Final Analysis of the Daliah Trial: A Randomized Phase III Trial of Interferon-α Versus Hydroxyurea in Patients with MPN

Abstract

Background: Hydroxyurea (HU) is the most commonly used first-line cytoreductive treatment option for patients with myeloproliferative neoplasms (MPN) worldwide. However, increasing evidence on the efficacy and safety of pegylated interferon-alpha2 (IFNα) is emerging, and optimal first-line treatment is to be established. Aims: To compare the efficacy and safety of HU vs. low-dose IFNα in patients with MPN over five years. Methods: DALIAH (NCT01387763) was a randomized phase III trial of HU vs. IFNα in newly diagnosed or untreated patients with MPN (essential thrombocythemia (ET), polycythemia vera (PV), prefibrotic myelofibrosis (PreMF), and primary myelofibrosis (PMF)). Patients > age 60 were randomized (1:1:1) to HU, IFNα-2a, or IFNα-2b whereas patients ≤ age 60 were randomized to IFNα-2a or IFNα-2b. The primary outcome was the JAK2V617F molecular response (MR) rate at 18, 36, and 60 months per 2009 European LeukemiaNetwork (ELN) (ET, PV, PreMF) or 2005 European Myelofibrosis Network (EUMNET) (PMF) criteria. Secondary outcomes included the complete clinicohematologic response (CHR) rate at 12 months. The JAK2V617F allele burden was measured using quantitative polymerase chain reaction (qPCR) on peripheral blood (assay sensitivity: 0.1%). Primary and secondary outcomes were compared between groups (HU vs. IFNα or HU vs. IFNα > 60 years) using Fisher's exact test (categorical variables) or Wilcoxon rank-sum test (continuous variables). Paired comparisons (within groups) were made using Wilcoxon signed-rank test. Serial JAK2V617F measurements were compared by unadjusted mixed-effects linear regression analysis. Results: We included 203 patients (ET: 73 (36%), PV: 89 (44%), PreMF: 16 (8%), and PMF: 25 (12%)) in the modified intention-to-treat (ITT) population. Baseline characteristics were well balanced except for median age (HU: 68 years vs. IFNα: 59 years, p<0.0001) ( Table 1). The MR rate by ITT analysis was similar between HU and IFNα (18 months: 19% vs. 21%, p=1.00; 36 months: 19% vs. 26%, p=0.64; 60 months: 23% vs. 24%, p=1.00) (Figure 1A). However, the JAK2V617F allele burden was significantly lower in the IFNα group at month 36 and beyond (Figure 1B) and the absolute median (IQR) change in JAK2V617F allele burden (baseline to 60 months) was greater with IFNα (-20% (-9;-49) vs. -7% (3;-15), p=0.0053) (Figure 1C). Two patients (IFNα: n=2) were in complete molecular remission (undetectable JAK2V617F) at 60 months. The CHR rate by ITT analysis was higher with HU at 18 months (58% vs. 38%, p=0.03) but similar at all other time points (12 months: 50% vs. 36%, p=0.21; 60 months: 24% vs. 22%, p=0.83) ( Figure 1D). A post hoc subgroup analysis comparing HU with IFNα > age 60 showed comparable efficacy results. Among patients remaining on treatment (per-protocol analysis), the MR and CHR rates were superior in the IFNα group compared to the HU group at 36 months and beyond. The MR and CHR rates (HU vs. IFNα) by per-protocol analysis were: MR at 36 months: 23% vs. 56%, p=0.01; MR at 48 months: 27% vs. 59%, p=0.02; MR at 60 months: 35% vs. 67%, p=0.03; CHR at 36 months: 33% vs. 67%, p=0.002; CHR at 60 months: 38% vs. 62%, p=0.05. Overall treatment discontinuation at 60 months was 60% (HU: 37% vs. IFNα: 65%, p=0.0019). The most common cause of treatment discontinuation was adverse events (HU: 6/38 (16%); IFNα: 71/165 (43%)). Adverse events ≥ grade 3 occurred in 46% (HU: 58% vs. IFNα: 45%, p=0.15). Nineteen major thrombotic events were reported in 16 patients (HU: 4 events in 4 patients; IFNα>60: 12 events in 10 patients; IFNα≤60: 3 events in 2 patients), corresponding to an incidence rate of 2.6 per 100 person-years for HU and 3.4 per 100 person-years for IFNα (IFNα>60: 6.2; IFNα≤60: 1.2). None of the patients transformed into secondary acute myeloid leukemia. Five patients died during follow-up (HU: 2; IFNα: 3). Bone marrow histologic remission rates at 36 and 60 months will be presented at the meeting. Conclusion: ITT analysis detected no significant difference in the MR or CHR rates between HU and IFNα with long-term treatment (5 years) among patients with MPN, reflecting a higher treatment discontinuation rate in the IFNα group (65%). Thus, using the per-protocol principle, the MR and CHR rates were superior in the IFNα group at 36 months and beyond.