Background: A selective group of patients, presenting with INTERMACS-1 cardiogenic shock due to acute ischaemic heart failure, can be considered for mechanical circulatory support (MCS). Patients with biventricular failure, severe septic shock or oxygenation problems should be selected for veno-arterial extracorporeal membrane oxygenation (VA-ECMO), although the percutaneous left Impella CP heart pump nowadays can be considered as a less-invasive alternative in supporting predominantly left ventricular failure. Bleeding issues are often the main concern in patients on MCS, especially in this group where triple anticoagulation therapy (unfractionated heparin [UFH] for the prevention of pump thrombosis and dual antiplatelet therapy [DAPT] after coronary stenting) is mandatory. We aim to investigate the bleeding and transfusion rate in DAPT patients on VA-ECMO
versus Impella.
Methods: We report single-centre data for 51 VA-ECMO and eight Impella patients between 2011 and 2019. Indication for MCS was acute ischaemic cardiogenic shock. Patient demographics, blood product transfusions and reported or radiographically diagnosed bleeding (BARC classification) complications were analysed. All patients received UFH and low-dose aspirin plus clopidogrel or ticagrelor. Impella CP flow was at least 2.5 l/min. Targets were haemoglobin 7 g/dl, fibrinogen 100 mg/dl (or 150 mg/dl when active bleeding) and platelet counts >50/fl.
Results: Patients supported by Impella were significantly older compared to the VA-ECMO group (VA-ECMO 52.8 years versus Impella 62.4 years; p=0.02). UFH levels and length of the MCS run were comparable in both groups. Total haemorrhage was comparable between both groups (mainly oozing from the insertion site in the Impella group; 63% versus VA-ECMO 72%; not significant), but major bleeds with BARC score ≥3 were significantly lower in the triple anticoagulated Impella group (DAPT 13% versus VA-ECMO 65%, p=0.005). Platelet and red blood cell transfusions were significantly lower in the Impella group (0.1 unit per day versus VA-ECMO 1.1, p=0.002, and 0.8 units per day versus VA-ECMO 2.6, p=0.02, respectively). Fresh frozen plasma transfusion rate was comparable.
Conclusion: Haemorrhage is frequent during extracorporeal support. However, in our cohort, triple anticoagulation in acute cardiogenic shock due to left ventricle failure resulted in a lower major bleeding rate when support was given by the left Impella device compared with VA-ECMO therapy. As a result, platelet and red blood cell transfusions were lower in the Impella group. These findings are probably partly explained by increased bleeding risk due to the cannula sizes and increased risk of consumptive coagulopathy due to the complexity and extensive foreign body surface of the ECMO circuit. We conclude that Impella support should be considered first choice in patients with indication for triple anticoagulant therapy and necessitating MCS towards recovery or other destination therapy.