Erythropoietin (EPO) plays an important role in the development and maturation of the gastrointestinal tract. Recombinant EPO (rEPO) has been used to prevent anemia of prematurity. The gastrointestinal trophic effects of EPO may reduce feeding intolerance and necrotizing enterocolitis (NEC) in preterm neonates. The aim of this systematic review of randomized controlled trials (RCTs) was to evaluate the effects of rEPO on clinical outcomes such as feeding intolerance, stage II or higher NEC, any stage NEC, sepsis, retinopathy of prematurity, and bronchopulmonary dysplasia in preterm neonates. Twenty-five RCTs (intravenous: 13; subcutaneous: 10; enteral: 2; n = 4025) were eligible for inclusion. Meta-analysis of data from 17 RCTs (rEPO compared with placebo) with the use of a fixed-effects model showed no significant effect of rEPO on stage II or higher NEC (RR: 0.87; 95% CI: 0.64, 1.19; P = 0.39). Meta-analysis of data from 25 RCTs (rEPO compared with placebo) showed that rEPO significantly decreased the risk of any stage NEC [cases/total sample: 120/2058 (5.83%) compared with 146/1967 (7.42%); RR: 0.77; 95% CI: 0.61, 0.97; P = 0.03]. Only one RCT reported on time to full feedings. Meta-analysis of data from 15 RCTs showed a significant reduction in late-onset sepsis after rEPO administration (RR: 0.81; 95% CI: 0.71, 0.94; P = 0.004). Meta-analysis of 13 RCTs showed no significant effect of rEPO on mortality, retinopathy of prematurity, and bronchopulmonary dysplasia. Prophylactic rEPO had no effect on stage II or higher NEC, but it reduced any stage NEC, probably by reducing feeding intolerance, which is often labeled as stage I NEC. Adequately powered RCTs are required to confirm these findings.