Abstract
Purpose: :
While early intervention with anti- VEGF drugs can significantly reduce visual loss from neovascular (wet) age-related macular degeneration (AMD), many patients do not present for treatment until months after conversion. We examined the likelihood of an eye developing neovascular AMD by retrospectively evaluating longitudinal changes in the morphological features of macular fundus photos and best corrected visual acuity.
Methods: :
A semi-automated analysis of fundus photos was conducted on 834 eyes, (513 subjects) enrolled in the Age-Related Eye Disease Study (AREDS) and Prophylactic Treatment of Age-related Macular Degeneration (PTAMD) at a single center. Images were assessed for specific morphological features in the central 1000 and 3000 micron diameter circles centered at the foveola. Parameters included macula hyperpigmentation, drusen area, and the number of small, intermediate, and large drusen. Other factors considered were age, days of follow-up, presence of CNV in fellow eye at baseline, and ETDRS best-corrected visual acuity (VA) measured by a rigorous protocol. Subjects were followed from 12 days to 8 years (mean 3.4 years). We used a longitudinal logistic regression with random intercepts model fitted to event occurrences to study the entire population of eyes over time. This model differs from one based on generalized estimating equations (GEE) and allowed us to include all eligible eyes in aggregate, regardless of whether the eyes were from unilaterally or bilaterally eligible subjects.
Results: :
33 of the 513 (6.4%) subjects developed CNV, with VA and aging variables showing a significant correlation. A three- letter drop in ETDRS VA occurring over time predicted a 50% increase in the probability for CNV developing in that eye (1.50, p < 0.0001). Statistical analysis showed that for single-letter drop in VA was associated with a 12% (1.12, p < 0.001) increase in the possibility of CNV event. Also there was a 0.3% increase each day the eye was in the study (aging). There was no difference in this effect if data from the 2 studies were assessed independently. We found that changes in drusen morphology offered no predictive value for eminent conversion to wet AMD.
Conclusions: :
Our analysis suggests that decrease in carefully measured VA over time increases the possibility of CNV event. Changes in morphological abnormalities showed no predictive value but this may be due to the limited number of CNV events.
Keywords: age-related macular degeneration • clinical (human) or epidemiologic studies: risk factor assessment • retina