Abstract
Purpose :
Recent cross-sectional studies have suggested a positive association between reticular pseudodrusen (RPD) presence and cardiovascular disease (CVD) in individuals with age-related macular degeneration (AMD). We longitudinally evaluated the risk of CVD events, stroke, and related death, according to RPD status, in the Age-Related Eye Disease Study (AREDS) and AREDS2.
Methods :
Post hoc analyses, where AREDS subjects were chosen for similarity to AREDS2 (bilateral large drusen or late AMD in one eye and large drusen in the other). RPD status was determined by reading center grading of fundus autofluorescence images (AREDS2) and by deep learning grading (DLG) of color fundus photographs (CFP) (AREDS). CVD events were determined prospectively using International Classification of Diseases codes (AREDS) and Cardiovascular Outcome Study procedures (AREDS2). Proportional hazards regression was performed for outcomes, according to RPD status, while adjusting for age, sex, BMI (in AREDS), smoking, diabetes, and baseline AMD simplified severity scale.
Results :
The study comprised 1290 (AREDS) and 1321 (AREDS2) subjects. Mean age was 71 and 73 years, with 56.4% and 57.2% female, respectively. RPD were present at baseline in 18.3% and 28.2%, respectively. Over mean follow-up of 9 and 3 years, respectively, the proportions of subjects with CVD events, stroke, and related deaths were 39.4%, 10.9%, and 12.2% (AREDS), and 4.9%, 1.4%, and 0.6% (AREDS2), respectively. In AREDS, RPD presence was nominally associated with higher CVD event risk (hazard ratio (HR), 1.41, 95% CI 1.14-1.75), but not stroke (HR 0.89, 0.56-1.41), or death (HR 1.30, 0.88-1.91). In AREDS2, RPD presence was not associated with increased CVD event risk (HR 1.23, 0.72-2.11). Stroke and death (19 and 8 events, respectively) were not analyzed. In sensitivity analyses of AREDS2, RPD status determined by the DLG of CFP (n=4105) was not associated with higher risk of CVD events (HR 1.05, 0.83-1.32) or death (HR 1.17, 0.55-2.49).
Conclusions :
Overall, the findings are not consistent with a large association between RPD presence and CVD events. This study benefitted from comprehensive longitudinal analyses, rigorous adjudication of CVD events, and adjustment for relevant covariates in statistical analyses that took time into account in two large AMD studies with clear eligibility criteria and long follow-up.
This abstract was presented at the 2023 ARVO Annual Meeting, held in New Orleans, LA, April 23-27, 2023.