Yoshinaga and Ogawa 2025 — Duplicate diet study: dietary metal exposure in Japanese adults

This duplicate diet study measured 12 metals (V, Co, Ni, Rb, Sr, Mo, Ag, Cd, Sb, Cs, Ba, and Pb) in 150 complete 24-hour dietary samples from adult Japanese, collected 2017–2019. Each participant provided a duplicate portion of every food and beverage consumed across the sampling day, which was homogenized, freeze-dried, acid-digested, and analyzed by ICP-MS. The duplicate diet approach measures what people actually consumed rather than estimating intake from food composition tables and market basket statistics. The authors also estimated the contributions of incidental soil and house-dust ingestion to total oral metal exposure and compared dietary intakes against US EPA Reference Doses.

Key numbers

Median daily intake from diet (Table 4, N=150):

*One-half the LLD substituted for non-detect samples in mean/SD; Cd values recapitulated from Yoshinaga and Ogawa, 2024. Non-detect counts: Ag n=83, Sb n=49, Pb n=65.

Median metal concentrations in dried duplicate diet (Table 3, N=150):

Moisture content of the 150 diets ranged 69.6–92.5 %; mean dry-matter fraction 0.158 (factor 0.842 reported in the paper for the conversion from dried to fresh basis).

Body-weight-corrected daily intake and RfD comparison (Table 5, N=150):

Median intakes were below the EPA RfD for all metals for which an RfD is available; Co, Rb, Cs, and Pb lack EPA oral RfDs. The maximum Mo intake (8.03 µg/kg/day) exceeded the RfD of 5 µg/kg/day; the population median (2.53 µg/kg/day) is ~51 % of the RfD, leaving a narrow margin. The authors also note that the minimum Mo intake among the 150 participants (30 µg/day, from the Table 4 range minimum of 30.3 µg/day) exceeded the Japanese MHLW daily Mo requirement of 25 µg/day, so every participant met or exceeded the nutritional requirement.

Diet vs soil and house-dust contributions (Table 6):

Soil and house-dust intakes computed from median environmental concentrations (Takeda et al. 2004 for soil; Yoshinaga et al. 2011 for house dust) and US EPA 2011 default adult ingestion rates (10 mg/day soil, 20 mg/day house dust). Diet contributes essentially all oral exposure for most metals; for V, Sb, and Pb, dust and soil together contribute 20–34 %.

Lower limits of detection (Table 2): in dried diet, µg/kg — V 1.8, Co 0.71, Ni 30, Rb 0.6, Sr 15, Mo 5.0, Ag 1.9, Cd 0.45, Sb 1.5, Cs 0.17, Ba 13, Pb 6.2. Corresponding daily-intake LLDs (µg/day) — V 0.68, Co 0.26, Ni 11, Rb 0.22, Sr 5.5, Mo 1.8, Ag 0.71, Cd 0.17, Sb 0.56, Cs 0.062, Ba 4.8, Pb 2.3.

Pb time-series context (Figure 1): Adult Japanese Pb intakes by sampling year — duplicate-diet studies (dark bars) reported 5.5 µg/day (2008), 4.5 (2015), 3.5 (2016), 2.24 (this study, 2018), nominal trend continued in 2019, ≈4.3 (2023); NIHS market-basket estimates (diagonal-hatched bars) consistently higher: ≈18 (2008), 12 (2015), 10 (2016), 10 (2018), ≈9 (2019). Methodological gap between approaches persists.

Method validation (Table 1): ICP-MS measurements against NIES CRM No. 27 (Typical Japanese Diet) — measured vs certified/reference (units per table): V 27.7±2.1 µg/kg (no cert), Co 20.3±0.7 µg/kg vs 22, Ni 380±25 µg/kg vs 390, Rb 4.64±0.08 mg/kg vs 4.7, Sr 4.97±0.20 mg/kg vs 4.9, Mo 0.444±0.009 mg/kg vs 0.43, Ag 5.88±0.41 µg/kg (no cert), Cd 0.063±0.003 mg/kg vs 0.069±0.009 (recapitulated from Yoshinaga and Ogawa 2024), Sb 4.98±0.58 µg/kg (no cert), Cs 20.1±0.7 µg/kg vs 20, Ba 1.14±0.03 mg/kg vs 1.1±0.1, Pb 0.592±0.025 mg/kg vs 0.62. Means agreed well with certified/reference values across the 13 repeat analyses.

Methods (brief)

24-hour duplicate diets were collected from 150 adult volunteers (65 males, 85 females, 19–83 years; 89 northern Kanto, 49 southern Kanto, 12 elsewhere in Japan) recruited 2017–2019 via newspaper advertisement and Toyo University staff/student networks. Each duplicate sample (sample weights 918–4,766 g/day; moisture 69.6–92.5 %) was homogenized in a 3 L domestic food processor (Cuisinart), freeze-dried, and pulverized. 50 mg of freeze-dried sample was nitric-acid digested (ultrapure HNO₃, Kanto Chemicals Ltd., Tokyo, Japan) in a 4.5 mL screw-capped Teflon vessel within a 25 mL Teflon outer vessel jacketed in stainless steel (Okamoto and Fuwa 1984 double-digestion method). Digests were analyzed by ICP-MS (Agilent 8800, Agilent Technologies Ltd., Japan) at the National Institute for Environmental Studies (NIES), Ibaraki, Japan, with helium collision gas (4.3 mL/min) and on-line indium (m/z 115) internal standardization. Each digestion series (n=20) included 2–3 procedural blanks, 2 sub-samples of NIES CRM No. 27 Typical Japanese Diet, and 15–16 participant samples; lower limit of detection was the 3-σ of the 16 blanks. Cd values were recapitulated from the same group’s prior study (Yoshinaga and Ogawa 2024). Distributions were non-normal (Kolmogorov–Smirnov test p<0.05); median (min–max) is the primary descriptor.

Implications

This duplicate-diet dataset provides Japan-specific adult dietary exposure benchmarks for 12 metals, computed from direct measurement of the consumed food matrix rather than from market-basket composites. The study reports diet-only oral exposure and separately estimates soil and house-dust contributions, giving an aggregated oral exposure estimate for each metal. For Pb, diet contributes 66 % of total oral exposure (2.24 µg/day from diet vs 3.4 µg/day total); for V and Sb, diet contributes 77–80 %; for the remaining metals diet is essentially the entire oral source. Maximum Mo intake exceeds the EPA RfD; the population median is at 51 % of the RfD, and median intake exceeds the Japanese MHLW daily requirement, so the margin is narrow on the high side rather than on a deficiency basis. The duplicate-diet Pb medians sit below the time-series of Japanese NIHS market-basket estimates by a factor of 2–4 across the years compared in Figure 1; the authors note that the methodology gap is unresolved.

The recruitment was non-random (newspaper plus university convenience sampling, mostly Kanto District) so the results are not a population mean; they are a snapshot of one regional adult sample. Pb was non-detect in 65 of 150 samples and the LLD (6.2 µg/kg dried diet) is high relative to clean-room work (Yoshinaga et al. 2017 reported 0.17 µg/kg under clean-room pretreatment), so the dietary Pb medians here are an upper-bound indicator rather than a definitive concentration measurement.

Wiki pages this source may touch

• lead
• cadmium
• nickel
• antimony
• vanadium
• cobalt
• molybdenum
• silver
• cesium
• barium

Verification notes

• 2026-05-25 (Claude session): Merge-enhance pass against the original PDF. Previous revision (dated 2026-05-14) had multiple defects: DOI was 10.5985/emcr.2025.0008 but the publisher record on page 1 of the PDF prints https://doi.org/10.5985/emcrr.20250008 (double-r in the journal abbreviation and yyyy0008 format). Corrected. raw_handle was manual-fetch-kimi (not a valid per-PDF handle); replaced with MFK_p0323-duplicate-diet-study-to-assess-dietary-expos. raw_path filename was truncated to P0323.pdf; replaced with the actual filename. Added raw_sha256 (computed locally), access_url, and no_doi_assigned: false. metals: was [Cd, Pb, Ni, Sb] (four of twelve measured); expanded to include V, Co, Mo, Ag, Cs, Ba (those with extant wiki/metals pages). Rb and Sr are measured but have no wiki/metals pages, so they are reported in Key numbers/Methods only.
• 2026-05-25 (Claude session): Key numbers expanded into the four published tables (Tables 3, 4, 5, 6) plus the validation table (Table 1) and LLDs (Table 2). Prior version condensed these into a bullet list and omitted the soil/house-dust contributions, the body-weight-corrected intakes, the moisture range, and the validation column. The exposure-vs-RfD comparison (Table 5) and aggregated oral exposure breakdown (Table 6) are the rows most useful for downstream synthesis pages and were missing from the prior revision.
• 2026-05-25 (Claude session): Removed prior policy-flavored editorializing in ## Implications (“Total diet study data from Japan is a relevant comparator for HMT&C exposure context”, “provides a baseline for how tightly the food supply needs to be controlled to achieve ongoing reductions”, course/app/certification audience sections). Replaced with neutral description of what the paper measured and what its limitations are (non-random sampling, high Pb LLD relative to clean-room work). Per Part 2, source pages report what one paper found; cross-source synthesis and HMT&C-context interpretation are downstream workflows.
• 2026-05-25 (Claude session): Renamed the legacy ## Wiki pages updated on ingest heading to ## Wiki pages this source may touch per current template convention. The list is candidate routing, not a record of completed downstream edits.
• 2026-05-25 (Claude session): Sample population string was generic; expanded with the actual recruitment breakdown (89/49/12 by region, age range 19–83, gender split) from page 132 of the PDF.
• DOI note: The publisher (Japan Society for Environmental Chemistry) uses the form 10.5985/emcrr.YYYYNNNN where NNNN is a 4-digit article-within-year identifier. Confirmed 10.5985/emcrr.20250008 resolves to this article via the DOI printed in the header of page 131.
• No-DOI cross-references: the authors cite Yoshinaga and Narukawa 2020 and Yoshinaga and Ogawa 2024 (the latter is the source of the recapitulated Cd values); these are prior duplicate-diet papers by the same group and may already be in the wiki under separate cite-keys. Not searched in this session; flagged for the synthesis pass.
• 2026-05-25 (audit subagent — applied): Fresh-context audit (v2 skill, general-purpose subagent) flagged three ❌ findings against the original PDF. All three independently re-verified against the source and applied:
  • Mo median misstatement on PDF p. 134 — wiki had written “Median Mo intake (30 µg/day, Table 4) also exceeded the Japanese MHLW daily requirement of 25 µg/day”. The PDF actually says “Even the minimum intake of the duplicate diet sample analyzed in this study (30 µg/day, Table 4) exceeded the requirement”. Corrected the wiki bullet to describe this as the population minimum, not the median; the Table 4 Mo median is 169 µg/day, which is correctly tabulated.
  • Pb censoring in Table 5 — wiki Table 5 row had (0.039–0.705) but PDF Table 5 shows (<0.039–0.705) with the lower bound marked non-detect. Restored the < censoring indicator (the * non-detect-substitution footnote in the mean column already applied; the censoring on the min in the median(min–max) column was the missing detail).
  • Sample population geography — wiki sample_population said “12 elsewhere (6 central Japan, 6 western Japan)“. PDF p. 132 reads “A total of 13 participants were from outside the Kanto District, i.e., six from central Japan and seven from western Japan.” Corrected to 13/6/7 to match the paper. Note that the three regional counts (89 + 49 + 13 = 151) sum to one more than the stated total N=150; this is a paper-internal off-by-one and is documented in the sample_population field rather than silently reconciled.
• 2026-05-25 (audit subagent — matrices vocabulary, noted only): Audit subagent flagged ⚠️ on matrices: [total-diet-study, duplicate-diet] because these bare-string matrix labels are not in the taxonomy snapshot provided to the audit subagent. The matrix vocabulary is documented in docs/gpt-collaboration/system-prompt.md, not in the taxonomy snapshot (which only enumerates ingredient/product/metal/regulation slugs). The label total-diet-study is in active use in the wiki (wiki/sources/spungen2019-fda-tds-children-lead-cadmium-2014-2016.md), and the routing layer reports this source as severity: advisory (missing optional products/ingredients fields), which is the correct treatment for a whole-diet study that does not measure individual products or ingredients. No change applied; the audit was flagging the absence of the matrix vocabulary in its reference file, not a defect in this page.

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.