Neurovascular Dysfunction Precedes Neural Dysfunction in the Retina of Patients with Type 1 Diabetes

Abstract

Purpose.: A variety of studies have shown that flicker-induced vasodilatation is reduced in patients with diabetes. It is, however, unclear whether reduced neural activity or abnormal neurovascular coupling is the reason for this phenomenon. In the present study, we hypothesized that retinal neurovascular dysfunction precedes neural dysfunction in patients with early type 1 diabetes.

Methods.: In the present study, 50 patients with type 1 diabetes without retinopathy and 50 healthy age- and sex-matched control subjects were included. The retinal vascular response to flicker stimulation was measured using the dynamic Retinal Vessel Analyzer. In addition, the response in retinal blood velocity to flicker stimulation as assessed with laser Doppler velocimetry was studied in a subgroup of patients. Pattern electroretinography (ERG) was used to measure neural retinal function.

Results.: The flicker responses of both retinal arteries and veins were significantly reduced in patients with diabetes (veins in the diabetic group: 3.5 ± 2.3% versus healthy control group: 4.6 ± 2.0%; P = 0.022 between groups, whereas arteries in the diabetic group: 2.0 ± 2.7% versus healthy control group: 3.8 ± 1.7%; P < 0.001 between groups). Likewise, the response of retinal blood velocity was reduced in patients with diabetes, although adequate readings could only be obtained in a subgroup of subjects (diabetic group [n = 22]: 19 ± 7%; healthy control group [n = 24]: 43 ± 19% P < 0.001 between groups). The parameters of pattern ERG were not different between the two groups.

Conclusions.: The study confirms that flicker responses are reduced early in patients with type 1 diabetes. This is seen before alterations in pattern ERG indicating abnormal neurovascular coupling. (ClinicalTrials.gov number, NCT00712842.)

Introduction

Classically, diabetic retinopathy is seen as a microangiopathy affecting endothelial cells and pericytes. ^1,2 Recently, however, it was recognized that the disease is also associated with early neurodegeneration. ³ Several large-scale epidemiologic studies indicate that wider retinal venous caliber is strongly associated with fasting glucose level as well as with diabetes. ⁴ In addition, lower arteriovenous retinal ratio is associated with the risk of diabetes, and both lower arteriovenous ratio and wider retinal venular caliber predict incident impaired fasting glucose levels. ⁴ This again indicates early vascular involvement in diabetic retinopathy. 

Another retinal vascular abnormality associated with diabetes is an abnormal retinal vascular response to flicker stimulation, as assessed for example with the Dynamic Vessel Analyzer (DVA; Imedos GmbH, Jena, Germany). ^5,6 The mechanism underlying this observation is unclear but may involve neuronal and vascular aspects. ⁷ The latter is supported by studies indicating that nitric oxide synthase (NOS) inhibition reduces the flicker response in animals and humans. ⁸ The former is supported by a recent study suggesting that reduced flicker response is correlated with inner retinal dysfunction as assessed with electroretinography (ERG). ⁹ Pattern ERG is a technique for assessing central retinal function and has been widely used for assessing retinal ganglion cell function in patients with retinal and optic nerve head disease. ¹⁰ Recent studies in mouse confirm that selective lesion of ganglion cells almost eliminates pattern ERG. In addition, the positive wave of mouse pattern ERG is dominated by contributions from ON pathway neurons, whereas the negative amplitude is affected mainly by spiking activity from the OFF pathway. ¹¹  

In the present study, we hypothesized that abnormal flicker-induced vasodilatation precedes an abnormal pattern ERG. This hypothesis was tested in a group of patients with early type 1 diabetes. Data were compared to results in a healthy age-matched control group. 

Materials and Methods

Subjects

The study protocol was approved by the Ethics Committee of the Medical University of Vienna and followed guidelines set forth in the Declaration of Helsinki. All patients signed written informed consent prior to inclusion and passed a screening examination that included physical examination, slit lamp biomicroscopy, funduscopy, and measurement of intraocular pressure (IOP) before the study day. One-hundred individuals aged >18 years old were included in this observer-blinded cross-sectional study. Fifty patients with type 1 diabetes with no diabetic retinopathy were included. Further inclusion criteria for type 1 diabetes patients were a serum cholesterol level <200 mg/dL, systolic blood pressure ≤140 mm Hg, and diastolic blood pressure ≤80 mm Hg. Exclusion criteria were any regular intake of medication except insulin, a hemoglobin (HbA₁c) value of more than 10% and smoking. Fifty healthy age- and sex-matched control subjects with systolic blood pressure (SBP) ≤140 mm Hg, diastolic blood pressure (DBP) ≤80 mm Hg, serum cholesterol level <200 mg/dL, and normal ocular findings were included. Exclusion criteria were regular drug intake and smoking. 

Further exclusion criteria for all subjects were ametropia >3 diopters, other ocular abnormalities that might interfere with the purposes of the present study, a clinically relevant illness prior to the study, pregnancy or lactation, and a patient or family history of epilepsy. All subjects were drug-free for at least 2 weeks prior to study, except for insulin in the diabetic group. Participants were required to abstain from beverages containing alcohol or caffeine for 12 hours before the study. 

Dynamic Vessel Analyzer

Diameters of the temporal retinal artery and vein between 1 and 2 disc diameters from the margin of the optic disc were continuously measured using the DVA. The DVA consists of a fundus camera (model FF 450; Carl Zeiss Meditec AG, Jena, Germany), a digital video camera, a real-time monitor, and software for analysis and determination of retinal vessel diameters from digitized images. The system provides excellent reproducibility and sensitivity. ^12,13 After selection of the measurement location, the DVA is able to follow the vessels during movements within the measurement window. Retinal vessel diameters were measured for 4 minutes. For the second minute, full-field flickering light with a frequency of 12.5 Hz was used for stimulation by square wave pattern modulation of the fundus camera illumination at a contrast ratio of 25:1. 

Laser Doppler Velocimetry

The principle of red blood cell velocity measurement by laser Doppler velocimetry (LDV) is based on the optical Doppler effect. Laser light, which is scattered by moving erythrocytes, is shifted in frequency. This frequency shift is proportional to the blood flow velocity in the retinal vessel. The maximum Doppler shift corresponds to the centerline erythrocyte velocity (Vmax). ¹⁴ In the present study, we used a fundus camera-based system with a single-mode laser diode at a center wavelength of 670 nm (Oculix Sarl; Arbaz, Switzerland). The Doppler shift power spectra were recorded simultaneously for two directions of the scattered light. Scattered light was detected in the image plane of the fundus camera. This scattering plane can be rotated and adjusted in alignment with the direction of Vmax, which enables absolute velocity measurements. ^15,16 In the present study, Vmax was determined for a major temporal vein before, during, and after stimulation with diffuse luminance flicker. All measurement locations were within 1 to 2 disk diameters from the center of the optic disk. 

For flicker stimulation, a custom-built device was used, stimulating with light flashes at a frequency of 12.5Hz. Flicker was generated by focusing the light of a 150-W-halogen light source on a rotating sector disc, producing a square wave light pattern with a modulation depth of 100%. Use of an optical fiber provided flicker stimuli that were delivered to the eye through the illumination pathways of the laser Doppler velocimeter, respectively. The flicker was centered on the macula with an angle of approximately 30°, producing a retinal irradiance of 300 μW cm⁻² (approximately 260 lux). 

Pattern ERG

Pattern ERG was performed according to the ISCEV standard for clinical patterns. ¹⁷  

Briefly, gold foil corneal recording electrodes were positioned directly under the center of the pupil so that there was no movement of the electrode when the patient blinked. Reference and ground electrodes were placed in the outer canthus and the forehead, respectively. A black-and-white reverse checkerboard pattern was used with an aspect ratio of width over height of the stimulus field not exceeding 4:3. The mean of the width and height of the stimulus field was 15°, with a check size of 0.8°.The contrast between black and white squares was close to 100%. The reversal rate was 2.2 Hz. A minimum of 100 artifact-free sweeps was collected and averaged. The P50 amplitude was calculated from the trough of N35 to the peak of P50, and the N95 amplitude was measured from the peak of P50 to the trough of N95. All data were recorded without pupil dilation. 

Measurement of IOP and Systemic Hemodynamics

IOP was measured with a slit lamp-mounted Goldmann applanation tonometer (Haag-Streit, Bern, Switzerland). Before each measurement, two drops of oxybuprocain hydrochloride combined with sodium fluorescein were instilled for local anesthesia. 

SBP, DBP, and mean arterial blood pressure were measured on the upper arm by an automated oscillometric device (HP-CMS patient monitor; Hewlett Packard, Palo Alto, CA). Pulse rate was automatically recorded by the same unit from a finger pulse oxymetric device. 

Experimental Paradigm

On the screening day, prestudy screening including an ophthalmic examination was performed. If the subject met inclusion criteria, a study day was scheduled within the next 14 days. On the study day, all subjects underwent pupil dilation after instillation with tropicamide (Mydriaticum Agepha-Gtt; Agepha, Vienna, Austria). After patients spent 20 minutes in a rest period in a sitting position, baseline measurements of arterial blood pressure and pulse rate were performed. Thereafter, noninvasive measurements of retinal vessel diameters and blood velocities, including flicker stimulation, were performed. All flicker stimuli were repeated three times. Pattern ERG measurements were scheduled afterward. Finally, intraocular pressure was measured using Goldman applanation tonometry at the end of the study day. 

Data Analysis

Baseline (BL) values of vessel diameters and blood velocities were calculated as an average of the last 20 seconds before start of the flicker stimulation. Values during flicker (FL) stimulation were calculated as an average of the last 20 seconds of the stimulation period. Flicker-induced changes in retinal vessel diameters and retinal blood velocities were expressed as percentages of change over baseline values, that is, (FL − BL) × 100/BL. The average percent of change was defined as the flicker response. Data during flicker stimulation were considered accurate only if the coefficient of variation of the flicker response in each individual was less than 20%. If this reproducibility criterion was not fulfilled, data were not considered for analysis. An unpaired t-test was used to compare data between subjects with diabetes and healthy control subjects. For all calculations, a P value of <0.05 was considered significant. 

Results

Characteristics of the subjects are shown in the Table. No differences were found with regard to sex, age, IOP, systemic blood pressure, and pulse rate. As expected, HbA₁c and plasma glucose levels were higher in patients with diabetes than in healthy control subjects. In 1 subject of the diabetes group and in 1 subject of the healthy control group, the reproducibility criterion for flicker responses in retinal veins was not fulfilled. Accordingly, data for both groups stem from 49 subjects only. The diameter of retinal veins was comparable between subjects with diabetes (156.9 ± 20.0 μm, n = 49) and healthy control subjects (154.9 ± 20.1 μm, n = 49; P = 0.62 between groups). Whereas the reproducibility criterion was fulfilled in all healthy subjects in retinal arteries, it was not fulfilled in 1 subject with type 1 diabetes. Retinal arteries were larger in the diabetic group (133.6 ± 17.5 μm, n = 49) than in the healthy control group (124.5 ± 15.3 μm, n = 50; P = 0.007 between groups). Reproducibility was generally weak in LDV measurements during flicker stimulation. As such, reproducibility criteria were fulfilled only for 22 patients with type 1 diabetes and 24 healthy control subjects. Baseline retinal blood velocities (2.27 ± 0.56 cm/s, n = 22) were higher In diabetic subjects than in healthy control subjects (1.89 ± 0.44 cm/s, n = 24; P = 0.024 between groups). 

Results comparing flicker response between subjects with diabetes and healthy control subjects are presented for vessel diameters and flow velocities in Figures 1 and 2, respectively. All flicker responses were significantly reduced in patients with diabetes. In retinal arteries, the response was 3.8% ± 1.7% in healthy subjects and 2.0% ± 2.7% in patients with diabetes (P < 0.001). In retinal veins, the response was 4.6% ± 2.0% in healthy subjects and 3.5% ± 2.3% in patients with diabetes (P = 0.022). With regard to the response of velocity responses to diffuse luminance flicker, it needs to be considered that less than half of the measurements fulfilled the reproducibility criteria. Nevertheless, we observed a significantly reduced flicker response in velocity between the two groups (healthy controls: 43 ± 19%; patients with diabetes: 19 ± 7%; P < 0.001). 

Figure 1. 

 

Flicker-induced vasodilatation in retinal arteries and veins in the two study groups (diabetic group n = 49; healthy group n = 50). Data are means ± SD. Asterisks indicate statistical significance.

Figure 2. 

 

Flicker induced increases in blood velocity in retinal veins in the two study groups (diabetic group n = 22; healthy group, n = 24). Data are means ± SD. Asterisk indicates statistical significance.

Results of pattern ERG implicit times and amplitudes are presented in Figure 3. Neither N50 implicit time (P = 0.37) nor N95 implicit time (P = 0.49) was significantly different between groups. Likewise, N50 amplitude (P = 0.26) and N95 amplitude (P = 0.31) were comparable between groups. 

Figure 3. 

 

Pattern ERG implicit times and amplitudes in the two study groups. Data are means ± SD.

Discussion

Several previous studies have shown that flicker-induced vasodilatation is reduced in patients with diabetes. ^5,6,18,19 One hypothesis is that reduced neuronal activity during flicker stimulation is responsible for this effect. This hypothesis is supported by several studies relating ERG measurements to the hyperemic response to visual stimulation. Recently, Lecleire-Collet and co-workers ⁹ studied retinal vascular responses to flicker stimulation in a mixed group of type 1 and 2 patients and related them to pattern ERG, rod ERG, and oscillatory potentials. Significant associations were found between measures of inner retinal neural impairment and reduced flicker responses. These findings are in good agreement with previous data in experimental animals and humans. Correlations were found between harmonic component amplitudes of the flicker ERG when the modulation or mean illumination of a diffuse photopic luminance flicker were varied from 10% to 100% in monkeys and humans. ^20,21 In glaucoma patients in whom the flicker response is also reduced, ²² Riva and co-workers ²³ have shown that neural activity and flicker-evoked changes in optic nerve head blood flow can be independently altered early in the disease process. 

Table. 

 

Subject Characteristics

Characteristic	Patients with Type 1 Diabetes	Healthy Subjects	P Value (Unpaired t-test)
Sex, males/females	20/30	20/30	1.0
Age, y	35.1 ± 7.6	34.0 ± 8.6	0.82
Systolic blood pressure, mm Hg	119 ± 11	118 ± 11	0.58
Diastolic blood pressure, mm Hg	64 ± 8	63 ± 10	0.68
Mean arterial pressure, mm Hg	83 ± 8	82 ± 9	0.56
Pulse rate, beats/min	68 ± 13	68 ± 11	0.92
Blood glucose, mmol/L	7.2 ± 3.4	4.7 ± 0.8	<0.001
HbA1c, %	7.5 ± 1.3	5.2 ± 0.3	<0.001
Diabetes duration, y	9.8 ± 3.4
Intraocular pressure, mm Hg	15 ± 2	15 ± 2	0.35

 Baseline data of the two study groups.

Our study indicates that the response to retinal vascular response to flicker stimulation is reduced in type 1 diabetes before changes in pattern ERG become evident. In comparison to the study by Lecleire-Collet and co-workers. ⁹ several differences in the study population need to be considered. In our study, only patients with type 1 diabetes were studied, whereas that previous study included patients with type 1 and type 2 diabetes. Diabetes duration was shorter in our study than for the type 1 patients in the previous study, and both HbA₁c and glucose plasma levels were lower. These factors may well explain why pattern ERG was abnormal in the previous study but not in our cohort. Whereas our study indicates that a reduced flicker response precedes an abnormal pattern ERG in patients with diabetes, it is likely that in later stages of the disease, reduced neural activity contributes to further decline of flicker-induced vasodilatation. ⁹ Present study evidence adds up to the fact that vascular abnormalities are an early process in diabetes, based on the analysis of retinal vessel calibers. Several studies have reported abnormalities in retinal blood flow ^24–28 and a supernormal retinal oxygenation response as early processes in diabetes. The ability to detect changes in retinal autoregulation in individual patients with diabetes depends on the reproducibility of the technique used. Currently, however, no technique for measuring retinal blood flow is available that can be used in population-based studies. 

The mechanism by which flicker stimulation increases retinal blood flow is not fully elucidated but appears to involve NO, ¹² which is a key regulator of vascular tone at the posterior pole of the eye. ²⁹ Neuronal NOS may be directly activated in neurons during neuronal activity, but endothelial NOS may play a role as well by regulating arachidonic acid metabolism and controlling flow-induced vasodilatation. ³⁰ Nguyen and co-workers ³¹ have shown that wider retinal arterial and venous calibers are associated with an impaired flicker response, hypothesizing that endothelial dysfunction is linking retinal vascular abnormalities with diabetes and diabetic retinopathy. We have previously shown that flicker-induced retinal vasodilatation is associated with an abnormal flow-induced forearm response, ³² again indicating vascular endothelial involvement. The hyperemic response to flicker stimulation is disturbed early in diabetes, when the endothelium-independent response to nitroglycerin is still normal, indicating that reduced vascular reactivity due to the morphological changes in the vessel wall are not associated with the phenomenon. ⁷ Finally, the response to flicker stimulation is also reduced in glaucoma, ^22,33 a disease that is also associated with abnormal NO production as well. ³⁴  

Whereas there is evidence that under physiological conditions, flicker-induced vasodilatation is mediated in part by NO from either endothelial or neural sources, Mishra and Newman ^35,36 have shown that inhibition of inducible NOS (iNOS) restores the response of retinal vessels to flicker stimulation in a streptozotocin-induced rat model of type 1 diabetes. This is highly compatible with previous data indicating that inhibition of iNOS also restores the retinal oxygenation response in diabetic rats, as assessed with magnetic resonance imaging. ^37,38 Evidence has accumulated showing that glial cells play a key role in the hyperemic response in the brain and retina. ³⁹ In the retina, the principal macroglial cells are the Müller cells, which are affected early in diabetes. They exhibit an altered fluid transport, become gliotic, and display altered potassium siphoning, glutamate, and gamma-aminobutyric acid uptake. ⁴⁰  

Most of the previously published studies relied on flicker-induced retinal vessel dilation as introduced more than 10 years ago, only. In the present study, we also measured retinal blood velocities during flicker stimulation. We have previously shown that flicker induces a pronounced increase in retinal blood velocities, which is in good agreement with data from the present study. ⁴¹ In addition, our data indicate that flicker-induced changes in retinal blood velocities are reduced in diabetes. Reproducibility using laser Doppler velocimetry during flicker stimulation is a problem because fixation becomes difficult. As such, reproducibility criteria were fulfilled in less than 50% of the participating subjects. There are, however, approaches to quantifying retinal blood velocity and retinal blood flow based on optical coherence tomography (OCT) techniques. ^42–44 Use of Doppler OCT preliminary data during flicker stimulation has been published, ⁴⁵ but no reproducibility data are available. 

Our study has several strengths and limitations. An advantage is related to the relatively strict inclusion/exclusion criteria that resulted in a relatively homogenous group of patients. Other strengths include the measurement of blood velocities and the check for reproducibility by using three consecutive flicker stimulation periods. Limitations include the relatively small sample size and the cross-sectional nature of the study. In addition, we did not measure flash ERG oscillatory potentials or multifocal ERG in the present study. Evidence has accumulated, however, showing that multifocal ERG is superior to other electrophysiological measurements in predicting the onset of diabetic retinopathy. ^46–48  

In conclusion, the present study indicates that the retinal vascular response to flicker stimulation is reduced before the reduction in pattern ERG in patients with type 1 diabetes. This indicates that in early diabetes, the abnormal retinal hyperemic response may not be a consequence of reduced neuronal activity. Altered endothelial function as well as diabetes-induced changes in glial cells may play an important role, but further studies are required to elucidate this issue in more detail. 

Footnotes

 Supported by a grant from the Christian Doppler Laboratory for Laser Development and Application in Medicine, and by Grant KLIF250 from the Austrian Science Fund (Fonds zur Förderung der Wissenschaftlichen Forschung).

Footnotes

 Disclosure: M. Lasta, None; B. Pemp, None; D. Schmidl, None; A. Boltz, None; S. Kaya, None; S. Palkovits, None; R. Werkmeister, None; K. Howorka, None; A. Popa-Cherecheanu, None; G. Garhöfer, None; L. Schmetterer, None