Determinants of reading performance in geographic atrophy secondary to age-related macular degeneration

Abstract

Purpose : To identify determinants of reading performance in geographic atrophy (GA) secondary to age-related macular degeneration as a prerequisite towards the application for reading performance as endpoint in clinical trials.

Methods : Best corrected visual acuity(BCVA) as logarithm of the minimum angle of resolution(LogMAR), low luminance visual acuity(LLVA[LogMAR]) and reading acuity(LogRAD) were assessed using the early treatment diabetic retinopathy study and Radner charts, respectively. Fundus autofluorescence, near-infrared reflectance and spectral-domain optical coherence tomography images were acquired to measure GA size(square root-transformed)(Spectralis, Heidelberg Engineering, Germany). Fixation stability(95%bivariate contour ellipse area[BCEA; log10-transformed]) was monitored using fundus-controlled perimetry(microperimetry). Linear mixed effect models were used in the analysis to consider the hierarchical nature of the data(eyes nested in patients).

Results : A total of 177eyes of 95patients [mean±SD] 78.0±7.2years with GA from the prospective natural history study DSGA(Directional Spread in Geographic Atrophy; NCT02051998) were included in the analysis. At baseline, the median [Q1;Q3] BCVA was 0.32LogMAR[0.2;1.0] with reading acuity 0.91[0.4;1.3] and square root GA area 3.09mm[2.07;3.94]. In the univariate analysis, BCVA, foveal sparing, LLVA, GA area (p<0.001 each) and BCEA (p=0.005) exhibited a significant effect on reading acuity. Using backwards selection, BCVA [effect estimate±standard error] [0.66±0.06LogRAD/LogMAR], presence of foveal sparing [-0.13±0.06LogRAD], LLVA [0.12±0.06LogRAD/LogMAR] and GA area [0.06±0.02LogRAD/mm] were identified to be significant predictors of reading acuity. Binocular reading acuity was significantly influenced by both, the better [0.87±0.04LogRAD/LogRAD] (p<0.001) and worse eye [0.13± 0.05LogRAD/LogRAD] (p=0.008).

Conclusions : In descending order in respect of the effect size, reading acuity depends on BCVA, presence of foveal sparing, LLVA and GA area. Hence, the rather low association between GA area and reading acuity highlights the challenge that both endpoints would demonstrate a therapeutic effect in the same direction. Therefore, a differential interpretation of (i) structural endpoints to substantiate biological efficacy and (ii) visual function endpoints to estimate patient-relevant treatment effects in clinical trials appears prudent.

This abstract was presented at the 2019 ARVO Annual Meeting, held in Vancouver, Canada, April 28 - May 2, 2019.