Treatment with recombinant human erythropoietin (rHu-EPO) modestly prevented packed red blood cell transfusions (pRBCTs) in preterm infants in studies performed several years ago. In France, some neonatal units stopped using rHu-EPO, while others continued. The aim of this study was to explore the role of rHu-EPO in the prevention of pRBCTs in a recent cohort of preterm infants. Preterm infants who met rHu-EPO indications and were hospitalised between 2018 and 2020 in two neonatal units -one that did not use rHu-EPO and another that did- were eligible. Data about the neonatal history, rHu-EPO and iron treatments and pRBCT indications and volumes were collected. Infants exposed and not exposed to rHu-EPO were compared in univariate and multivariate analyses using backward logistic regression and Cox proportional hazards regression. A total of 257 patients exposed to rHu-EPO and 285 patients who were not exposed were included. Three profiles emerged. In the infants with a gestational age <28 weeks, the cumulative pRBCT volume/kg was similar regardless of rHu-EPO exposure (mean difference -2.8 mL, 95% confidence interval -16.1, 10.5, p=0.68). In the infants born between 28 and 30 weeks, a late pRBCT was prevented in the rHu-EPO group (single pRBCT: no rHu-EPO 22.1% vs rHu-EPO 8%, p=0.003). However, rHu-EPO was not independently associated with avoidance of this pRBCT. Finally, the need for pRBCT was low in the infants born after 30 weeks of gestation, making rHu-EPO treatment futile. In contrast, early iron supplementation was revealed to be critical in preventing pRBCT. No benefit of rHu-EPO in preventing pRBCT was observed in our cohort. The place of rHu-EPO in future requires careful consideration of the population concerned, adjustment of the therapeutic schedule and evolution of the indications for pRBCT.