There is a number you have almost certainly seen: "59% of herbal supplements contain plants that are not on the label." It comes from a single 2013 study — and in July 2024 that study was retracted after an investigation found evidence of fabricated data. That is an awkward fact for everyone in this industry, including those of us who benefit when buyers worry about quality. But it matters, because the real question — how often is the plant on the label actually the plant in the bottle? — has a genuine answer, and it deserves better than a number nobody checked. This article gives that answer, explains why botanical identity is harder than it sounds, and covers the regulatory changes now moving through Washington in 2026.

The Number Everyone Quotes, and What Happened to It

In 2013, a team led by researchers at the University of Guelph published a DNA-barcoding study in BMC Medicine reporting that most of the North American herbal products they tested contained species not listed on the label. It travelled far: news coverage, industry presentations, and — in 2015 — a headline enforcement action in which the New York Attorney General sent cease-and-desist letters to four major retailers over supplements said to have failed DNA testing.

In July 2024 the paper was retracted. The retraction notice states that an investigation by the University of Guelph found evidence of fabricated data, and that the journal's editor no longer had confidence in the results.1 In fairness, several of the original authors publicly disagree with the retraction. But the paper is withdrawn, and the most-repeated statistic in this field no longer has a source behind it.

We are pointing this out even though the scary version of the story is commercially convenient for a company that tests every incoming botanical. A claim that turns out to rest on fabricated data does not become acceptable because it flatters our own practices.

So What Is the Real Number? Two Methods, One Answer

Retracting one study does not make adulteration imaginary — it just means we need evidence that survives scrutiny. There is some, and it is unusually tidy.

A 2019 meta-analysis pooled DNA-based authentication results for 5,957 commercial herbal products sold in 37 countries across six continents. It found that 27% were adulterated relative to their labelled species.2

Two years later, a companion review approached the same question from the opposite methodological direction — chemical authentication (HPLC, HPTLC, LC-MS, spectroscopy) rather than DNA — across 2,386 commercial products from 37 countries. It arrived at the same headline figure: 27% adulterated, 73% authentic.3 Among countries with more than 100 samples the spread was United Kingdom 37%, Italy 31%, United States 27%, China 21%.

Two different families of laboratory method, two different product sets, one number. That convergence is what makes 27% credible in a way 59% never was. One honest caveat: the two reviews share an author, so this is not two fully independent teams — it is one group deliberately checking its own conclusion with a different toolset, which is still a great deal better than a single unreplicated result.

So: roughly one product in four, globally, is not fully what its label says. That is a real problem worth fixing. It is also three in four that are fine — a materially different picture from "most supplements are fake."

Medicinal plant specimens in separate petri dishes for laboratory identification

What the evidence actually supports:

  • ~27% of herbal products globally are adulterated — found twice, by two different method families
  • The famous 59% figure came from a study retracted in 2024 for fabricated data
  • Rates differ by country: UK 37%, Italy 31%, US 27%, China 21%
  • DNA testing fails on extracts — chemistry does that job
  • Identity must be checked on incoming raw material, not the finished bottle

Why Botanical Identity Is Genuinely Hard

It is tempting to read "adulterated" as "someone cheated." Sometimes that is exactly right — a cheap filler standing in for an expensive root. But a large share of substitution is ordinary confusion, and understanding why makes you a better judge of who to trust.

  • Look-alike species in the same genus. Plants that a botanist separates by a seed capsule or a leaf hair look identical once dried, cut and sifted.
  • Common names that travel badly. One name can cover several unrelated species in different regions, and a single species can carry a dozen local names across its range.
  • Processing destroys the evidence. Powdering removes the microscopic features a trained analyst would use; extraction removes most of them and much of the DNA too.
  • Long, opaque supply chains. Material can change hands several times between a wildharvester and a manufacturer, and every step is a chance for a mix-up nobody intended.

A small illustration from our own work. While sourcing photographs for an article on pau d'arco, we searched the botanical name Tabebuia impetiginosa. A large share of results were a different species — Tabebuia rosea — and others were the yellow-flowered ipê, a different tree altogether. The pictures were labelled confidently and looked plausible. If that is how easily a photograph of a flowering tree gets mislabelled, it is not hard to see how a sack of dried bark does. The answer is not to trust harder; it is to test.

Why DNA Testing Is Not the Whole Answer

DNA barcoding sounds like it should settle everything, and for some product forms it nearly does. The catch is what happens to DNA during manufacturing. As the chemical-authentication review puts it, false negatives are expected when DNA "has been degraded or lost during post-harvest processing or manufacturing"3 — precisely the situation in a liquid extract, where the plant's genetic material is largely gone but its chemistry remains.

This was the substance of the industry's objection to the 2015 New York enforcement action: DNA barcoding had been applied to finished extracts, a product form in which a negative result may say more about the method than the material. That objection was technically sound, whatever one thinks of the products involved. The right test depends on the form:

  • Whole or cut-and-sifted herb — botanical microscopy and DNA methods both work well; the material still has intact structure and recoverable DNA.
  • Powders — microscopy weakens as particle size drops; DNA and chemistry both contribute.
  • Extracts, tinctures, glycerites — chemistry does the work. HPTLC produces a visual "fingerprint" of the plant's compounds; HPLC, LC-MS and infrared spectroscopy confirm identity against reference standards.

The practical implication is simple, and it is the most useful thing in this article: identity testing has to be done on the incoming raw material, before extraction, where every method still works. Testing only the finished bottle is testing at the point where the evidence is weakest.

What Actually Protects You Right Now

Contrary to the popular framing, dietary supplements are not unregulated in the United States. Since 2007, manufacturers have been bound by cGMP rules under 21 CFR Part 111, which require — among much else — that the identity of every incoming dietary ingredient be verified before use, that specifications be set and met, and that batch records document all of it.

The weak point is not the rule; it is verification that the rule is being followed. That is the difference between a company stating it follows cGMP and a company that has been audited and certified against it by an independent body. The first is a claim. The second is someone with no commercial stake examining the records.

→ This is what our own quality system and SGS cGMP certification look like in detail

2026: The Rules Are Moving

After nearly two decades of stability, supplement regulation in the United States is genuinely in motion this year, from several directions at once.

  • The medical profession has formally weighed in. In July 2026 the American College of Physicians published a position paper in Annals of Internal Medicine arguing that the present framework has allowed adulterated and mislabelled products onto the market, and calling for stronger FDA authority both before and after products go on sale, standardised terminology, and better data sharing.4
  • Mandatory product listing. A bill introduced in January 2026 would require manufacturers to list their products with the FDA — so that regulators know what is actually on the market, which today they largely do not.
  • One rulebook instead of fifty. A separate February 2026 bill would reaffirm federal primacy and head off a growing patchwork of state-level supplement laws.
  • The FDA is modernising definitions. The agency has been reconsidering what counts as a "dietary substance," including novel ingredients and those made with newer technologies.

The industry has pushed back, particularly against anything resembling pre-market approval, on the argument that supplements are foods rather than prescription drugs and that a drug-style regime would be disproportionate. That is a real debate with reasonable people on both sides, and we are not going to pretend it is settled.

Our own view is narrower and easier to state: whichever way the legislation lands, the direction of travel is toward proving what you claim. Identity testing, documented batch records and independent audit are the things that will still be there when the arguing stops.

How to Check a Brand in Five Minutes

  • Look for the botanical name, not just the common one. A label that says Hydrastis canadensis and not merely "goldenseal" is a label written by someone who knows which species they bought.
  • Ask which plant part was used. Root, leaf, bark and seed differ chemically. Vagueness here is a signal.
  • Ask whether identity is tested on incoming raw material, and by what method. "We test" is not an answer; "HPTLC and FTIR on every incoming lot" is.
  • Check whether cGMP compliance is certified or merely claimed, and by whom. Ask for the certificate number.
  • Ask about heavy metals and microbiological testing, and whether some of it is done by an accredited outside laboratory.
  • Be wary of a product that avoids specifics while making bold health claims. The two usually travel together.

A Note on the Evidence

The honest summary is that adulteration in the global herbal market runs around one product in four, that this figure is supported by two methodologically different reviews rather than one dramatic study, and that the most famous claim in this field was retracted for fabricated data. The problem is real and materially smaller than folklore suggests. It is also tractable: it is solved by testing the right thing at the right stage, and by someone independent checking that you did.

References

  1. Retraction Note: DNA barcoding detects contamination and substitution in North American herbal products. BMC Medicine, 2024.
  2. The DNA-Based Authentication of Commercial Herbal Products Reveals Their Globally Widespread Adulteration. Frontiers in Pharmacology, 2019.
  3. Chemical Authentication of Botanical Ingredients: A Review of Commercial Herbal Products. Frontiers in Pharmacology, 2021.
  4. Modernizing the Regulatory Framework for Dietary Supplements: A Position Paper From the American College of Physicians. Annals of Internal Medicine, 2026.

These statements have not been evaluated by the Food and Drug Administration. This article is provided for educational purposes only and is not medical advice, legal advice, or a description of any specific company's products other than where stated. Regulatory proposals described here were accurate at the time of writing and may change; consult primary sources for the current status of any legislation.