Highlight
• In a prospective, multicenter binational registry (EXCEL), neurologic complications occurred in 12% of adults supported with venoarterial (VA) ECMO and were associated with a 17.3% absolute increase in the composite outcome of death or new disability at 6 months (risk difference [RD] 17.28%; 95% CI 6.44–25.92%).
• The excess in the composite outcome was driven primarily by higher mortality at 6 months (RD 23.75%; 95% CI 12.06–34.75%) rather than by an increased rate of new disability among survivors.
• Patients who sustained neurologic complications were more likely to have ECMO discontinued because of a perceived poor prognosis (adjusted odds ratio 2.7; 95% CI 1.35–24.7), suggesting prognosis-driven treatment decisions may mediate some of the association with mortality.
Background
Venoarterial extracorporeal membrane oxygenation (VA‑ECMO) is an increasingly used rescue therapy for refractory cardiogenic shock and cardiac arrest. While it can be life‑saving, VA‑ECMO carries substantial complications, including neurologic injury such as diffuse cerebral ischemia, ischemic stroke, intracranial hemorrhage, and progression to brain death. Reported incidence rates vary across series owing to heterogeneous populations, definitions, and surveillance strategies. Most prior studies have focused on in‑hospital mortality or short‑term neurologic events; less is known about medium‑ to long‑term patient‑centered outcomes (survival with functional status and quality of life) after VA‑ECMO when neurological complications occur. Understanding these relationships is essential to inform clinical decision‑making, prognostication, monitoring strategies, and resource allocation for rehabilitation.
Study design and methods
Brown and colleagues report a prospective, multicenter observational analysis from the EXCEL registry involving 30 ECMO centers in Australia and New Zealand between February 2019 and December 2022. The cohort included 704 adult ICU patients supported with VA‑ECMO; patients receiving extracorporeal cardiopulmonary resuscitation (ECPR) were excluded. No experimental interventions were tested.
Neurologic complications were prespecified and included diffuse cerebral ischemia, ischemic stroke, cerebral hemorrhage, and brain death. The primary endpoint was a patient‑centered composite of death or new disability measured at 6 months. Disability assessments used validated instruments including the World Health Organization Disability Assessment Schedule (WHODAS) and health‑related quality of life using EuroQol (EQ‑5D‑5L). Analyses compared outcomes between patients who experienced neurologic complications during their index ECMO run and those who did not; secondary analyses adjusted for confounders and examined specific complication subtypes and processes of care such as withdrawal of ECMO for a perceived poor prognosis.
Key findings
Population: Median age was 54.5 years (IQR 42–64); 36.8% were female. Primary outcome data were available for 613 of 706 patients (86.2% follow‑up).
Incidence of neurologic complications: 85 patients (12%) experienced at least one neurologic complication during VA‑ECMO.
Primary outcome: Among the 613 patients with outcome data, 425 (69.3%) had either died or had new disability at 6 months. Patients with neurologic complications had an increased risk of the composite outcome compared with those without complications (risk difference 17.28%; 95% CI 6.44–25.92%).
Mortality versus disability: The excess in the composite outcome was driven mainly by higher mortality at 6 months in the neurologic complication group (RD 23.75%; 95% CI 12.06–34.75%). New disability among survivors did not appear to be higher in those with neurologic complications (RD −11.90%; 95% CI −30.58% to 13.56%), noting the confidence interval crossed zero and the study may be underpowered to detect smaller differences in functional outcomes among survivors.
Consistency and adjustment: The association between neurologic complications and worse 6‑month outcomes was consistent across different types of neurologic injury and after adjustment for measured confounders.
Treatment limitation: Patients with neurologic complications were more likely to have ECMO discontinued for a perceived poor prognosis (odds ratio 2.7; 95% CI 1.35–24.7), implying that decisions to limit or withdraw life‑sustaining therapy may account for a substantial portion of the higher mortality observed.
Interpretation and clinical implications
This large, prospectively collected, multicenter cohort provides robust evidence that neurologic complications during VA‑ECMO are a clinically important prognostic marker for worse 6‑month outcomes, chiefly due to increased mortality. Several implications follow for clinicians, centres, and researchers.
Clinical monitoring and prevention
Given a 12% incidence of recognized neurologic events and the clear association with worse outcomes, centres should maintain high vigilance for neurologic injury in VA‑ECMO patients. Prevention strategies are multifaceted and include careful anticoagulation management to balance thrombotic and bleeding risks, hemodynamic optimization to avoid cerebral hypoperfusion, meticulous cannula positioning to minimize embolic risk, and minimization of potentially neurotoxic exposures. Standardized protocols for neurologic surveillance — frequent neurologic examinations adapted to sedation level, early use of brain imaging when indicated, standardized use of neuromonitoring tools (continuous EEG, transcranial Doppler or near‑infrared spectroscopy where available) — may enable earlier detection and targeted interventions.
Prognostication and decision making
The finding that patients with neurologic complications had higher rates of ECMO withdrawal for perceived poor prognosis raises a critical and familiar issue in critical care: the interplay of objective neurologic injury, prognostic uncertainty, physician and surrogate expectations, and care limitation practices. Observational data are susceptible to confounding by indication and self‑fulfilling prophecy when care is withdrawn on the basis of perceived poor prognosis. Therefore, decisions regarding withdrawal should be made using standardized prognostic criteria, multidisciplinary input including neurocritical care when available, and with transparent shared decision‑making that communicates uncertainty and possible trajectories to families.
Rehabilitation and survivorship
Although this study did not show a higher rate of new disability among survivors with neurologic complications, follow‑up was limited to 6 months and missing data exist. Survivors of ECMO often face physical, cognitive and psychological sequelae; therefore, structured post‑ICU follow‑up and early referral to neurorehabilitation programs remain important components of care for patients with and without recognized neurologic injury.
Study strengths and limitations
Strengths: The prospective multicenter design, relatively large sample size for an adult VA‑ECMO cohort, use of validated patient‑centered outcome measures at 6 months, and adjusted analyses enhance the credibility and clinical relevance of the findings.
Limitations: As an observational study, residual confounding is possible and causal inference is limited. Neurologic events likely varied in ascertainment across centers (differences in monitoring intensity, imaging thresholds, and neurologic expertise), which could bias incidence estimates and associations. Missing outcome data (approximately 14%) may affect estimates if losses were nonrandom. The strong influence of treatment limitation practices complicates interpretation: higher mortality may reflect the severity of neurologic injury, but it may also be influenced by early prognostic pessimism and withdrawal decisions. Finally, functional outcomes were assessed to 6 months; longer‑term recovery trajectories remain unknown.
Future research directions
Priority areas include:
- Standardization of neurologic monitoring, event definitions, and reporting within ECMO registries to allow more precise incidence and outcome estimates.
- Development and validation of prognostic models that predict functional outcomes (not just mortality) after neurologic complications on ECMO, incorporating clinical features, neuroimaging, electrophysiology, and biomarkers.
- Interventional studies testing strategies to prevent neurologic complications (e.g., anticoagulation protocols, cannulation strategies) where equipoise exists; pragmatic trials or stepped‑wedge designs may be feasible.
- Prospective studies of neurorehabilitation interventions for ECMO survivors and longer‑term follow‑up studies focused on quality of life and cognitive outcomes.
Conclusion
In this large, prospective binational cohort, neurologic complications during VA‑ECMO were associated with a meaningful absolute increase in the composite of death or new disability at 6 months, largely explained by higher mortality. These findings underscore the need to prioritize prevention, harmonized neurologic monitoring, cautious prognostication, and organized survivorship care. Multicenter collaboration and standardized outcome measurement will be essential to reduce neurologic harm and to improve patient‑centered outcomes after ECMO.
Funding and clinicaltrials.gov
The authors report their funding and trial registration details in the primary publication (Brown et al., Crit Care Med. 2025). Readers should consult the original manuscript for specific funding sources and any clinical trial registry numbers.
References
1. Brown A, Dennis M, Rattan N, Nanjayya V, Burrell A, Serpa Neto A, Hodgson C; EXCEL Study Investigators and the International ECMO Network. Neurological Complications During Venoarterial Extracorporeal Membrane Oxygenation and Their Implications for 6‑Month Patient‑Centered Outcomes. Crit Care Med. 2025 Nov 10. doi: 10.1097/CCM.0000000000006938 . Epub ahead of print. PMID: 41212045 .
2. World Health Organization. WHO Disability Assessment Schedule (WHODAS 2.0). WHO; 2010. Available at: https://www.who.int/tools/who-disability-assessment-schedule
3. EuroQol Group. EuroQol — a new facility for the measurement of health‑related quality of life. Health Policy. 1990;16(3):199–208.
4. Extracorporeal Life Support Organization (ELSO). ELSO Guidelines and Resources. Available at: https://www.elso.org