Risk Assessment Model May Predict Advanced AMD

A risk assessment model may predict the development of advanced age-related macular degeneration (AMD), according to the results of a longitudinal analysis reported Online First August 8 in the Archives of Ophthalmology. "As progress in designing better preventive measures and earlier treatment options accelerates and new gene associations are identified that add to currently known risk factors, the desirability of having a reliable risk assessment model has become of considerable interest and potential value," write Michael L. Klein, MD, from the Casey Eye Institute, Oregon Health & Science University, Portland, and colleagues. "The optimal design might include known demographic and environmental risk factors, phenotypic risk factors derived from large population-based and interventional studies, and established genetic risk variants. The purpose of this article is to present a validated predictive model for AMD that incorporates these factors and can be used by the practicing physician."
The study sample consisted of 2846 participants in the Age-Related Eye Disease Study who had all levels of AMD at baseline, ranging from none to unilateral advanced AMD, defined as neovascular or geographic atrophy.
Cox proportional hazards analysis allowed evaluation of demographic, environmental, phenotypic, and genetic covariates and construction of a risk assessment model for the development of advanced AMD during follow-up (average duration, 9.3 years). The investigators used the C statistic and the Brier score to assess performance of the model, and they externally validated the model among participants in the Complications of Age-Related Macular Degeneration Prevention Trial.
Among participants with advanced AMD at baseline, 82% of those with geographic atrophy and 56% of those with neovascular AMD went on to have advanced AMD in the other eye.
Independent variables used in the final risk assessment model were age, smoking history, family history of AMD in a first-degree relative, phenotype based on a modified Age-Related Eye Disease Study simple scale score, and the genetic variants CFH Y402H and ARMS2 A69S.
Performance of the model was good, with a C statistic of 0.872 suggesting very good discrimination, and a Brier score of 0.08 at 5 years indicating excellent calibration and overall performance. The model was successfully validated in the external sample, and the investigators designed a risk assessment tool that could be used with or without the genetic component.
"We constructed a risk assessment model for development of advanced AMD," the study authors write. "The model performed well on measures of discrimination, calibration, and overall performance and was successfully externally validated. This risk assessment tool is available for online use."
Limitations of this study include an exclusively white study sample (ages 50 - 85 years) that was not derived from a population-based sample.
"We believe our current model is of substantial value in assessing AMD risk, and we expect that future advances will further improve its accuracy," the study authors conclude. "Unexplained heritability of AMD will be uncovered, studies on diet, other environmental factors, and serum biomarkers may identify new predictive factors, and better phenotyping methods are under development. As these new findings become available, we plan to update the model and maintain a current version for online use."
The Casey Eye Institute Macular Degeneration Fund; Research to Prevent Blindness, New York; the Bea Arveson Macular Degeneration Fund; and the Foundation Fighting Blindness, Columbia, Maryland, supported this study. A US patent entitled "Nutritional Supplement to Treat Macular Degeneration" was issued on December 9, 2003, and is owned by Bausch and Lomb. Study author Frederick L. Ferris III, MD, is one of the inventors; he has assigned his interest in the patent to the US government and receives government compensation.