Objective: Earlier studies have shown that erythropoietin (Epo) is beneficial in reducing erythrocyte transfusions given for symptomatic anemia among very low birth weight infants. The objective of this study was to determine if blood transfusions could be reduced using erythropoietin among stable preterm infants who are still at risk of donor blood exposure.Methods: A randomized, non-placebo controlled clinical trial was conducted on 28 preterm infants with birth weights ranging from 1000 to 1500 grams. Inclusion criteria also included being appropriately sized for gestational age and having an hematocrit ≥30% at study entry. Study infants were randomly assigned to receive Epo, 300 units/kg/day 3 times a week subcutaneously (n=13) or no treatment (n=15) starting at less than 2 weeks of age for 6 weeks. Criteria for transfusion was set by a standard protocol for both the treatment group and controls. Infants in both groups were supplemented with iron (Fer-In-Sol 0.3ml/day) beginning at 2 weeks of age. The dose of Fer-In-Sol was increased weekly by 0.3ml/day if serum ferritin level was noted to be less than 65.0ng/ml. The primary outcome variable was the total number and volume of transfusions given per group during the entire hospitalization. Results: Two infants required transfusions among the 13 Epo treated infants compared to 9 infants among the 15 control infants(p=0.02). The total number of transfusions given was 2 in the Epo group versus 18 in the controls (89% reduction, p=0.01). The total volume of transfusions given to the Epo group was 39 ml compared to 487 ml in the control group (92% reduction, p=0.007). We found no differences in the total white blood count, neutrophil count, platelet count, incidence of sepsis, necrotizing enterocolitis, seizures and hypertension between the groups. The mean hematocrit for the Epo group at 2 weeks into the study was 34.5±2.8 compared to 29.2±4.7 among controls (p=0.003). The mean hematocrit values at other study periods could not be meaningfully analysed because of transfusions given to control infants. Conclusions: These data suggest that use of Epo in stable preterm infants with birth weights of 1000 to 1500 grams is an effective means of reducing the need for transfusions.