All compounds

Phenibut

Also sold as β-phenyl-GABA, fenibut, Anvifen, Noofen, Fenibut, phenybut. GABA-B receptor agonist, structurally related to baclofen.

The evidence says don’t (5 of 5)

Phenibut works — it genuinely and noticeably reduces anxiety, which is exactly the problem, because tolerance builds within weeks and stopping after regular use can produce a withdrawal severe enough to require hospitalisation.

What it is

Phenibut is β-phenyl-GABA: the neurotransmitter GABA with a phenyl ring added so it can cross into the brain. It was developed in the Soviet Union in the 1960s at the Leningrad Pedagogical Institute, and it was carried in the Soviet cosmonaut medical kit as a tranquilliser that calmed without dulling performance.

It is a real medicine in the places that developed it. In Russia, Latvia, Ukraine and several other post-Soviet states it is a prescription drug sold as Anvifen, Noofen or Fenibut, used for anxiety and sleep.

In the English-speaking world it occupies a gap. There has never been a centralised European approval, and it has never been approved as a medicine in the United States, the United Kingdom, Canada or Australia — though it is worth being precise here rather than sweeping, because Latvia and Lithuania are EU member states where it is a nationally authorised prescription medicine. The US Food and Drug Administration has stated it does not meet the definition of a dietary supplement either, and has written to companies selling it as one. Australia has placed it under drug control, and it is restricted in the UK.

What that adds up to is a compound with genuine pharmacological weight and a real clinical history, sold in the West by companies that are not medicine manufacturers, to people who have no prescriber watching what happens next.

Structurally, the thing to hold onto is this: phenibut is a close relative of baclofen, a prescription muscle relaxant with a well-documented and dangerous withdrawal syndrome of its own. That family resemblance is the single most useful fact on this page, and it is the one to say out loud in a hospital.

What it actually does

Phenibut is an agonist at the GABA-B receptor, and it also acts at the α2δ subunit of voltage-gated calcium channels — the same site gabapentin and pregabalin work on.

What that feels like is the point. Phenibut reduces anxiety in a way people describe as clean: not sedated, not drunk, not blunted. Social situations get easier. The internal monologue quietens. People who have spent years managing social anxiety often describe the first experience as a revelation — this is what other people have felt the whole time.

We are stating that plainly because the sites that skip it lose the reader immediately. The confidence level belongs with the claim, though: that phenibut reduces anxiety rests on its pharmacology, on decades of prescribing in post-Soviet states, and on consistent user report — not on modern controlled trials, which do not exist. That is real evidence and it is not the strongest kind. What is not in doubt is that it works well enough that people end up dependent on it, which is the reason this page exists.

The effects come on slowly, over hours, which matters more than it sounds. A slow onset makes it easy to conclude nothing is happening and take more — and it means the dose you took is still arriving when you decide.

What the evidence actually shows

The evidence on phenibut is lopsided in an unusual way: we know a great deal about how it goes wrong and comparatively little about how well it works.

On efficacy: the supporting literature is largely Soviet-era, published in Russian, and does not meet modern standards for trial design, blinding or reporting. There is no large modern randomised trial in anxiety. So the strong claim “phenibut reduces anxiety” rests mainly on pharmacology, decades of prescribing in post-Soviet states, and consistent user report — which is real evidence, but not the kind that settles a question.

On harm: the literature is recent, English-language, and consistent. It consists of case reports and case series rather than trials, because you cannot run a randomised trial of dependence. Published reports describe tolerance developing within days to weeks in the people it happened to, dose escalation, and a withdrawal syndrome that presents like benzodiazepine or alcohol withdrawal: severe insomnia, tremor, sweating, agitation, hallucinations and in a number of cases frank psychosis and delirium requiring inpatient care. Two 2024 systematic reviews pulled these together and found the pattern held across dozens of independent reports from unrelated centres.

What nobody knows: how common dependence is among everyone who tries it. Case reports come from people who ended up in hospital. There is no denominator, so anyone who tells you the odds — in either direction — is guessing.

One practical fact worth its own line: phenibut does not appear on a standard urine drug screen. It is not structurally similar to anything those panels test for. A negative screen still tells a clinician something useful about the substances it does test for — but it says nothing whatever about phenibut, and clinicians have been misled by reading it as an all-clear.

The risk profile

The evidence says don’t (5 of 5) — The known downside is severe enough that we will say it plainly rather than hedge.

What going wrong looks like from the inside

The bad outcome is not a bad experience. It is that you stop being able to stop.

It usually goes like this. It works, and it works well, so you use it again. Within a couple of weeks the amount that worked does not work anymore, so you take more, and the gap between doses gets shorter. You start planning around it — keeping some back for a difficult day, noticing how much is left. Then a day comes when you do not take it, and the anxiety that comes back is far worse than the anxiety you started with. That is the moment that decides things, because it is almost impossible to read correctly from the inside. It feels like proof that you needed it. It is withdrawal.

That is the point where it stops being a decision. What follows in the published cases includes insomnia that does not respond to anything, tremor, sweating, a heart rate that will not settle, and agitation severe enough that people have arrived in emergency departments unable to be calmed, occasionally psychotic, occasionally with no memory of how they got there.

The other thing that catches people is the slow onset. Because it takes hours to arrive, it is easy to decide it is not working and take more. Then all of it lands at once, often after alcohol has been added on the assumption the phenibut did nothing.

If you are reading the escalation pattern above and recognising it, that is the thing to act on, and now is better than later — this gets harder the longer it runs, not easier. Do not stop abruptly on your own. Stopping phenibut explains why, and what to do instead.

Interactions and combinations

This is where phenibut contributes to deaths, and it is worth being blunt about it.

Phenibut is a central nervous system depressant. Combined with other depressants — alcohol above all, but also benzodiazepines, opioids, GHB, gabapentin, pregabalin, sedating antihistamines or zopiclone — the sedative effects can more than add up, and what can stop is breathing. In the published fatal cases we could find, another depressant was almost always present as well; that is a pattern across case reports rather than a measured rate, and there is no registry that would give you one. Alcohol is the common one, and it is common precisely because phenibut is taken in social settings where alcohol is also present.

The slow onset makes this worse than it would otherwise be. Someone who takes phenibut, concludes an hour later that it is not working, drinks, and then receives the full effect of both is describing the standard sequence in these reports, not an unusual one.

Baclofen deserves a separate mention. It acts at the same receptor, and clinicians have used it to manage phenibut withdrawal on that basis, with published cases where the substitution worked. Whether there is true cross-tolerance is inferred from the shared receptor and those cases rather than demonstrated directly. That is a fact about supervised treatment, not a suggestion. Substituting one for the other without supervision is dose-guessing between two drugs with dangerous withdrawal syndromes.

If you’re already using it

If you use it occasionally and are not escalating, the useful thing is simply to keep a record of when you take it. The shift from occasional to regular is gradual and is genuinely hard to see from the inside; a written log shows it when memory does not.

If you are using it regularly — most days, or needing it to feel normal — then the important message on this page is a narrow one: the dangerous move is stopping suddenly, not continuing another day while you arrange help. Withdrawal from a GABA-B agonist is not like stopping caffeine. Published cases include seizures, delirium and psychosis.

Stopping phenibut covers what the published literature says clinicians do, how to find someone to supervise it if you do not have a doctor, and what to expect. It gives no amounts and no schedule, and that is deliberate — coming off this drug without supervision is precisely the thing that goes wrong.

If you are unwell right now, start here. Poison Control will talk to you, free, at any hour, and they are not law enforcement.

If you have been using this regularly and want to stop: read this first. Stopping abruptly is the part that goes wrong.

This page has not yet been reviewed by a clinician. It was written from the published literature and every claim is cited, but no named medical professional has checked it. We say so here rather than let a missing byline read as an implicit one.

Last reviewed 7 September 2026. Found something wrong? Corrections are published, not silently edited.