Macular Edema Reflects Generalized Vascular Hyperpermeability in Type 2 Diabetic Patients With Retinopathy

A total of 20 type 2 diabetic patients with DMa, defined as retinal hemorrhages and/or microaneurysms combined with hard exudates and/or retinal edema in the macular area, were identified in the database of eye examinations in our screening clinic for diabetic retinopathy. For each patient in the case group, we defined the study eye as the eye with the most severe degree of retinopathy.

A total of 20 type 2 diabetic patients with no signs of retinopathy were identified in the above-mentioned database. For each patient in the DMa group, we identified and ranked the 10 patients in the database with no retinopathy that matched best with regard to age, sex, and known duration of diabetes. The patients were invited to participate in the study according to rank. For each control patient, we included the eye corresponding to the study eye of the matching case.

Patients were considered to have type 2 diabetes if they had onset of diabetes after age 30 years, had no need for insulin treatment for at least 1 year after the diagnosis of diabetes, and had no history of ketoacidosis. Because ACE inhibitors are known to affect UAE and have been suggested to affect TER_alb (12) and retinal vascular permeability (13), patients treated with ACE inhibitors or angiotensin II receptor antagonists were not included. Likewise, patients previously treated with laser photocoagulation were not included.

Leisure time physical activity was graded as follows: passive (not participants), moderate (physical exercise once or twice a week), and active (physical exercise more than twice a week). Patients were classified as nonsmokers (without daily use of tobacco for the preceding year), moderate smokers (<15 cigarettes per day), or heavy smokers (>15 cigarettes per day). Patients were classified as having macrovascular disease if one or more of the following were present: symptoms of angina pectoris; history of myocardial infarction, coronary artery bypass grafting, or percutaneous transluminal coronary angioplasty; symptoms of or operation for intermittent claudication; or history of amputation, transient ischemic attack, or stroke. All subjects included in the study were Caucasian.

The patients underwent routine ophthalmological examinations for diabetic retinopathy, including measurement of visual acuity, slit-lamp examination, fundus photography, fluorescein angiography, as well as OCT of the central retina.

Fundus photography was performed using a Canon 60UV fundus camera on Kodak Ektachrome 64 color diapositive film. In each eye, one 60° image centered on the fovea and a nasally displaced field centered on the optic disk were taken. Fluorescein angiography was performed using Ilford Delta 400 black/white film. A fast sequence was taken during the filling phase of the retinal vessels of the study eye, and late-phase images were taken of both eyes 5–10 min after injection of fluorescein. The fluorescein angiograms verified that all maculopathies were of an exudative type.

The OCT method is analogous to ultrasound B-scanning, in that distance information is extracted from the time delays of reflected signals. Optical (laser) instead of acoustic waves are used in OCT, which allows a much higher resolution (10 μm) and thereby a more precise assessment of retinal thickness (14). Therefore, this method is very sensitive and, in addition, noninvasive. In this study, we used the Humphrey optical coherence tomography scanner (version A4.1; Humphrey Instruments, San Leandro, CA). Six radial scans centered on the fixation point were performed with a 30° interval (Fig. 1). Each scan had a length of 2.83 mm, and the sequence of scanning was the same for all patients, as performed using the default settings of the scanner.

Analysis of OCT scans was performed by an ophthalmologist who had not participated in the examination of study patients. Therefore, this ophthalmologist had no knowledge of the patients’ clinical or laboratory data and did not know whether individual patients were case or control subjects. The scans were displayed on the OCT monitor, and the distance from the retinal surface to the zone with highest reflectivity at the choriocapillaris/pigment epithelium was measured with a ruler. The measured distances on the monitor were converted to distances on the retinal plane using the standard calibration of the OCT apparatus. For each scan, five thickness measurements were obtained, at points 10, 30, 50, 70, and 90% of the length of the scan from its beginning (Fig. 1) (15,16). The average thickness of a scan was calculated as the mean of the five measured thicknesses, and the average retinal thickness in the macular area was then calculated as the mean thickness of the six scans. One patient in the control group had no identifiable high reflectivity zone at the choriocapillaris/pigment epithelium, and for this reason, we were not able to measure retinal thickness in this patient. The OCT scans of this patient were excluded, leaving 20 sets of OCT scans in the case group and 19 sets in the control group for analysis.

In the first 15 patients included in each group, we measured the TER_alb as described by Parving et al. (17,18). We were unable to perform this examination in the last five patients in each group because of a sudden withdrawal of ¹²⁵I-labeled albumin from the market.

The examination was performed in the morning after an overnight fast. After resting for 1 h in the supine position, the patient received an intravenous bolus injection containing 0.2 MBq ¹²⁵I-labeled human albumin (HSA for metabolic studies, IsoPharma IT.23S). Blood samples were collected from a cubital vein in the other arm before and 10, 15, 20, 30, 40, 50, 55, and 60 min after the injection for counting of plasma radioactivity and measurement of total plasma protein concentration in duplicate by refractometry (Bellingham & Stanley, Tunbridge Wells, U.K.). Radioactivity was corrected for total plasma protein concentration, and the slope of the linear regression of radioactivity on time was used to calculate TER_alb (i.e., the plasma tracer disappearance rate). TER_alb measurements were accepted only if the correlation coefficient between the time points for blood collection and the corresponding values of a specific radioactivity exceeded 0.85. On the basis of this criterium, two TER_alb examinations (one case subject and one control subject) were excluded from the analyses.

AMBP was measured using an oscillometric technique (Spacelabs 90207, validated by the British Hypertension Society) (19). Readings were obtained at 20-min intervals over 24 h. Individually reported sleeping times were implemented in the calculation of day and night blood pressure (BP).

Three bedside cardiovascular reflex tests were performed: heart rate variation to deep breathing (inspiration/expiration difference, average of two determinations), heart rate response to standing up (30:15 ratio), and blood pressure response to standing up. These tests were performed and evaluated as described by Ewing et al. (20).

UAE was measured by radioimmunoassay and expressed as the geometric mean of three overnight collections made within 1 week. HbA_1c was determined by high-performance liquid chromatography (nondiabetic range 4.4–6.4%). Blood glucose was determined using the Reflolux II (Boehringer Mannheim, Mannheim, Germany).

All subjects included in the study gave their written informed consent to participate. The study was approved by the regional ethics committee.

To approximate normal distribution, UAE measurements were log-transformed before analysis. Differences between groups were tested by the unpaired Student’s t test. For noncontinuous variables, the χ² test with Yates’ correction was used. Correlations were analyzed using Pearson’s test. A two-tailed P value <0.05 was considered significant. Results for normally distributed variables are expressed as mean ± SD, whereas UAE values are expressed as geometric mean ×/÷ tolerance factor.

Clinical characteristics, retinal thicknesses, UAE, TER_alb, and AMBP values of the patients are shown in Table 1.

The groups were well matched with regard to age, sex, and known duration of diabetes. The median pairwise difference was 1.4 years for age and 1.2 years for duration of diabetes. Values of HbA_1c and total cholesterol were significantly higher in the patients with DMa than in the patients without retinopathy. Interestingly, the level of physical activity was higher in the control group than in the case group, whereas there were no differences in use of antidiabetic medication, prevalence of macrovascular disease, smoking habits, BMI, or serum creatinine level between the groups. The cardiovascular reflex tests revealed no statistically significant differences in autonomic function between the groups (data not shown).

Average UAE and retinal thickness, as measured by OCT, were higher in the group with DMa than in the control group. There was no significant difference in TER_alb between the groups. AMBP values, pulse pressures, and night/day ratios were uniformly higher in the case group than in the control group, although the differences between groups did not reach the level of statistical significance (all data not shown). Because more case subjects than control subjects received antihypertensive treatment, true differences in BP between the groups might have been missed. Restricting the analysis to the subgroup of patients not receiving antihypertensive treatment (10 case subjects, 14 control subjects), BP values were still uniformly higher in the case group than in the control group, but the differences were numerically larger, reaching statistical significance for night systolic BP (131 ± 14 vs. 117 ± 14 mmHg, P < 0.05) and night pulse pressure (60 ± 9 vs. 49 ± 9 mmHg, P < 0.01).

In the total group of patients (n = 40), retinal thickness (OCT), UAE, and TER_alb showed a significant correlation (OCT versus UAE: r = 0.56, P < 0.01; OCT versus TER_alb: r = 0.36, P = 0.069; UAE versus TER_alb: r = 0.38, P < 0.05). In a subgroup analysis, there was a strong correlation between OCT, UAE, and TER_alb in the group with DMa (\?\Figs. 2–4) (OCT versus UAE: r = 0.58, P < 0.01; OCT versus TER_alb: r = 0.55, P < 0.05; UAE versus TER_alb: r = 0.81, P < 0.01), whereas there were no correlations between these three parameters in the group without retinopathy (OCT versus UAE: r = 0.04, P = 0.86; OCT versus TER_alb: r = 0.37, P = 0.22; UAE versus TER_alb: r = 0.14, P = 0.63).

Our study is the first to examine associations between retinal thickness as assessed by OCT and measures of glomerular and vascular permeability in diabetes. In type 2 diabetic patients with DMa, we found significant correlations between retinal thickness, UAE, and TER_alb, whereas no such association was seen in type 2 diabetic patients without retinopathy. Increased retinal thickness has previously been shown to coincide with increased leakage of retinal blood vessels in type 2 diabetes (21). Therefore, our findings indicate that in type 2 diabetic patients with DMa, the pathologically increased permeability of the retinal vessels, as reflected by increased retinal thickness, corresponds to equivalent permeability changes in kidney capillaries and in small vessels throughout the circulation of the diabetic patient.

A decade ago, the hypothesis of albuminuria as a marker of widespread endothelial damage was introduced (22). Several studies have established a relationship between abnormal UAE and increased TER_alb (23,24,25). Because retinopathy and nephropathy are often coexisting in diabetes, it has been difficult to establish an independent association between the presence of retinopathy and increased TER_alb (24,25). In the present study, we were not able to demonstrate differences in TER_alb between the groups with and without retinopathy. This could be because the included cases had varying degrees of DMa, ranging from a single hard exudate to macular edema. However, when considering the varying degree of macular edema, as evaluated by OCT measurements of retinal thickness, we found a significant association between retinal thickening and TER_alb.

Several studies have established a statistical association between retinopathy and nephropathy in type 1 diabetes (26) as well as in type 2 diabetes (27), and these complications were also strongly correlated in the present study. Both DMa and diabetic nephropathy are characterized by a pathologically increased permeability of a vascular barrier. In DMa, leakage of lipoproteins and water from the retinal vessels results in formation of hard exudates and macular edema, whereas increased permeability of glomerular capillaries in diabetic nephropathy leads to an increased UAE. Hemodynamic as well as structural abnormalities are believed to initiate these changes. It is well known that systemic hypertension is a risk factor for development of both retinopathy and nephropathy in diabetes (28,29), conceivably because elevated blood pressure is transmitted to the microcirculation, thus resulting in capillary/glomerular hyperperfusion and hypertension (30,–32). Because the retinal vessels have no autonomic innervation, blood flow in response to increased blood pressure is dependent on local factors, e.g., autoregulation. As the autoregulatory capacity of the retinal circulation is impaired when blood glucose is elevated (33), simultaneous hypertension and hyperglycemia is anticipated to be particularly detrimental for the microvasculature.

In accordance with this hypothesis, our patients without retinopathy had significantly better glycemic control and also tended to have lower blood pressure than patients with DMa. These findings are in line with results from the U.K. Prospective Diabetes Study, in which poor glycemic control and elevated blood pressure have been pointed out as major risk factors for development of diabetic complications (34). We have recently shown that increased pulse pressure is associated with a higher frequency of complications in type 2 diabetic patients (35). Furthermore, lowering of blood pressure and glycemia has been proven to slow the development and progression of microvascular complications in diabetes (36,–40).

In several studies, treatment with ACE inhibitors has been proposed to provide specific organ-protective effects beyond what would be expected from the relatively modest reductions in BP achieved in these studies (38,41,42). A study in type 2 diabetic patients with nephropathy comparing 12 months of treatment with either lisinopril or atenolol showed a reduction in albuminuria and TER_alb in patients treated with ACE inhibitors, whereas this was not seen in the patients treated with atenolol, even though reduction in BP was similar in the two groups (12). This finding indicates that ACE inhibitors have direct effects on the microvasculature, thus reestablishing the barrier function of the leaky capillaries in diabetes. This theory is supported by the fact that retinal blood vessels have receptors for angiotensin II (43) and that ACE inhibition reduces the permeability of these vessels (13,44). Our patients with DMa seemed to have a systemic microvascular hyperpermeability, the degree of which was reflected by retinal thickness. Because treatment with ACE inhibitors is known to have beneficial effects on UAE and TER_alb, treatment with ACE inhibitors or angiotensin II receptor antagonists could be a promising treatment modality for this group of patients. OCT might then be the instrument of choice to evaluate the efficacy of such intervention in reducing macular edema.

In conclusion, our data show that retinal thickness in the macular area is associated with a glomerular and generalized vascular leakage in type 2 diabetic patients with DMa. Therefore, hyperpermeability of blood vessels seems to coexist in different organs in diabetic patients, and this pathological feature might represent a common target for medical intervention. Quantitative assessment of retinal thickness, e.g., by OCT, may become a useful instrument in evaluating such intervention effects in DMa.

Financial support for this study was received from the Sehested Hansen Foundation, the Novo Nordisk Foundation, the Beckett Foundation, the Jochum Jensen Memorial Grant, and the Illum Foundation.

We thank Theis Bacher, Merete Møller, Hanne Petersen, Lisbet Vester, and Helle Hedegaard for their excellent technical assistance.