DOI: 10.1161/jaha.126.049443 ISSN: 2047-9980

National Use Patterns and Comparative Effectiveness of Extracorporeal Membrane Oxygenation in Myocardial Infarction–Related Cardiogenic Shock

Ajar Kochar, Christina Lalani, Yang Song, Zaid Almarzooq, Andrew Farb, Changfu Wu, Bram Zuckerman, Navin K. Kapur, Jacob E. Moller, Eric A. Secemsky, Arthur R. Garan, Robert W. Yeh
Show PDF Cite

Background

Extracorporeal membrane oxygenation (ECMO) provides bi‐ventricular support for patients with myocardial infarction–related cardiogenic shock (MI‐CS); comparative effectiveness data relative to other mechanical circulatory support remain limited. We characterize contemporary ECMO utilization for MI‐CS and compare short‐term mortality across ECMO and non‐ECMO treatment strategies.

Methods

We compared baseline characteristics and temporal trends in mechanical circulatory support utilization among Medicare beneficiaries hospitalized with MI‐CS from 2016 to 2023 and treated with ECMO, Impella, intra‐aortic balloon pump, or medical management alone. Patient outcomes for each group were evaluated using inverse probability of treatment weighting, G‐formula approach, and grace period analysis. The primary endpoint was 30‐day all‐cause mortality.

Results

The cohort included 102 718 patients: 1165 (1.1%) treated with ECMO, 10 064 (9.8%) Impella, 24 631 (24.0%) intra‐aortic balloon pump, and 66 858 (65.1%) medical management alone. From 2016 to 2023, ECMO use for MI‐CS increased from 0.8% to 1.8%. The percentage of hospitals using ECMO increased from 9.2% to 17.1%. Unadjusted 30‐day mortality was higher for ECMO (69%) compared with Impella (60%), intra‐aortic balloon pump (39%), and medical management (46%) (all P <0.0001). Adjusted inverse probability of treatment weighting and G‐formula analyses revealed a 30‐day mortality rate of 68%–70% for ECMO, which was higher than that for other therapies. The grace period analysis revealed similar mortality between patients treated with ECMO (54%) and Impella (54%).

Conclusions

ECMO use in a contemporary Medicare cohort of patients with MI‐CS was infrequent (1.1%) and associated with high 30‐day mortality. Despite use of complementary causal inference methods, we were unable to establish valid causal inferences on ECMO use and patient survival.

More from our Archive