A Second Course of High-Dose Cytarabine before Autologous Stem Cell Transplantation Does Not Improve Outcome of Young Acute Myeloid Leukemia Patients in First Complete Remission: Results of the BGMT 95 Study.

Background. Once complete remission (CR) is achieved in young (18–60) acute myeloid leukemia (AML) patients (pts), the main question is about the best post remission treatment. For pts not eligible for allogeneic bone marrow transplantation (AlloBMT) high-dose cytarabine-based chemotherapy (HiDAC) followed by autologous stem cell transplantation (ASCT) is an interesting option. However, the number of courses of cytarabine to apply before ASCT is not determined. The aim of this randomized study was to evaluate the impact of a second HiDAC before ASCT.

Pts and methods. Pts received an induction by daunorubicine 60 mg/m2/d 3d + cytarabine 100 mg/m2/d (CI) 10 d. followed by a consolidation based on cytarabine 50 mg/m2/12h SC 7d + daunorubicine 60 mg/m2/d 2d. Pts < 45 with a sibling donor had an alloBMT (cyclophosphamide 60 mg/kg IV 2d + fractionated TBI 12 Gy). The others had a superconsolidation (cytarabine 3 g/m2/12h IV 4d + daunorubicine 45 mg/m2/d) followed by stem cell collection (either BM or PBSC). Then was the randomization between ASCT (busulfan 4 mg/kg/d po 4d + melphalan 140 mg/m2 on d5) immediately (SC1 group) or after 1 more course of HiDAC (cytarabine 3 g/m2/12h IV 4d) (SC2 group).

Results. Between 05.1995 and 02.2001 a total of 437 pts entered the study. The median age of these pts (M/F = 238/199) was 47 years (18–60) with 203 of them under 45. Median WBC count was 12.5 (0.2–335) and WBC was > 30 109/L for 140 pts. The FAB distribution was: M1/M2 = 227; M4/M5 = 151; M0/M6/M7 = 59. The cytogenetics risk groups distribution was: low risk = 56 (13.8%); intermediate risk = 265 (65.4%); high risk = 84 (20.8%). CR was achieved in 2 courses for 46 pts (10.5 %). Out of the 437 initial pts, 351 achieved CR (80 %), 65 were eligible for alloBMT, 277 for randomization, and 128 were randomized (65 for SC1, 63 for SC2). The only difference between randomized and non-randomized pts was for number of course to achieve CR (7 % vs 16.8 % respectively, p = .01). This turn into a difference for cumulative incidence of non relapse deaths (CINRD) (8.7 % vs 20.6 % ; p = .01) but not for cumulative incidence of relapse (CIR). Overall survival (OS), leukemia free survival (LFS) and CIR and CINRD were 57.7, 50.9, 35.5, 19.9 for alloBMT. For SC1/SC2 results were : 41.3/53.7 for OS (p = .10), 39.2/49.6 for LFS (p = .11), 56.7/45.1 for CIR (p = .08), 8.5/8.4 for CINRD (p = .95). Multivariate analysis for LFS showed the independent impact of initial WBC count, cytogenetics and number of course to achieve CR. However, type of treatment (SC1 vs SC2) had no impact on LFS (p = .41). This lead us to conclude that one course of HiDAC was enough before ASCT despite the rather low proportion of randomized pts.