Purpose
To investigate the progression rate and pattern of incident geographic atrophy (GA) associated with age-related macular degeneration.
Methods
Subjects included were those with bilateral large drusen who developed GA in the untreated eye during the course of the CAPT study. Stereoscopic color fundus photographs and fluorescein angiograms were reviewed from eyes that developed GA during the course of the CAPT study. Size, location, and number of distinct areas of atrophy were noted at each annual study visit. Analyses were performed to determine the average enlargement rate and factors affecting the progression of new areas of GA. The mean total size of GA at baseline in this study was 0.58 mm2, signifcantly smaller than those included in previous studies investingating the growth of GA.
Results
Incident GA remains unifocal in 43% and extrafoveal in 79% of subjects throughout the first 3-5 years of follow-up. When GA does not initially involve the central fovea, it encroaches on the fovea at a very slow rate (0.09 mm/year). Central visual acuity is not significantly affected early in the course, nor are basic visual function parameters, such as reading speed and contrast sensitivity. Early GA enlarges at an average rate of 0.49 mm2/year, though the rate of growth is highly variable amongst subjects and is dependent on baseline size. By using a square-root transformation, the growth rate of early GA is 0.20 mm/year. In contrast to previous reports, this square root transformation failed to eliminate the dependence on initial size. Use of a log-transformed growth model fit the data best, and yielded an enlargement rate of 56% per annum.
Conclusions
Early geographic atrophy progresses slowly and often remains unifocal and parafoveal thoughout the first 3-5 years. Our reported average growth rate (0.49 mm2/yr) is significantly slower than previous reports, which relied on larger areas of GA. This may be on account of the smaller size of early GA lesions, as we found growth rates based on the linear and square root models to be dependent on baseline size. Our data suggest that a log-transformed growth model best characterizes the growth of GA and eliminates the dependence on baseline size. To our knowledge, no previous studies have used a log-transformation for comparison. Using this model, our data suggest that early geographic atrophy grows at a rate of 56% per annum.