mammals or char eggs. sample_n set to null (multiple cohorts; see sample-structure table in Key numbers).
• Invalid matrices vocabulary — previous version used fish-tissue, broth, food. The matrices vocabulary recognizes fish, mollusk, seaweed, drinking-water; broth was added here because broth is the paper’s primary matrix and is not covered by the underlying ingredient slugs. Flagging broth as a candidate addition to the matrices controlled vocabulary for Karen’s review.
• Missing seaweed and bivalve data — the previous Key numbers focused only on mercury and arsenic and omitted the substantial seaweed/bivalve nutrient and Cd/Pb findings, which are a major part of the paper.
• Missing key finding — the previous version stated “iAs <6% in fish broth” but the <6% iAs figure is for fish tissue ingredients, not broth; the broth speciation values are different (AsB 85–88% in fish broth, As-Sug 65% in seaweed broth, iAs(V) 29% in clam broth — the last being the notable exception).
• Invented CRM — previous version cited DOLT-5 as a quality-control material; the paper uses DORM-2, DORM-4, TORT-3, SO-2, SRM-3232, and PTC. DOLT-5 is not mentioned.
• Lucky Iron Fish (LIF) is named once in Methods because the device is the explicit subject of one of the four experiments in the paper and the finding cannot be discussed coherently without the identifier. Per Part 12 strict reading this is a borderline case; the device is a social-enterprise public-health product evaluated against its own marketing claim, not a consumer-product brand under contamination ranking. Flagging for audit.
• Legacy headings — replaced ## Wiki pages updated on ingest (legacy) with a verification-notes section per current schema.
• bivalve-molluscs added to ingredients (paper substantively measures mussels and clams).
• Ca, Mg, Zn added to metals frontmatter (paper measures these alongside the heavy-metal panel as nutrient elements; consistent with the existing inclusion of Fe, K, Se).
• access_url added (CC BY 4.0 open access via DOI).

Audit application 2026-05-18 (fresh-context subagent, verdict REVISE):

Findings applied:

• Lake trout cheek/tongue/brain THg attribution corrected. Previous version had “Lake trout brain: 0.03–0.3 µg/g ww”; PDF p. 8 attributes 0.03–0.3 to Arctic char cheeks, not lake trout brain. Lake trout tongue and brain are reported only as a combined range (0.154 to 0.640 µg/g ww). Arctic char tongue/brain reported separately as <0.002 µg/g ww.
• Arctic char tongue/brain As-speciation row corrected. Previous version had labels inverted (“mainly AsB | with <13% As-Sug, <2% AsL”); PDF p. 8 reads “tongues and brains contained AsB (<13 %) and As-Sug (<2 %), but a large amount remains unknown.” Corrected to AsB <13%, As-Sug <2%, AsL/iAs not specified, unknown fraction noted.
• Pb broth exception language corrected: “except one lake trout broth” → “except a few lake trout broths” per PDF p. 8.

Findings rejected (false positives):

• Audit flagged “Cd in broths: all below the potable-water RMRV of 7 µg/L” as a missing exception. Verified against PDF p. 8: the source reads “THg (1 µg/L, except one lake trout broth), Cd (7 µg/L), and Pb (5 µg/L, except a few lake trout broths).” The “except one lake trout broth” parenthetical applies to THg, not Cd; Cd is parenthesized alone with no exception. Wiki claim is correct as written.
• Audit flagged the Arctic char broth row of Table 1 as having mismatched values (Ca, Fe, Zn). Verified against PDF p. 6 Table 1 directly: Arctic char Broth (NF), n=4: K 26 ± 5, Ca 12.0 ± 1.9, Mg 14.5 ± 0.8, Fe 4.4 ± 2.0, Zn 3.2 ± 2.0, Se 94 ± 20. Wiki matches exactly. The audit subagent appears to have parsed the table with a one-column offset.
• Audit flagged drinking-water as a non-snapshot matrices slug. Verified against docs/gpt-collaboration/system-prompt.md matrices vocabulary list: “drinking-water” is in the controlled vocab.
• Audit deferred the Lucky Iron Fish naming as a borderline Part 12 call. Decision retained: the device is the explicit subject of one of the four experiments in the paper and is being evaluated against its own marketing claim (not ranked against competitors for contamination). Without naming it, the finding cannot be traced to the cited Armstrong 2016 and Charles 2012 references. The Part 12 spirit (no brand-ranking-by-contamination) is preserved because the value attached is non-contamination (sub-marketed iron release). Karen-level policy override welcome.

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.