Our largest dataset comprehensively investigated the effect of donor characteristics on outcomes after TCR-haplo PBSCT with PTCy or low-dose ATG. We found that a low CD34+ cell dose was associated with poor OS after PTCy-haplo, whereas donor age, CMV-positive donors for CMV-negative recipients, and a larger number of HLA antigen mismatches were associated with poor OS after ATG-haplo. We provide foundational data for optimal donor choices in TCR-haplo PBSCT.
Several conflicting results have been reported regarding the effects of donor age after PTCy-haplo or ATG-haplo HSCT. Consistent with previous reports, we observed no impact of donor age on OS after PTCy-haplo,32–34 whereas it had a negative effect on OS after ATG-haplo. When focusing on older recipients, some reports have suggested that increased donor age is associated with inferior OS.35,36 Therefore, we checked the impact of donor age in spline terms in a Cox model using recipient age as one of the covariates. Increased donor age was not associated with the risk of death across all donor ages after PTCy-haplo, whereas there was a linear correlation after ATG-haplo. Our findings may differ from those reported by DeZern et al. who demonstrated that increasing donor age by decade was associated with poorer OS regardless of recipient age after PTCy-haplo.37 The discrepancy may be attributed to the different donor sources (only PBSCs in our study) and disease entities (AML over B-cell lymphoma in our study) between our study and the previous report. However, consistent with previous findings,38,39 increased donor age was a risk factor for acute GVHD after PTCy-haplo and ATG-haplo. Therefore, we recommend young donors over older donors not only for ATG-haplo PBSCT but also for PTCy-haplo PBSCT.
Different cut-off values for the number of CD34+ cells and the limited number of patients in the cohort made it difficult to determine the effect of CD34+ cell dose on transplant outcomes after TCR-haplo PBSCT. Consistent with previous studies, a low CD34 cell dose resulted in significantly worse OS and RFS after PTCy-haplo40–42 but not after ATG-haplo. Concerns about the risk of relapse and worse neutrophil engraftment after PTCy-haplo compared with ATG-haplo may have influenced the results.8,19,40 We also demonstrated an increased risk of grade III–IV acute GVHD and chronic GVHD in the high CD34+ cell dose group only after ATG-haplo. Because our previous studies suggested that ATG-haplo is more disadvantageous than PTCy-haplo in terms of suppressing severe GVHD,22,43,44 the benefit of an increased cell dose may only be confirmed after PTCy-haplo. Infusion of sufficient CD34+ cells (> 4 ×106/kg) after PTCy-haplo and not too many CD34+ cells (< 7 ×106/kg) after ATG-haplo was recommended in this study.
The effect of CMV serostatus after TCR haplo PBSCT has not yet been fully elucidated. For CMV-positive recipients, some reports suggested that CMV-negative donors increased the risk of CMV reactivation and worsened OS compared to CMV-positive donors.45 Therefore, the 2017 European Conference on Infections in Leukemia recommended a moderate level of CMV-positive donors for CMV-positive recipients after unrelated donor transplantation with MAC.46 In fact, CMV-positive donors showed favorable GRFS after ATG-haplo in our study. In contrast, CMV-positive donors did not affect OS, RFS, or GRFS after PTCy-haplo PBSCT, which was consistent with that observed in previous studies.47,48 This may be due to the fact that CMV-negative donors were associated with an increased risk of NRM, but significantly reduced the risk of relapse after PTCy-haplo PBSCT in this study. For CMV-negative recipients, donor CMV positivity is a known risk factor for CMV reactivation and mortality.49,50 This study confirmed that having a CMV-positive donor for a CMV-negative recipient after ATG-haplo is associated with worse GRFS. In contrast, CMV positivity did not affect the outcomes of CMV-negative recipients after PTCy-haplo. Different immunological activities of PTCy and low-dose ATG against CMV reactivation may have affected these results.
Our data also highlighted the impact of donor-recipient kinship on transplant outcomes. We applied IPTW analysis to reduce the effects of various confounding factors and avoid overcorrection in the multivariate analysis. We found that offsprings showed better RFS and GRFS than their siblings after ATG-haplo PBSCT in recipients over 40 years of age. As offspring are younger than siblings in most cases, this result may reflect the effect of donor age on OS after ATG-haplo PBSCT. In recipients under 40 years of age, there was no significant difference between parent and sibling donors, which contrasts with previous reports indicating that parent donors were associated with inferior outcomes compared to sibling donors after PTCy-haplo.34,37. This discrepancy may stem from previous reports that examined the impact of donor-recipient kinship in PTCy-haplo HSCT using bone marrow as the donor source (62% in Mariotti J et al. and 80% in DeZern AE et al.), where CD34+ cell dose was not included as a covariate in the multivariate analysis. Matching for NIMA correlates with a reduced incidence of GVHD after non-TCR-haplo HSCT.51–53 However, it has not been clarified whether this immunological tolerance is associated with better outcomes in TCR-haplo PBSCT using PTCy or low-dose ATG. In this study, we observed no significant differences between maternal and paternal donors after PTCy-haplo and ATG-haplo, although our study, with a small number of parent donors, may not have sufficient power to demonstrate a statistical difference. Further analysis in this transplant setting, including a comparison between NIMA mismatch siblings and non-inherited paternal antigen mismatch siblings, is warranted.
Finally, we demonstrated comparable outcomes between FD and NFD donors using PS matching. Elmariah et al. first demonstrated the acceptable safety outcomes after PTCy-haplo from NFD using BM with RIC regimens.54 Another study showed comparable results between FD and NFD donors in haplo-HSCT using a low-dose ATLG/G-CSF-mobilized PBSCs.55 A recent largest study from Global Committee and the Acute Leukemia Working Party of the European Society for Blood and Marrow Transplantation also showed similar outcomes following FD and NFD haplo-HSCT using ex-vivo T-cell depletion, ATG, PTCy, or the combination of ATG/PTCy for patients with AML or ALL in CR (n = 123).56 Our results from JDCHCT are consistent with these previous reports and extend previous findings in that our cohort includes different recipient entities including B-cell lymphoma or high-risk diseases, although larger studies are warranted.
In conclusion, using the largest dataset from the Japanese population, we comprehensively identified the donor characteristics that affected transplant outcomes after TCR-haplo PBSCT using PTCy or low-dose ATG. Although detailed HLA typing, killer cell immunoglobulin-like receptor typing33, and DSA57,58 should be considered, our data suggest the following; after PTCy-haplo PBSCT: 1) sufficient CD34+ cell numbers (> 4 ×106/kg), 2) younger donors over older donors; after ATG-haplo PBSCT: 1) younger donors over older donors, 2) less HLA mismatch donors, 3) not too many CD34+ cell numbers (< 7 ×106/kg), 4) offspring donors preferred over sibling donor, 5) CMV-negative donors for CMV-negative recipients, and 6) CMV-positive donors for CMV-positive recipients. This study provides comprehensive donor selection strategies according to GVHD prophylaxis using TCR-haplo PBSCT.