Katrynska et al. 2026 — Human milk as a biomonitor of toxic metal exposure

This structured narrative review with systematic search elements synthesizes 112 studies (2010–2025) on heavy metal contamination in human milk, with particular emphasis on maternal–infant transfer pathways, environmental and dietary determinants of exposure, and geographical variability. The review covers Pb, Cd, tHg (including MeHg in fish-eating populations), tAs, Cr, and Al. Under typical (low-exposure) conditions the authors report population-level concentration ranges of approximately 2–5 µg/L for Pb, 1.4–1.7 µg/L for tHg, and below 1 µg/L for Cd. Substantially higher values occur in industrialized and mining regions, with extreme isolated reports exceeding 1000 µg/L for Pb (Turkey, max 1515 µg/L) and ~104 µg/L for tHg (Amazon, max). The review frames human milk as a context-dependent biomarker reflecting both current environmental exposure and long-term mobilization of maternal stores during lactation (bone resorption for Pb, adipose mobilization for lipophilic organometals). Infants are described as a particularly vulnerable subpopulation due to immature hepatic and renal detoxification, higher per-body-weight intake, and increased gastrointestinal absorption (Pb absorption is several times higher in infants than adults). The authors conclude that breastfeeding’s benefits clearly outweigh its risks under typical exposure conditions but flag industrial/mining-area and chronic-water-contamination settings as warranting closer monitoring.

Key numbers

All concentrations µg/L breast milk unless otherwise noted. Values reproduced as reported by the review; original underlying study citations preserved in the review’s reference list (numbers 119–198).

Typical (background-population) ranges (review-wide summary):

Table 2 — Geographical variability of heavy metals in human milk (review’s own structured comparison):

Additional geographic and longitudinal datapoints reported in the body:

• Poland (Bzikowska-Jura group and others): mean breast-milk Pb 6.33 µg/L; mean Cd 2.11 µg/L. Maternal-age effect: mean Pb 4.76 µg/L in women aged 20–25 → 7.41 µg/L in women aged 36–40.
• Sweden, post-leaded-gasoline ban: Pb in breast milk declined from 17 µg/L (1989) to ~1 µg/L (2000–2009).
• Hungary (Budapest): Pb in breast milk declined ~90%, from 14.9 µg/L (1991) to 1.74 µg/L (2017), after leaded-gasoline phase-out, water-system modernization, and lead-pipe elimination.
• Taiwan (lactation-stage decline): Pb, Al, and tAs in milk may decline by 75–90% within the first two months of breastfeeding, attributed to declining protein content and reduced metal–protein binding.
• Chile (Arica, polymetallic waste site): mean B 270 µg/L (vs Santiago control 38 µg/L); mean tAs 0.36 µg/L.
• China (occupational lead exposure): mean Pb in colostrum 52.7 µg/L in occupationally exposed women vs 4.7 µg/L in non-exposed.
• Greece (urban vs rural): mean Pb 0.48 µg/L urban vs 0.15 µg/L rural.
• Nigeria (Owerri): mean Pb ~38 µg/L (sub-Saharan urban high).

Lifestyle and biological determinants (quantified associations):

• Maternal tobacco smoking: breast-milk Cd in smokers 37% to ~300% higher than in non-smokers. Single cigarette contains ~0.1–0.2 µg Cd. Passive smoking >15 min/day associated with elevated Al and Cd in milk.
• Fish/seafood consumption (Norway): regular marine-product consumers had mean tHg ~42% higher than non-consumers; lean-fish consumption associated with ~32% higher tHg; halibut and shellfish elevated tHg further.
• Cadmium–processed-potato correlation: r = 0.502 between Cd in milk and consumption of potato-based products (chips).
• Cosmetics/lipsticks–Pb correlation: regression coefficient β = 0.32 between use of lipstick/color cosmetics and breast-milk Pb.
• Dental amalgam: women with ≥1 amalgam filling had milk tHg ~3× higher than those without.
• Parity: primiparous women had higher milk Pb, Cd, and tHg than multiparous women (interpreted as gradual depletion of maternal stores across successive lactations).
• Iron-deficiency anemia: higher milk Pb due to shared DMT1 transporter (Fe deficiency upregulates absorption of divalent metals including Pb).

Lactation-stage patterns:

• Highest milk concentrations of Pb, tAs, Al, and tHg observed in colostrum, declining sharply through transitional and mature milk.
• Cd is the exception: stable or rising through mature milk, reaching highest values at 4–6 months of lactation (attributed to active Cd transport in mammary cells and competition with Zn/Ca on shared transporters).
• Trace nutrients also decline through lactation: Zn from ~4.9 mg/L (colostrum) to ~2.9 mg/L (transitional milk).

Extreme outliers (reported in the body):

• Highest single recorded Pb: 1515 µg/L (Turkey).
• Highest single recorded tAs: 149 µg/L (India).
• Highest single recorded tHg: 104.1 µg/L (Amazon, Brazil; mean 59.41 µg/L).

Methods (brief)

Design. Structured narrative review with systematic literature-search elements. Not registered, not formal PRISMA, no formal risk-of-bias assessment, no meta-analysis. Authors explicitly note “selection bias cannot be fully excluded.”

Databases and dates. PubMed, Scopus, Web of Science. Coverage 2010–2025. English-language only.

Search strategy. Boolean: (“human milk” OR “breast milk”) AND (“heavy metals” OR “toxic elements”) AND (“lead” OR “cadmium” OR “mercury” OR “arsenic” OR “aluminum” OR “chromium”) AND (“exposure” OR “transfer” OR “biomonitoring” OR “contamination”).

Inclusion / exclusion. Included: original research articles or reviews assessing heavy-metal occurrence/concentration in human milk from lactating women, with environmental, dietary, physiological, or lifestyle determinants. Excluded: animal-only studies, papers without quantitative or descriptive data, conference abstracts, editorials, non-peer-reviewed publications, and studies without full-text access.

Yield. ~520 records initial → 148 full-text-screened → 112 included in qualitative synthesis.

Analytical methods covered in the included studies (Section 4.6). ICP-MS treated as the gold standard for trace-element biomonitoring. Other methods discussed: flame AAS (FAAS), graphite-furnace AAS (GFAAS / ET-AAS), cold-vapor AAS (CV-AAS, for Hg specifically). Speciation: HPLC-ICP-MS hyphenation enables As (iAs vs organic As) and Hg (MeHg vs inorganic) separation.

Representative detection limits cited (ICP-MS, µg/L): Pb ~0.0051; Cd ~0.0100; tHg ~0.0060. GFAAS detection limits typically 0.01–1 µg/L. Recovery in validated methods 90–110%; acceptable CV 5–10%.

Limitations the authors call out. Lack of certified reference materials specifically for human milk; substantial inter-laboratory variability; lack of harmonization in lipid removal, digestion, and the fraction analyzed (whole milk vs aqueous phase); milk composition changes across lactation stages, affecting metal binding. The authors emphasize that direct comparison across studies is often limited.

Implications

Certification: This 2026 structured narrative review consolidates global breast-milk heavy-metal data through 2025 and provides typical-population reference ranges (Pb 2–5 µg/L, tHg 1.4–1.7 µg/L, Cd <1 µg/L) along with the geographic outlier landscape (Turkey, India, Iran, Nanjing, Ghana, Nigeria, Spain Murcia, Amazon). For breastmilk contamination-profile review, this is a high-value reference paper because it both anchors typical-condition baselines and documents the exposure scenarios (mining-area, leaded-water, fish-consumption, occupational, tobacco-smoking) that drive the upper tail.

Courses: Provides the integrative framework for teaching maternal–infant heavy-metal transfer via breastfeeding: environmental sources → maternal accumulation (bone for Pb, adipose for lipophilic organometals, kidney for Cd) → mobilization during lactation → infant exposure. The bone-mobilization mechanism for Pb is particularly important for teaching because it means women with historic occupational, environmental, or leaded-gasoline-era Pb exposure can transfer Pb to infants even when current dietary exposure is low. The Hungary-Budapest 90% decline and Sweden post-leaded-gasoline decline are good case studies for environmental-regulation effectiveness. The colostrum vs mature-milk pattern (Pb/Al/tAs decline 75–90% in first two months; Cd rises through 4–6 months) is a useful teaching point for understanding the protein-binding and active-transport mechanisms behind metal partitioning into milk.

App: Breast milk is not in the app’s packaged-food ingredient-list model but this review anchors the health-section content on infant exposure via lactation and the maternal-determinants picture (diet, smoking, parity, age, occupational exposure, iron status, dental amalgam, cosmetics).

Microbiome: Exposure of the infant gut to Pb, Cd, and Al during the early colonization window (0–6 months) via breast milk is relevant background for metal–microbiome interaction reviews; the colostrum-peak pattern means the highest exposure of the infant gut microbiome to Pb/tAs/Al typically occurs in the same window as the foundational microbial colonization events.

Wiki pages this source may touch

• breastmilk
• lead
• cadmium
• mercury-total
• arsenic-total
• aluminum
• chromium

Verification notes

2026-05-19 (Claude Code, merge-enhance from prior 2026-05-14 page). The prior revision had a thin Key numbers section that captured only the abstract-level qualitative ranges (2–5 Pb, 1.4–1.7 tHg, <1 Cd) and none of the structured numerical content from Tables 1 and 2 or the body. Enhancement:

(a) Key numbers expansion. Added Table 2 reproduction (13 geographic / exposure-context rows with metal-specific reported values from the review’s own structured comparison), plus the body-level numerical findings: Poland Pb 6.33 µg/L mean / Cd 2.11 µg/L mean / Pb age-effect 4.76 → 7.41 µg/L; Sweden Pb 17 → ~1 µg/L decline; Hungary Pb 14.9 → 1.74 µg/L decline (~90%); Taiwan 75–90% decline in Pb/Al/tAs in first two months of breastfeeding; Chile Arica B 270 µg/L and tAs 0.36 µg/L; China occupational Pb in colostrum 52.7 vs 4.7 µg/L; Greece urban-vs-rural Pb 0.48 vs 0.15 µg/L; Nigeria Owerri Pb ~38 µg/L. Added quantified lifestyle/biological associations: maternal-smoking Cd 37–300% higher; Norway fish consumption tHg ~42% higher (~32% for lean fish); Cd–chip correlation r = 0.502; cosmetics–Pb β = 0.32; dental-amalgam tHg ~3× higher; primiparous-vs-multiparous parity effect; Fe-deficiency–Pb DMT1 mechanism. Added the lactation-stage pattern (colostrum-peaks for Pb/Al/tAs; mature-milk-peak for Cd at 4–6 months; Zn decline 4.9 → 2.9 mg/L).

(b) Methods expansion. Added explicit Boolean search string, inclusion/exclusion criteria, study-yield funnel (520 → 148 → 112), the four analytical-method families covered by included studies (ICP-MS, FAAS, GFAAS/ET-AAS, CV-AAS), the HPLC-ICP-MS speciation note, ICP-MS detection limits cited by the review (Pb 0.0051, Cd 0.0100, tHg 0.0060 µg/L), recovery and CV norms (90–110%, 5–10%), and the authors’ own limitations statements (lack of human-milk CRMs, lipid-removal/digestion heterogeneity, lactation-stage variation).

(c) Sample population field. Expanded sample_population from a one-line summary to specify the database list, language restriction, the 520 → 148 → 112 funnel, and the global distribution of included studies.

(d) Speciation discipline (CHECK 3 of the audit prompt). The review uses “Hg” and “As” umbrella terms because included studies vary in whether they speciate. Frontmatter retains tHg and tAs per the rule that absent paper-wide speciation, the page uses the total-metal slug. MeHg is referenced narratively (fish-consuming populations) but is not the metric reported in most included studies.

(e) Brand firewall (CHECK 4). The review does not name commercial product brands in its included findings, so no brand-firewall scrubbing was required. Cosmetics/lipsticks are referenced as a generic product category (not by brand name); dental amalgam is referenced as a procedure type (not by manufacturer). Methods section names no commercial instrument vendors (the included studies do, but the review summarizes without vendor attribution).

(f) Wiki/HMTc firewall (CHECK 5). The Implications section limits itself to what this review contributes (baseline ranges, exposure-scenario framework, teaching anchors) and does not propose HMTc threshold values, make cross-source consensus claims, or issue consumer risk advisories. The review’s own Section 5 (“Public Health Implications”) contains recommendations (limit predatory fish, balance diet with iron/calcium, smoking cessation) but these are framed as the authors’ recommendations from their own paper, not adopted into this wiki page as HMTc guidance or consumer advice.

(g) Frontmatter preservation. cite_key, raw_handle, raw_path, license, metals, ingredients, products: [], matrices, jurisdictions, near_duplicates, evidence_tier preserved from prior revision. updated advanced to 2026-05-19; sample_population expanded as noted in (c).

(h) Products field preservation. products: [] is correct: human milk is not a commercial packaged-food product category in this wiki’s taxonomy (no products/breast-milk or equivalent), and the routing layer fans this paper out via the ingredients/breastmilk slug and the six metal pages. The page does not declare a product because the source is an ingredient/biomarker review, not a packaged-product analysis.

(i) Prior-revision verification note preserved. The 2026-05-17 Codex cross-vendor audit note (added H1, replaced invalid metal wikilinks, modernized heading, softened “authoritative” language, removed microbiome new-page instruction) is retained in spirit: those corrections are baked into the present page, and the same disciplines are applied here.

Speciation summary (frontmatter): Pb, Cd (no speciation), tHg (total — paper aggregates Hg-form across included studies), tAs (total — paper aggregates As-form across included studies; iAs vs organic discussed narratively but not reported as a uniform metric), Cr (total — paper does not separate Cr-VI), Al.

Sample size: sample_n: null is correct because this is a review of 112 included primary studies, not a single primary study with a defined biological sample size; per-paper sample sizes are heterogeneous and the review does not aggregate them quantitatively.

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.