Extracorporeal Membrane Oxygenation (ECMO) is an invasive and resource intense treatment used to support critically ill patients who have suffered severe cardiac arrest, cardiac failure or respiratory failure (including severe cases of COVID-19). ECMO acts as a mechanical circulatory support temporarily replacing the function of the heart or lungs by oxygenating blood and removing carbon dioxide, allowing time for these organs to recover. Many critically ill patients, including those on ECMO, have an increased risk of bleeding and reduced production/increased destruction of red blood cells (RBCs). This can lead to anaemia (haemoglobin levels <120 g/l), a condition where the body lacks enough healthy RBCs to carry enough oxygen to the body's tissues. Therefore, patients on ECMO frequently require RBC transfusion, with clinicians having to decide if administering an RBC transfusion (with its associated risks) is higher than tolerating complications of anaemia.

ROSETTA is a feasibility study that aims to determine the safety and feasibility of randomizing patients on ECMO to a restrictive RBC transfusion strategy (maintain Hb concentration above 70g/L) or to a more liberal transfusion strategy (maintain Hb concentration above 90g/L). Feasibility is defined as the ability to achieve a mean separation of at least 10g/L between the average lowest daily haemoglobin values in the two study groups.

A recent Cochrane analysis recommended a transfusion strategy that minimises the use of RBC transfusions in critically ill patients (by tolerating anaemia to avoid the adverse effects of an RBC transfusion). However, the analysis acknowledges that the degree of anaemia which can be tolerated by such patients is unknown, especially in patients suffering from conditions that limit oxygen delivery to the organs (like cardiac disease). As a result, the Australian Blood Authority's guidelines recommend an RBC transfusion to a patient at an Hb concentration of less than 70 g/L, while a transfusion at a Hb between 70 and 90 g/L should be based on the need to relieve clinical signs and symptoms of anaemia. However, this range is broad, and many studies in the general critically ill cohort have shown lower transfusion triggers are non-inferior to higher transfusion triggers.

No studies have been completed directly evaluating transfusion triggers in the ECMO patient cohort. ECMO patients differ to the general critically ill cohort as they have different physiological requirements, are at higher-risk for poor outcomes, and have an increased requirement for transfusions. Hb is a key driver of oxygen delivery (DO2), and critically ill ECMO patients are more commonly exposed to low DO2 due to low cardiac output and borderline oxygenation. Therefore, studies must be done to evaluate the optimal transfusion trigger/s (as determined by Hb concentration) that optimise mortality and long-term outcomes of ECMO patients.

Should the ROSSETTA Pilot results indicate adequate separation of at least 10g/L between the two study groups, and that patient safety has not been adversely affected by the trial methods, feasibility will be deemed confirmed and the protocol not in need of modification prior to full trial commencement. At this point the ROSETTA Pilot will be transitioned into the Red Blood Cell Transfusion Domain, within RECOMMEND Platform Trial. The Primary and Secondary outcomes of the trial at large, will be answered during this stage.

Inclusion Criteria:

• Patients receiving ECMO
• Age: 18 years or older

Exclusion Criteria:

• Contraindication to RBC transfusion (including known patient preference)
• Limitations of care put in place either through patient wishes or the treating medical teams
• ECMO treatment for more than 12 hours. The start of ECMO is defined as the time of initiation of extracorporeal blood flow unless ECMO was initiated during a surgical intervention in which case the start is defined as the arrival time into the initial ICU.
• The treating physician anticipates that ECMO treatment will cease before the end of tomorrow
• Where the treating physician deems the study is not in the patient's best interest
• Where the treating physician has concern regarding patient ability to tolerate restrictive or liberal transfusion trigger thresholds
• Patients actively listed for a solid organ transplant
• Patients who are suspected or confirmed to be pregnant
• Previous ECMO treatment during the same hospital admission

Patient data is de-identified and only aggregate summaries published.