The threshold for packed red blood cell (pRBC) transfusion has been an important focus in the management of critically ill patients since the publication of randomized control trials (RCTs) demonstrating no significant difference in mortality between liberal and conservative transfusion strategies.^1,2 Consequently, there has been a paradigm shift in clinical management favoring restrictive transfusion strategies in the last 2 decades.^8–14 However, it is notable that these studies did not compare outcomes in patients with normal hemoglobin levels to those with anemia or to those with impaired oxygen delivery. Nevertheless, transfusion of blood products has also been associated with poorer outcomes in surgical and nonsurgical patients in the intensive care unit (ICU).^3,4 Thus, hemoglobin thresholds for pRBC transfusion need to balance the risks and benefits of liberal compared to restrictive transfusion practices in similar settings.¹⁵ Although the current guidelines for blood transfusion in critically ill patients generally support a hemoglobin threshold of 7 g/dl in nonbleeding, hemodynamically stable patients in ICUs, the ability to oxygenate the blood directly during extracorporeal membrane oxygenation (ECMO) makes the physiology of oxygen delivery more nuanced in this patient population.¹⁶

Establishing guidelines for transfusion thresholds in ECMO patients has proven challenging due to lack of well-conducted RCTs, heterogeneity in patient characteristics and underlying conditions, ECMO configurations, and institutional practices. A global survey on hemoglobin thresholds and transfusion during venovenous ECMO showed wide variability in blood transfusion practice among institutions.¹⁷ The prospective multicenter observational study on transfusion practice in VV-ECMO patients (PROTECMO), the largest published multicenter prospective study that examined transfusion practices and outcomes during ECMO, demonstrated that patients with Hb <7 g/dl had an increased risk of 28 day mortality, and that blood transfusion was associated in those with a reduced risk of death.⁵ The two studies highlighted that for patients with Hb >7 g/dl, the decision to transfuse during ECMO runs cannot rely solely on a single value of Hb. The European Society of Intensive Care Medicine guidelines on transfusion strategies in critically ill adults did not make a recommendation for a restrictive (7 g/dl) versus a liberal transfusion (9 g/dl) threshold in critically ill adults undergoing venovenous or venoarterial ECMO, acknowledging that the quality of evidence was inadequate to make a recommendation.⁶ The Transfusion and Anemia Expertise Initiative guidelines for pediatric ECMO stated that there is insufficient evidence to support a specific hemoglobin or hematocrit threshold for red blood cell transfusion.⁷ An expert panel review recommended that the decision to transfuse pRBC in children needing ECMO should be based on inadequate cardiorespiratory support, or decreased systemic or regional oxygen delivery, rather than a fixed hemoglobin or hematocrit value.