ECMO use in a contemporary Medicare cohort of patients with MI-CS was infrequent (1.1%) and associated with high 30-day mortality, and despite complementary causal inference methods, the authors were unable to establish valid causal inferences on ECMO use and patient survival.
ECMO was used in a small minority of Medicare beneficiaries hospitalized with MI-CS from 2016 to 2023, while most received medical management alone.
ECMO utilization for MI-CS increased over the study period at both the patient and hospital levels.
Unadjusted 30-day mortality was highest among patients treated with ECMO.
Adjusted analyses using inverse probability of treatment weighting and G-formula still showed higher 30-day mortality with ECMO than with other therapies.
In the grace period analysis, 30-day mortality was similar between ECMO and Impella.
The authors could not establish valid causal inferences about ECMO use and survival despite using complementary causal inference methods.
This study looked at Medicare patients hospitalized with cardiogenic shock after a heart attack, a condition in which the heart cannot pump enough blood. It examined how often ECMO, a machine that takes over heart and lung function, was used between 2016 and 2023, and compared 30-day death rates with other approaches: the Impella pump, the intra-aortic balloon pump, or medical management alone. Of about 102,700 patients, only about 1 in 100 received ECMO. Its use roughly doubled over the period (0.8% to 1.8% of patients), and the share of hospitals offering it rose from about 9% to 17%. Patients who received ECMO had the highest death rate at 30 days (69%, compared with 60% for Impella, 39% for balloon pump, and 46% for medical management alone). Statistical adjustments for differences between patients still showed death rates of 68%-70% with ECMO. A different analysis, designed to handle timing of treatment, found similar death rates for ECMO and Impella (54% each). This research suggests ECMO is still rarely used and is given to very sick patients with high mortality. However, the authors concluded that they could not determine whether ECMO itself affects survival, because patients receiving different treatments differ in ways that are hard to fully account for in observational data. Better evidence would likely need to come from other study designs, such as randomized trials.
Kochar A, Lalani C, Song Y, Almarzooq Z, Farb A, Wu C, et al.. (2026). National Use Patterns and Comparative Effectiveness of Extracorporeal Membrane Oxygenation in Myocardial Infarction-Related Cardiogenic Shock.. Journal of the American Heart Association. https://doi.org/10.1161/JAHA.126.049443