UC Berkeley

Over last three decades, diabetes emerged as a leading cause of vision loss. Diabetic retinopathy and diabetic macular edema (DME) are common retinal complications in patients with diabetes. These end-organ complications have a sudden, and debilitating impact on vision eventually leading to blindness. The present standard of care for management and treatments of these complications are inherently destructive, associated with unavoidable side effects, and not universally effective in reversal of visual loss. They are expensive and invasive late stage treatments that are administered once structural retinal damage is clinically obvious. Early detection and better understanding of early stages of diabetic retinopathy is crucial to delay the onset and to slow progression of diabetic retinopathy.

Predictors of incipient retinopathy are needed for early diagnosis and for development of effective preventatives and curatives to save sight and to replace invasive treatments. This depends on the discovery and use of sensitive testing measures. In this thesis we use one such technique mfERG; it can be used to evaluate the health of retinal neurons at 103 retinal locations across the central 45 degrees of retina and is a very sensitive tool for detecting neural changes that are predictive of onset and progression diabetic retinopathy and edema.

Abnormalities revealed by the mfERG correlate well with the degree of clinical presentation of vascular retinopathy. Understanding the relationships between retinal neural changes and other retinal health indicator changes in diabetes may prove helpful to broaden knowledge about structural and functional changes in the retina that are exceptionally important for any new advances in curative as well preventative medicines and to define novel sensitive endpoints for clinical trials. In this dissertation I present four studies focused respectively on, revealing relationships between the mfERG measures of neural function and retinal thickness (Chapter 1), revealing relationships between mfERG measures of neural function and letter contrast sensitivity (Chapter 2), revealing early microvascular and photoreceptor changes in diabetes (Chapter 3) and finally, understanding how retinal health indicators (structural and functional) change in the patients with DME- a potentially sight threatening retinal vascular event.

The first study presented as the third chapter in this dissertation, tests if changes in retinal thickness and if mfERGs have any spatial correspondence in patients that are atan early stage of diabetes and do not have any retinopathy. Furthermore we are interested to see if there is any spatial agreement in the occurrence of abnormalities in both measures. It would be a very important finding if the study could shed some light on whether these two techniques could be used as surrogates of each other.

After testing the relationships of retinal neural function with retinal structure in the first study, I studied the possible relationships of retinal neural function with vision function assessed using letter contrast sensitivity in diabetes. The relationship between structure and function, a theme of the thesis, is also of interest in sensory function like letter contrast sensitivity. I examine whether retinopathy status influences the relationship between neural function and vision function. This study, as a first step, examines whether letter contrast sensitivity is different between participants with diabetes and the controls and whether the type, duration and control of diabetes and the presence of background retinopathy impacts the vision function measured here.

In the third study I examine, using adaptive optics scanning laser ophthalmoscopy (AOSLO), parafoveal cone photoreceptors and the retinal parafoveal capillary network in adult patients with type 2 diabetes and no retinopathy. Here again the rationale is to look at detailed and novel; structural aspects of the diabetic retina as an initial step in relating to function. Specifically, the focus of the study is to examine whether the cone photoreceptor spacing and retinal parafoveal capillary network are altered prior to the onset of vascular diabetic retinopathy. The study aimed to test whether AOSLO is a viable method for detection and characterization of microscopic signs of diabetic retinal damage in cone photoreceptors and capillaries in the parafovea before it is reflected in clinical examination.

Finally the fourth study is a longitudinal study to examine the vision, neuro- retinal function, retinal thickness, diabetes control and blood pressure of the subjects who are at risk for developing DME. Then the study proceeds to examine the natural history of the neural and vision function by tracking a group of participants beyond the onset of their DME; this is the main purpose in this study.

Collectively, across the 4 studies, I found the mfERG measures to be sensitive to diabetic changes in the function of neural retina at all stages of diabetic eye disease. However, there was no clear relationship between mfERG and other retinal health measures in the absence of clinical signs of retinopathy. AOSLO measures revealed clinically non-evident alterations in cone photoreceptor spacing and capillaries. The sensitive nature of the mfERG neural function measure and AOSLO structural assessment could eventually be used to help clinicians and researchers to develop the tools and means to preserve vision, in response to the burgeoning epidemic of blindness due to diabetes.