Survival After Extracorporeal Cardiopulmonary Resuscitation Based on In-Hospital Cardiac Arrest and Cannulation Location: An Analysis of the Extracorporeal Life Support Organization Registry

OBJECTIVES: Explore whether extracorporeal cardiopulmonary resuscitation (ECPR) mortality differs by in-hospital cardiac arrest location and whether moving patients for cannulation impacts outcome. DESIGN: Retrospective cohort study. SETTING: ECPR hospitals that report data to the Extracorporeal Life Support Organization (ELSO). PATIENTS: Patients having ECPR for in-hospital cardiac arrest between 2020 and 2023 with data in the ELSO registry. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Patient demographics, comorbidities, pre-cardiac arrest conditions, pre-ECPR vasopressor use, cardiac arrest details, ECPR cannulation information, major complications, and in-hospital mortality were recorded. Multivariable logistic regression model was used to examine the associations between in-hospital mortality and 1) cardiac arrest location and 2) moving a patient for ECPR cannulation. A total of 2515 patients met enrollment criteria. The adjusted odds ratio (aOR) for mortality was increased in patients who had a cardiac arrest in the ICU (aOR, 1.85; 95% CI, 1.45–2.38; p < 0.001) and in patients who had a cardiac arrest in an acute care bed (aOR, 1.68; 95% CI, 1.09–2.58; p = 0.02) compared with the cardiac catheterization laboratory. Moving a patient for cannulation had no association with mortality (aOR, 0.70; 95% CI, 0.18–2.81; p = 0.62). Advanced patient age was associated with increased mortality. Specifically, patients 60–69 and patients 70 years old or older were more likely to die compared with patients younger than 30 years old (aOR, 1.71; 95% CI, 1.17–2.50; p = 0.006 and aOR, 2.27; 95% CI, 1.49–3.48; p < 0.001, respectively). CONCLUSIONS: ECPR patients who experienced cardiac arrest in the ICU and in acute care hospital beds had increased odds of mortality compared with other locations. Moving patients for ECPR cannulation was not associated with improved outcomes.