Clinical Predictive Tool for Pediatric Cardiac Patients on Extracorporeal Membrane Oxygenation Therapy and

Jennifer Sedler1, Scott M Sutherland2, Amanda M Uber2

  • 1From the Department of Pediatric Hospital Medicine, Stanford University School of Medicine.

ASAIO Journal (American Society for Artificial Internal Organs : 1992)
|March 22, 2023
PubMed

Insights

Fluid overload in pediatric cardiac patients on extracorporeal membrane oxygenation (ECMO) requiring continuous renal replacement therapy (CRRT) is linked to worse outcomes compared to ultrafiltration (UF) alone. A predictive tool may help stratify these high-risk patients.

Area of Science:

  • Pediatric critical care medicine
  • Cardiorenal medicine
  • Extracorporeal life support

Background:

  • Fluid overload is a frequent complication in pediatric patients undergoing extracorporeal membrane oxygenation (ECMO).
  • In-line ultrafiltration (UF) and continuous renal replacement therapy (CRRT) are common treatments for fluid overload in this population.
  • The comparative outcomes and predictive factors for escalating therapy from UF to CRRT remain incompletely understood.

Purpose of the Study:

  • To compare outcomes between pediatric cardiac patients on ECMO treated with CRRT versus UF alone.
  • To identify clinical characteristics associated with the progression from UF to CRRT.
  • To develop a predictive tool for identifying patients likely to require CRRT during UF therapy.

Main Methods:

  • Retrospective chart review of 131 pediatric patients (age ≤18 years) treated with ECMO at a quaternary center.
  • Comparison of patient demographics, clinical characteristics, and outcomes between UF and CRRT groups.
  • Receiver operator curve (ROC) analysis to develop a predictive model for CRRT initiation.

Main Results:

  • Patients requiring CRRT exhibited higher creatinine and blood urea nitrogen levels at UF initiation, longer ECMO duration, lower renal recovery rates, and increased mortality.
  • The ROC analysis identified a predictive tool for CRRT need, with a specific risk variable threshold demonstrating significant predictive value (AUC 0.801, p = 0.002).
  • The presence of 3 or fewer of 7 risk variables predicted the use of UF alone with 87.5% positive predictive value.

Conclusions:

  • Pediatric cardiac patients on ECMO who require CRRT alongside UF experience significantly worse clinical outcomes, including higher mortality.
  • A novel clinical tool incorporating specific risk variables can aid in stratifying patients at the time of UF initiation, potentially guiding therapeutic decisions.
  • Early identification and risk stratification are crucial for managing fluid overload in critically ill pediatric cardiac patients on ECMO.