There Is Lead in Protein Powder. The Fight Is Over the Yardstick

Every couple of years a testing report lands, a screenshot of a ranked table circulates, and comment sections fill with people asking whether to bin a tub they have been working through since January. The claim underneath the alarm — that lead turns up in protein powder — is true, and two large testing rounds on the category landed within nine months of each other. What tends to get concluded from them is a good deal shakier than the measurements themselves.

What the Testing Actually Found

The numbers behind the January 2025 wave of coverage come from the Clean Label Project, a US non-profit that commissions contract laboratories to test consumer products and publishes the results. For its protein-powder report, dated January 2025, it says it purchased and tested 160 products from 70 of the top-selling brands, sourced from Nielsen and Amazon best-seller lists and supplemented with products from the natural and organic marketplace, and contracted the laboratory Ellipse Analytics to measure arsenic, cadmium, lead and mercury by inductively coupled plasma mass spectrometry. The headline results: 47 per cent of products “exceeded at least one federal or state regulatory set for safety, including CA Prop 65”; 21 per cent of samples were over twice Proposition 65 levels; and 77 per cent of plant-based powders exceeded the Proposition 65 threshold for lead, against 28 per cent of whey powders and 26 per cent of collagen ones Clean Label Project 2025. One wrinkle is worth knowing before anyone quotes the 47 per cent: the report's own executive summary describes the same figure more narrowly, as products exceeding Proposition 65 thresholds for toxic metals, so two different wordings of that statistic sit on different pages of one document.

Consumer Reports ran its own round nine months later, testing 23 protein powders and ready-to-drink shakes across dairy, beef and plant sources. It reported that for more than two-thirds of the products, a single serving contained more lead than its own food-safety experts say is safe to consume in a day — a threshold it set at 0.5 micrograms a day, taken from the Proposition 65 maximum allowable dose level. Lead in the plant-based products averaged nine times the level found in those made with dairy proteins. The two highest results were a vegan mass gainer at 7.7 micrograms of lead per serving and a second product at 6.3 micrograms; only one product in the set showed no detectable lead. Consumer Reports' own advice was blunt: “We advise against daily use for most protein powders, since many have high levels of heavy metals and none are necessary to hit your protein goals” Consumer Reports 2025.

The earlier round it compared against ran in 2010 and covered 15 protein drinks bought in the New York metro area, every one of which contained at least one of arsenic, cadmium, lead or mercury in at least one sample; three servings a day of one chocolate shake supplied an average of 16.9 micrograms of arsenic against a limit of 15 micrograms a day then proposed by the US Pharmacopeia Consumer Reports 2010. Consumer Reports says its worst 2025 product carried nearly twice as much lead per serving as the worst product in that 2010 round Consumer Reports 2025. It publishes no microgram figure for the 2010 product, and the two rounds were scored against different yardsticks — the Pharmacopeia's proposed supplement limits then, California's labelling level now — so the comparison is about products rather than about thresholds.

Two things are worth holding onto about this genre of report. The testing itself is not in dispute: these are real laboratory measurements on real products. But they are not peer-reviewed studies, the shopping baskets are best-seller and convenience selections rather than probability samples, and the Clean Label Project's whitepaper publishes no per-product results, no limits of detection and no methodology annex. Industry associations pushed back on exactly that. The Consumer Healthcare Products Association said the report “has not been peer-reviewed nor published in a scientific journal, and the full testing methods have not been provided, so the validity of any conclusions should be considered accordingly”; the Council for Responsible Nutrition urged the group to publish its findings in peer-reviewed journals; and a request from the Natural Products Association to disclose the methodology and the underlying data was declined NutraIngredients 2025. Those are trade bodies speaking for the brands being tested, so read them accordingly — but the auditability complaint stands on its own. A reader who wants to check how those percentages were produced cannot.

Detectable Is Not the Same as Dangerous

Lead is a naturally occurring element and a legacy pollutant, largely from decades of leaded petrol and lead-based paint, and a 2003 review in the British Medical Bulletin notes that the general population is exposed to it from air and food in roughly equal proportions Järup 2003. That is a twenty-three-year-old narrative review rather than new evidence, but it makes the relevant point: lead in food is a background condition, not a property of one product category.

Scale is what turns a detection into a judgement, and the Consumer Reports article lays out three different numbers for readers to choose between. Its own level of concern is 0.5 micrograms a day, from California's labelling threshold. The US Food and Drug Administration's interim reference levels are 2.2 micrograms a day for children and 8.8 micrograms a day for females of childbearing age — and an FDA spokesperson told Consumer Reports there is sufficient evidence that the 8.8 microgram benchmark should be applied to all adults, which is a statement to a journalist rather than a published adult reference level. Meanwhile the average American adult is exposed to up to 5.3 micrograms of lead a day through diet alone, according to a 2019 analysis by FDA scientists that the report cites Consumer Reports 2025.

Those three figures do different jobs. One triggers a warning label in a single US state. One is a regulator's interim benchmark, set for a specific population. One is a background estimate for the whole diet, and it is an upper bound rather than a central value. Consumer Reports drew its own conclusions from the set: more than two-thirds of products exceeded its 0.5 microgram level of concern; someone taking a daily serving of either of the two worst products is “likely exceeding the FDA's interim reference level for dietary lead”; and it advised against daily use for most powders, suggesting roughly weekly use even for several mid-range products Consumer Reports 2025. Anyone reporting the round as “only the two outliers matter” has substituted their own arithmetic for what the report actually says.

What the Risk Assessments Estimate

Testing reports measure concentrations. Risk assessments try to answer a different question: whether a modelled daily intake crosses a health-based guidance value. Several now exist for these products, and they repay reading in detail rather than in summary.

The first is a 2020 paper in Toxicology Reports. It is explicitly a screening-level assessment rather than laboratory work: the authors re-used published concentration data from Consumer Reports' 2010 round of 15 products and the Clean Label Project's 2018 study of 133 products — not the 2025 rounds this article opened with — and modelled a 70 kilogram adult consuming one or three servings a day, every day of the year. For arsenic, cadmium and mercury the summed hazard index stayed below one for all 15 Consumer Reports products and for the median, mean and 95th-percentile scenarios drawn from the Clean Label Project data — but one scenario crossed it. Three servings a day of the single product with the highest reported cadmium concentration, 13.18 micrograms a serving, “resulted in an HI > 1”, at a maximum hazard index of 1.757; the authors set that case aside as a product with “an exceedingly elevated Cd concentration” which “may not be representative”. Their stated conclusions were that heavy metal exposure from these products “does not pose an increased non-carcinogenic risk to human health” and that “no carcinogenic risk was expected from As via ingestion of protein powder supplements”, with arsenic intakes spanning 0.2 to 16.9 micrograms a day. Lead was handled separately, because there is no reference dose for it: the authors modelled blood lead instead, against a baseline adult level of 1.27 micrograms per decilitre from national survey data, and reported that no product under any scenario pushed modelled blood lead past the 5 microgram per decilitre guidance value they used, with background exposure the primary contributor. The highest hazard indices, approaching one, were in mass-gain powders; the lowest were in whey Bandara 2020.

Two caveats belong with that paper. Lead — the metal this whole debate is about — was never compared against an intake guidance value in it, because none exists; the reassurance for lead rests on a modelled blood level. And all three authors worked for the commercial risk-science consultancy Cardno ChemRisk, though they declared no competing financial interests. Their own recommendation was methodological: that the health risks of heavy metals in these products “should be conducted within the context of relevant background exposures and established health based standards instead of the presence of hazardous substances alone” Bandara 2020.

It is not the only assessment, and the independent ones are less soothing. Academic researchers at Munster Technological University, funded by Irish research councils, developed and validated a mass-spectrometry method and applied it to 36 protein powders sourced in Ireland, measuring 21 elements and running a consumer risk assessment. Their conclusion: “While the concentrations of potentially toxic elements were found to be at acceptable levels, the results suggest that excessive and prolonged use of some of these products may place consumers at a slightly elevated risk for developing cancer or other negative health impacts throughout their lifetimes” Ring 2021. That is the prolonged-daily-use scenario most readers of this article are actually in. A further assessment, published in the Journal of Food Composition and Analysis in 2025 by academic groups in Spain and Brazil, quantified lead and cadmium in protein supplements on the European market and carried out its own dietary exposure and health-risk assessment Bethencourt-Barbuzano 2025.

What all of these share is a limitation the headlines rarely carry: they model exposure against guidance values. None of them observed health outcomes in people who use protein powder, and none can tell you what one scoop a day did to anybody.

Where the Lead Actually Comes From

Nobody is adding lead to protein powder. The Clean Label Project's account is that “the two primary sources of contaminants in protein powders are the contaminated soils where ingredients are grown and the packaging used for these products” Clean Label Project 2025. On the packaging side its data show one clear improvement: bisphenol A or S was detected in only 3 of the 160 products, compared with 55 per cent in the group's previous testing Clean Label Project 2025.

The soil pathway is the one both testing programmes agree on. The Clean Label Project put 77 per cent of plant-based products over the Proposition 65 lead threshold against 28 per cent of whey products, and reported that organic products averaged three times more lead and twice the cadmium of non-organic ones, with 79 per cent of organic products over the threshold for lead Clean Label Project 2025. Consumer Reports found plant-based powders averaging nine times the lead of dairy-based ones Consumer Reports 2025. The 2020 risk assessment offers no independent check on that split, because it had no plant category: it classified its products only as weight-gainer or whey. It did report that whey powders had the lowest calculated hazard indices, and it speculated that plant-based powders, or ones intended for vegan consumers, “may contain a higher As and Cd content” — a remark about arsenic and cadmium rather than lead, and one for which it cited the Clean Label Project's own testing rather than any measurement of its own Bandara 2020. So this is two testing programmes pointing the same way, not three independent laboratories.

Nor does the pattern hold everywhere. The 2025 European assessment, which measured lead and cadmium in 90 animal- and plant-based protein supplements sold on the European market, found the split running the other way for lead: higher lead concentrations in the animal-based products, with cadmium the metal that was elevated in the plant-based ones Bethencourt-Barbuzano 2025. So the plant-versus-whey lead gap is a real feature of both 2025 North American shopping baskets rather than a settled property of plant protein, and it is worth holding loosely.

Flavour is murkier, and it is where the two 2025 reports part company. Cocoa is a documented carrier of metals: FDA scientists analysing a convenience sample of cocoa powder and chocolate products bought at US retail found cadmium ranging from 0.004 to 3.15 milligrams per kilogram and lead ranging from below the limit of detection to 0.38 milligrams per kilogram, with both metals significantly correlated with percentage cocoa and cadmium varying by growing region — higher in products reported as originating in Latin America than in Africa. Their own concluding point was about cadmium and international standards for chocolate, and no protein powder was tested Abt 2018. The Clean Label Project reported a flavour gap in its own protein data: 65 per cent of chocolate powders over Proposition 65 levels and 29 per cent over twice those levels, chocolate powders containing four times more lead than vanilla ones, and — its most dramatic statistic, and a cadmium one rather than a lead one — 110 times more cadmium than vanilla varieties Clean Label Project 2025. Consumer Reports, testing a much smaller set, found average lead concentrations of 17.3 parts per billion in chocolate products against 15.4 in vanilla ones, and called that no meaningful difference Consumer Reports 2025.

So the cocoa mechanism is documented in cocoa, and how much of it survives into a finished tub is contested. That matters for how you read the plant-versus-whey gap too. The whey-versus-plant argument is usually fought over amino acids; contamination is a different axis, and on the evidence above it behaves like a tendency across a shelf driven by where ingredients were grown, rather than a verdict on any two tubs.

Why “No Safe Level” Still Matters

This is where the reassurance has to stop short. In 2010 the European Food Safety Authority's contaminants panel reviewed lead in food and concluded that the provisional tolerable weekly intake of 25 micrograms per kilogram of bodyweight then in use — a figure derived by the Joint FAO/WHO Expert Committee on Food Additives and endorsed by the former Scientific Committee on Food, rather than one EFSA itself had set — “is no longer appropriate as there is no evidence for a threshold for critical lead-induced effects”. It did not issue a replacement tolerable intake; it moved to comparing exposures against reference points. The panel identified developmental neurotoxicity in young children and cardiovascular effects and nephrotoxicity in adults as the critical endpoints, and derived a reference point for developmental neurotoxicity of 12 micrograms per litre of blood lead, corresponding to a dietary intake of 0.50 micrograms per kilogram of bodyweight per day — a per-kilogram figure that is a different quantity from California's 0.5 micrograms per day in total, despite the coincidence of digits. Average adult dietary lead exposure in Europe, it reported, ranged from 0.36 to 1.24 micrograms per kilogram of bodyweight per day, up to 2.43 in high consumers, leaving margins of exposure at which the possibility of an effect in some consumers, particularly children aged one to seven, could not be excluded EFSA 2010.

That position rests on studies at exposures once considered unremarkable. A pooled analysis of individual data from seven prospective cohorts — 1,333 children followed from birth or infancy to ages five to ten — found a 6.9-point IQ decrement across an increase in concurrent blood lead from 2.4 to 30 micrograms per decilitre, with 3.9 of those points accruing between 2.4 and 10. The steepest declines were below the old 10 microgram action level, and the authors reported deficits in children whose maximum blood lead stayed under 7.5 micrograms per decilitre, with no evidence of a threshold Lanphear 2005. In adults, an observational cohort of 14,289 US survey participants whose blood lead was measured between 1988 and 1994 and who were followed for a median of 19.3 years found that an increase from 1.0 to 6.7 micrograms per decilitre was associated with all-cause mortality (hazard ratio 1.37), cardiovascular disease mortality (1.70) and ischaemic heart disease mortality (2.08) Lanphear 2018.

Both measure blood lead in people exposed mainly through their environment — old paint, old pipes, contaminated soil, historical air pollution — during a period of much higher population exposure than today's, and the adult study is observational, so it reports association rather than cause. Neither studied protein powder, and by design neither can be converted into a per-serving figure. What they establish is narrower and still important: below a regulatory limit is not a synonym for no effect. EFSA's panel put the practical implication plainly, recommending that protecting children against neurodevelopmental effects “would be protective for all other adverse effects of lead, in all populations” EFSA 2010.

The Proposition 65 Confusion

A great deal of the alarm traces back to one Californian number. The state's Office of Environmental Health Hazard Assessment lists a maximum allowable dose level for lead of 0.5 micrograms a day for reproductive toxicity, alongside a separate no-significant-risk level for cancer of 15 micrograms a day by the oral route, and states that exposures below a safe-harbour level are exempt from Proposition 65's requirements OEHHA. Under California's Proposition 65 regulations, a maximum allowable dose level is set at one one-thousandth of the level at which no reproductive effect is observed. A thousandfold margin is a defensible policy choice for a chemical with no identified threshold. It is not a line between safe and dangerous — and note that the figure both testing reports adopted is the reproductive one, thirty times stricter than the same agency's cancer level for lead.

Nor is it a sales ban. In the trade coverage of the Clean Label Project report, the Consumer Healthcare Products Association argued that “Prop 65 levels are not meant to represent a safety threshold or a cut-off for unsafe levels, but rather, a standard that triggers a label notification only in the state of California”, and the Council for Responsible Nutrition noted that the law “requires labeling of products over that amount, not a ban on them” NutraIngredients 2025. Read as “this crosses a threshold California set with a thousandfold margin”, a warning is accurate and worth knowing. Read as a poisoning claim, it is not.

What a Certification Seal Actually Proves

Third-party certification exists because supplement contamination has a documented history. Between October 2000 and November 2001, researchers at the Institute of Biochemistry of the German Sport University Cologne bought 634 non-hormonal nutritional supplements from 215 suppliers in 13 countries and analysed them by gas chromatography–mass spectrometry; 94 of them, 14.8 per cent, contained anabolic androgenic steroids not declared on the label. The rate varied sharply by dosage form: 19.6 per cent of capsules, 11.7 per cent of tablets and 6.9 per cent of powders — five of the 72 powders tested Geyer 2004. That study is about undeclared steroids rather than metals, its samples are a quarter of a century old, and powders were the least affected form. It is the origin story of batch testing, not evidence about lead.

Three marks are worth recognising, and they cover different ground. NSF certifies dietary supplements to NSF/ANSI 173, the American National Standard for the category, whose finished-product contaminant clause — 5.3.1.2 in the 2021 edition — sets per-daily-dose limits of 0.01 milligrams for inorganic arsenic, 0.0041 for cadmium, 0.02 for hexavalent chromium, 0.01 for lead and 0.002 for mercury; products in its Certified for Sport programme must also be certified to NSF/ANSI 173, NSF 229 or NSF 527, and are screened against the hundreds of substances banned by major athletic organisations NSF/ANSI 173. The US Pharmacopeia's Verified Mark covers ingredient identity and declared potency, requires that a product “does not contain harmful levels of specified contaminants” including heavy metals, checks that it breaks down in a specified time, and is backed by facility audits against current good manufacturing practices plus off-the-shelf retesting USP. Informed Sport, run by the laboratory group LGC, tests every batch of a certified product for 250 or more banned substances before it reaches retail, using ISO 17025 accredited methods, and continues testing blind samples after certification is awarded Informed Sport.

Read those specifications closely, because the ordering matters. Informed Sport is a doping-control programme: its standard certification is about banned substances, not heavy-metal content, so it is not evidence about lead at all. NSF/ANSI 173 and USP are the marks with contaminant limits — and their limits are permissive. NSF/ANSI 173 allows up to 0.01 milligrams of lead per daily dose, which is 10 micrograms, twenty times California's 0.5 microgram figure; the 2020 risk assessment records the US Pharmacopeia's permissible daily exposure for lead as also 10 micrograms a day, derived on a 50 kilogram bodyweight assumption Bandara 2020. So a seal tells you a batch was measured against a published standard by somebody other than the manufacturer, which is more than an uncertified tub offers. It does not tell you the powder is close to lead-free, and on the numbers above a certified product can sit well above both testing programmes' benchmarks and remain certified.

What This Means for a Daily Scoop

The honest position is neither “this is fine” nor “throw it out”. The cheapest lever is the one the testing organisation itself reached for: frequency. Consumer Reports recommended against daily use for most protein powders on the grounds that “none are necessary to hit your protein goals” Consumer Reports 2025, and the Irish academic assessment's caution was specifically about excessive and prolonged use Ring 2021. Both point the same way: the powder is a convenience, and convenience is the easiest thing in this article to change.

It helps to know what the supplement is buying you. A systematic review and meta-analysis of 49 randomised trials, 1,863 participants, found that protein supplementation during resistance training significantly increased one-repetition-maximum strength by 2.49 kilograms and fat-free mass by 0.30 kilograms — real but modest effects, with the authors noting that the training itself is the far more potent stimulus. The plateau figure people quote from that paper, at total protein intakes above about 1.62 grams per kilogram of bodyweight per day, comes from a segmental regression on a 42-arm subset that the authors present “despite not being statistically significant” at p equals 0.079, with a confidence interval running from 1.03 to 2.20 grams per kilogram. And their framing is the opposite of a warning about excess: they treat roughly 1.6 grams per kilogram a day as a sufficiency target, describe the 0.8 gram recommended dietary allowance as insufficient for people training for strength and size, and emphasise that older adults likely need more Morton 2018. So this is no licence to claim the second scoop is wasted. It is a reason to work out what your total daily protein intake actually is from food first, and to use powder for the gap rather than by habit.

After frequency, the filters run in a clear order. Source comes first, because it is the pattern both 2025 testing rounds agree on: if you use a plant protein daily, you are on the pathway that produced the higher lead numbers in both of them. Treat it as a tendency rather than a rule, though — the European assessment found lead running higher in animal-based supplements Bethencourt-Barbuzano 2025, and none of this testing tells you where an individual tub sits. Contaminant certification comes second, and specifically NSF/ANSI 173 or USP Verified rather than a banned-substance mark — while remembering how permissive those lead limits are. Flavour is the weakest lever of the three: the cocoa mechanism is documented in cocoa, but the two 2025 reports disagree about how much of it reaches a finished tub, and choosing vanilla is no substitute for choosing a tested one.

For anyone pregnant, breastfeeding, or stirring powder into a young child's food, the numbers to hold in mind are the regulators' own. The Food and Drug Administration's interim reference levels, as reported by Consumer Reports, are 2.2 micrograms of lead a day for children and 8.8 for females of childbearing age Consumer Reports 2025, and EFSA's panel concluded that protecting children against neurodevelopmental effects is protective across all populations EFSA 2010. None of the evidence cited here studied protein powder in pregnancy or childhood, so this is not an instruction to stop. It is an argument for buying the contaminant-tested tub, keeping the dose modest and the frequency lower than daily, and spending the remaining worry somewhere it changes an outcome.

Frequently asked questions

Should I throw out the protein powder I already have?

Nothing in the evidence cited here supports discarding an open tub. The screening-level risk assessment in Toxicology Reports modelled one to three servings a day, every day of the year, and reported a summed hazard index below one for arsenic, cadmium and mercury in every scenario but one: three servings a day of the single product with the highest reported cadmium concentration reached a hazard index of 1.757, which the authors set aside as unlikely to be representative. Modelled blood lead stayed below the guidance value the paper used in every scenario. The independent Irish assessment found concentrations at acceptable levels while cautioning that excessive and prolonged use of some products may carry a slightly elevated lifetime risk. The more useful response is to check whether you need the volume you are taking, and to make the next purchase one tested against a published contaminant standard.

Is plant protein worse than whey for heavy metals?

On the testing published so far, on average yes. The Clean Label Project's January 2025 report put 77 per cent of plant-based powders over California's Proposition 65 lead threshold, against 28 per cent of whey powders and 26 per cent of collagen ones, and Consumer Reports' 2025 round found lead in plant-based products averaging nine times the level in dairy-based ones. The Clean Label Project attributes contamination mainly to the soils ingredients are grown in and to packaging. That is a sourcing pattern across a shelf rather than a fixed property of any single tub — and it is not universal: a 2025 European assessment of 90 supplements found lead running higher in the animal-based products, with cadmium the metal elevated in the plant-based ones.

Does a Proposition 65 warning mean the product is unsafe?

Not on its own. California's Office of Environmental Health Hazard Assessment lists the maximum allowable dose level for lead at 0.5 micrograms a day for reproductive toxicity, plus a separate no-significant-risk level of 15 micrograms a day by the oral route for cancer, and exposures below a safe-harbour level are exempt from Proposition 65's requirements. The Consumer Healthcare Products Association's position is that Proposition 65 levels are not meant to represent a safety threshold but a standard that triggers a label notification only in California, and the Council for Responsible Nutrition notes that the law requires labelling of products over that amount rather than a ban.

What does an NSF or USP seal actually cover, and does Informed Sport cover lead?

NSF certifies supplements to NSF/ANSI 173, which sets per-daily-dose limits for contaminants including lead at 0.01 milligrams, or 10 micrograms; products in its Certified for Sport programme must also be certified to NSF/ANSI 173, NSF 229 or NSF 527 and are screened for hundreds of substances banned in sport. The USP Verified Mark covers ingredient identity, declared potency, limits for specified contaminants including heavy metals, disintegration, and facility audits. Informed Sport, run by LGC, tests every batch for banned substances before release; that is a doping-control programme, not a heavy-metals certification. None of these marks is a claim of zero lead: NSF's own lead limit is twenty times California's 0.5 microgram figure.

Are children and pregnant readers at higher risk?

The clearest evidence of harm at low exposure is developmental. A pooled analysis of seven prospective cohorts, 1,333 children, reported IQ decrements across blood lead concentrations below the old 10 microgram-per-decilitre action level, with the steepest declines below that level and no threshold identified. EFSA's contaminants panel concluded that protecting children against neurodevelopmental effects would be protective for all other adverse effects of lead in all populations. The Food and Drug Administration's interim reference levels, as reported by Consumer Reports, are 2.2 micrograms a day for children and 8.8 micrograms a day for females of childbearing age. None of that studied protein powder, but it is a reason to prefer contaminant-tested products and modest doses rather than to treat a labelling threshold as a safety line.

References