Patterns of Multiple Organ Dysfunction in Children on Extracorporeal Membrane Oxygenation Support

Sunday, March 22, 2026

9:45 AM - 10:45 AM Central Time

Location: Connections Central - RST 07

Victoria Huang, MD (she/her/hers)

Johns Hopkins Hospital

Disclosure(s): No relevant financial relationship(s) to disclose.
JH

Justin X. Huang, MD, MS

Resident Physician
University of California San Francisco

Disclosure information not submitted.
JR

Jennifer Roem, MSc

Senior Biostatistician
Johns Hopkins University

Disclosure information not submitted.
DN

Derek Ng, PhD

Associate Professor of Epidemiology
Johns Hopkins University

Disclosure information not submitted.
Melania M. Bembea, MD, MPH, PhD

Johns Hopkins University, Maryland

Disclosure information not submitted.

Introduction: In this retrospective cohort study of children on extracorporeal membrane oxygenation (ECMO) support, we characterized patterns of multiple organ dysfunction and their associations with in-hospital mortality.

Methods: Electronic health records data were extracted for all consecutive children < 18 years on ECMO at a single U.S. academic center between 2011 and 2024. Organ dysfunction (OD) was assessed using Pediatric Organ Dysfunction Information Update Mandate (PODIUM) criteria, as well as Pediatric Sequential Organ Failure Assessment (pSOFA) and Pediatric Logistic Organ Dysfunction-2 (PELOD-2) scores. These metrics were evaluated in 24-hour intervals from 14 days prior to cannulation through 28 days after cannulation.

Results: A total of 317 patients with a median age of 3.9 months were supported on ECMO with primary pulmonary (43%), cardiac (28%), or extracorporeal cardiopulmonary resuscitation (29%) ECMO indications. Cumulative incidence of in-hospital mortality was 46% and median time-to-death was 9 days [IQR 1-34]. Prior to cannulation, the median number of concurrent PODIUM OD was the same between survivors to hospital discharge and non-survivors (2 organs [IQR 1-4]). Accumulation of more OD among non-survivors vs survivors began shortly after cannulation. On ECMO day 0, a median of 7 organs [IQR 6-8] for non-survivors vs 6 [IQR 5-7] for survivors (p < 0.05) were observed in dysfunction. These differences persisted while on ECMO, with a median of 6 organs [IQR 5-7] for non-survivors vs 5 [IQR 4-6] for survivors (p < 0.05), and widened after decannulation, with a median of 4 organs [IQR 3-5] for non-survivors vs 1 [IQR 0-3] for survivors (p < 0.05). pSOFA and PELOD-2 scores followed the same pattern and were statistically significant, though differences between the two groups were smaller. Maximum number of concurrent PODIUM OD while on ECMO showed the best prognostic performance for in-hospital mortality at day 10 after cannulation (AUC 0.67).

Conclusions: Differences in the number of concurrent OD, as measured by PODIUM criteria, pSOFA and PELOD-2 scores, between survivors and non-survivors only began to emerge shortly after cannulation, suggesting that identifying intervenable risk factors pre- and on-ECMO could affect mortality.