Bunu et al. 2023 — Heavy metal occurrence and health-risk assessment of soft and energy drinks sold in Lokoja, Kogi State, Nigeria

Bunu and colleagues (Niger Delta University, Wilberforce Island; Madonna University, Elele; Bayelsa Medical University, Yenagoa) quantified arsenic, lead, zinc, aluminium, and magnesium in twelve packaged-beverage brands purchased from supermarkets in Lokoja, Kogi State, Southwest Nigeria using flame atomic absorption spectrometry after concentrated-nitric/perchloric-acid wet digestion. The authors then computed Estimated Daily Intake (EDI), Chronic Daily Intake (CDI), and Target Hazard Quotient (THQ) per sample against US EPA Reference Oral Dose comparators, separately for adults (70 kg body weight, 24-year exposure duration, 0.030 kg/person/day intake) and children (20 kg, 6 years, 0.012 kg/person/day). The principal occurrence findings are that arsenic exceeded the cited 0.01 mg L⁻¹ maximum contaminant level (MCL) in one of twelve samples (B9 at 0.0603 mg L⁻¹); lead exceeded the cited 0.01 mg L⁻¹ MCL in one sample the source flags as exceeding (B4 at 0.0234 mg L⁻¹), with two further samples (B6 at 0.01032 mg L⁻¹ and B8 at 0.01072 mg L⁻¹) reported above the same MCL but not flagged as exceeding by the source’s asterisk convention; zinc was below the 5.00 mg L⁻¹ MCL across the panel; aluminium exceeded the cited 0.02 mg L⁻¹ MCL in eleven of twelve samples; and magnesium exceeded the cited 0.02 mg L⁻¹ MCL in twelve of twelve samples. Despite uniform aluminium and magnesium exceedances against the cited 0.02 mg L⁻¹ MCLs, the authors conclude that aluminium is the metal of principal concern (eleven of twelve samples Al THQ > 1 for both adults and children) and report all five-metal THQ values acceptable for the remaining four analytes.

Key numbers

All concentrations are reported in the source as mg L⁻¹ on an as-sold beverage basis. The source does not declare an instrument limit of detection, but the value “0.0005 mg L⁻¹” appears repeatedly for arsenic, lead, zinc, and aluminium across multiple samples and is treated as the de facto reporting floor for this study. The abstract and discussion round this floor to “0.001 mg L⁻¹” when describing minimum-of-range values. The instrument is identified as “Solar Thermo Elemental Atomic Absorption Spectrometer, Model SN - SG 718960” (Methods, p. 3). The authors performed accuracy verification using a 10 mg L⁻¹ solution of As, Pb, Zn, Al, and Mg from another batch as a standards check; recovery values are not reported.

Element-symbol convention used by the source. Figure 1 (the diversity-profile plot, p. 4) uses the symbol “Pd” (palladium) in its legend to denote the analyte the source’s introduction, abstract, and tables explicitly identify as lead. Table 1 itself uses “Lead (Pb)” in the column header. The wiki page records the analyte as lead (Pb) throughout per the source’s Table 1 header convention. This is the same paper-internal element-symbol typographical error pattern observed in the companion Bunu et al. 2023 Bayelsa-State study (bunu2023-heavy-metals-energy-drinks-bayelsa), where “Pd” was used throughout for lead.

Heavy metal concentrations in twelve soft- and energy-drink samples (Table 1, p. 4, mg L⁻¹)

The source uses an asterisk convention in Table 1: * = “Slightly higher than expected level,” ** = “Greatly higher than expected level,” *** = “Significantly higher than expected level.” These asterisks are recorded below as the source presents them.

Per-analyte detected ranges across the panel of twelve samples (Table 1 values as authoritative):

• As: 0.0005–0.0603 mg L⁻¹ (eleven samples ≤MCL 0.01 mg L⁻¹; one sample exceedance — B9 at 0.0603 mg L⁻¹, flagged with ***). The source’s abstract and discussion report the range as “0.001–0.0603 mg L⁻¹”; the minimum 0.001 figure is a rounding of the 0.0005 reporting floor.
• Pb: 0.0005–0.0234 mg L⁻¹. The source’s text flags one sample (B4 at 0.0234 mg L⁻¹, with *) as exceeding the cited 0.01 mg L⁻¹ MCL. Two further samples (B6 at 0.01032 mg L⁻¹ and B8 at 0.01072 mg L⁻¹) are tabulated above the same 0.01 mg L⁻¹ MCL but are not flagged by the source’s asterisk convention; the source’s narrative (“within the MCL… except B4”) is therefore inconsistent with the tabulated values for B6 and B8.
• Zn: 0.0005–2.9071 mg L⁻¹ (twelve of twelve below the 5.00 mg L⁻¹ MCL). The source’s abstract reports the range as “0.5658 to 2.9071 mg/L” but this excludes B10’s tabulated 0.0005 mg L⁻¹ value, which is the censored reporting floor; Table 1 is recorded here as the authoritative dataset.
• Al: 0.0005–2.0491 mg L⁻¹ (eleven samples above the cited 0.02 mg L⁻¹ MCL; only B8 at 0.0005 mg L⁻¹ below). The source flags every above-MCL sample with asterisks, including B12 at ***.
• Mg: 0.6954–3.8216 mg L⁻¹ (twelve of twelve above the cited 0.02 mg L⁻¹ MCL, with B9 and B11 flagged ***). The source notes that no contemporary WHO Mg drinking-water recommendation exists; the 0.02 mg L⁻¹ ceiling cited in Table 1 is identified by the source’s discussion (p. 7) as “a Nigerian industrial standard… a desirable limit for magnesium at 0.02 mg/L in soft or energy drinks, but there is no recent WHO recommendation.”

Estimated Daily Intake (Table 2, p. 5, × 10⁻⁵ mg L⁻¹ as reported by source, split adult and child)

Per Table 2 header units, EDI values are reported in units of × 10⁻⁵ mg L⁻¹ (i.e., to convert to mg L⁻¹, divide by 10⁵). Sample B1 As Adult is therefore 0.0182 × 10⁻⁵ = 1.82 × 10⁻⁷ mg L⁻¹.

EDI formula reported by source (Methods, p. 3): EDI = Cm × Cfactor × Dfood-intake / Bw, with Cm in mg/kg, Cfactor = 0.085 (unitless), Dfood-intake = 30 cL/person/day (0.30 L), Bw = 20 kg for children and 70 kg for adults.

Chronic Daily Intake (Table 3, p. 5, × 10⁻⁵ mg L⁻¹ as reported by source, split adult and child)

CDI formula reported by source (Methods, p. 3): CDI = C × EDI × EFi × IR / (Bw × AT), with EFi = 365 d/y, IR = 0.012 kg/person/day for children and 0.030 kg/person/day for adults, AT = 365 × exposure-years (6 y for children, 24 y for adults).

Target Hazard Quotient (Table 4, p. 6, dimensionless, split adult and child)

THQ formula reported by source (Methods, p. 3): THQ = CDI / RfD. Per-analyte THQ ranges across the panel (acknowledging the unit-anomaly caveats below):

• As: Adult 7.00E-05 to 8.67E-03; Children 1.00E-03 to 1.20E-01 (all THQ < 1).
• Pb: Adult 6.00E-05 to 2.86E-03; Children 8.57E-05 to 4.00E-03 (all THQ < 1).
• Zn: Adult 0 to 0.0042; Children 0.000001 to 0.0058 (all THQ < 1).
• Al: Adult 0.018 to 73.333; Children 0.0250 to 100.00. Eleven of twelve samples exceed THQ = 1 for both adults and children (only B8 below). Twelve of twelve are above THQ = 1 for children except B8 (THQ = 0.025).
• Mg: Adult 0.00041 to 0.00222; Children 0.00054 to 0.00312 (all THQ < 1).

Regulatory and comparator anchors cited by the authors

• WHO/US EPA MCL (Table 1, p. 4): As 0.01 mg L⁻¹, Pb 0.01 mg L⁻¹, Zn 5.00 mg L⁻¹, Al 0.02 mg L⁻¹, Mg 0.02 mg L⁻¹. The source describes these jointly as “Maximum Concentration Level (mg/L) of potentially toxic metals in drinking water (US EPA, 2011).”
• US EPA Reference Oral Doses (RfDs) (Table 1, p. 4): As 3.0 × 10⁻⁴, Pb 3.5 × 10⁻³, Zn 3.0 × 10⁻¹, Al 1.2 × 10⁻⁵, Mg 7.38 × 10⁻¹ (mg/kg/day basis).
• Nigerian industrial Mg standard (Discussion, p. 7): 0.02 mg L⁻¹ desirable limit for magnesium in soft or energy drinks; the source notes no recent WHO recommendation for Mg in beverages.
• Body-weight and exposure parameters (Methods, p. 3): adult 70 kg / 24 y / 0.030 kg/day; child 20 kg / 6 y / 0.012 kg/day; daily-beverage-intake 30 cL/person/day (citing NSW Centre for Public Health Nutrition 2009 per source ref).
• The source cross-references prior Nigerian-beverage heavy-metals surveys (Adepoju-Bello et al. 2012 — Lagos soft drinks; Izah et al. 2016 — beverage review; Salako et al. 2016 — Idiroko alcohol and non-alcoholic canned drinks; Godwill et al. 2015 — soft drinks Nigeria), plus a Sudanese (Woyessa et al. 2015 — Ethiopia soft drinks) and Iranian (Balali-Mood et al. 2018 — Mashhad fruit juices) comparator.

Methods (brief)

Twelve packaged-beverage samples (labelled B1–B12) were purchased from the New Market (International Market), Old Market, and Kpata Market in Lokoja, Kogi State, Southwest Nigeria — a city located approximately 165 km southwest of Abuja and 390 km northeast of Lagos, at 7.8023° N, 6.7333° E. The source names the twelve brands collectively in its Methods section as a single-paragraph inventory; per CLAUDE.md Part 12 the wiki page does not reproduce that inventory (see Verification notes — Brand firewall).

Glassware was decontaminated by 24-hour soak in 10% v/v nitric acid solution followed by purified-water rinse, repeated deionized-water rinse, room-temperature drying, and clean-environment storage. Sample digestion followed the method of Otim et al. 2019 with modifications: 25 mL of well-mixed sample plus 20 mL of distilled water were transferred into a 150 mL beaker; 4 mL of concentrated HNO₃ and 2 mL of HClO₄ were added; the sample was heated until the volume was reduced to 15 mL; the cooled solution was filtered into a beaker for AAS aspiration.

Metal concentrations were quantified using a “Solar Thermo Elemental Atomic Absorption Spectrometer, Model SN - SG 718960.” A 10 mg L⁻¹ standard solution of As, Pb, Zn, Al, and Mg from a separate batch was analysed to verify instrument accuracy; recovery values for the verification check are not reported. The source does not specify whether flame-AAS or graphite-furnace-AAS was used, does not specify per-analyte wavelengths, does not report instrument limits of detection or quantification, does not report certified-reference-material recoveries, does not report blank-correction details, and does not specify whether each sample was run as a single measurement or in replicate.

Health-risk computations followed the US EPA framework: EDI from Cm × Cfactor × Dfood-intake / Bw with Cfactor = 0.085, Dfood-intake = 0.30 L/person/day, and Bw = 20 kg (children) or 70 kg (adults); CDI from C × EDI × EFi × IR / (Bw × AT) with EFi = 365 d/y, IR = 0.012 kg/person/day (children) or 0.030 (adults), AT = 6 × 365 (children) or 24 × 365 (adults); THQ = CDI / RfD using the US EPA Reference Oral Doses tabulated in Table 1. Data were analysed using PAST 3.01 (Paleontological Statistics package), GraphPad 9.0, and Microsoft Excel 2013; statistical significance was defined as P < 0.05.

No metal speciation was performed: arsenic, lead, zinc, aluminium, and magnesium are reported as total elemental concentrations (the arsenic measurement is reported here as total arsenic, tAs, per CLAUDE.md Part 14 speciation convention — without an iAs/oAs distinction).

Evidence Fitness

This source contributes direct primary occurrence values for finished packaged beverages (mixed soft-drink and energy-drink product class) sold in three markets in Lokoja, Kogi State, Nigeria. The principal limitations bearing on pooling eligibility and synthesis weight are:

(i) Small sample size and no declared replicate structure. Twelve samples are reported, but the source does not state whether each tabulated value is a single measurement, a duplicate mean, or a triplicate mean. This is adequate for an exploratory market-screening report but limits inference to the broader Southwest-Nigeria packaged-beverage population.

(ii) No declared analytical limit of detection or quantification. The reporting floor of 0.0005 mg L⁻¹ is inferred from its repeated appearance across As, Pb, Zn, and Al columns for multiple samples; the source does not explicitly identify this value as an LOD. The lack of declared LOD/LOQ and the lack of certified-reference-material recovery data prevent independent verification of analytical quality control. The 10 mg L⁻¹ standards-check value is reported as performed but no recovery percentages are provided.

(iii) No certified reference material or recovery data reported. A 10 mg L⁻¹ standards solution check is described but no spike-recovery, CRM (NRCC, NIST, or equivalent), or method-blank data are presented.

(iv) No metal speciation. Inorganic vs total Cr, MeHg vs total Hg, iAs vs tAs distinctions are not addressed; the arsenic value is recorded as total arsenic (tAs). The chromium, mercury, cadmium, tin, nickel, antimony, and uranium HMI analyte panels were not measured. Among the ten HMTc/HMI analytes, this source covers Pb, tAs, and Al (three of ten — and Al is the analyte with the principal source-identified exceedance pattern).

(v) EDI/CDI/THQ computations report values whose arithmetic relationship to the stated formulas is non-trivially obscured by a unit-scaling factor not stated in the Table 2 footnote. Applying the EDI formula (EDI = Cm × Cfactor × Dfood / Bw with Cfactor = 0.085, Dfood = 0.30, Bw = 70) to B1’s As concentration of 0.0005 mg L⁻¹ yields 1.82 × 10⁻⁷ mg/kg/day. Table 2 reports the B1 As Adult EDI as 0.0182 in units of ”× 10⁻⁵ mg/L” per the Table 2 header — i.e., 0.0182 × 10⁻⁵ = 1.82 × 10⁻⁷ mg/L, which (modulo a units-mismatch between “mg/kg/day” and “mg/L”) matches the formula’s expected output for B1 As Adult. The ”× 10⁻⁵ mg/L” header units are non-standard for an EDI calculation (EDI is conventionally a mg/kg/day quantity, not a mg/L quantity), and the same header unit is applied to the CDI table 3 even though CDI is conventionally also in mg/kg/day. The unit convention should be re-derived before pooling. The Table 1 raw concentration values are the authoritative dataset for pooling-eligibility decisions; Table 2 and Table 3 downstream computations are recorded as the source presents them but require careful unit-basis verification before being pool-eligible. Table 4 THQ values are dimensionless and self-consistent within the source’s framing.

(vi) Paper-internal numerical inconsistency between the abstract/discussion and Table 1 for Pb MCL flagging. The Discussion (p. 6) states “The lead contents of the soft drinks were within the MCL except in B4 which had 0.023 mg/L.” Table 1 reports B6 at 0.01032 mg L⁻¹ and B8 at 0.01072 mg L⁻¹, both above the cited 0.01 mg L⁻¹ MCL, but the source’s asterisk-flagging convention does not flag B6 or B8 as exceeding. The Table 1 figures are recorded as the authoritative dataset on the present page.

(vii) Paper-internal numerical inconsistency between the abstract and Table 1 for Zn range. The Abstract (p. 1) states the zinc range as “0.5658 to 2.9071 mg/L”; Table 1 reports B10 at 0.0005 mg L⁻¹ (the censored reporting floor), which the abstract’s range excludes. The Table 1 figures are recorded as the authoritative dataset.

(viii) Paper-internal element-symbol typographical error in Figure 1. Figure 1 (diversity-profile plot, p. 4) uses the symbol “Pd” in its legend for the analyte that Table 1 explicitly labels “Lead (Pb).” This is the same paper-internal element-symbol error pattern observed in the companion Bunu et al. 2023 Bayelsa-State study; in the present Kogi-State source, the error is restricted to the Figure 1 legend, with Table 1 using the correct “Pb” symbol throughout.

(ix) Paper-internal product-class label inconsistency between title and body. The title describes the subject as “Energy Drinks,” but the abstract, methods, results, and conclusion describe the samples consistently as “soft drinks.” The twelve brands named in the methods inventory (collectively, without per-sample mapping) span carbonated soft drinks, energy/sports drinks, fruit drinks, a milk-based drink, and a malt drink, with the carbonated-soft-drink class being the largest contributor to the panel. The present wiki page routes this source to both sports-energy-drinks and soft-drinks-carbonated-beverages to reflect the actual product class spanned by the panel.

(x) Conclusion characterizes the result as “all the heavy metals, except aluminum, were within acceptable THQ limits” while Table 1 reports twelve of twelve magnesium samples above the 0.02 mg L⁻¹ MCL (with the source’s own asterisk convention flagging every Mg sample as “greatly higher than expected level” or “significantly higher than expected level”). The Mg THQ values in Table 4 are below 1 because the US EPA Mg RfD (7.38 × 10⁻¹ mg/kg/day) is a large divisor; the THQ-vs-MCL distinction matters here and the conclusion’s framing collapses the two registers. The wiki page records the THQ values and the MCL exceedance pattern as reported, without reproducing the conclusion’s framing.

Evidence tier set to C. The source is primary research, peer-reviewed (Pharmacology and Toxicology of Natural Medicines, Vol 3, Issue 1, 2023; ISSN 2756-6838), but with the methodological caveats above (no declared replicate structure, no declared LOD/LOQ, no CRM recovery data, internal-arithmetic unit-basis ambiguities, internal-numerics inconsistencies between body text and tables, a paper-internal element-symbol typographical error confined to Figure 1, and the title/abstract product-class mismatch). Tier-B would require declared replicate structure, declared LOD/LOQ, ICP-MS or graphite-furnace AAS for trace analytes, CRM-anchored QC, and reconciled internal numerics. The Table 1 raw concentrations are pool-considerable for Pb, tAs, Al, and Zn at C-tier weight with explicit notation of the (i)–(x) caveats; the downstream EDI/CDI tables are not pool-eligible without re-derivation under a verified unit convention.

Implications

• Certification: contributes direct primary occurrence values for both the sports-energy-drinks HMTc category (Category 5 row 9) and the soft-drinks-carbonated-beverages HMTc category (Category 5) for the Nigerian retail market. The Pb column reports one source-flagged exceedance (B4 at 0.0234 mg L⁻¹) plus two further tabulated exceedances the source’s asterisk convention overlooks (B6 at 0.01032 and B8 at 0.01072 mg L⁻¹) against the cited 0.01 mg L⁻¹ MCL. The tAs column reports one exceedance (B9 at 0.0603 mg L⁻¹). The Al column reports eleven of twelve samples above the cited 0.02 mg L⁻¹ MCL, with the source’s own narrative identifying aluminium as the principal heavy-metal concern. These values complement the Polish-market ICP-MS data (Czarnek et al. 2024, czarnek2024-heavy-metals-energy-drinks), the Jordanian-market flame-AAS data (Al-Sayyed et al. 2024, alsayyed2024-heavy-metals-energy-drinks-jordan), the Nigerian Bauchi-State data (Babayo et al. 2026, babayo2026-heavy-metals-energy-drinks-nigeria), and the companion Nigerian Bayelsa-State data (Bunu et al. 2023, bunu2023-heavy-metals-energy-drinks-bayelsa, same first-author). The two Bunu 2023 datasets together (Bayelsa State and Kogi State, both Nigeria) will be considered by the synthesis pass for the Nigerian-packaged-beverage occurrence picture as a whole.
• Courses: useful as a teaching reference for (1) the limitations of single-laboratory AAS screening without declared LOD/LOQ or CRM recovery for trace metals in beverages; (2) the importance of distinguishing MCL exceedance from THQ exceedance — these are different epistemic registers and a uniform THQ < 1 result does not negate a uniform regulatory-MCL exceedance; (3) the importance of explicit replicate structure and below-detection censoring conventions for pool-eligibility decisions; (4) the importance of internal-consistency cross-checks (asterisk flagging vs tabulated values, abstract ranges vs table values, figure legends vs table headers) during peer review; (5) the importance of accurate product-class labelling (title vs body) in source pages and downstream synthesis.
• App: contributes Nigerian-market occurrence values for the mixed soft-drink and energy-drink product class. Per-sample brand identities are not disclosed by the source (samples are labelled B1–B12 with brand identities listed only as a collective inventory in the methods); no brand-firewall handling per CLAUDE.md Part 12 was required on the values side beyond not reproducing the inventory list.
• Discovery: the source’s reference list overlaps substantially with the Bayelsa companion paper; one new discovery lead is Otim et al. 2019 (PLoS ONE, “Assessing the health risks of consuming ‘sachet’ alcohol in Acoli, Uganda” — DOI 10.1371/journal.pone.0212938) which is the source of the digestion procedure used in the present study and a potential companion source for Ugandan-sachet-alcohol heavy-metal occurrence.

Provenance notes

Open-access article distributed under CC BY 4.0 (license declaration on p. 1 of the PDF: “This article by Pharmacology and Toxicology of Natural Medicines is licensed and published under the Creative Commons Attribution License 4.0 which permits unrestricted use, distribution, and reproduction in any medium, provided this article is duly cited”). Received 6 June 2022; Accepted 22 March 2023; published in Pharmacology and Toxicology of Natural Medicines, Volume 3, Issue 1, 2023, pp. 1–10. DOI: 10.52406/ptnm.v1i1.55. Online ISSN: 2756-6838. Published by Phytomedicine Research Group, Department of Pharmacology & Toxicology, Faculty of Pharmaceutical Sciences, University of Benin, Benin City 300001, Nigeria. Corresponding author: Samuel J. Bunu (pharmsamuelbunu@gmail.com; +234-806-9703966). Affiliations: 1Department of Pharmaceutical and Medicinal Chemistry, Faculty of Pharmacy, Niger Delta University, Wilberforce Island, Bayelsa State, Nigeria (Bunu, Kashimawo, Vaikosen); 2Department of Pharmaceutical Chemistry, Faculty of Pharmacy, Madonna University Elele, Rivers State, Nigeria (Ebeshi, Itodo); 3Department of Pharmacology & Toxicology, Faculty of Pharmaceutical Sciences, Bayelsa Medical University, Bayelsa State, Nigeria (Kpun). Conflict of interest: none declared. Funding: self-funded by the authors. Accessed via the Manual Fetch Discovery autopilot.

Wiki pages this source may touch

• lead
• arsenic
• aluminum
• zinc
• sports-energy-drinks
• soft-drinks-carbonated-beverages

Verification notes

Element-symbol convention. Figure 1 (the diversity-profile plot on p. 4) uses the symbol “Pd” (palladium) in its legend for the analyte the source’s Table 1 explicitly labels as “Lead (Pb)” and the source’s introduction and discussion identify as lead. This is a typographical error in Figure 1 only; Table 1, the abstract, the introduction, the methods, the discussion, and the conclusion all use “lead” or “Pb” correctly. The wiki page records the analyte as lead (Pb) throughout and uses Pb in the frontmatter metals: field per CLAUDE.md Part 14. This is the same pattern as the companion Bunu et al. 2023 Bayelsa-State source (bunu2023-heavy-metals-energy-drinks-bayelsa) but with much narrower scope (Figure 1 legend only, not paper-wide).

Speciation handling per CLAUDE.md Part 14. This source measures total elemental As, Pb, Zn, Al, and Mg by AAS without speciation. Arsenic is recorded as tAs (total arsenic) in the metals: frontmatter field, per CLAUDE.md Part 14’s “iAs/tAs and tHg/MeHg as non-negotiable distinctions” rule (no iAs vs oAs distinction is offered by the source, so tAs is the safe attribution). Pb, Al, and Zn are recorded as such. Mercury, cadmium, chromium, tin, antimony, nickel, and uranium were not in the source’s analyte panel and are not recorded. Magnesium is not in the HMI/HMTc analyte vocabulary and is therefore not recorded in the metals: frontmatter, although its values are preserved in the Key numbers section because the source’s own narrative makes magnesium the comparator for the aluminium THQ finding.

Products frontmatter. The products: frontmatter lists both sports-energy-drinks and soft-drinks-carbonated-beverages because the source’s twelve-sample panel mixes brands from the carbonated-soft-drink, energy/sports-drink, fruit-drink, milk-based-drink, and malt-drink commercial classes — with the carbonated-soft-drink class as the largest single contributor and the energy/sports-drink class as the second-largest. The source’s body text refers to the samples consistently as “soft drinks,” and the title’s “Energy Drinks” framing is a paper-internal mismatch with the actual panel composition. Routing to both product slugs is the appropriate level of specificity for a mixed-class market screen; the routing layer (CLAUDE.md Part 5b) will fan out to whichever product pages downstream synthesis weighs against. The fruit-drink, milk-based-drink, and malt-drink samples are not separately routed because their individual fractions of the panel are small enough that this paper is not a primary occurrence source for those product classes.

Ingredients frontmatter. The ingredients: field is empty. The twelve packaged beverages measured are finished consumer products, not single-ingredient measurements; the source does not measure heavy-metal content at the ingredient level.

Matrices frontmatter. The matrices: field uses both energy-drinks (established vocabulary, used in czarnek2024-heavy-metals-energy-drinks, alsayyed2024-heavy-metals-energy-drinks-jordan, babayo2026-heavy-metals-energy-drinks-nigeria, bunu2023-heavy-metals-energy-drinks-bayelsa) and carbonated-soft-drinks (established vocabulary, used in ahmed-soft-drinks Sudan and the koga2021-tin-arsenic-lead-soft-drinks comparator-only review). This reflects the mixed panel composition.

Jurisdictions frontmatter. NG (Nigeria). The sample population is restricted to three markets in Lokoja, Kogi State, Southwest Nigeria. The Bayelsa-State companion paper uses the same NG jurisdiction code; the synthesis pass will consider the two together for the Nigerian retail-market occurrence picture.

Brand firewall per CLAUDE.md Part 12 (strict reading, locked 2026-05-17). The source names twelve commercial brands in its Methods section as a collective inventory of the brands sampled (the twelve samples are labelled B1–B12 in all tables; no per-sample-to-brand mapping is disclosed, so no per-brand contamination value can be reported and no brand-by-brand ranking is implicit in the data). The strict Part 12 reading prohibits reproducing the inventory list of brands in wiki page content, since doing so converts source-disclosed brand identities into a wiki-side disclosure that the named brands were sampled and quantified for heavy-metal content. The wiki page therefore describes the sample population as “twelve commercially marketed soft-drink and energy-drink brands purchased from supermarkets in Lokoja, Kogi State, Southwest Nigeria” without naming the twelve brands. The methods section’s scientific-method vendor identities (Solar Thermo Elemental Atomic Absorption Spectrometer, Model SN - SG 718960; PAST 3.01; GraphPad 9.0; Microsoft Excel 2013) are retained under the Part 12 scientific-method-vendor exception locked 2026-05-17. Brand-by-brand handling, where it would be desired by the wiki/HMTc certification process, belongs in the private brand-intelligence build described in CLAUDE.md Part 26.

Wiki/HMTc firewall per CLAUDE.md Part 2. No HMTc threshold proposals, no consumer-audience risk advisories, and no synthesis claims of the form “this confirms the literature consensus that…” appear in this wiki page body. The observation that eleven of twelve samples exceed the cited 0.02 mg L⁻¹ Al MCL, that twelve of twelve samples exceed the cited 0.02 mg L⁻¹ Mg MCL, and that three Pb samples exceed the cited 0.01 mg L⁻¹ Pb MCL is reported as the source itself reports it — as a comparison against the regulatory comparator values the source itself cited — and is not framed as an HMTc threshold recommendation or as a consumer-safety claim. The Evidence Fitness section’s flagging of the source’s own conclusion language (“all the heavy metals, except aluminum, were within acceptable THQ limits”) is a methodological observation about the source’s internal logic (the MCL-vs-THQ register conflation), not a wiki-side risk advisory.

Paper-internal inconsistencies recorded faithfully and flagged.

1. Figure 1’s legend uses the symbol “Pd” for the analyte that Table 1 labels “Lead (Pb).” Recorded as Pb throughout this wiki page; Figure 1’s “Pd” usage is preserved in direct mention only.
2. The Discussion (p. 6) reports Pb “within the MCL except in B4” but Table 1 reports B6 at 0.01032 mg L⁻¹ and B8 at 0.01072 mg L⁻¹, both above the 0.01 mg L⁻¹ MCL. Table 1 is recorded as the authoritative dataset.
3. The Abstract (p. 1) reports the Zn range as “0.5658 to 2.9071 mg/L” but Table 1 reports B10 at 0.0005 mg L⁻¹ (the censored reporting floor). Table 1 is recorded as the authoritative dataset.
4. The Table 2 EDI and Table 3 CDI columns use the header units ”× 10⁻⁵ mg/L” rather than the conventional mg/kg/day used for EDI/CDI calculations. The arithmetic relationship to the formulas in the Methods section is consistent if ”× 10⁻⁵ mg/L” is interpreted as ”× 10⁻⁵ mg/kg/day” (i.e., a units-mismatch between the header and the underlying calculation). The Table 1 raw concentrations are the authoritative dataset; Tables 2 and 3 are recorded faithfully but flagged as requiring unit-basis re-derivation before pool-eligibility.
5. The title is “Energy Drinks” but the abstract, methods, results, and conclusion treat the samples as “soft drinks.” Of the twelve named brands, only three are commercially classified as energy/sports drinks; the rest are sodas, fruit drinks, a milk drink, and a malt drink. The routing reflects the actual panel composition (both sports-energy-drinks and soft-drinks-carbonated-beverages).
6. The Conclusion characterizes the result as “all the heavy metals, except aluminum, were within acceptable THQ limits” without addressing the twelve-of-twelve Mg MCL exceedance, the source-flagged eleven-of-twelve Al MCL exceedance, and the under-flagged Pb MCL exceedances at B6 and B8. Recorded faithfully; not framed as a wiki-side risk advisory.

Audit application (2026-06-06). Fresh-context Agent subagent audit (verdict PROMOTE; all five checks ✅ clean with one ⚠️ minor on Check 4 Part 12 brand-firewall scope into rationale paragraphs). The subagent flagged that Evidence Fitness (ix) and the Products frontmatter rationale paragraph as originally written named the twelve brand identifiers while explaining the panel composition — a borderline call under the strict 2026-05-17 Part 12 reading because the original page did not reveal a per-brand contamination value or ranking, but did reveal source-disclosed brand identities in a wiki-side disclosure. Independent re-verification against the companion Bunu Bayelsa-State page (bunu2023-heavy-metals-energy-drinks-bayelsa) confirmed that the prior convention does not name the eleven Bayelsa brands in any rationale paragraph; the strict-reading convention is therefore established for the Bunu-companion corpus. Corrected by redacting brand-name enumerations from Evidence Fitness (ix) and the Products frontmatter Verification note while preserving the class-composition summary. No other audit findings warranted page changes.

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.