CFIA 2020 — Toxic metals in selected foods, April 2018–March 2019

The Canadian Food Inspection Agency tested 985 samples of bran products, infant formula, meal replacement products, protein powders, and rice products collected from retail outlets in six Canadian cities (Halifax, Montreal, Toronto, Ottawa, Vancouver, Calgary) over fiscal year 2018–2019 for arsenic, cadmium, lead, and mercury. Across all 985 samples, 70% contained at least one of the four metals and 12% contained traces of all four; rice products had the highest mean arsenic level (0.263 ppm, max 1.13 ppm in brown rice bran/germ powder) and protein powders had the highest mean cadmium (0.060 ppm), lead (0.035 ppm), and mercury (0.003 ppm) levels. Infant formula had the lowest detection rates and the lowest mean values for all four metals; all arsenic and lead results in ready-to-serve infant formula and ready-to-serve meal replacement beverages met Canada’s existing tolerances, and Health Canada concluded that no sample in the survey posed a concern to human health.

Key numbers

Sample distribution (Table 1, p. 4): 985 total samples — 99 bran products (23 domestic, 41 imported, 35 unspecified origin), 395 infant formula (0 domestic, 386 imported, 9 unspecified), 198 meal replacement (16/135/47), 195 protein powders (72/34/89), 98 rice products (77/19/2).

Detected levels by product type (Table 2, p. 5; values in ppm = mg/kg; averages calculated using only quantifiable results; LOD 0.001–0.01 ppm):

Arsenic narrative (p. 5): overall detection 64%; meal replacement lowest (44%), rice products highest (95%). The two highest values (1.12 and 1.13 ppm) were both in brown rice bran and germ powder. All arsenic levels in ready-to-serve meal replacement beverages met the existing 0.1 ppm tolerance.

Cadmium narrative (p. 6): 596 of 985 samples (61%) below detection; range 0–0.219 ppm across the survey. Protein powders carried the highest detection rate, maximum, and average level; infant formula the lowest.

Lead narrative (p. 7): 656 of 985 samples (67%) below detection. Detection rate lowest in infant formula (13%) and highest in protein powders (71%); range 0–0.237 ppm. Lead in ready-to-serve infant formula and ready-to-serve meal replacement beverages met existing tolerances.

Mercury narrative (p. 8): overall detection 19%; only 2 of 395 infant formula samples were above LOD. Rice products had the highest detection rate (71%); protein powders had the highest concentrations. A single brown rice protein powder sample reached 0.022 ppm, the survey maximum.

Cross-year context (Table 3, p. 10): infant formula detection rates and means in 2018–19 (n=395) are similar to the 2017–18 survey (n=177); detections in 2018–19 are higher than in the 2012–13 survey (n=144) for As but lower for Pb. Protein powder mean cadmium fell from 0.181 ppm in 2012–13 (n=101) to 0.060 ppm in 2018–19 (n=195), which the report attributes to a smaller proportion of rice-based protein powders in the later sample. Rice product mean arsenic rose from 0.132 ppm in 2009–15 (n=975) to 0.263 ppm in 2018–19 (n=98); the report attributes this to two brown rice bran and germ powder outliers (1.12 and 1.13 ppm As).

Methods (brief)

Samples were analyzed by an ISO/IEC 17025 accredited food testing laboratory under contract with the Government of Canada or by a CFIA laboratory; the report does not specify the analytical technique (no ICP-MS / ICP-OES / AAS attribution given). LOD ranged 0.001–0.01 ppm depending on laboratory and analyte (some values in Table 3 use LOD as low as 0.0004 ppm). Results are expressed on a “food as sold” basis, not as consumed; reconstitution corrections are not applied. Average values were calculated using only the subset of samples with quantifiable metal levels (above LOD), not the full sample set; this convention inflates the reported means relative to a left-censored arithmetic mean and should be retained in any downstream comparison. Arsenic, cadmium, lead, and mercury are reported as total elements — the report does not speciate inorganic vs total arsenic, methylmercury vs total mercury, or hexavalent vs total chromium (chromium was not measured). Compliance was assessed against the 2014 Health Canada tolerances for arsenic and lead in ready-to-serve beverages and ready-to-serve infant formula; no Canadian regulatory maxima existed for metals in the other product categories sampled.

Limitations

The “food as sold” basis means powder protein and powder formula values cannot be directly compared to ready-to-serve liquid values without a reconstitution adjustment (powder values look numerically higher than the as-consumed equivalent because powders concentrate ingredients before water addition). Detection rates and means depend on LOD; cross-year comparisons in Table 3 are confounded by improved instrument sensitivity in later surveys (LOD ~0.0004 ppm in later years vs higher LODs in earlier surveys). Sample sizes are uneven across product categories (n=98 rice products vs n=395 infant formula). Country-of-origin is unspecified for 182 of 985 samples (18%). Analytical method, instrument vendor, sample digestion procedure, and reference materials are not disclosed.

Implications

• Certification (HMTc): Protein powders surface as the product class with the highest detection rates and maxima for cadmium, lead, and mercury — material input for Category 16 (Dietary Supplements) Row 20 (protein/collagen powders) threshold-setting. Rice products carry the highest mean and maximum arsenic (1.13 ppm); brown rice bran and germ powder products specifically should be flagged as a within-category outlier sub-form. Infant formula values met existing Canadian tolerances on both arsenic and lead, with infant formula detection rates the lowest across all four metals in the survey.
• Courses: Useful regulatory-compliance case study showing how Canada’s 2014 tolerances apply only to ready-to-serve beverages and ready-to-serve formula, leaving meal-replacement powders, protein powders, bran products, and rice products without category-specific Canadian maxima.
• App: Adds tAs, Cd, Pb, and tHg data points for ingredient rice (rice products and rice-based protein powder), and tAs, Cd, Pb, tHg for protein-powder product class generally. Note: total-arsenic reporting; not inorganic-arsenic-specific, which limits direct application to rice iAs thresholds.
• Microbiome: Not addressed.

Wiki pages updated on ingest

• arsenic
• cadmium
• lead
• mercury
• rice
• oat
• wheat
• corn
• infant-formula-powder
• infant-formula-rtf-liquid-non-soy
• supplements-protein-collagen-powders
• rice-bulk-grain
• other-grain-products

Verification notes

• Filename prefix is “HealthCanada” but the issuing body on the report cover and throughout the text is the Canadian Food Inspection Agency (CFIA). Health Canada conducted the final health risk assessment; CFIA conducted the sampling and chemistry. Cite key uses cfia2020- to match the existing cfia2025-toxic-metals-selected-foods-2022-23 series and the report’s true authorship.
• Publication year inferred as 2020 from PDF metadata (CreationDate 2020-08-10; no explicit publication date printed in the report body). The survey covers fiscal year 2018-04-01 through 2019-03-31.
• No DOI assigned (CFIA targeted-survey reports are not DOI-minted); access_url points to the inspection.canada.ca landing page convention used for prior-year reports.
• No exact taxonomy slug exists for “meal replacement products (ready-to-serve beverage mixes)” — the n=198 meal-replacement subset is left off products: rather than mis-routed to supplements-protein-collagen-powders (a different product class) or category-5-beverages (a Category 5 taxonomy index, not a routing destination). The meal-replacement values are still captured in this page’s Key numbers table; the routing layer will surface this as a candidate provisional scaffold if a second source touches the same category.
• The “average” values in Tables 2 and 3 are computed only over quantifiable samples, not over all samples in the cell. This convention inflates means versus a standard left-censored arithmetic mean and is preserved verbatim in the Key numbers table; downstream synthesis passes should not treat these means as ND-inclusive arithmetic means.
• Brand firewall: the survey reports values only at the product-category level (no brand names in the source PDF), so no brand redaction was required.
• Speciation: arsenic and mercury are total-only in this report (no iAs/MeHg breakout). Chromium and tin were not measured.

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.