Impella versus extracorporeal membrane oxygenation for acute myocardial infarction cardiogenic shock

Selected in Cardiovascular Revascularization Medicine by M. Alasnag , K. Al Faraidy

This study aims at evaluating the difference in baseline characteristics, procedural and clinical outcomes in those receiving Impella or ECMO in the United States in the setting of AMI and CS.

Reviewers

Mirvat Alasnag

@mirvatalasnag

Interventional cardiologist / Cardiologist

King Fahd Armed Forces Hospital - Jeddah, Saudi Arabia

Khalid Al Faraidy

Interventional cardiologist / Cardiologist

KFMMC - Dammam, Saudi Arabia

Our Comment

Why this study? – the rationale/objective

Acute myocardial infarctions (AMI) are complicated by cardiogenic shock (CS) in 5-10% of the cases. In spite of developments in inotropic regimens, the in-hospital mortality remains as high as 66%. As such, mechanical circulatory support (MCS) devices have been employed in an attempt to improve shock outcomes.

Intra-aortic balloon pump (IABP) counter-pulsation was amongst the first to be evaluated in the IABP-SHOCK II study as well as in several meta-analyses that failed to demonstrate a survival benefit with IABP. Therefore, the ESC/EACTS guidelines on myocardial revascularisation downgraded the recommendation for the use of MCS.

More recently, Impella (Abiomed, Danvers, Massachusetts, USA) has gained popularity given its greater cardiac output capability. Although several comparisons of IABP and Impella such as the matched-Pair IABP-SHOCK II Trial 30-Day by Schrage et al (Circulation. 2019;139:1249-1258) showed improvement in shock parameters, none demonstrated a reduction in mortality. Randomised trials including the IMPRESS in STEMI trial were underpowered to detect mortality differences.

Another device, is the extracorporeal membrane oxygenation (ECMO). Garan et al ( J Am Heart Assoc. 2019;8(9):e012171) examined the role of ECMO and a peripheral left ventricular assist device (pVAD) following CS in an AMI. Both VA-ECMO and pVAD had similar outcomes. Additionally, the use of both devices simultaneously was common, 50% of those with persistent CS.

The purpose of this study by Lemor et al is to evaluate the difference in baseline characteristics, procedural and clinical outcomes in those receiving Impella or ECMO in the United States in the setting of AMI and CS.

How was it executed? – the methodology

Hospital admissions that underwent percutaneous coronary intervention (PCI) and emergent Impella or ECMO placement for AMI-CS using ICD-10 codes were retrieved from the National Inpatient Sample database from October 2015 to December 2017. Propensity-score matching was performed to compare both groups for primary and secondary outcomes. The primary outcome was in-hospital mortality. The secondary outcomes were divided as procedural outcomes and clinical outcomes. Procedural outcomes included single or multivessel PCI, coronary artery bypass grafting (CABG), use of IABP, and pulmonary artery catheterisation. Clinical outcomes included acute kidney injury (AKI), acute ischemic stroke, blood transfusion, acute respiratory failure, acute liver failure, vascular complications, length of stay, and hospital costs.

Procedures that included both devices were excluded (n=305), as the investigators could not determine which device was used first or if both were used simultaneously. Other exclusion criteria included those who did not undergo PCI (n=34,500), those younger than 18 years of age (n=5), and those without mortality data (n=10).

What is the main result?

A total of 6,290 admissions for AMI-CS were included in this analysis. The Impella accounted for the majority of the admissions (91% compared to only 9% in the ECMO cohort). Those who received Impella support were older (65.6± 11.8 vs 61.2 ± 11.6 years old) and less frequently male (73.3% vs 83%). Baseline characteristics were similar between both groups; however, more of the Impella cohort had renal dysfunction and COPD (7.2% vs 1.8% and 16.8% vs 8.9% respectively).

The majority underwent single-vessel PCI (61.3% in the Impella cohort and 63.4% in the ECMO cohort). The use of IABP was higher in the ECMO cohort (16.5% vs 51.8%, p < 0.001 and there was no statistical difference in the use of pulmonary artery catheters (13.6% vs 14.3%%, p=0.84). The in-hospital mortality was significantly higher for the ECMO cohort (45.5% vs 41.4%, p=0.004). There were significantly lower rates of acute ischemic stroke, blood transfusion, vascular complications, acute respiratory failure and acute liver failure in the Impella cohort.

The in-hospital cardiac arrest and cardiopulmonary resuscitation were significantly higher in the ECMO cohort (17.5% vs 25.9% and 19.5% vs 31.3%, respectively). The hospital stay and the cost were significantly lower in the Impella cohort (7 days vs 11 days and $66,078 vs $122,996, respectively).

Propensity-score matched in-hospital mortality was significantly higher in the ECMO cohort (43.3% vs 26.7%, p=0.021). Secondary outcomes including vascular complications were also significantly higher in the ECMO cohort.

Comparison of cost and outcomes between Impella and ECMO

Comparison of cost and outcomes between Impella and ECMO

                                                           Illustration credit: Cardiovascular Revascularization Medicine

Critical reading and the relevance for clinical practice

In this propensity-matched cohort, the use of Impella was associated with lower in-hospital mortality, fewer complications, decreased hospital costs and decreased length of stay when compared to ECMO. The higher rate of ischemic stroke and vascular complications in the ECMO arm of this cohort will be explored further in the EUROSHOCK and ECLS-SHOCK studies on the use of ECMO. What is unique in this study by Lomar et al is the emphasis on propensity matching. As the investigators note, any study that does not include the use of vasopressors and inotropes, hemodynamics, lactate level, duration of shock has limited value.

To date, it remains difficult to discern which shock patients will succumb to shock and which will recover. As such, the world of interventional cardiology will continue to treat patients with AMI aggressively. Observational studies such as this will continue to provide guidance with respect to cost, hospital stay, as well as clinical and procedural outcomes until large randomised trials are available. However, a major limitation of this study is that the definition of shock was not assessed in the analysis. It is conceivable that advanced or refractory shock skewed the results and are a futile entity in whom MCS is not recommended. In addition, this study is not representative of all subgroups including women (17 and 26.7% in the ECMO and Impella arms respectively) and non-Caucasians (< 30% in each arm).

Finally, we remain in need of hub centers for shock. The study outcomes are difficult to apply universally as not all centers have streamlined protocols, expertise in the field and the necessary MCS devices. Escalation of therapy and use of multiple devices (eg Impella to vent the left ventricle in those with ECMO support) should be defined in a well-designed randomised trial.