Retinal microvascular abnormalities and cognitive function in older people with type 2 diabetes
The deleterious effects of Type 2 diabetes on the brain have been shown to result in a greater prevalence of age-associated cognitive impairment and an enhanced risk of age-related cognitive decline in older diabetic populations. Type 2 diabetes is a complex metabolic disorder. Apart from the negative impact of abnormalities intrinsic to diabetes, diabetes-associated cerebral microvascular disease may contribute to this accelerated cognitive ageing. Direct in vivo evaluation of the cerebral microcirculation is difficult in humans and the vessels themselves are too small to permit detailed visualisation with current neuroimaging methods. The microvasculature of the retina may offer a window into such vascular status of the brain as there is considerable homology between the retina and cerebral microcirculations. Moreover, the retinal vasculature is known to be affected by a wide range of systemic pathologies and is unique in that it is the only vasculature that can be directly visualised and photographed. Retinal microvascular abnormalities (RMAs) have been understudied risk factors in cognitive ageing epidemiological research. Few reports have comprehensively examined cognitive function in relation to diabetic retinopathy. Also the relationship between cognitive function and quantitative aspects of retinal vascular network geometry has not been investigated in people with Type 2 diabetes. The results of a systematic review reported in this thesis showed inconsistent findings on the importance of the association between retinal microvascular abnormalities and cognitive dysfunction in predominantly non-diabetic populations. This may have reflected substantial differences between studies regarding the choice of population under study, the methods applied for measuring and defining RMAs, the types of neuropsychological tests administered for assessing cognitive function, and the approach taken in data analysis. The principal aim of the original research described in this thesis was to examine the associations of cognitive test performance with severity of diabetic retinopathy and quantitative parameters of retinal vascular network in a population-based sample of older people with Type 2 diabetes. Objective, reproducible and computerized retinal image analysis was used to quantify retinal vessel calibres and arteriolar bifurcation geometry in order to detect subtle changes in retinal vascular network. A valid estimation of peak prior cognitive ability allowed the further exploration of the impact of retinal microvascular abnormalities on imputed cognitive decline from best-ever levels of cognitive function to that measured in old age. The analysis was based on a cohort of 547 men and 519 women aged 60-75 years with Type 2 diabetes, randomly sampled from the Lothian Diabetes Register, Scotland, in 2006/2007 (the Edinburgh Type 2 Diabetes Study). A battery of seven cognitive tests was administered and standard 7-field binocular digital retinal photography undertaken. The Mill Hill Vocabulary Scale was used to estimate pre-morbid cognitive ability. Diabetic retinopathy was evaluated independently by two optometrists using a standardised grading protocol (a modification of the Early Treatment of Diabetic Retinopathy Scale). Quantitative retinal vascular parameters were measured by myself from a digital image of field 1 using semi-automated, computer-based methods. Retinal vessel calibres were summarised as the central retinal arteriolar and venular equivalents (CRAE and CRVE, respectively) and arterio-venous ratio (AVR). Retinal arteriolar bifurcation geometry was expressed as arteriolar bifurcation angles (BA), arterial branching coefficient (BC), and sub-optimality (degree of deviation from optimality) of the retinal arteriolar angles. The statistical analyses were based on the 1,044 study participants who had both gradable retinal images and cognitive testing. Both general cognition, as indexed by a general cognitive factor reflecting the variance common to all the cognitive tests used, and most of the individual cognitive tests were negatively affected in participants with diabetic retinopathy relative to those without. These cognitive measures also showed a significant relationship with increasing severity of diabetic retinopathy (none, mild, and moderate-severe). Those with moderate-severe diabetic retinopathy had worst performances on general cognitive function, executive function, information processing speed, non-verbal memory and mental flexibility. When lifetime decline was estimated from peak, prior cognitive level, severity of diabetic retinopathy was significantly associated with a greater decline in information processing speed, non-verbal memory and mental flexibility and, in men for general cognition and executive function. The associations of severity of diabetic retinopathy with general cognition, executive function and information processing speed were independent of socio-demographic characteristics, cardiovascular risk factors, macrovascular disease, mood and hyperglycaemia. The associations with estimated decline in specific cognitive measures resulted principally from the impact of diabetic retinopathy on general cognitive ability. The study also showed that larger retinal arteriolar and venular calibres were both significantly associated with lower test scores on verbal memory in men. Multiple linear regression analyses demonstrated larger retinal arteriolar calibre was associated with a significantly greater decline in verbal memory after possible confounding by retinal venular calibre and vascular risk factors and disease was taken into account. In contrast, the study did not support an independent association between retinal venular calibre and cognitive decline in men or in women with Type 2 diabetes. Parameters of retinal arteriolar bifurcation geometry were not associated with cognitive outcome. Overall, these findings support the hypothesis that cerebral microvascular disease associated with Type 2 diabetes, reflected by the presence and severity of diabetic retinopathy, may exacerbate the effects of ageing on cognitive function. In particular, alterations in the blood-brain barrier may be an important pathophysiological mechanism in the occurrence of cognitive dysfunction in diabetic patients. They further may be added to the knowledge that gained from previous pathologic and brain imaging investigations demonstrating a relationship between markers of cerebral microvascular disease and cognitive dysfunction in diabetes. The role of quantitative parameters of retinal vascular network geometry in diabetes-related cognitive impairment is less clear. Prospective studies are required to clarify the temporal sequence of these associations and the eventual clinical significance of these small, early cognitive function changes. Such a follow-up project involving the present study population is underway. From a clinical perspective, if the above findings are substantiated, diabetes-associated cognitive dysfunction may be amenable to treatment and preventive strategies specifically targeted at protecting the cerebral microvasculature and reducing the risk of developing even mild microvascular disease in an ageing diabetic population.