Abstract： Objective:To investigate the effect of different timing of arterial -venous extracorporeal membrane oxygenation (VA-ECMO) on the prognosis of patients with acute myocardial infarction complicated with cardiogenic shock (AMICS).Methods:This study was a prospective cohort study. AMICS patients received VA-ECMO support primary percutaneous coronary intervention in Henan Provincial People's Hospital from May 2017 to July 2023 were divided into early VA-ECMO group and late VA-ECMO group. 64 AMICS patients who met the indications for VA-ECMO implantation, but did not revive VA-ECMO were included as control group. Demographic characteristics, coronary interventional (PCI) information and complications after VA-ECMO implantation were collected. The primary end points was 1-year survival, minor end point were in-hospital and perioperative death. Multivariate Logistic and Cox regression models were used to evaluate the effect of timing of VA-ECMO on prognosis of AMICS patients. Kaplan-Meier survival curve was used to analyze the 1-year survival outcome of the 3 groups.Results:A total of 143 AMICS patients were included, and materials of 136 patients entered in the final analysis, including 42 in the early VA-ECMO group, 34 in the late VA-ECMO group, and 60 in the non-VA-ECMO group. Compared with the late VA-ECMO group, the early VA-ECMO group had a higher ratio of PPCI after VA-ECMO, a longer D-to-B time, a shorter VA-ECMO support time, a higher success rate of VA-ECMO withdrawal, and a lower complication rate (all P<0.05). Compared with the early VA-ECMO group, the perioperative, in-hospital and 1-year mortality were significantly higher in Non-ECMO support (all P<0.05). There was no difference in perioperative and in-hospital mortality between the early VA-ECMO group and the late VA-ECMO group, but the 1-year mortality in the late VA-ECMO group was significantly higher ( P<0.05). Perioperative, in-hospital and 1-year mortality rates were lower in the late VA-ECMO group than in the no-VA-ECMO group, but the differences were not statistically significant. Multivariate Logistic and Cox regression analysis showed that after adjusting interference factors, early VA-ECMO was still a protective factor for in-hospital ( OR=0.244, P=0.015) and one year ( HR=0.308, P=0.001)mortality. Kaplan-Merier survival curve showed that compared with the late VA-ECMO group and the group without VA-ECMO, the early VA-ECMO group had the highest 1-year survival rate. Conclusion:Patients with AMICS may benefit more from early VA-ECMO than from late VA-ECMO support for PPCI.