Paudel et al. 2024 — Heavy metals (Fe, Cu, Zn, Pb) in 16 packaged fruit juices on the Kathmandu, Nepal retail market by FAAS

This open-access International Journal of Applied Sciences and Biotechnology study quantifies iron, copper, zinc, and lead in sixteen commercially packaged fruit juices — orange, apple, mango, and lychee flavours, four brands per flavour — purchased on the Kathmandu, Nepal retail market. The juices were wet-acid digested (aqua regia + concentrated nitric acid, hot-plate digestion to clarity) and measured by flame atomic absorption spectrophotometry on a Varian-style AA240FS instrument with an air–acetylene flame; calibration was constructed from stock standards and samples were prepared in analytical triplicate. Lead was below the analytical detection limit in every one of the sixteen samples. Iron, copper, and zinc were detected variably across flavours: Fe spanned 0.277–3.649 mg/L across all juices, with one apple-juice sample at 3.649 mg/L exceeding the 2 mg/L WHO/JECFA Provisional Maximum Tolerable Daily Intake (PMTDI)-derived drinking-water reference; Cu ranged from below detection (BDL) to 0.574 mg/L; Zn ranged from BDL to 0.2026 mg/L. The authors compare findings to WHO/FAO drinking-water guidelines (Fe 2 mg/L; Cu 2 mg/L; Zn 3 mg/L; Pb 0.01 mg/L) on the rationale that non-alcoholic beverages depend on water quality. No certified reference material recovery, LOD/LOQ tabulation, or procedural-blank values are reported, which constrains evidence tier to B despite peer-reviewed open-access publication.

Key numbers

Sixteen packaged fruit-juice samples (4 brands × 4 flavours), analysed in analytical triplicate by FAAS. Concentrations are reported in the source as mg/L of as-sold liquid juice; this page preserves the mg/L unit. WHO/FAO drinking-water reference levels are reproduced for comparison; the Codex maximum levels for Pb in fruit juice (0.03 mg/kg general, 0.04 mg/kg grape, 0.05 mg/kg berry/small-fruit) are noted but the source compared to the stricter drinking-water 0.01 mg/L on the rationale that beverages are water-dependent.

Iron (Table 2 of source, p. 161; mg/L)

¹ Flavour-level means are computed on this page from the four per-flavour values reported in source Table 2; the source itself does not report flavour means.

One of the four apple-juice samples carried Fe at 3.649 mg/L, exceeding the 2 mg/L WHO/JECFA PMTDI-derived water-derived reference; the other three apple-juice samples were ≤0.523 mg/L. The next-highest single Fe value across all sixteen samples was a lychee sample at 1.770 mg/L; the remaining fourteen samples were ≤0.762 mg/L.

Copper (Figure 5 of source, p. 161; mg/L)

All sixteen samples are well within the 2 mg/L WHO drinking-water guideline; the overall maximum (0.574 mg/L, one orange-juice sample) is ~29% of that guideline.

Zinc (Table 3 of source, p. 162; mg/L)

The overall Zn maximum (0.2026 mg/L, one orange-juice sample) is ~7% of the 3 mg/L WHO drinking-water guideline; all sixteen samples are well within that guideline.

Lead (Figure 6 of source, p. 162; mg/L)

Lead was below the detection limit in every one of the sixteen samples. The source reports R² = 0.978 for the Pb calibration curve (equation y = 0.039x + 0.030); the x-axis of Fig. 4 (p. 160) extends to ~10 mg/L, but the source does not explicitly state a Pb working range or a numerical LOD, so “below detection limit” cannot be converted to a numeric upper bound from the paper alone.

Calibration performance (Table 1 of source, p. 161)

Methods (brief)

Sample collection: sixteen commercially packaged fruit juices (4 brands × 4 flavours: orange, apple, mango, lychee) purchased from retail markets in Kathmandu. The brand identities are not disclosed in the source; the paper anonymises them as A, B, C, D. Sample preparation: 50 mL of each juice was combined with 5 mL aqua regia in a beaker, mixed, and heated on a hot plate; 5 mL of concentrated HNO₃ was added once boiling commenced to drive the digestion to a clear solution; the cooled digest was filtered through Whatman No. 42 filter paper and made to 100 mL with double-distilled water. Each sample was prepared three times (analytical triplicate). Instrumentation: flame atomic absorption spectrophotometry on a Model AA240FS instrument (the model number identifies a Varian-family FAAS) with air–acetylene flame; fuel pressure and hollow-cathode lamp were adjusted per element. Calibration: stock-solution dilutions; absorbance-vs-concentration curves measured before sample runs. The authors do not report procedural blanks, LOD/LOQ values, certified reference material recoveries, or matrix-matched standards.

Implications

Certification: Lead was below detection in all sixteen Kathmandu packaged fruit-juice samples (FAAS, no explicit numeric LOD reported). The dataset is small (n=16) and the analytical method (FAAS without GFAAS or ICP-MS confirmation, no CRM recovery reported) is not the most sensitive available for Pb at sub-ppb levels; “below detection” here bounds rather than excludes contamination.

Courses: Useful as an example of how drinking-water guideline values are commonly substituted for fruit-juice-specific maximum levels in regional studies when the Codex MLs (Pb 0.03 mg/kg general, 0.04 mg/kg grape, 0.05 mg/kg berry/small-fruit) are not invoked. Also illustrates why FAAS with R²=0.978 on the Pb curve is at the lower edge of usable for sub-ppb Pb work.

App: Supports a low-Pb signal for commercially packaged fruit juices (apple, mango, orange, lychee) on the Kathmandu retail market under the limitations above. Iron exceedance of the inferred 2 mg/L water-derived reference in one apple-juice sample is not a Pb-relevant signal for HMTc but is worth flagging on the Fe nutritional-balance side.

Wiki pages this source may touch

• apple-juice
• mango-juice
• orange-juice
• fruit-juice
• fruit-juice-not-canned
• fruit-juices-apple-containing
• fruit-juices-non-apple
• lead
• iron
• copper
• zinc

Verification notes

• 2026-05-19 merge-enhanced under the v2.0 ingest skill. Preserved cite_key, raw_handle: manual-fetch-kimi, and near_duplicates: [] per skill rule. Corrected the truncated raw_path (previously “Analysis and Detection of Heavy Metals Content in Some Selected P.pdf” → full filename ending “Flame Atomic Absorption Spectroscopy.pdf”). Added access_url, raw_sha256, ingest_method, and created_by fields; added the journal volume/issue/pages to publication. Re-set updated to 2026-05-19.
• Frontmatter metals expanded from [Pb] to [Pb, Fe, Cu, Zn] — the prior listing dropped three of the four metals the source actually measured. All four have wiki metal pages.
• Frontmatter ingredients previously listed [[ingredients/lychee-juice]], which is not a wiki page; “lychee-juice” is registered only as an alias on ingredients/fruit-juice.md. Replaced the invalid slug with [[ingredients/fruit-juice]] so the lychee evidence still routes to the parent ingredient. (Per CLAUDE.md Part 10 the lychee-juice page is below the 5-paper threshold for its own ingredient page; this is the routing the alias was created for.) matrices: similarly trimmed to remove the bare lychee-juice matrix that has no controlled-vocabulary entry.
• Evidence tier reduced from A to B: the source is peer-reviewed open-access, but the methods section reports no CRM recovery, no LOD/LOQ table, no procedural-blank values, and no matrix-matched standards. Pb calibration R²=0.978 is at the lower edge of accepted linear-range fit. Per CLAUDE.md Part 13 evidence-tier rules, those omissions place this paper in B rather than A.
• The source’s calibration values for Cu, Zn, and Pb in Table 1 use truncated decimal forms (y=0.173x+0.007, y=0.12x+0.008, y=0.039x+0.030) while the figures (Figs 2–4) carry one more decimal place. The Key numbers table reproduces the Table 1 forms because that is the source’s stated calibration record.
• WHO/FAO comparison values used by the authors are reproduced as the authors state them. The 2 mg/L Fe reference is explicitly an inference from the JECFA 1983 PMTDI, not a fruit-juice MAC; the 0.01 mg/L Pb reference is the WHO drinking-water guideline; the Codex MLs for Pb in fruit juice (0.03 mg/kg general; 0.04 mg/kg grape; 0.05 mg/kg berry/small-fruit) are noted on this page but were not used by the authors as the primary comparator.
• Audit subagent (2026-05-19) flagged Check 4 ❌ for per-brand contamination values reproduced under the source’s anonymous A/B/C/D codes throughout the Key numbers narrative — the original draft of this page named “brand C” three times as the highest-Fe, highest-Cu, and highest-Zn brand. Verified against Part 12 strict reading (locked 2026-05-17): even though the source’s A/B/C/D codes do not disclose real brand names, reproducing per-brand rankings on the wiki side is the brand-by-brand ranking pattern Part 12 prohibits. Finding applied — Key numbers tables and narrative now report ranges, wiki-computed flavour means, and counts of samples above thresholds, with no per-brand attribution. The previous Verification-notes paragraph that defended the per-brand reproduction has been removed.
• Audit subagent (2026-05-19) flagged Check 5 ⚠️ for a cross-source synthesis claim (“consistent with prior Nepali and South-Asian beverage studies…”) and an imperative voice (“the certifier should treat…”) in the Implications → Certification block. Verified against Part 2 firewall: both are real boundary slips. Synthesis sentence removed; “should treat” softened to descriptive “bounds rather than excludes contamination.”
• Audit subagent (2026-05-19) flagged Check 1 ⚠️ that “Pb working range covering 0–10 mg/L” was inferred from the x-axis of Fig. 4 rather than stated by the source. Verified against source p. 160 Fig. 4 and p. 159 Methods: the inference is reasonable but the source does not use the words “working range.” Wording corrected to “the x-axis of Fig. 4 (p. 160) extends to ~10 mg/L, but the source does not explicitly state a Pb working range or a numerical LOD.”
• Audit subagent (2026-05-19) flagged Check 1 ⚠️ that the “n=48 measurements over 16 samples” framing was wiki-derived from “samples were prepared three times.” Verified against source p. 159: source uses “prepared three times to minimize the possibility of error” without stating an n=48 count. Sample-population field rewritten to quote the source verbatim and drop the n=48 framing.
• Audit subagent (2026-05-19) flagged Check 1 ⚠️ that the Fe “Mean of four” column is wiki-computed because the source does not report flavour means. Verified: arithmetic checked, source Table 2 does not report means. Footnote added under the Fe table noting the means are wiki-computed.
• Audit subagent (2026-05-19) flagged Check 2 ❌ that ingredients/apple-juice and ingredients/mango-juice are not present in the taxonomy snapshot. Verified against docs/gpt-collaboration/taxonomy-snapshot.md: both slugs ARE present in the snapshot’s Ingredients line (alongside orange-juice, fruit-juice, grapefruit-juice, etc.); both have wiki pages at wiki/ingredients/apple-juice.md and wiki/ingredients/mango-juice.md. The audit finding is a false positive — the subagent missed these slugs while scanning the comma-separated list. No change applied.
• No HMTc threshold proposals, consumer risk advisories, or remaining cross-literature synthesis claims appear in the body — Part 2 firewall clean after audit-application edits. Part 12 brand firewall clean after audit-application edits.

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.