Long-Term Outcomes in Patients with Acute Myelogenous Leukemia (AML) Following Allogeneic Stem Cell Transplantation (AlloSCT) for Relapse After Initial Autologous Bone Marrow Transplant (ASCT).

Patients with low- and intermediate-risk AML have several options for consolidation therapy, including chemotherapy alone and ACST or AlloSCT. Since randomized studies comparing these approaches show no option to be clearly superior to the others, several centers focus on sequencing of therapies in terms of patient tolerance and toxicity. Since 1986 our preferred consolidation regimen at UCSF for these patients consists of high-dose chemotherapy with ASCT. Consequently, at the time of relapse, we are often presented with the need for a second transplant (usually AlloSCT), in the setting of prior ASCT. As there is a paucity of data describing the toxicity and efficacy of AlloSCT in this setting, we designed this study to evaluate these parameters.

This is a retrospective-cohort, single-institution study of patients with AML treated at UCSF between 1986 and 2008 and who received a second transplant at the time of relapse. Patients were identified through our prospective database, and data were collected from electronic medical records and primary clinical charts when available. Dates of death were corroborated with the social security database. Statistical analysis was performed using STATA (v9). Censoring date for all analyses was July 31st, 2009. This study was approved by the UCSF institutional review board.

Thirty-one patients with AML were identified who underwent an AlloSCT in the setting of relapse following a prior ASCT, with a median follow up of 54 months. The median time from the first to the second transplant was 17 months (range: 6 – 48). The median age at the time of the AlloSCT was 43 years (range: 20 – 64). Response status at the time of transplant was: 17 patients (55%) in complete remission (CR), 12 patients (39%) with less than CR, and 2 patients (6%) with response status unknown. The transplant was myeloablative in 14 patients (45%), non-myeloablative in 13 patients (42%), and cord-blood-based in 4 patients (13%). The donor was unrelated in 17 patients (55%). There were 7 (18%) treatment-related deaths in the first 6 months and 10 (32%) treatment-related deaths overall (pulmonary toxicity: 4, graft-versus-host-disease: 3, and infection: 3). Twelve patients (39%) died due to progressive disease. The median overall survival was 7 months, with 48% and 31% of patients being alive at 12 and 24 months respectively. Remission status prior to transplant was the most significant predictor of survival; the median survival among complete responders vs. all others was 19 months vs. 3.5 months respectively (p=0.006). Following transplantation, the median relapse-free survival among responders was 38 months, with 31% of patients being relapse-free at last follow-up. The time between transplants, as well as age, sex, and intensity of conditioning regimen had no effect on overall and relapse-free survival in our cohort. However, peripheral stem cell transplant was associated with improved overall survival (compared to bone marrow, p=0.02), though this likely reflects different eras of supportive therapy.

Our results suggest that AlloSCT is an effective approach for patients with AML with relapsed or refractory disease after a prior ASCT. In this cohort spanning 22 years, treatment-related mortality from AlloSCT was 32%, and 31% of responders remain free of relapse with long-term follow-up.

No relevant conflicts of interest to declare.