Cortexin

A Russian prescription injectable made from a water-soluble polypeptide fraction of cattle or pig cerebral cortex, marketed for stroke and encephalopathy — an undefined tissue extract rather than a single peptide, never reviewed by the FDA.

Also referenced as: Korteksin, Kortexin, Cortexin 10 mg, polypeptides of cattle cerebral cortex

Also appears in: Cognitive

COAPublic product evidenceSearch the public certificate ledger for this compoundNo exact compound records are currently indexed under this profile name.→

---

What is Cortexin?

Cortexin is not a peptide in the sense the rest of this library uses the word — it has no sequence, no single molecular weight, and no defined active ingredient. The approved labeling describes it as a lyophilizate obtained by acetic-acid extraction from the cerebral cortex of cattle or pigs, characterized only as a complex of water-soluble polypeptide fractions whose molecular weight does not exceed 10,000 Da. The fullest published description of the material — reported as the drug’s specification rather than as an independent analysis — puts it at 70–95% acidic and neutral polypeptides of 1,000 to 10,000 Da, with isoelectric points spanning pI 3.5–9.5 (Kurkin et al., PLoS One, 2021;16(7):e0254493). It is manufactured by Geropharm and is a registered prescription drug in Russia and other CIS states, sold as a vial of powder reconstituted for intramuscular injection.

It reaches Western readers for two reasons. The first is adjacency: it is the counterpart to cerebrolysin, the porcine-brain peptide mixture that dominates the same conversation, and the two are routinely compared head to head in the Russian literature. The second is that it is a finished foreign pharmaceutical that circulates through export resellers, so it lands in nootropic and research-market catalogs despite being a prescription drug in its home market and having no US regulatory status of any kind.

How it works

Because the product is a mixture, mechanism work describes what the extract does rather than what a molecule does. The published claims:

• Broad glutamatergic and GABAergic receptor binding. In an in vitro panel run against a wide receptor set, Cortexin at 10 µg/mL showed high or moderate binding to AMPA receptors (80.1%), kainate receptors (73.5%), mGluR1 (49.0%), GABA-A1 (44.0%), and mGluR5 (39.7%). The same work reported that radiolabeled Cortexin crossed the blood-brain barrier in mice, reaching brain concentrations equal to 6–8% of whole-blood levels (Kurkin et al., PLoS One, 2021;16(7):e0254493). This is a binding screen, not a demonstration that any of those interactions produces the clinical effects claimed for the drug.
• Selective caspase-8 inhibition. Cortexin inhibited brain caspase-8 while having much weaker or absent effects on caspase-1, -3, and -9, cathepsin B, and calpain. The authors isolated a simpler peptide fraction that retained the full inhibitory capacity of the whole drug, and both the drug and the fraction prevented neuronal death in a glutamate-toxicity culture model (Yakovlev et al., Biomeditsinskaia Khimiia, 2017;63(1):27–31).
• Antioxidant and anti-inflammatory effects, with named protein partners. A mechanism review summarized work in which Cortexin restored the balance of pro- and antioxidant systems in an accelerated-aging rat model and produced anti-inflammatory effects both centrally and systemically, and identified β5-tubulin, creatine kinase B, and protein 14-3-3 α/β as molecular partners of Cortexin peptides in the brain (Gulyaeva, Zhurnal Nevrologii i Psikhiatrii imeni S.S. Korsakova, 2018;118(10):93–96).
• Oxidative-stress and inflammatory-marker modulation after ischemia-reperfusion. In 35 rats subjected to cerebral ischemia-reperfusion injury, Cortexin at 1 and 2 mg/kg lowered serum total oxidant status, raised total antioxidant status, and reduced brain expression of OPG, RANK, RANKL, and TRPC1 (Guven et al., Neurological Research, 2026;48(4):510–521). This is one of the few Cortexin papers published outside the Russian-language literature.

The mechanistic picture is therefore pleiotropic by design and unfalsifiable in practice: with no defined active, there is no way to attribute an outcome to a component, and no pharmacokinetics in the ordinary sense — a point the manufacturer’s own labeling concedes, stating that the complexity of the composition does not allow conventional pharmacokinetic analysis. It is also worth noting what else is in the vial: the approved Russian labeling lists glycine as a stabilizer at 12 mg per 10 mg vial, so the excipient outweighs the labeled active by mass. Glycine is not inert filler in this context — it is itself one of the eleven compounds counted as a nootropic in Russian practice (Alexandrova et al., International Journal of Risk & Safety in Medicine, 2026;37(2):257–264).

Research status

Cortexin has a large clinical literature and almost none of it is independently verifiable. A PubMed title search returns 69 records, about ten of which belong to unrelated compounds that share the name; of the roughly 59 that concern the drug, 32 sit in a single Russian journal, Zhurnal nevrologii i psikhiatrii imeni S.S. Korsakova. The remainder scatter across mostly Russian-language titles, with only a handful — PLoS One, Biomedicines, Neurological Research, Neuroendocrinology — in English-language journals. A ClinicalTrials.gov intervention search for Cortexin returns zero registered studies, against 42 for cerebrolysin.

The one trial that cleared an international systematic review. The 2023 Cochrane update of Cerebrolysin for acute ischaemic stroke added a single RCT of what it called the “Cerebrolysin-like agent Cortexin,” contributing 272 participants to a review of seven trials and 1,773 participants total. The conclusion went against the class: at moderate certainty, the reviewers found that Cerebrolysin or Cerebrolysin-like peptide mixtures derived from cattle brain probably have no beneficial effect on preventing all-cause death after acute ischemic stroke (RR 0.96, 95% CI 0.65 to 1.41; 6 trials, 1,689 participants). The reviewers judged the Cortexin study at low risk of bias only for incomplete outcome data and unclear for every other domain. That trial is the multicenter double-blind placebo-controlled study of 272 acute ischemic stroke patients randomized to two 10-day Cortexin courses, one course followed by placebo, or two placebo courses over 70 days, which reported efficacy and safety in favor of the two-course arm (Aliferova et al., Zhurnal Nevrologii i Psikhiatrii imeni S.S. Korsakova, 2014;114(4):41–46).

The same review found something the Cortexin literature does not report on itself. Across the three Cerebrolysin trials that supplied safety data (1,335 participants), the reviewers found a probable increase in the total number of people with non-fatal serious adverse events (RR 2.39, 95% CI 1.10 to 5.23; moderate certainty), more prominent at the higher cumulative dose. That signal comes from Cerebrolysin arms, not Cortexin ones, and should not be transferred to Cortexin as if it had been measured there — but it is the only well-powered safety analysis anyone has run on this class, and it points the opposite way from the uniformly benign tolerability reported in the Russian Cortexin trials.

The larger recent trials are not placebo-controlled in the way that matters. A 2025 multicenter double-blind study enrolled 490 acute ischemic stroke patients, but it was a route-equivalence trial: both arms received Cortexin 10 mg — one intravenously with intramuscular placebo, the other intramuscularly with intravenous placebo — and both then received a second intramuscular course. It reported 93.64% versus 86.50% reaching modified Rankin 0–2 at day 90 and concluded the two forms were therapeutically equivalent (Fedin et al., Zhurnal Nevrologii i Psikhiatrii imeni S.S. Korsakova, 2025;125(12):60–67). A double-dummy design with no untreated arm can establish that two presentations behave alike; it cannot establish that either beats placebo. Similarly, a multicenter randomized dose-ranging study in 189 patients with chronic cerebral ischemia compared 20 mg, 10 mg, and basic treatment alone, reporting dose-dependent effects on neurological symptoms, asthenia, and sleep, an antioxidant effect independent of dose, and antidepressant and anxiolytic effects that were not significant (Fedin et al., same journal, 2018;118(9):35–42). The 2026 DIACORT trial randomized 110 patients with neurological complications of type 2 diabetes to background therapy with or without ten 10 mg intramuscular doses and reported broad advantages across cognitive, mood, and neuropathy scales — but the control group received no placebo injections, the report carries no author list in the PubMed record, and among the reported differences is a larger fall in HbA1c (7.3% versus 7.8%), a glycemic marker not obviously downstream of a ten-day course of a neuropeptide extract. Worth noting given the composition of the vial: every patient in both arms also received glycine and B-vitamin complexes for 30 days as background therapy (Zhurnal Nevrologii i Psikhiatrii imeni S.S. Korsakova, 2026;126(4):101–112).

Much of the remaining record is uncontrolled. A frequently cited pediatric study enrolled 635 children aged 3–7 with ADHD, speech delay, perinatal CNS injury, or asthenic syndrome, and gave all of them the same 10-injection course with no comparator (Zykov et al., same journal, 2018;118(3):27–31). An all-Russian screening program administered Cortexin to 50,000 patients with brain ischemia across 70 cities and analyzed before-and-after scores in 500 of them, again with no control group (Mashin et al., same journal, 2014;114(9):49–52). The CORTEX program in post-COVID syndrome followed 979 patients across four countries in an open observational design (Putilina et al., Neuroscience and Behavioral Physiology, 2022;52(6):836–841). Studies of this shape can describe tolerability and natural recovery; they cannot separate drug effect from either.

The one head-to-head comparison went against it. Forty patients in the early recovery period after ischemic stroke were randomized to Cellex or Cortexin alongside basic therapy. The Cellex group showed better NIHSS improvement (3 versus 5, p=0.03) and better MMSE scores (24 versus 22, p=0.04), with no significant difference on mobility or frontal assessment (Khabirov et al., Zhurnal Nevrologii i Psikhiatrii imeni S.S. Korsakova, 2020;120(12 Pt 2):11–15). The trial is small, but it is the comparison that exists, and Cortexin lost it.

Preclinical work is more consistent and largely sponsor-funded. Cortexin reduced necrosis volume and neurological deficit in rat middle-cerebral-artery-occlusion and carotid-stenosis models, performing comparably to cerebrolysin and better than Actovegin (Kurkin et al., PLoS One, 2021, above); reduced neurological deficit and cortical neuronal damage in rat models of developmental delay from prenatal ethanol and neonatal hypoxia-ischemia (Kurkin et al., Biomedicines, 2025;13(4):860); and protected cultured rat dorsal root ganglion neurons against high-glucose injury in a concentration-dependent way (Yazar and Ayar, Neuroendocrinology, 2023;113(9):924–929). The PLoS One and Biomedicines studies were sponsored by Pharm-Holding CJSC — the manufacturer’s own research organization — with three authors on each paper employed there.

No regulatory authority outside the CIS has approved Cortexin for anything, and it does not appear in US, EU, or UK clinical practice guidelines for dementia or Alzheimer’s disease. Russia’s own Alzheimer’s guidelines do not recommend it either: the two nootropics they name as adjuvant therapy are cerebrolysin and choline alfoscerate (Alexandrova et al., International Journal of Risk & Safety in Medicine, 2026;37(2):257–264).

Common dosage forms

• Lyophilizate vials at 10 mg (adult) and 5 mg (pediatric), typically boxed in tens, reconstituted in 1–2 mL of 0.9% sodium chloride, water for injection, or 0.5% procaine and given by intramuscular injection. This is the presentation used in essentially all of the clinical literature.
• An intravenous presentation, registered in Russia alongside the intramuscular form and the subject of the 2025 route-equivalence trial described above.
• A rectal suppository form used in the 2025 rodent developmental-delay study alongside the intramuscular route. This is a research formulation in animals, not a registered human presentation.

Two format notes distinguish Cortexin from most entries in this library. First, what circulates outside the CIS is the finished branded pharmaceutical moved through export channels, not a bulk research chemical — there is no sequence to synthesize, so the usual research-vendor supply model does not apply. Second, no FDA-approved product exists, Cortexin is not on the FDA’s section 503A bulk drug substances list, and no licensed US compounding route exists for an animal-tissue extract of this kind. This section describes formats only and is not dosing guidance.

Key considerations

• Registered abroad is not approved here, and the gap is wide. Cortexin is a prescription medicine in Russia and has been on the Russian list of vital and essential medicines (ZhNVLP) since January 2016. It is registered across the CIS, but per-country essential-medicines listing varies considerably: of the eleven nootropics surveyed in the CIS analysis, Russia’s list carried nine, Belarus and Kazakhstan eight, Uzbekistan four, and Armenia two, so registration in a country does not imply essential-medicines listing there. None of the class appears on the WHO Model List of Essential Medicines, and none is registered in the United States (Alexandrova et al., International Journal of Risk & Safety in Medicine, 2026;37(2):257–264). Foreign registration in this case reflects a regulatory tradition that accepts this class of drug, not an evidentiary bar an American reader would recognize.
• The evidence base is deep, narrow, and largely unblinded. Decades of published trials sit almost entirely in Russian-language journals, concentrated in one title, with no ClinicalTrials.gov registrations to check reported outcomes against pre-specified ones. The largest studies are observational or open-comparator; the one blinded placebo-controlled trial that reached an international systematic review contributed to a conclusion of no mortality benefit for the class.
• There is no defined active ingredient, which changes what verification can mean. For a synthetic peptide, a certificate of analysis can confirm mass and purity against a known sequence. For a sub-10-kDa cortical extract there is no such reference — identity and potency are questions of protein content and bioassay, and batch-to-batch variability is an intrinsic property of tissue-derived preparations rather than a manufacturing defect. Nothing an independent lab can run on a vial will tell a buyer it matches the material used in the published trials.
• The source tissue carries a category-level safety consideration, and the labeling does not address it. Cortexin is extracted from bovine and porcine central nervous tissue, the tissue class European regulators treat as highest-risk for transmissible spongiform encephalopathy agents. No transmission has ever been attributed to this product, and the published trials consistently report good tolerability with few adverse events. But the approved labeling documents no sourcing controls at all — no country of origin for the animals, no age limit, no TSE-clearance step in the described acetic-acid extraction and purification. Claims about young source animals circulate widely on reseller pages; they are not in the manufacturer’s leaflet. Tolerability data from short open-label courses does not address the sourcing question, and none of it was gathered under a regulator that would.
• The name collides with at least three unrelated things. “Cortexin” is also a neuron-specific 82-residue membrane protein enriched in rodent cerebral cortex, later assigned the gene symbol CTXN1 (Coulter et al., Journal of Neurochemistry, 1993;61(2):756–759); “cortexin 3” is a separate human gene, CTXN3, expressed in kidney and brain and studied in schizophrenia genetics (Wang et al., International Journal of Molecular Medicine, 2007;20(4):501–510); and “renal cortexin” is a proposed 43 kDa antihypertensive protein purified from goat kidney cortex, with no relationship to the brain preparation (Chakraborty et al., Journal of the American Society of Hypertension, 2009;3(2):119–132). Roughly one in six PubMed records with “cortexin” in the title belongs to one of these rather than to the drug, so any citation offered for Cortexin should be checked against the actual subject of the paper. Separately — and not a name collision, but a common confusion — Cortagen is a distinct synthetic tetrapeptide, Ala-Glu-Asp-Pro, obtained by directed synthesis based on amino-acid analysis of Cortexin and marketed under its own name (Anisimov et al., Neuro Endocrinology Letters, 2004;25(1–2):87–93).