Prediction of ROP Treatment and Evaluation of Screening Criteria in VLBW Infants–a Population Based Analysis

Background: The incidence of retinopathy of prematurity (ROP) and ROP screening criteria differ between countries. We assessed whether ROP screening could be reduced based on the local ROP incidence. Methods: Observational cohort study of infants born in Switzerland between 2006 and 2015 <32 0/7 weeks. Chronological and postmenstrual ages at ROP treatment were analyzed. A model to identify ROP treatment on patients born between 2006 and 2012 (training set) was developed and tested on patients born between 2013 and 2015 (validation set). Results: Of 7817 live-born infants, 1098 died within the first 5 weeks of life. The remaining 6719 infants were included into analysis. All patients requiring ROP treatment would have been identified if screening had been performed before reaching 60 days of life or 37 3/7 weeks postmenstrual age, whichever came first. The training and validation sets included 4522 and 2197 preterm infants encompassing 56 and 20 patients receiving ROP treatment, respectively. All patients would have required screening to reach 100% sensitivity. To reach a sensitivity of 95.0% and a specificity of 87.6%, we predicted a reduction in 13.2% of patients requiring screening (c-statistic = 0.916). Conclusions: A substantial reduction of infants requiring screening seems possible, but necessitates prospective testing of new screening criteria.