Speech Therapy Combined With Cerebrolysin in Enhancing Nonfluent Aphasia Recovery After Acute Ischemic Stroke: ESCAS Randomized Pilot Study

Clinical Question (PICO)

In right-handed, Romanian-speaking adults with left-middle-cerebral-artery ischemic stroke and nonfluent aphasia enrolled 3–5 days after stroke (P), does intravenous Cerebrolysin alongside structured speech and language therapy (I), compared with saline placebo alongside the same therapy (C), improve change from baseline in Western Aphasia Battery–Aphasia Quotient (WAB-AQ) at 30, 60, and 90 days (O)? Neurological impairment, disability, activities of daily living, and adverse events were secondary outcomes. No prespecified treatment-effect subgroup analysis is reported; the three visits are prespecified time points, not patient subgroups.

Bottom Line

In this 2025 double-blind pilot, Cerebrolysin plus speech therapy produced a 14.805-point greater mean WAB-AQ improvement at 90 days than placebo plus speech therapy (95% CI, 9.521–20.089; multiplicity-adjusted P<0.001, per-protocol analysis). The authors conclude that the combination is promising but requires confirmation in larger cohorts; this result alone does not establish routine clinical benefit.

Design

Trial type: Exploratory, phase 4, two-center, parallel-group randomized, placebo-controlled, double-blind pilot trial.
N: 132 enrolled/randomized, 66 assigned to Cerebrolysin and 66 to placebo in Table 1; the article reports 123 in its intention-to-treat (ITT) population and 120 in its per-protocol (PP) population (58 Cerebrolysin, 62 placebo at day 90). The efficacy numbers below come principally from the published PP analysis.
Randomization: 1:1 in blocks of four, using a pregenerated sequence and sealed opaque envelopes; separate nurses prepared infusions, and colored infusion lines masked Cerebrolysin’s amber appearance.
Setting: Two Romanian stroke centers, in Cluj-Napoca and Timișoara.
Enrollment: June 2020–October 2022.
Follow-up: Baseline at 3–5 days after stroke, then days 30±3, 60±3, and 90±3 after baseline.
Analysis: Between-group difference in mean change in WAB-AQ, with Wilcoxon comparisons and Bonferroni-adjusted P values; PP efficacy results in the article body, ITT results in supplementary tables, and a resampling-based repeated-measures analysis.
Primary outcome: Change in Romanian-validated WAB-AQ (0–100; higher is better) from baseline to each prespecified follow-up visit. These are continuous mean differences, not odds ratios or hazard ratios.

Population

Inclusion Criteria

Imaging- and clinically confirmed acute ischemic stroke in the left middle cerebral artery territory with nonfluent aphasia.
Enrollment 3–5 days after stroke; right-handed; Romanian used daily as the primary language; informed consent.

Exclusion Criteria

Previous symptomatic ischemic or hemorrhagic stroke; severe comprehension deficits, including global or fluent aphasia.
Epilepsy or documented epileptic discharges; severe renal or hepatic failure; life-threatening disease.
Uncorrectable hearing/vision problems or preexisting neurodegenerative or psychiatric disease.

Baseline Characteristics (randomized arms)

Cerebrolysin/placebo: 66/66 assigned; mean age 70.5/67.7 years; women 29/31 (43.9%/47.0%).
Mean WAB-AQ 50.8/56.0; mean NIHSS 9.2/8.6. The 5.2-point difference in baseline language score warrants attention in a change-score analysis.

Interventions

Cerebrolysin + speech therapy: 30 mL Cerebrolysin intravenously diluted with saline to 250 mL, once daily for ten treatment days in each of three cycles (days 1–14, 29–42, 57–70); one hour of speech and language therapy on each treatment day, 30 hours total.
Saline placebo + speech therapy: 250 mL saline intravenously on the same schedule, with the same 30-hour speech-therapy program. Some sessions were delivered remotely during pandemic restrictions.

Outcomes

Primary Outcome — change from baseline in WAB-AQ (PP population; prespecified visits, not subgroups):

Day 30: Mean between-arm advantage 7.099 points (95% CI, 3.478–10.721; adjusted P<0.001).
Day 60: Advantage 13.368 points (95% CI, 7.966–18.770; adjusted P<0.001).
Day 90: Cerebrolysin +35.579 points versus placebo +20.774; difference 14.805 points (95% CI, 9.521–20.089; adjusted P<0.001). The paper describes a >5-point within-person WAB-AQ change as clinically important, but that threshold is not a proven between-group treatment threshold.

Secondary Outcomes:

At day 90, NIHSS fell 2.085 points more with Cerebrolysin (95% CI for the between-arm difference in decreases, 1.076–3.094; adjusted P<0.001). NIHSS includes a language item, so this is not an independent validation of broad motor recovery.
The day-90 absolute mean modified Rankin Scale scores were 1.759 versus 2.258; adjusted between-arm P=0.081. Absolute mean Barthel Index scores were 82.586 versus 74.032; adjusted between-arm P=0.153. Neither absolute day-90 functional-scale comparison met the study’s adjusted significance criterion, although some change-from-day-30 comparisons did.

Adverse Events / Safety:

Any adverse event: 43.9% with Cerebrolysin versus 30.3% with placebo (P=0.105); mean events per patient 0.894 versus 0.636 (P=0.134). Serious adverse-event incidence did not differ significantly (P=0.381). This pilot was not powered to rule out uncommon harms; the full-text safety section does not give a separate symptomatic-intracranial-hemorrhage or mortality effect estimate.

Figures

ESCAS trial participant flow diagram — *Figure 1.* CONSORT (Consolidated Standards of Reporting Trials) flow diagram. SAE indicates severe adverse event.

Western Aphasia Battery Aphasia Quotient gains at follow-up visits — *Figure 2.* Boxplots of WAB-AQ change from baseline across visits; horizontal lines are medians, diamonds are means, solid whiskers show range excluding outliers, and dashed error lines show SD. Brackets show multiplicity-adjusted between-group Wilcoxon comparisons; *** denotes P<0.001.

NIH Stroke Scale change from baseline over time — *Figure 3.* Boxplots of NIHSS change from baseline across visits; lines are medians, diamonds are means, whiskers show range excluding outliers, and dashed error lines show SD. Brackets mark adjusted between-group comparisons, including *** for P<0.001.

Modified Rankin Scale distributions at successive visits — *Figure 4.* Boxplots of modified Rankin Scale scores across visits; lines are medians, diamonds are means, whiskers show range excluding outliers, and dashed error lines show SD. Brackets show adjusted group and within-group comparisons.

Barthel Index scores over the study visits — *Figure 5.* Boxplots of Barthel Index scores across visits; lines are medians, diamonds are means, whiskers show range excluding outliers, and dashed error lines show SD. Brackets show adjusted group and within-group comparisons.

Criticisms

The primary efficacy estimates in the article body are per-protocol (120/132 assigned); ITT estimates are relegated to the supplement. The text calls the ITT set 123, yet its day-30 arm counts (61+63) total 124; the analysis denominators need reconciliation before relying on exact ITT estimates.
The arms differed by 5.2 baseline WAB-AQ points (50.8 versus 56.0). Although a formal baseline comparison was not significant, residual differences in lesion volume, severity, and cognition were not controlled; regression to the mean can complicate interpretation of change scores.
The absolute day-90 disability and activities-of-daily-living comparisons were not significant after multiplicity correction (mRS P=0.081; Barthel P=0.153). A language-scale response is not proof of lasting independence or quality-of-life benefit.
Two centers, a Romanian-language test and therapy program, right-handed participants with nonfluent aphasia, three intensive IV treatment cycles, pandemic-era remote sessions, and only 90-day follow-up limit transportability and feasibility.
Any adverse events occurred in 43.9% versus 30.3% (P=0.105). Nonsignificance in a small pilot is not evidence of equivalent safety; manufacturer-related author disclosures also warrant transparent independent replication.

Funding

The authors report academic funding from the nonprofit Foundation of the Society for the Study of Neuroprotection and Neuroplasticity (SSNN) and Foundation of the Study of Nanoneurosciences and Neuroregeneration (FSNANO). The paper does not specify the funders’ roles in design, analysis, or publication. Several authors disclose previous trials or honoraria involving Cerebrolysin’s producer, EVER Neuro Pharma; these disclosures are not the same as a declaration of company funding for this trial. Prospectively registered as ISRCTN54581790 (not an NCT study).

The paper

Authors. Homberg V, et al.
Title. Speech Therapy Combined With Cerebrolysin in Enhancing Nonfluent Aphasia Recovery After Acute Ischemic Stroke: ESCAS Randomized Pilot Study.
Journal. Stroke.
Year. 2025.
DOI. 10.1161/STROKEAHA.124.049834
PMCID. PMC11932442

Deep Dive — click to expand

What this is

Published in 2025, ESCAS is a pilot test of whether adding 30 days of IV Cerebrolysin, spread across three cycles, to 30 hours of early speech therapy can improve nonfluent aphasia after left-MCA stroke. The comparison was placebo plus the same therapy, not no rehabilitation. At 90 days, the per-protocol between-group WAB-AQ change difference was 14.805 points (95% CI, 9.521–20.089); the trial is a signal worth testing, not a mandate to prescribe.

1. Shadow Audit

The headline language result does not carry over cleanly to absolute day-90 functional outcomes: adjusted P=0.081 for mRS and P=0.153 for Barthel Index. Safety has a clinically noticeable numerical imbalance—any adverse event in 43.9% versus 30.3%, a 13.6-percentage-point difference—despite P=0.105. More subtly, the main efficacy narrative centers PP n=120, while the stated ITT n=123 conflicts with the 61+63=124 day-30 counts printed in the text. Those issues do not negate the WAB-AQ finding, but they demand an auditable ITT replication with patient-centered endpoints.

2. Inversion Engine

To erase the observed day-90 point estimate, a well-controlled replication would need to shift the between-arm WAB-AQ effect downward by 14.805 points; to bring it below the paper’s cited 5-point within-person importance threshold, the point estimate would need to shrink by more than 9.805 points. The within-person threshold should not be treated as a validated between-arm prescribing cutoff. A less dramatic inversion is possible if the 5.2-point baseline WAB-AQ imbalance and unmeasured lesion burden explain some change-score advantage, or if the language gain fails to persist at six months. An adequately powered blinded ITT study with baseline-adjusted analysis could test those alternatives.

3. Second-Order Catalyst

The first plausible change is at stroke rehabilitation research units able to randomize patients within five days, standardize speech therapy, and arrange 30 infusions—not a general ward formulary change. Such units can add blinded WAB-AQ, patient-reported communication, 90-day disability, six-month durability, and adverse events to a confirmatory protocol. No new routine referral or infusion pathway is justified by a two-center pilot alone.

4. Asymmetric Leverage

A roughly 15-point language-scale advantage could be valuable to a person who cannot communicate fluently, but the trial addresses a narrow denominator: right-handed, Romanian-speaking patients with nonfluent aphasia after left-MCA ischemia who can start treatment at day 3–5. The proposed intervention is not a trivial add-on—it requires thirty IV administrations alongside thirty therapy hours. If larger trials replicate durable, patient-perceived communication gains, the clinical upside may justify that burden for a defined subgroup; extrapolating the effect to all aphasia, all stroke survivors, or all languages would overstate the evidence.

5. Paradigm Destroyer

The useful habit to discard is equating a statistically impressive language score with a proven gain in independent living—or treating an adjusted P>0.05 safety comparison as proof that harms are absent. A two-sentence protocol update for tomorrow: Continue evidence-based speech and language therapy for eligible stroke patients; do not add Cerebrolysin routinely on the strength of ESCAS alone. Offer trial enrollment where available, with blinded language assessment, prespecified ITT analysis, functional endpoints, and longer follow-up.

MVP — Minimum Viable Proof

Replicate a clinically meaningful between-arm WAB-AQ benefit in a larger randomized ITT cohort, show that patients experience better communication at six months, and rule out a material adverse-event penalty. ESCAS’s 90-day PP difference of 14.805 points supplies an effect-size hypothesis, not those three confirmations.

Best Combination

Keep structured speech and language therapy as the active foundation, then treat ESCAS and the earlier small Broca-aphasia Cerebrolysin study it cites as reasons to run an independent adjunct-therapy trial, not as replacements for rehabilitation. Predefine dose, therapy intensity, baseline-adjusted language and patient-reported outcomes, and long-term function before unblinding; the older supporting study and ESCAS share a narrow aphasia setting rather than establishing broad stroke-recovery efficacy.

Overvalue Warning

First, 14.805 WAB-AQ points is a mean PP change difference at 90 days, not the proportion who regained useful conversation, returned to work, or became independent. Second, P=0.105 for adverse events is not a clean safety bill: 43.9% versus 30.3% with only 132 assigned cannot rule out a meaningful increase in harm.

Educational use notice. The Clotbust Review is dedicated to medical education. Figures from published articles are embedded for teaching purposes, with attribution to the original publication. We do not claim copyright over any embedded figure; all rights remain with the original publisher and authors under the article's published license. If you are a rights holder and would like a figure removed, please contact the editorial team. See our editorial notice for the full policy.