Early functional changes in human diabetic retina : a multifocal electroretinogram study

Abstract

Diabetes mellitus (DM) is a set of metabolic disorders leading to chronic hyperglycemia. It has become one of the most concerning health problems around the world. DM can damage the ocular capillary circulation which develops into diabetic retinopathy (DR). DR is the most common cause of new incidence of blindness among the working population due to its sight-threatening complications. The current clinical screening tests involve the fundus camera and ophthalmoscopy. These two tests form an effective and accurate means of detecting DR; however, DR can only be confirmed when obvious retinal vascular defects or damage exist. Studying early retinal changes before there are visible signs of retinopathy can help to prevent the visual functional loss resulting from DR. This helps in relieving the heavy economic burden to society produced by DR. Multifocal electroretinogram (mfERG) is a functional test which can reflect subtle retinal changes before any visible retinopathy. By modifying the mfERG paradigm, retinal activity from different origins can be measured. The modified mfERG is likely to be a more suitable technique to investigate the underlying mechanism of DR at an early stage. In this study, modified mfERG paradigms (Global flash multifocal electroretinogram [MOFO mfERG] and "Long-duration" multifocal electroretinogram) were applied in human eyes with type II DM to study any early retinal changes. Comparisons were made with two other common clinical morphological and functional assessments (Humphrey perimetry and Stratus circumpapillary retinal nerve fiber layer (RNFL) thickness measurement) so as to study their correlations and their diagnostic sensitivity. Objectives: 1) To ascertain the characteristics of the response triggered by the global flash multifocal electroretinogram (MOFO mfERG) under various combinations of global and focal flash luminance, and to determine the optimal conditions for this measurement in healthy subjects 2) To investigate early functional changes of local retinal defects in type II diabetic patients using the global flash multifocal electroretinogram (MOFO mfERG) 3) To investigate the correlations of the global flash multifocal electroretinogram (MOFO mfERG) with two common clinical visual assessments Humphrey perimetry and Stratus circumpapillary retinal nerve fiber layer (RNFL) thickness measurement in type II diabetic patients 4) To investigate the characteristics of the multifocal on- and off-responses in the human diabetic retina by a "long-duration" multifocal electroretinogram paradigm. This was to evaluate any changes in the antagonistic interaction in the middle and inner retina in the early stage of Type II DM. Methods: Experiment 1: The global flash electroretinogram (MOFO mfERG) measurement was applied in this experiment with a visual stimulation consisting of a 103-hexagon pattern. The stimulation was displayed with four video frames (multifocal flashes, followed by a dark frame, a global flash and then another dark frame). The focal and global flash intensities were varied independently at four levels (50, 100, 200 and 400 cd/m2). Ten healthy young adult subjects were recruited and underwent MOFO mfERG measurements with sixteen combinations of focal and global flash luminance. The mfERG responses were grouped into central and peripheral regions for analysis. Experiment 2: The MOFO mfERG measurement was carried at high (98%) and low (46%) contrast conditions. A 103-hexagon pattern was used as the stimulus. The focal and global luminance at the high contrast condition was kept at the optimal ratio found in Experiment 1. Thirty-eight type II diabetic patients and fourteen age-matched controls were recruited. Nine of the diabetics were free from retinopathy, while the remainder had mild to moderate non-proliferative diabetic retinopathy (NPDR). The mfERG responses were grouped into 35 regions for comparison according to the DR classification at those locations. The diagnostic values of the MOFO mfERG parameters on DR were also evaluated by constructing the receiver-operating-characteristic (ROC) curve. Experiment 3: Three visual assessments were carried out: the MOFO mfERG measurement with 103-hexagon stimulus pattern at high (98%) and low (46%) contrast conditions, Humphrey perimetry and Stratus circumpapillary retinal nerve fiber layer (RNFL) thickness. Forty-two type II diabetic patients and fourteen age-matched controls were recruited for comparison. Ten of the diabetics were free from diabetic retinopathy, while the remainder had mild to moderate NPDR. Correlations between local values of mfERG responses, perimetric sensitivity and RNFL thickness were evaluated by mapping the localized responses for the three subject groups. This helped in evaluating the early functional and morphological changes at the early stage of DR. Experiment 4: A “long-duration mfERG paradigm was used to evaluate the multifocal on- and off-responses in the human diabetic retina. In this paradigm, the stimulus pattern contained eight successive multifocal flashes, and followed by eight successive dark frames. The mfERG stimulus was a 61-hexagon pattern with measurement carried out under two chromatic conditions --- white and blue conditions. Twenty type II diabetic patients with no or mild NPDR and twenty-one age-matched healthy controls were recruited. The mfERG responses were grouped into rings for analysis. Changes of the mfERG responses under the two chromatic conditions were used to investigate any changes in antagonistic interaction within the retina at the early stage of DR. Results: Experiment 1: The MOFO mfERG paradigm gave rise to two main components in the resultant waveform: the direct component (DC) and induced component (IC). The DC amplitude increased with the focal flash intensity with global flash held constant. The global flash and focal flash luminance (g/f ratio) was an important parameter in determining the optimal DC and IC responses; the IC amplitude reached the peak value when the g/f ratio was at about 2:1. Further increasing the global flash luminance did not enhance the IC amplitude. Keeping this ratio at 1:1 with the focal flash luminance set between 100 cd/m2 and 200 cd/m2 was recommended for subsequent experiments. Experiment 2: The MOFO mfERG paradigm helped in investigating the middle and inner retinal responses in terms of DC and IC respectively. Local reduction of the DC and IC amplitudes were found in diabetic patients with and without DR. With increasing severity of retinopathy, a further reduction of the mfERG amplitudes was found. The reductions in the DC and IC responses provided crucial evidence that the middle and inner retina are impaired at an early stage in diabetic patients. Under the MOFO mfERG paradigm, the amplitude of the high contrast DC was useful in screening for localized functional deterioration, even before the appearance of visible DR signs. Experiment 3: The MOFO mfERG was superior to the automated Humphrey perimetry and the Stratus circumpapillary RNFL thickness measurement in type II diabetic patients, in showing differences between the diabetic group and the controls. The MOFO mfERG parameters demonstrated a better correlation with the functional perimetric assessment than the RNFL thickness measurement. All the MOFO mfERG amplitudes (except IC amplitude at high contrast) correlated significantly with the perimetric sensitivity (Pearson's r ranged from 0.23 to 0.36, p<0.01) than did the mfERG implicit time at both high and low contrast conditions across all subject groups. No consistent correlation was found between the mfERG parameters and the RNFL thickness for any subject group. Experiment 4: The "long-duration" mfERG paradigm helped in minimizing the overlap between the on- and off-pathway activities of the retinal responses. The resultant mfERG waveform includes two main parts. The first part is predominantly from the on-pathway activity containing a negative trough (N1), a positive peak (P1) and then a trough (N2). Beyond N2, there is a plateau followed by a second peak (P2). The diabetic group showed significantly greater N2 amplitude than the controls under white stimulation in retinal regions Rings 2 and 4 (p<0.05). The blue stimulation generally triggered greater mfERG amplitudes in P1, N2 and P2 (p<0.05) than did the white stimulation for both diabetic and control groups. When the stimulus changed from white to blue, the diabetic group showed a smaller percentage change than the controls in the peripheral retinal region (Ring 5) (p<0.02). Conclusion: The MOFO mfERG indicated that the middle and inner retinal function has deteriorated before the existence of clinically visible DR lesions. Such functional impairment would be expected to continue with the existence of observable vascular lesions. The MOFO mfERG measurement correlated better with the functional perimetric sensitivity than the morphological RNFL changes. However, the deterioration of local luminance sensitivity (as measured using perimetry) could not fully explain the functional loss found by mfERG. The "long-duration" mfERG paradigm, under two different chromatic stimulation conditions, demonstrated an imbalance of lateral antagonism which is proposed to be at or near the middle retinal layer.