Dose and Time-Dependent Association of Red Blood Cell Transfusion with Increased Risk of GvHD and Worse Survival after Allogeneic HSCT

Background: Acute graft-versus-host disease (aGvHD) remains a critical barrier to the success of allogeneic hematopoietic stem cell transplantation (allo-HSCT). Nearly all patients who undergo allo-HSCT receive red blood cell (RBC) transfusions during the first 30 days post-transplant, but the long-term effects of transfusions on post-transplant outcomes remain unclear. Preclinical and clinical studies indicate that RBC transfusions can activate dendritic cells (DCs) sensitizing transplant recipients to minor histocompatibility antigens (miHA). Initial interactions between donor T cell and DC regulate donor T cell activation and proliferation. We hypothesized that RBC transfusion may contribute to aGvHD in allo-HSCT patients by activating DC and stimulating allo-reactive T cells or other inflammatory pathways.

Methods: We conducted a retrospective study of 336 adult allo-HSCT patients at Emory University Hospital from 2007 to 2013. In cases involving multiple transplants, only data relevant to the first allo-HSCT was included. Graft sources were restricted to bone marrow and G-CSF mobilized peripheral blood (excluded cord blood and T cell depleted grafts). 181 patients (54%) were male and 155 (46%) were female. Patients received transplants from matched related donors (n=123, 37%) or matched unrelated donors (n=213, 63%) for treatment of AML (n=132), ALL (n=40), acute leukemia (n=5), MDS (n=41), CML (n=22), CLL (n=15), HD (n=6), NHL (n=33), AA (n=10), MM (n=7) or other diseases (n=25). aGvHD with onset of up to 150 days after allo-HSCT was the primary end-point. Leuko-reduced and irradiated RBC transfusions administered during the week prior to transplant and 30 days post-transplant (while patients were closely monitored at the transplant center) were provided by a single Blood Bank. The median follow up time was 14.6 months post transplantation (range, 0.3 to 65.1 months). The relationship of RBC transfusions to aGvHD was determined as a time-dependent variable or as a function of the total number of RBC units transfused.

Results: 306 patients (91%) received RBC transfusions during the observation timeframe, while 30 (9%) did not require transfusion (median 6). 221 patients (66%) developed grade I-IV aGvHD with a distribution of 85 (25%), 72 (21%), 47 (14%), and 11 (3%) of patients with maximum grade of I, II, III and IV, respectively, while 115 (34%) patients did not show any signs of aGvHD. We compared transfusion history, prior to the diagnosis of GvHD, in patients who developed grade 0-II aGvHD (n=272, 81%) vs. grade III-IV aGvHD (n=64, 19%). Patients with grade III-IV aGvHD received more RBC units (median 8) than patients with grade 0-II aGvHD (median 4). In univariable Cox regression analysis, factors significantly associated with grade III-IV acute GvHD were HLA match (HR 2.22, p=0.004), number of RBC units per week (HR=1.26, p<0.001), cumulative number of RBC units (HR=1.09, p<0.0003), and maximum storage age of units transfused (HR per 5 days older =1.12, p=0.02). In multivariable analysis, HLA match (HR=2.148, p=0.008) and number of RBC units transfused weekly (HR=1.293, p=0.0001) were associated with development of grade III-IV aGvHD. Kaplan-Meier analysis showed significantly worse long-term survival for patients with grade III-IV aGvHD (p value from Log-rank test <0.0001). Increased numbers of transfused RBC units (HR per 2 units =1.17, p<0.0001) was also associated with worse long-term survival in univariable analysis, and patients who received 7 or more RBC units had poorer long-term survival (HR=3.56, p<0.0001). In multivariable analysis, grade III-IV aGvHD was associated with worse long-term survival.

Conclusion: We have demonstrated a significant association of RBC transfusions with the subsequent development of grade III-IV aGvHD for allo-HSCT patients and shown that patients with grade III-IV aGvHD have poorer long-term survival. While RBC transfusions are associated with worse survival in univariable analysis, based on our multivariable analysis, increased RBC transfusions are only associated with poorer survival in the context of severe aGvHD. These data are consistent with preclinical observation that allogeneic RBC can initiate inflammatory reactions to miHA. The data suggest that new strategies to improve RBC transfusion methods for allo-HSCT patients are needed to help reduce the risk of severe aGvHD and improve survival rates.