Cerebrolysin's human trials are real. Its animal papers are being retracted.

Why this note exists

Most evidence audits on this site end in the same place: the human data are thin, or old, or absent. Cerebrolysin is the opposite problem. There are large, multicentre, double-blind, placebo-controlled trials, published in Stroke and reviewed by Cochrane twice. That makes it a more interesting test of how to read evidence honestly, because the failure mode here is not a missing study — it is a reader taking the encouraging trials and skipping the discouraging ones, or the reverse.

In the index generated on August 15, 2026, 13 of the 112 vendors pepmg tracks carried a Cerebrolysin listing — 13 listings in all, fifty-ninth of 83 compounds by vendor coverage, so a modest corner of the market rather than a headline one.[22] Eleven of the 13 are labelled as a 60 mg vial; the listed prices on those run from $41.25 to $129.95.[22]

A note on what kind of evidence this is: every efficacy and safety figure in the trial sections below is human data, and each carries its design and participant count in the same sentence. The retraction section is about animal and transgenic-model studies, and says so every time. pepmg makes no claim about the contents of any vendor's vial, and does not convert a published trial regimen into a protocol.

What it is

Not a peptide, in the sense the market uses the word

Almost everything else in this index is a synthetic molecule with a published amino-acid sequence. Cerebrolysin is not. Cochrane's reviewers describe it as "a mixture of low-molecular-weight peptides and amino acids derived from porcine brain" — a hydrolysate, not a defined single entity — and note that it is widely used in the treatment of acute ischaemic stroke "in Russia, Eastern Europe, China, and other Asian and post-Soviet countries."[8][9] The Cochrane review of vascular dementia uses the same description: "a porcine brain-derived preparation that is said to have neurotrophic and neuroprotective activity," given as a series of daily intravenous infusions.[10]

The manufacturer, EVER Pharma, lists the licensed product as a "Solution for injection" at a strength of 215.2 mg/mL, indicated "[f]or treatment of cerebrovascular disorders" and specifically senile dementia of Alzheimer's type, vascular dementia, stroke and craniocerebral trauma.[1] That is a ready-made sterile solution of a stated concentration, dispensed in ampoules and infused — a different pharmaceutical article from a milligram-labelled vial, whatever is inside the vial.[1][22]

On US status the manufacturer is blunter than pepmg would need to be. Its product page carries a notice to U.S. visitors stating that "Cerebrolysin is not registered with the U.S. Food and Drug Administration (FDA) and is not approved for sale or distribution in the United States."[1] A Drugs@FDA query on September 1, 2026 returns no approved product under that name.[2] A ClinicalTrials.gov search the same day returned 42 registered studies of cerebrolysin, 37 of them interventional and four currently recruiting — and none with a listed location in the United States.[12]

The human record

Four randomized trials, and they do not all point the same way

This is the part that makes Cerebrolysin unusual in this index, and it deserves to be laid out with the numbers attached rather than summarized into a verdict.

Enrollment and results as reported in each trial's primary publication. Manufacturer involvement is disclosed in the papers themselves: three CARS authors give their affiliation as the Department of Clinical Research at EVER Neuro Pharma, and the ESCAS disclosure statement records investigators who have been principal investigators on Cerebrolysin trials funded by that company as well as on academically funded ones.[4][7]

Start with the largest, because it is the one that most often goes missing. CASTA randomized 1,070 patients with acute ischaemic hemispheric stroke within 12 hours of onset to 30 mL of Cerebrolysin daily or saline placebo by intravenous infusion for 10 days, on top of aspirin, and followed them for 90 days.[3] The primary endpoint was a combined global directional test of the modified Rankin Scale, Barthel Index and NIH Stroke Scale, and the published result is unambiguous: "[t]he confirmatory end point showed no significant difference between the treatment groups."[3] The paper then reports a post hoc analysis in the subgroup with NIHSS above 12 showing a trend in favour of Cerebrolysin, and the authors themselves write that this observation "should be confirmed by a further clinical trial."[3] A post-hoc subgroup in a neutral trial is a hypothesis, not a result.

CARS, published in Stroke in 2016, is the trial most often cited on the other side. It randomized 208 patients to 30 mL/day or placebo for 21 days alongside a standardized rehabilitation programme, with the Action Research Arm Test at day 90 as the primary endpoint, and reported a large nonparametric effect size favouring Cerebrolysin (Mann-Whitney estimator 0.71, 95% CI 0.63–0.79).[4] The authors' own conclusion contains the caveat that gets dropped in citation: "[b]ecause this study was exploratory and had a relatively small sample size, the results should be confirmed in a large-scale, randomized clinical trial."[4] Three of its authors list their affiliation as the Department of Clinical Research at EVER Neuro Pharma, the manufacturer.[4]

In traumatic brain injury, a prospective meta-analysis of the two randomized, double-blind, placebo-controlled CAPTAIN trials pooled 185 patients with admission Glasgow Coma Scores between 6 and 12 and reported a small-to-medium effect favouring Cerebrolysin on a multidimensional ensemble of functional and neuropsychological scales, statistically significant at day 30 and day 90, with comparable safety between groups.[5] A separate three-arm randomized trial combined Cerebrolysin with repetitive transcranial magnetic stimulation in TBI patients.[6]

The most recent randomized human data are from 2025. ESCAS enrolled 132 patients with left middle cerebral artery stroke and nonfluent aphasia at two Romanian centres, randomized them to Cerebrolysin or placebo alongside speech and language therapy, and analysed 123 in the intention-to-treat population.[7] At day 90 the Cerebrolysin group improved on the Western Aphasia Battery by a mean of 35.6 points against 20.8 on placebo, a difference of 14.8 points (95% CI 9.5–20.1), with no safety signal.[7] The authors describe it as a pilot and write that "[f]urther research with larger cohorts is needed to fully establish the efficacy of this combination therapy."[7] Its disclosure statement lists several investigators as principal investigators on manufacturer-funded Cerebrolysin trials.[7]

The systematic reviews

What happens when someone pools them and grades the certainty

Cochrane has reviewed Cerebrolysin in acute ischaemic stroke repeatedly since 2010; the current version is the 2023 update, which the reviewers describe as an update of a review first published in 2010 and last updated in 2020.[8][9] It includes seven randomized trials and 1,773 participants, and its findings are graded — which matters more than the direction of any single estimate.[8] One qualifier belongs in the same breath as that total: one of the seven trials, contributing 272 participants, tested Cortexin, a Cerebrolysin-like peptide mixture derived from cattle rather than pig brain, and the review's conclusions are written to cover both.[8] The 2020 version, which pooled seven Cerebrolysin trials and 1,601 participants, reported the same pattern.[9]

On all-cause death, the pooled risk ratio was 0.96 (95% CI 0.65 to 1.41) across six trials and 1,689 participants, moderate-certainty evidence.[8] On serious adverse events overall, moderate-certainty evidence of little to no difference (RR 1.16, 95% CI 0.81 to 1.66) — but that total decomposes, and the decomposition is the finding: fatal serious adverse events RR 0.90 (0.59 to 1.38), and non-fatal serious adverse events RR 2.39 (95% CI 1.10 to 5.23), also moderate certainty.[8] In the subgroup given 30 mL for 10 days — the CASTA regimen — the increase was more prominent, RR 2.87 (95% CI 1.24 to 6.69) across two trials and 1,189 participants.[8] The 2020 version reported the same pattern at RR 2.15 (95% CI 1.01 to 4.55).[9]

The review's authors also record who paid: "[t]he manufacturer of Cerebrolysin supported three multicentre studies, either totally, or by providing Cerebrolysin and placebo, randomisation codes, research grants, or statisticians," and they judged two of the seven studies to be at high risk of other bias and the remaining five at unclear risk.[8] None of the included studies reported the outcome most readers actually care about — death or dependence at the end of follow-up.[8]

The vascular dementia picture is different in shape and no more settled. The 2019 Cochrane review found six randomized trials with 597 participants, all in mild to moderate vascular dementia, and did find beneficial effects on cognition (SMD 0.36, 95% CI 0.13 to 0.58) and on global function (RR 2.69, 95% CI 1.82 to 3.98) — both graded very low-quality evidence.[10] The authors' conclusion is worth quoting because it is the honest form of a positive result: the data "are not definitive," the analyses were limited by heterogeneity and high risk of bias, and "[i]f there are benefits of Cerebrolysin, the effects may be too small to be clinically meaningful."[10] They also note that no new randomized trial in vascular dementia had appeared since the 2013 review — a seven-year gap in a condition the product is marketed for.[10]

Newer pooled work exists but is a weaker design, and should be read as such. A 2026 systematic review and meta-analysis of Cerebrolysin as an adjunct to mechanical thrombectomy reports favourable results on functional outcome, symptomatic intracranial haemorrhage and mortality — from three observational studies and 294 patients, with no randomization.[11] Observational comparisons of who received an add-on therapy cannot separate the drug from the reasons clinicians chose to give it.

Research integrity

Seven retractions in eighteen months, all of them animal work

Between January 2025 and June 2026, seven papers with Cerebrolysin in the title carried retraction notices, in BMC Neuroscience (two), Acta Neuropathologica, the Journal of Alzheimer's Disease, Molecular Neurobiology (two) and Neurochemical Research.[13][15][16][17][18][19][20] Every one of the seven is a preclinical study — transgenic mouse models of tauopathy, Alzheimer's disease, Parkinson's disease and Rett syndrome, and a rat nanoparticle-toxicity model. No human trial of Cerebrolysin has been retracted.[13][16][19][20]

One of those notices is available in full text, and it is specific rather than vague. Retracting a 2015 paper on Cerebrolysin in a triple-repeat tau model of Pick's disease, the editor writes that the article was retracted "because of concerns regarding figures presented in this work," that "[a]n investigation conducted after its publication discovered" five separate instances in which image panels appear to overlap — including panels labelled as vehicle-treated and Cerebrolysin-treated tissue from the same figure, which "represent tissues taken from animals which underwent different treatments" — and that the editor "no longer has confidence in the integrity of the research presented in this article."[13][14] The notice records that the authors did not reply to the publisher's correspondence, and that the corresponding author had died.[13]

Four of the seven retracted papers share a senior author, Eliezer Masliah, then at the University of California San Diego, and three of those four list co-authors whose stated affiliation is the manufacturer's research division — EVER Neuro Pharma or its predecessor EBEWE Pharmaceuticals.[13][15][16][17] Separately, NIH announced on September 26, 2024 that "[f]ollowing an investigation, the National Institutes of Health (NIH) has made findings of research misconduct against Eliezer Masliah, M.D., due to falsification and/or fabrication involving re-use and relabel of figure panels representing different experimental results in two publications," and that he was no longer serving as director of the Division of Neuroscience at the National Institute on Aging.[21]

Two facts, kept apart on purpose. The NIH finding names two publications and does not identify them in the statement; the Cerebrolysin retraction notices state figure concerns and do not say which investigation prompted them.[13][21] pepmg is not asserting a link between the two, and nobody should read one here. What is established, and is enough on its own, is narrower: a meaningful slice of the animal literature that made Cerebrolysin look neuroprotective has been withdrawn by its journals, and the remaining three retractions come from a different research group again.[18][19][20]

Why it matters for a reader of a price index rather than a reader of a journal: the case for most compounds sold as research peptides is preclinical. When human trials are neutral, the animal work is what supplies the story about mechanism and promise. Here, part of that supply has been formally withdrawn — while the human trials, which nobody has retracted, remain the mixed and heavily graded record described above.[8]

Three things this note is not saying

First, it is not saying Cerebrolysin does not work. Two Cochrane reviews found directionally positive effects in vascular dementia, a 208-patient randomized trial found a large effect on arm function after stroke, a pooled analysis of two randomized TBI trials found a small-to-medium effect, and a 2025 randomized pilot found a 14.8-point difference on an aphasia battery.[4][5][7][10] Those are real results in real patients.

Second, it is not saying the retractions prove the trials are wrong. Retracted animal papers and completed human trials are separate bodies of evidence, and the human trials stand or fall on their own methods. The honest inference is about how much weight the preclinical rationale can bear, not about what happened inside CASTA or CARS.[3][4]

Third, it is not saying a vial bought online is the product these trials used. The trials infused a licensed solution of stated concentration, in hospital, under supervision, for days at a time.[1][3] Eleven of the 13 listings in the index are milligram-labelled vials, and pepmg makes no claim about what is in them and does not translate a millilitre trial regimen into a milligram quantity.[22]

What it is saying is narrower: Cerebrolysin has the kind of evidence base most of this index lacks, and having read it, the strongest defensible summary is that the largest trial was neutral, the pooled certainty is moderate at best and very low in dementia, there is a signal of more non-fatal serious adverse events, and part of the animal literature underneath it has been retracted.[3][8][10][13]

Questions people are asking