Revisiting the Universal Donor: Does Exposure to O Blood Products Affect Patient Outcomes?

Overview

In a recent analysis of a large transfusion database (Transfusion Research Utilization, Surveillance and Tracking database \[TRUST\]), the investigators found that the transfusion of ABO non-identical RBCs to group A individual was associated with an increased risk of death in-hospital compared to patients transfused with ABO identical RBCs (Red Blood Cells). Our finding was corroborated in a separate study of low birth weight neonates who received only group O RBCs (e.g., group O neonates received ABO identical RBCs but group A, B, and AB neonates received ABO non-identical RBCs). A subgroup of neonates who received ABO non-identical transfusions had higher mortality (Z. Sohl, personal communication, April 30th, 2020). Similar adverse clinical outcomes have been reported in a number of studies where patients have received ABO non-identical RBCs and/or platelets. Together, these findings raise the concern that the longstanding policy of transfusing group O non-identical RBCs and platelets may increase the risk of harm for some patients. In Hamilton, Ontario hospitals, approximately 20% of transfused patients receive ABO non-identical RBCs every year because of inventory shortages, urgent requests, and specific phenotype requirements. The negative impact of this practice could have widespread national and international implications for transfusion policy. The ability to undertake critical exploratory analyses in transfusion medicine is enabled by large research and administrative data sets that include all Hamilton hospitals. The initial finding of potential harm with ABO non-identical RBCs is hypothesis-generating and requires confirmation through external datasets and translational studies to support a biological mechanism. If confirmed, this hypothesis can then be tested in a clinical trial.

Why ABO non-identical RBCs may cause harm: In addition to the two exploratory studies published by our group (Z. Sohl, personal communication, April 30th, 2020), other investigators have also suggested that ABO non-identical transfusions could be harmful. A retrospective study by Heal et al., evaluated the effects of a policy change to provide only ABO identical RBCs and platelets for patients undergoing stem cell transplantation and patients receiving treatment for hematological malignancies. A historical control group was used for comparison. The ABO identical policy resulted in less bleeding (5% vs. 15-20%) and improved survival. Other retrospective studies have also shown ABO non-identical platelet transfusions to be associated with an increased risk of platelet refractoriness and that refractory patients had circulating immune complexes for several days. Post transfusion platelet count increments were also higher when ABO identical platelets were transfused. In a retrospective study of 153 patients undergoing primary coronary artery bypass graft or coronary valve replacement surgery, the transfusion of at least one ABO non-identical platelet pool was associated with an increased hospital stay, more days with fever, and more RBC transfusions. Other outcomes (mortality in hospital, length of stay in the intensive care unit, antibiotic days, and a total number of platelet transfusions) were not statistically different. A subgroup analysis (n=139) of patients who received at least two platelet pools showed a non-significant trend towards increased morbidity and mortality (8.6% vs. 1.9%; p=0.10) in recipients of ABO-matched platelets. A retrospective study by Lapierre et al. analyzed data from 186 consecutive children with neuroblastoma or brain tumors who were treated with high-dose chemotherapy followed by hematopoietic stem cell transplantation. The primary endpoint was hepatic veno-occlusive disease. In their multivariate analysis, two factors significantly increased the risk of this outcome: transfusion of platelet concentrates containing ABO-incompatible plasma and use of melphalan in the conditioning regimen. They concluded that transfusion of platelet concentrates containing ABO-incompatible plasma increases the risk of hepatic veno-occlusive disease and hypothesized that passive antibody binding to A and/or B antigens expressed on the surface of hepatic endothelial cells could be involved in the pathophysiology. It is important to emphasize that a publication bias probably exists in this literature with primarily positive studies being reported, and most of the studies are observational (lower quality evidence). Many of the platelet studies also included both minor incompatibilities (plasma in the platelet product has ABO antibodies that react with the recipient's RBCs) and major incompatibilities (recipient's plasma has ABO antibodies that react with ABO antigens on the transfused platelets); however, this literature combined with our preliminary exploratory analyses (Z. Sohl, personal communication, April 30th, 2020) raises the hypothesis that ABO non-identical transfusions (whether minor or major) could impact patient outcomes and should be further explored. Possible biological mechanism: The concept of transfusion-related immune modulation (TRIM) was defined over 30 years ago. More recent evidence suggests that biological mechanisms leading to TRIM can be the heterogeneous involving donor, product, and/or patient factors that contribute to patient morbidities and mortality. A conceptual framework for two possible mechanisms that could lead to harm post-transfusion are a proinflammatory pathway and an immunosuppression pathway. For both pathways, inflammation is one of the prime targets that contribute to the adverse events seen in recipients. In this study, the investigators will use biomarkers of inflammation to determine if the differences are seen between patients who receive ABO identical RBC or platelet transfusions compared to those receiving ABO non-identical blood products. The investigators hypothesize that passive anti-A and anti-B (from group O donors) can bind to recipients' endothelial cells or soluble antigen causing circulating immune complexes that can signal cytokine generation and release causing a "cytokine storm". The severity of the storm may be tempered or enhanced by the secretor status of the recipient and possibly the donor, the titre of the passive antibody transfused, and Group A or AB recipients' subgroup status (A1/A2). Antibody incompatibility could also lead to small amounts of hemolysis, which could trigger an inflammatory response. The biomarkers frequently used to detect inflammation include interleukin-6 (IL-6); tumor necrosis factor-alpha (TNF-α); interleukin-8 (IL-8) and interleukin-1 beta (IL-1β); CD40 Ligand and, C-reactive protein (CRP). Markers of hemolysis in patients post-transfusion include bilirubin, haptoglobin, and lactate dehydrogenase. The investigators will also measure circulating immune complexes as these have been linked to inflammation and a serological profile of the donor/product and the patient will also be performed. A complete list of biomarkers is included below. summary of testing to be performed at various time points during each transfusion episode: Patient Tests: 1. C-Reactive Protein 2. Circulating Immune Complexes 3. IL-6 4. IL-1β 5. TNF-α 6. IL-8 7. CD40 Ligand 8. Complete Blood Count 9. Bilirubin 10. Lactose Dehydrogenase 11. Haptoglobin 12. Anti-A titre (for A and AB group) 13. Anti-B titre (for B and AB group) 14. A1 Phenotyping 15. Lewis Phenotyping Product Tests: 1. Anti-A titre (for A and AB group) 2. Anti-B titre (for B and AB group) 3. A1 Phenotyping 4. Lewis Phenotyping Patients tests will be done at baseline (right before the transfusion starts), 1 hour after transfusion, and 12-24 hours after transfusion Product tests will be done one time on each RBC unit that will be used in the transfusion episodes.