Abstract Or108: Divergent Biventricular Mechano-Energetic Responses to VA-ECMO Flow Ramping After Cardiac Arrest: A Preclinical ECPR Swine Model

Abstract

Background: Venoarterial extracorporeal membrane oxygenation (VA-ECMO) is increasingly used in cardiac arrest (CA) resuscitation, yet its effects on biventricular mechanics remain poorly defined. Given that early on-ECMO hemodynamics associate with survival, we used a swine model with real-time pressure-volume (PV) loop analysis to assess ventricular-arterial coupling and mechano-energetics during graded ECMO flow after CA. Methods: In a ventricular fibrillation swine model, high-fidelity conductance catheterization was performed to measure real-time PV loops in left ventricle (LV) and right ventricle (RV). Baseline (pre-arrest) data were acquired, followed by 20 minutes of CA and VA-ECMO initiation. PV data were obtained across multiple ECMO flow rates (Q [cc/kg/min]). Ventricular elastance (Ees), arterial elastance (Ea), Ees/Ea coupling ratio, stroke work (SW), potential energy (PE), pressure-volume area (PVA), and mechanical energy efficiency (MEE = SW/PVA) were quantified. Results: Both ventricles demonstrated acute post-arrest dysfunction; however, the trajectories diverged thereafter. RV Findings: Ees/Ea declined from 1.7 (baseline) to <0.4 across ECMO flows, reflecting uncoupling. This was driven by disproportionate increases in Ea and declines in stroke volume, despite modest changes in contractility (Ees). RV volumes (EDV/ESV) remained elevated, indicating sustained dilation. SW remained flat while PE and PVA increased, resulting in a drop in MEE from >80% to ~40%. LV Findings: The LV exhibited partial recovery in Ees with ECMO ramping. Despite increased Ea (likely from retrograde flow and reduced ejection), Ees/Ea remained stable (~0.7). SW and PVA both declined, consistent with unloading, yet MEE was preserved. PV loop morphology showed progressive LV volume reduction with preserved loop geometry, while RV loops demonstrated early dilation, impaired ejection, and rising energetic burden. Conclusions: This model shows that following CA and VA-ECMO initiation, the RV and LV display distinct mechano-energetic and coupling patterns. The RV demonstrated persistent uncoupling and inefficiency, potentially due to post-arrest pulmonary changes, while the LV showed partial recovery and preserved efficiency. These results support the need for individualized biventricular assessment during ECMO-supported resuscitation.