This study was performed to identify whether cytogenetics, International Prognostic Scoring System (IPSS), or World Health Organization Classification-Based Prognostic Scoring System are predictive of the efficacy of azacitidine in patients with myelodysplastic syndrome (MDS). We retrospectively reviewed the clinical records of 113 patients with MDS treated with azacitidine. The "response alternating disease natural history," "cytogenetic response," and "hematologic improvement" were assessed by serial bone marrow biopsy, cytogenetic study, and hemogram analyses. The complete and partial remission rates were 17.6% and 3.9% in 51 evaluable patients. There were no significant differences in response rate in the different cytogenetic/IPSS/WPSS groups. The overall hematologic response (HR) rate was 49.6%, and the HR rate was significantly greater in patients classed as "very high" risk according to the WPSS compared with other patient groups. The 1-year overall survival (OS) rate was higher among patients with HR compared with those without HR (80.9% vs 63.3%, p = 0.046), and the 1-year OS rate among patients classed as being at high risk by each criteria was similar to that of patients classed as being at low risk. The hazard ratio of death among patients with HR compared with those without HR was 0.17 (95% CI 0.04-0.69) for high + very high risk group based on WPSS. Patients in the WPSS high-risk group had an increased HR rate compared with other patient groups, and the achievement of HR was associated with a significant increase in OS. Azacitidine showed similar efficacy in all patient groups, even in patients with poor cytogenetics and in high-risk groups.