Incidence, predictors and consequences of poor graft function after post-transplant cyclophosphamide (PTCy)-based allogeneic transplantation

Abstract

Abstract Prolonged cytopenias requiring transfusion and growth factor support are a known complication of allogeneic hematopoietic stem cell transplantation (allo HCT). There is heterogeneity in studies characterizing poor graft function (PGF) due to lack of consensus on a clinical definition. The American Society for Transplant and Cellular Therapy (ASTCT) defined PGF as frequent dependence on blood and/or platelet transfusions and/or growth factor support in the absence of other explanations and in the presence of adequate donor myeloid and lymphoid chimerism. While this definition seeks to bring homogeneity in clinical practice, there is still a knowledge gap in terms of predictors for PGF, its impact on transplant outcomes and management of this complication. Furthermore, there are few studies characterizing PGF in the context of PTCy-based GVHD prophylaxis. To this end, we performed a retrospective single institution analysis of 239 consecutive patients with acute leukemia or myelodysplastic syndromes (MDS), undergoing a first PTCy-based allo HCT between 2016 and 2024. We defined PGF as ≥2 cytopenic lineages (ANC <500 or need for growth factor support, platelets <20,000 or requiring platelet transfusions to maintain a higher platelet threshold, or transfusion dependent anemia defined as ≥2 RBC units in 1 month) for at least 2 weeks from day +28 until day +100, in the presence of >95% donor chimerism and in the absence of relapse or severe GVHD. The median age was 58 (range 19-80) with a diagnosis of AML, MDS or ALL in 48%, 29% and 24% respectively. A haploidentical donor was used in 79%, PBSC was the stem cell source in 98%, and myeloablative conditioning was used is 36%. PGF was identified in 115 (48%) patients. Patients who had PGF were more likely to be CMV positive, ABO incompatible, had microangiopathic hemolytic anemia (MAHA), acute GVHD, or had been treated with ruxolitinib or sirolimus. In multivariable analysis (MVA), ABO incompatibility (OR 2.15, p=0.01), use of sirolimus (OR 3.5, p=0.005), presence of MAHA (OR 5.52, p=0.014), or acute GVHD (grade 2, OR 2.0, p=0.029; or grade 3-4, OR 8.24, p<0.001) were predictors for PGF. Patients with or without PGF had comparable transplant outcomes including overall and disease-free survival, as well as chronic GVHD incidence. Unexpectedly, patients with PGF had a significantly lower cumulative incidence of relapse (p=0.012), however this effect was balanced by higher NRM (p=0.001). In MVA, controlled for age, DRI, acute GVHD, CMV reactivation, and year of transplant, PGF retained a significant association with lower relapse risk (HR 0.37, 95% CI 0.21-0.67, p<0.001) and higher NRM (HR of 4.14, 95% CI 1.36-12.6, p=0.013). In conclusion, PGF is a common problem after PTCy-based allo HCT, occurring in almost half of patients. Predictors of PGF include baseline factors such as ABO incompatibility and early post-transplant factors such as use of sirolimus or the development of MAHA or acute GVHD. Although PGF significantly increased the risk of NRM, it had no impact on survival due to a corresponding decrease in relapse incidence. The positive effect of PGF on relapse reduction, which is independent of disease-related factors or the development of CMV reactivation or GVHD, deserves further study to understand the basis of this finding.