One metal per product: protein powders, greens and honey — which heavy metal decides your batch?
Share
One metal per product: which heavy metal decides your protein powder, greens or honey
For each product a supplement brand sells, one element decides whether the batch passes, and it enters by one route. This article names that element for protein powders, chocolate-flavoured proteins, greens powders and bee products, gives its source, states the EU number in plain words, and marks the point in the process where a single test settles the question.
1One element per product
A heavy-metals panel reports four elements: lead, cadmium, mercury and arsenic. A brand with twelve products does not have forty-eight questions. For each product one of the four is the one that fails, it comes from one identifiable place, and Regulation (EU) 2023/915 gives it one number or, where it gives none, one rule for deriving it.
The figure below puts six measured products against the row that applies to each of them, expressed as a multiple of that row so that different elements and different limits can share one axis. Everything to the right of the gold line is above its limit.
2What the four numbers protect against
The limits are not arbitrary and they are not all of one kind. Each element has a different target organ, a different reference value behind its row, and a different relation between what the panel measures and what the toxicologist means.
| Element | Where it acts | Reference value behind the rows | What the panel reports |
|---|---|---|---|
| Lead | Developing nervous system; blood pressure | No tolerable intake. The FAO/WHO expert committee withdrew it in 2011 because no intake without effect could be identified; EFSA works with a benchmark dose and the principle of keeping exposure as low as achievable. | Total lead. Identical to the regulatory figure. |
| Cadmium | Kidney; bone. Biological half-life 10 to 30 years. | EFSA tolerable weekly intake 2.5 µg per kg body weight. | Total cadmium. Identical to the regulatory figure. |
| Mercury | Methylmercury: developing brain. Inorganic mercury: kidney. | Separate EFSA tolerable intakes for methylmercury and inorganic mercury. | Total mercury, which is the figure the rows for these products use. |
| Arsenic | Inorganic forms: carcinogen at chronic low intake (skin, bladder, lung). Organic forms in seafood: far less toxic. | EFSA reassessed dietary exposure to inorganic arsenic in 2024 using a margin-of-exposure approach. | Total arsenic, the upper bound on the inorganic fraction. Matters here for rice-based protein and for algae. |
Why a daily product is judged more strictly in practice. Cadmium accumulates. A protein shake or a greens powder taken every day delivers its cadmium every day for years, and the tolerable weekly intake is a lifetime figure. Two products at the same concentration are not the same exposure if one is taken daily and the other occasionally. The regulation applies one row to both; the risk assessment does not.
3What an exceedance costs the producer
Three sentences of the regulation set the cost.
Article 2(1). Food listed in Annex I "shall not be placed on the market and shall not be used as a raw material in food or as an ingredient in food where it contains a contaminant at a level which exceeds the maximum level".
Article 2(2). Food complying with the maximum levels "shall not be mixed with food which exceeds these maximum levels".
Article 4. Food containing the listed contaminants "shall not be deliberately detoxified by chemical treatments".
In plain words: a batch above its number cannot be sold, cannot be used as an ingredient, cannot be blended with a clean batch to bring the average under the line, and cannot be treated to remove the metal. If it has already been sold it is withdrawn or recalled and reported through the Rapid Alert System for Food and Feed, whose portal is public and is read by buyers as well as by authorities. The finding is per batch: one high lot does not condemn the product line, and the next lot is judged on its own result.
Products with no row of their own. Article 3 covers food that is dried, diluted, processed or compound and has no specific row. The maximum level of the nearest listed food is applied after accounting for concentration by drying or processing, for the proportions of the ingredients, and for the limit of quantification. Under official control the operator must provide and justify those factors. This is the article that governs a protein isolate and a flavoured blend, and it is why a certificate for the raw grain does not answer for the isolate made from it.
4Plant protein powders: pea, soy, rice
The metal
Cadmium. Arsenic as well when the base is rice.
Where it comes from
The plant takes the metal from the soil, and isolating the protein concentrates part of it with the protein. Rice is the exception among the three: it takes up arsenic from flooded paddies and stores most of it in the bran. Rice-bran powders and tablets have been measured at 284 to 767 µg/kg of inorganic arsenic, 1.4 to 3.3 times the limit for white rice, and the bran of a grain holds more arsenic than the kernel. Whey is not a plant and does not take metal from soil, which is why it is the comparator for this category rather than a member of it.
The number
No row exists for a protein isolate. Sold as a food supplement, the product takes the supplement rows: lead 3.0 mg/kg, cadmium 1.0 mg/kg, mercury 0.10 mg/kg, and no arsenic row. Sold as a food, Article 3 applies; for rice protein the reference is the rice-flour row of 0.25 mg/kg inorganic arsenic together with the concentration factor from grain to isolate, which the operator must be able to justify.
When to test
Per supplier lot of isolate. The certificate for the grain does not cover the isolate, because the concentration step is where the number changes.
5Chocolate-flavoured proteins and mass gainers
The metal
Cadmium, from the cocoa powder in the flavour system, not from the protein.
Where it comes from
The cacao tree takes cadmium from the soil, and beans from Latin American origins are the highest. In a 2026 survey of 103 commercial cocoa powders from twelve countries, cadmium ranged from below detection to 1.30 mg/kg and 15.5 % of samples were above the EU cocoa-powder limit. One unit of soil pH changes bean cadmium by a factor of 1.6, and about 30 % of Ecuador's cocoa area produces beans above 0.6 mg/kg. The number changes with origin and with lot.
The number
The cocoa rows are set by cocoa content, and the one that matters to a blender is the last.
| Row | Product | Cadmium, mg/kg |
|---|---|---|
| 3.2.15.1 | Milk chocolate with less than 30 % total dry cocoa solids | 0.10 |
| 3.2.15.2 | Chocolate with less than 50 % cocoa solids; milk chocolate with 30 % or more | 0.30 |
| 3.2.15.3 | Chocolate with 50 % or more cocoa solids | 0.80 |
| 3.2.15.4 | Cocoa powder for the final consumer, or as an ingredient in sweetened cocoa powder or drinking chocolate | 0.60 |
| 3.2.21.1 | Food supplements (the finished chocolate-flavoured product, if sold as a supplement) | 1.0 |
Under Article 2(1) a cocoa lot above 0.60 mg/kg cannot be used as an ingredient at all, whatever the finished product's status. The finished product then takes the supplement row of 1.0 mg/kg if it is sold as a supplement, or Article 3 if it is sold as a food.
When to test
The cocoa lot, before it goes into the blend. A new origin is a new lot.
6Greens and superfood powders: spirulina, chlorella, wheatgrass, barley grass
The metal
Lead, and arsenic for algae grown in open ponds.
Where it comes from
The water and the dust at the growing site. Spirulina is grown in open ponds, chlorella usually in closed systems, and the grasses in fields; the site changes with every supplier. A survey of thirteen spirulina and four chlorella products on the Polish market found lead from below 0.02 to 0.38 mg/kg, cadmium up to 0.078 mg/kg and mercury up to 0.017 mg/kg, all below the supplement rows. The spread between products was a factor of twenty for lead, which is the point: the category is safe on average and variable by supplier.
The number
The supplement rows: lead 3.0 mg/kg, cadmium 1.0 mg/kg, mercury 0.10 mg/kg. The 3.0 mg/kg cadmium row for seaweed-based supplements does not apply to spirulina or chlorella, which are not seaweed. There is no arsenic row; the panel's total-arsenic result is the figure to keep on file.
When to test
Per algae lot. A blend of several greens is tested as the blend, and the result is read against the supplement rows.
7Bee products: honey, pollen, propolis, royal jelly
The metal
Lead.
Where it comes from
The surroundings of the apiary. Bees forage up to 10 km from the hive and bring back what settles on flowers: lead from road traffic and from industrial dust. Honey from urban apiaries is higher than from rural ones; in one Italian survey the two elevated samples came from hives next to a busy road and an industrial site; honey from industrial areas of Romania has measured 0.19 to 0.20 mg/kg. Pollen and propolis concentrate more than honey: 0.27 mg/kg of lead in pollen from an urban Italian apiary, and 0.046 to 0.96 mg/kg in propolis from Turkey.
The number
Honey has its own row, and it is one of the strictest in the regulation: 0.10 mg/kg of lead. Pollen, propolis and royal jelly have no rows of their own; sold as supplements they take the supplement rows, lead 3.0 mg/kg, cadmium 1.0 mg/kg, mercury 0.10 mg/kg. The same hive can therefore produce a honey that fails and a pollen that passes with a wide margin, because the two are judged thirty-fold apart.
When to test
Per apiary site and per season for honey; per supplier for pollen and propolis, because a brand rarely knows where its supplier's hives stand.
8One table to keep
The four sections reduce to one table and one diagram. The table gives the element, the source, the row and the sampling point; the diagram shows where in each product's route the sample is drawn.
| Product | The metal | Where it comes from | EU row, mg/kg | Test this |
|---|---|---|---|---|
| Powders | ||||
| Plant protein isolate (pea, soy) | Cadmium | Soil, concentrated with the protein | Supplement: Cd 1.0, Pb 3.0, Hg 0.10. As food: Article 3. | Each supplier lot of isolate |
| Rice protein | Inorganic arsenic (and cadmium) | Flooded paddies; bran | Rice-flour row 0.25 iAs via Article 3; supplement rows if a supplement | Each supplier lot of isolate |
| Chocolate-flavoured protein, gainer | Cadmium | The cocoa powder | Cocoa powder 0.60 (ingredient); finished supplement 1.0 | Each cocoa lot, before blending |
| Greens powder (spirulina, chlorella, grasses) | Lead; arsenic for pond algae | Water and dust at the growing site | Supplement: Pb 3.0, Cd 1.0, Hg 0.10; no As row | Each algae lot; a blend as the blend |
| Bee products | ||||
| Honey | Lead | Road and industrial dust within 10 km of the hive | Honey 0.10 | Each apiary site, each season |
| Pollen, propolis, royal jelly | Lead | As honey, more concentrated | Supplement: Pb 3.0, Cd 1.0, Hg 0.10 | Each supplier |
Heavy metals testing from NorrChemica
We measure lead, cadmium, mercury and arsenic by ICP-MS in an EU laboratory accredited to ISO/IEC 17025. Tell us the product and how it is sold and we apply the row that fits; the report states the figure against that row. You send the lot sample, we arrange shipping and customs in both directions. Anyone can order, and no company is required.
Request testing → Part 1: the EU levels Part 2: lead in wine, beer and spirits Part 3: reading the report Part 5: tablets and capsules Part 6: hemp and derivatives Part 7: botanicalsSources
- Commission Regulation (EU) 2023/915 on maximum levels for certain contaminants in food, consolidated text of 8 October 2025: Articles 2, 3 and 4; Annex I rows 3.1.17 (honey, lead), 3.1.28 (food supplements, lead), 3.2.15.1 to 3.2.15.4 (cocoa and chocolate, cadmium), 3.2.21.1 and 3.2.21.2 (food supplements, cadmium), 3.3.2 (food supplements, mercury), 3.4.1.3 (rice flour, inorganic arsenic). EUR-Lex
- S. Y. Kim, D. Seo, M. Yoo & K. Lee, Analysis of arsenic species in processed rice bran products using HPLC-ICP-MS, Journal of Food Science 83 (2018) 2682. doi
- M. Cano-Lamadrid et al., Inorganic and total arsenic contents in rice and rice-based foods consumed by sportspeople, Journal of Food Science 81 (2016) T1031 — arsenic in bran against kernel. doi
- M. B. Silva et al., Mineral profile of cocoa powder: analytical approach in commercial samples, Journal of the Science of Food and Agriculture 106 (2026) 7056 — 103 commercial cocoa powders from twelve countries. doi
- D. Argüello et al., Surface soil liming reduces cadmium uptake in cacao seedlings, Journal of Environmental Quality 49 (2020) 1359 — bean cadmium against soil pH; survey means for Central and South American beans. doi
- J. L. Vázquez-deCastro et al., The economic costs of cadmium non-tariff measures for smallholder cocoa farmers in Ecuador, The World Economy 47 (2023) 1147 — share of Ecuador's cocoa area above 0.6 mg/kg. doi
- M. Ćwieląg-Drabek et al., Health risks from consumption of medicinal plant dietary supplements, Food Science & Nutrition 8 (2020) 3535 — spirulina and chlorella products on the Polish market, with per-sample values. doi
- H. R. El-Seedi et al., Bee products: safety measures and new technologies to secure their daily consumption, Journal of the Science of Food and Agriculture 106 (2026) 4550 — foraging range, urban against rural honey, the Italian road and industrial samples, propolis values. doi
- K. R. Hladun, N. Di, T. Liu & J. T. Trumble, Metal contaminant accumulation in the hive, Environmental Toxicology and Chemistry 35 (2016) 322 — honey from industrial areas of Romania; pollen from an urban area of Italy. doi
- A. Silva Chaves et al., Exposure-based toxicological evaluation of lead and elemental impurities in encapsulated green tea dietary supplements, Journal of Applied Toxicology (2026) — the JECFA 2011 withdrawal of the lead tolerable intake, the EFSA 2010 lead opinion, the EFSA cadmium tolerable weekly intake and the EFSA 2024 inorganic-arsenic reassessment, as cited in section 2. doi
About this article
The rows and articles of the regulation were read in the consolidated text on 2 September 2026. The measured values in sections 1 and 4 to 7 were read in the tables or text of the papers listed, not from abstracts. The EFSA and JECFA reference values in section 2 are cited as reported in source 10 and were not opened for this article. We found no peer-reviewed survey of plant protein isolates as sold; section 4 is written from the rice data and from the regulation, and will be revised when one is found.
This article explains what the rules say and what the published surveys have measured. It is not legal advice and it does not assess any particular product. We will date any revision.
© 2026 NorrChemica™ · Content may not be reproduced without written permission