Abstract
Abstract: :
Purpose: To assess the components of variability in retinal vessel measurements controlling for pulsation Methods: Digital images of the fundus, centered on the optic disc, were taken at 3 distinct times during the cardiac cycle of a 19 year old man. The pulse sensor was on the earlobe. Ten images were taken at 0%, 30%, and 70% delay from the onset of the pulse cycle over a one-hour period. Two trained graders measured the widths of 1 large retinal arteriole, 1 large retinal venule, 1 small retinal arteriole, and 1 small retinal venule. The two graders measured these retinal vessels in each image 10 times using semi-automated computer software. Results: The mean retinal vessel widths were 77.0µ, 92.6µ, 113.2µ, and 144.1µ for the small arteriole, small venule, large arteriole, and large venule, respectively. The large retinal venule had the smallest intragrader variability within the same image yielding an average standard deviation of 0.95µ, compared to 1.04µ, 1.09µ, and 1.12µ in the small venule, large arteriole, and small arteriole, respectively. Among images taken at the same point in the pulse period, the percent change from the minimum to maximum size was between 6% and 17% for arterioles and between 2% and 11% for venules. Small vessels had bigger percent changes than large vessels. Also, no point in the pulse period was more variable than any other point. Mixed-Effect models were fit for each of the vessel types to determine the greatest source of variability. Grader and pulse period were considered fixed effects. From these models, the largest source of variability for all 4 vessels measured was between images, accounting for over 50% of the total amount of variability. Conclusions: Grading of retinal venules is generally less variable than retinal arterioles. Further, grading of larger vessels is less variable than grading of smaller vessels. After controlling for pulsation (a known source of variation from previous research), the largest source of variation is due to the variability between images. Understanding the components of variability in grading retinal vascular caliber is important as these techniques are applied in epidemiological studies.
Keywords: clinical research methodology • clinical (human) or epidemiologic studies: bio • retina