Safety, Adulteration and Product Quality
Most herb-safety pages are about the herb. This one mostly is not, and that is the finding, not an evasion.
Epimedium leaf itself appears to be reasonably well tolerated at ordinary doses over short periods. It carries some real cautions — oestrogenic activity, liver-injury reports, a single cardiac case report — and those are covered in full below. But the dominant risk attached to buying horny goat weed is not pharmacological. It is that the product category it is sold into is systematically adulterated with prescription drugs, in quantities that can exceed the maximum approved dose, with no warning label, sold to people who have never been medically screened.
The numbers below are not estimates or scare copy. They are counts from published analytical surveys and from the FDA’s own database. Read them before you read anything else on this page.
Table of Contents
- What the Analytical Surveys Actually Found
- One Product, Fifteen Years, Eight Drugs
- Why a Hidden PDE5 Inhibitor Can Kill
- How to Spot a Suspect Product
- The Herb’s Own Risks
- Liver Injury
- Oestrogenic Activity
- Drug Interactions
- Who Should Avoid It Entirely
- Reading a Label
- Key Research Papers
- Connections
What the Analytical Surveys Actually Found
The French academic survey: 150 products, NMR spectroscopy
A group at the Université Paul Sabatier in Toulouse bought 150 dietary supplements marketed for sexual performance, every one of which claimed to contain only natural compounds, plant extracts or vitamins, and analysed them by proton NMR with mass spectrometry confirmation. The results, published in 2015:
- 61 percent were adulterated with PDE5 inhibitors. Twenty-seven percent with the actual medicines — sildenafil, tadalafil, vardenafil — and 34 percent with structurally modified analogues designed to evade routine screening.
- Among the adulterated products, 36 percent contained a mixture of two, three or even four different PDE5 inhibitors.
- A quarter of the drug-tainted products exceeded the maximum recommended dose of the medicine they secretly contained.
- A further 5.5 percent contained other sexual-dysfunction drugs: yohimbine, flibanserin, phentolamine, DHEA or testosterone.
- Only 31 percent could be considered genuine herbal or natural products.
- Following individual products over time showed manufacturers quietly changing what was inside. The authors also documented plain manufacturing failure: impurities, degradation products, and different compositions within the same stated batch number.
One detail is worth pausing on. A small fraction — 2.5 percent — contained genuine plant-derived compounds including icariin. In a survey of 150 products sold on the strength of botanical claims, the botanical was the rare finding and the pharmaceutical was the common one.
The US convenience-store survey
A separate 2013 study collected 91 samples representing 58 distinct products, almost all bought from convenience stores and filling stations in two American metropolitan areas, priced between $2.99 and $17.99, and analysed them by liquid chromatography–mass spectrometry. No sample claimed to contain any synthetic substance.
- 74 of 91 samples (81 percent) contained PDE5-inhibitor pharmaceutical ingredients.
- Forty contained tadalafil and/or sildenafil — and 18 of those contained more than 110 percent of the highest approved drug strength.
- Thirty-four contained PDE5-inhibitor analogues.
- Content varied markedly between samples of the same product, indicating essentially no quality control.
- Only 14 of the 91 samples carried any warning against taking them with nitrates.
An honest disclosure about that study: it was coordinated by Pfizer Global Security, and Pfizer markets sildenafil. That is a genuine conflict of interest and you should weigh it. It matters less than it might, because the independent French academic group with no commercial stake found the same pattern by a different analytical method on a different continent, and because the FDA’s own regulatory record — below — is a third, independent line of evidence pointing the same way.
The FDA’s own record
Investigators at the California Department of Public Health analysed the FDA’s Tainted Products Marketed as Dietary Supplements database for 2007 through 2016 and published the summary in JAMA Network Open:
- 776 adulterated dietary supplements were identified, implicating 146 different companies.
- Sexual enhancement was the single largest category: 353 products, 45.5 percent of the total — ahead of weight loss and muscle building.
- Sildenafil was the adulterant in 166 of those 353 sexual-enhancement products (47 percent).
- One in five adulterated products contained more than one unapproved drug.
- Twenty-eight products were the subject of two or three separate FDA warnings more than six months apart — and 19 of those 28 (68 percent) contained a new unapproved ingredient by the later warning. Being caught changed the drug, not the practice.
One Product, Fifteen Years, Eight Drugs
The Toulouse group later published a case history that captures the whole problem in one product line. They analysed ten samples of a supplement sold as a 100 percent natural aphrodisiac, purchased between 2007 and 2022.
- Every single sample was tainted. Not most. All ten, across fifteen years.
- Eight different adulterants were identified over that period: sildenafil itself, six sildenafil analogues, and a vardenafil analogue.
- Concentrations ranged from 15 to 145 mg per capsule. Four samples contained at least 100 mg per capsule — the maximum recommended single dose of prescription sildenafil, in a capsule labelled as a herb.
- The identity of the adulterant kept changing, apparently to stay ahead of regulators’ routine screening tests.
- Despite repeated warnings and seizures by European food and health authorities, the product remained on sale over the internet throughout.
The authors’ conclusion was blunt: this demonstrates the impossibility of controlling this market. They called for a European public database of tainted supplements modelled on the FDA’s, with product names, photographs and named adulterants.
Why a Hidden PDE5 Inhibitor Can Kill
Sildenafil is a good drug. Prescribed properly, it is safe, cheap as a generic, and has been used by tens of millions of men. The danger is not the molecule. It is receiving the molecule without knowing it, at an unknown dose, without any of the screening that precedes a prescription.
The nitrate interaction
This is the one that kills people. Nitrate medicines — nitroglycerin spray or tablets, isosorbide mononitrate, isosorbide dinitrate — work by donating nitric oxide, which drives production of cGMP and relaxes blood vessels. A PDE5 inhibitor blocks the enzyme that destroys cGMP. Give both and cGMP accumulates without a brake: vessels throughout the body dilate at once, blood pressure collapses, and the fall can be profound, prolonged and resistant to treatment. The consequences include myocardial infarction, stroke and death. The contraindication on every PDE5 inhibitor label is absolute, not relative.
Now consider who takes nitrates: people with angina — that is, people with coronary artery disease. And consider who has erectile dysfunction: disproportionately, people with coronary artery disease, because both are manifestations of the same endothelial pathology. The overlap between the group at risk from this interaction and the group buying these products is not small. It is the same population.
Recreational nitrites — amyl, butyl or isopropyl nitrite, sold as “poppers” — carry the identical interaction, and the overlap with sexual-enhancement supplements is obvious.
The other hazards of an unscreened dose
- Alpha-blockers. Tamsulosin, doxazosin and similar drugs for prostate symptoms or hypertension combine with PDE5 inhibitors to cause symptomatic hypotension and fainting. Prescribers separate the doses and start low; a supplement does neither.
- Cardiac disease that makes sex itself risky. Prescribing a PDE5 inhibitor involves assessing whether the patient’s heart can tolerate sexual activity. Recent myocardial infarction or stroke, unstable angina, severe aortic stenosis and uncontrolled arrhythmia all change that answer. Nobody performs that assessment at a petrol station.
- Overdose effects. At more than the approved maximum — which nearly a quarter of tainted products delivered — headache, flushing, visual disturbance, hypotension and priapism become more likely. Priapism, an erection lasting more than four hours, is a urological emergency; untreated it causes permanent tissue damage.
- Rare but serious label warnings. Sudden hearing loss, and non-arteritic anterior ischaemic optic neuropathy (NAION), a form of sudden vision loss. Rare with a prescription, and impossible to attribute or report when the drug was never declared.
- Analogues nobody has tested. A third of the French samples contained modified PDE5 inhibitors — molecules invented to evade detection. These have no toxicology, no human safety data, no known dose and no antidote. You cannot be warned about a compound that has never been studied.
How to Spot a Suspect Product
You cannot detect adulteration by taste, appearance or price. But the pattern of a spiked product is recognisable, and these signals track closely with what the analytical surveys found.
| Signal | Why it matters |
|---|---|
| Sold at a filling station, convenience store, adult shop or vending machine | The exact sampling frame in which 81 percent of products tested drug-positive |
| Single-serve sachet or blister of one or two capsules | A drug-dose format, not a supplement format. Nobody sells a two-day supply of a herb |
| Promises results in 30–60 minutes, or effects lasting days | Describes PDE5 pharmacokinetics. A leaf extract cannot do this |
| Aggressive brand name evoking hardness, power, rhinos, tigers or numbers of hours | Consistently over-represented in FDA warning lists |
| “Proprietary blend” with no per-ingredient milligrams | Legally shields the formula from disclosure, which is the point |
| No lot number or expiry date; inconsistent packaging | Documented in the adulteration surveys as a marker of no quality system |
| No warning about nitrates anywhere on the label | Only 14 of 91 tested samples carried one — and 74 of them contained a PDE5 drug |
| Bought from an overseas web shop or arriving by international mail | A third of recent FDA-identified adulterated products came from online sampling or mail interception |
| Sold with the claim that it is “like Viagra but natural” | Half of that sentence is often literally true |
Before buying anything in this category, search the FDA’s Tainted Sexual Enhancement Products database for the brand name. It is free, searchable, and lists the product, the company and the hidden drug found.
The Herb’s Own Risks
Set the adulteration problem aside and consider genuine Epimedium leaf.
The reassuring data point is the 2021 Singapore trial: 58 postmenopausal women took 740 mg per day of a purified prenylflavonoid extract for six weeks under double-blind conditions, with no significant adverse symptoms and no changes in liver, kidney or blood parameters. That is a real, monitored human safety signal and it is worth having. It is also six weeks in 29 treated women, which is a long way from establishing safety for indefinite daily use.
Cardiac effects
There is one published case report: a man who developed a tachyarrhythmia together with hypomania while taking horny goat weed, reported in Psychosomatics in 2004. A single case report proves nothing about causation — the report itself does not claim to. But it is not nothing either, given that this is a plant with documented vascular activity, and it is the reason anyone with a known arrhythmia or structural heart disease should leave it alone.
Commonly reported side effects
At higher intakes, users report dizziness, dry mouth, thirst, nosebleeds, nausea and a racing heartbeat. These are not well quantified because no adequately powered long-term trial has collected them systematically.
Liver Injury
This deserves its own section because it is the herb’s most substantiated intrinsic risk and it is not widely known in the West.
A 2024 study in Biomedical Chromatography from Beijing University of Chinese Medicine opens with a statement that should be on every label: Epimedium has been the subject of numerous adverse-reaction reports in recent years, and the most common of these is liver injury. The researchers profiled the herb’s constituents, identified those absorbed into blood, and used network analysis plus metabolomics in human liver cells to nominate a likely culprit — icaritin, one of the sugar-stripped metabolites that human pharmacokinetic work shows actually circulates after an oral dose. A 2024 systematic review of Epimedium koreanum similarly gives toxicity its own treatment rather than a footnote.
Two things follow. Herb-induced liver injury is usually idiosyncratic: unpredictable, not dose-related in any simple way, and not preventable by taking a smaller amount. And the mechanism proposed here implicates a metabolite that is more abundant after processing or prolonged use, not less.
Stop the herb and seek medical care for: yellowing of the skin or the whites of the eyes, dark urine, pale stools, persistent nausea, unusual fatigue, itching without a rash, or pain under the right ribs. Anyone with existing liver disease, or taking other hepatotoxic drugs, should not take Epimedium at all.
Oestrogenic Activity
Epimedium flavonoids selectively activate oestrogen receptor alpha. This has been shown in bone cells, and it is the leading explanation for the herb’s effect on bone density. The same property is a contraindication elsewhere.
The 2007 bone trial provides real reassurance for healthy women: over 24 months, neither serum oestradiol nor endometrial thickness changed in either group. That is a meaningful negative safety finding in exactly the tissue people worry about. But the trial enrolled healthy postmenopausal women with no hormone-sensitive disease. It cannot be extrapolated to someone with an oestrogen-driven tumour, and no study has tried.
Until better data exist, avoid Epimedium if you have or have had: oestrogen-receptor-positive breast cancer, endometrial cancer, ovarian cancer, endometriosis, or uterine fibroids. If you take tamoxifen or an aromatase inhibitor, do not add a phyto-oestrogen without oncology input — the theoretical interference has never been studied and the downside is not worth the experiment.
Drug Interactions
| Drug or class | Concern | Recommendation |
|---|---|---|
| Nitrates — nitroglycerin, isosorbide mono/dinitrate; recreational nitrites | Catastrophic hypotension if the product contains a PDE5 inhibitor, declared or not | Absolute avoidance |
| PDE5 inhibitors — sildenafil, tadalafil, vardenafil, avanafil | Additive effect; unknown total dose if the supplement is spiked | Do not combine |
| Alpha-blockers — tamsulosin, doxazosin, alfuzosin | Symptomatic hypotension, dizziness, falls | Avoid |
| Antihypertensives generally | Additive blood-pressure lowering; the herb has vascular activity and spiked products far more so | Medical supervision; monitor blood pressure |
| Anticoagulants and antiplatelets — warfarin, apixaban, rivaroxaban, clopidogrel, aspirin | Interaction data are sparse rather than reassuring. Bleeding risk cannot be excluded | Discuss with a pharmacist before starting; do not self-combine |
| Antiarrhythmics, or any known arrhythmia | Case report of tachyarrhythmia | Avoid |
| Hepatotoxic drugs — methotrexate, isoniazid, high-dose paracetamol, some statins | Additive liver risk | Avoid, or monitor liver enzymes |
| Oestrogens, tamoxifen, aromatase inhibitors | ERα activation may interfere | Specialist advice required |
| Drugs cleared by glucuronidation | Icariin has been shown to affect UDP-glucuronosyltransferase activity in mouse liver — an animal signal, not a demonstrated human interaction, but a reason for caution on a narrow-therapeutic-index drug | Mention it to your pharmacist |
| Surgery | Blood-pressure and bleeding uncertainty | Stop at least two weeks beforehand and tell the anaesthetist |
Who Should Avoid It Entirely
- Anyone taking a nitrate for angina, or using recreational nitrites
- Anyone with coronary artery disease, recent heart attack or stroke, unstable angina, severe aortic stenosis, or an arrhythmia
- Anyone on alpha-blockers or multiple blood-pressure medications
- Anyone with a hormone-sensitive cancer or condition, or on endocrine therapy
- Anyone with liver disease or unexplained abnormal liver enzymes
- Anyone taking an anticoagulant or antiplatelet drug without pharmacist review
- Pregnancy and breastfeeding — no safety data, demonstrable hormonal activity
- Children and adolescents — no data, and hormonal activity during development is not something to experiment with
- Anyone within two weeks of scheduled surgery
- Anyone whose erectile dysfunction has not yet been medically evaluated — see the sexual-function article on why this is a cardiovascular warning sign first
Reading a Label
If you have read all of the above and still want to try genuine Epimedium — a defensible decision, made with open eyes — the label is your only tool.
| What to require | Acceptable | Reject |
|---|---|---|
| Latin binomial | Epimedium brevicornu, E. sagittatum, E. koreanum or E. pubescens | “Horny goat weed” with no species |
| Plant part | Leaf / aerial parts (Epimedii Folium) | Unspecified, or “whole plant” |
| Marker | Icariin, given as a percentage and mg per capsule | “Standardised” with no marker named |
| Formula | Single herb, or every ingredient with its own mg | Any “proprietary blend” |
| Testing | Named third-party laboratory; certificate of analysis on request; ideally explicit screening for PDE5-inhibitor adulterants | A GMP logo alone |
| Traceability | Manufacturer address, lot number, expiry date | Any of these missing |
| Claims | Restrained, specific, no timing promises | Anything implying drug-like speed or reliability |
| Retailer | Established pharmacy or supplement retailer with a returns policy | Petrol station, adult shop, marketplace listing, unbranded overseas seller |
And two rules that override all of the above: tell your doctor and pharmacist you are taking it, and stop immediately if you develop chest pain, palpitations, fainting, unusual bruising or bleeding, jaundice, dark urine, or an erection lasting more than four hours — the last of which is a hospital visit, not a wait-and-see.
Key Research Papers
Each identifier below was verified live against NCBI E-utilities — first author, title, journal and year all had to match. All figures quoted in this article come from the abstracts of these papers.
Adulteration surveys
- Gilard V, Balayssac S, Tinaugus A, Martins N, Martino R, Malet-Martino M. Detection, identification and quantification by 1H NMR of adulterants in 150 herbal dietary supplements marketed for improving sexual performance. Journal of Pharmaceutical and Biomedical Analysis. 2015;102:476–493. 61 percent adulterated with PDE5 inhibitors; only 31 percent genuinely natural.
- Balayssac S, Danoun S, Gilard V, Martino R, Malet-Martino M. The POWER saga from 2007 to 2022: an example of a sexual enhancement dietary supplement tainted by different adulterants and still on the market. Journal of Pharmaceutical and Biomedical Analysis. 2023;227:115283. Ten samples over fifteen years, eight adulterants, 15–145 mg per capsule.
- Campbell N, Clark JP, Stecher VJ, Thomas JW, Callanan AC, Donnelly BF, Goldstein I, Kaminetsky JC. Adulteration of purported herbal and natural sexual performance enhancement dietary supplements with synthetic phosphodiesterase type 5 inhibitors. The Journal of Sexual Medicine. 2013;10(7):1842–1849. 81 percent of 91 convenience-store samples drug-positive. Industry-coordinated — see the disclosure above.
- 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;1(6):e183337. 776 adulterated supplements, 2007–2016; sexual enhancement the largest category.
- Patel DN, Li L, Kee CL, et al. Screening of synthetic PDE-5 inhibitors and their analogues as adulterants: analytical techniques and challenges. Journal of Pharmaceutical and Biomedical Analysis. 2014;87:176–190. Why detecting designer analogues is so difficult.
The herb’s own safety record
- Wang J, Cao Y, Sun M, et al. Integrating metabolomics and bioinformatics to reveal the mechanism of Epimedium-induced liver injury. Biomedical Chromatography. 2024;38(9):e5948. States that liver injury is the most common reported adverse effect; nominates icaritin as the hepatotoxic component.
- Qian HQ, et al. A systematic review of traditional uses, phytochemistry, pharmacology and toxicity of Epimedium koreanum Nakai. Journal of Ethnopharmacology. 2024;318(Pt B):116957.
- Partin JF, Pushkin YR. Tachyarrhythmia and hypomania with horny goat weed. Psychosomatics. 2004;45(6):536–537. The only published case report of a cardiac event attributed to this herb.
- Yong EL, Cheong WF, Huang Z, Thu WPP, Cazenave-Gassiot A, Seng KY, Logan S. Randomized, double-blind, placebo-controlled trial to examine the safety, pharmacokinetics and effects of Epimedium prenylflavonoids, on bone specific alkaline phosphatase and the osteoclast adaptor protein TRAF6 in post-menopausal women. Phytomedicine. 2021;91:153680. The one monitored human safety dataset — clean, but only six weeks.
- Xu SF, et al. Effect of icariin on UDP-glucuronosyltransferases in mouse liver. Planta Medica. 2014;80(5):387–392. An animal signal for a possible metabolism-based interaction.
- Xiao HH, et al. Flavonoids from Herba epimedii selectively activate estrogen receptor alpha (ERα) and stimulate ER-dependent osteoblastic functions in UMR-106 cells. Journal of Steroid Biochemistry and Molecular Biology. 2014;143:141–151. The basis of the hormone-sensitivity caution.
- Zhang G, Qin L, Shi Y. Epimedium-derived phytoestrogen flavonoids exert beneficial effect on preventing bone loss in late postmenopausal women: a 24-month randomized, double-blind and placebo-controlled trial. Journal of Bone and Mineral Research. 2007;22(7):1072–1079. Includes the reassuring 24-month endometrial-thickness and oestradiol findings.
Why the drug interaction matters to this population
- Thompson IM, Tangen CM, Goodman PJ, Probstfield JL, Moinpour CM, Coltman CA. Erectile dysfunction and subsequent cardiovascular disease. JAMA. 2005;294(23):2996–3002.
- Vlachopoulos CV, Terentes-Printzios DG, Ioakeimidis NK, Aznaouridis KA, Stefanadis CI. Prediction of cardiovascular events and all-cause mortality with erectile dysfunction: a systematic review and meta-analysis of cohort studies. Circulation: Cardiovascular Quality and Outcomes. 2013;6(1):99–109.
Live PubMed Searches
- Sildenafil adulteration of supplements
- Designer PDE5 analogues in supplements
- Epimedium and liver injury
- Herb-induced liver injury
- Nitrates and PDE5 inhibitors
- Priapism and PDE5 inhibitors
- Icariin and oestrogen receptors
- Supplement-related emergency visits
Connections
- All Herbs
- Epimedium Benefits — hub
- Epimedium (Horny Goat Weed)
- Sexual Function and the Icariin PDE5 Question
- Bone Density and Menopause
- Cardiology — who must never take a PDE5 inhibitor
- Tongkat Ali — sold into the same adulterated category
- Tribulus
- Maca