Red blood cell transfusion thresholds in acute leukemia: Current practice and physician perspectives from a national survey

Abstract Introduction Red blood cell (RBC) transfusion guidelines for patients with acute leukemia are lacking and transfusion practices are based on recommendations for other populations. Patients with acute leukemia often experience the combination of severe anemia and thrombocytopenia, placing them at heightened risk of bleeding. Given that red blood cell (RBC) rheology may enhance platelet function, patients with acute leukemia may benefit from more aggressive correction of anemia. To inform the design of a future randomized clinical trial (RCT) evaluating a liberal RBC transfusion strategy in patients with acute leukemia, we conducted a national survey in Canada to understand current RBC transfusion practices and physician perspectives on the feasibility of a clinical trial. Methods We developed a web-based survey using an electronic data capture platform (REDCap). The survey adopted Dillman's tailored design method, which entailed personalized communication, reminders, and a user-friendly interface. Experts in qualitative research, transfusion medicine, and acute leukemia care reviewed the survey questionnaire. The target sample was members of the MYELO-CAN network, a group of leukemia-treating physicians engaged in leukemia research in Canada. Participants received a personalized secure survey link via email. We collected data on current hemoglobin thresholds for RBC transfusion in clinical practice and attitudes regarding participation in, and feasibility of an RCT evaluating a liberal (Hg = 110 g/L) RBC transfusion threshold. Results Of 43 leukemia physicians from the MYELO-CAN network, 5 were excluded (4 did not treat leukemia and one could not be contacted) and 34 completed the survey (response rate = 89%). Respondents were from 7 Canadian provinces. Most (94%) practiced in academic hospitals and 2 (6%) in community hospitals. For stable, non-bleeding patients undergoing induction chemotherapy for acute leukemia, 27/34 (79%) used a RBC transfusion threshold of 70 g/L; 6/34 (18%) used 80 g/L; and 1/34 (3%) used 90 g/L in current practice. Of 34 respondents, 11 (32%) were willing to participate in a future RCT evaluating a liberal threshold (110 g/L); 21 (62%) were possibly willing; and 2 (6%) were unwilling to participate. The most frequent concern was the investigational hemoglobin threshold of 110 g/L (n=10) and several respondents favored a lower threshold (e.g., 90–100 g/L). Feasibility concerns raised by respondents were: 1) the ability to achieve and maintain the high hemoglobin level (n=7, 21%); 2) the potential impact of the trial on blood bank inventory (n= 5, 15%); 3) justification for the high hemoglobin target (n= 5, 15%); and 4) the rarity of major bleeding as the primary outcome. Three respondents (9%) raised concerns about the risk of transfusion reactions, particularly transfusion-associated circulatory overload (TACO) in older patients. Conclusion This survey identified variation and uncertainty in RBC transfusion practice among physicians treating patients with acute leukemia. There is equipoise for an RCT to examine a liberal RBC transfusion strategy on bleeding outcomes. A hemoglobin target of 100 g/L is likely reasonable for the intervention; the impact on blood bank inventory should be anticipated; and the rarity of severe bleeding will require a large sample size.