Abstract
Introduction:
Haploidentical bone marrow transplantation has emerged as the predominant form of allogeneic transplantation in Brazil, driven by lower costs, increased donor availability, and the potential for long-term donor cell reuse in immunotherapy. However, the influence of HLA and non-HLA factors on donor selection remains unclear, particularly within Brazilian genetically admixed population. This study aims to evaluate the impact of HLA and non-HLA variables on outcomes of haploidentical transplants performed at a quaternary care center in Brazil.
Methods:
We retrospectively analyzed 88 adult patients with onco-hematological diseases who underwent T-cell replete haploidentical transplantation with post-transplant cyclophosphamide between 2010 and 2023. High-resolution HLA typing was performed, including manual alignment of HLA-DRB1 residue 86 (Gly/Val). B leader status, HLA-DQ (G1/G2) heterodimers and HLA-DPB1 expression and permissiveness using the TCE version 2 model were assessed using online tools. Non-HLA variables included donor/recipient age, sex, relationship, ABO compatibility, and CMV serostatus. Disease and transplant-related factors included Disease Risk Index (DRI), conditioning regimen, disease status, CD34+ cell count, and graft source. The primary endpoint was overall survival (OS); secondary endpoints included progression-free survival (PFS), non-relapse mortality (NRM), relapse incidence (RI), engraftment, and incidence of acute and chronic graft-versus-host disease (GVHD). OS and PFS were estimated using the Kaplan-Meier method and compared via log-rank test. Cumulative incidence curves were used for NRM, RI, engraftment, and GVHD. Multivariable analysis was performed using a Cox regression model with a stepwise variable selection approach based on the lowest Akaike Information Criterion.
Results:
Median patient age was 50 years (range 36–61); 51% were male; 53% had AML or MDS; 80% had Karnofsky scores ≥90%; and 56% received reduced-intensity conditioning. Donors had a median age of 36 years (range 27–46); 59% were male; and 55% donated bone marrow. CMV seropositivity was present in 72% of pairs; 55% were ABO compatible. Median follow-up was 31 months. Four patients died before engraftment, all of whom had active disease at the time of transplantation. Median neutrophil and platelet engraftment occurred at 18 and 21 days, respectively. Two-year OS, PFS, NRM, and RI were 55%, 47.2%, 25.5%, and 27.3%, respectively. One-year cumulative incidence of grade 2–4 acute GVHD was 31%; two-year chronic GVHD was 5.6%. No significant OS difference was observed between transplants with 5/10 (43 patients) versus >5/10 (45 patients) HLA matches (p=0.9). Most matches occurred at HLA-A (22%), followed by HLA-DQB1 (17%), HLA-C (16%), HLA-DRB1 (10%), and HLA-B (6.8%). Among transplant pairs, 67% had matched B leaders, 52% were HLA-DPB1 permissive, and 41% had favorable expression mismatches. HLA-DQ heterodimers were G1G2 in 56% of recipients and 53% of donors. HLA-DRB1 dimorphism was predominantly GlyVal in recipients (56%) and donors (52%). No significant associations were found between outcomes and HLA-DPB1 permissiveness or expression, B leader compatibility, HLA-DRB1 dimorphism, or HLA-DQ heterodimers. HLA-DRB1 mismatch, B leader match, and HLA-DPB1 non-permissive status were not associated with improved OS (p=0.4) or PFS (p=0.3), even in additive models. In multivariable analysis, only active disease at transplant was significantly associated with inferior OS (HR 3.0; 95% CI, 1.4–6.5), PFS (HR 2.7; 95% CI, 1.3–5.7), and RI (HR 3.1; 95% CI, 1.2–8.0).
Conclusions:
This single-center study confirms the feasibility and favorable outcomes of haploidentical transplantation in patients with onco-hematological diseases in Brazil. However, previously described HLA mismatch models were not predictive in our population. Larger multicenter studies are warranted to refine donor selection algorithms tailored to diverse genetic backgrounds.
