Abstract
Background: Bleeding is a common complication of extracorporeal membrane oxygenation (ECMO) for pediatric cardiac patients. We aimed to identify anticoagulation practices, cardiac diagnoses, and surgical variables associated with bleeding during pediatric cardiac ECMO by combining two established databases, the Collaborative Pediatric Critical Care Research Network (CPCCRN) Bleeding and Thrombosis in ECMO (BATE) and the Extracorporeal Life Support Organization (ELSO) Registry. Methods: All children (<19 years) with a primary cardiac diagnosis managed on ECMO included in BATE from six centers were analyzed. ELSO Registry criteria for bleeding events included pulmonary or intracranial bleeding, or red blood cell transfusion >80 ml/kg on any ECMO day. Bleeding odds were assessed on ECMO Day 1 and from ECMO Day 2 onwards with multivariable logistic regression. Results: There were 187 children with 114 (61%) bleeding events in the study cohort. Biventricular congenital heart disease (94/187, 50%) and cardiac medical diagnoses (75/187, 40%) were most common, and 48 (26%) patients were cannulated directly from cardiopulmonary bypass (CPB). Bleeding events were not associated with achieving pre-specified therapeutic ranges of activated clotting time (ACT) or platelet levels. In multivariable analysis, elevated INR and fibrinogen were associated with bleeding events (OR 1.1, CI 1.0–1.3, p = 0.02; OR 0.77, CI 0.6–0.9, p = 0.004). Bleeding events were also associated with clinical site (OR 4.8, CI 2.0–11.1, p < 0.001) and central cannulation (OR 1.75, CI 1.0–3.1, p = 0.05) but not with cardiac diagnosis, surgical complexity, or cannulation from CPB. Bleeding odds on ECMO day 1 were increased in patients with central cannulation (OR 2.82, 95% CI 1.15–7.08, p = 0.023) and those cannulated directly from CPB (OR 3.32, 95% CI 1.02–11.61, p = 0.047). Conclusions: Bleeding events in children with cardiac diagnoses supported on ECMO were associated with central cannulation strategy and coagulopathy, but were not modulated by achieving pre-specified therapeutic ranges of monitoring assays. Copyright © 2022 Ankola, Bailly, Reeder, Cashen, Dalton, Dolgner, Federman, Ghassemzadeh, Himebauch, Kamerkar, Koch, Kohne, Lewen, Srivastava, Willett and Alexander.