Correlation of Coagulation Parameters in Patients on ECMO: A Retrospective Cohort Study

Introduction: Extracorporeal membrane oxygenation (ECMO) is frequently complicated by hemorrhagic and thrombotic events, making anticoagulation management challenging. Unfractionated heparin (UFH) is widely used, but practice guidelines do not provide a preferred monitoring method. While point-of-care activated clotting time (ACT) is commonly measured, there are limited data to show that this method correlates with activated partial thromboplastin time (aPTT) and anti-factor Xa (anti-Xa) levels. The objective of this study was to evaluate the concordance between ACT and aPTT levels and between ACT and anti-Xa levels in patients on ECMO.


Methods: This was a single-center, retrospective cohort study of adult patients on ECMO support from January 2018 to December 2023. An appropriate window of 30 minutes before or after the time of ACT collection was established for pairing of coagulation labs with ACT. The primary outcome was the concordance of the relationships between ACT and anti-Xa as well as ACT and aPTT, which was determined by utilizing bootstrap re-sampling to calculate the R2. Cohen’s Kappa was used to assess agreement of therapeutic and non-therapeutic values between ACT, anti-Xa, and aPTT.


Results: A total of 47 patients were included in the study, with 292 paired ACT-anti-Xa labs and 152 paired ACT-aPTT labs. 51.1% of patients underwent veno-arterial ECMO. The mean time for ECMO support was 221.6 hours (SD 147), and all patients received intravenous heparin infusion for prevention of circuit thrombosis. For the correlation of anti-Xa versus ACT levels, the R2 was 0.082 (95% CI 0.0041-0.18). The weighted kappa value was 0.068 (95% CI -0.011 – 0.15), indicating no to slight agreement. For the correlation of aPTT versus ACT levels, the R2 was 0.50 (95% CI 0.39-0.59). The weighted kappa value was 0.18 (95% CI 0.0047-0.36), indicating slight agreement.


Conclusions: This single-center retrospective cohort study demonstrates that ACT and aPTT monitoring methods often agree and are modestly correlated, while ACT and anti-Xa often disagree and are poorly correlated. Clinicians should avoid assumptions that these three coagulation assays are interchangeable, as utilizing ACT over an institution’s preferred method of monitoring intravenous heparin can lead to misinformed dosing decisions.