Human Recombinant Erythropoietin and the Incidence of Retinopathy of Prematurity: A Multiple Regression Model

Abstract

Introduction: Recombinant human erythropoietin (rhEPO) has been used for the treatment of anemia of prematurity, effectively decreasing blood transfusion requirements. However, it has also been found to have properties similar to vascular endothelial growth factor (VEGF), the major angiogenic factor implicated in the pathogenesis of retinopathy of prematurity (ROP). Methods: Data were analyzed from 322 infants admitted to a tertiary medical center’s neonatal intensive care unit (NICU) in 1994 and 2002. The infants were analyzed for demographic data, incidence of markers of major morbidity, rhEPO treatment, number of red blood cell (RBC) transfusions received, and incidence and severity of ROP. A multiple logistic regression model was used to determine the relation of the studied risk factors to the outcome measures of incidence (any stage) and severity (threshold ROP requiring cryotherapy or laser photocoagulation) of ROP. Results: The risk of developing ROP was increased in infants who received >20 doses of rhEPO (OR, 3.53; 95% CI, 1.59 to 7.85). These infants were also more likely to require laser photocoagulation (OR, 4.31; 95% CI, 1.99 to 9.33). The age at which rhEPO was started was also significant, with those starting rhEPO after 20 days of age having increased risk of ROP (OR, 3.57; 95% CI, 1.59 to 8.03). Discussion: The addition of rhEPO, an angiogenic factor, to the already accumulated VEGF in the premature retina may cause a threshold level of angiogenic factors to be reached, leading to abnormal neovascularization and severe ROP. Conclusion: These findings represent the first use of multiple logistic regression to demonstrate rhEPO as an independent risk factor for ROP.