Purchase this article with an account.
Charlotte Joslin, Thasarat Vajaranant, Lawrence Ulanski; Age-related macular degeneration and mortality in AREDS: the effect of sex and time. Invest. Ophthalmol. Vis. Sci. 2013;54(15):245.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Age-related macular degeneration (AMD), stroke and cardiovascular disease (inclusive of all circulatory disease) share risk factors and a similar pathogenesis. Previous studies yield mixed associations, potentially due to differences in follow-up time that minimize sample size and study power or alter competing risks, as well as an absence of sex-stratified results masking sex/gender differences.
The Age-Related Eye Disease Study (AREDS) dataset was used for secondary data analysis. Standard AREDS categories (1, 2, 3 and 4) were used to assess the hazard of baseline AMD status on all-cause and circulatory disease mortality. Cause of death was based on hospital records and death certificates collected during AREDS (11/92 - 12/05), and a search of the National Death Index (NDI) after the study conclusion. The proportionality hazard assumption was tested by including time-dependent covariates in the model using PROC PHREG to assess the hazard ratio (HR) and 95% confidence interval (95% CI) for the effect of AMD on mortality in whites (SAS v9.3, Cary, NC). Multivariable Cox Proportional Hazard regression models were fit adjusting for covariates and stratifying on non-proportional predictors to estimate the HR overall at multiple time points, including survival at 5-, 7-, and 10-years for 1) any death; and, 2) death due to circulatory disease.
The mean follow-up among 4116 subjects was 9.6 years, with 967 of 1096 deaths identified during AREDS (Table 1). Regression models (Table 2) generally show a significant HR between AMD and circulatory system death, in which the magnitude of association is stronger 1) in men than women; 2) with increasing AMD severity; and, 3) with shorter vs. longer survival analysis. Among all-cause mortality, a similarly increased HR, dampened vs. circulatory deaths, exists that is stronger in men and increases with increasing AMD severity; HR differences by survival duration are less apparent.
Results suggest a significant association between increasing AMD severity and mortality. Variation in the strength of association exists by sex for both all-cause and circulatory disease mortality. Convergence in the HR between AMD and mortality by length of survival analysis comparing circulatory disease vs. all-cause mortality suggests the importance of competing risks in longer survival analyses.
This PDF is available to Subscribers Only