Randomized Phase II Study of Combined Epigenetic Therapy: Decitabine Vs. Decitabine and Valproic Acid in MDS and AML

Introduction: Inhibition of DNA methylation (DNMTi) by decitabine (DAC) or azacitidine is a form of epigenetic therapy that is clinically effective in the treatment of MDS and AML. In-vitro, histone deacetylase inhibition (HDACi) following DNA hypomethylation induction results in synergistic enhancement of gene expression activation, but the effects of HDACi on the cell cycle can also interfere with DNTMi activity, resulting in schedule dependent antagonism. Phase I/II studies of the combination of DNMTi and HDACi have shown some promise, triggering randomized studies.

Methods: We conducted a randomized phase II study of DAC at 20 mg/m² IV/1 hour daily ×5 q4 weeks vs. DAC at a similar dose + Valproic acid (VPA) 50 mg/kg PO daily ×7 starting on day 1 of DAC. Eligibility included MDS (FAB), IPSS>0 or AML, age >60 (excluding APL and CBF AML). An adaptive randomization design based on a composite score of CR, response and survival was used after the 40^th patient to assign patients to the superior arm. DNA methylation was measured by bisulfite pyrosequencing on peripheral blood mononuclear cells prior to and during treatment.

Results: 76 patients were enrolled on the study, 2 of whom received no therapy and are excluded from analysis. These included 8 patients with CMML (median age 72), 23 patients with AML (median age 71 (63–81), median BM blasts 40% (30–87), median WBCs 5.4 (1.1–97)) and 43 patients with MDS (median age 66 (36–89), IPSS Int1 (10), Int2 (19) and high (14)). Cytogenetics were abnormal in 40 patients (54%), most with complex or poor risk karyotypes. 42 patients (57%) were randomized to DAC alone. Overall, the median number of courses given so far is 4 (1–17) and 27 patients (36%) remain on therapy, at a median follow-up of 14 months. Response data are available for 67 patients (7 are too early). Overall, responses were seen in 31 patients (46%), with CR in 23 (34%) and other responses in 8 (12%). Overall response rate was 39% in AML, 71% in CMML and 46% in MDS. In patients receiving decitabine alone, the overall response rate was 17/40 (43%), compared to 14/27 (52%) in those randomized to DAC+VPA (p=NS). Median time to first response was 64 days (18–194) with DAC alone compared to 57 days (23–123) with DAC+VPA (p=NS). VPA added significant neurotoxicity to the regimen, with several patients discontinuing the drug due to somnolence or confusion. Median survival was 8.7 months in AML and 14.9 months in MDS (p=0.04). Kaplan-Meier analysis showed no difference in survival between DAC and DAC+VPA in the first year after therapy. DNA methylation analysis showed a similar degree of LINE demethylation in both arms,

Conclusions: Preliminary analysis of this randomized study suggests that adding VPA to DAC only marginally improves response rate and time to first response and has no impact on survival in MDS and AML. It remains to be seen (in randomized studies) whether more potent HDACi will show greater evidence of clinical synergy with DNMTi.