TOP 75 - Implications of Inflammatory State on Hemorrhagic and Thrombotic Complications in Pediatric ECMO

Publication Number: 1833.TOP 75

Rachelle Wood, Phoenix Children's Hospital, Tempe, AZ, United States; Ali McMichael, Phoenix Children's Hospital, Phoenix, AZ, United States; Erin Thornley, Phoenix Children's Hospital, Phoenix, AZ, United States; Caitlin Toohey, Phoenix Children's Hospital, Phoenix, AZ, United States; Hamy Temkit, Phoenix Children's Hospital, Phoenix, AZ, United States; Nathaniel P. Grennan, Phoenix Children's Hospital, Phoenix, AZ, United States

Background: Initiation of extracorporeal membrane oxygenation (ECMO) provokes inflammation and dysregulates coagulation for approximately 48 hours before homeostasis is restored. The etiology of coagulopathy for pediatric patients on ECMO is complex and multifactorial. There is growing evidence that elevations in specific inflammatory cytokines that promote coagulopathy may be predictive of mortality for patients on ECMO. However, investigation into the connection between inflammatory state and the incidence of hemorrhage and/or thrombosis has been limited to the initial dysregulation of coagulation and inflammation that occurs within the first 48 hours of ECMO. There is minimal pediatric data on the association between inflammatory state and thrombotic or hemorrhagic complications beyond 48 hours on ECMO. We hypothesize that fluctuation in the inflammatory state after 48 hours on ECMO is associated with hemorrhagic and thrombotic complications in pediatric ECMO.

Objective: To describe the association between inflammation and hemorrhagic and/or thrombotic events after 48 hours on pediatric ECMO.

Design/Methods: Single center retrospective cohort study of 100 pediatric patients on ECMO over a 3-year period. This study has been approved by our institutional IRB. Demographic characteristics, markers of inflammation, ECMO data (type, duration, etiology), and hemorrhagic and thrombotic event data extracted from the electronic medical record will be summarized overall and by whether a subject has experienced a hemorrhagic or thrombotic event or not using frequencies and proportion for categorical variables and mean with standard deviation or median with interquartile range for continuous measures. The comparison between the two groups will be conducted using chi-squared or Fisher’s exact test for categorical variables and t-test or Wilcoxon rank-sum test for continuous variables depending on distribution. The association between each risk factor and the incidence of an event will be investigated using a univariate logistic repeated measures model and the results summarized using odds ratio, corresponding 95% confidence intervals and p-value for significance. An exploratory multivariable logistic repeated measures model to predict the odds of having an event will be investigated. A p-value < 0.05 will be considered statistically significant. Analysis will be performed using SAS version 9.4 and will be completed by 1/20/2026.