Semax has something most research peptides lack: a country where it is a registered medicine, and a published clinical literature going back to the 1990s. That makes its side-effect record look as if it should be full. It is not. When FDA's reviewers worked through everything they could find for the July 2026 compounding advisory committee, they concluded that the clinical information was insufficient to characterise semax's safety at all. This page sets out what that review found, what FDA's adverse-event database holds tonight, and the two animal findings FDA singled out. It reports; it does not reassure or warn, and a decision about use or about a symptom belongs with a clinician. What doses the literature describes, and why its percentages will not convert to a vial, is on our semax dosage page; how semax sits among the other cognitive peptides, and how the committee voted against FDA's recommendation, is on our cognitive peptides page. How brands appear on this site is on our disclosure page.
The review this page rests on
The source is FDA's briefing document on semax-related bulk drug substances for the Pharmacy Compounding Advisory Committee meeting of July 23–24, 2026 (FDA briefing document, evaluation dated 2026-05-11). Its reviewers searched PubMed, Embase, the Cochrane database, FAERS and ClinicalTrials.gov and read the nominators' submissions. Two limits shape everything below. The references do not say whether the semax used was the free base or the acetate salt, which are the two forms nominated. And a large Russian-language literature was set aside for lack of verified translations, so the record is what an English-reading regulator could check — which is also what an English-reading buyer can check.
Eight studies, one that reported on side effects
FDA's safety table lists every reference in which a person received semax. Reduced to what each says about harm:
| Study (as FDA lists it) | Who | What was given | Adverse events |
|---|---|---|---|
| Koroleva 1996 | 37 adults with trigeminal neuralgia, dental pain or migraine | Intranasal, 0.5 mg/kg once | Authors reported none |
| Kaplan 1996 | 19 healthy men | Intranasal, 1 mg daily for 2 days, or 0.25 mg once | Not discussed |
| Cherkasova 2002 | Adults with chronic ischaemic brain disease (number not given) | Intranasal, printed as "600 mg daily for 10 days" | Not discussed |
| Ivanikov 2002 | 32 adults with peptic ulcer | 1% drops, 2–4 per nostril three times a day, 10 days, plus standard ulcer drugs | Not discussed |
| Proskurina 2018 | 120 children aged 12–14 with depression | 0.1% semax plus cognitive behavioural therapy and physiotherapy | Not discussed |
| Aminova 2016 | 331 children with tics and Tourette syndrome | Semax with another nootropic and neuroleptics | Not discussed |
| Lebedeva 2018 | 14 healthy adults | Intranasal, 1.2 mg | Not discussed |
| Panikratova 2020 | 14 healthy adults (possibly the same people) | Intranasal, 1.2 mg | Not discussed |
Three things stand out once the rows are read together. Seven of eight references say nothing about adverse events — not that there were none, but nothing. Most of the people documented were children: 451 of roughly 550, in two studies that were meeting abstracts without full reports, and in both semax was one of several treatments given at once, so any effect, good or bad, cannot be separated out. And the Cherkasova figure of 600 mg a day, which FDA's table reproduces as printed, is far out of line with the rest of the table — the healthy-volunteer doses are 1 to 1.2 mg, and even the single 0.5 mg/kg dose is about 35 mg for a 70 kg adult; the same figure appears in a separate indexed abstract where it cannot describe a 0.1% nasal solution, as our dosage page sets out. A figure reproduced in a regulator's table is still an unresolved unit, not a dose.
FDA's conclusion from the table: most references "did not discuss safety", no study assessed immune responses or aggregation, and "there are no safety data" for the subcutaneous route at all.
What FDA's adverse-event database holds
FDA's Office of Surveillance and Epidemiology searched FAERS through 2025-12-03 and found one report. A consumer reported that in May 2024 they had ocular pain and eye burning after using semax 0.1% nasal drops bought online; they reported a hospital stay, and said the pain had not resolved when they filed in May 2025. FDA's literature searches through the same date found no published case report of harm.
We ran the openFDA drug-event query ourselves on 2026-09-27 (data exported 2026-09-23). It returns two records with "semax" in the product field. One is the eye-pain report (safety report 25343150, received 2025-05-20), which lists eighteen other products alongside semax, among them naltrexone, pregabalin, alprazolam, tizanidine and heparin cream, and codes the reactions as "eye pain" and "product advertising issue". The other, received 2021-12-15, names a product the database's dictionary codes as "SEMAX [CITALOPRAM HYDROBROMIDE]" — the antidepressant citalopram — in a list of fourteen other medicines, most of them psychiatric, with drowsiness, breathlessness and a prolonged QT interval. Whether the reporter meant the peptide cannot be told from the record, and FDA's own count excluded it. A count built by searching the name alone would have doubled the number of semax reports. Either way the total is tiny, and it measures reporting, not harm: research-market buyers and 503A compounders rarely file.
The two animal findings FDA flagged
With so little human data, FDA's safety discussion leans on pharmacology, and two findings recur in its conclusions.
Clotting. In rats, semax at 1 mg/kg by the nose or a vein increased fibrin-dissolving activity and tissue plasminogen activator and reduced platelet aggregation by about 35%; in another rat study, semax-treated animals were less prone to clot than untreated ones. Its breakdown product Pro-Gly-Pro showed anticoagulant activity of its own. One human abstract (Cherkasova 2002) described possible anti-thrombotic properties. FDA's reviewers wrote that this "raises concern about the risk of bleeding, particularly in certain populations at risk for bleeding or if combined with other medications that increase bleeding risk", and that compounded products carry no labelling to warn of it. No study has measured bleeding in people.
Dopamine. In mice, semax raised brain serotonin turnover — the proposed basis for its pain-relieving effect — but also "potentiated amphetamine-induced dopamine release in the striatum and amphetamine-induced locomotor activity". FDA called this "concerning because increased dopaminergic tone in the striatum is a response typically induced by drugs of abuse such as cocaine", and noted that nothing had been published on semax's abuse potential. It is one animal finding, not a demonstration of addiction; it is also the only data on the question.
The toxicology file
FDA found no acute toxicity study, no repeat-dose toxicity study, no genotoxicity battery and no developmental or reproductive study of semax. The one cancer-related study it found gave semax to female mice of a strain bred to develop mammary tumours, late in life and for less than three months on average; the tumours were smaller and the mice lived longer, but FDA judged the design inadequate to say anything about carcinogenic potential either way. For a peptide nominated for daily injection, FDA also restated its general immunogenicity concern: peptide aggregation and synthesis impurities can provoke immune responses that range from harmless antibodies to severe reactions, "enhanced when peptides are given via the SC or nasal spray" route, and there were "insufficient data to conclude" that semax does not present that risk. The broader pattern across research peptides is on our peptide side effects overview; its closest comparator, Selank, has a record of the same shape, covered on our Selank dosage page.
Where the online side-effect lists come from
Semax pages commonly list nasal irritation, headache, anxiety or restlessness, and sleep changes. Nasal irritation is plausible for any nasal drop, but we could not open the Russian state drug register or its leaflet from here to check, so this page does not state what it says. The others do not trace to any adverse-event table FDA or this census found. A list that cites human data for them is citing something the regulator could not locate.
What the record does not contain
Any human pharmacokinetic study, by any route; any study of subcutaneous injection; any adverse-event table from a controlled trial; any registered trial on ClinicalTrials.gov; any study of bleeding, abuse potential or immune response in people; any use in pregnancy; a standard toxicology package; and any analysis of what research-market vials or imported drops actually contain. Semax has no FDA approval, and the committee's July 2026 vote in favour of listing it is advisory, not an approval. Anyone weighing its use — especially alongside blood thinners, stimulants or psychiatric medicines — or wondering whether a symptom is related to it, should take the question to a clinician who can see their history.
Sources and dates
Read 2026-09-27: FDA briefing document for semax-related bulk drug substances, July 23–24, 2026 PCAC (evaluation dated 2026-05-11) — recognition, nonclinical pharmacology and toxicology, human safety, immunogenicity and conclusion sections, including Table 3; openFDA drug event endpoint, search on "semax" (two records, both read in full). We could not reach the Russian drug register or a Russian drug-reference site from here, so no Russian label wording is quoted. Corrections go to the contact page.