TOP 67 - Factors Associated with the Utilization of Extracorporeal Life Support in Patients with High-Risk Diagnoses

Publication Number: 1826.TOP 67

Zachary A. Rumlow, University of Michigan Medical School, Ann Arbor, MI, United States; Ryan Barbaro, University of Michigan Medical School, Ann Arbor, MI, United States; Stephen Gorga, University of Michigan Medical School, Ann Arbor, MI, United States; Erin Carlton, University of Michigan Medical School, Ann Arbor, MI, United States; Acham Gebremariam, University of Michigan Medical School, Ann Arbor, MI, United States

Background: Extracorporeal life support (ECLS) utilization is increasing as indications broaden and contraindications shrink; updated Extracorporeal Life Support Organization (ELSO) guidelines avoid the use of absolute ECLS contraindications. For many conditions with historical contraindications, there was near equipoise among surveyed providers regarding whether they would offer ECLS. Moreover, recent data suggests that there has been increasing utilization of ECLS in patients with diagnoses which had previously been considered absolute contraindications to ECLS (trauma, trisomy 13 and 18, and oncologic diagnoses including hematopoietic stem cell transplant (HSCT)). Altogether this evolution belies broadening ECLS utilization and a lack of uniformity in ECLS patient selection provides an opportunity to identify factors associated with ECLS utilization in patients traditionally thought to not be ECLS candidates.

Objective: We aim to evaluate if there is hospital variation in the proportion of patients with a trauma, trisomy 13, trisomy 18, cancer or HSCT diagnosis receiving ECLS. We also aim to test to test if patient and hospital factors are associated with differential utilization of ECLS in these patients.

Design/Methods: We utilized the Pediatric Health Information System (PHIS) database of 49 children’s hospitals to evaluate ECLS cases between January 1, 2016 and December 31, 2024. A total of 10,816 patients at all 49 PHIS sites were identified as eligible for this analysis. Patients were eligible for inclusion if they had diagnostic codes consistent with the diagnoses of interest and received ECLS. Patients and ECLS use were identified using established diagnostic and procedural codes in the 10th iteration of the International Statistical Classification of Diseases and Related Health Problems (ICD-10). We will use multilevel regression analyses to identify factors that are associated with differences in the ratio of ECLS administered to patients with diagnoses of interest.