Blood Management During ECMO for Cardiac Support

Background: Extracorporeal Membrane Oxygenation (ECMO) is the most frequently used mechanical circulatory support (MCS) for severe forms of cardiogenic shock (CS) when simpler measures like inotropes and the intra-aortic balloon pump fail. Despite improvements in technology and growing clinical experience mortality among patients supported with ECMO remains above 40%. In addition to bloodstream infections and circuit-related complications, disorders of coagulation which can result in life-threatening hemorrhage, thrombosis and thromboembolic events remain the most frequent and dangerous complications. Significant perturbations in hemostatic function are seen during extracorporeal therapies. The mechanisms of these are complex and relate to underlying pathology, blood-surface contact and pump mechanics. They are evidenced by platelet dysfunction and consumption, hyperfibrinolysis or coagulopathy secondary to disseminated intravascular coagulation (DIC). Given the multifactorial disturbances of hemostasis that can result in hemorrhage during ECMO, standard laboratory hemostatic testing may not enable the clinician to specify the underlying cause of the bleeding. Consequently, bleeding events are typically managed with transfusion of coagulation factors guided by clinician judgment. Anticoagulation practice itself may also determine the bleeding risk and it has been noted that practice is variable. Heparin dosage and titration as well as stopping rules are not well defined and vary vastly in the literature. Historically, bleeding had been reported in clinical trials where high anticoagulation targets were used which contributed to excessive bleeding. Bleeding and the utilization of blood products have frequently been described as a predictor of poor outcome in the general ICU population. Guidelines suggesting high hemoglobin targets to optimize oxygen delivery in ECMO have subsequently been challenged to avoid potential risk associated with blood product transfusions. Excessive transfusion of blood products, which increases the risk of complications such as transfusion-related acute lung injury (TRALI), transfusion-related circulatory overload (TACO) and transfusion-related immunomodulation (TRIM) may outweigh the potential benefit. Unnecessary exposure to procoagulant factors also increases the risk of ECMO-circuit malfunction and thromboembolic complications. These additionally lead to increased ICU and hospital length of stay, and markedly increased costs. Improvement in the management of anticoagulation and hemostatic therapy in this setting may alter outcome in ECMO. Current knowledge is insufficient to provide high level evidence-based recommendations on anticoagulation, management of bleeding or transfusion triggers. Published literature is mostly drawn from retrospective single centre reports over different time periods, case mix, ECMO equipment and data points. In the absence of randomized controlled trials (RCT) addressing these multitude of issues, best practice should be based on multicenter cohort studies. A standardized multicenter approach is needed to observe associations between variations in practice and clinically relevant outcomes. These observations will support hypothesizes for future RCT and is crucial to propose feasible interventions. Hypothesis: • Bleeding, thromboembolic complications and blood product usage in ECMO for MCS * Vary among centers providing the service * Differ with different anticoagulation strategies * Are dependent on diagnostic grouping and presentation category * Are independently associated with morbidity and mortality Design: • An international, multicenter, observational cohort study. Aim: * To prospectively describe bleeding, thromboembolic complications, blood management and anticoagulation in ECMO for MCS. * Investigate its associations with morbidity and mortality. * Identify risk factors for bleeding, thromboembolic complications and blood product usage. * To provide information on current standard of care for future prospective interventional trials. Significance The Extracorporeal Life Support Organization (ELSO) maintains a registry and reports 4297 annual adult ECMO runs worldwide in 2016. This number has more than doubled since 2012. Since this registry relies on voluntary contributions it is almost certainly only reporting a fraction of all global treatments. Using national insurance datasets McCarthy reports almost 3000 usages in the US in 2012 (ELSO registry: less than 1000 for North America). Using a similar approach Karagiannidis found almost 4500 annual runs in Germany in 2014 (ELSO registry: 921 for Europe in total during the same year). Using the Japanese Diagnosis Procedure Combination national inpatient database 5263 patients were treated in that country for cardiac support only over a period of 3 years from 2010 to 2013 (ELSO registry: less than 500 for the whole Asia-Pacific area). No global statistic of ECMO usage exists but it can be assumed that the largest registry probably underestimates the true utilisation by a factor of 5-20. This would translate into an annual number of patients treated with ECMO of around 20.000 to 100.000. The mortality attributed to bleeding and thromboembolic complications is also unknown. Based on ELSO data Lorusso found the incidence of stroke, cerebral haemorrhage, brain death and seizures to be 7% with an associated mortality of more than 70% in patients treated with ECMO for respiratory support. Given that these patients were primarily diagnosed with a pulmonary disease and a low risk profile for cerebral complications, ECMO may be a contributor to this negative outcome which may potentially be avoidable. Other causes of bleeding and thromboembolic complications during ECMO for MCS are more frequent and reported as: * Clots: oxygenator 8.2% * Other clots in the ECMO system 5.6% * Gastrointestinal haemorrhage 4.4% * Cannulation site bleeding 17.5% * Surgical site bleeding 19.2% * Haemolysis 5.5% * Disseminated intravascular coagulation (DIC) 3.5% * Cardiac tamponade: blood 5% * Pulmonary haemorrhage 2.8% * Limb Ischemia 3.6% * Compartment syndrome 1% * Limb Fasciotomy 1.5% * Limb Amputation 0.5% While these registry data do not allow insight into a causative correlation with mortality, bleeding and thromboembolic complications almost certainly contribute to morbidity and treatment costs. The need for blood transfusions has been quantified to be between 1 and 4 Units per day of ECMO treatment which may further contribute to mortality and morbidity through direct and indirect transfusion related complications. Best strategies to avoid blood transfusions in ECMO have not been identified. The costs directly related to ECMO are difficult to assess. However, so far the largest randomized trial has calculated the additional costs of patients treated with the modality at about 40.000 pounds while other observational studies measured a direct treatment related cost of 73.000 USD and 106.000 Euro. Even a conservative estimation of the annual case load would quantify the global treatment related costs at more than 1.5 billion USD. Using the same conservative approach and assuming an average ECMO treatment duration of 10 days would translate into 200.000 units of blood transfused. Given the reported direct costs of a blood transfusion of USD 265 the costs would be about 50 Million USD per annum for red blood cells alone. Indirect costs related to TRALI and other complications as well as administration costs cannot be assessed but are most likely much higher. Specific data on blood management and anticoagulation have only been published in single centre retrospective trials. This study's targeted recruitment of 500 patients would represent by far the largest contemporary sample in this area worldwide.