Higher Graft Failure after Haploidentical in Comparison to Matched Related and Unrelated Stem Cell Transplantation for Severe Aplastic Anemia: A Single-Center Retrospective Study

Introduction: Hematopoietic stem cell transplantation (HSCT) with haploidentical donors using post-transplant cyclophosphamide (PT-Cy) has become an alternative for severe aplastic anemia (SAA) patients without a matched sibling donor (MSD) or an unrelated donor (UD). However, direct comparisons of outcomes among these different donor modalities are emerging and are necessary to guide therapeutic decisions. Objectives: To compare outcomes between MSD plus UD and haploidentical HSCT with PT-Cy in patients with SAA. Patients and Methods: This is a single-center retrospective study. Thirty-two patients with SAA were submitted to HSCT between January 2018 and December 2023. There were 15 patients who underwent MSD (n= 9) and UD HSCT (n= 6), designated as the non-haplo group, and 17 underwent haploidentical HSCT, designated as the haplo group. Donor-specific anti-HLA antibodies (DSA) screening test was performed for all patients with HLA incompatibility. Patients with high DSA levels (mean fluorescence intensity (MFI) >2000) underwent desensitization therapy. Conditioning regimens and graft-versus-host disease (GVHD) prophylaxis were cyclophosphamide 200 mg/Kg plus rabbit ATG (Thymoglobulin) with a calcineurin inhibitor (CNI) plus methotrexate for MSD HSCT and fludarabine, cyclophosphamide plus TBI 400 cGy with post-transplant cyclophosphamide, CNI plus mycophenolate mofetil for UD and haploidentical HSCT. Results: Non-haplo and haplo groups were similar regarding: median age (16 vs. 20 years, p= 0.94), female sex (53.3% vs. 29.4%, p= 0.16), diagnosis of very SAA (46.7% vs. 52.9%, p= 0.72), first-line HSCT (60% vs. 64.7%, p= 0.55), total nucleated cell dose (3.32 vs. 3.96 x 108/Kg, p= 0.25), and use of bone marrow as graft source (100% vs. 82.4%, p= 0.08). Patients in the haplo group had higher pre-HSCT ferritin levels (2098 vs. 1289 ng/mL). Graft failure occurred only in the haplo group (23.5% vs. 0%, p= 0.04): 2 primary and 2 secondary graft failures. All patients who rejected had low level of DSA (MFI 500 - 1300). The incidence of grade II-IV acute GVHD at day 100 was similar between the non-haplo and haplo groups: 7.7% vs. 18.6% (p= 0.22), respectively. However, patients in the haplo group had more moderate-to-severe chronic GVHD at day 180: 19.6% vs. 0% (p= 0.05). The 2-year overall survival (OS) was 86.7% vs. 73.7% in the non-haplo and haplo groups, respectively (p= 0.26). The 2-year OS was dismal in patients who had graft failure, 25% vs. 87.1% (p= 0.002). Of the patients in the haplo group, 29.4% underwent salvage HSCT versus 0% in the non-haplo group (p= 0.02). Graft failure was associated with a higher mortality risk in Cox regression analysis (HR 8.85, p= 0.008). No association was found with mortality in Cox regression analysis for the following variables: age (p= 0.65), very SAA (p= 0.26), pre-HSCT ferritin (p= 0.11), second-line HSCT (p=0.37), haploidentical HSCT (p= 0.27), peripheral blood as graft source (p= 0.72), grade II-IV acute GVHD (p= 0.16), and moderate-severe chronic GVHD (p= 0.47). Conclusions: Patients with SAA who underwent haploidentical HSCT with PT-Cy had more chronic GVHD than those who underwent MSD and UD HSCT. Graft failure was more common after haploidentical HSCT despite careful DSA screening and desensitization therapy when needed. Graft failure was associated with an increased risk of mortality after HSCT for SAA. Our results are in line with a recent EBMT retrospective study that also demonstrated more graft failure after haploidentical graft failure compared with UD HSCT for SAA (Montoro et al., Blood 2024; 144 (3): 323-333). Prospective controlled studies are urgently needed to address the role of haploidentical HSCT for SAA.