Volume-outcome relationships for extracorporeal membrane oxygenation in acute myocardial infarction

Acute myocardial infarction (MI) is one of the major scenarios of extracorporeal membrane oxygenation (ECMO) use. The utilization of mechanical circulatory support systems including ECMO varies widely at the hospital level, while whether ECMO volume per hospital is associated with outcomes in acute MI is unclear. Using a Japanese nationwide administrative database, a total of 26,913 patients with acute MI undergoing percutaneous coronary intervention from 154 hospitals were included. The relations among PCI volume for acute MI, observed and predicted in-hospital mortality, and observed and predicted rates of ECMO use were evaluated at the hospital level. Of 26,913 patients, 423 (1.6%) were treated with ECMO, and 1561 (5.8%) died during the hospitalization. Median ECMO use per hospital per year was 0.5. An observed rate of ECMO use was linearly correlated with the predicted probability of ECMO use and was not associated with the observed/predicted in-hospital mortality ratio. The observed/predicted mortality ratio was lowest in hospitals with the observed/predicted ECMO use ratio of around one. In conclusion, ECMO was infrequently used in a setting of acute MI at each hospital annually. An observed rate of ECMO use was not associated with observed/predicted in-hospital mortality ratio, while the observed/predicted in-hospital mortality ratio was lowest when ECMO was used as predicted, suggesting that standardized ECMO use may be an institutional quality indicator in acute MI.