Between 2007 and 2016, the Food and Drug Administration identified 776 dietary supplements that contained unapproved pharmaceutical drugs. Seven of those products drew a warning letter to the firm responsible.
Seven. Out of 776. That ratio, published in JAMA Network Open in 2018 by a team from the California Department of Public Health’s Food and Drug Branch, is the single most useful fact about supplement contamination in the United States, because it tells you something the individual horror stories do not. The problem is not that nobody finds these products. Somebody does find them. What happens next is the problem.
Contamination in this category divides into four kinds: pharmaceutical drugs that were never declared, heavy metals, pesticide residues, and microbes. They have different causes, different detection records, and very different amounts of evidence behind them. This article covers what is documented about each, including where the documentation runs out.
The biggest category is prescription drugs
When people picture a contaminated supplement, they often picture dirt, mould or a trace of metal. The best-documented contaminant is a drug.
The California team pulled every entry from FDA’s tainted-supplements database for 2007 through 2016 and, after removing duplicates, counted 776 adulterated products. Three marketing categories accounted for almost all of them: sexual improvement at 353 products, weight loss at 317, and muscle building at 92. In 757 of the 776 cases, which is 97.6 percent, the drug was not declared on the label.
The drugs tracked the category. Among the sexual improvement products, 287 of 353 contained sildenafil, the active ingredient in a prescription erectile dysfunction medicine, or one of its chemical analogues. Among the weight-loss products, 269 of 317 contained sibutramine, an appetite suppressant FDA removed from the US market in 2010 over cardiovascular risk, and 75 contained phenolphthalein, a laxative pulled in 1999 over a possible cancer risk. Among the muscle-building products, 82 of 92 contained synthetic steroids or steroid-like compounds. A small “other” group of 14 products, mostly marketed for joint and muscle pain, included seven containing the prescription anti-inflammatory diclofenac and five containing the steroid dexamethasone.
Then there were combinations. One hundred and fifty-seven products, just over a fifth, contained more than one drug. Thirty-three contained three or more. Two products contained six drug ingredients each.
That last detail is the clinical point. A prescription medicine comes with a dose, a warning list and a set of contraindications. A supplement containing the same molecule at an unknown concentration comes with none of those. Someone taking nitrates for chest pain is told not to combine them with sildenafil, because the combination can drop blood pressure dangerously. Nothing on a sexual improvement supplement’s label would tell that person they were about to do exactly that.
Two numbers for the same problem, and both are right
The 2018 paper and the invited commentary published alongside it in the same journal issue report different denominators, and a careful reader will hit the discrepancy immediately.
The research paper reports 776 adulterated products, of which 360, or 46.4 percent, were associated with a voluntary recall. The commentary, by Pieter Cohen of Cambridge Health Alliance and Harvard Medical School, reports 746 brands with 360 recalled, or 48 percent.
Both are correct about different objects. The research paper states that its 776 entries covered 746 distinct products, because a handful of products were identified more than once. The commentary divides the same 360 recalls by the count of distinct products. Nothing is wrong with either figure, and this article is not going to pretend one supersedes the other.
There is a second, smaller inconsistency inside the research paper itself. Its abstract says 146 supplement companies were implicated; its results section says 147 companies were named in total. One unit apart, in the same publication. The honest summary is that roughly 150 companies were involved.
The reason to report all of this rather than smooth it over is that the headline number gets quoted constantly without its denominator, and a reader who later encounters the other version deserves to know why they differ.
What a warning does
Once FDA knows a product contains a drug, several things can happen, and the weakest of them happens most.
The 2018 analysis sorted every product by the most serious action associated with it. Voluntary recalls, initiated by the company rather than ordered by the agency, covered 360 products. Public notifications, which are essentially a posted alert, covered 342. News releases covered 58, consumer updates 8, and a Department of Justice press release 1. Warning letters to the firm covered seven products, under 1 percent. Over the final five years of the period the trend ran further toward the informal end: in both 2015 and 2016, more than 80 percent of products drew a public notification rather than a recall.
The obvious question is whether the notice works. A 2022 research letter in JAMA, written by the same Cambridge Health Alliance physician with three co-authors at the University of Mississippi, answers it with purchases rather than assumptions. The team took 31 supplement products named in FDA warning letters issued in 2015, 2016 and 2019 over three prohibited stimulants, and went looking for them online.
One product had been recalled. Nine were still available to buy, a mean of six years after the warning letter. On chemical analysis, five of those nine still contained at least one FDA-prohibited ingredient. One of the five still held the very drug its own warning letter had named, and still declared it on the label. The other four had swapped in a different prohibited ingredient, and three of those four listed the replacement openly.
A separate study in Clinical Toxicology in 2022 found something stranger. Researchers bought four brands of supplements labelled as containing phenibut, a drug approved in Russia but never in the United States and linked to more than 80 reported cases of coma and death, then bought them again after FDA warned that it does not belong in supplements. Before the warning, two of the four contained it, at 484 mg and 487 mg per serving. After the warning, all four contained it, ranging from 21 mg to 1,164 mg per serving. The quantity went up in three of the four products.
A warning letter is a letter. That is the measured result, not a cynical reading.
It is worth being precise about what those studies do and do not establish. They tracked specific named products through specific enforcement actions. They do not estimate how common contamination is across the market, and neither does the 776 figure, because, as the authors of the 2018 analysis stated plainly, the total number and variety of products FDA tests each year is unknown. A rising count of adulterated products could reflect a worsening market or a more active agency, and the published data cannot separate the two.
Bacteria, and a recall clock that runs late
Microbial contamination is less common in the record than drug adulteration, but when it happens the timeline is instructive, and there is a very recent case to work from.
The Centers for Disease Control and Prevention closed its investigation into a Salmonella outbreak tied to greens supplement powders and moringa powders and capsules with a final update on 22 July 2026, stating that the outbreak was over. The final count was 131 cases across 38 states. Of the 121 people for whom information was available, 36 were hospitalised. No deaths were reported. Three serotypes were involved: Typhimurium, Newport and Richmond.
Now look at the dates. Illnesses began on 22 August 2025. The first voluntary recall, of a greens powder, began on 14 January 2026, which is roughly five months after people started getting sick. A nationwide expansion followed on 20 January and a moringa capsule recall on 28 January. Then, on 26 May 2026, a different company recalled two more moringa products. The last illness onset was 14 June 2026. CDC’s investigation was closed in March 2026 and later reopened.
So the recalls began about five months after the first illness, and they kept being added for another four and a half months after that. A further recall on 30 June 2026 does not belong in that count: CDC says the ingredient samples behind it carried Salmonella that did not match the outbreak strain. That is a story about how long a contaminated shelf-stable powder can keep circulating, not a story about one bad batch.
Running alongside it was a second, smaller and harder case. FDA investigated an outbreak of extensively drug-resistant Salmonella linked to a brand of moringa powder capsules, with 10 illnesses and 3 hospitalisations across eight states, onsets between 26 September 2025 and 8 January 2026, and a recall on 13 February 2026. Of eight people interviewed, seven reported eating the capsules. FDA described the outbreak strains as “resistant to all first-line and alternative antibiotics” normally recommended for treating Salmonella infection.
Shelf-stable botanical powders are a plausible vehicle for this. They are dried plant material, often imported, consumed without cooking, and frequently mixed into a cold drink. Nothing in that sequence kills a pathogen.
Heavy metals and pesticides: what has been measured, and what has not
This is the section where the honest answer is partly about missing data.
There is no current, peer-reviewed US survey that measures heavy metals across the whole supplement market and reports concentrations metal by metal. Several non-US market studies exist, and extrapolating from them to American shelves would not be legitimate. What exists instead are two solid substitutes.
The first is a government laboratory survey. In testimony to the Senate Special Committee on Aging on 26 May 2010, the Government Accountability Office reported results from testing 40 single-ingredient herbal supplement products at an accredited lab. Thirty-seven of the 40 contained trace amounts of at least one potentially hazardous contaminant. All 37 tested positive for trace lead. Mercury turned up in 32, cadmium in 28 and arsenic in 21. Eighteen products carried residues from at least one pesticide.
Two caveats matter and GAO stated both. None of the contaminant levels were at amounts considered to pose an acute toxicity hazard, and the heavy-metal levels did not exceed FDA or Environmental Protection Agency limits for supplements. Separately, FDA agreed that 16 of the 40 products would be considered in violation of US pesticide tolerances if the agency’s own testing confirmed GAO’s results. Those two pesticide numbers answer different questions, one about presence and one about legal violation, and they should not be merged.
The second substitute is narrower and more current. A 2022 narrative review in Annals of Pharmacotherapy by C. Michael White of the University of Connecticut School of Pharmacy searched the literature from 1980 to May 2021 on lead in mineral and multivitamin-multimineral products. The pattern it describes is encouraging in the aggregate and unhelpful at the individual level. Lead was historically higher in calcium drawn from dolomite, bone meal and unrefined carbonate than in the refined salts, and that difference has shrunk substantially since the turn of the century. Where multivitamins have been tested from 2007 onward, the central values came back low. But the review also found large differences between products, with many containing very little lead and some containing concerning amounts, and it concluded that neither a pharmacist nor a consumer can tell which is which without independent third-party testing. Note the direction of that finding: the average is reassuring and the outliers are the problem, which is the opposite of how contamination stories are usually told.
Pesticides have almost no routine federal surveillance in this category at all. FDA’s pesticide residue monitoring programme has no dedicated dietary supplement commodity group, and the handful of supplement samples it does test are counted in the companion article on what regulators verify. Protein powders are covered separately in this library, as are contamination and identity problems specific to herbal products.
The lists, and what they admit they leave out
FDA publishes a Health Fraud Product Database, built from warning letters, online advisory letters, recalls, public notifications and press announcements. It is genuinely useful for checking a specific product name.
It is also candid about its limits. The page states that the list includes only a small fraction of the potentially hazardous products marketed to consumers online and in retail stores, and adds: “Even if a product is not included in this list, consumers should exercise caution before using certain products.” A product’s absence from the database means nothing.
Two practical notes about using it. First, the separate database of tainted products that researchers analysed through 2016 no longer exists as a standalone resource; that web address now redirects into the Health Fraud Product Database. Any older guide that sends you to a dedicated tainted-products list is out of date. Second, the database does not display a running total of entries, so no current count can be given here. The two defensible figures are the 776 products for 2007 through 2016 from the peer-reviewed analysis, and FDA’s own 2025 statement, in guidance for retailers and distributors, that in recent years it has identified over a thousand products marketed as supplements or conventional foods carrying hidden drugs and chemicals.
FDA has also stated the structural reason the list is incomplete. On its page about products contaminated with hidden ingredients, current as of 10 June 2025, the agency says federal regulations prevent it from testing these products before they reach the market, and that it tests as many as it can once they are on sale. Everything in the database is therefore a product that was already available to buy, sometimes for years, before anyone measured what was in it.
Who can order a recall, and since when
Most people assume FDA can simply order a contaminated supplement off the shelf. For most of the history of the modern supplement market, it could not.
FDA received mandatory recall authority over most regulated foods, including dietary supplements, only when the Food Safety Modernization Act was signed on 4 January 2011. Section 206(a) of that law added section 423 to the Federal Food, Drug, and Cosmetic Act. Before that date, a recall depended on a company agreeing to one.
The agency’s reports to Congress record how often the new power was used. Its 2013 report describes the first use, a notification letter sent on 13 February 2013, and the product was not a supplement: it was pet treats. Its 2014 report calls the action it describes the second time the agency had sent such a letter, and that one did involve a dietary supplement firm, whose products were tied to a cluster of liver illnesses in Hawaii. FDA issued a warning letter in October 2013, notified the company in November that there was a reasonable probability the products were adulterated, and a company-initiated recall followed within days.
Twice in four years, across every food FDA regulates. That is the enforcement backdrop behind a voluntary recall system.
Signs that need a clinician, not a search engine
Contamination becomes a personal problem through symptoms, and the useful thing is knowing which symptoms to treat as urgent.
Seek emergency care for chest pain, a racing or irregular heartbeat, fainting, severe shortness of breath, a sudden severe headache, confusion or one-sided weakness that begins after taking a product marketed for sexual improvement, weight loss, energy or muscle building. The undeclared drugs documented in those three categories act on blood pressure, heart rhythm and the nervous system, and they interact with common prescriptions.
Seek medical care, and say what you took, for yellowing of the skin or eyes, dark urine, pale stools or unexplained severe fatigue, which can signal liver injury. For a greens powder, botanical powder or capsule, treat bloody diarrhoea, a fever above 102 degrees Fahrenheit, diarrhoea lasting more than three days, or signs of dehydration as reasons to be seen rather than to wait, and mention the product by name. A drug-resistant organism changes what treatment will work, and a clinician cannot factor that in without knowing what was consumed.
Three things worth doing in any of those situations:
- Keep the container, including the lot number and expiration date. It is the only record of what was taken, and it is what an investigator would need.
- Tell the clinician the product category, not just “a supplement.” Sexual improvement, weight loss and muscle building are the categories with the documented drug problem, and naming the category changes what a clinician looks for.
- Report the problem. Consumer reports are one of the inputs FDA uses to decide what to sample, and products enter the fraud database partly because somebody reported them.
The strongest single takeaway is the one the numbers keep pointing at. Finding a contaminated product is not the hard part, and a public notice is not the same as removal. Nine of 31 products named in FDA warning letters were still for sale six years later, and five of those still contained a banned drug. If a product sits in one of the three categories where undeclared pharmaceuticals cluster, the fact that it is still being sold tells you very little about whether anyone has looked at it.
Sources
- JAMA Network Open. “Tucker J, Fischer T, Upjohn L, Mazzera D, Kumar M. “Unapproved Pharmaceutical Ingredients Included in Dietary Supplements Associated With US Food and Drug Administration Warnings.” JAMA Network Open, 2018. PubMed. DOI: 10.1001/jamanetworkopen.2018.3337.” doi.org
- JAMA Network Open. “Cohen PA. “The FDA and Adulterated Supplements—Dereliction of Duty.” Invited commentary, JAMA Network Open, 2018. PubMed. DOI: 10.1001/jamanetworkopen.2018.3329.” doi.org
- JAMA. “Cohen PA, Avula B, Katragunta K, Khan I. “Recalls, Availability, and Content of Dietary Supplements Following FDA Warning Letters.” JAMA, 2022. PubMed. DOI: 10.1001/jama.2022.9734.” doi.org
- Clinical Toxicology. “Cohen PA, Ellison RR, Travis JC, Gaufberg SV, Gerona R. “Quantity of phenibut in dietary supplements before and after FDA warnings.” Clinical Toxicology, 2022;60(4):486-488. PubMed. DOI: 10.1080/15563650.2021.1973020.” doi.org
- Annals of Pharmacotherapy. “White CM. “Lead in Mineral or Multivitamin-Multimineral Products.” Annals of Pharmacotherapy, 2022. PubMed. DOI: 10.1177/10600280211023328.” doi.org
- US Centers for Disease Control and Prevention. “Investigation Update: Salmonella Outbreak Linked to Super Greens and Moringa Supplement Powders (final update 22 July 2026).” cdc.gov
- US Food and Drug Administration. “Outbreak Investigation of Extensively Drug-Resistant Salmonella: Moringa Powder (February 2026).” fda.gov
- US Government Accountability Office. “GAO-10-662T, Herbal Dietary Supplements: Examples of Deceptive or Questionable Marketing Practices and Potentially Dangerous Advice (26 May 2010).” gao.gov
- US Food and Drug Administration. “Health Fraud Product Database.” fda.gov
- US Food and Drug Administration, Office of Inspections and Investigations. “Tainted Products Marketed as Dietary Supplements and Foods: Responsibilities of Retailers and Distributors (2025).” fda.gov
- US Food and Drug Administration. “Avoiding Products Contaminated with Hidden Ingredients.” fda.gov
- US Food and Drug Administration. “Annual Report to Congress on the Use of Mandatory Recall Authority – 2014.” fda.gov
- US Food and Drug Administration. “Annual Report to Congress on the Use of Mandatory Recall Authority – 2013 (first use: 13 February 2013 notification letter, pet treats).” fda.gov