Semax: the monograph

Semax is a synthetic chain of seven amino acids, Met-Glu-His-Phe-Pro-Gly-Pro, with the molecular formula C37H51N9O10S and a molecular weight of 813.9 ⁵. It is built by taking the fragment of adrenocorticotropic hormone known as ACTH(4-7) and attaching the tripeptide Pro-Gly-Pro to its tail, which is why the literature calls it both an ACTH(4-10) analogue and ACTH(4-7)PGP. Those two names describe the same molecule ^5,33,26.

It was designed in the Soviet Union and then Russia as a neuroprotective and nootropic drug, and it kept the peptide backbone of a hormone while shedding the hormone's activity: the indexed literature describes it as completely devoid of hormonal activity ²⁴. It is given intranasally, as drops or spray, and in the Russian ophthalmology literature by endonasal electrophoresis ^17,15,21.

In the United States semax is not a medicine. No FDA-approved product contains it, FDA lists it among bulk substances that may present significant safety risks in compounding, and in July 2026 FDA proposed that it not be added to the list of substances pharmacists may compound with ^3,1,2. It reaches American buyers as a research chemical.

This monograph grades every claim by the species it was tested in and, for human studies, by the language it was published in. That second column matters more here than for almost any other compound in this fleet, and the section below explains why.

Class: synthetic heptapeptide, an analogue of ACTH(4-10) with Pro-Gly-Pro replacing residues 8 to 10 ^5,26.

US status: not FDA approved for any use; on FDA's compounding safety-risks list; FDA proposed in July 2026 that it not be added to the 503A Bulks List ^3,1,2.

Route: intranasal in essentially all published work ^17,15.

Evidence base: 206 records in Europe PMC, 83 of them Russian-language, 50 indexed with an English abstract only, and exactly one carrying the Randomized Controlled Trial tag ⁷.

Human pharmacokinetics: none published ⁸.

Registered trials: none on ClinicalTrials.gov ⁴.

Semax is not FDA approved for any use in the U.S. It is sold as a research chemical; FDA has proposed not to allow it in compounded drugs, and efficacy has not been established to FDA standards.

That sentence is the whole US position, and each half of it is separately checkable. A query of FDA's Drugs@FDA database for semax as an active ingredient returns no matches at all ³. FDA lists it on its page of bulk drug substances that may present significant safety risks when used in compounding, under the name Semax (heptapeptide), stating verbatim: "Compounded drugs containing semax (heptapeptide) may pose risk for immunogenicity for certain routes of administration due to the potential for aggregation and peptide-related impurities. FDA has no, or limited, safety-related information for proposed routes of administration. Therefore, the agency lacks sufficient information to know whether the drug would cause harm if administered to humans." ¹.

One precision note, because the peptide pages on the rest of the internet blur this. FDA's semax entry does not accuse the substance of poor API characterization, and it does not say FDA found no human exposure data. Those sentences appear on the same FDA page against other peptides such as PEG-MGF, MOTS-c and TB-500, and it would be easy to borrow them. FDA's actual complaint about semax is narrower: immunogenicity risk from aggregation and peptide-related impurities, and too little safety information about the routes people propose to use ¹.

Nothing about this status is a statement that semax does not work. It is a statement that the American regulator has not been shown enough to say either way.

On July 23 and 24, 2026, FDA's Pharmacy Compounding Advisory Committee reviewed seven peptides for possible inclusion on the 503A Bulks List, the roster of bulk substances a compounding pharmacist is permitted to use. Semax was the last item on the agenda. FDA's briefing document states, verbatim, as points 13 and 14: "FDA is proposing that Semax (free base) NOT be included on the 503A Bulks List." and "FDA is proposing that Semax acetate NOT be included on the 503A Bulks List." ².

The detail almost nobody reports is in the footnotes. Both semax nominations had already been withdrawn by the people who submitted them, LDT Health Solutions on behalf of the International Peptide Society, and Wells Pharmacy Network. FDA brought the substance to the committee anyway, writing: "This nomination was withdrawn by the nominator (FDA-2015-N-3534-0485/ FDA-2015-N-3534-0487). However, FDA is electing to proceed with the presentation of Semax-related bulk drug substances (Semax (free base) and Semax acetate) to the PCAC." ². The industry that wanted semax on the list stopped asking, and the regulator finished the sentence regardless.

FDA judges these nominations against four criteria from its February 2019 final rule: "(1) The physical and chemical characterization of the substance; (2) Any safety issues raised by the use of the substance in compounded drug products; (3) The available evidence of the effectiveness or lack of effectiveness of a drug product compounded with the substance, if any such evidence exists; and (4) Historical use of the substance in compounded drug products", weighed against each other in a balancing test ².

Two things this does not mean. It does not mean semax was declared unsafe, and it does not mean the opposite either: the briefing document is an agenda and points-to-consider paper that links out to separate FDA review documents, and it does not contain FDA's reasoning about semax. This site could not resolve those linked reviews from the document, so it does not attribute any specific rationale to the proposal beyond FDA's own safety-risks wording ². It also does not mean the question is closed: FDA states that it "does not intend to issue a final determination on the issues at hand until input from the advisory committee process has been considered and all reviews have been finalized" ².

And if semax had been added to the list, it still would not have become an approved medicine. The 503A Bulks List governs what a compounding pharmacist may work with; it sits inside the exemptions from new drug approval, from labeling with adequate directions for use, and from current good manufacturing practice requirements ². Listing is permission to compound, not evidence of efficacy.

This is the single most repeated claim about semax, and it deserves a careful answer rather than a confident one.

What is verifiable from sources this site fetched and read: the peer-reviewed literature indexed in PubMed repeatedly describes semax as a Russian drug used clinically in Russia. A 2000 paper in Vestnik oftalmologii is titled "Evaluation of therapeutic effect of new Russian drug semax in optic nerve disease"; a 2001 paper in the same journal calls it "a new Russian neuropeptide" used in glaucoma patients; a 2001 paper in Brain Research calls it "a novel Russian compound"; and a 2020 paper in the journal Genes states that semax "has been used successfully in the treatment of patients with severe impairment of cerebral blood circulation" ^19,20,31,29. A 1997 review from the group that developed it describes fifteen years of design and study and calls semax "one of the rare analogues of regulatory peptides which underwent all stages from fundamental investigations to practical usage" ²⁵.

What this site did not verify, and therefore does not assert: that semax holds a current Russian marketing authorisation. The obvious way to check that is the Russian state register of medicines. We tried. The register's search interface requires an interactive postback; a direct query returned the empty search shell after consuming its full time budget and contained no semax record at all ⁶. So the registration itself is unconfirmed here.

The honest formulation is therefore narrow and worth stating plainly: semax is described in the peer-reviewed literature as a Russian clinical drug, and this site could not independently confirm its registration status against the Russian regulator. That gap is not a hint that the claim is false. It is a refusal to launder a widely repeated claim into a verified one, which is the same discipline our sister site applies to selank ^19,6.

None of it changes the American position. A drug in clinical use in one country is not approved in another, and the two facts sit side by side without contradiction.

Most compounds in this fleet have a thin evidence base. Semax has a thick one that most readers cannot read, which is a different problem and a more interesting one.

A Europe PMC search for semax in the title or abstract returns 206 records ⁷. Of those, 83 are flagged Russian-language and 120 English, so Russian-language work is 40.3 percent of everything published on this peptide ⁷. Fifty records carry the PubMed publication type English Abstract, the tag applied when a non-English article is indexed with an English abstract and nothing more ⁷. For those fifty papers, a reader without Russian gets the authors' summary of their own results and never sees the methods, the results tables, the group sizes or the limitations.

The literature also has a shape in time: 18 records from the 1990s, 91 from the 2000s, 62 from the 2010s and 35 from the 2020s ⁷. The peak was twenty years ago.

Then there is the publication-type census, which is the number that ends most arguments. Across all 206 records, 3 carry the Clinical Trial tag, 2 carry Controlled Clinical Trial, and exactly 1 carries Randomized Controlled Trial ⁷. One. And that one is not a stroke trial or a cognition trial: it is a study of chronic partial denervation in 27 patients with motor neuron disease, published in Russian in 2007 ¹³.

A small honesty note about our own number, because a round count is easy to hide behind. Not all 206 records are about the peptide. At least one indexed paper, on respiratory acoustic thoracic imaging, uses semax as the name of a statistical variable and writes phrases like "semax = 0.005" ³⁵. We report the raw search count and the false positives rather than quietly filtering to a tidier figure, because the filtering is exactly where a citation count becomes a marketing number.

The table below is the whole evidence base this site is willing to stand behind, with a species column on every row and a language column on every human row.

Read from the top, the pattern is consistent: a moderately large body of Russian clinical work in stroke, cerebrovascular disease and ophthalmology, almost none of it randomized or blinded and most of it readable only as an abstract; two placebo-controlled brain-imaging studies in healthy volunteers, published in English, that measure network topography rather than any clinical outcome; a mechanistic animal literature in rats and mice that is genuinely substantial and genuinely modern, including 2024 and 2025 papers in mainstream Western journals; and then four rows that are documented absences.

The absences are the load-bearing part. There is no double-blind semax study, no meta-analysis, no systematic review and no Cochrane review anywhere in the indexed literature ⁹. There is no published human pharmacokinetic study ⁸. There is no registered trial on ClinicalTrials.gov ⁴. There is no dedicated human safety or toxicity study ¹⁰. Those four rows are printed in the table with the same weight as the positive findings, because a reader deciding what to believe needs them just as much.

Stroke is the indication semax is genuinely built around, and it is where the honest grading matters most.

The human side is real but thin by Western trial standards. The 1997 study that anchors the reputation gave semax to 30 patients in the acute phase of hemispheric ischemic stroke and compared them with 80 patients on conventional therapy, reporting faster regression of cerebral and focal disorders, especially motor disorders ¹¹. It was not randomized, was not blinded, and its abstract carries no effect size, no confidence interval and no p value. The 2018 rehabilitation study followed 110 post-stroke patients on a regimen of 6000 micrograms per day for ten days repeated after a twenty-day gap, and reported higher plasma BDNF and faster Barthel index improvement in the semax subgroups; again, allocation is not described as randomized and no blinding is described ¹². A 2005 report on 187 patients with cerebrovascular insufficiency describes reduced risk of stroke and transient ischemic attacks, but gives no control-group composition, so the comparison behind that risk claim cannot be checked ¹⁴.

The animal side is stronger, more recent and much more transparent. In a rat transient middle cerebral artery occlusion model, RNA sequencing found 394 genes differentially expressed at 24 hours under semax versus saline, with inflammation-related genes suppressed and neurotransmission-related genes activated, precisely the reverse of what ischemia-reperfusion did on its own ²⁹. A 2024 follow-up in the penumbra-associated frontal cortex found ischemia disrupting 3774 genes and semax shifting 1539 of them back toward the uninjured profile ³⁰. An older rat study showed semax abating the ischemia-driven rise in nitric oxide and restoring neurological function where glycine failed ³¹.

So the fair summary is this: the mechanism for a stroke effect is well developed in rats, and the human evidence that it translates is non-randomized, largely unblinded, mostly untranslated, and never once replicated in a Western trial. Both halves of that sentence are true at the same time, and a page that gives you only one half is selling something.

Semax is sold as a cognitive enhancer. Here is what the indexed literature actually contains on that point.

In healthy people, two placebo-controlled studies exist, both using resting-state functional MRI, both published in English. The 2018 study imaged 24 healthy volunteers, 14 on intranasal semax and 10 on placebo, and found a larger rostral subcomponent of the default mode network in the semax group ¹⁵. The 2020 study looked at 52 healthy participants across semax, selank and placebo and found connectivity differences involving the right amygdala and neighbouring temporal regions ¹⁶. These are the best-designed human semax studies in existence, and it is important to be exact about what they measured: brain network topography, minutes after a dose. Neither reported a memory score, an attention test result, a reaction time, or any clinical outcome.

The claims about memory and attention themselves trace mostly to the developer group's own 1997 narrative review, which asserts improved operative memory and attention in animals and in humans, effects lasting 20 to 24 hours after intranasal doses of 0.015 to 0.050 milligrams per kilogram, and no side effects in any case ²⁵. That is a review written by the people who created the compound, not a controlled trial, and the no-side-effects-in-any-case claim is exactly the kind of statement a controlled study exists to test.

The mechanistic backing is rat work, and it is good rat work. A single 50 microgram per kilogram dose raised hippocampal BDNF protein 1.4-fold, trkB phosphorylation 1.6-fold, exon III BDNF mRNA 3-fold and trkB mRNA 2-fold, alongside more conditioned avoidance reactions ²⁶. Semax binds specifically to basal forebrain membranes and raises BDNF there after intranasal dosing ²⁷, and changes NGF and BDNF gene expression in vivo within an hour ²⁸. When a vendor page says semax raises BDNF, this is the evidence, and it is measured in rat brain tissue, which nobody can measure in a customer.

Three claims travel with semax across supplement sites. Each one has a traceable origin, and the origins are not what the claims imply.

"Studied for ADHD." This traces to a 2007 paper proposing semax as a candidate for attention-deficit hyperactivity disorder and Rett syndrome. It was published in Medical Hypotheses, a journal whose purpose is to publish untested ideas, and it reports no trial and no patients of its own; its own language is that "it is proposed in this paper that Semax may have good therapeutic potential in ADHD" ²⁴. A hypothesis is a reasonable thing to publish. It is not a study, and a page that lists it among the evidence for ADHD is miscounting a suggestion as a result.

"Raises BDNF." True, in rats, at 50 micrograms per kilogram, measured in hippocampal and basal forebrain tissue ^26,27. In humans, the one measurement of a BDNF-related outcome is plasma BDNF in a non-randomized Russian rehabilitation study ¹². Plasma BDNF and brain-tissue BDNF are not interchangeable, and no human study has measured the latter.

"Decades of clinical research." Also true in a literal sense, and the shape of those decades is the point: 206 indexed records, peaking in the 2000s, 40 percent of them Russian-language, one carrying a randomized-controlled-trial tag, none double-blind, none replicated in the West, and zero registered on ClinicalTrials.gov ^7,9,4. Decades of research and decades of confirmed results are different sentences.

There is also a set of indexed patent documents in the semax literature covering treatment methods in children and other applications ⁷. A patent describes a claimed invention. It is not evidence that the invention works.

Thirteen human study rows appear in the table above. Ten of them come from Russian-language publications and three from English-language ones. Beyond stroke and cerebrovascular disease, the human literature scatters across indications in a pattern that is itself informative: optic nerve disease, glaucomatous optic neuropathy, diabetic retinopathy, optic neuropathy of vascular origin, posthypoxic encephalopathy, motor neuron disease, refractory peptic ulcer, and psoriasis with metabolic syndrome ^{19,20,21,23,18,13,17,22}.

Several of those studies cannot support a semax-specific conclusion even on their own terms. The 2022 optic neuropathy study delivered semax as one of four simultaneous interventions alongside transcranial magnetic electrical stimulation, an oxybaric chamber and acupuncture, so nothing in its result can be attributed to the peptide ²³. The 2002 peptic ulcer report gives a striking healing difference, 89.5 percent versus 30.8 percent at day 14, but never states how many patients were enrolled, and its authors close by asking for clinical studies of the very activity they just reported ¹⁷. The 2015 psoriasis study states a dose of 600 mg per day of a 0.1 percent solution, which is not reconcilable with the microgram-scale regimens reported everywhere else in this literature; we reproduce it as published rather than silently correcting a source ²².

What does not exist is easier to list. No double-blind study. No meta-analysis. No systematic review. No Cochrane review. No registered trial. No human pharmacokinetic study. No dedicated human safety study. No Western replication of any finding ^9,4,8,10.

There is no human pharmacokinetic profile of semax. A search of the indexed literature for semax together with pharmacokinetic returns zero records, and the single record pairing semax with half-life or bioavailability is a rat forebrain membrane-binding study ⁸.

This is not a small gap. Nobody has published how much of an intranasal dose reaches a human bloodstream, how long it stays, what its half-life is, or what plasma concentration any of the regimens in the table above actually produce. The peptide is short and peptides of that size are typically degraded quickly, and a good deal of the Russian chemistry literature is devoted to making semax analogues that resist proteolysis, which tells you the parent compound's stability was considered a problem worth solving. But that is inference, not a measured human profile, and this site will not print inference as a number.

What can be said with sources is limited to animals: after intranasal dosing in rats, semax reaches brain tissue in quantities sufficient to bind basal forebrain membranes and change BDNF and NGF gene expression within an hour ^27,28. That is a demonstration that it gets in, in a rat. It is not a pharmacokinetic profile, and it does not transfer to people.

Semax is an intranasal compound in essentially all published work. The concentrations that appear in the literature are 0.1 percent and 1 percent aqueous solutions, given as nasal drops or spray ^17,15,21,13.

The Russian ophthalmology studies use a second route that has no equivalent in the consumer market: endonasal electrophoresis, in which the solution is driven across the nasal mucosa with a small electric current in a clinic ^21,23. In the 2013 diabetic retinopathy study, that route outperformed simple intranasal instillation of the same drug, with effects lasting up to twelve months ²¹. Anyone citing those ophthalmology results as evidence for a nasal spray bought online is citing a different intervention.

Injectable semax is sold by research-chemical vendors. This site found no published human study administering semax by injection; the human literature it located is intranasal or endonasal throughout. FDA's own safety concern is explicitly route-conditional, noting immunogenicity risk "for certain routes of administration" ¹.

There is no established human dose of semax, because there is no approved product anywhere this site can verify and no dose-finding programme has produced one ^3,4.

The regimens that appear in the published literature do not agree with each other. The 2018 stroke rehabilitation study used 6000 micrograms per day for ten days, repeated after a twenty-day interval ¹². The 2002 peptic ulcer report used a 1 percent solution, 2 to 4 drops three times daily for ten days ¹⁷. The developer group's 1997 review describes doses of 0.015 to 0.050 milligrams per kilogram ²⁵. The 2015 psoriasis study states 600 mg per day, which is three orders of magnitude away from the others and is very likely a units error in the source ²².

Those numbers span a range too wide to reconcile, and the tool on this site presents each one attached to the study it came from rather than blending them into a recommendation. A regimen used in a Russian hospital in a non-randomized study is a historical fact about that study. It is not a dose, and this site does not convert it into one.

The honest answer is that nobody has established this to a standard worth relying on, and the confident safety claims attached to semax come from a narrower evidence base than they appear to.

No dedicated human safety or toxicity study of semax was located. Ten indexed records pair semax with toxicity, adverse events or safety, and every one is chemistry, in vitro work, animal work or a narrative review ¹⁰. There is no long-term human safety data at any dose, and there cannot be, because there is no long-term human study.

The frequently quoted reassurance, that semax produces no side effects, traces to the developer group's own 1997 narrative review, which states that in the majority of cases the peptide had positive effects and "in no case it produced negative side actions or complications connected with its administration" ²⁵. That is a claim made by the compound's originators in a review article, and its structure, an absence of harm observed in uncontrolled use, is the weakest form such a claim can take.

Against it sits a specific observation from the human literature that vendor pages do not carry. In a 1999 Russian study of 73 patients with posthypoxic encephalopathy, EEG showed episodes of paroxysmal activity after semax injection in some cases, and the authors concluded that the first dose should be given while monitoring the brain's bioelectric activity ¹⁸. A recommendation to give a first dose under EEG monitoring is not the profile of a substance with no effects worth watching.

FDA's position is a third kind of statement. It concerns the product rather than the molecule: compounded semax "may pose risk for immunogenicity for certain routes of administration due to the potential for aggregation and peptide-related impurities", and the agency "lacks sufficient information to know whether the drug would cause harm if administered to humans" ¹.

There is no adverse-event table for semax of the kind an approved drug's label carries, because there is no label and no trial programme that systematically collected one.

What the indexed literature contains is the paroxysmal EEG activity noted after dosing in some patients with posthypoxic encephalopathy, which prompted its authors to advise EEG monitoring for the first dose ¹⁸, and the developer group's blanket statement that no negative side actions or complications were observed across their fifteen years of work ²⁵. Those two statements come from different kinds of source and should not be averaged.

The risks FDA names are product risks rather than reported reactions: immunogenicity from aggregation and peptide-related impurities in compounded preparations ¹. For material bought from a research-chemical vendor, with no approved product and therefore no product-level quality standard, those risks are unquantified rather than absent ^3,1.

No interaction study of semax was located in the indexed literature, so this section is a list of unknowns rather than a list of interactions.

The populations who appear in the human literature are adults with stroke, cerebrovascular disease, eye disease, encephalopathy, motor neuron disease, peptic ulcer and psoriasis, in Russian clinical settings ^{11,12,21,18,13,17,22}. There is no published study in pregnancy, in breastfeeding, in children outside patent documents, in liver or kidney impairment, or in healthy people taking it long term for cognitive enhancement, which is the way it is actually marketed ⁷.

One combination is worth naming because the literature raises it in animals: semax has been reported to augment the effects of psychostimulants on central dopamine release ²⁴. That report appears in the same hypothesis paper discussed above and is not an interaction study, but it is the closest thing to a drug-interaction signal in this literature and it points at a widely used drug class.

It comes from the research-chemical market, and that market has no product-level quality standard behind it. Because no semax product holds an FDA approval, nothing about a vial or a spray bottle sold to an American buyer has been reviewed for identity, purity, strength or sterility ³.

FDA's compounding safety-risks entry names the specific technical worries: aggregation and peptide-related impurities, which are the failure modes that carry immunogenicity risk with a peptide ¹. Those are manufacturing-quality problems, and they are the reason a certificate of analysis supplied by the seller is not the same thing as a quality standard.

With FDA proposing that semax stay off the 503A Bulks List, the compounding pharmacy route is the one that was under discussion in July 2026 ². That leaves no regulated US supply path, and the material circulating is what it has always been: a research chemical sold with a label that says so.

This site has no verified storage instruction for semax, because storage instructions come from an approved product's label and no such label exists ³.

What the literature indicates indirectly is that stability was a real concern for the chemists working on this peptide: a recurring line of Russian research produced N-terminally modified and acetylated semax analogues aimed at resisting proteolysis in biological media ⁷. Work to make a molecule last longer is evidence that the original did not last long.

Any specific temperature, shelf life or reconstitution instruction on a vendor page is the vendor's assertion. It is not a pharmacopoeial standard and this site does not repeat it.

This section is the honest core of the monograph and it should be read before anything else on the site is acted on.