Abstract 15696: Outcomes With Intra-Aortic Balloon Pump versus Percutaneous Ventricular Assist Device for Left Ventricular Venting in Adults Receiving Venoarterial Extracorporeal Life Support: An ELSO Registry Analysis

Background: Insufficient left ventricular (LV) unloading during venoarterial extracorporeal life support (VA-ECLS) can result in poor LV recovery and inability to wean from support. Published data suggest a survival benefit with LV mechanical venting (MV), but there is limited data comparing modalities. Methods: We queried the ELSO registry from 2010-2019 for adults undergoing VA-ECLS with MV and stratified them by intra-aortic balloon pump (IABP) or percutaneous ventricular assist device (pVAD). We excluded patients with pulmonary embolism, heart transplant, congenital and valvular heart disease, aortic disease, and central cannulation. We performed a subgroup analysis excluding extracorporeal cardiopulmonary resuscitation (ECPR). The primary outcome was in-hospital mortality. Secondary outcomes were on-support mortality, medical and cannula site bleeding, hemolysis, ischemic stroke, limb ischemia, and renal injury. We used multivariable logistic regression modeling to adjust for clinical covariates. Results: Among 3353 adults with MV on VA-ECLS, 2782 (83%) were vented with IABP and 571 (17%) with pVAD. IABP patients were less likely to be supported for ventricular arrhythmia (9.6% vs 14.3%) and had lower rates of pre-ECLS arrest (49% vs 66%) and concomitant renal (13.5% vs 34.5%), liver (3.4% vs 9.3%), and respiratory (17.6% vs 37.0%) failure, p<0.001 for all. On-support (40.8% vs 45.5%, p=0.037) and in-hospital (55.9% vs 60.6%, p=0.038) mortality were lower with IABP vs pVAD. In multivariate modeling, IABP was associated with a significantly lower odds of medical bleeding with a trend towards lower in-hospital mortality that was significant after excluding ECPR patients (Figure). Conclusion: In adults with MV on VA-ECLS, IABP compared to pVAD is associated with a significantly lower risk of bleeding and a strong signal of improved survival. We cannot exclude residual unmeasured confounding, so prospective studies are needed to compare MV devices.