How finasteride works on DHT, and what else that enzyme does

Every claim made for and against this drug traces back to one biochemical step. Understanding the step does not settle the arguments, but it explains why they take the shape they do — and why a question about hair keeps turning into a question about mood.

One enzyme, two types

5-alpha-reductase converts testosterone into dihydrotestosterone. The label describes two distinct isozymes in humans: type I, found mainly in the sebaceous glands of most regions of skin including the scalp, and in the liver; and type II, found mainly in the prostate, seminal vesicles, epididymides and hair follicles, as well as the liver. Type I accounts for roughly a third of circulating DHT; type II for about two-thirds.

Finasteride is described as a competitive and specific inhibitor of type II, with about a hundredfold selectivity for type II over type I in binding studies. Dutasteride inhibits both, which is the whole of the difference between them — covered here.

How much DHT goes, and how fast

Quickly. The label reports a rapid reduction in serum DHT reaching 65% suppression within 24 hours of a single oral dose. The enzyme complex it forms turns over slowly — a half-life of about 30 days for the type II complex — even though the drug itself clears from plasma in hours, with a terminal half-life of around five to six hours in men aged 18 to 60.

That mismatch is worth holding on to. The molecule leaves quickly; the inhibited enzyme does not.

Why less DHT helps a follicle

In men with male pattern hair loss, the label says, the balding scalp contains miniaturised hair follicles and increased amounts of DHT compared with hairy scalp. Administration of finasteride decreases both scalp and serum DHT in these men. By this mechanism, it says, the drug appears to interrupt a key factor in the development of the condition in those genetically predisposed.

Two pieces of honesty are built into that passage. The relative contributions of the scalp reduction and the blood reduction have not been defined. And the language is about interrupting a key factor in people who are predisposed — not about curing a condition or removing its cause. What that looks like as an outcome is in the trial figures.

What the label says does not change

This part is routinely misrepresented in both directions, so it is worth listing exactly.

• Testosterone does not fall. Mean circulating testosterone and estradiol rose by about 15% from baseline, remaining within the physiologic range.
• The drug is not a hormone itself. It has no affinity for the androgen receptor and no androgenic, antiandrogenic, estrogenic, antiestrogenic or progestational effects.
• The pituitary axis was not affected. No clinically meaningful changes were detected in luteinising hormone, follicle-stimulating hormone or prolactin, and in healthy volunteers the LH and FSH response to gonadotropin-releasing hormone was unaltered.
• Nor were several other systems. No effect on cortisol, thyroid-stimulating hormone or thyroxine, none on the plasma lipid profile, and none on bone mineral density.

So "it lowers your testosterone" is not what the label describes. The specific claim the label supports is narrower and stranger: a large reduction in one downstream androgen, with the upstream hormone slightly higher.

The sentence the debate is built on

One line sits in the distribution section of the pharmacokinetics, between the volume of distribution and the semen measurements, and it is the most consequential sentence in the document: finasteride has been found to cross the blood-brain barrier.

The label states it and leaves it there. It does not build a theory on it, and neither will we — but it is the fact that makes the psychiatric reports pharmacologically discussable rather than dismissible, and it is why the regulators' wording about mood has been argued over so carefully. The MHRA's 2026 update added that sexual dysfunction may contribute to mood disorders and has also been reported without mood alterations: two different routes, both in the reports. What is and is not known about persistent symptoms is the page for that argument.

Why one enzyme produced two licences

The type II isozyme sits in the prostate and in hair follicles. Both tissues respond to DHT — the prostate by growing, a susceptible follicle by shrinking — so a single inhibitor has two entirely separate clinical uses, granted five years apart. That is the whole explanation for a fact people find suspicious: that the hair drug and the prostate drug are the same molecule.

It is also the origin of most of the bad statistics in this field. Trials done in older men at five times the hair-loss strength get quoted at young men taking the hair-loss tablet, and findings from the hair-loss trials get quoted back at prostate patients. The label keeps the two apart carefully, and so does this site.

What the label does not measure

This is the honest edge of the mechanism. The label tells you what happened to testosterone, estradiol, LH, FSH, prolactin, cortisol, thyroid hormones, lipids and bone density, because those were measured. It does not tell you what happened in every tissue where 5-alpha-reductase is expressed, because that was not the question the dossier was assembled to answer.

An absence of measurement is not evidence of harm, and pretending otherwise is how this subject goes wrong in one direction. But it is also not the same as a finding of no effect, and pretending otherwise is how it goes wrong in the other. The regulators have been sitting in exactly that gap since 2024, which is why their language keeps getting more careful rather than more confident.

The useful summary

Finasteride holds one enzyme shut. The enzyme's best-known job is making DHT, which is the pressure behind pattern hair loss and behind prostate growth — hence two licences for one molecule. The benefit follows directly from the interruption, which is why it lasts only while the drug is taken. And the open questions all take the same form: what else was that enzyme doing, in tissues nobody was measuring? The label answers part of that and is silent on the rest.

The whole medicine, from licence to warnings, is on the finasteride page.