Can Surrogates Predict VA for Patients With Down Syndrome?

Journal Highlights

Translational Vision Science & Technology
2023;12(9):11

Download PDF

Individuals with Down syndrome (DS) have unique visual systems and a high risk of refractive errors, strabismus, nystagmus, and severe higher-order aberrations. The intellectual disability that accompanies DS can hinder the vision-testing process. Therefore, it would be prudent to develop methods to judge image quality on behalf of patients with DS. Schneider et al. aimed to determine whether control observers without DS could serve as surrogates to predict VA gains for patients with DS. Although their results indicated a possible trend toward habitual corrections having the worst performance and metric-optimized corrections having the best, the acuity gains for metric-optimized corrections as interpreted by the control observers were much greater than the actual gains.

Thirty adults with DS participated. The refraction methods tested were standard clinical, pupil fraction tessellated optimized, and visual Strehl ratio (VSX) optimized. Monocular VA was obtained from habitual refraction and each experimental refraction method. Five control observers without DS, whose BCVA was 20/20 or better, viewed acuity charts simulating the retinal image produced for each DS eye by each correction method. The observers provided input on VA and ranked the quality of each chart, noting perceptions of ghosting, blur, and contrast. Their assessments were used to identify aspects of image quality related to VA performance. The group median VA (DS vs. controls) was compared for each refraction type, and the quality scores given by controls were compared with corresponding VA results across refraction methods.

In the DS cohort (mean age, 29 years), the refractive error based on dry auto­refractor readings was large (−15.25 to +6.00 D), and all but two participants had −1.00 D cylinder power or worse in at least one eye. Nine had strabismus. Median VA of the DS group ranged from 0.46 logMAR (habitual correction) to 0.36 logMAR (VSX correction). The median VA as judged by the observers ranged from 0.37 logMAR (habitual) to 0.01 logMAR (VSX). In general, image quality was best for metric-optimized methods and showed a strong linear correlation with control-assessed VA (r = −0.91; p < .001) and a weaker correlation with DS VA (r = −0.33; p < .001).

Even though VA gains could not be determined by observers, the VA predictions and image quality were markedly better with metric-optimized corrections. Hence, the authors sur­mised that nonoptical factors may contribute to the VA deficits of people with DS.

The original article can be found here.