This US retail study measured total arsenic, cadmium, lead and total mercury in 51 maintenance or all-life-stage dry dog foods. Each formulation contributed one production lot. The authors report results per unit of food energy and compare three formulation groups, each containing 17 products.
Key numbers
Table 1 reports medians and ranges in mg per 1,000 kcal of metabolizable energy (mg/Mcal).
| Formulation group | Total As | Cd | Total Hg | Pb |
|---|---|---|---|---|
| Poultry, N=17 | 0.054 (0.007–0.133) | 0.015 (0.005–0.063) | 0.0008 (0.0004–0.0023) | 0.037 (0.019–0.305) |
| Red meat, N=17 | 0.037 (0.007–0.134) | 0.013 (0.008–0.047) | 0.0012 (0.0004–0.0064) | 0.091 (0.032–1.621) |
| Fish, N=17 | 0.343 (0.025–1.104) | 0.027 (0.014–0.215) | 0.0082 (0.0010–0.0139) | 0.049 (0.018–0.325) |
Fish-based foods had higher As, Cd and Hg than the other two groups (source-reported p<0.01). For Pb, the results and Figure 4 identify higher red-meat values than poultry; fish differs significantly from neither group. The abstract’s broader Pb comparison is not supported by that pairwise result.
Methods (brief)
Forty-seven foods came from an online retailer, one from a grocery, one from a local pet retailer and two from a veterinary center. Samples were analyzed within a week of acquisition. The source does not give the calendar sampling period. Triplicate 5 g aliquots from each submitted 100 g sample were analyzed; these are analytical replicates, not additional products.
Microwave-assisted nitric-acid digestion and oven drying at 40°C preceded ICP-MS using an Agilent 7500ce and Mira Mist nebulizer in hydrogen mode. The reported quantitation limit was 0.001 mg/kg as fed for all four metals. The authors state that all samples were above detection; they do not give a separate numeric detection limit. Intra-assay CV was below 5%, and inter-assay CV below 7%.
The authors used manufacturer-provided metabolizable energy to express the average of triplicate readings per 1,000 kcal. Product-level energy values are not supplied, so the reported group summaries cannot be reversed into mg/kg group concentrations. Shapiro-Wilk testing was followed by Kruskal-Wallis and Dunn comparisons, with medians and ranges reported.
Evidence fitness
Supports per-energy occurrence in finished dry dog foods and associations with formulation groups. The four analytes were measured without speciation; total arsenic and total mercury do not establish inorganic arsenic or methylmercury. Formulation labels do not constitute separate measurements of raw fish, poultry or red meat.
Limitations
One lot per formulation does not characterize batch variability. Ingredient provenance, acquisition dates, individual concentrations and exact graphical quartiles are absent. The study measured foods, not animal health outcomes or contaminant bioavailability. Explanations involving fishmeal, game-meat exposure, processing water or equipment remain hypotheses.
The cited human-intake comparisons use historical studies rather than concurrent controls. The prose arsenic maximum-to-median ratio has an inconsistent denominator. Figure 4’s human lead reference line also appears inconsistent with the numeric comparator in Table 1 and the results text. Figure 3’s poultry-mercury upper whisker differs from the maximum in both Table 1 and the text; the exact tabulated value is retained here. Neither these comparisons nor cross-species toxicity references establish lifetime pet safety or current regulatory compliance.
Related evidence
Update history
No substantive edit history is available in this build. The full commit record is available in git.