Jason M Huang1, Xihui Lin1, Yu-Guang He1, Jennifer H Cao2. 1. Department of Ophthalmology, University of Texas Southwestern Medical Center, Dallas, Texas. 2. Department of Ophthalmology, University of Texas Southwestern Medical Center, Dallas, Texas. Electronic address: Jennifer.cao@utsouthwestern.edu.
Abstract
PURPOSE: The Colorado Retinopathy of Prematurity Screening Algorithm (CO-ROP) recommends screening for infants meeting the following criteria for retinopathy of prematurity (ROP): gestational age ≤30 weeks, birth weight of ≤1500 g, and net weight gain of ≤650 g between birth and 4 weeks of age. This study was performed to evaluate the validity of CO-ROP in a tertiary referral county hospital. METHODS: CO-ROP was used to retrospectively analyze the data from consecutive newborns screened for ROP using national screening guidelines at Parkland Hospital, Dallas, Texas, between April 1, 2009, to August 30, 2013. Sensitivities and specificities for identifying ROP were calculated. RESULTS: A total of 374 infants were included, of whom 29 (7.8%) developed type 1 ROP and 12 (3.2%) developed type 2 ROP. The CO-ROP model would have decreased number of infants screened by 34% compared to current national screening criteria. CO-ROP had sensitivity of 93.1% (95% CI, 77.2-99.1) and 92.7% (95% CI, 61.5-99.8) for identifying type 1 and type 2 ROP, respectively. Of 29 patients who developed type 1 ROP, 2 were not identified using CO-ROP. CONCLUSIONS: The CO-ROP model significantly reduced total number screened but failed to detect 2 infants with type 1 ROP, suggesting the need for further modification of the algorithm.
PURPOSE: The Colorado Retinopathy of Prematurity Screening Algorithm (CO-ROP) recommends screening for infants meeting the following criteria for retinopathy of prematurity (ROP): gestational age ≤30 weeks, birth weight of ≤1500 g, and net weight gain of ≤650 g between birth and 4 weeks of age. This study was performed to evaluate the validity of CO-ROP in a tertiary referral county hospital. METHODS: CO-ROP was used to retrospectively analyze the data from consecutive newborns screened for ROP using national screening guidelines at Parkland Hospital, Dallas, Texas, between April 1, 2009, to August 30, 2013. Sensitivities and specificities for identifying ROP were calculated. RESULTS: A total of 374 infants were included, of whom 29 (7.8%) developed type 1 ROP and 12 (3.2%) developed type 2 ROP. The CO-ROP model would have decreased number of infants screened by 34% compared to current national screening criteria. CO-ROP had sensitivity of 93.1% (95% CI, 77.2-99.1) and 92.7% (95% CI, 61.5-99.8) for identifying type 1 and type 2 ROP, respectively. Of 29 patients who developed type 1 ROP, 2 were not identified using CO-ROP. CONCLUSIONS: The CO-ROP model significantly reduced total number screened but failed to detect 2 infants with type 1 ROP, suggesting the need for further modification of the algorithm.