ABSTRACT
Background: The characteristics of premature infants vary from country to country rendering it challenging to apply retinopathy of prematurity (ROP) screening algorithm globally. The screening criteria for postnatal growth and ROP (G-ROP) for preterm infants are known to be beneficial, but it is not clear whether these criteria can be used universally. Aim: The aim of this study is to validate the accuracy of the G-ROP criteria in screening preterm infants in Saudi Arabia. Subjects and Methods: This single-center retrospective study included 300 premature infants (mean gestational age [GA], 28.72 ± 2.2 weeks; range, 21-36 weeks) at a referral center who were screened for ROP between 2015 and 2021. The inclusion criteria were the availability of data on ROP outcome and body weight up until day 40 after birth. The G-ROP 1 and G-ROP 2 models were examined for their ability and accuracy in identifying infants with any stage ROP and treatable ROP. Results: The G-ROP 1 and G-ROP 2 models identified 233 and 255 infants for screening, respectively. The sensitivity of G-ROP 1 and G-ROP 2 for detecting treated ROP was 96.7% and 100%, respectively, and the specificity for detecting treatable ROP was 24.4% and 16.7%, respectively. Incorporation of the G-ROP 2 model, which did not miss any infant with type 1 ROP, would have reduced the number of screened infants by 15%. Conclusion: G-ROP 2 was more sensitive than G-ROP 1 for identifying infants who required treatment and could potentially reduce the burden of ROP screening.
