Optimal patient and mechanical circulatory support device selection in acute myocardial infarction cardiogenic shock

Jacob C Jentzer, Benjamin Hibbert

Lancet 2024; 404: 992-993

Temporary mechanical circulatory support (TMCS) has been used in patients with acute myocardial infarction and cardiogenic shock (AMICS) since the introduction of passive intra-aortic balloon counterpulsation more than 50 years ago.^1,2Despite the development and expanding use of more sophisticated active TMCS devices, such as percutaneous ventricular assist devices (PVADs) and venoarterial extracorporeal membrane oxygenation devices (VA-ECMO), outcomes of patients with AMICS remain unacceptably poor.^1–6 Although proponents have argued that tailored use of the correct device in the correct patient can save lives, randomised clinical trials of TMCS have uniformly failed to show improved survival in broad cohorts with cardiogenic shock.^2,3,7–9 The recent Danger-Shock trial reported that early use of PVAD in certain patients with cardiogenic shock due to ST-elevation acute myocardial infarction (STEMI) without anoxic brain injury improved survival despite more non-fatal complications.¹⁰ The finding of better survival with TMCS in selected patients with AMICS in Danger-Shock has led to uncertainty about whether TMCS can be beneficial in other subgroups of patients with AMICS, and whether the potential benefit varies between different TMCS devices.

In The Lancet, Holger Thiele and colleagues¹¹ addressed these questions in a meta-analysis using individual patient data from 1059 patients (median age 65 years [IQR 57–73], 845 [79·9%] of 1058 patients with data were male and 213 [20·1% were female) from nine randomised controlled trials of TMCS in AMICS to explore the effects of loading versus unloading TMCS devices on 6-month mortality, including important patient subgroups; as expected, the findings mostly recapitulate the findings of the largest included studies.^9,10 When all patients with AMICS and who received TMCS devices were analysed together, no improvement in survival was observed at 6 months with TCMS (50·7% [95% CI 46·5–55·0; 268 deaths observed] vs 55·9% in the control group 51·7–60·2; 293 deaths; hazard ratio [HR] 0·87, 95% CI 0·74–1·03; p=0·10).¹¹ The effect was similar when examining both unloading (PVAD) and loading (VA-ECMO) devices, although the confidence interval for the effect of PVAD on mortality suggested a probable benefit (47·3% [95% CI 41·3–53·7] 6-month mortality [117 deaths] vs 56·9% [50·8–63·3] 6-month mortality [137 deaths] in the control group; HR 0·80, 95% CI 0·62–1·02; p=0·075). The important subgroup of patients with STEMI-related cardiogenic shock without coma (including those enrolled in Danger-Shock) had significantly reduced mortality if they received TMCS of either type (132 [44·3%] of 298 patients who received TMCS had died at 6 months vs 161 [55·3%] of 291 patients in the control group; HR 0·77, 95% CI 0·61–0·97; p=0·024).^10,11 Overall, the results were consistent across other prespecified subgroups according to age and sex, with beneficial effects of TMCS observed in patients with poor reperfusion (thrombolysis in myocardial infarction flow grade <3 after percutaneous coronary intervention) or low blood pressure. Regardless of the device or patient population, use of TMCS substantially increased both bleeding (odds ratio 2·64, 95% CI 1·91–3·65) and vascular complications (4·43, 2·37–8·26).¹¹ Exclusion of the Danger-Shock trial weakened the findings by reducing statistical power but did not markedly change the point estimates.

So how should clinicians apply the available evidence in the context of the current analysis? Broadly speaking, in neurologically intact patients with AMICS and STEMI, every ten patients treated with early TMCS (especially with PVAD) could result in one additional survivor at 6 months, despite an increased risk of serious complications. By contrast, routine use of TMCS in all patients with AMICS will probably result in net harm, including substantially increased risk of bleeding.

Despite being the best evidence available in this context, this well-conducted meta-analysis has relevant limitations that readers should consider when applying the results into clinical practice. Each study enrolled a different patient population at expert centres predominantly in Europe, with variable treatment protocols before and after random assignment. The outcomes of using TMCS devices in centres with less expertise are likely to be worse than those observed in these randomised controlled trials, which already showed high complication rates.³ Importantly, the included randomised trials focused on early up-front TMCS versus delayed selective TMCS, suggesting that selective use of TMCS as a rescue therapy for refractory cardiogenic shock remains reasonable in certain patients. These trials primarily enrolled patients with severe, but not refractory, AMICS and do not provide insights into the care of patients with lesser or greater severity of cardiogenic shock, nor patients with cardiogenic shock due to causes other than acute myocardial infarction, who account for most cases.^4–6

Patients with preceding cardiac arrest and coma (accounting for the majority of patients enrolled in several studies) appeared to derive less benefit from TMCS than patients without cardiac arrest or those who had a brief arrest without anoxic brain injury—this was expected, as the former group often dies from anoxic brain injury regardless of haemodynamic status.^3,12 Only about 20–25% of all patients with cardiogenic shock in contemporary cohorts would be included in the subgroup that benefited from TMCS in this meta-analysis (ie, STEMI without coma).^4,5 The analysis of unloading TMCS devices and loading TMCS devices was primarily a comparison of VA-ECMO versus the Impella PVAD, and there are several important differences between these devices without a true direct comparison in the same randomised controlled trial.¹¹Crucially, the disparate study populations enrolled in Danger-Shock and ECLS-SHOCK could explain the divergent results in the unloading (PVAD) group versus loading (VA-ECMO) group, rather than unique features of the studied TMCS devices.^9,10The Danger-Shock trial stands as an outlier in many ways and is the only randomised controlled trial showing a significant benefit of TMCS; the overall benefits of TMCS and PVADs observed in this meta-analysis are largely driven by the findings of Danger-Shock.¹⁰

Ultimately, this analysis suggests that indiscriminate use of TMCS in patients with AMICS will not improve survival but will increase bleeding and vascular complications at the expense of greater resource use. Higher survival might be observed when early TMCS is used selectively in patients with STEMI without a major competing risk of death from anoxic brain injury. The premature termination of the confirmatory RECOVER IV trial (NCT05506449) of PVAD in AMICS will result in continued uncertainty about the efficacy of TMCS, as this meta-analysis might have been underpowered. In our opinion, it remains ethical, feasible, and necessary that future randomised controlled trials corroborate the findings of the current analysis and, perhaps more importantly, evaluate the efficacy of TMCS in broader populations with cardiogenic shock, in whom it continues to be used inconsistently in the absence of definitive evidence yielding poor clinical outcomes.

更多精彩内容，敬请登录CSCCM官方网站（www.csccm.org.cn）或点击屏幕最下方“阅读原文”