ExtracorporeaL life support and Modification Of Hemostasis: the ELMOH trial: hemostatic changes during the first 48 h of VV- and VA-ECMO: a prospective multicenter cohort study

Background: Anticoagulation during extracorporeal membrane oxygenation (ECMO) requires balancing thrombotic and bleeding risks, yet early coagulation dynamics may differ between venovenous (VV) and venoarterial (VA) support. This study prospectively characterized standard coagulation tests and viscoelastic profiles during the first 48 h of ECMO and related these to early bleeding and transfusion need. Methods: Multicenter, prospective cohort in four Belgian ICUs (03/2021-01/2023; NCT04912336). Adults initiated on VV- or VA-ECMO were enrolled immediately before cannulation and sampled at inclusion, + 2 h, + 24 h, and + 48 h. Laboratory tests (aPTT, PT/INR, fibrinogen, platelets, anti-Xa, D-dimer, AT), ROTEM, bleeding (BARC, GUSTO), and transfusions were recorded. Results: Forty-three patients were included (23 VV, 20 VA). At inclusion, platelet counts were higher in VV than in VA (median 292 vs. 145·109/L). Thrombocytopenia was present at enrollment in 30% and developed in 56% during the first 48 h (mild 14%, moderate 35%, severe 7%); both modes showed significant platelet decline (p < 0.001). Fibrinogen remained higher in VV, whereas VA exhibited lower fibrinogen and higher INR and aPTT at all time points; anti-Xa was higher at baseline in VA and declined over 48 h. CRP was higher in VV but rose in VA after initiation. ROTEM showed initial hypocoagulability in VA (prolonged CT/CFT, reduced A10/MCF) that normalized by 48 h; while VV remained hypercoagulable. Early bleeding occurred in 37% by BARC ≥ 2 (VA 55% vs. VV 22%) and 21% by GUSTO moderate/severe (VA 35% vs. VV 9%); between-group differences were nonsignificant. Transfusion occurred in 58%; no fresh frozen plasma was given to VV patients (significantly fewer than VA). One-year survival was 87% for VV and 60% for VA. Conclusions: Early hemostatic profiles differed markedly between patients requiring ECMO for circulatory failure and those supported for isolated respiratory failure. VA-ECMO showed lower platelet counts and a more hypocoagulable profile early after initiation that partially improved within 48 h, whereas VV-ECMO remained relatively hypercoagulable. These findings support the importance of individualized hemostatic monitoring and anticoagulation management.