Factors That Contribute to Long-Term Survival In Patients with Leukemia Not In Remission at Allogeneic Hematopoietic Cell Transplantation

Allogeneic hematopoietic cell transplantation (HCT) may even cure leukemia following relapse or primary induction failure. Several pre-transplant variables including age, duration of remission, poor-risk cytogenetics, tumor burden at HCT, type of donor, and performance status reportedly affected the post-HCT prognosis of leukemia that is not in remission. However, there has been insufficient examination of the factors required to achieve long-term survival or cure of leukemia that is not in remission at HCT. We might consider long-term survival without relapse, particularly for more than 5 years, as ‘likely cure' of leukemia. Therefore, we evaluated the factors that contribute to long-term survival (for more than 5 years) in patients with active leukemia at HCT.

We retrospectively performed an analysis of leukemia not in remission at HCT performed at our single institute between January 1999 and July 2009. Forty-two patients aged from 15 to 67 years (median age: 39 years) received intensified myeloablative (n=9), myeloablative (n=11) or reduced-intensity conditioning (n=22) for HCT. Twelve patients received individual chemotherapy for cytoreduction within the three weeks before reduced-intensity conditioning for HCT. Diagnoses included de novo AML (n=17), ALL (n=12), CML-AP (n=2), MDS/AML (n=10) and plasma cell leukemia (n=1). In those with acute leukemia, cytogenetic abnormalities were intermediate (n=17, 44%)or poor (n=22, 56%). Seven patients were primarily refractory to induction chemotherapy. The other patients relapsed after conventional chemotherapy or the first HCT. The median number of blast cells in bone marrow (BM) was 26.0% (range; 0.2–100) before the start of chemotherapy for HCT. Six patients had leukemic involvement of the central nerve system. Stem cell sources were related BM (n=3, 7%), related peripheral blood (n=13, 31%) unrelated BM (n=20, 48%) and unrelated cord blood (CB) (n=6, 14%). Thirty-one pairs were matched for HLA-A, B and DRB1 antigens. Three patients were mismatched for one HLA antigen (two at HLA-A, one at HLA-B), and seven were mismatched for two (two at HLA-A and B, five (all CB) at HLA-B and DRB1). The remaining patient was mismatched for all three antigens. Prophylaxis for acute GVHD consisted of calcineurin alone (n=5), calcineurin combined with short-term methotrexate (n=32), calcineurin combined with mycophenolate mofetil (n=2) or none (n=3). In this study, we defined long-term survival as survival without relapse for more than 5 years.

Engraftment was achieved in 33 (79%) of 42 patients. Median time to engraftment was 17 days (range: 9–32). Five patients died early after HCT (range 4–20 days). Twenty four (65%) of 37 evaluable patients developed acute GVHD (eight grade I, nine grade II, five grade III, two grade IV), and 12 (50%) of 24 evaluable patients developed chronic GVHD (1 limited, 11 extensive). With a median follow up of 85 months for surviving patients, the five-year Kaplan-Meier estimates of leukemia-free survival rate and overall survival (OS) were 17% and 19%, respectively. At five years, the cumulative probability of non-relapse mortality was 38%. In the univariate analyses of impact of pre-transplant variables on OS, poor-risk cytogenetics, number of BM blasts (>26%), MDS/AML and CB as stem cell source were significantly associated with worse prognosis (p=.03, p=.01, p=.02 and p<.001, respectively). In addition, the five-year Kaplan-Meier estimates of OS in patients with and without cGVHD were 66.7% and 0% (p<.001) respectively.

Graft-versus-leukemia effects mediated by cGVHD may have played a crucial role in long-term survival in, or cure of active leukemia. We speculate that effective cytoreduction by individual chemotherapy and/or conditioning for HCT to control disease until cGVHD subsequently occurred might be also important, particularly in leukemia with rapid proliferation. However, intensive conditioning for HCT did not appear to be indispensable in relatively indolent leukemia, even with non-remission status at HCT. In addition, based on our results, CB might be unsuitable as a source of stem cells for leukemia that is active at the time of HCT.

No relevant conflicts of interest to declare.