Overview
Nickel behaves unlike lead, cadmium, or arsenic in the food supply. For those metals, an elevated concentration in a commodity is usually a signal of contamination somewhere in the supply chain: contaminated soil, an industrial point source, an adulterated spice, or a processing-side transfer. For dietary nickel the elevated concentrations sit predominantly in a specific and predictable set of plant staples (cocoa and chocolate, soybeans and other legumes, oats and whole-grain cereals, and tree nuts and peanuts), and they are present because those plants take up and concentrate nickel as an ordinary feature of their physiology, not because a pollution source has been introduced. Six independent datasets spanning four decades and four jurisdictions converge on this pattern: the European Union CONTAM Panel’s 44,585-sample occurrence and exposure assessment (efsa2015-nickel-food-water), a Danish literature-plus-primary compilation from 1984 (flyvholm1984), a Polish cereal-grain survey from 2020 (mania2020), a Chinese provincial total-diet study from 2024 (han2024), a Polish tree-nut survey from 2025 (cwielag-drabek2025), and an Italian muscle-meat survey from 2025 (manfredi2025). The five plant-focused datasets agree on which foods carry nickel, in what rank order, and at what magnitude; the meat dataset supplies the confirmatory contrast that animal muscle is, by comparison, a minor nickel source.
The non-obvious synthesis has two parts. First, the same food ranking recurs across geographies and across forty years of independent measurement, which is the signature of an intrinsic plant trait rather than of localized contamination. Second, in the European occurrence and exposure assessment, background dietary nickel intake reaches or exceeds the tolerable daily intake for younger age classes with no pollution source involved at all, which means that treating nickel as a contaminant to be sourced out of a supply chain is a category error: the metal is endogenous to the commodities that dominate exposure. This is the structural reason a simple maximum-level ceiling is a poor primary instrument for dietary nickel, which is a regulated food-contaminant metal. One important honesty caveat runs through the whole synthesis and is stated in its own section below: the operative tolerable daily intake for nickel was revised upward by a factor of roughly 4.6 between 2015 and 2020, and whether background diets exceed the reference point depends on which reference point is applied.
The European occurrence and exposure baseline (EFSA CONTAM Panel 2015)
efsa2015-nickel-food-water is the largest and most authoritative anchor. The EFSA Panel on Contaminants in the Food Chain integrated 44,585 admitted occurrence records (18,885 food and 25,700 drinking-water samples) submitted by 15 European Member States for 2003 to 2012, combined them with the EFSA Comprehensive European Food Consumption Database across 26 chronic surveys in 17 countries, and characterized dietary nickel exposure by age class. Analytical results were expressed on a whole-weight (as-analyzed) basis. The occurrence distribution is strongly food-specific. The highest-concentration subgroups were cocoa beans and cocoa products (solid) at a mean of 9,528 micrograms per kilogram (n=238), soybeans at 4,624 to 4,685 micrograms per kilogram, peanuts at 3,537 to 3,569, dried beans at 3,055 to 3,077, chocolate (cocoa) products at 3,231 to 3,236 (n=490), and dry tea and herbs for infusion at 761 to 762. The aggregate legumes, nuts and oilseeds group had a mean of 1,862 (lower bound) to 1,880 (upper bound) micrograms per kilogram with a 95th percentile of 7,000 (n=1,218). Grains and grain-based products had a mean of 271 to 321 micrograms per kilogram with a 95th percentile near 1,069 to 1,078 (n=4,291). At the opposite end, drinking water was the lowest group at a mean of 1 to 2 micrograms per kilogram (n=25,700, 89 percent left-censored), meat and meat products sat at 191 to 239, fish at 77 to 112, and milk and dairy at 71 to 93.
The exposure result is the load-bearing finding. Under the tolerable daily intake the Panel derived in this opinion, 2.8 micrograms of nickel per kilogram of body weight per day, mean chronic dietary exposure was at or above the tolerable daily intake for toddlers, other children, infants, and adolescents, and the 95th-percentile exposure was above the tolerable daily intake for every age class (mean estimates ranged from 2.0 to 13.1, and 95th-percentile estimates from 3.6 to 20.1 micrograms per kilogram of body weight per day across surveys and age classes). The dominant chronic contributors were grains and grain-based products, non-alcoholic beverages (cocoa-based drinks in the young and coffee in adults), sugar and confectionery (chocolate), the legumes-nuts-oilseeds group, and vegetables. In one young-population survey, cocoa-driven non-alcoholic beverages alone accounted for 41.8 percent of total nickel exposure. Drinking water, by contrast, contributed between 0.0005 and 1.7 percent of total dietary nickel. The Panel recorded that no European maximum levels for nickel in food existed and did not propose any, a point the regulatory-implications section returns to.
The Danish precedent (Flyvholm, Nielsen, and Andersen 1984)
flyvholm1984-nickel-content-food-dietary-intake establishes that the same food ranking was already visible four decades before the EFSA opinion, from an entirely separate body of measurement. Working at the Danish National Food Institute and the Danish Toxicology Centre, the authors compiled 2,221 food samples from the post-1969 literature (restricted to samples analyzed by atomic absorption spectrometry or PIXE for reliability) supplemented with their own analyses, and modeled the Danish average diet. The concentration rank order, reported in micrograms per gram (equivalent to milligrams per kilogram), is the same one the EFSA dataset would later produce: cocoa at a mean of 9.8 (n=7), soybeans at 5.2 (n=3), soy products at 5.1, walnuts at 3.6, peanuts at 2.8, oats at 2.3 (n=37), bitter chocolate at 1.9, hazelnuts at 1.8, dried legumes at 1.7, and almonds at 1.3. Within the modeled average diet, oatmeal was the standout per-gram contributor, with a load factor of 24 (the load factor being the ratio of a food’s share of total nickel intake to its share of total consumption; any value above 1 means the food contributes nickel disproportionately to its dietary weight).
The intake estimate anchors the exposure side. The Danish average diet delivered 150 micrograms of nickel per person per day from food (over 2,099 grams of food per day, exclusive of drinking water), with tap water adding roughly 16 micrograms per day under normal draw. Substituting high-nickel foods into the average diet raised modeled intake to 900 micrograms per day or more, which the authors placed against the 600 to 5,600 microgram single oral dose range that provoked hand-eczema flare in nickel-sensitized patients under clinical challenge. That a 1984 Danish compilation and a 2015 pan-European assessment, using different samples, different methods, and different populations, produce the same commodity ranking and the same conclusion that ordinary diets carry biologically meaningful nickel loads is the strongest available evidence that the pattern is an intrinsic property of the foods rather than an artifact of any one survey or era.
The Polish cereal-grain signal (Mania 2020)
mania2020-nickel-cereal-grains-poland provides a focused, recent, primary occurrence dataset for the cereal end of the pattern. The authors measured nickel by graphite-furnace atomic absorption spectrometry in 56 cereal-grain and cereal-based product samples from the Polish retail market in 2019 to 2020, reporting concentrations in milligrams per kilogram on an as-purchased product basis with middle-bound substitution for the 27 percent of results below the 0.06 milligrams per kilogram matrix limit of quantification. Whole and less-refined fractions carried the nickel: cereal grains (millet, rye, wheat, barley) had a mean of 1.16 and a 95th percentile of 3.91 milligrams per kilogram, with millet reaching 4.80 and oat flakes 2.53; bran had a mean of 1.34 and flakes 0.93; groats sat at 0.63, flour at 0.35, and pasta at 0.26. The processing gradient is explicit in the data: whole-grain flour ranged up to 2.12 milligrams per kilogram while refined wheat flour fell below the limit of quantification, and whole-grain pasta reached 1.79 while regular wheat-flour pasta ranged only 0.03 to 0.08. This is consistent with nickel being concentrated in the bran and germ fractions that refining removes, which is a milling-physics observation rather than a contamination observation. The authors assessed exposure against the revised 2020 tolerable daily intake of 13 micrograms per kilogram of body weight per day, and under that reference point cereal grains excluding rice contributed 13.4 percent of the tolerable daily intake for adults at mean contamination, while a child 95th-percentile cereal-product scenario reached 11.8 percent.
The Chinese total-diet signal (Han 2024)
han2024-nickel-zhejiang-china extends the pattern to an East Asian total-diet context and to a different regulatory frame. The authors measured nickel by ICP-MS in 2,628 food samples collected across 11 cities of Zhejiang Province in 2018 to 2019 and paired the concentrations (reported in milligrams per kilogram, fresh-food basis) with a consumption survey of roughly 19,000 individuals. Beans were the clear high-nickel outlier: the bean category (n=5) had a mean of 3.094 milligrams per kilogram, all five samples exceeded the Chinese national limit of 1 milligram per kilogram under GB 2762-2022, and a single soybean sample reached 9.21 milligrams per kilogram (the small sample size is a genuine limit on the strength of the bean inference and is noted as such in the source). The rest of the categories fell in a much lower band: meat and products (n=690) at a mean of 0.258, vegetables (n=365) at 0.185 (with legume vegetables the highest vegetable subgroup at 0.813), aquatic foods (n=1,405) at 0.261, cereal products (n=27) at 0.136, and fruits (n=136) at 0.097. The exposure model found that children aged 0 to 6 were the only population segment reaching an unacceptable cumulative target hazard quotient (1.078) under the high-concentration, high-consumption assumption (95th-percentile concentration multiplied by 95th-percentile consumption), at a modeled total nickel intake of 21.57 micrograms per kilogram of body weight per day. The authors noted that this figure exceeds both the 2015 EFSA tolerable daily intake (2.8) and the revised 2020 value (13), and that beans and vegetables were the dominant contributors. That a Chinese provincial diet, a Danish diet, and a European multi-country diet all place legumes and plant staples at the top of the nickel contribution list is the cross-cultural confirmation of the trait.
The Polish tree-nut signal (Ćwieląg-Drabek 2025)
cwielag-drabek2025-nuts-cd-pb-cr-ni-poland isolates the tree-nut and peanut arm of the pattern with recent market data. The authors measured cadmium, lead, chromium, and nickel by electrothermal atomic absorption spectrometry in 69 nut samples (16 peanuts, 15 hazelnuts, 15 almonds, 8 cashews, 15 walnuts) from the Polish market sourced from 13 countries, reporting concentrations in milligrams per kilogram on a wet-weight basis. Nickel was the most abundant of the four toxic elements measured in nuts and the primary driver of the modeled health risk. Mean nickel concentrations were 6.434 milligrams per kilogram in cashews, roughly 4 in peanuts, and 1.299 in almonds, with an all-nut mean of 3.21; the highest single value was 11.20 milligrams per kilogram in a peanut sample, the highest nickel value across the studies the authors compared. The nickel rank order was cashews above peanuts, then walnuts, hazelnuts, and almonds. The health-risk assessment found that at higher consumption scenarios (50 to 100 grams per day, modeled for children) the cumulative hazard index exceeded 1, driven primarily by nickel. This dataset is independent of Mania 2020 in authorship and institution despite the shared Polish market, and it confirms that nuts belong in the same high-nickel plant-staple tier as cocoa, soy, and oats.
The Italian meat contrast (Manfredi 2025)
manfredi2025-nickel-meat-italy is the confirmatory negative case that sharpens the whole synthesis. The authors measured nickel by ICP-MS in 809 official-control muscle-meat samples collected in Italy from 2011 to 2023, reporting concentrations in milligrams per kilogram. The means were low across all three matrices: poultry (n=156) at 0.016, beef (n=306) at 0.010, and pork (n=347) at 0.009 milligrams per kilogram, with maxima of 0.583, 0.606, and 0.158 respectively and minima below the limit of quantification. The exposure assessment found that meat consumption contributed less than 1 percent of the EFSA nickel tolerable daily intake across all age groups, with the single highest contribution being 0.86 percent from pork in toddlers. Animal muscle, in other words, is roughly two to three orders of magnitude below cocoa, soy, and cashews on nickel, which is exactly what the plant-accumulation hypothesis predicts: nickel enters the food supply through plant physiology, and tissues that do not accumulate it from a plant-uptake pathway stay low. The meat contrast is what distinguishes an intrinsic-plant-trait explanation from a generic “everything is a little contaminated” explanation.
The mechanistic explanation
The reproducibility of the food ranking across jurisdictions and decades is most parsimoniously explained by nickel being taken up from soil by plant roots and translocated into seeds, beans, and grain fractions through the plant’s ordinary micronutrient and metal-handling physiology; none of the six anchor datasets measures this uptake pathway directly, so the mechanism is an inference from the reproducible occurrence pattern rather than a measured finding of these sources. Certain taxa appear to concentrate nickel markedly: legumes (including soybeans and peanuts, which is a legume), cocoa, oats and whole-grain cereals, and tree nuts recur at the top of every dataset synthesized here. The concentration is highest in the metabolically active storage fractions, which is why the milling gradient in Mania 2020 is so clean: bran, germ, and whole-grain fractions carry the nickel and refined white flour does not, because refining physically removes the fractions where the plant deposited it. This is mechanistically parallel to the biological-accumulation logic documented for cadmium and arsenic in seaweed, where the elevated metal is a feature of the organism’s physiology rather than a quality-control failure, and to the multi-metal uptake documented in tea. The distinguishing evidence that this is uptake physiology and not soil-pollution geography is the reproducibility of the food ranking: a contamination explanation would predict that the high values track polluted regions and shift with sourcing, whereas the observed pattern is that the same foods rank high in Denmark, across the European Union, in Poland, and in China regardless of origin, and that a compositionally different tissue (animal muscle, Manfredi 2025) stays low even in the same food system. Nickel concentration in cocoa or cashews is therefore a property of what the commodity is, not of where a particular lot came from.
Why conventional contaminant surveillance misframes nickel
Conventional heavy-metal surveillance is built around the premise that an elevated result flags a problem to be traced and removed: find the contaminated field, the leaded spice, the industrial outfall, and the concentration falls. That premise is productive for lead, cadmium, and arsenic in many matrices, and it underlies the maximum-level ceiling as the default regulatory instrument. It misframes nickel in two ways. First, the highest-nickel foods are high because of what they are, so there is no upstream source to remove; a cocoa lot, a soybean lot, or a cashew lot that a supplier could realistically deliver will carry nickel in the multi-milligram-per-kilogram range as a baseline. Second, because nickel is distributed across many regularly consumed staples rather than concentrated in a few flagged commodities, aggregate exposure is driven by the ordinary diet, which is precisely why the EFSA assessment found background mean intake at or above the tolerable daily intake for younger age classes with no contamination event anywhere in the picture. A surveillance program that treats each high nickel result as a defect to be sourced out will chase values that cannot be removed without removing the food, and will still not address the aggregate-exposure problem, which lives in the diet as a whole rather than in any single lot. This is the structural reason a single across-the-board maximum-level ceiling is a poor instrument for dietary nickel; the EFSA opinion recorded that no European maximum levels for nickel in food existed and declined to propose any, and a ceiling set below the intrinsic content of cocoa or soy would function as a ban rather than a contamination control.
The unresolved tolerable-intake reference point
An honest account states plainly that the exceedance claim depends on which health-based reference point is applied, and that this reference point moved substantially. In the 2015 opinion, the EFSA CONTAM Panel derived a chronic tolerable daily intake of 2.8 micrograms of nickel per kilogram of body weight per day, from a benchmark dose lower confidence limit of 0.28 milligrams per kilogram of body weight per day for post-implantation fetal loss in rats, divided by an uncertainty factor of 100. Under that value, background European diets exceeded the reference point at the mean for young age classes and at the 95th percentile for all age classes, and the finding that ordinary diets already breach the tolerable intake is robust. The Panel subsequently revised the chronic tolerable daily intake upward to 13 micrograms per kilogram of body weight per day in its 2020 update, roughly 4.6 times higher, on reconsideration of the dose-response endpoint. Under the 2020 value, most background mean intakes fall below the reference point, and the datasets here reflect this: Mania 2020 finds cereal grains contributing only about 13 percent of the 13-microgram tolerable daily intake for adults, and Han 2024 finds an exceedance (21.57 micrograms per kilogram of body weight per day) only for children aged 0 to 6 under the joint high-concentration, high-consumption assumption. The honest synthesis is therefore that background diets clearly reach or exceed the 2015 reference point and reach it primarily for young children and high-consumption scenarios under the 2020 reference point. Two further caveats belong alongside this. The EFSA opinion also derived a separate acute reference point of 1.1 micrograms per kilogram of body weight for nickel-sensitized individuals (an estimated prevalence of up to 15 percent of the general population with allergic contact dermatitis), against which calculated margins of exposure were far below the low-concern threshold at ordinary intakes; that sensitized-subpopulation concern is not resolved by the upward revision of the chronic tolerable daily intake and connects to Single-metal regulatory limits systematically underprotect against dietary heavy-metal exposure. And several of the high-magnitude subgroup means rest on small samples (cocoa n=238 is robust, but Han’s bean category is n=5 and Flyvholm’s cocoa and soy figures rest on fewer than ten samples each), so the concentration ranking is far better supported than any single subgroup point estimate.
What the evidence implies for testing, regulation, and consumption
For testing, nickel screening should be targeted by commodity composition rather than applied as a uniform contamination screen. The foods that carry nickel are predictable (cocoa and chocolate, soy and other legumes, peanuts and tree nuts, oats and whole-grain and bran cereal fractions), and a product’s likely nickel load can be estimated from its ingredient list before any measurement, with whole-grain and less-refined fractions weighted higher than refined ones and animal-muscle ingredients weighted very low. Direct measurement remains the confirmation, but the prior is strong and stable across geographies.
For regulation, the evidence argues against a single across-the-board maximum-level ceiling as the primary instrument for dietary nickel. Because the high concentrations are intrinsic to the commodity, a ceiling set at a contamination-control level would fall below the natural content of cocoa, soy, or cashews and operate as a prohibition of those foods rather than as a contamination limit. The EFSA opinion itself recorded that no European maximum levels for nickel in food existed at the time and declined to propose any, which is consistent with the difficulty of ceiling-setting for an endogenous constituent. A nickel limit for cocoa, legume, or nut products is setting a level against an intrinsic content: a limit set below the commodity’s natural distribution would function as a prohibition of those foods rather than as a contamination-removal target, so it represents a policy choice about acceptable exposure, not the same kind of number as a lead or cadmium contamination limit. The same reasoning is why organic certification does not lower nickel content, for the reason developed in Organic certification is not protective against heavy-metal contamination, and in some food matrices is associated with higher loads: the metal enters through plant uptake, not through synthetic agricultural inputs.
For consumers, the actionable statements are specific and dose-anchored rather than framed as “avoid nickel.” Cocoa and dark or bitter chocolate, soybeans and soy products, peanuts and tree nuts (cashews in particular, at a mean near 6.4 milligrams per kilogram wet weight), and oats and whole-grain cereals are the dietary staples that contribute most nickel, at multi-milligram-per-kilogram concentrations, whereas meat, fish, dairy, fruit, and refined grains are minor contributors below a few tenths of a milligram per kilogram. For the general population these foods are not a reason for concern under the 2020 reference point at typical intakes, and they carry well-established nutritional value. The two populations for whom the load is more than academic are young children eating high-nickel staples at high frequency (the group for which Han 2024 modeled an exceedance) and the subset of individuals with diagnosed nickel allergy or systemic nickel allergy syndrome, for whom the relevant clinical literature describes low-nickel dietary management that targets exactly these staples; the single-dose oral-challenge range that provoked flare in sensitized patients in Flyvholm 1984 (600 to 5,600 micrograms) is reachable by substituting high-nickel foods into a day’s diet. A consumer-facing claim about nickel should always specify which of these populations it addresses, at what intake, and at what frequency, because a statement that is accurate for a nickel-allergic toddler eating oatmeal and cocoa daily is not accurate for a general-population adult.
Provisional status
This synthesis was established 2026-07-27 on six independent anchor sources: one pan-European A-tier occurrence and exposure assessment, one foundational Danish A-tier compilation, two Polish A-tier market surveys (cereals and nuts, from distinct author groups), one Chinese A-tier total-diet study, and one Italian A-tier meat survey supplying the low-nickel contrast. The concentration ranking and the intrinsic-plant-trait conclusion are robust across four decades and five jurisdictions (the EU-wide occurrence assessment plus Denmark, Poland, China, and Italy), with the Italian meat survey serving as the low-nickel contrast case. The two live uncertainties, carried explicitly in the body, are the roughly 4.6-fold upward revision of the chronic tolerable daily intake between the 2015 and 2020 EFSA opinions (which determines whether background diets are described as exceeding or merely approaching the reference point) and the small sample sizes behind several individual high-magnitude subgroup means. Resynthesis triggers per CLAUDE.md Part 9 fire when the corpus gains a source that either contradicts the commodity ranking or resolves the sensitized-subpopulation acute-exposure question, and if any future European maximum level for nickel is proposed, since the anchor opinion recorded that none existed and none was proposed, so its arrival would need testing against the intrinsic-content problem described here.
Peer review state
This synthesis claim has not yet been evaluated by external reviewers. Verdicts will be added here as named domain experts (listed at Curators and conflict-of-interest disclosure) complete their review. The verdict log is data/peer-review/<reviewer-slug>.jsonl and is part of the public corpus.
| Reviewer | Verdict | Review date | Notes |
|---|---|---|---|
| no reviews yet |
The Heavy Metal Index publishes synthesis claims as preprints — before external review completes — with the review state visibly tracked. Until at least one external verdict is recorded below, a synthesis page is a preprint, not a peer-reviewed work. External review accumulates over time, and the credibility of the claim is partly the cumulative result of that visible review.
Update history
The five most recent substantive edits to this page, classified major (evidence or structure moved), correction (a published value or statement was wrong and has been fixed), or minor (narrative rewritten without changing the underlying evidence). Each description is derived from what the edit did to this page; the linked commit is the authoritative record, routine regeneration passes are excluded, and 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.
| Commit | Date | Change | Description |
|---|---|---|---|
| 030d8b4 | 2026-07-27 | major | 12 sections added; narrative text revised |