Blunden & Wallace 2003 — Tin in canned food: occurrence and effect review

Blunden and Wallace (ITRI Ltd, UK) review the literature on tin dissolution from tinplate food and beverage cans, surveyed occurrence values in retail canned foods, published acute clinical and human-incident reports of gastrointestinal effects following ingestion of tin-rich canned products, and the chronic-effect and absorption literature for divalent inorganic tin. The review’s central conclusion is that adverse gastrointestinal effects have been clinically documented at tin concentrations of approximately 700 mg/kg and above in canned foods or beverages, but that two human studies at higher concentrations reported no adverse effects, and that no published reports of acute adverse effects attributable to tin contamination in the 100-200 mg/kg range have appeared in the 25 years preceding the review. The authors estimate that approximately 4% of UK plain (unlacquered) tinplate food cans contain over 150 mg/kg tin and infer that the contents of roughly 25 million such cans containing at least 150 mg/kg tin are consumed annually in the UK alone without published acute-effect reports. The review supports the position that tin contamination at concentrations up to about 200 mg/kg in canned products is not associated with reported acute gastrointestinal effects, while noting that controlled clinical studies at these concentrations are absent. ITRI Ltd is the tin-industry research institute and the sponsorship context of the review is relevant when interpreting its overall framing.

Key numbers

Tin can production and use (pp. 1651-1652)

Regulatory limits cited (p. 1653, Section 2)

Occurrence in canned foods (Section 4)

Table 1 of the review (p. 1654, citing Ratana-Ohpas et al. 1996) reports concentrations of total tin in conserves of acidic fruit:

Additional occurrence values cited in the review body:

Storage-temperature effect on tin dissolution (Section 3, citing Calloway and McMullen, 1966)

Acute clinical studies (Section 6.1.1, Primary reports)

Acute clinical studies (Section 6.1.2, Secondary reports)

Acute episode reports (Section 6.2)

Reported general dose range associated with episode reports: 250-2,000 mg/kg, predominantly in acidic fruits and fruit juices. Symptom frequency (combined across reports): nausea 97%, abdominal cramps 87%, vomiting 70%, headaches 57%, diarrhoea 33%, fever 13%. Incubation 15-30 minutes, duration 30 minutes to 3 weeks (Piscator, 1979; Schafer and Femfert, 1984; Dewitte et al., 2001; Barker and Runte, 1972).

Non-human oral studies (Section 6.2.3, Table 2)

Acute oral LD50 of divalent inorganic tin compounds:

Cats given orange juice containing 540 or 1,370 mg/kg tin by stomach tube at 5 or 10 ml/kg bw showed vomiting; apple juice at 605 mg/kg and pear nectar at 750 mg/kg were not emetic (Benoy et al., 1971). A Japanese study reported vomiting in 5 of 6 cats given 10 ml/kg bw orange juice at 452 mg/kg tin but not at 337 mg/kg (Omori et al., 1973).

Chronic effects and absorption (Section 7)

Overview-section worst-case dose calculation (Section 8)

If all canned foods and beverages contained 200 mg/kg tin and a 1 kg daily diet consisted of 4-8% canned foods (MAFF 1998), average daily intake would be 8-16 mg/day or 0.13-0.3 mg/kg bw/day for a 60 kg adult, which is approximately 7-fold below the JECFA reference of 2 mg/kg bw/day (the daily equivalent of the JECFA Provisional Tolerable Weekly Intake of 14 mg/kg body weight per week; JECFA 1988a, 1988b). The UK is described as the highest per capita European consumer of canned food and the largest consumer of plain unlacquered cans (~2,500 million annually, MPMA 2000); approximately 4% probably contain at least 150 mg/kg tin, implying contents of ~25 million such cans containing this concentration are consumed annually in the UK without published acute-effect reports in the prior 25 years.

Evidence Fitness

This is a narrative literature review without explicit search-and-selection protocol or PRISMA-style inclusion criteria, sponsored by ITRI Ltd (the International Tin Research Institute). It contributes no primary heavy-metal measurements and no occurrence distributions of its own; all quantitative figures it presents are second-hand summaries of the underlying primary studies that should be obtained as primary evidence. Public evidence label “Context only” is appropriate.

The source supports literature-context claims about (i) the magnitude and variability of tin dissolution from unlacquered tinplate into canned acidic foods and beverages, (ii) the approximate concentration threshold (700 mg/kg and above) at which controlled clinical studies have detected acute gastrointestinal effects from divalent inorganic tin, (iii) the absorption-limited and local-irritation mechanism of inorganic tin GI effects, and (iv) the historical regulatory framework (JECFA PTWI 14 mg/kg bw, Codex 250 mg/kg solid / 150 mg/kg beverage, UK 200 mg/kg statutory). It should not be used as the basis for HMTc threshold work for tin in canned product categories, which should reference the primary occurrence studies (Biégo 1999, Greger and Baier 1981, Ratana-Ohpas 1996, MAFF/FSA UK surveys) and the primary clinical studies (Benoy 1971, Calloway and McMullen 1966, Johnson 1982) directly, and should account for the review’s industry-sponsored framing when interpreting its overall position.

Methods (brief)

Narrative literature review with no stated search strategy, inclusion/exclusion criteria, quality appraisal, or quantitative pooling. The authors restrict the review (with stated exceptions for particular relevance) to reports published after 1970, citing changes in can-making technology (elimination of soldered side-seams, improvements in filling procedures) as the rationale for the cutoff. The review organises evidence into seven thematic sections: introduction and packaging context, background tin chemistry, factors affecting tin dissolution, surveyed levels in foods, toxicology overview, acute effects (clinical studies and consumption episodes), and chronic effects (including absorption, distribution and excretion). No new measurements are reported. Quantitative figures in tables and prose are reproduced verbatim from cited primary sources.

Implications

Certification: Process and toxicology context only. The review supports the general claim that tin in canned acidic foods and beverages is associated with acute gastrointestinal irritation at concentrations of approximately 700 mg/kg and above in controlled studies, with episode reports at 250-2,000 mg/kg, and that the JECFA PTWI of 14 mg/kg bw/week is approximately 7-fold above the worst-case daily intake estimable from a diet at the 200 mg/kg statutory limit. HMTc threshold-setting for tin in canned product categories should reference the primary occurrence and clinical studies directly rather than this review, and should treat the review’s “little evidence of acute effects at up to 200 mg/kg” framing with the caution appropriate to industry sponsorship.

Courses: A useful entry point to the tin-canned-food acute-effect literature for course modules covering the local-irritation versus systemic-absorption distinction (tin is one of the few heavy metals where the principal acute effect is gastrointestinal irritation and absorption is below ~3% for divalent forms), the relationship between can construction (unlacquered tinplate, nitrate-driven detinning, storage temperature) and tin migration, and the regulatory landscape for tin in canned foods. The review provides convenient pointers to the primary studies and should be presented to learners alongside source-sponsorship context.

App: No ingredient contamination_profile impact directly from this review. Per-product caution flags for canned acidic juices and fruits in unlacquered tinplate, particularly after extended storage at warm temperatures, are supported by the underlying primary literature this review compiles.

Wiki pages this source may touch

• tin
• tin-inorganic
• tomato
• canned-tomatoes
• orange-juice
• grapefruit-juice
• pineapple
• peach
• asparagus
• guava
• mango
• canned-foods-general
• canned-fruit
• canned-tomatoes
• canned-vegetables
• fruit-juices-non-apple
• fruit-juices-apple-containing
• codex-cxs-193-1995-tin-canned-foods

Verification notes

• 2026-06-04 fresh ingest (Claude Opus 4.7, manual-fetch ingest from raw/Manual Fetch Kimi /June 3 Folder/): DOI 10.1016/S0278-6915(03)00217-5 verified against the article’s first page; received 24 April 2003, revised 1 July 2003, accepted 5 July 2003; published in Food and Chemical Toxicology 41 (2003) 1651-1662. The article is an Elsevier copyrighted review (no Creative Commons licence statement on the PDF); recorded as subscription-access. SHA256 of the source PDF recorded in frontmatter.
• Author affiliation: ITRI Ltd, Unit 3, Curo Park, Frogmore, St Albans, Hertfordshire AL2 2DD, UK. ITRI Ltd is the International Tin Research Institute, the trade-funded research body of the tin industry. The review acknowledges assistance from APEAL (European Steel Association), Ledepp-Arcelor Group, BIBRA International, Crown Cork & Seal, Shell International, and others. The sponsorship context is preserved in the page text (Evidence Fitness and Implications sections) so downstream readers can weight the review’s framing accordingly. This is a structural source-context disclosure, not a brand attribution to a contamination value; it does not violate Part 12.
• Evidence tier B reflects the narrative-review character of the source (no primary measurements, no PRISMA protocol, no quality appraisal of cited primary studies, industry sponsorship). evidence_fitness: EF-4 and public_evidence_label: Context only reflect that the review contributes no occurrence values usable for HMTc threshold work.
• No brand names are reported in the review’s body content for canned foods being measured (the implicated products are described by category: canned peaches, canned tomato juice, canned salmon, canned rhubarb, orange juice, grapefruit juice, etc.). The review does name two industrial container vendors in the Acknowledgements (Crown Cork & Seal, Shell International) and the Metal Packaging Manufacturers Association data source (MPMA 2000); these are not brand-attributions of contamination values and are retained only as bibliographic context.
• The Kuwait outbreak (Metal Box Co., 1967) and the Washington/Oregon canned tomato juice incident (1969, Barker and Runte 1972) are public-record regulatory events documented in the review. Brand names are not provided in the review for these episodes; the review uses geographic and product-form descriptors throughout.
• Speciation convention: the review treats inorganic tin (divalent Sn²⁺ as primary, with Sn⁴⁺ noted for absorption studies) throughout; organotins are explicitly excluded as outside the scope of canned-food tin migration (Section 5, p. 1655). Frontmatter records Sn only with species: [Sn2+, Sn4+].
• jurisdictions: [UK, EU, US, international] reflects the geographic scope of cited primary studies: UK MAFF/FSA surveys, EU/European production figures, US FDA/Greger/Calloway data, JECFA international guidance.
• matrices: [canned-food, canned-beverage, canned-fruit-juice] follows the convention established by eu-scf-2002-acute-tin-canned-foods.md and jecfa-64th-cadmium-tin-2006.md.
• Ingredient slug selection: only slugs that exist in the current taxonomy are referenced. The review names canned products of pumpkin, salmon, rhubarb, and gooseberries that do not have current ingredient pages; these are not invented as new slugs here and will be picked up by the routing layer if and when those pages are created (see auto-stub policy in CLAUDE.md Part 10).
• Product slug selection: canned-foods-general for the umbrella scope, plus the specific product slugs that the review’s evidence supports (canned-fruit, canned-tomatoes, canned-vegetables, fruit-juices-non-apple, fruit-juices-apple-containing). canned-fish and canned-seafood are not added despite the canned-salmon episode (Kwantes 1966) because tin migration from modern fish cans is generally low and the review does not develop the fish-can occurrence picture beyond the one salmon outbreak; over-broad product fan-out would weaken row-fit downstream.
• The review’s Section 8 “200 mg/kg up to 200 mg/kg” framing — that there is “little evidence” for an association between canned-food tin at concentrations up to 200 mg/kg and acute GI effects — is preserved in the page as the review’s own claim, with the industry-sponsorship caveat. This is the literature-native view; HMTc threshold-setting work should evaluate this position alongside the primary clinical and outbreak evidence (some of which is below 250 mg/kg, e.g., the 156-247 mg/kg Kojima 1969 tomato juice episode and 154-289 mg/kg family outbreaks in Barker and Runte 1972) directly rather than relying on the review’s interpretation.
• 2026-06-04 fresh-context audit (general-purpose subagent) — verdict REVISE. Applied findings:
  • Benoy et al. 1971 dose-structure correction (Check 1 ❌): Audit subagent flagged that the Key Numbers Acute Clinical row had merged the no-nitrate arm (0, 498, 540, 1370 ppm tin) with the nitrate-supplemented arm (730 ppm tin + 50 mg/kg nitrate). Verified against PDF p. 1655 (“0, 498, 540 or 1370 ppm tin, or 730 ppm tin and 50 ppm nitrate … Nausea and/or diarrhoea were reported by all participants, following ingestion of the highest tin concentration. Lower concentrations were apparently without effect.”). Corrected: split into three rows showing no-nitrate-no-effect (0/498/540), nitrate-supplemented-no-effect (730 + 50 nitrate), and high-dose-effect (1370).
  • Barker and Runte 1972 banquet-count framing (Check 1 ❌): Audit flagged that “(43, 22 affected)” was misleading — PDF p. 1656 reports these as total cases at the two banquets, not affected counts; the ~65% affected applies separately. Corrected the row to read “43 and 22 cases respectively (~65% of those consuming juice at banquets affected)“.
  • Johnson and Greger 1982 absorption-range sign error (Check 1 ❌): Audit flagged that the wiki had written the absorption ranges with multiple negative signs (“range -7 to -9 to +9%” and “range -4 to -71%”). Verified against PDF p. 1658 (“the range was −7-9%” and ”(−4-71% of dose)”); the en-dash is a range separator and the second number is positive. Corrected to “(range -7% to +9%)” and “(range -4% to +71%)“.
  • Winship 1988 human clearance phases (Check 1 ❌): Audit flagged that the wiki wrote “First 20% in 4 days; second 20% in 25 days” but the PDF p. 1658 reports half-lives (“The first 20% is cleared with a half-life of 4 days and the second 20% with a half-life of 25 days. The final 60% has a much longer half-life of 400 days”). Corrected to “First 20% with half-life 4 days; second 20% with half-life 25 days; final 60% with half-life ~400 days” — the half-life-versus-clearance-time distinction matters because the half-life expresses an exponential-decay parameter, not a complete-elimination time.
  • JECFA “PTWI of 2 mg/kg bw/day” unit basis (Check 1 ⚠️): Audit flagged that calling 2 mg/kg bw/day a “PTWI” confuses units. The PTWI is the weekly value (14 mg/kg bw/week, per JECFA 1988a, 1988b); the 2 mg/kg bw/day is its daily equivalent (14 ÷ 7). PDF Section 8 (p. 1659) actually says “current JECFA figure of 2 mg/kg body weight/day” rather than “PTWI.” Corrected the Overview-section paragraph to refer to “the JECFA reference of 2 mg/kg bw/day (the daily equivalent of the Provisional Tolerable Weekly Intake of 14 mg/kg body weight per week)“.
  • Greger and Baier 1981 matrix labelling (Check 1 ⚠️): Audit flagged that the wiki listed matrices as “grapefruit, orange, tomato sauce, pineapple” while PDF p. 1654 has “grapefruit juice, orange juice, tomato sauce and pineapple in unlaquered or partially lacquered cans”. Corrected to “grapefruit juice, orange juice, tomato sauce, and pineapple” and added the “partially lacquered” qualifier.
• 2026-06-04 audit findings rejected as false positives:
  • Check 2 invented slugs (audit ❌; false positive): Audit claimed ingredients/pineapple and products/canned-foods-general are not in the current taxonomy and were invented. Verified against the live filesystem: wiki/ingredients/pineapple.md and wiki/products/canned-foods-general.md both exist as real wiki pages. The audit subagent appears to have read a stale or partial copy of docs/gpt-collaboration/taxonomy-snapshot.md that does not reflect the current ingredient and product directories. Slugs retained without change.
  • Check 1 European beverage cans figure (audit ⚠️; clarification rather than correction): Audit noted that the abstract says “15,000 million tinplate beverage cans” while the body (p. 1652) says “approximately 40,000 million [beverage] cans per annum, of which almost half are made of tinplate and all are internally lacquered.” The wiki’s row “European beverage cans per annum: >15,000 million tinplate” matches the abstract and is consistent with “almost half of ~40,000 million total” in the body. The two figures are not in conflict; they refer to total beverage cans (40,000 million) and tinplate beverage cans specifically (>15,000 million). No correction applied.
  • Check 1 cat-dose orange juice “498 ppm not emetic” omission (audit ⚠️; clarification only): Audit noted that the page omits the PDF detail that 498 ppm orange juice did not produce emesis in cats and that dogs (groups of four) did not vomit at 540 or 1370 ppm. Omission rather than error; the page’s purpose is to summarise tin-occurrence and acute-effect ranges, not reproduce every animal-dosing data point. Not applied; flagged here for future readers.
• Audit-application summary: 6 ❌/⚠️ findings applied (Benoy dose structure; Barker-Runte case framing; Johnson-Greger absorption sign; Winship half-life framing; JECFA PTWI unit basis; Greger-Baier juice labelling); 2 findings rejected as false positives (invented-slug claim contradicted by filesystem; European-beverage-can figure consistent with abstract). Subagent’s other ✅ confirmations on Checks 3 (speciation/methods), 4 (Part 12 brand firewall — sponsorship-context disclosure correctly handled), and 5 (Part 2 wiki/HMTc firewall — source-attribution framing correctly preserved) recorded.

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.