Extracorporeal membrane oxygenation in the therapy of cardiogenic shock: 1‐year outcomes of the multicentre, randomized ECMO‐CS trial

Aims Among patients with cardiogenic shock, immediate initiation of extracorporeal membrane oxygenation (ECMO) did not demonstrate any benefit at 30 days. The present study evaluated 1‐year clinical outcomes of the Extracorporeal Membrane Oxygenation in the therapy of Cardiogenic Shock (ECMO‐CS) trial. Methods and results The ECMO‐CS trial randomized 117 patients with severe or rapidly progressing cardiogenic shock to immediate initiation of ECMO or early conservative strategy. The primary endpoint for this analysis was 1‐year all‐cause mortality. Secondary endpoints included a composite of death, resuscitated cardiac arrest or implantation of another mechanical circulatory support device, duration of mechanical ventilation, and the length of intensive care unit (ICU) and hospital stays. In addition, an unplanned post‐hoc subgroup analysis was performed. At 1 year, all‐cause death occurred in 40 of 58 (69.0%) patients in the ECMO arm and in 40 of 59 (67.8%) in the early conservative arm (hazard ratio [HR] 1.02, 95% confidence interval [CI] 0.66–1.58; p = 0.93). The composite endpoint occurred in 43 (74.1%) patients in the ECMO group and in 47 (79.7%) patients in the early conservative group (HR 0.83, 95% CI 0.55–1.25; p = 0.29). The durations of mechanical ventilation, ICU stay and hospital stay were comparable between groups. Significant interaction with treatment strategy and 1‐year mortality was observed in subgroups according to baseline mean arterial pressure (MAP) indicating lower mortality in the subgroup with low baseline MAP (<63 mmHg: HR 0.58, 95% CI 0.29–1.16; p interaction = 0.017). Conclusions Among patients with severe or rapidly progressing cardiogenic shock, immediate initiation of ECMO did not improve clinical outcomes at 1 year compared to the early conservative strategy. However, immediate ECMO initiation might be beneficial in patients with advanced haemodynamic compromise.