Contribution of HLA Incompatibility on Clinical Outcomes of CBT in Patients with MDS Including Secondary AML in Japan: Multivariate Analysis Using Competing Risk Regression Model

[Study purpose] We have reported that overall results of cord blood transplantation (CBT) were inferior to those of bone marrow transplantation (BMT) from unrelated donor in myelodysplastic syndrome (MDS) including transformed acute myelogenous leukemia (AML/MDS) (ASH 2007). Poor engraftment and higher incidence of relapse in CBT patients were serious problems. We now show here the impact of conditioning regimen and GVHD prophylaxis method on clinical outcomes of CBT, and then show the results when we looked for the appropriate graft selection in terms of human leukocyte antigen (HLA) compatibility and cell dose.

[Patients and Methods] Clinical data of 333 patients with MDS including AML/MDS who received unrelated CBT without prior transplant history between 1998 and 2006 in Japan were collected by the Japan Cord Blood Bank Network (JCBBN). The median period of follow-up for survivors (n=148) after transplants was 13 (range, 1–99) months. We analyzed the hematopoietic recovery, incidences of acute and chronic graft-versus-host disease (GVHD), risks of transplant-related mortality (TRM) and relapse, and disease-free survival (DFS) using competing risk regression models.

[Results] Both myeloablative conditioning regimen including 8Gy or more dose of total body irradiation (TBI) and GVHD prophylaxis as calcineurin inhibitor (cyclosporin or taclorimus) plus methotrexate (MTX) significantly (p<0.05) contributed to the lower risks of relapse and TRM and better DFS in univariate analysis. In multivariate analysis, TBI including regimen contributed the better engraftment of platelet (p=0.009) and lower risk of relapse (p=0.011). Calcineurin inhibitor plus MTX method was associated with the lower risks of grades III+IV acute GVHD (p=0.006) and TRM (p=0.002). Next, we analyzed 116 patients who used TBI including regimen and calcineurin inhibitor plus MTX combination and looked for significant graft factors for better clinical outcomes of CBT. In multivariate analysis, one antigen or more of HLA disparity by high resolution typing in either GVHD or rejection direction was significantly associated with lower risk of relapse (p<0.001). On the other hand, there was no event in HLA matched recipients by high resolution typing (n=5) and thus HLA compatibility seemed to be implicated in low risk of TRM. Finally, 2 antigens mismatch by low resolution typing affected better DFS result in multivariate analysis (p=0.01). Those advantages in HLA mismatched cord blood recipients might be induced by allo-reactivity against MDS. In contrast to HLA disparity, cell numbers of nucleated cells, of CD34 positive cells and of CFU-GM at the cryopreservation did not affect to the results in our analysis probably because of the selection bias as bellow: As the cell dose is a top priority of graft selection in most centers and majority (87%) of patients was adult (>15 years old), grafts had relatively high quantity with narrow range, which the median numbers of nucleated cells, CD34 positive cells and CFU-GM were 2.57 (74% patients in 2.0–3.5 dose range) ×10⁷/kg, 0.85 (76% in 0.5–2.0 dose range) ×10⁵/kg and 25.4 (77% in 10–50 dose range) ×10³/kg, respectively.

[Conclusion] Myeloablative conditioning regimen with TBI and calcineurin inhibitors plus MTX for GVHD prophylaxis in CBT might be promising for MDS patients and HLA disparity possibly contributed on lower risk of relapse and better DFS in this setting.