What does human evidence show about Semax?

Semax has been given to people. That does not mean the popular cognitive claims have been demonstrated.

The human record includes Russian stroke reports, two short-term brain-imaging papers in healthy volunteers, an uncontrolled pain report and a small ulcer study. Some reported promising signals. None provides a replicated, well-controlled demonstration that Semax improves memory, focus or everyday cognition in healthy people.

FDA’s July 2026 review found no human pharmacokinetic study by any route and insufficient evidence for cerebral ischemia, migraine or trigeminal neuralgia. An advisory committee then recommended possible section 503A listing. That vote changed the regulatory debate—not the clinical evidence.

Human evidence is more than human exposure

Three questions are often collapsed into one:

1. Has a person received Semax in a study?
2. Did the study detect a biological or clinical difference?
3. Does the design show that Semax caused a useful benefit?

For Semax, the answer to the first question is yes. The second is mixed and outcome-specific. The third remains unresolved.

A brain-imaging measure can differ without a person thinking better. A blood biomarker can move without recovery improving because of it. A treated group can outperform a comparison group even when treatment selection, rehabilitation timing or background care created the difference.

That is why human study and human proof are not synonyms.

What is Semax?

Semax is a synthetic seven-amino-acid peptide derived from an ACTH fragment. It is commonly described by the sequence Met-Glu-His-Phe-Pro-Gly-Pro.

FDA evaluated two related but distinct bulk drug substances: Semax free base and Semax acetate. Many publications name only Semax, so it is not always possible to tell which chemical form was studied. Product identity, purity, delivery device and finished formulation can also differ.

Semax is registered as a drug in Russia. It is not an FDA-approved drug in the United States. Russian registration, US drug approval and eligibility to use a bulk substance in certain compounded drugs are three different regulatory states.

What human studies actually measured

The stroke literature reports signals, but the methods remain thin

Stroke is where Semax has the most visible human clinical history.

A 1997 report described 30 people who received Semax during acute hemispheric ischemic stroke and 80 conventionally treated controls with strokes described as similar in severity and location. Its English abstract reported faster regression of neurological problems, especially motor impairment.

The abstract does not say allocation was randomized or blinded. It does not provide a prespecified primary endpoint, a complete comparative estimate or the information needed to reproduce the analysis. Calling this a modern randomized controlled trial goes beyond the accessible record.

A 2018 paper followed 110 people after ischemic stroke. Participants were split by earlier or later rehabilitation and then by whether they received Semax. The authors reported higher plasma BDNF and faster or better recovery on motor and daily-function measures in association with Semax and earlier rehabilitation.

That is more clinically relevant than an animal mechanism. It is still not a clean causal test. The abstract does not describe random assignment, blinding or a placebo. Rehabilitation timing, treatment selection, background care and natural recovery could all influence the result.

FDA reached a narrower evidence set. The agency found additional Russian- language stroke reports, but did not use untranslated full papers in its effectiveness evaluation. Its only usable cerebral-ischemia item was a meeting abstract that lacked relevant clinical-function outcomes.

Both observations can be true: PubMed shows that human stroke cohorts were reported, while the accessible evidence remains inadequate to establish efficacy.

Healthy brains were scanned—not shown to work better

The strongest English-language healthy-volunteer record is a short-term fMRI study of 24 adults. Fourteen received Semax and ten received placebo. Researchers scanned participants before and shortly after administration and reported a difference in part of the brain’s default mode network.

That result shows a measured imaging difference under the study conditions. It does not show better memory, attention, decision quality, productivity or daily function because those were not the reported outcomes.

A later paper examined resting-state connectivity in 52 healthy participants across Semax, Selank and placebo conditions. It reported short-term connectivity differences involving the amygdala and temporal cortex. Again, connectivity was the outcome—not useful cognition.

FDA noted that the 14 Semax recipients in the two imaging reports may have been the same people. Until that overlap is resolved, they should not be counted as two independent Semax cohorts.

An older human EEG paper is also indexed, but PubMed exposes no abstract. The index proves that an EEG study exists. It does not supply enough information to claim a cognitive benefit.

BDNF is the mechanism story—not the human verdict

Much of Semax’s online appeal comes from BDNF, TrkB, neuroplasticity, inflammation and neuroprotection. Those pathways are scientifically relevant. The best-known experiments behind them are largely in rats or cells.

For example, animal studies have reported changes in BDNF signaling, gene expression and inflammatory pathways after experimental cerebral ischemia. Those findings can justify a clinical hypothesis. They cannot reveal whether a healthy person will focus better, whether a stroke patient will regain function or whether a finished human product is safe.

The 2018 rehabilitation paper measured plasma BDNF in people, but a blood biomarker does not by itself prove brain target engagement or explain a functional result. A biomarker becomes useful evidence only when its relation to a meaningful outcome is validated in the relevant setting.

The pain evidence is one small uncontrolled report

FDA found one report covering migraine, typical trigeminal neuralgia and dental plexalgia.

The migraine group contained 12 people and no untreated, placebo or blinded control. Four reported headache cessation; the majority did not report resolution. Baseline migraine details, actual scale values and variation were inadequately reported.

The same report included 16 people with typical trigeminal neuralgia and nine with dental plexalgia. The typical-neuralgia group showed no relevant change. The dental subgroup reported improvement, but the pain scale, units and clinical meaning were unclear. The paper’s authors did not conclude that Semax itself had analgesic activity.

FDA found no additional clinical reference for either use and concluded the evidence was insufficient.

One ulcer study broadens exposure—not cognitive evidence

A 2002 study compared background ulcer therapy with the same therapy plus Semax in 32 adults with refractory peptic ulcers. Its abstract reported more healing in the Semax group and called for further clinical studies.

That preliminary result matters because it documents administered-human exposure outside neurology. It cannot be converted into evidence for memory, stroke recovery or general neuroprotection. Multiple background therapies and a small sample also make the independent contribution of Semax difficult to isolate.

What does the human evidence table show?

The table does not say Semax is biologically inactive. It shows that biological activity and useful clinical benefit have not been connected by strong human evidence.

What remains unresolved

Human pharmacokinetics are missing

FDA found no human pharmacokinetic study of Semax free base or Semax acetate by any route.

That gap matters. Without measured human exposure over time, confident claims about absorption, brain delivery, duration, dose-response or equivalence between forms and products lack a direct human foundation.

The existing brain-imaging differences may show short-term biological activity. They do not substitute for pharmacokinetics and do not establish how much of a characterized substance reached which tissue.

The safety denominator is unclear

FDA located records involving dozens of adults and several hundred children, but much of the safety reporting was missing or difficult to interpret. Some studies were available only as meeting abstracts. Several combined Semax with other treatments. Most did not discuss adverse events.

The agency found one spontaneous consumer report describing severe eye pain and burning after an online-purchased nasal product, hospitalization and symptoms reported as unresolved one year later.

One report cannot show that Semax caused the event or estimate how often it happens. It also cannot create a reassuring safety rate: exposure is unknown, reporting is incomplete and product identity may be uncertain.

FDA additionally raised unresolved concerns about possible bleeding effects, peptide aggregation, impurities and immune reactions. No human study resolved those issues for the free base or acetate. A ClinicalTrials.gov search on 3 August 2026 returned no registered Semax study.

Unknown safety is not proof of danger. It is a reason not to claim safety.

What did the 2026 FDA hearing change?

FDA staff recommended against adding Semax free base and Semax acetate to the section 503A Bulks List. Staff cited weak characterization, insufficient effectiveness evidence and inadequate safety information.

On 24 July 2026, the Pharmacy Compounding Advisory Committee voted 8–5–1 to recommend listing each form.

That vote matters, but its meaning is narrow:

• it was an advisory recommendation to FDA;
• it concerned possible use of bulk substances under section 503A;
• it did not approve a Semax drug;
• it did not validate a nootropic claim;
• it did not add a clinical result;
• it did not itself change the operative list.

The panel disagreed with staff about the policy recommendation. It did not make the evidence gaps disappear.

Does Semax improve cognition in healthy people?

The current published record does not establish that it does.

The fMRI studies reported short-term resting-network differences after administration. An indexed human EEG paper exists, but its PubMed record does not expose enough information to characterize the result. The available record has not demonstrated a meaningful, replicated improvement in memory, attention or daily cognitive performance. Stroke and ulcer findings come from people with medical conditions and cannot be transferred to healthy enhancement.

This is the central correction to the search market: mechanism, foreign registration and human exposure are real. A proven healthy-cognition benefit is not.

Primary sources

• FDA evaluation of Semax free base and Semax acetate
• FDA July 2026 meeting page
• FDA webcast with the Semax votes
• Gusev et al., 1997 acute-stroke report
• Miasoedova et al., 1999 stroke-mechanism report
• Gusev et al., 2018 rehabilitation report
• Lebedeva et al., 2018 default-mode-network study
• Panikratova et al., 2020 connectivity study
• Koroleva et al., 1996 human EEG index
• Ivanikov et al., 2002 peptic-ulcer study
• ClinicalTrials.gov Semax search

Independent editorial summary. The authors, agencies and organizations are not affiliated with LifeScore.