Background: Regular transfusion dependence is seen in almost every MDS patient (pt) during the natural course of disease. Due to deleterious sequelae of chronic transfusions, including iron-overload, alloimmunization, and infections, delaying transfusion dependence or attainment of transfusion-independence are important MDS treatment goals. The role of epoetin alfa (EPO) is well established for anemia treatment prior to transfusion dependence, although its utility in transfusion-dependent pts is not well understood. The present meta-analysis was conducted to assess the efficacy of EPO in achieving transfusion independence in transfusion-dependent MDS pts.Methods: Data was extracted from studies from PubMed, ASCO and ASH proceedings from 1990–2006 in transfusion-dependent MDS pts treated with EPO±G/GM-CSF. Studies using International Working Group (IWG) response and IWG-like criteria with EPO±G/GM-CSF were compared with EPO monotherapy studies not using IWG criteria (Non-IWG). Pooled estimates of erythroid response (ER) rates were calculated using random-effects models. Where only pre-therapy values were available, values for post-therapy reductions in RBC requirements (units/month) were imputed based on response criteria stated in the study.Results: Included were 578 evaluable transfusion-dependent pts representing 58% of the total study population in 25 studies. All studies collectively showed significant treatment effects from baseline (BL) as evidenced by a mean overall ER rate of 34.5% and a mean proportion of pts achieving complete transfusion independence of 23.3% with EPO±G/GM-CSF (p<.001 for both). A significant decrease was observed in mean RBC units/month after treatment (2.55 pre-therapy vs 1.96 post-therapy,p=.0002,n=17 studies). IWG studies showed numerically higher ER rates and lower transfusion requirements than Non-IWG studies (Table) (p>.05). Among IWG studies, EPO monotherapy as compared to EPO±G/GM-CSF, showed comparable ER rates (41.6% vs 36.4%,p=.643), a comparable proportion of pts achieving transfusion-independence (28.8% vs 24.8%,p=0.705), and a greater reduction in RBC units/month (34.3% vs 10.4%,p=0.339). Additionally, EPO monotherapy IWG studies showed dose response with significantly higher overall ER rates in pts receiving 60-80,000 units/week vs 30-40,000 units/week (45.3% vs 30.8%,p=.041). Multivariate regression did not yield significant association of any BL study characteristics with overall ER rate.Conclusion: Findings from this meta-analysis showed EPO±G/GM-CSF treatment is effective in reducing transfusion requirements in transfusion-dependent MDS pts. Identification of predictors of EPO response in this population may offer significant clinical benefit to this subset.Study GroupNon-IWG StudiesIWG StudiesIWG StudiesEPOMonotherapyEPOMonotherapyEPO +G/GM-CSFNo. of studies (evaluable pts)12(268)7(195)6(115)BL sEPO levels, mU/mL (Mean±SD)687.2±684.7374.3±72.2288±172.1BL Hb, g/dL (Mean±SD)8.5±0.67.9±0.48.1±0.4Initial weekly dose of EPO, Units (Mean±SD)48,310±41,34549,326±22,75234,789±12,083Mean monthly pre-therapy RBC Units2.712.372.33Pooled ER rate, % (95%CI)27.8(20.4–35.2)41.6(24.5–58.6)36.4(22.5–50.2)Post-therapy transfusion-independent pts, % (95%CI)17.9(10.9–24.9)28.8(12.6–45.1)24.8(11.8–37.8)Mean post-therapy reduction in RBC units/month, % (Median)17.3(17.1)34.3(24.7)10.4(11.5)