Anastario et al. 2025 — Metal profiling of Hass avocados: ICP-MS and pXRF

This pilot study quantified 24 elements across 24 Hass avocado samples using ICP-MS (primary) and portable XRF (pXRF, screening tool validation), stratifying by country of origin (California, USA vs. Michoacán, Mexico) and agricultural practice (organic vs. conventional). It provides baseline data on elemental distribution between the edible mesocarp (pulp) and exocarp (peel), finding that metals generally accumulate more in the exocarp. Notably, mercury was elevated in avocado tissues relative to comparable tropical fruits, particularly in exocarp and in USA-sourced samples, which the authors attribute to possible Hg accumulation in lipophilic matrices. Cd, Pb, and As were detected at relatively low concentrations in mesocarp.

Key numbers

Concentrations are reported as geometric mean (95% CI) per Table 1. The basis is µg/g dry weight for all elements except Hg, which is in ng/g dry weight per the Table 1 footnote. Each stratum is n=6 (4 strata × 2 tissues = 48 sample-measurements).

Mesocarp (edible pulp), HMTc-priority analytes — geometric means in µg/g dry weight unless noted:

*The Pb mesocarp Cal-USA organic value of 0.14 µg/g represents a single sample above LOD (range not reportable); it drove the Pb loading on mesocarp PC1 and is described by the authors as the “second cluster… notable for moderate Pb.” Treat as suggestive of an organic-USA Pb outlier, not as a stratum geometric mean.

Exocarp (peel, generally not eaten), HMTc-priority analytes — geometric means in µg/g dry weight unless noted:

Mercury fresh-weight reconstruction (from Discussion). Maximum mesocarp Hg after moisture-content adjustment to fresh weight: approximately 0.012–0.021 mg/kg. Exceeds the stringent EFSA-cited produce reference value of 0.01 mg/kg, well below the FAO/WHO Codex CXS 193 fish range of 0.5–1.0 mg/kg. The authors flag this as warranting follow-up via Cold Vapor Atomic Fluorescence Spectroscopy and speciation work (no speciation was performed in this study).

Macro-elements (mesocarp, µg/g dry weight): K 18,084–27,042; Mg 1,155–1,392; Ca 580–669 (lowest of all groups in mesocarp). Exocarp values are markedly higher: K up to 39,111, Mg up to 2,484, Ca up to 2,785 (all in Cal-USA conventional). Higher exocarp concentrations are consistent with direct environmental exposure or mesocarp-to-exocarp uptake.

Other detected elements (selected, dry weight µg/g): Fe (mesocarp 12.96–18.72; exocarp up to 53.90), Mn (mesocarp 4.61–6.84; exocarp 7.48–15.01), Cu (mesocarp 8.48–9.96; exocarp 15.53–32.31, with highest in Mexico conventional exocarp), Zn (mesocarp 16.13–19.05; exocarp 17.89–29.53), Co (mesocarp 0.03–0.04; exocarp 0.06–0.15), Mo (mesocarp 0.02–0.11; exocarp 0.07–1.06, with highest in Cal-USA conventional exocarp).

Not detected in any sample (LOD given in µg/L of digestate): Be (LOD 0.005), Ag (LOD 0.014), Sn (LOD 0.010). Sn non-detection across all 48 sample-measurements is the only HMTc-analyte concrete null.

Metal-mixture signatures (PCA + K-means, k=3).

• Mesocarp PC1: Pb, Ni, Na (high loadings). PC2: Se, As, Zn, K. PC3: Ni, Co.
• Exocarp PC1: Fe, Mo, Mg, Hg. PC2: Sb, Na, Al, As. PC3: Ni, Ba, Mn, Na.
• Mesocarp Co–Ni co-occurrence (both >1 SD above mean) appeared only in organic Cal-USA avocados, attributed by the authors to ultramafic/serpentine soil geology in California.
• Exocarp Mg–Fe–Mo–Hg signature was more frequently present in USA-sourced samples than Mexican-sourced (χ² = 4.29, p = 0.038).
• Cluster membership in mesocarp was significantly associated with country of origin (χ² = 8.19, p = 0.017); weakly associated with farming type.

Correlation highlights (Pearson r ≥ 0.50): Hg and Zn (mesocarp r=0.70, exocarp r=0.71, cross-tissue r=0.52); Hg and K (mesocarp r=0.55, exocarp r=0.87, cross-tissue r=0.74); K and Mo (mesocarp r=0.62, exocarp r=0.72, cross-tissue r=0.67); Co and Ba (mesocarp r=0.62, exocarp r=0.72, cross-tissue r=0.59). The overall MANOVA test (Pillai’s trace=0.84, F(20,21)=4.9, p<0.001) confirms a significant interaction of origin × practice × tissue on the elemental profile.

pXRF calibration vs ICP-MS. Significant linear relationships in mesocarp for K (β=1.12, 95% CI 0.48–1.77), Ca (β=0.85, 95% CI 0.52–1.17), and Zn (β=1.26, 95% CI 0.86–1.66). In exocarp, K, Ca, Fe, Cu, and Zn showed the strongest fits. Mg and Al are outside the Niton XL5 Plus Soils-Mode analyzer range. Trace elements with frequent non-detects (V, Cr, Mn, Co, Ni, Se, Ba) could not be calibrated. pXRF is positioned as a field-screening prioritization tool, not a quantitative substitute for ICP-MS.

Sample size and design. n=24 fruits (6 per origin × practice stratum), each split into mesocarp and exocarp = 48 sample-measurements. Sampling window: November 11, 2024 – January 15, 2025. Commercially sourced from US grocery stores and wholesale suppliers; origin and practice assigned from supplier/store labeling (no farm-level provenance verification). Pilot study explicitly positioned as a precursor to a larger biosocial study with paired farm and soil sampling. The small per-stratum n and the absence of paired soil composition data limit interpretability of co-occurrence patterns. Authors note that grinder blanks were not analyzed, so trace contamination during sample homogenization cannot be fully excluded.

Methods (brief)

Cross-sectional pilot, n=24 Hass avocados commercially sourced in the US (12 Cal-USA, 12 Michoacán-Mexico; 6 organic + 6 conventional per origin), November 2024 – January 2025. Each fruit split into mesocarp and exocarp for separate analysis. Exocarp dehydrated at 57.2°C × 48 h; mesocarp × 96 h; ground with ceramic mortar and pestle (no grinder-blank validation).

ICP-MS. Microwave-assisted acid digestion in a MARS 6 system (CEM): ~0.1 g dried sample + 2 mL trace-metal-grade HNO₃ in MARSXpress vessels, ramped to 200°C over 20 min and held 10 min. Diluted to 15 mL with ultrapure water and spiked with 0.15 mL concentrated HCl. Two replicates of NIST 1515 Apple Leaves CRM included for QC. Concentrations reported in µg/g dry weight for all elements except Hg, which is reported in ng/g dry weight (Table 1 footnote). 24 elements assessed; LOD given per element in µg/L of digestate.

pXRF. Thermo Niton XL5 Plus in Soils Mode under factory calibration, 180 s per sample, four Thermo Scientific CRMs (QC Standard Multi Element RCRA, QC Standard Soil, QC Standard Blank SiO₂, Control Sample Soil) verified ±20% of certified values. Three unused 5 mL Eppendorf tubes used as background. Mg and Al fall outside the analyzer’s range.

Speciation. No arsenic speciation (total As reported only). No mercury speciation; authors flag this as a follow-up priority and suggest Cold Vapor Atomic Fluorescence Spectroscopy for elemental Hg.

Statistics. Geometric means with 95% CI per stratum; MANOVA (Pillai’s trace) for the origin × practice × tissue interaction; Pearson and Spearman correlations; PCA with eigenvalue >1 retention; K-means clustering (k=3) with elbow and silhouette method; chi-square tests for cluster–attribute association.

Implications

Certification: Contributes occurrence data for tHg in Hass avocado mesocarp and exocarp, including a moisture-adjusted fresh-weight maximum (0.012–0.021 mg/kg) that exceeds the EFSA-cited 0.01 mg/kg produce reference value while remaining below the FAO/WHO Codex CXS 193 fish range (0.5–1.0 mg/kg). Mesocarp values for Pb, Cd, and tAs are at low µg/g dry-weight levels with high non-detection rates; the most notable Pb observation is a single Cal-USA-organic detect at 0.14 µg/g dry weight that drives PC1 loadings but does not represent a stratum geometric mean. Origin-linked Co/Ni co-occurrence (present only in organic Cal-USA mesocarp) is a candidate signal for origin-aware research, attributed by the authors to ultramafic/serpentine soil geology rather than to farming practice.

Courses: The organic/conventional difference for Ni and Co (linked to soil geology, not farming practice) is a strong illustration of why origin matters more than label for certain metals. The pXRF vs. ICP-MS calibration data are relevant to field-screening curriculum in testing methods.

App: n=24 with 6 per origin×practice stratum is below the threshold for app-facing concentration distributions; the page should be treated as provisional until larger-n occurrence data are available. The tissue split (mesocarp vs. exocarp) matters for dietary-exposure inputs: consumers do not eat the exocarp, so the mesocarp values — not the higher exocarp values — are the appropriate basis for ingestion estimates.

Wiki pages updated on ingest

• avocado
• mercury
• nickel
• lead

Verification notes

• 2026-05-25 (Claude, merge-enhance): Re-read the source PDF in full including Table 1 (continued) on page 5. Extracted precise per-stratum geometric means (95% CI) for Pb, Cd, tAs, Hg, Ni, and Al in both mesocarp and exocarp. Prior version of this page declared these “not fully resolvable from text extraction”; that gap is now closed.
• 2026-05-25: Corrected the Hg unit. Table 1 footnote: “Values for Hg represented in ng/g,” whereas all other elements are µg/g dry weight. Prior version of the Methods section did not flag this distinction.
• 2026-05-25: Flagged Pb mesocarp Cal-USA organic = 0.14 µg/g as a single-detect outlier (only 1 of 6 samples above LOD), not a stratum geometric mean. This value drove the Pb loading on mesocarp PC1 and the “moderate Pb” cluster identified in the K-means analysis. Correct interpretation: organic Cal-USA mesocarp has at least one fruit with anomalously elevated Pb relative to all other strata; the geometric-mean value is not representative.
• 2026-05-25: Corrected raw_path frontmatter (previously truncated to P0309.pdf; actual file is P0309_Metal profiling of Hass avocados.pdf).
• 2026-05-25 (audit-application): Fresh-context audit subagent (REVISE verdict) flagged the n>LOD (of 48) column header as potentially misleading because Table 1’s n>LOD column is a single combined-tissue count, not per-tissue. Verified against Table 1 — the column shows e.g. Pb=12 with no tissue split. Relabeled column to n>LOD (combined tissues, of 48) in both mesocarp and exocarp tables.
• 2026-05-25 (audit-application): Subagent flagged three phrases in Implications/Certification as drifting toward HMTc programmatic recommendation rather than literature reporting (“supporting inclusion of tHg in risk monitoring,” “actionable for supply-chain screening guidance,” “Flag as provisional”). Verified against CLAUDE.md Part 2 — these phrases do cross from “what the literature reports” into “what HMTc should do.” Rewrote the Certification paragraph to factual occurrence-data framing and softened App-section advisory language. Preserved the substantive findings (Hg fresh-weight reconstruction, EFSA/Codex comparison, Pb single-detect outlier, origin-linked Co/Ni signature) without programmatic verbs.
• 2026-05-25 (audit-application note): Subagent flagged avocado-mesocarp and avocado-exocarp matrix slugs as outside the documented snapshot. These were preserved from the prior version of the page; the current routing audit accepts them and the source is not in routing_malformed.csv or routing_unresolved.csv. Tissue-level matrix slugs are the correct granularity for this paper because the mesocarp/exocarp distinction is foundational to its findings (different metal-mixture signatures by tissue, only mesocarp is consumed). Slug-vocabulary governance for tissue-level matrices is a routing-layer concern, not a per-source fix; leaving slugs as-is.

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.