Testing the use of retinal microvascular parameters to identify individuals with type 2 diabetes with mild cognitive impairment

Abstract

Objective To investigate the use of retinal microvascular parameters, as determined by color fundus photographs (CFP), to detect individuals with type 2 diabetes with mild cognitive impairment (MCI). Methods The current study is based on cross sectional data obtained prospectively in RECOGNISED, a European Union-funded multicenter observational study involving individuals 65 years of age or older with type 2 diabetes. The diagnosis of MCI was established based on cognitive screening, followed by a detailed neuropsychological test battery. CFP were analyzed to quantify retinal vessel width, arteriovenous ratio, tortuosity, and fractal dimension. Crude and multivariable logistic regression models were used to investigate the association between MCI and retinal microvascular parameters and the ability of the latter to identify individuals with MCI using both eyes in our analysis if eligible. Results We included 578 eyes of 307 individuals with type 2 diabetes of whom 182 had MCI and 125 had normal cognition. Median age was 72 (IQR 69–77) years, 65 % were male, and median diabetes duration was 19 (IQR 12–24) years. We found arteriolar tortuosity was associated with MCI in the multivariable analysis (OR 1.44 [95 % CI 1.08–1.92]; the less the tortuosity the higher the risk). When estimating the discriminative ability of the retinal microvascular parameters, we found an AUC of 0.73. Conclusion In people with type 2 diabetes, a lower retinal arteriolar tortuosity was associated with MCI. In addition, we found retinal microvascular parameters to have an acceptable discrimination ability when paired with patient’s characteristics.