Efficacy Of Azacitidine Versus Low-Dose Cytarabine In Patients With Acute Myeloid Leukemia - A Retrospective Single Center Experience

Azacitidine (AZA) treatment has been shown to be superior to conventional care regimens including low dose cytarabine (LD-Ara-C) in acute myeloid leukemia (AML) patients with low bone marrow (BM) blast counts (20-30%). In contrast, data on efficacy of AZA in patients with blast counts exceeding 30% are scarce. Here we present a retrospective, single center analysis, comparing the efficacy and toxicity of AZA versus LD-Ara-C in AML patients with high BM blast counts (≥30%) prior to treatment.

Twenty-seven patients receiving AZA and 38 patients receiving LD-Ara-C met the eligibility criteria for the analysis (age ≥18 years, documented BM blast count ≥30% prior to start of treatment and administration of at least one complete therapy cycle). Patients who underwent allogeneic transplantation following AZA treatment or received stem cell support following LD-Ara-C therapy were excluded from this analysis. Overall survival (OS) was estimated using the method of Kaplan and Meier. Comparison of OS between the AZA and LD-Ara-C group was done using the logrank test and by Cox regression adjusting for known confounders.

Patient (age, ECOG status) and diseases characteristics (type of AML, cytogenetics, pretreatment, number of treatment cycles) did not differ significantly between the treatment groups, except for BM blast count (median 44% vs. 60% in the AZA and LD-Ara-C group, respectively; p=0.03) and peripheral blood blast count (median 6% vs. 56% in the AZA and LD-Ara-C group, respectively; p<0.01). Response rates to AZA treatment according to international working group (IWG) criteria were low with two patients achieving a complete remission (CR) and one patient showing partial remission (PR) after AZA treatment. In the LD-Ara-C cohort no CR was observed and two patients experienced a PR. Hematologic improvement (HI) rates according to IWG criteria did not differ between both treatment groups (any type of HI 33% vs. 24% in the AZA and LD-Ara-C group, respectively; p=0.41). In both cohorts, most common non-hematologic toxicities (CTCAE grade≥3) included febrile neutropenia, pneumonia and bleedings without significant differences regarding frequencies. As expected, skin involvement was more commonly observed in the AZA group (19% vs. 3%, p=0.04). One year survival rates were only 15% (95% CI 8-22%) and 13% (95% CI 7-19%) in the AZA and LD-Ara-C group, respectively. There was no statistically significant difference between the treatment groups (HR 1.2, p=0.41). Furthermore, there was no difference in hospitalization time (total days spent in hospital during treatment per patient-year of follow-up 29.4 vs. 27.2 in the AZA and LD-Ara-C group, respectively; RR 1.07 95% CI 0.95-1.21, p=0.23). In a multivariate analysis with OS as endpoint adverse cytogenetics (HR 2.24 95% CI 1.17-4.70, p<0.02) were significantly associated with inferior survival, whereas the treatment had no impact (AZA vs. LD-Ara-C HR 1.27 95% CI 0.67-2.40, p=0.46).

In our center, treatment with AZA showed limited efficacy and no superiority to LD-Ara-C treatment in AML patients with BM blasts ≥30%.

Dreger:Riemser Pharma : Consultancy, Honoraria, Research Funding.