Critical Care Unit Characteristics and Extracorporeal Cardiopulmonary Resuscitation Survival in the Pediatric Cardiac Population: Retrospective Analysis of the Virtual Pediatric System Database

OBJECTIVES: Existing literature provides limited data about ICU characteristics and pediatric extracorporeal cardiopulmonary resuscitation (E-CPR) outcomes. We aimed to evaluate the associations between patient and ICU characteristics, and outcomes after E-CPR in the pediatric cardiac population. DESIGN: Retrospective analysis of the Virtual Pediatric System database (VPS, LLC, Los Angeles, CA). SETTING: PICUs categorized as either cardiac-only versus mixed ICU cohort type. PATIENTS: Consecutive cardiac patients less than 18 years old experiencing cardiac arrest in the ICU and resuscitated using E-CPR. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Event and time-stamp filtering identified E-CPR events. Patient, hospital, and event-related variables were aggregated for independent and multivariable mixed effects logistic regression to assess the association between ICU cohort type and survival. Among ICU admissions in the VPS database, 2010-2018, the prevalence of E-CPR was 0.07%. A total of 671 E-CPR events (650 patients) comprised the final cohort; congenital heart disease (84%) was the most common diagnosis versus acquired heart diseases. The majority of E-CPR events occurred in mixed ICUs (67%, n = 449) and in ICUs with greater than 20 licensed bed capacity (65%, n = 436). Survival to hospital discharge was 51% for the overall cohort. Independent logistic regression failed to reveal any association between survival to hospital discharge and ICU type (ICU type: cardiac ICU, odds ratio [OR], 1.01; 95% CI, 0.71-1.44; p = 0.95). However, multivariable logistic regression revealed an association between cardiac surgical patients and greater odds for survival (OR, 2.03; 95% CI, 1.40-2.95; p < 0.001). Also, there was an association between ICUs with capacity greater than 20 (vs not) and lower survival odds (OR, 0.65; 95% CI, 0.43-0.96). CONCLUSIONS: The overall prevalence of E-CPR among critically ill children with cardiac disease observed in the VPS database is low. We failed to identify an association between ICU cohort type and survival. Further investigation into organizational factors is warranted.