Highlight
In a multicenter Japanese registry of pediatric out-of-hospital cardiac arrest, extracorporeal cardiopulmonary resuscitation was associated with higher 1-month survival and better neurologic outcome than continued conventional CPR among patients considered at risk for receiving ECPR.
The apparent benefit was concentrated in a highly selected population: mostly adolescents, predominantly witnessed arrests, and largely cardiogenic etiologies treated at pediatric ECPR-capable centers.
The study used time-dependent propensity score risk-set matching, an important design feature that attempts to reduce resuscitation time bias, but the final ECPR sample was small and confidence intervals were wide.
These findings support clinical equipoise for prospective studies and may help shape inclusion criteria for future pediatric ECPR trials rather than justify indiscriminate expansion of ECPR programs.
Background
Pediatric out-of-hospital cardiac arrest remains an uncommon but devastating emergency. Survival is generally poor, and favorable neurologic recovery is even less frequent, especially when return of spontaneous circulation is not achieved rapidly. Conventional CPR can sustain limited organ perfusion for a period of time, but prolonged low-flow states substantially increase the risk of irreversible hypoxic-ischemic brain injury, multiorgan dysfunction, and death.
Extracorporeal cardiopulmonary resuscitation, typically referring to venoarterial extracorporeal membrane oxygenation initiated during refractory cardiac arrest, has emerged as a rescue strategy intended to restore systemic perfusion when standard resuscitation fails. In adults, especially in selected cases of presumed reversible or cardiogenic arrest, the evidence base for ECPR has grown considerably. However, pediatric evidence has lagged behind. Most existing pediatric data come from in-hospital cardiac arrest cohorts, single-center experiences, or mixed observational series. Comparative data for pediatric out-of-hospital cardiac arrest have been particularly sparse.
This gap is clinically important. The pathophysiology of pediatric arrest differs from adult arrest, with a higher proportion of respiratory, asphyxial, drowning, trauma, congenital heart disease, and nonshockable rhythm etiologies. Pediatric out-of-hospital arrest also presents substantial logistical barriers for ECPR: transport delays, low case volumes, need for specialized cannulation expertise, uncertainty about candidacy, and concerns regarding neurologic futility after prolonged downtime. As a result, clinicians need better evidence to identify whether there is a subgroup in whom ECPR offers meaningful benefit.
Study Design and Methods
Design and data source
Amagasa and colleagues conducted a retrospective cohort study using a multicenter Japanese out-of-hospital cardiac arrest registry from 2014 to 2022. The analysis focused on patients younger than 18 years who were transported to pediatric ECPR-capable institutions.
Population
The cohort included 799 pediatric OHCA patients treated at centers with the capability to initiate pediatric ECPR. Of these, 27 ultimately received ECPR. Because the key clinical question is whether ECPR improves outcomes compared with ongoing conventional CPR in patients who could still plausibly receive ECPR, the authors did not simply compare all ECPR patients with all non-ECPR patients. Instead, they constructed an at-risk comparison group consisting of patients who had not yet received ECPR and remained eligible to receive it later during resuscitation.
Exposure and comparator
The exposure was initiation of extracorporeal cardiopulmonary resuscitation. The comparator was continued conventional CPR among patients who were at risk of receiving ECPR at the same time point in the resuscitation course.
Outcomes
The primary outcomes were 1-month survival and favorable neurologic outcome at 1 month, defined as Pediatric Cerebral Performance Category 1 to 3. This neurologic endpoint captures normal to moderate disability and is commonly used in pediatric resuscitation research.
Analytic approach
A major methodological challenge in ECPR research is timing. Patients must survive long enough during conventional resuscitation to receive ECPR; therefore, naive comparisons can introduce substantial resuscitation time bias or immortal time bias. To address this, the investigators used risk-set matching with time-dependent propensity scores and full matching with up to four controls per ECPR case. This is an appropriate and sophisticated observational approach because it attempts to compare patients at similar moments during arrest management, rather than comparing early deaths with patients who had enough physiologic reserve and logistical opportunity to reach extracorporeal support.
After matching, 27 ECPR patients were compared with 108 at-risk controls.
Key Results
Who received ECPR?
The ECPR group represented a highly selected phenotype. Median age was 14 years, indicating that adolescents accounted for much of the treated population. Arrests were witnessed in 70.4% of ECPR patients, and cardiogenic causes accounted for 74.1%. These details matter because they suggest ECPR was not being broadly applied across the full pediatric OHCA spectrum. Rather, it was used in patients whose arrest circumstances may have been considered more potentially reversible and more technically suitable for extracorporeal rescue.
The authors report that baseline characteristics were similar after matching, which supports the internal validity of the comparison, although unmeasured confounding remains possible.
Survival
One-month survival was 25.9% in the ECPR group, corresponding to 7 of 27 patients, versus 11.1% in matched controls, corresponding to 12 of 108 patients. The reported risk difference was 17.3% with a 95% confidence interval of -0.9% to 35.6%. The risk ratio was 3.56 with a 95% confidence interval of 1.37 to 9.28.
Clinically, this is an important signal. A near 15-point absolute difference and more than threefold relative increase in survival would be highly meaningful in a condition with otherwise poor outcomes. Statistically, however, interpretation requires nuance. The confidence interval for the risk ratio excludes 1, whereas the confidence interval for the risk difference crosses 0. This pattern reflects the small sample size and instability of absolute effect estimates. The direction of effect is encouraging, but the precision is limited.
Neurologic outcome
Favorable neurologic outcome at 1 month occurred in 18.5% of ECPR patients, or 5 of 27, compared with 6.5% of controls, or 7 of 108. The risk difference was 13.9% with a 95% confidence interval of -2.9% to 30.8%. The risk ratio was 3.78 with a 95% confidence interval of 1.19 to 11.99.
This finding is arguably more important than survival alone. ECPR is meaningful only if it can preserve neurologically acceptable survival in at least some children. The observed absolute rate of favorable neurologic outcome remained modest, but in refractory pediatric OHCA even this magnitude of benefit may be clinically significant if reproducible. Again, the wide confidence intervals underscore uncertainty.
Interpretive table
Outcome comparison after matching:
1-month survival: ECPR 25.9% versus continued CPR 11.1%; risk difference 17.3% (95% CI, -0.9 to 35.6); risk ratio 3.56 (95% CI, 1.37 to 9.28).
Favorable neurologic outcome: ECPR 18.5% versus continued CPR 6.5%; risk difference 13.9% (95% CI, -2.9 to 30.8); risk ratio 3.78 (95% CI, 1.19 to 11.99).
Clinical Interpretation
This study provides one of the clearest comparative signals to date that pediatric ECPR may improve outcomes after out-of-hospital cardiac arrest in carefully selected children. The most important practical implication is not that all pediatric OHCA patients should be considered for extracorporeal support, but that there may be a definable subgroup in whom ECPR is worth mobilizing.
The observed treatment profile offers clues to that subgroup. Patients were largely adolescents with witnessed arrest and presumed cardiogenic etiology. That resembles the adult literature more than the broader pediatric OHCA population. In real-world terms, an older child or teenager with a witnessed collapse, rapid EMS activation, suspected cardiac cause, and ongoing refractory arrest despite high-quality CPR may be the clinical scenario in which ECPR is most biologically plausible and operationally feasible.
The mechanism is intuitive. ECPR can re-establish near-normal systemic perfusion, decompress failing cardiopulmonary physiology, and buy time to treat reversible causes such as acute myocarditis, primary arrhythmia, coronary anomaly, pulmonary embolic physiology, drug toxicity, or other treatable cardiogenic states. In contrast, prolonged conventional CPR often delivers marginal cerebral and coronary perfusion, particularly outside the hospital environment.
Still, enthusiasm should be tempered. The final treated sample was only 27 patients over eight years across multiple centers. This indicates how uncommon pediatric ECPR for OHCA remains and how difficult it will be to generate robust evidence. It also raises concerns about center effect. Institutions capable of offering ECPR may differ in many ways beyond the technology itself, including pediatric intensivist staffing, cannulation expertise, postarrest neurocritical care, rehabilitation pathways, and patient selection discipline.
Methodological Strengths
The study’s strongest feature is its attempt to emulate a target trial using time-sensitive matching. Much of the earlier ECPR literature has been vulnerable to bias because controls included patients who died before ECPR could ever be initiated. By using risk-set matching with time-dependent propensity scores, the investigators reduced, though did not eliminate, this bias.
The multicenter registry design also increases relevance compared with a single-center case series. The focus on clinically meaningful patient-centered outcomes, especially neurologic status at 1 month, is another strength.
Limitations and Sources of Uncertainty
Residual confounding
As the authors appropriately note, observational comparisons cannot fully eliminate confounding. Clinicians likely selected ECPR for patients perceived to have greater reversibility, better premorbid status, more favorable transport conditions, or better CPR quality. Even advanced matching cannot account for unmeasured factors such as low-flow quality, biochemical severity, no-flow duration, cannulation success, or subtle bedside judgments about futility.
Small sample size
Only 27 ECPR-treated patients were available. This severely limits precision and makes subgroup analyses difficult or impossible. A few additional events could meaningfully shift the effect estimates. The wide confidence intervals reflect this fragility.
Generalizability
These data come from Japan and from pediatric ECPR-capable institutions. Systems of emergency response, transport geography, prehospital care, availability of pediatric surgeons or ECMO teams, and thresholds for ECPR initiation differ across countries. The findings may not translate directly to regions without centralized pediatric critical care pathways or rapid extracorporeal deployment teams.
Population selection
The cohort appears enriched for adolescents and cardiogenic arrest. Younger children, infants, unwitnessed arrests, asphyxial mechanisms, drowning, trauma, and prolonged no-flow intervals may have very different risk-benefit profiles. The study should not be interpreted as evidence for routine ECPR in those populations.
Outcome timeframe
The primary endpoint was 1 month. While appropriate and commonly used, longer-term neurologic, developmental, and quality-of-life outcomes would provide a more complete picture in pediatric survivors.
How This Fits With Existing Literature and Guidelines
Current pediatric resuscitation guidelines have generally supported considering ECPR for in-hospital cardiac arrest in experienced centers when reversible causes are present, but evidence for pediatric OHCA has been much weaker. The 2020 American Heart Association Pediatric Advanced Life Support guidance emphasized that ECPR may be considered for patients with cardiac diagnoses in settings with appropriate expertise and systems, while acknowledging limited evidence, particularly outside the hospital.
Observational studies in pediatric in-hospital cardiac arrest have suggested benefit from ECPR in selected cardiac patients, especially those with congenital heart disease or postoperative arrests. By contrast, pediatric OHCA evidence has mostly consisted of small series and mixed cohorts with uncertain comparators. This new study therefore fills an important gap by offering a more methodologically rigorous comparison.
The adult ECPR literature is relevant but cannot be imported wholesale into pediatrics. Adult randomized and high-quality observational studies have mainly focused on refractory ventricular fibrillation or cardiogenic arrest, often with streamlined transport and protocolized cannulation pathways. Pediatric OHCA is more heterogeneous, and therefore likely requires narrower inclusion criteria for any future trial.
Implications for Practice and Research
For clinicians, the study supports development of systems-based thinking rather than case-by-case improvisation alone. Centers interested in pediatric ECPR should define candidacy criteria prospectively, integrate EMS and transport pathways, and ensure that extracorporeal initiation is linked to comprehensive postarrest care. ECPR is not a stand-alone intervention; its success depends on the entire chain of survival.
For researchers, the findings help identify a plausible enrichment strategy for future trials: adolescents or older children with witnessed, likely cardiogenic OHCA, rapid transport, minimal no-flow time, and persistent refractory arrest despite high-quality CPR. A pragmatic randomized trial may be difficult given low case numbers, but international collaboration, registry-based trial infrastructure, and standardized data definitions could make it feasible.
Key unanswered questions include optimal patient selection, timing thresholds, transport versus on-arrival cannulation strategies, role of mechanical CPR during transport, integration with targeted temperature management and neuroprognostication, and how to define futility without excluding potentially salvageable patients.
Funding and Trial Registration
The abstract provided does not report funding information. No ClinicalTrials.gov registration applies to this retrospective observational registry study.
Conclusion
This multicenter Japanese registry analysis suggests that extracorporeal cardiopulmonary resuscitation may improve both survival and neurologic outcome after pediatric out-of-hospital cardiac arrest when compared with continued conventional CPR among patients at risk of receiving ECPR. The signal is clinically meaningful, but the evidence remains preliminary because the number of treated patients was small and residual confounding cannot be excluded.
The study’s main contribution is not definitive proof of efficacy, but a stronger rationale for disciplined patient selection and prospective evaluation. At present, pediatric ECPR for OHCA should be viewed as a specialized rescue strategy for carefully chosen patients in experienced systems, rather than a broadly generalizable standard of care.
References
Amagasa S, Iwamoto S, Okubo M, Utsumi S, Kashiura M, Yasuda H, Kishihara Y, Uematsu S. Extracorporeal Cardiopulmonary Resuscitation for Pediatric Out-of-Hospital Cardiac Arrest. Ann Emerg Med. 2026 May 4. PMID: 42084583.
Topjian AA, Raymond TT, Atkins D, et al. Part 4: Pediatric Basic and Advanced Life Support: 2020 American Heart Association Guidelines for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care. Circulation. 2020;142(16_suppl_2):S469-S523.
Holmberg MJ, Wiberg S, Ross CE, et al. Extracorporeal cardiopulmonary resuscitation for cardiac arrest: A systematic review. Resuscitation. 2018;131:91-100.
Ortmann L, Prodhan P, Gossett J, et al. Outcomes after in-hospital cardiac arrest in children with cardiac disease: a report from Get With The Guidelines-Resuscitation. Circulation. 2011;124(21):2329-2337.
Lasa JJ, Rogers RS, Localio R, et al. Extracorporeal cardiopulmonary resuscitation in the pediatric cardiac population: in-hospital outcomes and predictors of mortality. Pediatr Crit Care Med. 2016;17(8):698-706.
