Indication for ECMO predicts time to first actionable bleeding complication.

PURPOSE: Bleeding is a major complication of patients requiring extracorporeal membrane oxygenation (ECMO). Several risk factors have been identified; however, there remains a paucity of evidence for optimal management of anticoagulation and bleeding in ECMO patients.

METHODS: A total of 255 patients required ECMO from January 1996 to December 2021 at a single institution. The Bleeding Academic Research Consortium (BARC) Score was used for defining actionable bleeding. Univariate and multivariate testing were used for outcome analysis. Kaplan-Meier survival curves were plotted for time-to-event analysis.

RESULTS: Of the 255 patients, 147 patients had no actionable bleeding complications, while 108 had at least one actionable bleeding complication. Duration of support ( p <0.001) and total number of transfusions ( p <0.001) differed between the two groups significantly, with no significant difference in survival to discharge ( p =0.894). On multivariate regression, significant predictors for actionable bleeding complications included diabetes (OR 2.01, p =0.03), precannulation hematocrit (OR 0.97, p <0.001), length of support (OR 1.00, p <0.001), use of warfarin (OR 2.28, p =0.03), and post-cardiotomy indication for ECMO (OR 0.77, p =0.02). The median time to first actionable bleeding complication after cannulation was 141.2 h. When stratified by indication for ECMO or type of ECMO circuit, there was a significant difference in time to first actionable bleeding complication ( p =0.001, p=0.018).

CONCLUSIONS: Indication for ECMO and type of ECMO circuit both are predictive of timing to first actionable bleeding complication in our study. Further data are needed to reliably establish individualized anticoagulation strategies and bleeding management based on indication and circuit setup.

SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1007/s12055-023-01601-9.