Cerebrolysin research and evidence overview

Cerebrolysin has more human trial data than almost any compound covered on this site — dozens of randomised studies, pooled analyses running to thousands of patients — and the field still cannot agree on whether it works. The pattern in the data is consistent and telling: early neurological scores improve reliably, while functional independence at 90 days does not reach statistical significance. Nearly all of the pivotal work involves the manufacturer. That combination — real trials, real effect sizes on softer endpoints, persistent sponsor entanglement — is why the drug is approved in some countries and absent from guidelines in others.

Cerebrolysin is not a single peptide. It is a parenterally administered preparation of low-molecular-weight neuropeptides and free amino acids derived from purified porcine brain protein, delivered by daily intravenous infusion at doses around 30 mL. This matters for interpreting the literature: unlike a defined molecule, a biological extract has batch-to-batch composition that cannot be fully specified, and there is no generic equivalent against which independent groups could run a replication.

The best-known individual trial is CARS (Cerebrolysin and Recovery After Stroke), a prospective, randomised, double-blind, placebo-controlled, multicentre study (Muresanu et al., Stroke 2016;47(1):151–159, PMID 26564102). Patients received Cerebrolysin 30 mL/day or saline placebo once daily for 21 days starting 24 to 72 hours after stroke onset, alongside a standardised 21-day rehabilitation programme in both arms. The primary endpoint was the Action Research Arm Test at day 90.

The result favoured Cerebrolysin strongly — a Mann-Whitney effect size of 0.71 (95% CI 0.63–0.79, p<0.0001) on the ARAT, and 0.62 (95% CI 0.58–0.65, p<0.0001) on a multivariate global-status measure across 12 outcome scales. Dropout was under 4% and safety matched placebo.

The authors' own conclusion is the honest one and is often dropped when this trial is cited: the study was exploratory with a relatively small sample, and the results "should be confirmed in a large-scale, randomized clinical trial." A pre-planned individual-patient meta-analysis combining CARS-1 and CARS-2 reproduced the direction of effect (Guekht et al., Neurological Sciences 2017, PMID 28707130). Both papers include authors employed by the manufacturer, EVER Neuro Pharma.

The most useful recent synthesis pools 14 randomised controlled trials comparing Cerebrolysin against placebo in acute ischaemic stroke, totalling 2,884 patients, with risk of bias assessed by Cochrane RoB 2.0 and certainty graded by GRADE (Patel et al., Cureus 2025, PMID 41018475).

Its findings, in order of importance:

• Neurological recovery improved: mean difference in NIHSS change +1.39 (95% CI 0.53–2.25, p=0.020).
• Functional independence (modified Rankin Scale 0–2) did not reach significance: risk ratio 1.31 (95% CI 0.90–1.91).
• No difference in serious adverse events (RR 1.08, 95% CI 0.84–1.40) or mortality (RR 0.86, 95% CI 0.68–1.09).
• Hemorrhagic transformation: RR 0.55 (95% CI 0.32–0.92).

A 1.39-point difference on a 42-point stroke scale is small. The authors themselves call for further high-quality trials to confirm any impact on long-term functional outcomes.

Three structural features keep this literature contested. First, the split between endpoints: a drug that moves an examiner-rated neurological score but not the patient-relevant measure of whether someone can live independently is exactly the profile that attracts scepticism. Second, sponsor involvement runs through the pivotal trials and their analyses, and truly independent replication is scarce. Third, much of the trial base comes from a limited set of regions and centres, which raises questions about generalisability that pooling cannot resolve.

Beyond stroke, Cerebrolysin has been studied in traumatic brain injury and dementia, generally in smaller trials with mixed results. Those literatures are thinner than the stroke one and share its structural problems. Anyone reading a strong claim about Cerebrolysin should check which endpoint it rests on and who funded the trial.