One metal per product: mineral tablets, multivitamins, creatine and fish oil — which heavy metal is most likely to be over the limit?
Share
One metal per product: which heavy metal is most likely to be over the limit in mineral tablets, multivitamins, creatine and fish oil
In a tablet or capsule, the metal most likely to be over the limit is usually not in the active ingredient. It is in the mineral salt or the filler that makes up most of the tablet, and in a fish oil capsule it is in the crude oil before refining. This article names that element for mineral tablets, for multivitamins and pre-workouts, for amino acids and creatine, and for omega-3 capsules. For each one it gives the source of the metal, states the EU limit in plain words, puts that limit next to the number the ingredient supplier certifies against, and shows the point in the process where one test settles the question.
1One element per tablet
Part 4 looked at powders and bee products. For each product it named the one metal, out of the four on the panel, that is most likely to go over the limit, where that metal comes from, and the limit that Regulation (EU) 2023/915 sets for it. Tablets and capsules follow the same pattern, with one difference. A vitamin dose is a few milligrams or less, and a tablet weighs about a gram. The rest of the gram is filler, binder, glidant and lubricant. Article 2(3) applies the limits "to food as placed on the market", which means to the whole tablet. So the metal in a tablet usually comes from the filler and not from the active ingredient.
Part 4, section 2, explains what each of the four limits protects against, and section 3 explains what an exceedance costs the producer under Articles 2, 3 and 4: a batch above its limit cannot be sold, cannot be used as an ingredient, cannot be blended down and cannot be treated to remove the metal.
The figure below puts two numbers side by side for each of six ingredient-and-metal pairs. The first is the ceiling on the supplier's certificate: the maximum concentration of that metal that the food-additive specification in Regulation 231/2012 allows in the ingredient. That is the navy bar. The second is what real lots actually contained: the average measured in a survey of 190 excipient samples by the US FDA. That is the grey bar. Both are drawn as a fraction of the EU supplement limit, so that lead and cadmium can share one axis. The gold line at 1× is the limit itself, and anything to the right of it is above the limit. Section 2 gives the specifications and the survey in full.
2The number the supplier certifies against
Calcium carbonate, dicalcium phosphate, talc, magnesium stearate and most other mineral salts and fillers used in tablets are also approved food additives, each with an E-number. For each of them, Regulation (EU) 231/2012 sets purity criteria: the limit for lead, cadmium, arsenic and mercury in the material. The manufacturer tests each lot against those criteria, and that is what the certificate sent with the lot confirms. Your tablet, however, is a supplement, and Regulation 2023/915 judges it against the supplement limits, which are different numbers. The same calcium carbonate is therefore covered by two sets of limits, and the table puts them side by side.
| Additive | Arsenic | Lead | Cadmium | Mercury |
|---|---|---|---|---|
| Mineral sources, purity criteria in mg/kg | ||||
| E 170 calcium carbonate | 3 | 3 | 1 | none set |
| E 341(ii) dicalcium phosphate | 1 | 1 | 1 | 1 |
| E 341(iii) tricalcium phosphate | 1 | 1 | 1 | 1 |
| E 530 magnesium oxide | 3 | 2 | none set | none set |
| E 504(i) magnesium carbonate | 4 | 2 | none set | 1 |
| E 579 ferrous gluconate | 3 | 2 | 1 | 1 |
| E 585 ferrous lactate | 3 | 1 | 1 | 1 |
| Fillers, glidants and lubricants, purity criteria in mg/kg | ||||
| E 460(i) microcrystalline cellulose | 0.1 | 0.1 | 0.1 | 0.1 |
| E 551 silicon dioxide | 3 | 5 | none set | 1 |
| E 553b talc | 10 | 2 | none set | none set |
| E 470b magnesium stearate | 3 | 2 | 1 | 1 |
| The rows the finished supplement is judged against, mg/kg | ||||
| Food supplements, rows 3.1.28, 3.2.21.1 and 3.3.2 of Regulation 2023/915 | no row | 3.0 | 1.0 | 0.10 |
Three points stand out. A calcium carbonate that meets E 170 may contain up to 3 mg/kg of lead and 1 mg/kg of cadmium, and those are the supplement limits themselves. A lot that only just meets its specification is already at the limit before anything else goes into the tablet. Dicalcium and tricalcium phosphate are held to 1 mg/kg for all four elements, so their lead ceiling is three times below the supplement row. Among the fillers, talc has the loosest arsenic criterion, 10 mg/kg, and no criterion at all for cadmium or mercury. Silicon dioxide may contain lead up to 5 mg/kg, which is above the supplement row. Microcrystalline cellulose is held to 0.1 mg/kg for all four elements, thirty times below the lead row. Zinc salts used as nutrient sources are not food additives, so 231/2012 sets no criteria for them.
What lots actually contain is known from one survey. In 2015 the US FDA's Division of Pharmaceutical Analysis analysed 190 samples of 31 excipients by ICP-MS, supplied by the manufacturers through the excipient industry council. The averages, in mg/kg, were: calcium carbonate (26 samples) cadmium 0.77 ± 0.10, lead 0.18 ± 0.59, arsenic 0.04; calcium phosphates (13 samples) cadmium 0.03, lead 0.03, arsenic 0.24 ± 0.18; talc (11 samples) lead 0.19, arsenic 0.17; silicon dioxide (3 samples) lead 0.12. In the authors' own summary, materials made from mined raw materials contained lead, cadmium, arsenic and mercury in the range of 100 to 900 µg/kg. Synthetic and cellulose-based materials contained 1 to 10 µg/kg or less. Calcium carbonate was above the ICH Q3D oral concentration limit for cadmium.
What a certificate says. A certificate that reads "conforms to E 170" tells you that lead is not above 3 mg/kg and cadmium is not above 1 mg/kg. It does not tell you what the lot measured, and it allows the lot to be at 100 % of the supplement row. The useful thing to ask a supplier for is the measured value of each element on the lot, together with the laboratory's limit of quantification. On your side, the useful control is to measure the lot and read the result against the row.
3Mineral tablets: calcium, magnesium, zinc, iron
The metal
Which of the four metals is most likely to be over the limit in a mineral tablet depends on which mineral salt the tablet is made of. In a calcium carbonate tablet it is cadmium. In the survey of excipients described in section 2, calcium carbonate lots contained on average 0.77 mg/kg of cadmium, which is 77 % of the supplement limit of 1.0 mg/kg, while their lead averaged 0.18 mg/kg, only 6 % of the lead limit of 3.0 mg/kg. Of the two, cadmium is the one closer to its limit. Lead is still the metal most people associate with calcium supplements, because the products of the 1980s and 1990s made from bone meal, dolomite and oyster shell contained more of it, and because California's Proposition 65 sets a very low daily lead level that US brands test against. For a tablet made from calcium phosphate the metal is arsenic, because phosphate rock contains arsenic. The EU sets no arsenic limit for supplements, so arsenic will not fail the batch, but it is the highest figure on the panel for this material and the one a buyer will ask about. For magnesium, iron and zinc salts there is no survey with enough samples to name one metal, so the panel result is the answer.
Where it comes from
The metal comes with the ore. The specification defines calcium carbonate as "the product obtained from ground limestone or by the precipitation of calcium ions with carbonate ions". Limestone is a marine deposit, and the cadmium and lead in it are geological. In the FDA survey the 26 calcium carbonate samples averaged 0.77 mg/kg of cadmium with a standard deviation of only 0.10, so this figure describes the material rather than one lot. Lead averaged 0.18 mg/kg with a standard deviation of 0.59, which means most lots were low and a few were high. Two US surveys of calcium supplements as sold, 70 brands in 1993 and 136 brands in 2000, put the sources in the same order: dolomite, bonemeal and natural-source "oyster shell" carbonate had the most lead, refined carbonate and chelates the least.
Phosphate rock contains arsenic, and the phosphate salts made from it do too. In the FDA survey the calcium phosphates averaged 0.24 mg/kg of arsenic, against 0.03 mg/kg each for lead and cadmium. For magnesium and iron the survey had one sample of each material: a magnesium carbonate at 0.43 mg/kg of cadmium, a magnesium hydroxide at 0.84 mg/kg, and a carbonyl iron with lead below detection. One sample says nothing about the range. For zinc salts we found no survey we could read.
The number
Sold as a supplement, the tablet is judged against the supplement rows: lead 3.0 mg/kg, cadmium 1.0 mg/kg, mercury 0.10 mg/kg, and no row for arsenic. Article 2(3) applies them to the tablet as sold. The arithmetic for a calcium carbonate tablet is short. Take a 1,450 mg tablet with 1,250 mg of calcium carbonate in it, which is 500 mg of calcium. At the survey average of 0.77 mg/kg of cadmium in the carbonate, the tablet contains 0.66 mg/kg from the carbonate alone, two-thirds of the row. At the survey average for lead, it contains 0.16 mg/kg, five per cent of the lead row. On this evidence, cadmium rather than lead is the element closest to its limit in a calcium carbonate tablet sold in the EU.
For the same tablet sold in California, lead is the element closest to its limit. Proposition 65 sets a maximum allowable dose level for lead of 0.5 µg per day; above that daily exposure a warning is required. A 1,250 mg dose of carbonate at the survey average gives 0.23 µg of lead per day. The same dose at the E 170 ceiling of 3 mg/kg gives 3.75 µg per day, seven and a half times the level. A certificate to E 170 does not answer for a product sold in the United States.
When to test
Each lot of mineral salt on arrival, before blending, with the result read against the supplement rows rather than against the E-number. Part 3 explains how to read a result against a row once the measurement uncertainty is taken into account.
4Multivitamins, pre-workouts and the excipient question
The metal
In a multivitamin or a pre-workout tablet, the metal most likely to be over the limit is the one that comes with the filler, because the filler is most of the tablet. With a calcium carbonate filler that metal is cadmium. With talc it is arsenic and lead, though the measured averages for talc are low. With dicalcium phosphate it is arsenic, and arsenic has no supplement limit to fail. With cellulose as the filler, none of the four is close to a limit.
Where it comes from
The label lists the active ingredients, but most of the mass is excipient. In a multivitamin tablet the vitamins are dosed in micrograms and milligrams and the minerals in tens of milligrams. The rest of the gram is dicalcium phosphate, calcium carbonate or cellulose as filler, silica or talc as glidant, magnesium stearate as lubricant, and a coating. A pre-workout, electrolyte or fat-burner tablet is built the same way. So the excipient averages in section 2 are the tablet's starting point. A tablet filled with calcium carbonate starts at 0.77 mg/kg of cadmium multiplied by the carbonate's share of the mass. A tablet filled with cellulose starts near zero.
The one survey of finished sport products we could read measured mercury. Fifty supplements were bought on the Polish market in 2025 and 2026: pre-workouts with caffeine, protein, energy gels, non-stimulant ergogenics, amino acids, electrolytes, collagen and fat-burners. The median was 9.0 µg/kg of total mercury and the range 0.4 to 51.7 µg/kg. Every product was below the 100 µg/kg row, and the highest was at 52 % of it. Electrolyte and amino-acid products had the highest category medians, 29 and 22 µg/kg, and protein and gels the lowest, 7 µg/kg; the differences between categories were not statistically significant. A separate study of 22 protein powders on the Hungarian market, whey, vegan and beef, found arsenic, cadmium, mercury and lead below the ICP-MS detection limit in every product. Plant protein isolates, with their cadmium and arsenic, are covered in Part 4.
The number
The supplement rows, applied to the tablet. A brand that also sells outside the EU has two more numbers to meet, and both are daily doses rather than concentrations. USP ⟨2232⟩ sets permitted daily exposures for supplements of 5 µg of lead, 5 µg of cadmium and 15 µg of inorganic arsenic. Proposition 65 sets 0.5 µg of lead per day. The conversion is the same in each case: the concentration in mg/kg multiplied by the grams taken per day gives the micrograms per day. The Polish survey shows why this matters. Three gel products with maximum daily portions of 300 g were under the mercury row on concentration, yet on dose they were above a screening hazard quotient of 1, under the bounding assumption that all of their mercury was methylmercury.
Concentration and dose are two different tests. A row in the regulation is a concentration. A health-based limit is a dose. A product at 30 µg/kg of mercury is at 30 % of the row whether the serving is 1 g or 300 g, but the exposure differs by a factor of 300. For a tablet the serving is a gram or two and the row is the stricter test. For a gel, a drink powder or anything taken by the hundred grams, the dose is the stricter test.
When to test
Each excipient lot from a new supplier, and the finished product once for each formulation. The filler sets the metal content, so a change of filler counts as a new product here.
5Amino acids and creatine
The metal
For creatine, branched-chain amino acids and glutamine we cannot name one metal, because we found no published survey of these products as sold that we could read in full. This section therefore explains how these products are made, which is where any metal would come from, and states the limits that apply. The one measured figure we have is for mercury: the four amino-acid products in the Polish survey of sport supplements contained 0.5 to 44 µg/kg, median 22, all under the limit of 100 µg/kg.
Where it comes from
Creatine monohydrate is synthetic: it is made from sarcosine and cyanamide and recovered by crystallisation. The branched-chain amino acids and glutamine are fermentation products, recovered from the broth and crystallised. Neither route starts from a mineral or from a plant grown in soil, so there is no ore and no soil uptake. What can get in is what the process adds: the metal content of the process water, of the acids and bases used in recovery, of the activated carbon and filter aids, and of the equipment. Crystallisation leaves most of that behind in the mother liquor. The figure a panel reports on a creatine or amino-acid lot is therefore a measure of the manufacturer's process. It changes when the manufacturer changes, and not with the season.
The number
The supplement rows, applied to the capsule or the powder as sold. A blend with flavour, sweetener and anticaking agent is judged as the blend.
When to test
Once per manufacturer and per bulk lot. A change of manufacturer is what changes the number.
6Fish oil and omega-3 capsules
The metal
In a fish oil capsule the metal a buyer asks about is mercury, because mercury is the metal associated with fish. The metal the laboratory is most likely to report at a high figure is arsenic, and arsenic has no supplement limit, so the number needs an explanation rather than a pass or fail.
Where it comes from
Refined marine oil contains very little mercury. A survey of 36 oil supplements on the Polish market, analysed with a direct mercury analyser, found 0.030 to 0.207 µg/kg in twelve cod-liver oils, 0.023 to 0.087 µg/kg in six shark-liver oils, and 0.131 to 0.427 µg/kg in eighteen vegetable oils, so the vegetable oils contained more than the fish oils. In 24 fish oil supplements on the Korean market, cadmium, lead and arsenic were below 0.05 mg/kg and total mercury was not detected in any product. Against the 100 µg/kg row, refined marine oil is about a thousand times below.
Arsenic is different. Marine oils contain arsenic in fat-soluble forms. A crude cod-liver oil contained 5 µg/g of arsenic, part of it as arsenic-containing fatty acids. Six fish oils sampled for the Norwegian feed monitoring programme contained 9.1 to 13.5 mg/kg of total arsenic, and in the feeds made from them inorganic arsenic was under 1.2 % of the total. In marine foods generally, arsenolipids make up 10 to 50 % of the arsenic. The arsenic-containing hydrocarbons among them were as toxic to cultured human cells as arsenite, and EFSA asked in 2009 for a risk assessment, but there is still no row. The refined supplement oils in the Korean survey measured below 0.05 mg/kg. The two sets of figures describe two different materials: crude and feed-grade oil on one side, refined supplement oil on the other.
The number
The regulation has no fish-oil row for any metal. The capsule, sold as a supplement, is judged against the supplement rows: lead 3.0 mg/kg, cadmium 1.0 mg/kg, mercury 0.10 mg/kg. The bulk oil, sold as an ingredient, comes under the general row for fats and oils, 0.10 mg/kg of lead (row 3.1.18), and under Article 2(1) it cannot be used as an ingredient above that. Neither has an arsenic row. Our panel reports total arsenic, not the inorganic fraction. On a marine oil that figure is mostly organic arsenic, and it is the figure to keep on file, with its explanation, before a buyer asks for it.
When to test
The refined oil lot, before encapsulation. A new species, a new refiner, or an oil sold as "virgin" or "unrefined" is a new question, and for that oil the arsenic figure is the first one to check.
7One table to keep
The four sections reduce to one table and one diagram. The table gives the element, its source, the row and the sampling point. The diagram shows where in each product's route the sample is taken.
| Product | The metal | Where it comes from | EU row, mg/kg | Test this |
|---|---|---|---|---|
| Minerals | ||||
| Calcium carbonate tablet | Cadmium in the EU; lead for US sales | Marine limestone; survey mean 0.77 Cd, 0.18 Pb | Supplement: Cd 1.0, Pb 3.0, Hg 0.10. E 170 permits Pb 3, Cd 1. | Each carbonate lot, against the row |
| Calcium phosphate tablet, phosphate filler | Arsenic, no row | Phosphate rock; survey mean 0.24 As | Supplement rows. E 341 permits 1 mg/kg for all four. | Each lot; keep the arsenic figure on file |
| Magnesium, iron, zinc salts | The panel; no survey range | The ore or the precipitation route | Supplement rows. E 530, E 504, E 579, E 585 criteria; none for zinc. | Each salt lot from a new supplier |
| Vitamins and sport formulas | ||||
| Multivitamin, pre-workout, electrolyte tablet | The filler's metal | Filler mass share × survey mean | Supplement rows on the tablet. USP ⟨2232⟩ and Proposition 65 by daily dose. | Each excipient lot; product per formulation |
| Amino acids | ||||
| Creatine, BCAA, glutamine | The panel; no survey read | Process water, reagents, equipment | Supplement rows; a blend as the blend | Per manufacturer, per bulk lot |
| Oils | ||||
| Fish oil, omega-3 capsule | Mercury has the row; total arsenic has the figure | Crude marine oil; refined oil measures low | Supplement rows on the capsule; bulk oil Pb 0.10 (3.1.18); no As row | Refined oil lot before encapsulation |
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 4: powders and bee products 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 3.1.18 (fats and oils, lead), 3.1.28 (food supplements, lead), 3.2.21.1 and 3.2.21.2 (food supplements, cadmium), 3.3.2 (food supplements, mercury); section 3.4 (arsenic), which contains no row for food supplements. EUR-Lex
- Commission Regulation (EU) No 231/2012 laying down specifications for food additives, consolidated text of 18 August 2026: purity criteria for E 170, E 341(ii), E 341(iii), E 460(i), E 470b, E 504(i), E 530, E 551, E 553b, E 579 and E 585. EUR-Lex
- G. Li, D. Schoneker, K. L. Ulman, J. J. Sturm, L. M. Thackery & J. F. Kauffman, Elemental impurities in pharmaceutical excipients, Journal of Pharmaceutical Sciences 104 (2015) 4197 — 190 samples of 31 excipients by ICP-MS; the means in sections 2 to 4 and Figure 1. doi
- A. Puścion-Jakubik, D. Jabłonka, D. Wojciuk, R. Markiewicz-Żukowska & K. Socha, Mercury content and screening exposure assessment in sports food supplements, Toxics 14 (2026) 680 — 50 sport supplements on the Polish market, by category. doi
- I. L. Horváth, G. Kajner, G. Galbács & D. Csupor, Analysis of heavy metal content in protein powders available on the Hungarian market, Journal of Nutritional Science 14 (2025) e49 — 22 protein powders by ICP-MS. doi
- B. Brodziak-Dopierała, A. Fischer, M. Chrzanowska & B. Ahnert, Mercury exposure from the consumption of dietary supplements containing vegetable, cod liver, and shark liver oils, International Journal of Environmental Research and Public Health 20 (2023) 2129 — 36 oil supplements on the Polish market. doi
- J. Lee, M. K. Kim, B. Kim, J. Kim & K. Lee, Analysis of polychlorinated biphenyls, heavy metals and omega-3 fatty acids in commercially available Korean functional fish oil supplements, International Journal of Food Science & Technology 51 (2016) 2217 — 24 fish oil supplements. doi
- A. Rumpler et al., Arsenic-containing long-chain fatty acids in cod-liver oil, Angewandte Chemie International Edition 47 (2008) 2665 — crude cod-liver oil at 5 µg As/g. doi
- J. J. Sloth, K. Julshamn & A.-K. Lundebye, Total arsenic and inorganic arsenic content in Norwegian fish feed products, Aquaculture Nutrition 11 (2005) 61 — six fish oils, Table 4; inorganic arsenic as a share of total. doi
- S. Meyer et al., Arsenic-containing hydrocarbons and arsenic-containing fatty acids: transfer across and presystemic metabolism in the Caco-2 intestinal barrier model, Molecular Nutrition & Food Research 59 (2015) 2044 — share of arsenolipids in marine food, cytotoxicity, the EFSA 2009 request. 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 USP ⟨2232⟩ permitted daily exposures as reported in its Table 4. doi
- California Office of Environmental Health Hazard Assessment, Proposition 65, Lead and lead compounds: maximum allowable dose level 0.5 µg/day; oral no significant risk level 15 µg/day. OEHHA
- B. P. Bourgoin, D. R. Evans, J. R. Cornett, S. M. Lingard & A. J. Quattrone, Lead content in 70 brands of dietary calcium supplements, American Journal of Public Health 83 (1993) 1155 — abstract only, see About. doi
- G. M. Scelfo & A. R. Flegal, Lead in calcium supplements, Environmental Health Perspectives 108 (2000) 309 — abstract only, see About. doi
About this article
We read the rows, articles and purity criteria of both regulations in the consolidated texts on 2 September 2026. The measured values in sections 2, 3, 4 and 6 and in Figure 1 come from the tables or text of the papers listed, not from abstracts, with two exceptions. The order of calcium sources in section 3 comes from the abstracts of sources 13 and 14; their full texts are scanned pages we did not read, and no figure from either appears in the body. The USP ⟨2232⟩ limits in section 4 are cited as reported in source 11; we did not open the USP text. The survey in source 3 analysed pharmaceutical-grade excipients, which are the grades used in supplement tablets, and where it had one sample of a material we say so. The production routes in section 5 are stated from general process chemistry. We found no peer-reviewed survey of creatine or amino-acid supplements as sold that we could read in full, and we will revise that section when one is found. Our panel reports total arsenic only.
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