Predictors of Mortality in Venoarterial Extracorporeal Membrane Oxygenation Regardless of Early Left Ventricular Unloading: A National Experience

Abstract

Objective

The use of an intra-aortic balloon pump (IABP) has been suggested to unload the left ventricle while on venoarterial extracorporeal membrane oxygenation (VA-ECMO) for cardiogenic shock (CS), leading to possibly improved in-hospital mortality. However, the predictors of mortality on dual mechanical circulatory support have not yet been evaluated, especially in real-world clinical settings. Therefore, a case-control study was conducted to determine the rate of all-cause mortality associated with VA-ECMO use regardless of left ventricular (LV) unloading, and with early LV unloading in the setting of CS, and to identify the predictors of mortality associated with VA-ECMO, with concurrent early LV unloading.

Design

Retrospective observational case-control study.

Setting

National tertiary cardiology center.

Participants

All patients with CS requiring VA-ECMO cannulation during the index admission between January 06, 2016, and January 0, 2022.

Intervention

VA-ECMO with or without IABP

Measurements and Main Results

Patient- and disease-related characteristics associated with in-hospital 30-day mortality following VA-ECMO with and without IABP support were assessed using multivariate logistic regression. Results are presented as odds ratio (OR), and a p-value < 0.05 indicates statistical significance. A total of 110 patients were included. Most were male (90%) with a mean age of 53 ± 11 years. Around 67% were Asian. The majority of patients were admitted with ST-elevation myocardial infarction (87%), with 26% presenting with left main disease. In-hospital 30-day mortality occurred in 42.7% of those who received VA-ECMO support regardless of IABP use, while it was 46.9% among those receiving early LV unloading with IABP. Significant positive predictors of mortality with VA-ECMO regardless of IABP in CS were cardiopulmonary resuscitation (CPR) >20 minutes (adjusted OR 14.74, 95% confidence interval 2.02-107.41, p-value = 0.008), older age (ie, >55 years) and left main disease of more than 50% stenosis were associated with a fourfold increase in the odds of mortality while on VA-ECMO. Conversely, CPR >20 minutes (adjusted OR 12.45, 95% confidence interval 1.79-86.36, p-value = 0.011) was the only significant positive predictor of mortality with VA-ECMO and IABP.

Conclusion

The mortality rate in CS requiring VA-ECMO, regardless of IABP use, remains high. However, only one predictor (ie, prolonged CPR) was found to increase the likelihood of 30-day mortality with early LV unloading, suggesting that concomitant IABP use might minimize the effect of mortality predictors.