Introduction: The need for preoperative extracorporeal membrane oxygenation (ECMO) support prior to left ventricular assist device(LVAD) therapy has been associated with increased mortality. Our study investigated the impact of de-escalating from ECMO to Impella prior to implantation.

Methods: We conducted a retrospective review of all LVAD implantations from 2017-2023. We stratified patients into 4 groups: those requiring ECMO at the time of LVAD; those requiring ECMO who were de-escalated to Impella only support; those requiring Impella support who never required ECMO; and those who did not require temporary mechanical circulatory support(tMCS). The primary outcome was Kaplan-Meier survival.

Results: From 2017-2023, 146 of our patients underwent LVAD implantation. Preoperative mechanical support included ECMO(4, 2.74%), ECMO then Impella only(11, 7.53%), Impella only(46, 31.51%), and no tMCS(85, 58.22%). Overall operative mortality was 8.22% with a 1-year survival of 85.40%. When stratified by the need for preoperative tMCS, the need for preoperative ECMO support was associated with significantly decreased 1-year survival(50.00%) compared to ECMO de-escalated to Impella only(80.81%), Impella only(80.88%), and no tMCS(89.91%, p=0.02). Notably, patients who initially required ECMO support but were subsequently de-escalated to Impella only had clinically and statistically similar survival compared to those requiring Impella only(p=0.92) or those who did not require tMCS(p=0.36).

Conclusions: Indirectly bridging from ECMO to LVAD by de-escalating to Impella only prior to LVAD implantation may mitigate the increased risk associated with using ECMO as a direct bridge to LVAD therapy.