Effect of Stem Cell Source From Unrelated Donors on Transplant Outcomes In Severe Aplastic Anemia (SAA): a Comparison of Unrelated Bone Marrow (BM) and Peripheral Blood Progenitor Cells (PBPC)

Abstract 531

Unrelated donor HSCT has been considered a therapy of last choice in SAA, because of poor overall survival reported prior to 1998 (32% @ 5-years). This is due in large measure to low resolution donor-recipient HLA typing and the selection of patients. Survival after unrelated donor BM HSCT for SAA has improved in recent years with survival rates of approximately 75% @ 5-years when the donor and recipient are matched at HLA-A, -B, -C and –DRB1. Furthermore, improvements in conditioning regimen may also have reduced the rate of graft rejection. In recent years PBPC instead of BM are increasingly used and PBPC grafts now account for 25% of unrelated donor HSCT for SAA. An earlier report from the CIBMTR and the EBMT identified higher chronic graft-versus-host disease (GVHD) incidence and mortality after transplantation of PBPC from HLA-matched siblings compared to BM in patients with SAA aged ≤20 years. GVHD rates are higher after unrelated adult donor HSCT regardless of graft type, but the effects of PBPC on survival in patients with SAA are not known. Two hundred and forty nine patients transplanted in 2000–2007 received either BM (n=186) or PBPC (n=63) from unrelated adult donors matched at HLA-A, -B, -C, -DRB1. Median follow up was 3 years. Compared with recipients of BM, PBPC recipients were older (median age: 19 vs. 35 years), transplanted in 2006–2007 (37% vs. 46%), more likely to be male (47% vs. 65%), receive non-TBI containing transplant conditioning (fludarabine + cyclophosphamide or busulfan or melphalan ± anti-thymocyte globulin; 34% vs. 55%) and tacrolimus-containing GVHD prophylaxis (29% vs. 57%). Among patients who received TBI-containing regimens, approximately 90% of BM and PB recipients received TBI 200 cGy. There were no differences in patient performance score at HSCT, immunosuppressive treatment prior to HSCT and interval from diagnosis to HSCT (median time to HSCT was 13 months in BM recipients and 11 months, in PBPC recipients). As shown after HLA-matched sibling transplantation, the cumulative incidence of neutrophil recovery (³0.5 × 10⁹/L at day-28) and platelet recovery (≥20 × 10⁹/L) was similar in recipients of BM (92% and 83%, respectively) and PBPC (95% and 90%, respectively). The day-100 probability of grades 2–4 acute GVHD was higher after PBPC than after BM (51% vs. 33%, p=0.01). The 3-year probability of chronic GVHD was higher after PBPC than after BM, in patients older than 20 years at HCT (60% vs. 41%, p=0.001) but not in younger patients (23% vs. 26%, p=NS). In multivariate analysis, the risk of mortality was higher after PBPC than after BM (HR 1.79, p=0.02). Other factors that predicted mortality included poor performance score (HR 1.95, p=0.009) and non-TBI transplant conditioning regimens (HR 1.96, p=0.006). Although patient age (>20 vs. ≤20 years) was significantly associated with chronic GVHD there was no significant effect of age on mortality (HR 1.21, p=0.47). The 3-year probability of overall survival adjusted for performance score and transplant conditioning regimen was 76% for BM and 60% for PBPC recipients (p=0.02). These data suggest that: 1) outcomes after URD HSCT for SAA are now similar to outcomes observed in MRD HSCT, 2) BM is the preferred graft source when considering unrelated donor HCT in patients with SAA and 3) the data also support low dose TBI in the preparative regimen for transplant conditioning.