Du et al. 2025 — Toxic metals in commercial cat and dog food, Chinese market

This study measured Pb, Cd, Cr, Hg, and As in 93 imported cat and dog food products collected from Chinese supermarkets and store chains (Beijing, Shenzhen, Wuhan) in 2021-2022 and conducted an acute hazard index (aHI) assessment for cats and dogs across body-weight categories. Flame atomic absorption spectrometry (FAAS) quantified Pb, Cd, Cr; atomic fluorescence spectrometry (AFS) quantified Hg and As. Cr and As were detected in 100% of samples; Hg in 97.85%. Cr had the highest average concentrations of the five metals. Against the Chinese Hygienical Standard for Pet Feed (Announcement No. 20 of 2018), 8.60% of samples exceeded the Pb limit and 26.88% exceeded the Cr limit; Cd, Hg, and As remained within their limits across all 93 samples. Dry food showed higher contamination rates and higher mean concentrations of Pb, Cd, Cr, and As than canned food. Combined contamination by three or more metals was the norm; Cr+Hg+As was the dominant triad. aHI values for Cr, Pb, and As exceeded the safety threshold (>1) across all simulated body weights.

Key numbers

93 samples total: 45 cat food, 48 dog food; 79 dry, 14 canned. Imports from USA (32), Spain (29), Thailand (17), Germany (10), New Zealand (5).

Chinese Hygienical Standard for Pet Feed legal limits (mg/kg, per published Table 2): Pb 5, Cd 2, Cr 5, Hg 0.3, As 10.

All samples (n=93; published Table 2):

• Pb: 39 contaminated (41.94% above LOQ 2 mg/kg); 8 above legal limit (8.60%); min 2.07, max 5.80, mean 3.63 mg/kg
• Cd: 29 contaminated (31.18% above LOQ 0.20 mg/kg); 0 above limit; min 0.0200, max 0.279, mean 0.138 mg/kg
• Cr: 93 contaminated (100% above LOD 150 µg/kg); 25 above limit (26.88%); min 0.48, max 15.47, mean 3.47 mg/kg
• Hg: 91 contaminated (97.85% above LOD 0.15 µg/kg); 0 above limit; min 0.001, max 0.027, mean 0.011 mg/kg
• As: 93 contaminated (100% above LOD 0.010 mg/kg); 0 above limit; min 0.02, max 1.89, mean 0.21 mg/kg

Cat food (n=45) vs dog food (n=48), contamination rate / mean (mg/kg):

• Pb: cat 42.22% / 3.93; dog 41.67% / 3.34. Exceedance: cat 8.89%, dog 8.33%.
• Cd: cat 33.33% / 0.142; dog 29.17% / 0.134. Exceedance: 0 / 0.
• Cr: cat 100% / 3.31; dog 100% / 3.62. Exceedance: cat 22.22%, dog 31.25%.
• Hg: cat 95.56% / 0.011; dog 100% / 0.012. Exceedance: 0 / 0.
• As: cat 100% / 0.16; dog 100% / 0.26. Exceedance: 0 / 0.

Mean Pb and Cd were higher in cat food than dog food; mean Cr and As were higher in dog food than cat food.

Canned cat food (n=12) vs dry cat food (n=33) (published Table 4):

• Pb: canned 0% contamination; dry 57.58% contamination, 12.12% above limit.
• Cd: canned 16.67%; dry 39.40%; 0 above limit either form.
• Cr: 100% both; canned 0 above limit; dry 30.30% above limit.
• Hg: canned 83.33%; dry 100%.
• As: 100% both.

Canned dog food (n=2) vs dry dog food (n=46) (published Table 5):

• Pb: canned 0%; dry 43.48%, 8.70% above limit.
• Cd: canned 0%; dry 30.43%; 0 above limit either form.
• Cr: 100% both; canned 0 above limit; dry 32.61% above limit.
• Hg: 100% both.
• As: 100% both.

(The Research Square preprint reports the canned/dry split pooled across cat+dog: canned n=14 with 0% Pb, 14.29% Cd, 100% Cr, 85.71% Hg, 100% As contamination; dry n=79 with 49.37% Pb [10.13% above limit], 34.18% Cd, 100% Cr [31.65% above limit], 100% Hg, 100% As. The published version splits these by species, as captured above.)

Combined contamination (n=93; Fig 1 published):

• All samples contained at least two metals.
• 2.15% contained two metals (Cr + As).
• The largest fraction contained three metals (Cr, Hg, As dominant).
• 19.35% contained four metals (Pb, Cr, Hg, As).
• 26.88% contained all five metals.
• In dog food specifically, every sample contained ≥3 metals.

Acute hazard index (aHI = ESTI / ARfD; US EPA non-carcinogen RfDs as ARfD surrogate) for the highest observed concentration of each metal (published Table 6, dimensionless; threshold of concern = 1):

• Pb (RfD 6 µg/kg/day): cat 2 kg 21.82, cat 5 kg 16.13; dog 5 kg 17.36, dog 10 kg 14.60, dog 30 kg 11.09.
• Cd (RfD 1 µg/kg/day): cat 2 kg 5.86, cat 5 kg 4.33; dog 5 kg 5.41, dog 10 kg 4.55, dog 30 kg 3.45.
• Cr (RfD 0.3 µg/kg/day): cat 2 kg 1040.14, cat 5 kg 768.73; dog 5 kg 404.03, dog 10 kg 339.75, dog 30 kg 258.15.
• Hg (RfD 0.1 µg/kg/day): cat 2 kg 13.28, cat 5 kg 9.81; dog 5 kg 4.32, dog 10 kg 3.78, dog 30 kg 2.88.
• As (RfD 0.3 µg/kg/day): cat 2 kg 124.97, cat 5 kg 92.36; dog 5 kg 94.50, dog 10 kg 88.20, dog 30 kg 67.20.

Ranking by aHI (highest concentration scenario): Cr ≫ As > Pb > Hg > Cd across both cats and dogs. The authors explicitly state As ranks second after Cr, followed by Pb. Lighter-weight animals had higher aHI values across all metals.

Methods (brief)

Sample preparation: Dry samples ground to pass 20-mesh sieve (TAISITE FW100 mill); canned samples homogenized (Precision Labs HM-7300). Pb (GB/T 13080-2018), Cd (GB/T 13082-2021), Cr (GB/T 13088-2006), As (GB/T 13079-2006) digested by dry ashing (5 g sample, DK-98-II furnace, then Nabertherm L9/11/B170 muffle furnace). Hg (GB/T 13081-2006) digested by microwave (0.20-1.0 g sample, 69-70% HNO₃ + 30% H₂O₂ in SINEO MASTER 40 system, three-stage program to 190 °C at 2100 W).

Analytical instruments: Pb, Cd, Cr by flame atomic absorption spectrometry (ZEEnit 700 P, Analytik Jena AG; wavelengths Pb 283.3 nm, Cd 228.8 nm, Cr 359.3 nm). Hg, As by atomic fluorescence spectrometry (AFS-8530, Beijing Haiguang Instrument Co.). Calibration standards from Guobiao (Beijing) Testing & Certification Co. (1000 µg/mL).

LOD/LOQ (published Table 1): Pb LOQ 2 mg/kg; Cd LOD 0.08 / LOQ 0.20 mg/kg; Cr LOD 150 µg/kg; Hg LOD 0.15 µg/kg; As LOD 0.010 mg/kg.

Quality control (Sci Reports published version only — not in preprint): each batch of 21 samples included parallel samples, reagent blanks, and certified reference materials. Spike recovery rates: Pb 105.28%, Cr 117.93%, Cd 102.52%, Hg 89.9%, As 107.6%. RSD across 4 replicate determinations: 2.88-4.16%.

No As speciation reported (total As only). No Hg speciation reported (total Hg only). The authors discuss methylmercury toxicity in cats but do not measure MeHg.

Health risk model: ESTI = HER × K where HER is the highest observed concentration in the sample set for each metal and K is the manufacturer’s recommended daily food intake. K values (Table 2 preprint, published Table not directly accessible but consistent): 33 g/day (2 kg cat), 63 g/day (5 kg cat), 106 g/day (5 kg dog), 179 g/day (10 kg dog), 408 g/day (30 kg dog). aHI = ESTI / ARfD; ARfD = US EPA non-carcinogen RfD (mammalian; no pet-specific ARfD exists for these metals).

Statistical analysis: Microsoft Excel 2016 and PowerPoint 2016 for descriptive statistics (contamination rate, min, max, mean of positive samples).

Implications

Certification: HMT&C does not certify pet food. The relevance to HMT&C threshold work is indirect: pet-food ingredient supply chains (meat and bone meal, fish meal, poultry by-products, cereal grains, tubers) draw from the same primary commodity streams as human-food ingredients; metal levels in pet food provide an indirect indicator of supply-chain contamination for those shared inputs. The 26.88% Cr exceedance rate against a 5 mg/kg pet-feed limit is striking but uses a much looser limit than human-food Cr standards.

Courses: Useful as a case study of (1) total vs speciated metal reporting (the paper measures total Cr without Cr-VI speciation despite explicitly noting Cr-VI’s toxicity), (2) processing-form effects on contamination (dry kibble carries more metal than canned wet food across Pb, Cd, Cr, As; the authors attribute this to puffing/granulating/baking and to packaging considerations), and (3) the acute-hazard-index framework with mammalian RfD surrogates when species-specific reference doses are absent.

App: Pet-food matrix; not applicable to the human-food contamination-profile data the app consumes.

Verification notes

2026-05-18 (Claude Code, merge-enhance pass on PCMF preprint ingest):

This wiki page exists for the published Scientific Reports version of the study (FM_12006530, Marker source). The Research Square preprint of the same study (PCMF_individual-and-combined-contamination-of-the-heavy; DOI 10.21203/rs.3.rs-4838876/v1, posted 2024-09-08) was added to the manual-fetch queue. Cross-checking the two versions against the page:

• The published Sci Reports version has the same authors, sample size (n=93), the same overall contamination rates (Table 2: 41.94% Pb, 31.18% Cd, 100% Cr, 97.85% Hg, 100% As), and the same exceedance rates (8.60% Pb, 26.88% Cr). The published version splits the canned/dry breakdown by species (Tables 4-5: 12+2 canned, 33+46 dry); the preprint pools across species (Table 5: 14 canned, 79 dry). Pooled numbers reconcile exactly.
• The published version adds a Quality Control section with spike recoveries and RSD; the preprint does not.
• The published aHI values (Table 6) differ substantially from the preprint Table 6 — published values are higher across the board (e.g., Cr cat 2 kg: 1040.14 published vs 182.050 preprint; Pb cat 2 kg: 21.82 published vs 10.808 preprint). The K (daily food intake) values appear unchanged, so the difference must stem from a revised HER (highest observed concentration) input or a unit-handling fix between preprint and publication. The published values are used here as the canonical, peer-reviewed numbers.
• The aHI ranking statement in the published Discussion is explicit: “In both cats and dogs, As ranks as the second highest in aHI value, followed by Pb.” The preprint Discussion ranks Pb second and As third; the published version corrects this. The prior wiki page mirrored the preprint ranking (Pb > As). Corrected to Cr > As > Pb > Hg > Cd.

Defects corrected from the prior wiki revision (updated 2026-05-13):

• Pb contamination rate was reported as “10.75%” — wrong. Actual: 41.94% above LOQ (Table 2). The 10.75% appears to be a misread; 10.13% is the % above legal limit in dry food only (Table 5 preprint), not the overall contamination rate.
• Cd contamination rate was reported as “17.20%” — wrong. Actual: 31.18% above LOQ (Table 2). No clear source for the 17.20% figure.
• Chinese Hygienical Standard limits were stated as “Pb ≤5, Cd ≤0.5, Cr ≤5, Hg ≤0.1, As ≤4 mg/kg” — three of five were wrong. Actual (published Table 2 row “Legal limits”): Pb 5, Cd 2, Cr 5, Hg 0.3, As 10 mg/kg.
• aHI ranking stated as “Cr poses highest aHI followed by Pb and As” — wrong direction between Pb and As. Corrected to “Cr ≫ As > Pb > Hg > Cd” per the published Discussion.
• Products frontmatter listed only pet-food-dry-kibble; added pet-food-canned-wet since the paper covers 14 canned samples (12 cat, 2 dog) explicitly. Both product pages exist in current taxonomy.
• Added preprint_doi: 10.21203/rs.3.rs-4838876/v1 to capture the Research Square preprint citation. The published Sci Reports DOI is not in the Marker extraction and cannot be reconstructed without going to the Nature URL; left as null.
• Added PCMF preprint handle to near_duplicates so the routing audit can track that both raw sources point to the same study.
• Kept no_doi_assigned: true (the published Sci Reports version surely has a DOI but it was not extracted from the Marker PDF and cannot be reconstructed without going to the Nature URL; the preprint DOI is captured in the separate preprint_doi field).
• Sample population description corrected: the published version specifies samples were collected in supermarket/store chains across Beijing, Shenzhen, and Wuhan; the preprint says “China’s supermarket and store chains” without naming cities. Used the published version’s more specific sourcing.
• Expanded Methods with explicit GB/T national standard numbers, instrument vendor model numbers, LOD/LOQ values per Table 1, and the QC recovery/RSD details from the published Quality Control section.
• Expanded Implications to address why pet food bears on HMT&C’s human-food remit (ingredient supply-chain overlap) rather than asserting it directly.

No brand-firewall violations to remove (the study does not name brands; sample identification is by country of origin only). No HMT&C threshold proposals or consumer risk advisories in the page.

2026-05-18 (fresh-context audit subagent, verdict REVISE):

Checks 1, 3, 4, 5 returned ✅. Check 2 raised a ⚠️ finding that products/pet-food-dry-kibble and products/pet-food-canned-wet are not present in the 2026-05-17 taxonomy snapshot at docs/gpt-collaboration/taxonomy-snapshot.md. Verified false positive: both pages exist in live wiki/products/ (ls wiki/products/ | grep '^pet-food-' returns the expected files), and the routing audit successfully generated direct-evidence rows for both targets without raising unresolved entries. The snapshot is stale relative to live state — pet-food product pages were added after the snapshot’s generated: 2026-05-17 timestamp. No correction to the source page is needed; flagged here so the snapshot can be regenerated in a separate housekeeping pass.

All other findings ✅: full Table 1-6 numerical fidelity verified, Methods/speciation correct (tHg + tAs, no Cr-VI speciation noted), brand firewall clean (Methods instrument/reagent vendors are Exception 2 permitted), wiki/HMTc firewall clean (Implications appropriately scopes to indirect supply-chain framing without proposing thresholds).

Wiki pages this source may touch

• lead
• cadmium
• chromium
• mercury-total
• arsenic-total
• pet-food-dry-kibble
• pet-food-canned-wet

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.