Selecting between HLA-Matched Siblings and HLA- Haploidentical Related Donors for Acute Leukemia in the Era of Post-Transplant Cyclophosphamide: The Center for International Blood and Marrow Transplant Registry and the Acute Leukemia Working Party of the European Society for Blood and Marrow Transplant

An HLA-matched sibling has been considered the optimal donor for allogeneic hematopoietic cell transplantation (HSCT). However, several potential transplant recipients may have a haploidentical sibling or an offspring who may also serve as donors. In this study, we sought to determine the optimal alternative related donor (haploidentical sibling or an offspring) as compared to an HLA-matched sibling. The primary objective was comparison of chronic graft-versus-host disease (GVHD). Secondary outcomes included acute GVHD, non-relapse mortality (NRM), relapse, treatment failure (relapse or death, inverse of relapse-free survival) and overall mortality. The study population included 4540 donor-recipient pairs (n=218 haploidentical sibling; n=218 offspring; n=4104 HLA-matched sibling) with acute myeloid (n=3617) and acute lymphoblastic (n=923) leukemia transplanted in 2008 to 2015. There were few offspring (n=87) who donated to patients aged 18-54 years and haploidentical siblings (n=61) who donated to patients aged 55-76 years and were excluded from the analysis. Post-transplant cyclophosphamide (PT-Cy) with calcineurin inhibitor (CNI) and mycophenolate was used for GVHD prophylaxis for all haploidentical HSCTs. HLA-matched siblings received CNI-containing GVHD prophylaxis; CNI with methotrexate was the predominant prophylaxis regimen. Patient age was correlated with donor age and donor-recipient relationship. Bone marrow was the predominant graft for haploidentical HSCTs (64%) and peripheral blood (89%) for HLA-matched sibling HSCTs. Younger patients (age 18-54 years) of were more likely to receive reduced intensity conditioning regimens for haploidentical HSCTs compared to HLA-matched sibling HSCTs (22%). Among older patients, conditioning regimen intensity did not differ by donor type. Exploratory analysis confirmed differences in survival by patient age. Therefore, based on differences in survival by patient age, two age groups were created: 18 - 54 years and 55 - 76 years and within each patient age group, donor-recipient relationship and its effect on transplant outcomes were tested using Cox regression models. In addition to the standard Cox regression model, we also performed a matched-pair analysis. Recipients of haploidentical HSCT (cases; n=436) were matched to HLA-matched siblings (controls; n=1707) on age, disease and disease risk index. 88% of patients aged 18-54 years were matched to 4 controls and 96% of patients aged 55-76 years were matched to 4 controls. The median difference in age between cases and controls were 0.08 (range 0-9.9) years for patients aged 18-54 years and 0.05 (0 - 9.5) years for patients aged 55-76 years. The results of multivariate Cox regression and matched-pair analysis are shown in Tables 1, 2. Among patients aged 18-54 years, chronic GVHD risks were lower after haploidentical sibling compared to HLA-matched sibling HSCT. There were no differences in acute GVHD, NRM, relapse, treatment failure or overall mortality. Among patients aged 55-76 years, acute and chronic GVHD risks were lower after offspring compared to HLA-matched sibling HSCT. But risks for NRM, treatment failure and overall mortality were higher after offspring compared to HLA-matched sibling HSCT. The 2-year probabilities of overall survival, adjusted for disease risk index and sex are shown in Figure 1. In summary, chronic GVHD rates were lower after haploidentical HSCT with PT-Cy containing GVHD prophylaxis regimens compared to HLA-matched sibling HSCT for all patients. Whether this can be attributed solely to the GVHD prophylaxis regimen or in part attributed to use of peripheral blood grafts for HLA-matched sibling HSCT must be studied further in a setting in which PT-Cy containing GVHD prophylaxis is used for HLA-matched sibling HSCT. Despite lower chronic GVHD with haploidentical sibling donor HSCT with PT-Cy, NRM and overall survival did not differ by donor type for patients aged 18-54 years. However, for patients aged 55 - 76 years, despite lower chronic GVHD with offspring donor HSCT with PT-Cy, NRM was higher and overall survival was lower compared to HLA-matched sibling HSCT. Definitive proof of these findings would require a prospective randomized trial.

View largeDownload slide

View largeDownload slide

Close modal