Peptide library

Semax claims and the evidence:
what studies measured

Few studies have given Semax to people, and each measured something narrow: muscle strength and a blood marker, resting brain activity, rating scales. For each one: the model, the sample, what was measured and the limit.

Short answerCompound research · 6 min read · Updated 21 Nov 2026

Semax is a synthetic peptide of seven amino acids, and its human studies are few and mostly small. One randomly divided 110 adults in rehabilitation into Semax and no-Semax groups, without a placebo. Another, with a placebo group, measured resting brain activity in 52 healthy people. None shows whether, when or for how long an effect occurs in people.

Key takeaways

  • Gusev and colleagues’ 2018 study randomly divided 110 adults in rehabilitation into Semax and no-Semax groups; it had no placebo, and the paper does not say how the randomisation was done.
  • Panikratova and colleagues’ 2020 study enrolled 52 healthy people across three groups and measured resting brain activity, not health outcomes.
  • The 5- and 20-minute brain scans in the 2020 study tell you when researchers looked, not whether or when a meaningful effect begins.
  • In its July 2026 briefing, FDA staff found insufficient evidence of effectiveness for the three proposed uses they could evaluate.

01Which Semax studies randomised people or used a placebo?

Two features make a comparison in people stronger: random assignment, and a placebo group so that participants cannot tell who received the peptide. Among the Semax studies cited here, one has the first and one has the second; we could verify none that has both.

Gusev and colleagues (2018) divided 110 adults in rehabilitation into early and late rehabilitation groups. Their Methods say each group was randomly split into a Semax and a no-Semax subgroup, without saying how. The no-Semax subgroups received no placebo. Researchers measured muscle strength, independence in daily activities and the blood level of a protein called brain-derived neurotrophic factor.

Panikratova and colleagues (2020) had a placebo group beside Semax and Selank groups, in 52 healthy people, but the abstract does not say how participants were assigned. They measured resting brain activity on scans, not a health outcome.

FDA staff also list a 1996 paper by Kaplan and colleagues with 19 healthy men. FDA’s summary does not describe its design, and we could not open the paper’s methods, so this page gives it no tier.

02What each study measured

Read across for the model, sample size, year, measure, and main limitation. Here, n is the number of participants or animals in the comparison described, not necessarily the number in the Semax group.

Measure areaStudy and yearModel and nDesign and tierWhat was measuredMain limitation
Brain activity at restPanikratova et al., 2020Human: 52 healthy participants across all groupsSemax, Selank, or placebo; T2 here because the abstract does not describe assignmentHow closely activity in different brain regions tracked together on functional magnetic resonance imagingBrain scans, not a health outcome. Material sold online is not established to be the same as the study material.
Physical function and a blood markerGusev et al., 2018Human: 110 adults in rehabilitationEarly or late rehabilitation, each with a Semax and a no-Semax group; T2 here because the paper does not explain how it randomized participantsMuscle strength, independence in daily activities, and blood levels of brain-derived neurotrophic factor, a protein measured as a markerNo placebo comparison. The blood marker and the functional scores are different measures. Material sold online is not established to be the same as the study material.
Pain ratingsKoroleva et al., 1996Human: 37 adults across the study populationsOpen-label, no control group; T2Among the measures: pain ratings and electrical responses to facial nerve stimulationWithout a control group, the study cannot separate the contribution of Semax from other explanations. Material sold online is not established to be the same as the study material.
Muscle electrical activity and quality of lifeSerdiuk et al., 2007Human: 27 adultsOpen-label, sequential groups as described in the abstract; T2 hereAmong the measures: muscle electrical activity, functional ratings, and quality of lifeThe abstract does not explain random assignment despite PubMed’s randomized-trial label. Material sold online is not established to be the same as the study material.
Gene activity in brain tissueMedvedeva et al., 2014Animal: male rats, at least 5 at each sampling timeControlled animal experiment; T3Activity of genes in brain tissue after an experimentally blocked arteryGene activity in this rat model is not evidence of an effect in people.
Enzyme activity in serumKost et al., 2001Laboratory: human serum outside the body; sample count not stated in the abstractTest-tube experiment; T4Activity of enzymes that break down peptides called enkephalinsHuman serum in a test tube is not evidence of an effect in people.

03What the study clock can tell you

In Panikratova and colleagues’ 2020 study, the clock marked when the researchers scanned 52 healthy participants across the Semax, Selank, and placebo groups. It did not mark when a meaningful effect began.

  1. BaselineFirst scan

    The researchers recorded resting brain activity for comparison.

  2. 5 minutesSecond scan

    This was the first follow-up imaging time point, not a measure of when participants noticed a change.

  3. 20 minutesThird scan

    This was the last imaging time point in the study, not a measure of how long a meaningful effect would last.

There is a separate question: how Semax levels change in the body over time. That is what a pharmacokinetic study measures. FDA staff reported that their review did not find a human pharmacokinetic study of Semax, including Semax acetate. A scan time is not a measurement of the peptide’s level in the body.

04How to read the evidence

If you searched “semax benefits”, these three checks show what any Semax study can and cannot tell you.

1
Check what the comparison controls for

A placebo comparison and a comparison against rehabilitation alone answer different questions. Read how people were assigned, not just how many groups the paper lists.

2
Check whether the measure matches the claim

A brain scan, a blood marker, and a score for independence in daily activities are not interchangeable. Ask which one the researchers actually recorded before drawing a conclusion.

3
Keep lab documentation separate

A certificate of analysis documents tests on a sample. It does not establish a health outcome or show that the sample is equivalent to material used in a study. See how to read a certificate of analysis.

For the peptide’s identity and regulatory background, see the Semax guide.

05What we don’t know

  • Whether narrow measures tell us anything broader. A blood marker, a strength score and a brain scan in small groups do not show how Semax would affect other people, settings or measures.
  • Whether or when a meaningful effect occurs, and how long it lasts. The 2020 scan schedule and the human pharmacokinetic gap identified by FDA staff do not give a reliable answer.
  • Whether the proposed uses are supported. In its July 2026 briefing, FDA staff found insufficient evidence of effectiveness for the three uses they could evaluate. Only two references were available to them, with “insufficient details on the design and conduct of the studies”.

This is a plain-English summary of studies and public documents checked on 21 November 2026. It is not medical advice.

RefSources

  1. U.S. Food and Drug Administration. [Semax and Semax acetate briefing for the Pharmacy Compounding Advisory Committee.] July 2026, pp. 10–11, 25–29, 38–41, 43–44; includes the review of Koroleva et al. and the search for human pharmacokinetic studies. www.fda.gov/media/193348/download
  2. Kaplan AYA, Kochetova AG, Nezavibatko VN, et al. [Study of Semax in healthy men, listed by FDA staff with 19 participants.] Neuroscience Research Communications, 1996;19(2):115–123. Full text not read. onlinelibrary.wiley.com/doi/abs/10.1002/%28SICI%291520-6769%…
  3. Panikratova YR, Lebedeva IS, Sokolov OY, et al. [Resting-state brain imaging with Semax, Selank, or placebo.] Doklady Biological Sciences, 2020;490:9–11. PMID 32342318. link.springer.com/article/10.1134/S001249662001007X
  4. Gusev EI, Martynov MY, Kostenko EV, et al. [Rehabilitation groups, blood marker measurements, and functional assessments.] Zh Nevrol Psikhiatr Im S S Korsakova, 2018;118(3 Pt 2):61–68. PMID 29798983. DOI 10.17116/jnevro20181183261-68. Allocation wording checked in the publisher’s Russian-language Methods section, not the PubMed abstract. pubmed.ncbi.nlm.nih.gov/29798983/
  5. Koroleva MV, Meizerov EE, Nezavibat’ko VN, et al. [Pain ratings and neurophysiological measurements in adults.] Bulletin of Experimental Biology and Medicine, 1996;122:1107–1109. Sample and design from the FDA briefing. link.springer.com/article/10.1007/BF02447659
  6. Serdiuk AV, Levitskiĭ GN, Miasoedov NF, Skvortsova VI. [Muscle electrical recordings and quality-of-life assessments.] Zh Nevrol Psikhiatr Im S S Korsakova, 2007;107(4):29–39. PMID 18379501. pubmed.ncbi.nlm.nih.gov/18379501/
  7. Medvedeva EV, Dmitrieva VG, Povarova OV, et al. [Gene activity in rat brain tissue.] BMC Genomics, 2014;15:228. Sample wording from Methods, Experimental groups. link.springer.com/article/10.1186/1471-2164-15-228
  8. Kost NV, Sokolov OY, Gabaeva MV, et al. [Enzyme activity in human serum outside the body.] Russian Journal of Bioorganic Chemistry, 2001;27:156–159, English edition. link.springer.com/article/10.1023/A%3A1011373002885

Plain-English summary of public sources, not legal or medical advice. Last checked 21 Nov 2026.

Questions

Was Gusev’s 2018 Semax study randomised?

Its Methods say each rehabilitation group was randomly split into Semax and no-Semax subgroups, without saying how. There was no placebo, so it shows less than a placebo-controlled trial would, and we grade it T2.

Do the scan times in Panikratova’s 2020 study establish an onset?

No. The researchers scanned participants at baseline and at 5 and 20 minutes. Those were imaging measurements, not a record of whether or when participants noticed a meaningful change.

Why is Serdiuk’s 2007 study graded T2 here?

PubMed labels it a randomized controlled trial, but the abstract describes an open-label study in sequential groups and does not explain random assignment. We grade the design described in the abstract rather than the database label.

Does Kost’s 2001 serum experiment count as a human trial?

No. It measured enzyme activity in human serum outside the body. That is laboratory evidence, not evidence of an effect in people.

Evidence before hype.

One email a week: new research explained and regulatory changes. No prices, no hype.

Get the 6:00 brief

Free. Unsubscribe any time.

Written by 6AM Research. Drafted from primary sources; every claim links to its source, and the page is dated. Editorial policy · How we check every page