Abstract 3995 Background: Azacitidine (5-AzaC) and decitabine (DAC) are two hypomethylating agents commonly used in treatment of Myelodysplastic Syndromes (MDS). The comparative effectiveness of 5-AzaC versus DAC is not known. Both agents have been tested against supportive care in controlled settings, but not head-to-head. The most comprehensive assessment of the efficacy of 5-AzaC versus DAC was done by Kumar et al. (Haematologica 95 : 340-2) by performing an indirect meta-analysis using data from controlled trials, RCTs. The real world effectiveness of the 5-Aza-C versus DAC has not been evaluated. In response, we have performed a comparative effectiveness study of 5-AzaC versus DAC for treatment of MDS. Methods: All MDS patients treated with 5-AzaC or DAC at Moffitt Cancer Center from December 1999 to December 2009 were included in the analysis. Benefits and harms associated with 5-AzaC and DAC were assessed. The primary outcome was overall survival, OS. The secondary outcomes were progression free survival (PFS), defined as progression to Acute Myeloid Leukemia (AML), response rate evaluated according to IWG 2006 criteria, reduction in IV antibiotic use and transfusion dependence. For harms assessment, we extracted data on grade 3/4 toxicities and treatment related mortalities (TRM). OS and PFS were estimated using the Kaplan-Meier method and difference between treatments was calculated using the log-rank test. Cox proportional hazard model was used to estimate hazard ratios. Competing risk regression was applied in estimating PFS for time to AML or death, since the occurrence of AML may be unobservable due to patients' death. Covariate matching analysis (CMA) was applied for the outcomes of OS and PFS to estimate differences in treatment due to the non-random nature of the data. Each patient was matched to four other patients on prognostically significant covariates of age, MDS type, IPSS score, and FAB class. Differences in treatment effects for dichotomous outcomes were assessed using Mann-Whitney test. Results: One hundred seventy four MDS patients met the inclusion criteria (121 patients treated with 5-AzaC and 53 with DAC). Patient characteristics are summarized in table below. The unadjusted results showed a statistically significant OS benefit with 5-AzaC versus DAC (hazard ratio (HR) was 1.496 (95% Confidence Interval (CI), 1.005 to 2.226; p=0.047). However, the CMA showed a statistically non-significant gain of 1.088 months of survival with use of 5-AzaC versus DAC (95% CI, -4.219 to 6.395; p=0.688). For PFS, a statistically significant benefit was associated with 5-AzaC versus DAC (unadjusted HR=1.471 (95% CI 1.000 to 2.165; p=0.049). Nevertheless, CMA results showed a statistically non-significant gain of 1.060 months of PFS with use of 5-AzaC versus DAC (95% CI, -4.391 to 6.512; p=0.703). The overall response rate (20.0% vs 6.3%; p=0.029) and hematological response rate (24.4% vs 6.0%; p=0.006) was significantly better with 5-AzaC versus DAC. However, differences in cytogenetic response rate (23.4% vs 33.3%; p=0.398), bone marrow blast response rate (43.3% vs 54.5%; p=0.370), reduction of IV antibiotic drug use (9.5% vs 12.0%; p=0.780), and RBC (15.3% vs 12.2%; p=0.609) or platelet (5.0% vs 9.8%; p=0.291) transfusion requirements were statistically non-significant with 5-AzaC versus DAC. There was a statistically non-significant difference in occurrence of grade 3/4 toxicity (24.6% vs 37.6%; p=0.094) or TRM (1.0% vs 0.0%; p=.512) between patients treated with 5-AzaC versus DAC. Conclusions: Results from first retrospective population based study assessing the effectiveness of 5-AzaC versus DAC for treatment of MDS showed no significant difference in OS and PFS. However, 5-AzaC was associated with higher overall response rates compared with DAC with no significant difference in harms associated with the two treatments. Due to the limitations of retrospective analysis, these results warrant a prospective direct comparison of 5-AzaC versus DAC in a RCT. View this table: Disclosures: Lancet: Eisai: Consultancy; Celgene: Honoraria.

中文翻译：

---

阿扎胞苷与地西他滨治疗骨髓增生异常综合征的疗效比较

摘要 3995 背景：阿扎胞苷 (5-AzaC) 和地西他滨 (DAC) 是两种常用于治疗骨髓增生异常综合征 (MDS) 的低甲基化药物。5-AzaC 与 DAC 的比较有效性尚不清楚。这两种药物都在受控环境中接受了支持性护理的测试，但没有进行正面对战。Kumar 等人对 5-AzaC 与 DAC 的功效进行了最全面的评估。(Haematologica 95 : 340-2) 通过使用来自对照试验、RCT 的数据进行间接荟萃分析。尚未评估 5-Aza-C 与 DAC 的真实世界有效性。作为回应，我们对 5-AzaC 与 DAC 治疗 MDS 的有效性进行了比较研究。方法：1999 年 12 月至 2009 年 12 月在 Moffitt 癌症中心接受 5-AzaC 或 DAC 治疗的所有 MDS 患者都包括在分析中。评估了与 5-AzaC 和 DAC 相关的利弊。主要结果是总生存期，OS。次要结果是无进展生存期 (PFS)，定义为进展为急性髓性白血病 (AML)，根据 IWG 2006 标准评估的反应率，减少静脉注射抗生素的使用和输血依赖。对于危害评估，我们提取了 3/4 级毒性和治疗相关死亡率 (TRM) 的数据。使用 Kaplan-Meier 方法估计 OS 和 PFS，使用对数秩检验计算治疗之间的差异。Cox比例风险模型用于估计风险比。竞争风险回归用于估计 AML 或死亡时间的 PFS，因为 AML 的发生可能由于患者死亡而无法观察到。对 OS 和 PFS 的结果应用协变量匹配分析 (CMA) 以估计由于数据的非随机性而导致的治疗差异。根据年龄、MDS 类型、IPSS 评分和 FAB 等级等具有预后意义的协变量，每位患者与其他四名患者相匹配。使用 Mann-Whitney 检验评估二分类结果的治疗效果差异。结果：174 名 MDS 患者符合纳入标准（121 名患者接受 5-AzaC 治疗，53 名患者接受 DAC 治疗）。下表总结了患者特征。未经调整的结果显示，5-AzaC 与 DAC 相比具有统计学上显着的 OS 益处（风险比 (HR) 为 1。496（95% 置信区间 (CI)，1.005 至 2.226；p=0.047）。然而，CMA 显示使用 5-AzaC 与 DAC 相比，1.088 个月的生存率在统计学上无显着增加（95% CI，-4.219 至 6.395；p=0.688）。对于 PFS，5-AzaC 与 DAC 的统计学显着获益相关（未经调整的 HR=1.471（95% CI 1.000 至 2.165；p=0.049）。然而，CMA 结果显示 PFS 的 1.060 个月无统计学意义增加使用 5-AzaC 与 DAC（95% CI，-4.391 至 6.512；p=0.703）。总体反应率（20.0% 对 6.3%；p=0.029）和血液学反应率（24.4% 对 6.0%；p= 0.006) 显着优于 5-AzaC 与 DAC。然而，细胞遗传学反应率（23.4% 对 33.3%；p=0.398）、骨髓原始细胞反应率（43.3% 对 54.5%；p=0.370）、减少IV抗生素药物使用（9.5% vs 12.0%；p=0.780），并且 RBC（15.3% 对 12.2%；p=0.609）或血小板（5.0% 对 9.8%；p=0.291）输血需求在 5-AzaC 与 DAC 的情况下在统计学上不显着。5-AzaC 与 DAC 治疗的患者在 3/4 级毒性（24.6% 对 37.6%；p=0.094）或 TRM（1.0% 对 0.0%；p=.512）的发生率上存在统计学非显着差异. 结论：第一项基于人群的回顾性研究评估了 5-AzaC 与 DAC 治疗 MDS 的有效性，结果显示 OS 和 PFS 没有显着差异。然而，与 DAC 相比，5-AzaC 与更高的总体反应率相关，两种治疗相关的危害没有显着差异。由于回顾性分析的局限性，这些结果保证了在 RCT 中对 5-AzaC 与 DAC 进行前瞻性直接比较。查看此表：披露：柳叶刀：卫材：咨询；Celgene：荣誉金。