Can a Risk Factor Based Approach Safely Reduce Screening for Retinopathy of Prematurity?

Abstract

Objective. Current American retinopathy of prematurity (ROP) screening guidelines is imprecise for infants ≥ 30 weeks with birth weights between 1500 and 2000 g. Our objective was to evaluate a risk factor based approach for screening premature infants at low risk for severe ROP. Study Design. We performed a 13-year review from Intermountain Health Care (IHC) data. All neonates born at ≤32 weeks were reviewed to determine ROP screening and/or development of severe ROP. Severe ROP was defined by stage ≥ 3 or need for laser therapy. Regression analysis was used to identify significant risk factors for severe ROP. Results. We identified 4607 neonates ≤ 32 weeks gestation. Following exclusion for death, with no retinal exam or incomplete data, 2791 (61%) were included in the study. Overall, severe ROP occurred in 260 (9.3%), but only 11/1601 ≥ 29 weeks (0.7%). All infants with severe ROP ≥ 29 weeks had at least 2 identified ROP risk factors. Implementation of this risk based screening strategy to the IHC population over the timeline of this study would have eliminated screening in 21% (343/1601) of the screened population. Conclusions. Limiting ROP screening for infants ≥ 29 and ≤ 32 weeks to only those with clinical risk factors could significantly reduce screening exams while identifying all infants with severe ROP.

Figure 1

Study participants flow diagram. Infants were excluded due to lack of ROP screening data because of death, discharge, transfer, or ROP exam deemed unnecessary. Infants with severe ROP were categorized by gestational and/or birth weight group and then again by those in each group treated with laser surgery.

Figure 2

Receiver operating characteristic (ROC) curves showing sensitivity and specificity of predictive modeling for both diagnosis of severe ROP (a) and for treatment with laser therapy (b) using either the logistic regression model or the ROP Risk Factor Score.