Retinal vessel caliber and cognitive performance: the multi-ethnic study of atherosclerosis (MESA)

Abstract Retinal vessel calibers share anatomic and physiologic characteristics with the cerebral vasculature and can be visualized noninvasively. In light of the known microvascular contributions to brain health and cognitive function, we aimed to determine if, in a community based-study, retinal vessel calibers and change in caliber over 8 years are associated with cognitive function or trajectory. Participants in the Multi-Ethnic Study of Atherosclerosis (MESA) cohort who completed cognitive testing at Exam 5 (2010–2012) and had retinal vascular caliber measurements (Central Retinal Artery and Vein Equivalents; CRAE and CRVE) at Exam 2 (2002–2004) and Exam 5 were included. Using multivariable linear regression, we evaluated the association of CRAE and CRVE from Exam 2 and Exam 5 and their change between the two exams with scores on tests of global cognitive function (Cognitive Abilities Screening Instrument; CASI), processing speed (Digit Symbol Coding; DSC) and working memory (Digit Span; DS) at Exam 5 and with subsequent change in cognitive scores between Exam 5 and Exam 6 (2016–2018).The main effects are reported as the difference in cognitive test score per SD increment in retinal vascular caliber with 95% confidence intervals (CI). A total of 4334 participants (aged 61.6 ± 9.2 years; 53% female; 41% White) completed cognitive testing and at least one retinal assessment. On multivariable analysis, a 1 SD larger CRAE at exam 5 was associated with a lower concomitant CASI score (− 0.24, 95% CI − 0.46, − 0.02). A 1 SD larger CRVE at exam 2 was associated with a lower subsequent CASI score (− 0.23, 95%CI − 0.45, − 0.01). A 1 SD larger CRVE at exam 2 or 5 was associated with a lower DSC score (− 0.56, 95% CI − 1.02, − 0.09) and − 0.55 (95% CI − 1.03, − 0.07) respectively. The magnitude of the associations was relatively small (2.8–3.1% of SD). No significant associations were found between retinal vessel calibers at Exam 2 and 5 with the subsequent score trajectory of cognitive tests performance over an average of 6 years. Wider retinal venular caliber was associated with concomitant and future measures of slower processing speed but not with later cognitive trajectory. Future studies should evaluate the utility of these measures in risk stratification models from a clinical perspective as well as for screening on a population level.