Abstract
OBJECTIVE: To describe the incidence of bleeding and thrombotic complications in VA-ECMO according to anticoagulation strategy. DESIGN: This systematic review and meta-analysis included randomised controlled trials (RCTs) and observational studies reporting bleeding and thrombotic complications in VA-ECMO. The incidence of primary outcomes according to anticoagulation drug and monitoring test was described. DATA SOURCES: CENTRAL, MEDLINE, Embase and CINAHL (2010-January 2024). REVIEW METHODS: Data was extracted using Covidence. A meta-analysis of proportions was performed using STATA MP v18.1 metaprop. RESULTS: We included 159 studies with 21,942 patients. No studies were at low risk of bias. The incidence of major bleeding or thrombotic events was similar among heparin-, bivalirudin- and anticoagulation-free cohorts. The pooled incidence of major bleeding and thrombotic complications were 40% (95%CI 36-44, I(2) = 97.12) and 17% (95%CI 14-19, I (2) = 92.60%), respectively. The most common bleeding site was thoracic. The most common ischaemic complication was limb ischaemia. The incidences of major bleeding or thrombotic events, intracranial haemorrhage and ischaemic stroke were similar among all monitoring tests. Mechanical unloading was associated with a high incidence of major bleeding events (60%, 95%CI 43-77, I(2) = 93.32), and ischaemic strokes (13%, 95%CI 7-19, I(2) = 81.80). CONCLUSIONS: Available literature assessing the association between anticoagulation strategies in VA-ECMO, and bleeding and thrombosis is of limited quality. We identified a substantially higher incidence of major bleeding events than a previous meta-analysis. Limited numbers of patients anticoagulated with alternatives to heparin were reported. Patients with additional mechanical LV unloading represent a cohort at particular risk of bleeding and thrombotic complications.