Efficacy and Safety of a 5-Day Regimen of Azacitidine for Patients with High-Risk Myelodysplastic Syndromes

Background: In the AZA-001 trial, azacitidine (AZA; 75 mg/m²per day for 7 consecutive days of every 28-day cycle) demonstrated improved survival compared with conventional care regimens in patients with high-risk myelodysplastic syndromes (MDS). Although a 7-day regimen of AZA is the standard treatment of high-risk MDS, it is difficult to administer AZA treatment in outpatient settings on weekends. In this study, we investigated outcomes of a 5-day regimen of AZA in patients with high-risk MDS.

Methods: Clinical data were retrospectively collected from consecutive high-risk MDS patients treated with AZA at Fujisawa City Hospital. “High-risk MDS” was defined as MDS with transfusion dependency or MDS with an International Prognostic Scoring System (IPSS) risk of intermediate-2 or high. Every month, AZA was administered at 75 mg/m²per day: for 5–7 days in the hospitalization for the first cycle, and for 5 days in an outpatient center for all subsequent cycles. Patients with neutropenia received antimicrobial prophylaxis with levofloxacin. Responses to AZA were evaluated according to the International Working Group 2006 criteria.

Results: Between April 2011 and December 2013, 50 patients with MDS (40 men and 10 women; age, 40–86 years; median age, 73 years) initiated the AZA treatment. All patients had an Eastern Cooperative Oncology Group performance status of 0–2. Thirty-nine patients had primary MDS and 11 patients had secondary MDS. According to the World Health Organization MDS classifications, 13 patients had RCMD, 12 had RAEB-1, 15 had RAEB-2, 1 had CMML and 9 had AML with 20–30% bone marrow blasts. The IPSS risk was intermediate-1 in 19 patients, intermediate-2 in 19 patients, and high in 11 patients. The IPSS of 1 RAEB-2 patient could not be estimated due to insufficient data of chromosome analysis. The cytogenetic risk, known for 49 patients, was good for 23 patients, intermediate for 14, and poor for 12. Thirty-six patients were red blood cell transfusion-dependent and 17 patients were platelet transfusion-dependent. With the exception of 1 AML patient who received AZA as consolidation therapy after induction of chemotherapy, all other 49 patients received AZA as first-line therapy. The median time to treatment from diagnosis was 1.5 months (range, 1–153). The median number of AZA cycles was 7 (range, 1–27). AZA was discontinued before the third cycle in 12 patients due to: 3 cases of hematopoietic stem cell transplantation, 2 cases of pneumonia, 2 cases of insufficient blood cell recovery, 1 case of cerebral hemorrhage, 1 case of gastrointestinal hemorrhage, 1 case of stomach cancer, 1 case of rash, and 1 case involving a geographical move. Among 37 patients who received more than 3 cycles of AZA as fist-line therapy, 23 (62%) achieved any hematological improvement. The median time to first response was 3 cycles (1–5). Hematological improvement was obtained irrespective of IPSS risk or of primary and secondary MDS classifications. Grade 3/4 neutropenia occurred in 39 patients, and grade 3/4 thrombocytopenia occurred in 38 patients. Most of these hematological adverse events occurred during cycles 1–2. The most common grade 3 or higher non-hematological adverse events were pneumonia (9 patients), febrile neutropenia (8 patients), and sepsis (5 patients). According to IPSS risk, median overall survival was 12 months (range, 1–24) in high risk MDS and 16 months (range, 1–28) in intermediate-2 risk MDS. Median overall survival in intermediate-1 risk MDS had not yet been reached (range, 6–29 months, p = 0.03). Median overall survival in primary and secondary MDS was 24 months (range, 1–29) in primary MDS and 12 months (range, 1–19) in secondary MDS (p = 0.05). Twenty-four patients died: 9 of overt leukemia, 9 of pneumonia, 2 of cerebral hemorrhage, 1 of sepsis, 1 of gastrointestinal hemorrhage, 1 of stomach cancer, and 1 of senility.

Conclusion: In the AZA-001 trial, the median time to first response was 2 cycles. In our study, the median time to first response was 3 cycles. Although the first response was delayed, the 5-day regimen of AZA showed respectable continuation rates lasting longer than 3 cycles with equivalent efficacy to the 7-day regimen. Future studies comparing the 5-day regimen and the 7-day regimen would provide valuable follow-up information to the findings in this study.