Anticoagulation and Transfusions Management in Veno-Venous Extracorporeal Membrane Oxygenation for Acute Respiratory Distress Syndrome: Assessment of Factors Associated With Transfusion Requirements and Mortality

Purpose: We describe an approach for anticoagulation and transfusions in veno-venous-extracorporeal membrane oxygenation (VV-ECMO), evaluating factors associated with higher transfusion requirements, and their impact on mortality.

Methods: Observational study on consecutive adults supported with VV-ECMO for acute respiratory distress syndrome (ARDS). We targeted an activated partial thromboplastin time of 40 to 50 seconds and a hematocrit of 24% to 30%. Univariate and multiple analyses were done to evaluate factors associated with transfusion requirements and the influence of increasing transfusions on mortality during ECMO.

Results: In a cohort of 82 VV-ECMO patients (PRedicting dEath for SEvere ARDS on VV-ECMO [PRESERVE] score: 4, Interquartile range [IQR]: 3-5, Respiratory Extracorporeal Membrane Oxygenation Survival Prediction [RESP] score: 2, IQR: 2-4), 76 (92.7%) patients received at least 1 unit of packed red blood cells (PRBCs) during the intensive care unit stay related to ECMO (median PRBC/d 156 mL, IQR: 93-218; median ECMO duration 14 days, IQR: 8-22). A higher requirement of PRBC transfusions was associated with pre-ECMO hematocrit, and with the following conditions during ECMO: platelet nadir, antithrombin III (ATIII), and stage 3 of acute kidney injury (all P < .05). Sixty-two (75.6%) patients survived ECMO. Pre-ECMO hospital stay, PRBC transfusion, and septic shock were associated with mortality (all P < .05). The adjusted odds ratio for each 100mL/d increase in PRBC transfusion was 1.9 (95% confidence interval [CI]: 1.1-3.2, P = .01); for the development of septic shock it was 15.4 (95% CI: 1.7-136.8, P = .01), and for each day of pre-ECMO stay it was 1.1 (95% CI: 1-1.2, P = .04).

Conclusion: Implementation of a comprehensive protocol for anticoagulation and transfusions in VV-ECMO for ARDS resulted in a low PRBC requirement, and an ECMO survival comparable to data in the literature. Lower ATIII emerged as a factor associated with increased need for transfusions. Higher PRBC transfusions were associated with ECMO mortality. Further investigations are needed to better understand the right level of anticoagulation in ECMO, and the factors to take into account in order to manage personalized transfusion practice in this select setting.