The Specificity of the WINROP Algorithm Can Be Significantly Increased by Reassessment of the WINROP Alarm

Abstract

Background: Retinopathy of prematurity (ROP) is a sight-threatening disease affecting extremely preterm infants. The introduction of new ROP screening surveillance systems, with higher sensitivity and specificity than established ROP screening guidelines, has the potential to reduce the number of stressful eye examinations in these infants.

Objectives: To improve the specificity of the WINROP (Weight, Insulin-like growth factor-I, Neonatal, ROP) surveillance system, identifying extremely preterm infants requiring treatment for ROP.

Methods: Two cohorts that had previously been subjected to WINROP analyses were included and reevaluated in this study. The weight at WINROP alarm for extremely preterm infants, born at gestational age <27 weeks, was reevaluated and by establishing 'safe' WINROP alarm weight limits, an intersample reassessment of WINROP alarm was performed. The two cohorts were as follows: (1) the Extremely Preterm Infants in Sweden Study (EXPRESS) cohort, infants born in Sweden during 2004-2007 (n = 407), and (2) extremely preterm infants in a North American cohort, born during 2006-2009 (n = 566).

Results: In the EXPRESS cohort, 12.5% (40/319) of the infants who previously received a WINROP alarm were now reassessed as having no alarm; the specificity of WINROP in EXPRESS increased from 23.9% (86/360) to 35.0% (126/360). In the North American cohort, 15.4% (81/526) were reassessed as having no alarm; the specificity increased from 8.5% (38/447) to 26.6% (119/447). The sensitivity persisted as 97.5% in EXPRESS (45/47) and 98.3% (117/119) in the North American cohort.

Conclusions: The specificity of the WINROP surveillance system for extremely preterm infants can be significantly improved by reassessment using the weight at WINROP alarm.