Cardiogenic Shock Device Choice Depends on Phenotype and Destination

Short-term mechanical circulatory support can restore systemic perfusion in cardiogenic shock, but randomised studies have not shown a uniform survival advantage from routine use. A 2026 mini review argues that device selection should be guided by the failing ventricle, respiratory needs, vascular access, reversibility, neurological prognosis and a realistic therapeutic destination: myocardial recovery, definitive repair, transplantation, durable ventricular support or temporary stabilisation while decisions are made. The strongest randomised evidence concerns selected patients with acute myocardial infarction-related shock, where one trial found a survival benefit with microaxial flow-pump support while larger trials of veno-arterial extracorporeal membrane oxygenation found no overall benefit. Evidence for heart failure-related shock and other causes remains much thinner.

Cardiogenic shock is a heterogeneous syndrome rather than a single physiological state. It may involve predominant left ventricular failure, right ventricular failure, both ventricles, respiratory failure or a mixed cardiogenic–vasodilatory pattern in which inadequate vascular tone contributes to shock. Systolic pressure alone is therefore an unreliable guide: increased systemic vascular resistance can preserve blood pressure above 90 mmHg despite tissue hypoperfusion. Echocardiography, lactate trends, vasoactive-drug requirements, congestion, oxygenation and, when available, invasive haemodynamics should be interpreted repeatedly as the patient’s phenotype evolves. The 2022 SCAI SHOCK classification, ranging from stage A, “at risk,” to stage E, “extremis,” adds severity information but does not determine whether support will benefit an individual.

The physiological rationale for support must also be separated from evidence of clinical benefit. Greater haemodynamic severity identifies patients with the greatest need, but it can also identify irreversible neurological injury, advanced multiorgan failure, severe frailty or comorbidity that makes meaningful recovery unlikely. Isolated thresholds for age, lactate concentration or shock stage should not be regarded as absolute determinants of candidacy. Support is most rational when it can contribute to a credible therapeutic destination. If the destination remains uncertain, it may function as a bridge-to-decision while prognosis and reversibility are assessed. Patients resuscitated after cardiac arrest require particular caution because hypoxic-ischaemic brain injury may govern the outcome independently of cardiac recovery, while sedation, temperature management and metabolic disturbance can make early neurological assessment unreliable.

Randomised trials for routine device use have produced largely neutral results. In IABP-SHOCK II, 600 patients with acute myocardial infarction-related cardiogenic shock who underwent early revascularisation were assigned to an intra-aortic balloon pump or standard care; 30-day mortality was 39.7% versus 41.3%, respectively, with a relative risk of 0.96 and 95% confidence interval of 0.79–1.17. Trials of routine veno-arterial ECMO were also neutral. In ECLS-SHOCK, the largest of these with 420 participants, early veno-arterial ECMO did not lower 30-day mortality, which was 47.8% with ECMO and 49.0% with control, but it increased moderate or severe bleeding from 9.6% to 23.4% and peripheral vascular complications requiring intervention from 3.8% to 11.0%.

A contrasting result came from DanGer-Shock, in which 360 patients with ST-elevation myocardial infarction-related cardiogenic shock at 14 centres in Denmark, Germany and the United Kingdom were randomised to the Impella CP microaxial flow pump or standard care, with 355 analysed. At 180 days, all-cause mortality was 45.8% in the pump group and 58.5% in the control group, producing a hazard ratio of 0.74 and 95% confidence interval of 0.55–0.99. The benefit came with more device-related adverse events and longer intensive-care and hospital stays. The population was highly selected: about 20% had preceding cardiac arrest, compared with 78% in ECLS-SHOCK. Differences in eligibility, shock phenotype, competing neurological risk, randomisation timing, support and care setting mean the results do not establish direct superiority of microaxial pumps over ECMO.

An individual-patient-data meta-analysis of nine randomised trials involving 1,059 patients found no significant overall reduction in six-month mortality with mechanical support. In a prespecified subgroup of patients with STEMI-related shock and either no resuscitation or less than 10 minutes of resuscitation, used as a proxy for lower hypoxic brain-injury risk, support was associated with lower six-month mortality, with a hazard ratio of 0.77 and 95% confidence interval of 0.61–0.97. The analysis did not prospectively establish that patients respond differently according to a defined shock phenotype, so the finding remains hypothesis-generating rather than validation of a phenotype-guided treatment rule. Evidence in heart failure-related shock is more limited still: Altshock-2 stopped for futility after comparing early balloon-pump support with standard care in 101 patients, with no significant improvement in 60-day survival or bridging to heart replacement therapy.

Timing may also affect results, although no major randomised trial has directly compared early with delayed implantation. Restoring cardiac output sooner may reduce catecholamine exposure, improve oxygen delivery and limit multiorgan injury. An observational meta-analysis of 13 studies and 6,810 patients with infarction-related shock reported lower short-term mortality when Impella support began before, rather than during or after, percutaneous coronary intervention: 37.2% versus 53.6%, with a relative risk of 0.70 and 95% confidence interval of 0.56–0.88. Because the evidence was observational, selection bias and residual confounding remain possible. The overall implication is selective rather than routine support, with device physiology matched to shock phenotype and trajectory, an approach grounded in haemodynamic principles and subgroup observations rather than prospective validation.

The devices differ materially in flow, unloading, respiratory support and access requirements. An intra-aortic balloon pump provides modest counterpulsation, improves diastolic coronary perfusion and lowers left ventricular afterload, but depends on native cardiac function and pulsatility. It may suit selected left ventricular shock or provide a bridge to repair of post-infarction ventricular septal rupture or acute severe mitral regurgitation. Microaxial flow pumps actively support and unload the left ventricle: the Impella CP, delivered through large-bore arterial access, typically provides about 3.5–3.6 litres per minute for short-term support, while the Impella 5.5 can provide up to 5.5 litres per minute and is intended for longer support, usually through a surgically implanted axillary graft. These systems require adequate right ventricular output and do not provide oxygenation; left ventricular thrombus and a mechanical aortic valve also preclude transvalvular pump support.

Right ventricular failure calls for a different strategy because isolated left ventricular pumping depends on blood crossing the pulmonary circulation. Devices such as the Impella RP or a ProtekDuo centrifugal right ventricular assist configuration bypass or unload the right ventricle and increase pulmonary flow and left ventricular preload; an oxygenator can be incorporated in selected circuits. Veno-arterial ECMO offers high-flow circulatory support and gas exchange for profound or refractory failure, particularly with biventricular dysfunction or severe respiratory failure. Peripheral femoral cannulation is common, but retrograde aortic flow can increase left ventricular afterload, causing distension, elevated filling pressure, pulmonary congestion and poor aortic-valve opening. Additional unloading, using a balloon pump, microaxial pump, transseptal drainage or surgical venting, may be needed, although a universal survival benefit from routine unloading has not been established. Every choice must also accommodate vascular disease, coagulopathy, anatomical constraints, expertise, expected duration and the therapeutic destination.

Complications can offset physiological improvement. In DanGer-Shock, moderate or severe bleeding occurred in 21.8% of patients receiving microaxial support and 11.9% of controls, while limb ischaemia occurred in 5.6% and 1.1%. Blood contact with artificial surfaces activates coagulation, platelets and complement, while prolonged exposure consumes platelets and clotting factors; high shear can impair platelet function and cause loss of high-molecular-weight von Willebrand factor multimers. Anticoagulation therefore requires repeated individual assessment of bleeding and thrombotic risk. Large-bore access can injure vessels or compromise distal perfusion, and antegrade distal perfusion cannulation may be considered during peripheral ECMO. High-speed microaxial pumps can cause haemolysis, especially with malposition or suction, requiring surveillance of pump position, haemoglobin, lactate dehydrogenase, plasma-free haemoglobin and haptoglobin.

Prolonged support also introduces infection, thrombosis, stroke and device-specific mechanical risks. Sepsis with a positive blood culture occurred in 11.7% of microaxial-support recipients in DanGer-Shock, compared with 4.5% of controls. During peripheral ECMO, poorly oxygenated native left ventricular output may supply the upper body while oxygenated circuit flow perfuses the lower body, producing differential hypoxaemia known as Harlequin syndrome. Inadequate left ventricular ejection may also encourage intracardiac stasis and thrombosis. Prevention requires aseptic access management, vascular and haemostatic assessment before implantation, continuous monitoring during support and daily reconsideration of myocardial recovery, end-organ perfusion, complications and readiness for de-escalation. Once adequate recovery is established, support should be reduced in a device-specific, stepwise manner and removed as soon as clinically appropriate.

Subject of Research: Phenotype-guided selection and complication risks of short-term mechanical circulatory support in cardiogenic shock

Article Title: Update in short-term mechanical circulatory support: indications, devices and complications

Article References: Moreno-Coca, J. M., Moreno-Monterde, E., Manzano, A., Andrea, R., & Giménez-Milà, M. (2026). Update in short-term mechanical circulatory support: indications, devices and complications. Frontiers in Cardiovascular Medicine, 13, 1994496. https://doi.org/10.3389/fcvm.2026.1994496

Image Credits: AI Generated

DOI: 10.3389/fcvm.2026.1994496

Keywords: cardiogenic shock, mechanical circulatory support, device selection, phenotype-guided care, VA-ECMO, microaxial flow pump, intra-aortic balloon pump, bleeding, haemolysis, timing of support, Update, short-term