ID:121 OA1-7 Phenotype-Guided Timing Identifies Survival Benefit of Left Ventricular Unloading During VA-ECMO

Background: The survival benefit of left ventricular unloading during VA-ECMO remains inconsistent, suggesting treatment response depends on physiologic phenotype and timing. We evaluated whether organ viability, baseline severity, and early ECMO trajectory identify unloading-responsive phenotypes. Methods: This retrospective cohort included 425 adult VA-ECMO patients. Organ viability was defined by bilirubin ≥1.0 mg/dL and urine output <500 mL/day. Baseline severity was assessed using the SAVE score and early trajectory using the RESCUE-24 score. LV unloading included IABP, transseptal drainage, Impella, pulmonary artery drainage, or surgical LVAD. The primary outcome was survival to hospital discharge, including recovery, durable LVAD, or heart transplantation. Phenotype interaction and landmark timing analyses were performed. Results: Overall survival was 46%. Survival declined stepwise with worsening organ viability, from 69–100% in organ viable to 16–39% in organ non-viable patients. SAVE and RESCUE-24 scores alone did not predict unloading benefit. In contrast, unloading significantly improved survival in organ non-viable patients (OR 7.56, p<0.001), with the greatest benefit in SAVE high-risk with organ non-viable phenotype (OR 39.0, p=0.008) and RESCUE-24 intermediate-risk with organ non-viable phenotype (OR 9.0, p=0.002). Early unloading (≤24 hours) conferred the greatest benefit. Conclusion: Integrated phenotype assessment identifies patients most likely to benefit from LV unloading and defines the optimal timing for intervention.