Risk Factors Associated with Retinopathy of Prematurity in Very and Extremely Preterm Infants.

Background and Objectives: Retinopathy of prematurity (ROP) is the leading cause of blindness in preterm infants. We studied the relationship between different perinatal characteristics, i.e., sex; gestational age (GA); birth weight (BW); C-reactive protein (CRP) and lactate dehydrogenase (LDH) concentrations; ventilation, continuous positive airway pressure (CPAP), and surfactant administration; and the incidence of Stage 1-3 ROP. Materials and Methods: This study included 247 preterm infants with gestational age (GA) < 32 weeks that were successfully screened for ROP. Univariate and multivariate binary analyses were performed to find the most significant risk factors for ROP (Stage 1-3), while multivariate multinomial analysis was used to find the most significant risk factors for specific ROP stages, i.e., Stage 1, 2, and 3. Results: The incidence of ROP (Stage 1-3) was 66.40% (164 infants), while that of Stage 1, 2, and 3 ROP was 15.38% (38 infants), 27.53% (68 infants), and 23.48% (58 infants), respectively. Following univariate analysis, multiple perinatal characteristics, i.e., GA; BW; and ventilation, CPAP, and surfactant administration, were found to be statistically significant risk factors for ROP (p < 0.001). However, in a multivariate model using the same characteristics, only BW and ventilation were significant ROP predictors (p < 0.001 and p < 0.05, respectively). Multivariate multinomial analysis revealed that BW was only significantly correlated with Stage 2 and 3 ROP (p < 0.05 and p < 0.001, respectively), while ventilation was only significantly correlated with Stage 2 ROP (p < 0.05). Conclusions: The results indicate that GA; BW; and the use of ventilation, CPAP, and surfactant were all significant risk factors for ROP (Stage 1-3), but only BW and ventilation were significantly correlated with ROP and specific stages of the disease, namely Stage 2 and 3 ROP and Stage 2 ROP, respectively, in multivariate models.