Background: Recombinant human erythropoietin (rhEPO) lost its role in minimizing red blood cell transfusion in very preterm infants after it had been associated with severe retinopathy of prematurity (ROP). Previous systematic reviews did not stratify ROP by gestation and birth weight (BW). Objectives: The aim of this study was to investigate the effect of early prophylactic rhEPO on ROP in a stratified meta-analysis of randomized controlled trials (RCTs). Methods: The databases EMBASE, MEDLINE, and the Cochrane Central Register of Controlled Trials were searched in January 2022 and complemented by citation searching. RCTs comparing early rhEPO treatment with no treatment or placebo were selected if they were published in a peer-reviewed journal and reported ROP outcomes. Previously unpublished data were requested from the study authors to allow stratified analyses by gestational age (GA) and BW. Data were extracted and analyzed using the standard methods of the Cochrane Neonatal Review Group. Pre-specified outcomes were "ROP stage =3" (primary outcome) and "any ROP." Results: Fourteen RCTs, comprising 2,040 infants of <29 weeks of GA, were included for meta-analysis. Data syntheses showed no effects of rhEPO on ROP stage =3 or on any ROP, neither in infants of <29 weeks GA, nor in infants of <1,000 g BW, nor in any GA strata. The risk ratio (95% confidence interval) for ROP stage =3 in infants of <29 weeks of GA was 1.13 (0.84, 1.53), p = 0.41 (quality of evidence: moderate). Conclusions: The present meta-analysis detected no effects of early rhEPO on ROP in any comparison, but most stratified analyses were limited by low statistical power.
