Progression of Geographic Atrophy with Subsequent Exudative Neovascular Disease in Age-Related Macular Degeneration

Written by: Mehnaz Khan, MD

March 2021

Hwang CK, Agrón E, Domalpally A, Cukras CA, Wong WT, Chew EY, Keenan TDL; AREDS2 Research Group. Progression of Geographic Atrophy with Subsequent Exudative Neovascular Disease in Age-Related Macular Degeneration: AREDS2 Report 24. Ophthalmol Retina. 2021 Feb;5(2):108-117. doi: 10.1016/j.oret.2020.10.008. Epub 2020 Oct 16. PMID: 33075546; PMCID: PMC7870515.

One of the more frustrating diagnoses for a vitreoretinal surgeon to give his/her patient with age-related macular degeneration (AMD) is that of geographic atrophy (GA). GA can progress over time, but the rate of progression/enlargement greatly varies between patients and has been shown to be associated with various clinical and genetic factors. At present, there are no clinically approved therapies for treatment or stabilization of GA. Gaining a better understanding about the pathophysiology of GA enlargement can help in development of therapies treatments to halt its progression. To that end, the purpose of this study (AREDS2 report 24) was to determine whether the presence of exudative neovascular age-related macular degeneration (nAMD), and indirectly, nonexudative nAMD, may slow GA enlargement.

This study is a post hoc analysis of the AREDS2 (a multicenter, phase 3, randomized control trial designed to assess the role of nutritional supplements in progression of AMD) clinical trial data. The design of the study is based on the well-accepted definition of nAMD as the presence of macular neovascularization (MNV), which can subsequently be complicated by exudation or development of fibrovascular scar. The advent of OCT angiography (OCTA) has helped us greatly in understanding the progression of MNV; it has shown that in most cases of MNV there likely exists a prolonged and variable period of nonexudation before the onset of exudation. Since OCTA was not available as an imaging modality during the design of this study, development of exudative AMD was used as a proxy for presuming the prior presence of nonexudative MNV. The AREDS2 trial had a 5-year follow-up time-point during which eyes with GA showed presence of MNV, allowing an opportunity to determine GA progression in the setting of MNV. Overall, the purpose of this study was to use the AREDS2 dataset to determine whether the subsequent development of exudative MNV (and indirectly nonexudative AMD) influenced the rate of GA progression.

Participants in this post-hoc analysis included participants of the AREDS2 trial, 50-85 years of age. Baseline and annual color fundus photographs and fundus autofluoresence (FAF) images were obtained and graded centrally at the Fundus Photograph Reading Center. Standardized definitions of GA were used while reading the photographs. Additionally, exudative nAMD was defined as a history of treatment for nAMD and the presence of two of the following features on fundus photography – serous detachment of the retina, retinal hemorrhage, RPE detachment, fibrovascular tissue, or hard exudates. Having identified patients with GA from the AREDS2 trial, this study constructed two separate cohorts: 1. Prevalent cohort (patients who had GA at time of baseline assessment without presence of prior or current nAMD) 2. Incident cohort (patients who developed GA during the study follow-up period without presence of prior or current nAMD). Overall, 456 eyes qualified for the prevalent cohort, and 757 eyes qualified for the incident cohort.

Subsequently mixed-model regression was performed to determine correlation between the square root transformation of GA area in the presence or absence of nAMD. This was done separately for the prevalent and incident cohort. Additionally for each cohort, the baseline demographic (such as age, gender, smoking status), clinical (such as first appearance of GA on fundus photos, central involvement, configuration, number of eyes with GA) and genetic ( 3 single neucleotide polymorphisms) characteristics were compared between the groups with and without subsequent development of nAMD.

Differences in the baseline characteristics for groups in the prevalent cohort were not statistically significant. In the incident cohort however, the proportion of former smokers was higher in the group with subsequent nAMD (p=0.02). As for GA progression rates, in the incident cohort, GA enlargement was found to be significantly slower in eyes with subsequent nAMD development when compared to eyes without subsequent nAMD development (p=0.037). However, no significant correlation was noted in the prevalent cohort, whose eyes showed overall larger GA lesions. One explanation for this may be the idea that MNV can slow growth of GA in its local vicinity which allows for detectable showing of growth of smaller GA lesions of the incident cohort but no detectable slowing of larger GA lesions of the prevalent cohort.

The authors offer the following explanation as a potential mechanism via which MNV slows progression of GA lesions: Type I MNV penetrates the Bruch’s membrane and provides a capillary rich network in in the sub-RPE space which in turn results in efficient delivery of nutrients and oxygen to the RPE and choroid. This likely compensates for areas of choriocapillaris loss which is known to play a role in GA progression. Overall, the results from this study are in line with prior studies suggesting association between presence of nonexudative MNV and slower GA enlargement.

One of the strengths of study is the large sample size of eyes with GA followed by nVAMD development (136 eyes in both cohorts). Yet another strength is the prospective nature of the data acquisition with standardized criteria for review of all the fundus photographs by masked readers at the reading center. Additionally, this study was able to account for various baseline characteristics (demographic, clinical and genetic) that have been shown in the past to play a role in GA progression/enlargement.

One of the important limitations of this study acknowledged by the authors is the use of exudative MNV as a proxy for presence of nonexudative NAMD. OCTA was not available during the AREDS2 trial to confirm this assumption. By the same token, the duration of presence of nonexudative NAMD for this study was unknown. However, studies such as this one are very important for understanding the pathophysiology of GA progression that will not only play a vital role in development of therapeutic interventions but also help clinicians in determining long term prognosis while counseling patients with GA.