Noopept is among the most widely sold compounds marketed as a peptide nootropic. Its dose figure circulates freely and is usually given without a source.
There is a source. There is exactly one, it is a translated abstract of an unblinded study, and the useful work on this page is establishing that and then saying what surrounds it.
The census
Run on 18 September 2026, and reproducible by anyone:
| Query | Where | Records |
|---|---|---|
| noopept OR omberacetam | ClinicalTrials.gov | 0 |
| noopept (all fields) | PubMed | 106 |
| GVS-111 (development code) | PubMed | 97 |
| omberacetam | PubMed | 91 |
| N-phenylacetyl-L-prolylglycine | PubMed | 17 |
| noopept AND Humans[MeSH] | PubMed | 28 |
| noopept AND Randomized Controlled Trial[pt] | PubMed | 0 |
| noopept AND Clinical Trial[pt] | PubMed | 3 |
| noopept | FDA Drug Master File list | 0 |
The three clinical-trial-tagged papers are all Russian-language: PMID 22500312 (2011) and PMID 19008801 (2008), both in Zhurnal Nevrologii i Psikhiatrii imeni S.S. Korsakova, and PMID 18318195 (2007) in Eksperimental'naia i Klinicheskaia Farmakologiia.
The figure, and where it comes from
The one dose statement in that set is in the 2011 stroke paper, and its abstract puts it in a single sentence:
"noopept, used in dose 20 mg daily during 2 months, improves cognitive functions in stroke patients and has a high level of safety"
That is the figure. 20 mg daily, two months, PMID 22500312, 2011.
What the study was
The same abstract describes the design, and the description is what qualifies the figure:
- "an open prospective study" — unblinded. Patients and investigators knew who was being treated.
- 60 patients with stroke, followed over 12 months.
- Cognitive function assessed by neuropsychological tests before and after treatment.
- Significant improvement on MMSE and association tasks reported after 2 months against controls.
- Global assessment of efficacy: "mild improvement" in the treated group, no change in controls.
Two things are absent and both matter. The abstract does not state that participants were randomised, and PubMed's index carries no Randomized Controlled Trial tag on the paper. It reports controls without describing in the abstract how they were assigned.
Cognitive testing is among the outcomes most sensitive to expectation — from the patient, and from the assessor scoring the test. An unblinded design is the one least able to separate the drug's effect from the effect of being treated and repeatedly assessed. That does not make the result wrong. It makes it the weakest design that still produces a publishable number, and it is the sole human dose figure this compound has.
The second clinical-trial-tagged paper (PMID 19008801, 2008) adds an EEG signal — increased alpha and beta power, reduced delta, in post-traumatic and vascular cognitive disturbance — and states no dose in its abstract. A change in an EEG rhythm is evidence that something pharmacological is happening. It is not a cognitive outcome, and the paper does not present it as one.
Why zero registered trials is the more important number
Trial registration exists to fix what a study intends to measure before it measures it. Without a registered protocol there is no pre-specified endpoint, no public record of what was planned, and no way to know how many studies were started against how many were published.
Noopept has no record in ClinicalTrials.gov under any of its names. Part of that is historical: its clinical development happened in a Russian research system that predates and sits outside the registration regime, so the absence is not evidence of anything being hidden. The consequence is nonetheless unavoidable — nothing about this compound's human use can be checked against a registered protocol, and the 20 mg figure cannot be traced back to a plan.
This catalogue has met the shape before, and twice with compounds from the same research tradition. Semax and Selank were both developed in the same Moscow institutional world, and both carry human literature that is real, Russian and awkward to convert into a usable figure — Semax because its published numbers are solution concentrations rather than masses, Selank because its randomised evidence sits in two papers in one journal. Noopept is the same tradition at its extreme: the literature is the largest of the three and the registry presence is nil.
What it is, chemically
N-phenylacetyl-L-prolylglycine ethyl ester. PubChem CID 180496, C17H22N2O4, 318.4 g/mol.
Two residues — proline and glycine — with a phenylacetyl group on one terminus and an ethyl ester on the other. Both modifications are load-bearing: they are what let it survive the gut and work as a swallowed capsule, which an ordinary peptide does not.
So it is a peptide-derived small molecule, not a peptide drug, and it is the smallest thing this market sells under the peptide heading. Peptide Lexicon's noopept entry covers the structure and the Moscow design history, which worked backwards from piracetam to the dipeptide fragments implicit in it.
What is not in the record
No approval in the United States or the European Union. No registered US drug-substance filing — zero Drug Master Files of any type, which for a compound with no US application pending is the expected result rather than a finding, on the reasoning set out on the sources page. No dose-ranging study establishing why 20 mg. No second figure from an independent group to compare it against.
The honest summary is short. One number exists, from one unblinded study of sixty stroke patients, published in Russian in 2011, and it is 20 mg a day for two months. Everything else in circulation is that figure repeated, or an inference from rodent work.
Reported here with its source and its date, as this catalogue reports every figure. None of it is offered as a figure to use, and decisions about administering anything to a person belong with a clinician.
