Unrelated Allogeneic Stem Cell Transplantation with Myeloablative Conditioning in Adults with Acute Lymphoblastic Leukemia in Remission: Peripheral Blood or Bone Marrow?.

With the objective to compare outcomes of peripheral blood stem cells (PBSC) and bone marrow (BM) in patients with ALL in remission, we have studied 388 adults with ALL given an unrelated BMT and 337 PBSCT reported to EBMT from 1998 to 2004. The donors were HLA identical (HLA-A and -B low resolution and HLA-DRB1 allelic typing). Median age was 27 years for BM (range 16–63, male 60%) and 30 years for PBSC (range 16–63, male 62%), respectively. Fifty nine and 68% of patients were transplanted in first complete remission (CR1) in the BM and PBSC groups. Female donor to male recipient constellation was 18% in the BM and 15% in the PBSC group. Total Body Irradiation (TBI)-based preparative regimens were given in 85% and 88% of patients in the BM and PBSC group respectively. GvHD-prophylaxis consisted of cyclosporine A (CsA) (6% and 4%), CsA and methotrexate (77% and 58%) or in vivo T-cell depletion (17% and 38% in BM and PBSC group). With the median follow-up of 19 and 13 months, the incidence of GvHD grade II–IV was 37% in the BM group and 39% in the PBSC group (p=0.39). Probabilities of LFS, NRM and relapse were calculated using the Kaplan-Meier estimate. LFS was 45+/−3% vs 36+/−3% (p=0.03) nonrelapse related mortality (NRM) was 30+/−3% and 34+/−4 (p=0.39) while incidence of relapse (RI) was 24+/−4% and 30+/−5% (p=0.13), in the BM and PBSC groups respectively. Three years LFS for the patients transplanted in CR1 was 48% (BM) and 42% (PBSC, p=0.48). In a multivariate analysis adjusting for differences in patients with CR1 in both groups showed not statistically difference for RI, NRM and LFS. However, 3-years LFS for patients transplanted in CR2 was significantly higher in the BM (n=159) compared to the PBSC group (n=106) (40% and 27%, p<0.04). The RI was significantly lower in BM compared to PBSC (33% vs 54%, p<0.01) without there being a detectable difference in NRM at 3 years (40% vs 40%, p= 0.59). This finding led us to look in more details CR2 group of ALL patients. There was no statistically difference between patients receiving BM or PBSC regarding the incidence of Ph’+ or t(4;11) (20% vs 37% in BM and PBSC group), B-lineage (80% vs 76%, respectively) and T-lineage ALL (20% vs 24%, respectively), but patients were older in the PBSC group at the time of transplant (24y vs 27y, p= 0.02) and they received more in vivo T-cell depleted transplants by using antithymocyte globulin (p=0.02). The interval from diagnosis to transplant was longer in BM group (977d vs 687 d in BM and PBSC group, p=0.02). Three years LFS with an interval from diagnosis to Tx > 30 months was not different in BM and PBSC group (48% vs 42%, p=0.65). However, there was a significant difference in LFS in BM group (32%) in comparison to PBSC group (17%) if an interval from diagnosis to Tx <30 mo (p=0.001). The incidence of chronic GvHD in adult ALL in CR2 was 53% and 40% in PBSC and BM, respectively (p=0.09). After statistical adjustments for differences between the two groups, the interval from diagnosis to Tx <30 mo was a risk factor for LFS. In addition, TBI-based conditioning and the interval from diagnosis to Tx <30 mo were risk factors for RI. In conclusion, UD-SCT with either BM or PBSC as a graft source is an effective treatment for adult ALL patients in CR1 and CR2 when the interval between diagnosis and transplant is superior to 30 months. In patients transplanted in CR2 with an interval from diagnosis to Tx<30 months outcomes are worse independent of the cell source used.