Abstract 1922

In allogeneic hematopoietic stem cell transplantation (HSCT) for recipients with acute myeloid leukemia (AML), cyclophosphamide (Cy) combined with total body irradiation (TBI) (Cy+TBI) is the most common myeloablative conditioning (MAC) regimen, but busulfan (Bu) in combination with Cy (Bu+Cy) has been an alternative to Cy+TBI since early 1980s. But as oral Bu has a problem of interpatient variation in intestinal absorption, intravenous Bu (ivBu) has been developed and substituted for Bu in conditioning regimens for HSCT. For the last decade, fludarabine (Flu)-based regimens with the addition of cytotoxic agents such as Bu or melphalan (L-PAM) have been developed as reduced-intensity conditioning (RIC) regimens. After the introduction of ivBu, Flu+ivBu has become one of the common RIC regimens. In Japan, ivBu was introduced in 2006 and have been widely used as a part of conditioning regimens. In this nationwide retrospective study, we evaluated the clinical outcomes of allogeneic HSCT for AML, especially focusing on ivBu-based conditioning regimens.

The study population included HSCT recipients reported to the Japan Society for Hematopoietic Cell Transplantation. From this database, we extracted the data of adult patients with AML who received first allogeneic HSCT between 1975 and 2010. There were 9,396 recipients selected according to this criterion. Then, we excluded 345 (3.7%) cases from the study because of missing key variables.

A total of 9,051 recipients were evaluated in this study. Median age at transplant was 43 years (range, 16–82), and 41.8% (n=3,785) were female. Types of transplant included bone marrow transplantation from sibling donor (RBMT) (n=1,978, 21.9%), peripheral blood stem cell transplantation from sibling donor (RPBSCT) (n=1,411, 15.6%), bone marrow transplantation from unrelated donor (UBMT) (n=3,321, 36.7%) and cord blood transplantation from unrelated donor (CBT) (n=1,728, 19.1%). MAC regimens were applied to 80.2% (n=7,259) of recipients and RIC regimens to 19.8% (n=1,792), according to the definitions proposed by the NMDP and the CIBMTR in 2007. These MAC regimens included Bu+Cy-based (12.4% of all MAC regimens), Cy+TBI-based (50.0%) and ivBu+Cy-based (5.6%) regimens. RIC regimens consisted mainly of Flu+Bu-based (27.6% of all RIC regimens), Flu+L-PAM-based (24.1%) and Flu+ivBu-based (19.5%) regimens. Median follow-up of survivors was 1,437 days (range, 26–8,344). In MAC setting, overall survival (OS) of HSCT recipients with ivBu+Cy-based regimens did not show the significant difference between that with Bu+Cy or Cy+TBI-based ones in RBMT (p=0.168), RPBSCT (p=0.236) and UBMT (p=0.604). But in CBT, Cy+TBI was significantly superior to Bu+Cy (p=0.004). Though the cumulative incidences of relapse (RI) were similar among recipients with these three regimens, the cumulative incidence of non-relapse mortality (NRM) with Bu+Cy was significantly higher than with Cy+TBI in CBT (p=0.049). In RIC setting, OS of recipients with Flu+ivBu-based regimens was comparable to that with Flu+Bu or Flu+L-PAM-based ones regardless of the type of transplant. RIs with these three regimens were almost equivalent, but NRM with Flu+ivBu-based was significantly lower than that with Flu+L-PAM-based in UBMT (p=0.023). In the multivariate analysis for OS, ivBu+Cy-based regimens did not have significant impacts regardless of the type of transplant, but Flu+ivBu-based regimen had a significantly favorable impact in RBMT (hazard ratio [HR] 0.48, 95% confidence interval [CI] 0.23–0.99). In the multivariate analysis for NRM, Flu+ivBu-based regimen had a significantly reduced risk compared with Flu+L-PAM in RBMT (HR 0.32, 95%CI 0.11–0.95) and UBMT (HR 0.46, 95%CI 0.25–0.83).

These data indicates that ivBu+Cy-based and Cy+TBI-based MAC regimens have almost equivalent efficacy profiles for OS, RI and NRM, and Flu+ivBu-based RIC regimens can reduce the risk of NRM compared with Flu+Bu and Flu+L-PAM-based ones in allogeneic HSCT for recipients with AML.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.