Venting About Left Ventricular Unloading

To the Editor: Over the past decade, lumped parameter-based modeling studies have predicted that veno-arterial (V-A) extracorporeal membrane oxygenation (ECMO) should increase left ventricular (LV) afterload and preload, such that pressure–volume area (PVA [stroke work plus unavailable potential energy], which is proportional to and the principal per beat causative input into myocardial oxygen consumption [VO2M]—PVA/VO2M being an expression of myocardial chemo-mechanical efficiency) increases. In the setting of antecedent LV failure, these effects are predicted to occur universally. However, this is untrue; most patients treated with V-A ECMO do not develop LV distension.1 Human studies demonstrate that V-A ECMO results in decreased LV end-diastolic pressure and volume and PVA,2 consistent with different theoretical approaches preceding these studies.3 More recent theoretical work4 has also suggested that V-A ECMO-induced increased coronary blood flow may enhance LV contractility (Gregg effect) and result in “auto-unloading.” In ASAIO Journal, Natov et al.5 comment on the results of the EVACS study.6 Briefly summarized, in a retrospective multicenter study of patients with cardiogenic shock supported with V-A ECMO with or without concomitant Impella microaxial endovascular left ventricular assist devices (LVADs), Piperata et al.6 found that survival outcomes were similar between the two groups. First, Natov et al.5 distinguish LV venting from unloading. However, the validity of this distinction is tenuous. The authors state that the goal of venting is “…to reduce LV cardiac filling pressures and prevent stasis in the LV and aortic root,” whereas the goal of unloading is “to reduce LV pressure and volume, thereby lowering native LV wall stress and myocardial oxygen consumption.” “Vents” originate in cardiac surgery, evacuating liquid (blood) and gas (air that enters the operative field). A vented LV has lower pressure and volume, and in the beating heart, this is generally true in systole and diastole (nonlinearities being discussed in Rajagopal3). For the terms to be used nonsynonymously, there must be circumstances in which reduced LV pressure and volume do not relate to reduced VO2M. Of the strategies available to vent/unload the LV—inclusive of auto-unloading—only inotropes reduce LV load while augmenting VO2M, as they enhance auto-unloading through increased LV contractility. Intra-aortic balloon pumps (IABPs) mechanically enhance auto-unloading via LV afterload reduction, whereas vent catheters (flow-through driven by the ECMO circuit paracorporeal pump) and endovascular LVADs both directly remove blood volume independent of (but assisted by, in the case of LV drainage, or left atrial drainage in the presence of mitral valve regurgitation) LV systolic action. Second, it is important to directly examine the studies that Natov et al.5 cite is warranted. None of the studies in the table provides data on the physiological efficacy of unloading—not even LV dimensions and pulmonary artery occlusion pressures, which are often clinically obtainable. If there is a benefit to LV unloading, it ought to be mechanistically linked to its physiological effects. In addition, if unloading efficacy were to drive benefits, one would predict that direct unloading (most notably, Impella) would be superior to facilitated auto-unloading (IABP). However, even the studies selected for analysis fail to show this. The largest study cited, by Grandin et al.,7 an analysis of the Extracorporeal Life Support Organization Registry, demonstrated worse unadjusted survival in Impella-treated compared with IABP-treated patients. Only post hoc statistical manipulations yielded equivalent results. Another study cited by Russo et al.8 performed a subgroup analysis that failed to show any benefit to Impella-based LV unloading, consistent with a recent single-institution study.9 In their commentary, Natov et al.5 group Impella and IABP together, neglecting the subgroup analyses of the articles that they cite. Finally, a large study of the National Inpatient Sample10 not cited by Natov et al.5 failed to demonstrate a survival benefit to LV unloading in the setting of V-A ECMO, with the Impella-treated subgroup exhibiting the worst survival. Last, these studies generally state that they included both pre-V-A ECMO LV support and post-V-A ECMO LV support. These groups of patients are almost certainly extremely and importantly different from one another. Escalation of LV-only support to the addition of V-A ECMO is clearly dissimilar to LV unloading following V-A ECMO institution. I agree that prospective and randomized studies will provide great insights. It is also clear that there are clinical scenarios and patients who would physiologically benefit from LV unloading on V-A ECMO. What is far from clear is whether LV unloading ought to be uniformly or even widely applied, or whether it ought to be selectively used, based on identified risk factors or post-ECMO initiation physiological alterations.