Bivalirudin Versus Heparin in Adults Requiring Venoarterial ECMO for Surgical Indications

Introduction: Extracorporeal membrane oxygenation (ECMO) cannulation is thrombogenic, requiring systemic anticoagulation while balancing bleeding risk. Despite recognition of this challenge, current literature is largely retrospective and evaluates anticoagulants across mixed ECMO configurations, age groups, and indications. This leaves uncertainty regarding the optimal anticoagulation strategy for adult veno-arterial (VA) ECMO in the surgical population. We hypothesized that bivalirudin would reduce in-circuit thrombosis and bleeding events compared to unfractionated heparin (UFH).


Methods: This IRB-approved, retrospective, single-center study evaluated adult surgical VA-ECMO cases from July 2018 to September 2024 at a large academic center. Patients were excluded if they received no anticoagulation for >48 hours after cannulation, were cannulated at an outside institution, or were pregnant. If patients had crossover from one agent to another for >36 hours, they were excluded. The primary outcome was in-circuit thrombosis. Secondary outcomes included time to and in therapeutic range. Safety outcomes included bleeding per ELSO criteria, transfusion needs, and thrombotic complications. Statistical analysis was performed using Microsoft Excel.


Results: Thirty-eight patients were included: 14 received bivalirudin and 24 received UFH. Two patients had agent crossovers. The cohorts were similar in baseline characteristics, including heart transplant (71% vs 63%; p > 0.05) and median SOFA score at cannulation (12 vs 10; p > 0.05). There was no difference in in-circuit thrombosis (7% vs 8%, p > 0.05). Time to therapeutic partial thromboplastin time was significantly shorter with bivalirudin (4 hr vs 12 hr; p < 0.001), and time in therapeutic range was higher (76% vs 50%; p < 0.05). Major bleeding in the bivalirudin group was 14% vs 38% in the heparin group, although not statistically significant. All other bleeding events and transfusion requirements were similar between groups.


Conclusions: In this cohort, bivalirudin did not significantly reduce thrombosis, bleeding, or transfusion needs compared to UFH. However, it achieved therapeutic levels faster and had a higher proportion of time in therapeutic range, suggesting more predictable anticoagulation in surgical VA-ECMO patients.