Authors: Fatma Zaguia, Philippe Lamer, Jiaru Liu, Martine Claveau, Therese Perreault, Robert Koenekoop, Daniela Toffoli, Ayesha Khan

Author Disclosure Block: F. Zaguia: None. P. Lamer: None. J. Liu: None. M. Claveau: None. T. Perreault: None. R. Koenekoop: None. D. Toffoli: None. A. Khan: None.

Abstract Body:

Purpose: Retinopathy of prematurity (ROP) is a leading cause of preventable blindness worldwide. Early detection and treatment are crucial in maintaining vision. Current screening techniques include routine dilated retinal exam of premature newborns by trained ophthalmologists. This can often be burdensome and distressing to infants and their families. While using the current screening criteria, less than 10% of screened patients will go on to require treatment. In recent years, Loqvist et al have developed WINROP (Weight, IGF, Neonatal ROP), as an alternative to traditional screening methods. WINROP is a free online surveillance algorithm that uses longitudinal postnatal weight measurements to better predict early risk for developing severe ROP that will require treatment. The WINROP algorithm has been validated in several countries around the world, however scarce data is available to support the program in geographically and racially diverse populations such as ours in Montreal. The aim of our study is to validate the WINROP algorithm as an ROP screening tool in our North American cohort of preterm infants. 

Study Design: Study is a retrospective chart review between May 1st 2015 and August 1st 2017. Methods: Study was conducted in the neonatal intensive care unit at the Montreal Children’s Hospital at the McGill University Health Centre, Montreal, Quebec. 351 infants were eligible for ROP screening, and following our exclusion criteria, 240 infants were included in the study. Gestational age, birth weight and weekly weights were recorded and entered in the WINROP online surveillance algorithm. 

Results: The prevalence of sight-threatening ROP requiring treatment was 1.7% in our population. The median time from birth to WINROP alarm was 16 days (range: 7-20 days). The median time from the alarm to the time of diagnosis of sight-threatening ROP was 13 weeks (range: 11.4 to 14.6 weeks), with infants being treated an average of 1.5 days following diagnosis. In our cohort, the sensitivity of the WINROP algorithm to detect vision-threatening ROP was 100%, with a specificity of 57%. The positive predictive value was 3.8% and negative predictive value was 100%. In total, WINROP alarm was activated for 106 infants, which would represent a decrease of 56% in dilated fundus examinations. 

Conclusions: WINROP algorithm detected 100% of infants who developed ROP requiring treatment, suggesting optimal sensitivity to be used clinically in our population. Furthermore, time of alarm to proliferation ROP suggests early detection. Using this screening tool, stressful eye exams could be markedly reduced in our population.