A26-15 Dynamic Biomarker Changes Following Cardiopulmonary Bypass and Red Blood Cell Transfusions Are Associated With Organ-Dysfunction Trajectories in Critically Ill Pediatric Patients

Rationale Dynamic inflammatory responses to cardiopulmonary bypass (CPB) and red blood cell (RBC) transfusion may contribute to variable outcomes in critically ill pediatric patients. While previous studies have linked plasma biomarkers with pediatric organ-dysfunction scores such as PELOD-2, few have examined how CPB- and transfusion-related proteomic changes influence PELOD-2 trajectories. We aim to investigate the relationship between dynamic biomarker patterns post transfusion and CPB and organ-dysfunction trajectories which may inform future targeted interventions in critically ill pediatric patients undergoing CPB and RBC transfusions. Methods Using multiplex and single-plex platforms (NULISA, Alamar Biosciences; Luminex and ELISA) we analyzed pre- and post-transfusion plasma samples from critically ill pediatric patients who underwent CPB and received a RBC transfusion (n = 86) covering inflammatory (n = 249), thrombosis (n = 9), and heme (n = 1) markers. Baseline-adjusted biomarker changes were calculated from residuals of a linear regression using post- versus pre-transfusion values. PELOD-2 scores from the first 7 days were used to derive organ-dysfunction trajectories (Class 1 = improving, Class 2 = worsening) via growth mixture modeling. Since the biomarkers (n = 260) largely exceed the sample size, principal component analysis (PCA) was performed on baseline-adjusted biomarker change values to reduce dimensionality and summarize major trends in the data. PCA identified principal components (PCs), which represent distinct patterns of correlated biomarker changes. Logistic regression was used to test associations between these PCs and PELOD-2 trajectories. The most highly correlated markers (|loading|≥0.1) in significant PCs (p < 0.05) were extracted and mapped to biological pathways using the Database for Annotation, Visualization, and Integrated Discovery (DAVID). Results Differential analysis of pre- and post-transfusion biomarkers identified multiple markers significantly associated with PELOD-2 trajectory outcomes (false discovery rate [FDR]<0.05, see Figure). PCA identified 5 PCs, explaining 44% of the total variance in the biomarker changes. A logistic regression with these 5 PCs was able to discriminate PELOD-2 trajectories with an AUC = 0.888 (p = 0.02, 95% CI: 0.7-1). PC3 was found to be the most strongly associated with improving PELOD-2 trajectories (p = 0.01, OR = 0.65, 95% CI: 0.42-0.87). The top PC3 loadings identified 21 key biomarkers, which mapped to inflammatory and growth-factor signaling pathways, including cytokine-cytokine receptor interactions, JAK-STAT signaling, and PI3K-Akt signaling. Conclusions Analysis of a comprehensive plasma biomarker panel following RBC transfusion in CPB patients reveals a set of markers corresponding to inflammatory and growth-factor signaling pathways that may be associated with improving PELOD-2 trajectories and suggest that regulated activation of cytokine and growth-factor signaling may support post-transfusion recovery. This abstract is funded by: 1RO1HD092471