The hypomethylating drugs azacitidine and decitabine have shown efficacy in myelodysplastic syndromes and acute myeloid leukaemia, but complete tumour responses are infrequent and of short duration, possibly because of the short half-lives and suboptimal bone marrow exposure of the drugs. Guadecitabine, a next-generation hypomethylating drug, has a longer half-life and exposure than its active metabolite decitabine. A phase 1 study established 60 mg/m² guadecitabine for 5 days as an effective treatment schedule. In this phase 2 study, we aimed to assess the safety and activity of two doses and schedules of guadecitabine in older (≥65 years) patients with treatment-naive acute myeloid leukaemia who were not candidates for intensive chemotherapy.

We did a multicentre, randomised, open-label, phase 1/2 study of guadecitabine in cohorts of patients with treatment-naive acute myeloid leukaemia, relapsed or refractory acute myeloid leukaemia, and myelodysplastic syndromes; here we report the phase 2 results from the cohort of treatment-naive patients with acute myeloid leukaemia. We included patients aged at least 65 years from 14 US medical centres (hospitals and specialist cancer clinics) who were not candidates for intensive chemotherapy and randomly assigned them (1:1) using a computer algorithm (for dynamic randomisation) to guadecitabine 60 or 90 mg/m² on days 1–5 (5-day schedule) of a 28-day treatment cycle. Treatment allocation was not masked. We also assigned additional patients to guadecitabine 60 mg/m² in a 10-day schedule in a 28-day treatment cycle after a protocol amendment. The primary endpoint was composite complete response (complete response, complete response with incomplete platelet recovery, or complete response with incomplete neutrophil recovery regardless of platelets). Response was assessed in all patients (as-treated) who received at least one dose of guadecitabine. We present the final analysis, although at the time of the database lock, 15 patients were still in follow-up for overall survival. This study is registered with ClinicalTrials.gov, number NCT01261312.

Between Aug 24, 2012, and Sept 15, 2014, 107 patients were enrolled: 54 on the 5-day schedule (26 randomly assigned to 60 mg/m² and 28 to 90 mg/m²) and 53 were assigned to the 10-day schedule. Median age was 77 years (range 62–92), and median follow-up was 953 days (IQR 721–1040). All treated patients were assessable for a response. The number of patients who achieved a composite complete response did not differ between dose groups or schedules (13 [54%, 95% CI 32·8–74·4] with 60 mg/m² on the 5-day schedule; 16 [59%; 38·8–77·6] with 90 mg/m² on the 5-day schedule; and 26 [50%, 35·8–64·2] with 60 mg/m² on the 10-day schedule). The most frequent grade 3 or worse adverse events, regardless of relationship to treatment, were febrile neutropenia (31 [61%] of 51 patients on the 5-day schedule vs 36 [69%] of 52 patients on the 10-day schedule), thrombocytopenia (25 [49%] vs 22 [42%]), neutropenia (20 [39%] vs 18 [35%]), pneumonia (15 [29%] vs 19 [37%]), anaemia (15 [29%] vs 12 [23%]), and sepsis (eight [16%] vs 14 [27%]). The most common serious adverse events, regardless of relationship to treatment, for the 5-day and 10-day schedules, respectively, were febrile neutropenia (27 [53%] vs 25 [48%]), pneumonia (14 [27%] vs 16 [31%]), and sepsis (eight [16%] vs 14 [27%]). 23 (22%) patients died because of adverse events (mainly from sepsis, eight [8%]; and pneumonia, five [5%]); four deaths were from adverse events deemed treatment-related (pneumonia, two [2 …