Abstract
Purpose :
Distance visual acuity (VA) testing is commonly performed in children to screen for significant refractive error; however, the diagnostic value of distance VA remains unclear. We examined associations between VA and different types of cycloplegic refractive error in children.
Methods :
Unaided distance VA (logMAR) was assessed monocularly in 234 children using single-surround ATS-HOTV (5-6 years) or eETDRS (7-9 years) protocols. Better-eye VA was used for analyses. Cycloplegic refractive error was determined using the Grand Seiko WR-5100K autorefractor. Refractive error was classified as emmetropia (EMM; -0.75D< sphere (sph) <+0.75D & cylinder (cyl) <0.75D), simple myopia (SM; sph≤-0.75D & cyl<0.75D), simple hyperopia (SH; sph≥+0.75D & cyl<0.75D), compound myopic/hyperopic astigmatism (CMA/CHA; both meridians myopic/hyperopic & cyl≥0.75D), simple myopic/hyperopic astigmatism (SMA/SHA; one myopic/hyperopic & one emmetropic meridian), or mixed astigmatism (MA; one hyperopic & one myopic meridian). Receiver operating characteristic (ROC) analysis was conducted to identify optimal logMAR VA thresholds for the detection of each refractive error type (if N>10) using the Youden Index.
Results :
VA ranged from -0.3 to 1.2 logMAR (median=0.0, IQR=0.3), SE from -3.15D to +10.45D (mean=+1.00D, SD=1.79D), and astigmatism from 0.03D to 4.24D (mean=0.75D, SD=0.75D). Refractive error distribution was SH (N=90), EMM (N=52), SHA (N=27), CHA (N=26), SM (N=16), SMA (N=13), CMA (N=7), and MA (N=3). Analyses found low optimal logMAR VA thresholds across refractive error types (SM=0.15, SMA=0.05, SH=-0.25, SHA=-0.05, CHA=0.05), with high sensitivity (SM=94%, SMA=85%, SH=100%, SHA=85%, CHA=81%) but low specificity (SMA=63%, SH=1%, SHA=39%, CHA=66%) except for SM (79%). Analyses were repeated with clinically significant hyperopia (sph≥+2.00D; SH: N=13, CHA: N=13), also yielding low optimal thresholds (SH=-0.25, CHA=0.05), high sensitivity (SH=100%, CHA=92%), and low specificity (SH=2%, CHA=55%). ROC analysis on any clinically significant refractive error type showed a low optimal logMAR threshold (0.05) with 78% sensitivity and 87% specificity.
Conclusions :
Distance VA had poor diagnostic value in the detection of clinically significant refractive error in children 5 to 9 years of age. Optimal logMAR VA thresholds based on the Youden Index were not informative in discriminating refractive error type.
This abstract was presented at the 2022 ARVO Annual Meeting, held in Denver, CO, May 1-4, 2022, and virtually.