Optimizing Extracorporeal Cardiopulmonary Resuscitation Candidate Selection in out-of-Hospital Cardiac Arrest: A
Chi-Hsin Chen1,2,3, Edward Pei-Chuan Huang1,3,4, Chih-Wei Sung1,3
1Department of Emergency Medicine National Taiwan University Hospital Hsin-Chu Branch Hsinchu Taiwan.
Journal of the American Heart Association
|June 15, 2026
Summary
Machine learning models can identify patients most likely to benefit from extracorporeal cardiopulmonary resuscitation (ECPR) for out-of-hospital cardiac arrest. This data-driven approach shows greater survival benefit than current selection criteria.
Area of Science:
- Cardiology
- Intensive Care Medicine
- Data Science in Medicine
Background:
- Extracorporeal cardiopulmonary resuscitation (ECPR) improves survival in select out-of-hospital cardiac arrest (OHCA) patients.
- Optimal ECPR candidate selection criteria remain uncertain, highlighting the need for improved strategies.
- Machine learning-based individualized treatment effect (ITE) modeling offers a potential solution by capturing treatment response heterogeneity.
Purpose of the Study:
- To develop and evaluate a machine learning-based ITE model for predicting ECPR survival benefit.
- To compare the performance of the ITE model against current rule-based criteria for ECPR candidate selection.
Main Methods:
- Retrospective analysis of adult, non-traumatic OHCA patients from four Taiwanese tertiary centers (2016-2024).
- Propensity score matching for shockable rhythm and witnessed arrest.
- Development of a gradient-boosted trees-based causal forest model to estimate ITE and predict ECPR survival benefit.
Main Results:
- Among 1953 matched patients, 977 received ECPR; overall survival was similar (11.1% vs 12.8%).
- In the top 10% predicted benefit group, ECPR yielded 50.0% survival vs 20.0% without (absolute observed treatment effect 30.0%, P=0.042).
- ITE model identified subgroups with higher survival benefit than rule-based criteria; factors included lower pH/PaCO2, higher lactate, younger age, and bystander CPR.
Conclusions:
- The ITE-based model demonstrated greater observed survival benefit compared to current rule-based criteria.
- This data-driven framework shows potential for optimizing ECPR candidate selection.
- Further validation in prospective studies is warranted.
Related Concept Videos
01:25Cardiopulmonary Resuscitation IV: Pharmacological Management
Pharmacologic intervention is crucial in treating cardiac arrest patients during ACLS or Advanced Cardiovascular Life Support. The ACLS algorithms guide the administration of specific drugs based on the patient's cardiac arrest rhythm, which includes pulseless ventricular tachycardia (VT), ventricular fibrillation (VF), asystole, and pulseless electrical activity (PEA).EpinephrineIndication: Epinephrine is the first-line drug for all cardiac arrest rhythms.Mechanism of Action: Epinephrine...
01:23Cardiopulmonary Resuscitation III: AED Use
Introduction to AEDAn Automated External Defibrillator (AED) is a portable medical device that analyzes the heart's rhythm and, if necessary, delivers an electrical shock to help the heart re-establish an effective rhythm during sudden cardiac arrest (SCA). SCA occurs when the heart suddenly and unexpectedly stops beating, leading to a loss of blood flow to the brain and other vital organs. In such emergencies, time is of the essence, and using an AED, combined with Cardiopulmonary...
01:21Cardiopulmonary Resuscitation I: Adult
Cardiopulmonary resuscitation, or CPR, is a life-saving emergency procedure performed when a person's heart has stopped beating or they are no longer breathing. The foundation of CPR is Basic Life Support (BLS), which focuses on the early recognition of cardiac arrest, the immediate start of high-quality chest compressions, and the timely use of an automated external defibrillator (AED).Assessing Responsiveness and Checking the Carotid PulseWhen approaching an unresponsive person, first ensure...