Allogeneic Stem Cell Transplantation for Acquired Aplastic Anemia: Better Outcome with Bone Marrow as Compared to Peripheral Blood in HLA-Matched Sibling Donor Transplantation and Improved Outcome Over Time After Matched Unrelated Donor Transplantation. A Retrospective Analysis of Transplants Reported to the German Registry for Stem Cell Transplantation (DRST).

Transplantation of bone marrow (BM) from a HLA-matched sibling donor is an effective treatment for severe aplastic anemia (AA) with long-term survival in excess of 80%. In the recent years there were two trends in allogeneic stem cell transplantation (SCT) for AA: (1) increasing proportion of transplants performed from matched unrelated donors (MUD) and (2) increasing proportion of transplants using peripheral blood progenitor cells (PBSC) as stem cell source instead of BM. A similar switch to PBSC over BM grafts is reported in leukemia transplants. PBSC grafts for leukemia are associated with higher rates of chronic graft-versus-host disease (cGVHD). This adverse consequence may be offset by lower rates of leukemia relapse in some settings. In contrast, there is no perceived benefit of cGVHD for AA. A recent retrospective analysis of EBMT/CIBMTR reported worse outcome after PBSCT compared to BMT in young patients after HLA-matched sibling (sib) transplantation. The impact of stem cell source on outcome after MUD transplants has not been studied in detail so far.

Some reports on favourable results after MUD SCT prompted the discussion whether MUD SCT should be performed earlier in the treatment algorithm in AA instead of considering it as salvage treatment after failure of immunosuppressive treatment. Therefore we performed a retrospective study on the dataset of the German Registry of Stem Cell Transplantations (DRST) to (1) analyze outcome after MUD compared to sib SCT and (2) to study impact of stem cell source (PBSC vs BM) on both sibling and MUD SCT and (3) to analyze impact of transplant period (1998-2002 vs. 2003-2008).

We analyzed 182 sib SCT and 114 MUD SCT (first transplants only). The interval between diagnosis and transplant was significantly longer for MUD SCT as compared to sib SCT (median 98.5 days vs. 511.5 days; p<0.001 ) suggesting that the majority of sib SCT were performed upfront whereas MUD SCT in the majority of cases was performed after failure of other treatments. Median age was 28.5 years (sib SCT) and 30 years (MUD) years (p=0.41). PBSC were used as stem cell source in 50.5% of sib SCT and 61.4% of MUD (p=0.097). 5-year probability of survival was 84.6% (95%-CI: 79.3-90.3%) after sib SCT and 70.1% (95%-CI. 61.8-79.6%) after MUD (p<0.003). In univariate comparison age at transplant has significant impact on survival: After sib SCT 5-year probability of survival in patients ≤ 30 years vs. > 30 years was 94.5% (95%-CI: 90.0-99.3%) vs. 73.5% (95%-CI: 64.3-84.1%)(p<0.001) . 5-year probability of survival after MUD SCT in patients ≤ 30 years vs. > 30 years was 77.7% (95%-CI: 67.3-89.7%) and 60.6 (95%-CI: 48.2-76.2%) (p=0.044). In the most recent period (2003-2008) 2-year probability of survival was 81.6% (95%-CI: 72.9-91.3%) after sib SCT (n=80) and 75.6 (66.1 – 86.5%) after MUD SCT (n=73) (p=0.34). After sib SCT survival was significantly better with BM as compared to PBSCT (5-yr. prob. 95.3%; 95%-CI: 90.9-99.9%; n=89 vs. 74.1%, 95%-CI: 65.1-84.2%; n=92; p<0.001) and cumulative incidence of chronic GvHD was significantly higher with PBSCT as compared to BM (47.0% vs. 18.4%; p<0.01). Cumulative incidence of acute GvHD II-IV did not differ significantly between BM and PBSC. In contrast, stem cells source did neither significantly affect overall survival nor cumulative incidence of acute or chronic GvHD after MUD. In multivariate analysis of the sib SCT older age (>30 years) and use of PBSC were significant risk factors for mortality (Hazard Ratio (HR) 3.4 (1.2-9.3) and HR 4.0 (1.3-12.2). Other variables in the model (conditioning regimen; GvHD prophylaxis; sex match; time diagnosis to transplant and transplant period) were not significant. For MUD SCT none of these variables were significant in a multivariate model.

These results indicate that BM grafts should be preferred to PBSC in patients undergoing HLA-identical sib SCT for AA. In contrast, no negative impact of stem cell source on outcome after MUD SCT could be demonstrated. Results of MUD SCT substantially improved over time. In the most recent period from 2003- 2008 the probability of survival after MUD and HLA-identical sib SCT was no longer different. So far majority of MUD transplants were performed late after diagnosis, mostly after failure of immunosuppression. Re-assessment of both the indication of MUD SCT for AA and the timing of MUD SCT is warranted.

Supported by the Deutsche José Carreras Leukämie-Stiftung.

No relevant conflicts of interest to declare.