Open Access Highly Accessed Research article

Multivariate and subgroup analyses of a randomized, multinational, phase 3 trial of decitabine vs treatment choice of supportive care or cytarabine in older patients with newly diagnosed acute myeloid leukemia and poor- or intermediate-risk cytogenetics

Jiří Mayer1*, Christopher Arthur2, Jacques Delaunay3, Grzegorz Mazur4, Xavier G Thomas59, Agnieszka Wierzbowska6, Farhad Ravandi7, Erhan Berrak8, Mark Jones8, Yuhan Li8 and Hagop M Kantarjian7

Author Affiliations

1 Masaryk University Hospital Brno and Central European Institute of Technology (CEITEC), Masaryk University, Brno, Czech Republic

2 Royal North Shore Hospital, St. Leonards, New South Wales, Australia

3 Nantes University Hospital, Nantes, France

4 Wroclaw Medical University, Wroclaw, Poland

5 Affiliation at time of study: Hôpital Edouard Herriot, Lyon, France

6 Medical University of Lodz, Lodz, Poland

7 The University of Texas MD Anderson Cancer Center, Houston, TX, USA

8 Eisai Inc., Woodcliff Lake, NJ, USA

9 Current address: Hôpital Lyon-Sud, Lyon, France

For all author emails, please log on.

BMC Cancer 2014, 14:69  doi:10.1186/1471-2407-14-69

Published: 6 February 2014



Compared with younger patients, older adults with acute myeloid leukemia (AML) generally have poorer survival outcomes and less benefit from clinical trials. A recent phase 3 trial demonstrated a trend toward improved overall survival (OS) with decitabine, a hypomethylating agent, compared with treatment choice of either cytarabine or supportive care (7.7 months, 95% CI: 6.2–9.2 vs 5.0 months, 95% CI: 4.3–6.3, respectively) in older adults with newly diagnosed AML. The current analyses investigated prognostic factors for outcomes in this trial and examined OS and responses in prespecified subgroups.


A multivariate Cox proportional hazards model was used to investigate effects of demographic and baseline characteristics, including age, sex, cytogenetic risk, AML type, ECOG Performance Status, geographic region, bone marrow blasts, platelets, and white blood cells on OS, based on mature data. Similar analyses were conducted with a logistic regression model to predict response rates. Prespecified subgroup analyses were performed for OS and response rates, also using mature data.


Patient characteristics that appeared to negatively influence OS included more advanced age (hazard ratio [HR] 1.560 for ≥75 vs <70 years; p = 0.0010), poorer performance status at baseline (HR 0.771 for 0 or 1 vs 2; p = 0.0321), poor cytogenetics (HR 0.699 for intermediate vs poor; p = 0.0010), higher bone marrow blast counts (HR 1.355 for >50% vs ≤50%; p = 0.0045), low baseline platelet counts (HR 0.775 for each additional 100 × 109/L; p = 0.0015), and high white blood cell counts (HR 1.256 for each additional 25 × 109/L; p = 0.0151). Regarding geographic regions, patients from Western Europe had the longest median OS. Response rates favored decitabine for all subgroups investigated, including patients ≥75 years (odds ratio 5.94, p = 0.0006).


Response to decitabine in AML is associated with known prognostic factors related to both patient demographics and disease characteristics.

Trial registration identifier NCT00260832

Decitabine; Acute Myelocytic Leukemia; Elderly; Treatment