Kangre 2024 — Lead and tin in canned tomato mix, Ghana (storage display conditions)

An MPhil thesis from the University of Education, Winneba (Ghana) examined Pb and Sn in 32 cans of canned tomato mix (a single canned tomato paste brand) under two retail display conditions: sunlight-exposed outdoor stalls vs enclosed shops. Samples were collected from the Effutu Municipality and analysed by ICP-MS at the Ghana Standards Authority. Both metals were below Codex limits on a mean basis. Pb mean was numerically higher in sunlight-exposed product (0.052 ± 0.12 mg/kg, SE) than enclosed-shop product (0.038 ± 0.02 mg/kg, SE), and Sn mean was numerically lower in sunlight-exposed product (2.304 ± 0.83 mg/kg, SE) than enclosed-shop product (2.656 ± 0.79 mg/kg, SE). An independent-samples t-test found no statistically significant difference between display models for either metal (p > 0.05); the null hypothesis of equal means was retained. A dietary survey of 400 university students supplied consumption parameters, and Monte Carlo simulation (Palisade @Risk, 100,000 iterations) produced incremental lifetime cancer risk (ILCR) for Pb and margin-of-exposure (MoE) for chronic kidney disease, with the 95th-percentile ILCR for Pb exceeding the US EPA de minimis 1×10⁻⁶ in all four gender × display strata, and 5th-percentile Pb and Sn MoE values below 1 in every stratum.

Key numbers

Pb and Sn concentrations in canned tomato mix (mg/kg, mean ± SE; n = 16 per display model):

Per-sample distribution (Table 3, mg/kg):

The Pb maximum of 0.51 mg/kg in a sunlight-exposed sample exceeds the 0.1 mg/kg Codex maximum level for Pb in processed tomato products by a factor of five, even though the group mean of 0.052 mg/kg sits below it. This single high value drives the large standard error (0.12) in the sunlight stratum.

Analytical performance:

• LOD: Pb = 0.0002 mg/kg; Sn = 0.0005 mg/kg.
• LOQ: Pb = 0.01 mg/kg; Sn = 0.10 mg/kg.
• CRM recovery (FAPAS 07390 tomato paste): Pb 107.6 % (assigned 0.447 ± 0.161 mg/kg; found 0.481 mg/kg); Sn 95.6 % (assigned 0.181 ± 0.26 mg/kg; found 0.173 mg/kg).
• Calibration r² > 0.99; duplicate analysis per sample.

Chronic daily intake (CDI, mg/kg-d):

Hazard quotient (HQ), reference doses Pb 3.5 × 10⁻³ mg/kg-d, Sn 0.6 mg/kg-d:

All HQ values are below 1 (the threshold for systemic non-cancer risk concern).

Incremental lifetime cancer risk for Pb (PF = 0.0085 (mg/kg-d)⁻¹), Table 7:

The maximum ILCR exceeded the US EPA de minimis 1 × 10⁻⁶ in all four strata, and the 95th-percentile ILCR exceeded 1 × 10⁻⁶ in all four strata (the thesis flags these with the “significant cancer risk” footnote in Table 7).

Margin of exposure to chronic kidney disease (BMDL₁₀ Pb = 0.00063 mg/kg-d; BMDL₁₀ Sn = 0.46 mg/kg-d), Table 7:

Mean MoE values are above the protective threshold of 1, but every 5th-percentile MoE value falls below 1 (the thesis flags these as “significant public health concern”). The thesis interprets this as substantial sub-population risk of nephrotoxicity from Pb and Sn through canned tomato mix consumption, even though group means imply protection.

Consumption survey (n = 400 students, Winneba campus, multi-stage cluster sample):

• Canned tomato mix is consumed primarily in jollof rice (highest per-sitting rate), stew, and soup.
• Per-sitting consumption (both models combined): jollof 114.15 g, stew 80.59 g, soup 77.72 g.
• Total daily consumption across all respondents and all dishes: 1450 g/day; total annual 2332.37 kg/year.
• Females consumed more canned tomato mix per day than males in both display strata.
• 41.5 % of respondents (n = 166) bought from enclosed shops; 6.25 % (n = 25) from sunlight-exposed shops; 52.25 % (n = 209) from both. Exposed-shop-only consumers had higher per-sitting volume but lower headcount.
• Body weight (mean): males 59–61 kg; females 60–64 kg, consistent with prior Ghanaian student surveys.

Methods (brief)

Cross-sectional analytical study combining chemical analysis of canned product with a dietary-survey exposure model. Systematic sampling drew 32 cans of one canned tomato mix brand from 3000 estimated Winneba retail shops (40 % sunlight-exposed, 60 % enclosed); sampling intervals were k = 75 (exposed) and k = 112 (enclosed) for 16 cans each. Samples were coded and sent to the Ghana Standards Authority. Microwave digestion (Milestone MA079) used 5 mL HNO₃ (Suprapur, 65 %) + 3 mL H₂O₂ (30 %) on ~1.0 g sample at 170 °C, 1000 W, 50 bar, 50 min. Pb and Sn quantification used a Perkin Elmer NexION 2000P ICP-MS following BS EN 15763:2009. Quality control comprised method blanks, a FAPAS 07390 tomato paste CRM, and duplicate analysis with r² > 0.99 calibration. Statistical analysis used SPSS v20: one-sample one-tailed t-tests against Codex MLs and an independent-samples t-test between display models, p < 0.05.

The exposure model followed Anyimah-Ackah et al. (2022). CDI was computed as CDI = (C × MF × EF × ED) / (BW × AT × 1000), with averaging time 70 years for Pb (carcinogenic) and 30 years for Sn (non-cancer). HQ used CDI / RfD; ILCR used PF × CDI with PF(Pb) = 0.0085 (mg/kg-d)⁻¹; MoE used BMDL₁₀ / CDI with BMDL₁₀(Pb) = 0.00063 mg/kg-d and BMDL₁₀(Sn) = 0.46 mg/kg-d, both for chronic kidney disease (Alexander et al. 2010; Jayasumana et al. 2015). Monte Carlo simulation in Palisade @Risk v8.1 ran 100,000 iterations across fitted distributions for concentration, mass consumed, exposure frequency, exposure duration, body weight, and averaging time.

The student survey used Yamane’s formula on the UEW undergraduate population of 35,947 (margin of error 5 %), yielding n ≈ 395 raised to 400 for non-response. Multi-stage cluster sampling: 4 of 9 faculties drawn by lottery, 1 department per chosen faculty, then 25 students per academic level (100–400) by fishbowl draw. The questionnaire collected demographics plus dietary recall of canned tomato mix in jollof rice, stew, and soup, pilot-tested at University of Ghana Legon (n = 30; Cronbach’s α = 0.78). Ethics approval: Ghana Health Service Ethics Review Committee, GHS-ERC:043/06/23.

Limitations: single brand, single municipality, student population not representative of the general Ghanaian adult population, only Pb and Sn measured, no analyte speciation, and a sampling design that pulls 16 cans from each display model without explicitly controlling for time-on-shelf or temperature/UV dose. The single high-Pb sample (0.51 mg/kg in sunlight-exposed) drives the large SE in that stratum and is the likely numeric origin of the higher group mean rather than a systematic sunlight-exposure effect; the t-test failing to reject the null is consistent with that reading.

Implications

Certification. Group-mean Pb and Sn in canned tomato mix from Winneba retail are both well below Codex MLs (0.1 mg/kg for Pb in processed tomato; 250 mg/kg for Sn). However, the Pb maximum of 0.51 mg/kg in a sunlight-exposed sample exceeds the Codex ML by 5×, and the Monte Carlo–derived ILCR and MoE values flag tail-risk sub-populations in every display × gender stratum, even where group means imply protection. Certification programs reading this study should not over-index on group means: a percentile-based certification (the HMTc P97 framework for clean categories per standards-methodology) would catch the 0.51 mg/kg outlier in the sunlight stratum, whereas a mean-based certification would miss it entirely. The display-condition contrast did not produce a statistically significant mean shift in this single-brand study, so claims that sunlight exposure measurably raises Pb in canned tomato paste require larger or multi-brand replication.

Courses. Useful teaching case for probabilistic risk assessment in a developing-country retail context: the gap between group-mean concentrations (which all sit below regulatory limits) and 5th-percentile MoE / 95th-percentile ILCR values (which all flag concern) is a clean illustration of why dietary risk cannot be reduced to “is the mean below the limit”. The single-high-Pb sample driving group SE is a teaching case for why robust statistics matter when a category contains rare extreme values.

App. Canned tomato mix (Ghana, single brand, n = 32): Pb 0.038–0.052 mg/kg mean (range 0.00–0.51 mg/kg, P95 0.03–0.05); Sn 2.3–2.7 mg/kg mean (range 0.65–4.01 mg/kg). B-tier (MPhil thesis, single brand, single site, West Africa); not generalisable beyond the studied brand and municipality.

Wiki pages updated on ingest

• tomato
• tomato-paste
• canned-tomatoes
• lead
• tin

Verification notes

2026-05-28 enhance pass (Claude): re-read source thesis (pp. 88–130 verified in full; abstract p. xiii). Schema reconciled to the post-2026-05-14 source template and several substantive defects corrected against the source. Specifically:

• Frontmatter products corrected from [products/canned-tomato, products/tomato-sauce] (the first slug does not exist in wiki/products/; the second slug names a thinner product than the “tomato mix” the source describes) to [products/tomato-paste, products/canned-tomatoes]. “Tomato mix” in the Ghanaian / West African retail context is a concentrated canned tomato paste; routing to the paste page is the literal match, and the canned-tomatoes page captures the can-lining migration angle that the source emphasises.
• Body language “Tasty Tom” / “Tasty Tom brand” dropped from prose and frontmatter sample_population per CLAUDE.md Part 12 brand firewall. The fact that the study examines a single brand is preserved (as a methods detail and a limitation), but the brand name is not propagated through the wiki.
• ”±” labeling corrected from SD to SE throughout. Table 4 (p. 103) and Table 3 (p. 118) of the thesis both label the dispersion as standard error, not standard deviation.
• Pb display-model claim corrected. The prior wiki said “Pb was significantly higher in sunlight-exposed products”; the source (p. 107) explicitly reports the independent-samples t-test as p > 0.05 and states “the null hypothesis was maintained”. The means are numerically higher in sunlight-exposed product but not statistically significantly so; the wiki now reflects that.
• Sn values added. The prior wiki recorded Sn as “consistent” with no numeric values; Table 4 reports enclosed-shop mean 2.656 ± 0.79 mg/kg (SE) and sunlight-exposed mean 2.304 ± 0.83 mg/kg (SE), and Table 3 reports the underlying per-sample range and percentiles. Sn was numerically lower (not equal) in the sunlight-exposed stratum, and the independent t-test for Sn was also non-significant (p > 0.05).
• Pb maximum value (0.51 mg/kg in a sunlight-exposed sample, 5× the Codex ML) added. The prior wiki described “Pb concentrations well below most international regulatory limits” without noting that the maximum single value exceeded the Codex ML; the per-sample distribution (Table 3) makes this clear, and it is the most certification-relevant detail in the chemistry.
• HQ for Sn added (the prior wiki only carried HQ for Pb). All four HQ-Sn values are <0.005, well below 1.
• ILCR by gender × display added in full (the prior wiki only quoted a coarse range “3.28×10⁻⁶ to 6.29×10⁻⁶”; Table 7 has eight discrete values worth showing because every stratum exceeds the 1×10⁻⁶ de minimis at maximum and at P95).
• MoE results added (entirely absent from the prior wiki). Every 5th-percentile MoE for both Pb and Sn falls below the protective threshold of 1, which the thesis flags with a footnote as a significant public-health concern.
• “Certification” implications block reworked to literature-native phrasing per Part 2 firewall: the prior wiki claimed retail display “measurably affects” Pb migration; the source’s own statistical test does not support a measurable shift between display models, so the certification implication is reframed around the percentile-vs-mean distinction (which the data actually support) instead of the display-model effect (which they do not).
• raw_path, cite_key, raw_handle, evidence_tier, license (“unknown” — the IR.uew.edu.gh institutional repository serves the thesis without an explicit reuse licence), near_duplicates, and sample_n preserved from the prior revision.

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.