The delicate balance of ECMO: LV unloading and bleeding complications

Abstract Background Bleeding is one the most frequent and fatal complications in patients undergoing extracorporeal membrane oxygenation (ECMO). Despite its clinical significance, reliable prognostic factors for predicting hemorrhage risk in ECMO patients remain undefined. This study aims to identify clinical and laboratory parameters routinely measured in a cardiac intensive care unit (CICU) to improve hemorrhage risk prediction. Methods We conducted a retrospective analysis of 58 patients treated with veno-arterial ECMO (VA-ECMO) in the CICU of a tertiary hospital between 2020 and 2024. Significan bleeding was defined as a hemoglobin drop ≥2 g/dL, blood loss 20 mL/kg, transfusion requirement ≥10 mL/kg within 24 hours, retroperitoneal or pulmonary bleeding, or the need for surgical intervention to control hemorrhage. Results The cohort consisted of 58 patients with a mean age of 53 (±1.62) years; 41% were male. Among these, 47% received VA-ECMO alone, 24% VA-ECMO combined with Impella (ECPELLA), and 29% VA-ECMO with an intra-aortic balloon pump (IABP). The primary indications for ECMO were cardiogenic shock (67%), refractory cardiac arrest (26%), and arrhythmic storm (7%). Significant hemorrhage occurred in 31% of patients (n = 18). A statistically significant correlation was observed between the type of circulatory support and bleeding risk (p = 0.024). Specifically, the combination of VA-ECMO and Impella was associated with a higher hemorrhage risk compared to VA-ECMO alone (p = 0.003) or VA-ECMO with IABP (p = 0.005). No significant difference was found between VA-ECMO alone and VA-ECMO with IABP (p = 0.12). Furthermore, the presence of significant pericanular bleeding (defined as the need for gauze changes every four hours) was predictive of major hemorrhage in the following hours (p = 0.015). Pericanular bleeding was more frequently associated with the use of ECPELLA compared to VA-ECMO alone (p = 0.49) or VA-ECMO with IABP (p = 0.013). However, no differences were observed between VA-ECMO alone and VA-ECMO with IABP. Other factors, including age (p = 0.255), Charlson Index (p = 0.37) and maximum TTPa value (p =0.096),did not demonstrate significant associations with hemorrhage risk, despite being identified as predictors in previous studies. Conclusions In CICU patients receiving ECMO, the combination of MCS devices has been shown to have a significant impact on haemorrhage risk. The highest bleeding risk was associated with VA-ECMO combined with Impella. .These findings suggest that for patients requiring LV unloading who also have a high bleeding risk, clinicians should consider choosing VA-ECMO + IABP over ECPELLA. Additionally, close monitoring for pericanular bleeding may help identify patients at imminent risk of significant bleeding. Other clinical or analytical factors did not achieve statistical significance. Further research is needed to refine haemorrhage risk stratification and reduce morbidity and mortality in ECMO patients.Summary of key results Association between support and bleeding