Takanashi et al. 2018 — Aluminium in baking-powder-containing processed foods, Japan 2015–2016

A survey by the Tokyo Metropolitan Institute of Public Health, Tama Fuchu Public Health Center, and the Tokyo Metropolitan Bureau of Social Welfare and Public Health measured aluminium (Al) content in 123 processed-food products (bakery products, confectionery, flour-processed foods other than bakery/confectionery, and pre-mix powders) distributed in Tokyo markets in 2015–2016. Products were selected because they were likely to use baking powder containing aluminium-potassium-sulfate (alum). Al was detected in 41 of the 123 samples at concentrations of 0.01 (limit of quantitation) to 0.40 mg/g wet weight. Four samples reached levels at which a single 1-piece serving consumed once per week would exceed the JECFA-2011 provisional tolerable weekly intake (PTWI) of 2 mg/kg body weight/week for a 16-kg young child (32 mg/week). The detection rate in non-Japanese confectionery had declined from 80% in the authors’ 2009 survey to 14% in 2015–2016, consistent with industry substitution of alum-free leavening agents, while Japanese confectionery (wagashi) remained at 56% detection and flour-processed foods such as okonomiyaki, chijimi, and takoyaki at 60% detection. Imported takoyaki — all of Chinese origin in this dataset — had higher Al than domestic takoyaki (0.09–0.38 mg/g vs. 0.01–0.10 mg/g among detects).

Key numbers

• Total samples: 123 products in 4 groups — bakery products (n=15), confectionery (n=71; subdivided into Japanese 36, Western 25, Chinese 8, snack 2), flour-processed foods other than bakery/confectionery (n=20; chijimi 2, okonomiyaki 5, takoyaki 13), and pre-mix powders (n=17).
• Detection (≥0.01 mg/g LOQ): 41 of 123 products (33%).
• Concentration range across detects: 0.01–0.40 mg/g wet weight.
• Highest values: imagawayaki-1 (Japanese pancake/cake, Japan origin) 0.40 mg/g; takoyaki-1 (China origin) 0.38 mg/g; takoyaki-2 and takoyaki-3 (China) 0.35 mg/g; deep-fried bread-1 (China, “no labelling”) 0.22 mg/g; chijimi-1 (Vietnam) 0.22 mg/g.
• LOQ: 0.01 mg/g (per the Shokuhin Eiseigaku Kensa Shishin Shokuhin Tenkabutsu-hen 2003, “alum-class testing method”). Values <0.01 mg/g reported as ND.
• Detection rate by group: bakery products 13% (2/15); confectionery 35% (25/71); flour-processed foods 60% (12/20); pre-mix powders 12% (2/17).
• Detection rate within confectionery sub-classes (2015–2016): Japanese confectionery 56% (20/36); non-Japanese confectionery — Western, Chinese, and snack pooled — 14% (5/35).
• Trend in non-Japanese confectionery detection rate: 2009 80% (20/25) → 2011–2014 28% (18/65) → 2015–2016 14% (5/35).
• Four samples whose single 1-serving-per-week intake exceeds the 16-kg-child PTWI of 32 mg Al/week: deep-fried bread-1 (0.22 mg/g × 192 g/piece × 1 piece = 42 mg); takoyaki-1 (0.38 × 19 × 6 = 43 mg); takoyaki-2 (0.35 × 19 × 6 = 40 mg); takoyaki-3 (0.35 × 18 × 6 = 38 mg). Takoyaki serving size assumed at 6 pieces (typical retail portion of 6–8 pieces).
• Five additional samples whose single 1-serving-per-week intake approaches the child PTWI and whose 2-serving/week intake exceeds it: imagawayaki-1 (0.40 × 76 × 1 = 30 mg); chijimi-1 (0.22 × 104 × 1 = 23 mg); takoyaki-4 (0.20 × 22 × 6 = 26 mg); takoyaki-5 (0.18 × 26 × 6 = 28 mg); takoyaki-6 (0.12 × 34 × 6 = 24 mg).
• Reference exposure benchmarks cited by the source: JECFA-67 (2006) PTWI 1 mg/kg bw/week (revised down from the 33rd JECFA 7 mg/kg bw/week); JECFA-74 (2011) revised PTWI to 2 mg/kg bw/week (NOAEL 30 mg/kg bw/day basis); Japan Food Safety Commission set TWI 2.1 mg/kg bw/week in December 2017. For a 16-kg child the 2 mg/kg/week PTWI corresponds to 32 mg/week. The Ministry of Health, Labour and Welfare 2011–2012 market-basket survey reported background Al intake from food at 0.863 mg/kg bw/week for 1–6-year-olds, below the PTWI on average.
• Imported takoyaki concentrations (China-origin Al-detected): 0.09–0.38 mg/g; domestic Japanese takoyaki concentrations (Japan-origin Al-detected): 0.01–0.10 mg/g.
• Same-product re-purchase (2009 vs 2015–2016): 4 products re-tested. Steamed bread and sablé were essentially unchanged; both snack-food products had dropped below the LOQ (one product removed baking-powder labelling entirely; the other retained baking-powder labelling but no longer reported detectable Al).
• China’s per-food-class Al limits cited in the discussion (per JETRO summary the authors reference): 100 mg/kg (0.1 mg/g) for bean products, batters, fried noodles, shrimp crackers, and bakery products; 500 mg/kg (0.5 mg/g) for salted jellyfish. Takoyaki is not on the listed regulated-food schedule.
• Source-side reformulation projection: if MHLW’s then-proposed maximum aluminium-ammonium-sulfate use rate of 0.1 g Al/kg in bread and confectionery were applied, no product in this survey would exceed the 16-kg-child PTWI even at 2 servings/week (worked example: imagawayaki-1 0.1 mg/g × 76 g/piece × 2 pieces = 15 mg vs 32 mg PTWI).

Methods (brief)

About 0.5 g of homogenized sample was weighed into a microwave-digestion vessel with 4 mL of HNO3 (sp. gr. 1.42, Wako Pure Chemical Industries, “tokkyu” grade) and 4 mL of ultrapure water, digested in a Milestone General ETHOS One microwave digestion system, brought to exactly 50 mL with water, and filtered through a 0.45 µm filter. Frozen foods were thawed per package instructions and allowed to cool before sampling. Al was measured by inductively coupled plasma optical emission spectrometry (ICP-OES; PerkinElmer Optima 7300DV). Calibration used Al standard (Kanto Chemical, 1000 µg/mL atomic absorption grade) diluted in HNO3 (4→50) across five concentrations covering 0.1–10 µg/mL Al with r² > 0.999; routine quantitation used a two-point calibration with HNO3 (4→50) blank and a 10 µg/mL Al standard. Method validation followed the MHLW Guidelines for Evaluation of Analytical Methods for Metal in Foods (Notification Shokuan-Hatsu 0926001, 26 September 2008). The 0.01 mg/g LOQ is the quantitation limit specified in the Shokuhin Eiseigaku Kensa Shishin Shokuhin Tenkabutsu-hen 2003 alum-class testing method.

Implications

This survey contributes occurrence data for aluminium in baking-powder-containing processed foods in a high-income market (Japan) where regulator-led use-reduction guidance has been in place since 2013. It documents two distinct patterns within one category: (a) substantial reduction of Al-containing food-additive use in Western confectionery, Chinese confectionery, and snack-food formats over 2009 → 2015–2016, consistent with industry response to the MHLW reduction notice; and (b) persistent high detection in Japanese confectionery (wagashi) and in flour-processed foods such as takoyaki, okonomiyaki, and chijimi, where the reduction had not propagated. Within flour-processed foods, Chinese-origin imported takoyaki concentrations were substantially higher than Japanese-origin product, consistent with origin-country differences in alum permissibility for that product class.

For threshold-setting work the page contributes Al occurrence data in mg/g wet weight by product type, the per-piece serving sizes used by the authors for child-PTWI exceedance calculation (deep-fried bread 192 g, imagawayaki 76 g, chijimi 104 g, takoyaki 18–34 g across products), and the source-author worked example of a 0.1 g Al/kg use cap producing no PTWI exceedance even at 2 servings/week.

For app work the presence of baking powder, alum, potassium alum, or aluminium-potassium-sulfate on an ingredient list — or the generic “leavening agent” without speciation — is a label-derivable marker for potential Al exposure in baked, fried-flour, and pancake-format products. Confectionery sold as Japanese-style (wagashi) and fried-flour products such as takoyaki, okonomiyaki, and chijimi carry higher prior probability of Al detection in this dataset than Western confectionery or pre-mix powders.

Wiki pages this source may touch

• aluminum
• bread-and-baked-goods
• snacks-crackers-biscuits

Verification notes

• 2026-05-29 merge-enhance: existing page (dated 2026-05-14) was structurally correct but carried several frontmatter and routing defects; the underlying numerical content was verified against the PDF and is preserved.
  • raw_handle corrected from generic manual-fetch-kimi to the per-file MFK_ form MFK_survey-of-aluminium-content-of-processed-food-usin (matches the abdolahpour2023 / aguilar-miranda2024 MFK_ convention adopted post-2026-05-17).
  • raw_path filename corrected from “…Baking Powder.pdf” to the actual filename “…Baking Powder (2015-2016).pdf” — the prior path did not resolve.
  • Added access_url to the JSTAGE DOI URL.
  • Added raw_sha256: 6f74f4a2ab52ffe079fd5d4c25dc2da8442276a30347d3166040734848004a1f for the verified PDF.
  • source_type changed from non-vocabulary government-report to peer-reviewed. The paper is a peer-reviewed article in Shokuhin Eiseigaku Zasshi (the same journal as watanabe2017-retail-fish-mercury-methylmercury-japan, which uses peer-reviewed); the conducting institution being TMIPH does not change the publication venue’s classification.
  • Added sampling_year_range: "2015-2016" per the title and Methods.
  • sample_population clarified to name all four product groups and the alum-selection criterion.
  • Body link [[metals/aluminium]] corrected to [[metals/aluminum]] — the existing wiki metal page is wiki/metals/aluminum.md; no aluminium.md exists.
  • Body link [[regulations/jecfa-al-ptwi-2011]] removed from “Wiki pages this source may touch” because no such regulation page exists. The JECFA-74 (2011) PTWI revision and the Japan FSC 2017 TWI setting are documented in the Key numbers section; a JECFA-Al PTWI regulation page is a Karen-approved page-creation (Part 10 / skill stop condition), not auto-created during ingest.
  • Heading ”## Wiki pages updated on ingest” renamed to ”## Wiki pages this source may touch” per current convention (the prior heading is a legacy phrasing that implied the ingest did downstream edits, which the routing layer now owns).
  • Implications section reworded to remove the prior version’s two Part 2 wiki/HMTc-firewall slips: “HMT&C product standards for baked goods should flag Al as a process-additive analyte” and “requiring Al-free leavening agents in certified products is feasible.” Both proposed HMTc design choices rather than reporting what the source contributes. Reworked to describe what occurrence data and serving-size assumptions the source contributes for downstream threshold work, without proposing a threshold or a supplier-specification lever.
  • Methods section expanded to name the digester (Milestone General ETHOS One), the HNO3 vendor (Wako Pure Chemical Industries), the standard vendor (Kanto Chemical), and the specific MHLW guideline citation (Shokuan-Hatsu 0926001 of 2008). These are scientific-method vendor names (Part 12 Exception 2), not contamination-value brand attributions.
  • Key numbers expanded with the within-confectionery sub-class breakdown (Japanese 56% vs non-Japanese 14%), the full 2009 → 2011–2014 → 2015–2016 trend percentages, the 4 PTWI-exceeding-at-1-serving samples with per-piece weights, the 5 PTWI-approaching samples, the China per-class limits the authors cite, and the source-author’s worked example of a 0.1 mg/g Al use cap.
• 2026-05-29 retained: cite_key, doi, all measured Al values, metals: [Al], matrices, jurisdictions: [JP], sample_n: 123, evidence_tier: A. The `

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.