Progress in the treatment of acute myeloid leukemia (AML) in older patients (pts) is still limited with poor complete remission (CR) rate and overall survival (OS). This is attributed to a variety of reasons including an inherently poor biology, especially a higher incidence of poor-risk karyotypes, co-morbidities, and an age-related functional impairment.

In our randomized AML HD98B trial, the addition of all-trans retinoic acid (ATRA) to intensive chemotherapy resulted in an increased CR rate, event-free (EFS) and OS (Schlenk et al Leukemia 2004). More recent reports on in vitro studies indicated a synergistic action of the histone deacetylase inhibitor valproic acid (VPA) when associated with ATRA plus cytarabine and anthracyclines. In the randomized AMLSG 06-04 trial, therefore, VPA was evaluated in combination with intensive induction therapy plus ATRA in older pts (>60 years) with newly diagnosed AML. In first analyses, the addition of VPA did not provide a significant advantage in OS and EFS after a median follow-up of 47 months (Tassara et al, ASH 2010, abstract #185). This was mainly due to increased hematological toxicity by VPA after the second induction therapy. Here we provide updated analyses especially on survival outcome data based on mature follow-up.


To evaluate VPA in combination with intensive induction therapy and ATRA in older patients with newly diagnosed AML.


Between August 2004 and February 2006 186 patients were randomized (standard-arm, n=93; experimental-arm, n=93) in the AMLSG 06-04 study (ClinicalTrials.gov Identifier: NCT00151255); median age was 68 years (60-84). The first 77 pts were randomized to receive 2 induction cycles (idarubicin 12 mg/m2 i.v. days 1-3, cytarabine 100 mg/m2 cont. i.v. days 1-5, ATRA 45 mg/m2 days 3-5 and 15 mg/m2 days 6-28) with or without VPA (days 1-28; started at 400 mg bid and then adapted in order to obtain a serum level of 60-150 mg/l). After an interim analysis the study was amended; for the following 109 patients idarubicin was dose-reduced to day 1 and 3 and VPA only added during the first induction cycle. All patients were intended for consolidation. Molecular diagnostics were performed as previously published (Schlenk et al, Haematologica 2009)


Details of the response rate and toxicity of the induction treatment have already been presented (Tassara et al, ASH 2010, abstract #185). To summarize, CR rates after double induction were in trend higher in the standard-arm (52% vs. 40%; p=0.10), and early death rate higher in the experimental-arm (14% vs. 26%; p=0.06). The main toxicities attributed to VPA were grade 3/4 infections and delayed hematologic recovery (leukocytes, neutrophils and platelets) observed after the second induction cycle. Therapy (i.e. double induction and consolidation) was completed by 37/93 (40%) of patients in the standard arm and 19/93 (20%) in the experimental arm (p=0.01) After a median follow up of 84 months, analysis of the primary endpoint EFS revealed no differences between the two arms (EFS at 5 years, standard arm 2.3%, experimental arm 7.6%; p=0.95); similarly OS was not different (OS at 5 years, standard arm 11.7%, experimental arm 11.4%; p=0.57). However, pts in the experimental arm had a significantly better relapse-free survival (RFS at 5 years, standard arm 6.4%, experimental arm 24.0%, p=0.02). In explorative subset analyses superior RFS (p=0.03) and OS (p=0.03) of CR-patients were observed in AML patients with mutated NPM1 randomized into the experimental arm (RFS at 5 years, standard arm 8%, experimental arm 42%; OS at 5 years, standard arm 37%, experimental arm 52%). In contrast no differences were seen in AML patients with NPM1 wild-type for RFS (p=0.13) and OS (p=0.87) of CR-patients (RFS at 5 years, standard arm 7%, experimental arm 20%; OS at 5 years, standard arm 15%, experimental arm 22%). Due to a low frequency of FLT3-ITD (9/72) in this patient subset meaningful analyses were not possible.


In older patients with AML, the addition of VPA to standard induction treatment was associated with severe hematological toxicity as well as higher rates of infections and did not improve EFS and OS. However, after a long follow-up VPA was associated with a significantly improved RFS, which might be related to the mutated NPM1 genotype.


Schlenk:Celgene: Honoraria, Research Funding; Pfizer: Honoraria, Research Funding; Chugai: Research Funding; Amgen: Research Funding; Novartis: Research Funding; Ambit: Honoraria.

Author notes


Asterisk with author names denotes non-ASH members.