Rationale Extracorporeal membrane oxygenation (ECMO) may provide mechanical pulmonary and circulatory support for patients with cardiogenic shock refractory to conventional medical therapy. Prediction of survival in these patients may assist in management of these patients and comparison of results from different centers. Aims To identify pre-ECMO factors which predict survival from refractory cardiogenic shock requiring ECMO and create the survival after veno-arterial-ECMO (SAVE)-score. Methods and results Patients with refractory cardiogenic shock treated with veno-arterial ECMO between January 2003 and December 2013 were extracted from the international Extracorporeal Life Support Organization registry. Multivariable logistic regression was performed using bootstrapping methodology with internal and external validation to identify factors independently associated with in-hospital survival. Of 3846 patients with cardiogenic shock treated with ECMO, 1601 (42 ) patients were alive at hospital discharge. Chronic renal failure, longer duration of ventilation prior to ECMO initiation, pre-ECMO organ failures, pre-ECMO cardiac arrest, congenital heart disease, lower pulse pressure, and lower serum bicarbonate (HCO3) were risk factors associated with mortality. Younger age, lower weight, acute myocarditis, heart transplant, refractory ventricular tachycardia or fibrillation, higher diastolic blood pressure, and lower peak inspiratory pressure were protective. The SAVE-score (area under the receiver operating characteristics [ROC] curve [AUROC] 0.68 [95 CI 0.64-0.71]) was created. External validation of the SAVE-score in an Australian population of 161 patients showed excellent discrimination with AUROC = 0.90 (95 CI 0.85-0.95). Conclusions The SAVE-score may be a tool to predict survival for patients receiving ECMO for refractory cardiogenic shock (www.save-score.com).