Outcomes after bone marrow versus peripheral blood mismatched unrelated donor hematopoietic cell transplantation using posttransplant cyclophosphamide-based GVHD prophylaxis

Abstract BACKGROUND: Posttransplant cyclophosphamide (PTCy) is increasingly utilized as graft-versus-host disease (GVHD) prophylaxis in allogeneic hematopoietic cell transplantation (allo-HCT), including in the mismatched unrelated donor (MMUD) setting. While bone marrow (BM) has traditionally been favored in this context, peripheral blood stem cells (PBSC) remain a widely used alternative. This study aimed to evaluate post-transplant outcomes associated with PBSC versus BM grafts in MMUD allo-HCT recipients receiving PTCy-based GVHD prophylaxis. METHODS: We conducted a retrospective multicenter analysis of adult MMUD allo-HCT recipients from 2017 to 2021 using the publicly available P5891 dataset from the Center for International Blood and Marrow Transplant Research (CIBMTR), as published by Shaffer et al. Patients received PTCy-based GVHD prophylaxis. Outcomes included overall survival (OS), disease-free survival (DFS), relapse, non-relapse mortality (NRM), acute and chronic GVHD, and GVHD-free, relapse-free survival (GRFS). Median follow-up was estimated using the reverse Kaplan-Meier method. Univariable and multivariable Cox proportional hazards models were used to assess associations between graft source and outcomes, adjusting for variables with p < 0.2 in univariable analyses or known clinical relevance. Hazard ratios (HRs) and 95% confidence intervals (CIs) were reported. Statistical significance was defined as p < 0.05. RESULTS: A total of 559 adults underwent MMUD allo-HCT with PTCy-based GVHD prophylaxis; 466 (83.4%) received PBSC grafts and 93 (16.6%) received BM. Median recipient age was 58 years (IQR 46–66), and 55.8% were female. Nearly half had Karnofsky performance status <90 (45.8%) and high comorbidity burden (HCT-CI ≥3 in 56.4%). The primary diagnoses included acute myeloid leukemia (AML, 54.0%), myelodysplastic syndromes (MDS, 24.7%), and acute lymphoblastic leukemia (ALL, 21.3%), with most patients in first complete remission (55.6%) at the time of transplant. Conditioning intensity was myeloablative in 40.6%, reduced-intensity in 42.9%, and non-myeloablative in 16.5%. CMV mismatch was present in 46.3%, and median donor age was 29 years (IQR 24–36). At a median follow-up of 35.5 months, the median GRFS was 17.0 months (95% CI, 10.7–26.5). In univariable analysis, BM grafts were associated with lower NRM compared to PBSC (HR 0.49; p = 0.040); however, this association was not significant in multivariable analysis (HR 0.64; p = 0.241). Graft source had no significant association with GRFS (HR 0.99; p = 0.951), DFS (HR 1.01; p = 0.976), OS (HR 0.98; p = 0.931), relapse (HR 1.16; p = 0.549), or GVHD outcomes, including grade II–IV acute GVHD (HR 1.08; p = 0.705), grade III–IV acute GVHD (HR 0.49; p = 0.178), and moderate/severe chronic GVHD (HR 0.82; p = 0.530). Several clinical factors were significantly associated with outcomes. MDS was linked to inferior GRFS (HR 1.49; p = 0.005), DFS (HR 1.39; p = 0.041), OS (HR 1.40; p = 0.044), and higher NRM (HR 2.27; p = 0.001) compared to AML. Karnofsky score <90 predicted worse GRFS (HR 1.38; p = 0.008) and OS (HR 1.46; p = 0.009), with a trend toward inferior DFS (HR 1.26; p = 0.077). HCT-CI ≥3 was associated with worse GRFS (HR, 1.70; p = 0.009) and a higher risk of relapse (HR, 1.85; p = 0.046). Age ≥70 was associated with significantly worse DFS (HR 2.20; p = 0.036), OS (HR 2.41; p = 0.027), and a markedly higher NRM (HR 5.27; p = 0.020). CMV mismatch (+/− vs. +/+) was linked to inferior DFS (HR 1.53; p = 0.024) and higher relapse risk (HR 1.85; p = 0.017). Compared to AML, ALL was associated with a lower relapse risk (HR, 0.48; p = 0.014), but a significantly higher NRM (HR, 3.19; p < 0.0001). Donor aged 25–29 (vs. 18–24) was associated with a higher risk of moderate/severe chronic GVHD (HR 2.04; p = 0.026). Non-myeloablative conditioning was associated with a reduced incidence of grade II–IV acute GVHD (HR 0.52; p = 0.015), without significantly affecting relapse or survival.CONCLUSIONS: In mismatched unrelated donor allo-HCT with PTCy-based GVHD prophylaxis, graft source was not significantly associated with survival, relapse, NRM, or GVHD outcomes after adjusting for clinical variables. Instead, patient age, comorbidity burden, performance status, disease type, and CMV mismatch were the primary determinants of outcome. These findings support the use of PBSC grafts when BM is not feasible and underscore the need for individualized transplant planning.