Evidence check

Horny Goat Weed, Icariin And The Human Trial That Does Not Exist

The mechanism is genuine and it is the same enzyme that prescription erectile medicines act on. That fact gets repeated across this entire category in language borrowed from clinical medicine. What almost nobody mentions is that nobody has ever run a randomised trial of it in men.

By VapoFil Team · Last updated November 4, 2026 · 8 min read · Evidence check · Reviewed by Dr. Erin Victor, ND

The mechanism, stated accurately

An erection is a blood-flow event. Nitric oxide relaxes smooth muscle in the arteries supplying erectile tissue, which lets blood in. An enzyme called phosphodiesterase type 5, or PDE5, breaks down the second messenger that carries the nitric oxide signal, which ends it. Inhibit the enzyme and the signal lasts longer. That is what sildenafil and tadalafil do.

Icariin, the marker compound in Epimedium, inhibits human PDE5 in a laboratory assay. That was published in 2008 in the Journal of Natural Products and the finding is real. It is why this plant appears on almost every label in this aisle and why the marketing language around it is so confident.

Everything in the paragraph above is accurate. Notice what it does not say: that anybody took icariin, or Epimedium, or a capsule containing either, and measured what happened to a person.

How strong an inhibitor icariin actually is

The follow-up literature answers this and the answer is not flattering.

A 2019 study in PLOS One explored the structural space around icariin, building and testing analogues to find out which features drive potency against human PDE5. The useful finding for a label reader is what it implies about the starting point: the interesting analogues were orders of magnitude more potent than icariin itself. Chemists modify a molecule when the original is too weak to be useful.

The animal work runs in the same direction. A 2018 study in Andrologia reported improved erectile function in spontaneously hypertensive rats and proposed a mechanism involving nitric oxide synthase. It is a reasonable piece of science and it was done in rats, at rat doses, in an animal model of hypertension.

• The enzyme result is genuine and was measured on isolated human PDE5.
• The compound is weak by the standards of the drugs it is compared with.
• The erectile evidence is an animal model, not a human outcome.
• The raw material varies considerably between batches and sources.
• No trial has measured what any of this does in a man.

The evidence ladder, and where this sits

Evidence about a compound climbs in rungs, and a claim can only honestly stand on the highest rung reached.

A claim is only as strong as the highest rung it has actually reached. For this ingredient that is rung two.

Icariin has reached rung two. That is not nothing and it is not a criticism of the science, which is perfectly good science. It is a statement about what can be claimed.

The problem in this category is that the language used about rung two is borrowed from rung four. A phrase like improves blood flow to intimate regions describes a clinical outcome, and the evidence behind it is an enzyme in a dish and a rat.

The fourth row is the one that matters and it is the one almost never printed.

Why the trial has not been done

Three reasons, none of them sinister.

Trials are expensive and nobody owns the plant. A pharmaceutical company running a phase III trial expects a patent at the end of it; nobody can patent Epimedium, so nobody has an economic reason to spend several million dollars finding out whether it works.

The comparator is brutal. Any trial of a botanical PDE5 inhibitor would be measured against drugs that are cheap, generic and highly effective. A result showing a small effect would be commercially useless and a result showing none would be worse.

And the material is inconsistent. The standardisation review sets out how variable Epimedium preparations are between sources, which makes designing a trial genuinely hard: you have to be able to say what was administered.

The row on this panel

On the bottle this website sells, horny goat weed appears at 8 mg of a 10:1 extract, stated as equivalent to 80 mg of dry herb. There is no standardisation and no icariin percentage, which is normal for a consumer label and which means the actual icariin content is unknown.

Two things follow. The first is that the row cannot be called under-dosed, because there is no trial dose to be under. The second is that it cannot be called adequate either, for the same reason.

This website therefore refuses the claim that goes with it. The seller's page says the ingredient helps maintain firmer and longer-lasting erections; that is a clinical statement resting on an enzyme assay, and it is one of five claims recorded as refused on the disclaimer page with the reason printed beside it.

What the traditional-use argument is worth

The fallback position, when a trial is absent, is that the plant has been used for centuries. It is worth taking that argument seriously rather than dismissing it, because it does carry some information — just not the information it is usually offered as.

What long traditional use establishes reasonably well is tolerability. A plant consumed by large numbers of people over a long period is unlikely to be acutely dangerous at customary amounts, because that would have been noticed. That is a genuine and useful inference, and it is the reason nobody demands toxicology data before drinking tea.

What it does not establish is effect. Traditional use records that people believed something worked, which is a fact about people rather than about the plant. The history of medicine is substantially a history of remedies used confidently for centuries and found, on testing, to do nothing. That is not a slight on traditional knowledge; it is the reason testing was invented.

There is also a form question hiding inside the argument. Traditional Epimedium use means decoctions and wines made from the whole herb, in quantities measured in grams. A 10:1 ethanol or water concentrate at 8 mg in a capsule is a different preparation at a different amount, and whatever the tradition supports, it does not obviously support that.

What to do with an ingredient like this

A practical test that works on any label in this aisle.

1. Find the highest rung. Ask what kind of study the claim rests on: a dish, an animal, or a person. The answer is usually findable in one search.
2. Watch for clinical language on preclinical evidence. Words like improves, maintains and enhances describe outcomes. Enzyme assays do not produce outcomes.
3. Treat a very confident mechanism as a warning sign. The more precisely a page can describe how something works, the more likely it is that the mechanism is all there is.
4. Check the category's quality record separately. Sexual-health supplements are the most adulterated category on the market, and a product promising a drug-like effect from a botanical is exactly the profile that attracts spiking.
5. Do not conclude it is useless. Untested is untested. The honest position is that nobody knows, and a page telling you either way with confidence has gone past its evidence.

Applied to the rest of this bottle, the same test produces a mixed answer: two ingredients with real human trials, three with chemistry and no trial, and two in between. The ingredient guide sorts all of them, and the ingredients page does it row by row for this label specifically.

References