Computer-assisted measurement of retinal vessel diameters in the Beaver Dam Eye Study: methodology, correlation between eyes, and effect of refractive errors

Abstract

Objective Narrowed retinal arterioles may independently predict cardiovascular disease. We describe a computer-assisted method to measure retinal vessel diameters from digitized fundus photographs in a large population, and examine the correlation of retinal vessel diameters between eyes and whether refractive errors affect these measurements. Design and population Population-based, cross-sectional study in Beaver Dam, Wisconsin (n = 4926; age, 43–84 years). Methods Retinal photographs were digitized, and all arterioles and venules located in an area 0.5 to 1 disc diameter from the optic disc were measured with the aid of computer software. Main outcome measures Summary of retinal arteriolar and venular diameters, and the ratio of their diameters (arteriole:venule ratio [AVR]). Results Correlation between right and left eyes was substantial for retinal arteriolar diameters (Pearson correlation coefficient, ρ = 0.71) and venular diameters (ρ = 0.74), and moderate for the AVR (ρ = 0.49). The inverse association of higher blood pressure and smaller retinal arteriolar diameters was similar using data from either one eye or two eyes; arteriolar diameters decreased by 4.1 μm (right eyes), 4.0 μm (left eyes), and 4.0 μm (mean of both eyes) with each 10-mmHg increase in mean arterial blood pressure. A myopic refraction was associated with smaller retinal vessel diameters; arteriolar diameters decreased by 2.8 μm and venular diameters by 3.3 μm with each −1.0-diopter shift towards myopia. However, the pattern and strength of the association of blood pressure and retinal vessel diameters were not altered by variations in refractive errors. Conclusions There is good correlation of retinal vessel diameters between eyes. A myopic refraction is associated with smaller retinal vessel diameters. The association of retinal arteriolar diameters and blood pressure seems similar using data from either one eye or two eyes, and is minimally affected by refraction. These data suggest that measurement of retinal vessel diameters from one eye without regard to its refractive status may provide adequate information indicative of a person's retinal vessel caliber if information from two eyes and refraction is unavailable.