Neuroretinal Degeneration in Relation to Vasculopathy in Diabetes

Diabetic vasculopathy is still, despite all efforts to treat its late complications, the leading cause of blindness among the working-age adults in the Western world, and in the developing countries, a similar situation is created due to the catching up in welfare and in its wake an identical unhealthy diabetes-promoting lifestyle (1).

The definition of diabetic retinopathy is based on observation of vascular changes. The first recognizable vascular abnormalities are microaneurysms and small hemorrhages, followed by more severe signs of vascular leakage, such as hard exudates and larger hemorrhages; vascular dropout, such as cotton wool spots; more widespread hemorrhages; and neovascularizations. The severity of the vascular changes has been graded to guide ophthalmologists in monitoring patients and treating patients in a timely fashion to prevent loss of vision (2).

It has been long known that the neuroretina is affected at an early stage by diabetes-induced metabolic changes. This diabetic neuroretinal degeneration has been demonstrated in histological studies and through the measurement of functional loss with a number of functional tests (3), such as contrast vision, color vision, visual field, and dark adaptation. An important test demonstrating neuroretinal degeneration is the multifocal electroretinogram (mfERG) (4,5). Recently, with spectral domain optical coherence tomography, structural changes could be shown with decreased thickness of the ganglion cell layer and retinal nerve fiber layer (6).

Signs of neuroretinal degeneration, measured with all these methods, were present in eyes even before any visible vascular pathology could be detected. This highlights the interesting concept of neuroretinal degeneration playing a role in the development of diabetic retinal vasculopathy (7–9). If neuroretinal degeneration does play a role in the subsequent development of diabetic vasculopathy, one would expect neuroretinal degeneration to precede vasculopathy.

Currently, the screening diagnostic gold standard for the presence of diabetic vasculopathy is automated reading of fundus photographs (10). With stereoscopic fluorescein angiography, the sensitivity to detect minor vascular changes increases (11). Recently, adaptive optics microvascular imaging found key vascular differences in patients without changes detected through clinical assessment (12). This implies that even though diabetic vascular pathology is not clinically visible, it still could be present.

In this issue, Reis et al. (13) investigated the order of events in diabetic retinopathy using a different approach to measure the intactness of the retinal vasculature. By measuring the ratio of plasma concentration of fluorescein and the concentration of fluorescein in the posterior vitreous following an injection of an individualized dosage of fluorescein (14 mg/kg) and thereby correcting for background signal, a penetration ratio can be measured, which increases with loss of the blood-retinal barrier (BRB) function. The authors assume a loss of the BRB is the first sign of loss of vascular integrity and is more sensitive than previous clinical methods to detect vasculopathy (14).

To detect neuroretinal dysfunction, Reis et al. used three types of tests: chromatic contrast sensitivity measuring color vision, achromatic contrast sensitivity with frequency-doubling technology, and mfERG. From these tests, several parameters were calculated and compared between healthy control subjects (n = 25), type 1 diabetic patients without visible vasculopathy and normal BRB function (n = 14), and type 1 diabetic patients with mild vasculopathy and/or loss of BRB function (n = 28).

The authors found significant differences for all functional tests, although not for all parameters, between healthy control subjects and both groups of diabetic patients. They could not find a difference between the groups of diabetic patients. The authors probed the sensitivity of the used functional tests to discriminate between healthy control subjects and the patients without visible diabetic retinopathy and normal BRB, using area under the curve calculations at a fixed specificity of 90%. The mfERG amplitudes P1 (positive wave) and N1 (negative wave) proved to be the parameters with the highest sensitivity to make this distinction. Reis et al. concluded that this is proof of the existence of a “neural phenotype,” with neural changes occurring in addition to, and independently of, vascular lesions in diabetic patients.

The number of patients in the study by Reis et al. is rather small, and although significant differences in even a small group of subjects tested could hint at a robust difference, the generalizability becomes less. Nevertheless, the findings of the authors are in line with and confirmed by many other articles on the subject (4–7,9,15–21).

The authors claim to have used a more sensitive criterion to detect vasculopathy compared with clinical visualization of vasculopathy with photographs. However, in the 17 patients without visible vasculopathy, only 3 did not show visible vasculopathy while their BRBs showed a loss of function. Perhaps results would not have been so different if the authors would have used the visibility of vasculopathy criterion alone. The authors ascribe differences between their results and other studies to differences in the population definition, especially the use of a more sensitive test to detect vasculopathy, and this seems to be questionable for the above-mentioned reason.

Two important risk factors for the development of vasculopathy, and probably also for neuroretinal degeneration, were analyzed: HbA_1c and duration of diseases. For HbA_1c, Reis et al. used a single measurement and found no statistical significant correlation with BRB function, but they did not perform an analysis of a possible relation with the neurophysiological test results. Duration of disease was very weakly related to changes in neurophysiological tests and was tested separately for patients without detectable vasculopathy and those with vasculopathy, instead of for the whole group of patients. In the patients with vasculopathy, duration was related to some parameters of the mfERG, and in the patients without vasculopathy, duration was associated with changes in color vision. The authors suggest this to be an argument for a different pathophysiological mechanism for neuroretinal degeneration in the presence or absence of vasculopathy. This seems to be an unlikely assumption. Accepting the presence of neuroretinal degeneration in the absence of vasculopathy that precedes later to development of vasculopathy, one could expect additional damage to an already disturbed neuroretina, not a complete different type of neural damage.

The results of the study by Reis et al. have confirmed the notion that neuroretinal degeneration leading to functional loss can be detected in patients with diabetes at a time that neither visible vasculopathy nor loss of BRB function is present, a method perhaps even more sensitive to detect any vasculopathy. Neuroretinal degeneration could be a very sensitive biomarker for subsequent vascular damage, which often will lead to vision-threatening complications, such as diabetic macular edema or proliferative disease. Neuroretinal degeneration could play an important permissive or even causative role in the subsequent development of vasculopathy. Treatment aimed at preventing neural degeneration through neuroprotection could become an interesting new strategy (22–24). Last but not least, neuroretinal degeneration could be a sign of more widespread damage to the neural system of patients, such as peripheral neuropathy or neuropsychological disturbances (25–28).

Ignoring this very early sign of damage caused by diabetes in patients could rob the patients of an opportunity to improve the control of the disease in a timely manner, thereby preventing further damage in and outside the eye, which could improve their quality of life (Fig. 1). The methods to measure neurophysiologic functional loss and the method to detect a disturbed BRB function used in Reis et al. (13) are not particularly useful to be incorporated into clinical practice. There is a need to develop tests to detect neuroretinal degeneration in a simple and reliable way. The challenge is the very subtle changes in neuroretinal structure and function in the eyes of diabetic patients and the large variability in the outcomes of the tests that can be used to detect early neuroretinal damage. As rightly pointed out by the authors, this could be overcome by performing longitudinal studies, in which each individual patient serves as his or her own control subject, so that change over time could become a measure of ensuing damage, taking into account the changes induced by physiological aging.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Central in the complex pathophysiology are the changes in the metabolism caused by diabetes. These disturbances, in a multitude of metabolic pathways, lead to neurodegeneration and vasculopathy. Once established, both processes will negatively influence each other. At present, the string of events is unknown: Does vasculopathy lead to neurodegeneration or the other way around? Are both processes independent of each other? Another important question is the relationship between neurodegeneration in and outside the eye. Perhaps the eye can be a proxy for neurodegeneration outside the eye, specifically in the brain.

• © 2014 by the American Diabetes Association. Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered.