BACKGROUND:

Blood vessels in the retina of the eye provide a valuable window for evaluating quantitatively small vessel arteriosclerosis (arteriolar narrowing, arterio-venous nicking and arterio-to-venule ratio (AVR)). Previous studies have suggested that these optic fundi vascular parameters may be useful measures of generalized vascular disease and may have important prognostic implications. This study addressed this issue in a quantitative way using an established cohort with previously acquired fundic photographs. The study addressed an important public health problem with findings that may identify at-risk patients who would benefit from earlier intervention.

The established cohort was from the Beaver Dam Eye Study cohort, a well-characterized population of predominantly white persons aged 43-86 years at the baseline in 1988-1990. The cohort was originally evaluated in 1988-1990 with routine clinical and laboratory evaluation plus high-quality optic fundi photos. The population had been followed over a 10-year period with identified deaths from ischemic heart disease or stroke.

DESIGN NARRATIVE:

This was a population-based case-cohort study to determine whether retinal arteriolar changes (generalized narrowing, focal narrowing and arterio-venous nicking) and retinopathy were associated with 10-year stroke-and ischemic heart disease-related mortality. The study population was selected from participants of the Beaver Dam Eye Study, a well-characterized population of predominantly white persons aged 43-86 years at the baseline examination in 1988-90. Cases were defined as participants who had died from either stroke or ischemic heart disease since the baseline examination.

Three participants per case were selected from the cohort at baseline as controls, matched on gender and 5-year age intervals to cases. Focal arteriolar narrowing, arteriovenous nicking and retinopathy were graded to baseline using a standardized photographic grading protocol. To evaluate generalized arteriolar narrowing, a method modified from the Atherosclerosis Risk in Communities study was used. Retinal photographs were digitized and processed using a high-resolution scanner. Retinal arteriole and venule widths were then measured with the help of a computer program based on pixel density contrast between the vessel and the background retina. Finally, the average width of the arterioles was summarized as a retinal arteriole to venule ratio (AVR). Standard case-control methodology were applied to calculate the relative odds of association between generalized retinal arteriolar narrowing (using different categories of AVR) and other changes with stroke-and ischemic heart disease-related mortality. Logistic regression models controlling for potential confounders (e.g., blood pressure, serum lipid levels) were employed to evaluate the independent association between retinal arteriolar characteristics and stroke- and ischemic heart disease mortality.