We thank Dr. Bahr and colleagues for their interest in our study [1]. We agree that there is a hierarchical structure regarding the level of evidence, in which systematic reviews of randomized controlled trials (RCTs) provide the highest level of evidence. Therefore, we should carefully interpret the results of our cohort study considering the results of previous high-quality RCTs.

We consider the level of evidence for our cohort study to be 2B, based on the system used by Dr. Bahr [1]. Previous studies have reported no association between erythropoietin use and retinopathy of prematurity (ROP) [2], although the RCT by Ohls et al. [2] included in Dr. Bahr’s letter [1] had not been formally published as a full-text article at the publication of our study [3]. In contrast, our cohort study found an association between erythropoietin and ROP [3]. The possibility of confounding bias cannot be ruled out. However, the following two points in our study, which differ from the existing RCTs, may be related to the differences in the study findings.