Abstract Funding Acknowledgements Type of funding sources: None. Background The availability and accessibility to short-term mechanical circulatory support (ST-MCS) in each hospital may condition the choice of the device regardless of characteristics of patients with refractory cardiogenic shock (CS). Purposes The objectives were to stratify the risk of patients with CS assisted with percutaneous or minimally invasive support (VA-ECMO, Impella CP and 5.0), to analyze whether the choice of ST-MCS was adequate according to the characteristics of the patient (severity, respiratory failure and right ventricular dysfunction) and differences in prognosis. Methods Patients with CS requiring percutaneous or minimally invasive ST-MCS at a tertiary center from 2020 to August 2022 were prospectively included. Patients requiring immediately escalating from any device to VA-ECMO were included in the ECMO group for analysis. Early escalation from Impella CP to a more potent support (< 24 hours) was considered as inappropriate device selection. Bleeding was classified according to ELSO. Results A total of 40 patients with CS were supported: 15 with VA- ECMO (37,5%), 7 with Impella 5.0 (17,5%) and 18 with Impella CP (45%). Impella 5.0 was mostly placed in patients with a less severe cardiogenic shock (SCAI-C 71,4%, median lactate: 1,2 mmol/l (1-1,9), median VIS-score: 7,3 (3,1-26), median SOFA 5 (3,75-7)), mainly due to decompensated chronic heart failure (71%) as a bridge to heart transplant (71%). VA-ECMO was mostly canulated after acute myocardial infarction (AMI)-related cardiogenic shock (46,7%), mainly in patients with a profound shock (SCAI D 40% and E 53%, lactate 3(6-10,25), VIS score 80 (40-80), SOFA 8 (6-11)). Impella CP was mainly inserted after AMI (77,8%) in patients with intermediate severity (SCAI D (55,6%), lactate 3,9 (2,3-5,8), VIS 66 (29,4-95), SOFA: 7 (6-8). No differences were found in the need for mechanical ventilation (p= 0,426), nor the proportion of RV dysfunction (p= 0,635). Upgrading from Impella CP to VA-ECMO was required in 7 patients (38%), 4 immediately in the cath-lab (22%) and 2 within the first 24 hours of support (11,1%). Escalation from Impella CP to 5.0 was needed in 11,1%, all after > 24 hours. Mortality during support was significantly higher in VA-ECMO group (40 vs 0 vs 11%; p=0,041) and also in-hospital mortality (60 vs 14 vs 22%; p=0,035), but no differences were found in first 24-hours mortality and 30-day mortality. There were not differences in complications (Table 1). Conclusions Selection of first ST-MCS seems appropriate in 85% patients. Early escalation from Impella CP to VA-ECMO was necessary in 6 patients (15%, 33% Impella CP), probably implying inappropriate selection of first assistance probably due to faster and easier access to Impella CP. Higher in-hospital mortality and mortality during ST-MCS in VA-ECMO group, are probably explained by the greater severity of shock rather than the increase of complications during ST-MCS, which were similar among groups.