G-ROP criteria for predicting retinopathy of prematurity among neonates with different birth weight percentiles

BACKGROUND The Postnatal Growth and Retinopathy of Prematurity (G-ROP) Study showed that adding postnatal weight gain to birth weight and gestational age detected 100% of cases with type 1 retinopathy of prematurity (ROP) while reducing the ROP examinations by 30%. The purpose of this study was to investigate whether being small for gestational age (SGA) affects the sensitivity and specificity of the G-ROP model. METHODS We applied the G-ROP criteria for premature infants. The infants were classified as three subgroups: SGA, appropriate for gestational age (AGA), and large for gestational age (LGA). The performance of G-ROP criteria was assessed for each group for ROP. RESULTS There were 41 (10.5%) SGA, 312 (80%) AGA, and 37 (9.5%) LGA neonates. Twenty-six (6.7%) neonates were treated for ROP, and the G-ROP model identified all of them. The sensitivity of the model for treatment-requiring ROP (TR-ROP) was found to be 100% in the whole patient group and for each subgroup. The specificity for TR-ROP was 46.4% for the whole group, 50% for SGA, 44% for AGA, and 63.6% for LGA. By applying the G-ROP model, the number of ROP examinations could be reduced by 25% for the whole group, 27% for SGA, 24% for AGA, and 31% for LGA, without missing TR-ROP. CONCLUSIONS The sensitivity and specificity of the G-ROP model for TR-ROP in SGA infants were similar to the whole group. The model did not miss any cases of TR-ROP.