Revolutionizing cardiogenic shock treatment: Is the microaxial flow pump the ultimate game changer?

Acute myocardial infarction-related cardiogenic shock (AMICS) is associated with persistent high early mortality rates ranging from 40% to 60%.1-6 Despite advancements in AMICS knowledge, few positive studies have impacted patient outcomes. Indeed, landmark trials such as SHOCK3 and CULPRIT-SHOCK4 provided valuable insights into cardiogenic shock (CS) management, but did not highlight the potential benefit of temporary mechanical circulatory support (tMCS). Recently, four randomized controlled trials (RCTs) investigating extracorporeal life support (ECLS) in CS have failed to demonstrate any benefits in terms of survival compared with optimal medical treatment.5, 7-9 Thus, the consideration of tMCS in CS as a bridge-to-recovery/-to-decision/-to-bridge stays with a low level of recommendation (class IIC and IIaC, respectively, for the International Society for Heart and Lung Transplantation/Heart Failure Society of America and the European Society of Cardiology).10, 11 In contrast, the recent DanGer Shock trial6 offered for the first-time robust evidence in favour of the percutaneous microaxial flow pump (MaFP) (Impella, Abiomed) in AMICS. Here, we provide a summary of the two most extensive prospective randomized clinical trials in tMCS (ECLS-SHOCK5 and DanGer Shock6) aiming to comprehend the differences in their outcomes (Figure 1). Key differences emerge between the two study populations. DanGer Shock is notable for its inclusion of a more homogeneous group of patients, specifically limited to those with ST-elevation myocardial infarction (STEMI) contrasting with only 67.2% of STEMI cases in the ECLS-SHOCK trial. This strict criterion, alongside others such as arterial lactate elevation (>2.5 mmol/L), left ventricular (LV) ejection fraction < 45%, absence of severe right ventricular (RV) failure (confirmed by echo on admission), and absence of persistent coma in case of cardiac arrest, led to the identification of a selected AMICS population with profound cardiac failure and organ hypoperfusion, most likely to benefit from tMCS. Firstly, patients included in the DanGer Shock trial seem less profoundly shocked with lower rates of patients being in stage D and E of the Society for Cardiovascular Angiography & Interventions (SCAI) classification, and lower lactate levels (4.6 vs. 6.9 mmol/L) on admission. This is partly attributed to the variances in CS definitions employed in the two trials. Indeed, CS was defined by the combination of systolic blood pressure <90 mmHg persisting for 30 min or necessitating vasopressors, along with lactate levels >3 mmol/L and additional signs of organ hypoperfusion such as confusion, oliguria, and cold or clammy skin and limbs in the ECLS-SHOCK trial. These criteria likely resulted in a delay in initiating tMCS for patients with deeper CS. Furthermore, the DanGer Shock trial excluded patients with persistent coma following resuscitation (coma Glasgow scale <8), whereas the ECLS-SHOCK trial enrolled those with prolonged resuscitation (up to 45 min), resulting in a higher proportion of post-cardiac arrest patients (77.7% vs. 21.9%) with prolonged resuscitation (median time 20 min). Cardiac arrest induces ischemia–reperfusion injury triggering a systemic inflammatory response syndrome leading to mixed cardiogenic/vasoplegic shock. Here, tMCS often fails to yield significant benefits due to the elevated risk of severe multi-organ failure. Additionally, post-arrest myocardial stunning further contributes to myocardial dysfunction and impairs positive outcomes. In this negative spiral, mortality is more often driven by non-cardiac complications, such as post-anoxic encephalopathy (26.2 vs. 2.7% of deaths, after 30 days in the ECLS-SHOCK and DanGer Shock trials, respectively), rather than cardiac failure on its own. In association with a lower rate of resuscitation, the immediate LV unloading with MaFP in the DanGer Shock versus ECLS-SHOCK trial may explain a lower requirement of mechanical ventilation (74.3% vs. 90.1%) and a lower duration of respiratory support (5 vs. 7 days), facilitating early patient rehabilitation. Secondly, DanGer Shock focused predominantly on LV-associated CS, excluding patients with severe RV/biventricular failure by pre-randomization echocardiography (RV dilatation and tricuspid annular plane systolic excursion <10 mm), conditions associated with a poorer prognosis.12 Thirdly, an important contrast implicates the timing of device implantation. Whereas most patients underwent device implantation before revascularization in the DanGer Shock trial (84.8%), only few did in the ECLS-SHOCK trial (21.1%). This emphasizes DanGer Shock's focus on early haemodynamic support to counteract myocardial ischaemia during high-risk percutaneous coronary intervention, herewith preventing haemodynamic deterioration and cardiac arrest, and allowing complete and effective revascularization associated with better outcomes in tMCS patients.13 Finally, high rates of device-associated complications were observed in both trials with a predominance in the ECLS-SHOCK trial, including bleeding (23.4% vs. 21.8% in DanGer Shock) and limb ischaemia (11% vs. 5.6%), possibly attributed to the larger cannulas used in ECLS patients (16-18 Fr vs. 14 Fr for MaFP). It highlights the challenges associated with tMCS and emphasizes the importance of a structured approach to anticoagulation and antiplatelet therapy to optimize outcomes and minimize complications in these critically ill patients.14 Withal, renal replacement therapy was significantly higher in DanGer Shock (41.9% vs. 8.1%), possibly due to high rotation/shear stress-induced haemolysis in MaFP-supported patients (up to 10 000–50 000 rpm), highlighting the need to minimize the duration of MaFP support as much as possible. The contrast in early mortality rates between DanGer Shock and ECLS-SHOCK (39.6% vs. 47.8% at 30 days, respectively) may stem from various factors, as previously discussed. Intriguingly, this trend persisted at day 180, enhancing confidence in the observed outcomes and relevant translation into clinical practice. However, the substantial crossover rate (26% in ECLS-SHOCK and 24% in DanGer Shock received tMCS in the control group) adds a layer of complexity to interpreting the results. The utilization of a composite recovery endpoint in DanGer Shock, including treatment escalation, heart transplantation, or death, highlights its significance in contemporary medical research. Indeed, while short-term all-cause mortality is commonly used, experts suggest exploring alternative composite endpoints in CS that encompass both mortality and morbidity for a more nuanced evaluation.15 This approach provides a comprehensive understanding of patient outcomes, not only considering survival but also quality of life and the intervention-associated adverse events. Also, intrinsic device characteristics may explain the divergent trial results. The MaFP is positioned through the aortic valve, continuously ejecting a forward blood from the left ventricle into the ascending aorta. MaFP technology induces LV unloading, reduces ventricular preload and myocardial oxygen consumption but lacks oxygenation/ventilatory capabilities.16 However, in CS, respiratory failure primarily manifests as pulmonary oedema resulting from LV overload. Given that pulmonary oedema can often be resolved by proper intensive care unit care (higher positive end-expiratory pressure, afterload reduction, restrictive fluid management, etc.), the question arises as to whether an oxygenator in the circuit (with all possible complications) is always necessary.17 LV unloading makes a triangularization of the pressure–volume curve (reducing the isovolumetric contraction and relaxation phase), and decreases intraventricular pressures while it increases aortic pressure through anterograde aortic flow, resulting in decoupling of the ventriculo-aortic system. Additionally, MaFP-supported cardiac assistance enhances coronary perfusion and may improve myocardial ventricular function,18 but meticulous monitoring for haemolysis is needed. Over the past decade, ECLS has been increasingly used in AMICS. Peripheral ECLS induces significant reduction in RV preload and transpulmonary flow, along with an increase in LV afterload due to retrograde reinjection into the aorta. This increase in afterload affects myocardial contractility and increases myocardial work (potential energy, stroke work, and pressure–volume area), leading to an unfavourable balance between myocardial oxygen supply and demand.19 The alterations induced by peripheral ECLS in RV preload, transpulmonary flow, and LV afterload highlight the complexity of its impact on cardiac function and underscore the need for meticulous monitoring, including pulmonary artery catheters and echocardiography. Moreover, the potential decrease in coronary blood flow coupled with unchanged myocardial work underscores the necessity for ECLS introduction. Combining ECLS with ventricular unloading techniques like intra-aortic balloon pump or MaFP (ECMELLA) may offer a more balanced haemodynamic support strategy, addressing both perfusion needs and myocardial workload. Surprisingly, in the ECLS-SHOCK trial, only 5.8% of patients in the ECLS group received an LV unloading strategy, compared to up to 30–40% in recent registries.20 In conclusion, it is likely that initiation of LV support by MaFP in AMICS patients emerges as a first-line tMCS, even before revascularization. Its ease of deployment, 24/7 accessibility by interventional cardiologists, and the possibility of upgrading to ECMELLA in cases of persistent hypoperfusion makes it the strategy of choice. Indeed, ECLS should be considered as an add-on to MaFP therapy, particularly when MaFP alone proves insufficient in providing haemodynamic support, such as lung failure or biventricular dysfunction. In the absence of local expertise, transfer after MaFP initiation or need for tMCS escalation to a hub centre is recommended. The discussed trial results emphasize the importance of a balanced and thoughtful strategy in the selection and utilization of tMCS, with the aim of optimizing patient outcomes and improving the overall management of AMICS. Ongoing RCTs such as RECOVER IV (NCT05506449) and ULYSS (NCT05366452) will provide further insights into the role of MaFP in AMICS. These studies aim to confirm the findings of DanGer Shock in larger (cumulated inclusion planned of 764 patients) and/or different populations (non-STEMI) to elucidate the position of MaFP in the contemporary management of AMICS. Additionally, the ongoing UNLOAD ECMO (NCT05577195) and REMAP ECMO (NCT05913622) trials could offer critical insights into the effectiveness of immediate LV unloading with MaFP and intra-aortic balloon pump, respectively, in patients supported by ECLS. These trials represent pivotal opportunities to enhance treatment approaches and improve critically ill patient care. Nevertheless, it is crucial to acknowledge that outcomes observed in RCTs may not entirely mirror those in real-world settings due to highly selective patient criteria.1 Notably, the DanGer Shock and ECLS-SHOCK criteria only match 24.9% and 40.9% of patients' phenotypes observed in real-world AMICS data.20 Finally, the high rates of complications observed in both trials with these devices underscore the urgent need for the development of proper and standardized intensive care unit management protocols. Conflict of interest: A.U. has nothing to disclose. C.D. and C.V. report lectures fees from Abiomed.