Beyond flow: a ventricular-load framework for VA-ECMO in cardiogenic shock

Cardiogenic shock still carries a 30-day mortality of approximately 40%–50% despite reperfusion, vasoactive support and modern intensive care. Veno-arterial extracorporeal membrane oxygenation (VA-ECMO) can rapidly restore systemic perfusion, but recent randomised evidence has reframed its role. In infarct-related shock, randomised trials have not shown a clear survival benefit from routine early VA-ECMO; ECLS-SHOCK and pooled analyses also found more bleeding and vascular complications. In the same period, DanGer Shock showed that a primary microaxial-pump (Impella CP) strategy improved survival in a carefully selected population with ST-elevation myocardial infarction complicated by cardiogenic shock, although it was tested against standard care rather than head-to-head against VA-ECMO and does not isolate unloading as the causal mechanism. Read together, these data motivate the central argument of this review: peripheral VA-ECMO is usefully managed not only as a problem of circulatory flow, but also as a dynamic problem of ventricular loading. We advance this as an organising framework rather than as a demonstrated mechanism, since differences in patient selection between these trials offer a competing and equally tenable explanation. Retrograde aortic return increases effective LV afterload, while venous drainage reduces preload; the clinical result depends on the interaction between circuit flow, native contractility and biventricular loading conditions. This narrative, non-systematic review synthesises contemporary trial evidence, the physiology and treatment of LV loading, and approaches to serial assessment of the supported heart, from haemodynamics and bedside echocardiography to emerging deformation and advanced imaging tools. VA-ECMO should therefore be viewed as a time-limited bridge whose net benefit depends on reversible disease, early recognition of clinically relevant ventricular overload and repeated assessment of recovery, escalation options and futility.