What is Cerebrolysin?
Cerebrolysin is an injectable mixture made by breaking down proteins from pig brain tissue. It contains peptide fragments and amino acids, rather than one defined peptide. The Austrian product label describes a prescription preparation given by intramuscular injection or intravenously, including infusion. Austrian label.
The interest comes from a practical question: can a medicine improve recovery alongside rehabilitation after brain injury? Human trials have tested that question. Some found better recovery scores, while others did not meet their main outcome. The answer depends on the condition, the outcome being measured and the quality of the comparison.
What did the acute-stroke trials find?
The CASTA trial enrolled 1,070 people with acute ischemic stroke and began intravenous treatment within 12 hours. Its main 90-day outcome combined disability, daily activities and neurological impairment. Cerebrolysin did not significantly improve that combined outcome. A later analysis of the more severely affected subgroup suggested possible benefit, including a mortality difference, but it was a post hoc finding that needed confirmation. CASTA trial.
The 2023 Cochrane review, which searched through 2022, included seven trials with 1,773 participants. One tested Cortexin, a different brain-derived preparation. Across six trials reporting deaths, Cerebrolysin or Cortexin probably made little or no difference to mortality. The reviewers could not establish the effect on their defined outcome of death or dependence because none of the included studies reported that outcome. That is a reporting gap, not proof that every aspect of disability was unchanged. Cochrane review.
The same review found a possible increase in nonfatal serious adverse events with Cerebrolysin: a risk ratio of 2.39, with a 95% confidence interval from 1.10 to 5.23, across three trials. Total serious adverse events did not clearly differ. These results do not support describing it as harmless or as a proven way to improve survival after stroke.
Why are some rehabilitation results more encouraging?
CARS-1 randomized 208 people to Cerebrolysin or placebo alongside a standardized rehabilitation programme. The treated group improved more on an arm-function test at 90 days. This was an exploratory trial, and its authors called for larger confirmation. CARS-2 enrolled another 240 people with milder starting deficits and showed much smaller differences. The combined analysis favored Cerebrolysin using its planned method, but the overall day-90 result was not significant under a random-effects model. Those are two trials analyzed together, not three independent demonstrations. CARS-1; CARS combined analysis.
The 2025 ESCAS pilot adds a specific language-recovery finding. It enrolled 132 people with nonfluent aphasia after an ischemic stroke in two Romanian centres; 123 entered the intention-to-treat analysis. Both groups received speech therapy. The main reported per-protocol comparison, involving 120 people, found a 35.6-point improvement in the Western Aphasia Battery score with Cerebrolysin versus 20.8 with placebo at 90 days. Intention-to-treat results were reported as similar. Adverse events occurred in 43.9% versus 30.3%, without a statistically significant difference. This supports further study of this particular rehabilitation setting, without establishing a benefit for every stroke survivor. ESCAS trial.
A larger 2025 observational study, C-REGS2, enrolled 1,865 people and analyzed a target population of 1,769 with moderate ischemic stroke. Cerebrolysin treatment was associated with better disability scores at 90 days than standard care alone. Its prospective design and efforts to account for differences between patients make this useful evidence from clinical practice. Treatment was not randomized, so remaining differences between groups could contribute to the result. C-REGS2 study.
What about traumatic brain injury and dementia?
The CAPTAIN trials studied Cerebrolysin added to standard care after moderate-to-severe traumatic brain injury. CAPTAIN I enrolled 46 people and did not meet its primary combined outcome in the intention-to-treat analysis. Some individual cognitive tests and the per-protocol analysis favored treatment. CAPTAIN II enrolled 142 people, with 139 analyzed, and found a statistically significant small-to-medium benefit on a combined set of 13 outcome scales at 90 days. It was a single-centre trial. A positive composite does not mean every component improved or that complete recovery was demonstrated. CAPTAIN I; CAPTAIN II.
In vascular dementia, a 242-person randomized trial found better cognitive and clinician-rated global outcomes at 24 weeks. The cognitive score improved by 10.6 points with Cerebrolysin versus 4.4 with placebo. That is a favorable symptomatic result, not evidence that the underlying vascular disease was reversed. Vascular dementia trial.
Alzheimer's disease has also been studied. In a 279-person trial comparing three intravenous doses with placebo, the lowest dose improved cognition and global clinical ratings at 24 weeks. The two higher doses improved global ratings but did not significantly improve cognition. This uneven response does not establish that a larger dose works better, or that treatment prevents dementia. Alzheimer's trial.
What does the mechanism tell us?
Descriptions such as neuroprotective or neurotrophic refer to proposed effects on nerve-cell survival and recovery. Laboratory findings cannot establish that an injection repairs a person's brain. Parts of the preclinical literature also need particular care: a mouse tauopathy paper was retracted in December 2025 because the editor lost confidence in its figures. That notice concerns that animal study, rather than the human trials above. We do not use it as evidence of benefit. Retraction notice.
What is promising, and what still needs an answer?
Observed findings
- Arm-function and language-recovery benefits in selected stroke rehabilitation trials.
- Positive cognitive and global scores in some dementia trials.
- A positive 90-day composite in CAPTAIN II.
Important limits
- The large CASTA trial missed its main outcome.
- Results vary across trials and ways of combining them.
- A nonfatal serious-adverse-event signal remains relevant.
Questions to bring
- Does the study match my diagnosis and stage of recovery?
- What useful change would be measured, and when?
- What product, authorization and monitoring would apply?
Where is it approved, and what are the risks?
Authorization is country-specific. Austria's label includes supportive treatment for Alzheimer-type and vascular dementia, deficits after stroke and traumatic brain injury in adults. Cerebrolysin is not FDA approved in the United States. An Austrian authorization does not establish permission to prescribe, compound or import a preparation elsewhere. The FDA has previously identified compounded Cerebrolysin products as unapproved drugs; that enforcement record is separate from Austria's licensed product. Austrian label; FDA regulatory record.
The Austrian label contraindicates use in status epilepticus, severe kidney impairment and hypersensitivity. It calls for caution in other seizure disorders because seizure frequency may increase, and identifies possible additive effects with antidepressants and MAO inhibitors. Allergic reactions, agitation, dizziness and administration-related reactions are among the reported risks. Pregnancy, breastfeeding and other medicines require individual review. Trial reports of similar average adverse-event rates do not remove these concerns.
How should the evidence guide a conversation?
Acute stroke
CASTA was neutral; pooled evidence did not show a mortality benefit.
Rehabilitation
Some arm-function and language outcomes improved alongside rehabilitation.
TBI and dementia
Selected trials show benefits on specific scales, with limits to generalization.
Research in progress
The CADASIL registry lists an estimated 30 participants and no posted results.
The CERICA record for CADASIL still lists active, not recruiting, with estimated enrollment of 30 and no posted results as of November 9, 2026. A registry entry is a research plan and status record, not a positive finding. CERICA registry.
For someone considering treatment, the useful discussion is about their diagnosis, the rehabilitation or established care already in place, and a realistic way to measure benefit. These selected trials do not establish cognitive enhancement in healthy people. Related records are searchable on our Science page.
Questions people ask
Is Cerebrolysin a single peptide?
No. It is a mixture of peptide fragments and amino acids derived from pig brain proteins. The licensed Austrian preparation is given by injection or infusion. Austrian label.
Has it been shown to help after a stroke?
Some rehabilitation trials found improvements in arm function or language. The larger CASTA acute-stroke trial did not improve its main combined outcome, and the Cochrane review did not show a mortality benefit. The findings need to be matched to the type and timing of treatment. CASTA; Cochrane; ESCAS.
Is Cerebrolysin approved in the United States?
It is not FDA approved. Austria has a prescription product label for specified neurological conditions, but that does not establish authorization in another country. Austrian label; FDA record.
Can it be called safe because some trials found similar side effects?
That would be too broad. The Cochrane stroke review found a possible increase in nonfatal serious adverse events. The Austrian label also lists contraindications and precautions involving seizures, kidney function, allergies and other medicines. Cochrane review; Austrian label.
The evidence
Selected original studies, trial records and regulatory sources checked November 9, 2026. Study doses describe the research and are not instructions for use.
This article is for educational purposes only and is not medical advice, a diagnosis, or a treatment recommendation. Evidence and approvals apply to the products and populations described. Any treatment decision should be made with a qualified physician. Individual results vary.