European Association for the Study of Diabetic Eye Complications

Annual Meeting Amsterdam May 30th to June 1st, 2008

Free Paper Abstracts


H. Zavrelova1, B.C.P. Polak1,2, A.C. Moll1,2, G. Nijpels1,2, J.M. Dekker1,2

1VU University Medical Center and 2 Institute for Research in Extramural Medicine, Amsterdam - The Netherlands

PURPOSE/BACKGROUND. Hyperglycaemia may exist for several years before symptoms develop in type 2 diabetes mellitus (T2DM) patients. Such patients can be detected by screening. However, it is not clear whether screening-detected T2DM carries a lower risk of secondary T2DM complications
such as retinopathy and ischemic heart disease to that of T2DM diagnosed in usual care. The authors studied this issue in the Hoorn Study.

METHODS . In the Hoorn Study, a population-based cohort study (n=2484, age 50-75) (QUERY: Are ranges correct throughout abstract?) of diabetes and diabetes complications, 95 patients were classified as having newly diagnosed T2DM after screening at baseline. The 10-year incidence of complications of these screening-detected patients was compared to the 10-year incidence of complications, as estimated from cross-sectional annual medical examinations in symptom-detected T2DM patients in usual care (n=1001). The regression coefficient of the relation between the proportion of complications and the T2DM duration was used to compute an estimated 10-year incidence of T2DM complications in the usual care detected T2DM patients. Both groups underwent fundus photography to assess retinopathy and ECG recording to assess ischemic heart disease.

RESULTS. Of the 95 screening detected T2DM patients, 22 (23%) had died, and 37 participated in the follow-up. The 10 year incidence of retinopathy and ECG abnormalities was 11.4% and 21.9%, respectively. For both retinopathy and ECG abnormalities no linear association of prevalence and T2DM duration was found in usual care detected T2DM patients. Thus, the 10-year incidence could not be estimated for this group and the risk difference between groups not computed.

CONCLUSIONS. In contrast to what was expected no linear association was observed between duration and progression of complications. The improvement of glucose control in a managed diabetes care system may explain these results.


E.S. Soto-Pedre1, M.C. Hernaez-Ortega2

1European Innovative Biomedicine Institute, Castro-Urdiales Spain.
2Hospital of Cruces, Baracaldo (Vizcaya) - Spain

PURPOSE. Several clinical practice recommendations for subsequent screening for retinopathy in patients with diabetes mellitus have been developed. However, these recommendations were mostly based on expert consensus opinion rather than direct evidence. This study estimated the most appropriate eye screening intervals for early detection of sight-threatening retinopathy (STDR) in patients with diabetes mellitus.

SETTING. Funduscopic results of patients attending Hospital of Cruces (Vizcaya, Spain) were reviewed between 1998 and 2004.

METHODS. A six-year retrospective follow-up study to review screening results of two cohorts of diabetic patients (i.e., cohort free of diabetic retinopathy and cohort with mild non-proliferative diabetic retinopathy at baseline) was conducted. Patients had been screened by means of a non-mydriatic retinal camera for the presence of retinopathy. Baseline information related age, sex and diabetes mellitus (type, duration,
treatment and metabolic control) was collected. Statistical analysis was based on life table method of risk estimation.

RESULTS. A total of 286 diabetic patients free of retinopathy and 144 patients with mild non-proliferative retinopathy at baseline entered the study. The probability of remaining free of STDR for patients free of retinopathy at baseline was 97% at the end of the fourth year, and those with type 2 diabetes
were more likely to progress to STDR (p<0.01). For patients with mild non-proliferative retinopathy at baseline, the probability dropped to 94% at the end of the second year and it was still high only for those patients with a glycated hemoglobin level less or equal to 7.5% at baseline.

CONCLUSIONS. The recommended interval for screening of STDR in patients free of diabetic retinopathy was 4 years in those who had type 1 diabetes and 3 years for those who had type 2 dia-
betes. In those who had mild non-proliferative retinopathy, the recommended screening interval was one year or less.


S. Koshy, E. O'Sullivan, D. De Alwis

Mayday University Hospital, CROYDON, United Kingdom

PURPOSE. To determine when it is safe to discharge patients with treated stable proliferative diabetic retinopathy back to retinal screening.

METHODS. The authors retrospectively looked at the notes of 88 patients who had pan-retinal photocoagulation [PRP] for proliferative diabetic retinopathy whose last treatment had been at least over 1 year previously. Notes were looked at one, two three and five years after PRP to determine if they
had further neovascular growth and/or vitreous haemorrhage after treatment had been considered adequate. Patients who no longer attended regular follow-up, attended other hospitals or were deceased were excluded.

RESULTS. Of the 88 patients, 35 were stable for 3 years and only 5 were not stable for 3 years. Of the 35 patients, 24 were male and 11 were female and their age ranged from 38 to 85 years. Insulin dependent diabetics were 24 and their duration of diabetes ranged from 6 years to more than 50 years. Reasons for laser were NVD for 29, NVE for 36, vitreous hemorrhage for 21 and rubeosis for 3 patients. Duration of laser treatment ranged from one month to 10 years. Total laser which the patients had ranged from less than 1000 to a maximum of 10534, the mean of which was 3266. After 3 years 4 had NVD and 4 had NVE. However they were stable. After 5 years, 27 were stable. One had mild vitreous hemorrhage, 7 did not have follow up and one died during the period.

CONCLUSIONS. At both one year and two years follow-up, after previously completed PRP, 5 patients needed further laser for re-growth of neovascularisation and/or vitreous haemorrhage. Out of the 35 patients who were stable for 3 years or more, only one had mild vitreous hemorrhage (which did not require treatment) during the five year period of follow-up. We conclude that it appears to be safe to discharge patients back to retinal screening if, after completed PRP, they have not had neovascular re-growth or vitreous haemorrhage for at least three years.


E.S. Stefánsson, E. ‘lafsdóttir University of Iceland, Reykjavík - Iceland

AIMS. To evaluate the safety of every-other-year eye screening for patients with diabetes without retinopathy.

M E T H O D S . Since 1994, patients with diabetes without retinopathy in Iceland have received eye screening every other year. 296 patients with diabetes who had no diabetic retinopathy in 1994/95 were followed with biennial eye examinations until they had developed retinopathy. The 10-year experience of this approach is reviewed.

RESULTS. Out of the 296 diabetic individuals, 172 did not develop diabetic retinopathy during the 10-year observation period. 96 patients developed mild non-proliferative retinopathy, six developed clinically significant diabetic macular oedema, 23 developed preproliferative retinopathy, and four developed
proliferative diabetic retinopathy during the 10-year observation period. All the patients who developed macular oedema or proliferative retinopathy had already been diagnosed as having mild nonproliferative retinopathy and entered an annual screening protocol before the sight-threatening retinopathy developed. No patient had any undue delay in treatment.

CONCLUSIONS. Every other year screening for diabetic eye disease seems to be safe and effective in diabetics without retinopathy. Such an approach will reduce the number of screening visits more than 25%. This reduces health costs and strain on resources considerably and relieves the patients with diabetes from unnecessary clinic visits and examinations.


T. Peto, I. Leung, N. Harris, W. Xing, C. Bunce

Moorfields Eye Hospital, London - United Kingdom

PURPOSE. The National Screening programme for diabetic retinopathy in the United Kingdom is likely to change the patient profile of diabetes eye clinics; only patients with complications are likely to attend, while others will remain in photographic screening. Patients with diabetic retinopathy requiring treatment are more likely to have other complications and these are likely to influence the resources needed to treat
them. The purpose of this study was to survey the complication profile of patients attending the Moorfields Diabetes Eye Clinics before screening was introduced in the neighbourhood.

RESULTS. Patients with diabetes were identified from clinic records during a 4 weeks period; those attending were interviewed by a team of trained interviewers. Overall, 830 patients were identified from the records, 4.3% were registered partially sighted and 4.7% registered blind. Altogether 580 interviews were conducted; the rest either did not attend (141), or could/would not be interviewed (109). Those interviewed were more likely to have had laser treatment in the past (p=0.016); there was no other significant difference between the groups. Mean duration of diabetes was 10 years (3 months to 52 years); 74.3% had Type 2 diabetes; 64.8% reported hypertension. Diabetes related complications were reported by a substantial proportion of patients, with diabetic neuropathy being the most prominent (48.9%), followed by
heart disease (24%), diabetic nephropathy (13.4%) and stroke (9.7%). On questioning, 13.1% did not understand the relationship between these complications, diabetes and diabetic retinopathy. On the day of the interview 5.8% had FFA and 13.4% had laser treatment with more than half of the patients expecting it to work immediately (51.6%).

CONCLUSIONS. High percentage of our patients had diabetes related complications, even before the implementation of the National Screening Programme in the area. With patients who require no intervention and have minimal diabetic eye disease being removed from the eye clinics, this is expected
to rise together with the percentage of patients requiring investigations and treatment. Multiple diabetes related co-morbidity will require special care from the eye units and the needs of the patients will need to be monitored carefully if quality of care is to remain high.


J. Grauslund1, A. Green2, A.K. Sjølie2

1University of Southern Denmark, Odense - Denmark
2Odense University Hospital, Odense - Denmark

PURPOSE/BACKGROUND. Diabetic retinopathy remains a leading cause of blindness in the working population in Europe. The aim of this study was to examine the prevalence of blindness and to evaluate the relationship between blindness and mortality in a 33-year follow-up of a population-based cohort of Danish type 1 diabetic patients.

METHODS. In the Danish County of Fyn, we have previously identified all insulin-treated diabetic patients with an onset before the age of 30 as of 1 July 1973 (n=727). In November 2006 survival status was evaluated by means of the Danish Central Office of Civil Registration and data on blindness was
collected from the Danish Association of the Blind. The Danish Association of the Blind is a voluntary organization open for all patients with a best-corrected visual acuity of the best eye at or below 20/200 (0.1).

RESULTS. Of the 727 patients identified in 1973 72 (9.9%) were at some point registered as blind. Median age at registration at the Danish Association of the Blind was 45.5 years (range 21-70 years) corresponding to a median duration of diabetes of 29.5 years (range 14-64 years). Cumulative incidence proportion of blindness was 11.9% and 8.0% (p=0.08) for patients with an onset before and after 1960, respectively. Cumulative incidence proportion of blindness was higher among women (11.8%) than men (8.5%) but this did not reach statistical significance (p=0.14). Although statistical significance was not reached, patients who became blind had a higher mortality (62.5%) than patients who did not (52.0%), p=0.09. For patients who became blind and died, median time from blindness to death was 10.7 years.

CONCLUSIONS. Despite recent improvements in care, blindness is still a major concern for diabetic patients. In a population-based cohort of Danish type 1 diabetic patients we found a 33-year cumulative incidence proportion of blindness at 9.9%. There was a tendency toward a higher cumulative inci-
dence proportion of blindness among women and among patients with an onset of diabetes before 1960, and a higher mortality was found in legally blind patients.


V.S. Vujosevic1, C.M. Cortese2, M.S. Miotto2, B.E. Benetti2,
M.E. Midena1,2

1Fondazione G.B. Bietti, IRCCS, Roma - Italy
2Department of Ophthalmology University of Padova, Padova Italy

PURPOSE. The cornea, the most innervated human tissue, may be affected by a specific diabetic neuropathy: corneal diabetic neuropathy (CDN), mainly involving the corneal subbasal nerve plexus.The aim of this study was to investigate, in diabetic population, the reproducibility and reliability of quantification of corneal subbasal nerve plexus parameters, using in vivo corneal confocal microscopy.

METHODS. Sixty two consecutive diabetic patients (107 eyes) were investigated using corneal confocal microscopy (Confoscan, Nidek, Japan). A validated technique for subbasal nerve plexus detection was applied. Corneal confocal microscopy parameters for subbasal nerve plexus evaluation were: number and density of nerve fibers, nerve tortuosity, nerve branching, and number of nerve beadings. Two masked examiners evaluated and quantified corneal confocal microscopy images. Peripheral diabetic neuropathy was also assessed using the Michigan Neuropathy Screening Instrument (MNSI). Fifty three normal subjects (96 eyes) served as controls.

RESULTS. Corneal confocal microscopy allowed in all cases a quantitative analysis of subbasal nerve plexus, and showed the presence of CDN even before peripheral diabetic neuropathy was detected by standard non invasive methods (p<0.001). No side effects were documented. Intra and inter–examiner agreement for confocal microscopy images were almost perfect (k= 0.95 and 0.92 , respectively). Signifi-
cant decrease of nerve beadings (p<0.001) and increase of nerve tortuosity (p<0.001) were the most distinctive parameters of CDN. Decrease of nerves density and fibers, and decrease of nerve branching were also documented (p< 0.001, <0.001 and <0.005, respectively).

CONCLUSIONS. Corneal confocal microscopy is a key diagnostic technique in evaluating and monitoring corneal disorders, namely corneal diabetic neuropathy. Quantification of corneal subbasal nerve plexus parameters allows a correct, reproducible and objective, in vivo, non invasive approach to CDN, allowing to characterize and follow (untreated and treated) peripheral diabetic neuropathy, a potentially highly
disabling complication of diabetes.


N.G.M. Wiemer 1,2, M. Dubbelman 2,3, P.J. Ringens 1,2, B.C.P. Polak1,2

1 Department of Ophthalmology, VU University Medical Center, Amsterdam - The Netherlands
2 Institute for Research in Extramural Medicine, VU University Medical Center, Amsterdam - The Netherlands
3 Department of Physics and Medical Technology, VU University Medical Center, Amsterdam - The Netherlands

PURPOSE. To examine the influence of diabetes mellitus (DM) type 1 on the thickness, radius of curvature, equivalent refractive index and power of the crystalline lens.

METHODS. 114 patients with DM type 1 and 75 healthy control subjects were examined by means of Scheimpflug imaging, to determine the thickness and shape of the lens, and with Hartmann-Shack aberrometry, to determine ocular refraction. The equivalent refractive index and the power of the lens were calculated from these parameters. Subjects with cataract or any other ocular pathology were excluded.

RESULTS. The lenses of the patients with DM type 1 were significantly thicker and more convex, compared to those of the control group (p<0.001). Furthermore, there was a significant decrease in the equivalent refractive index of the diabetic lenses, compared to the control group. The duration of DM type 1 was an important determinant of lens biometry; the independent effect of the duration of DM per year on lens
thickness, anterior radius, posterior radius and equivalent refractive index was respectively 95%, 88%, 207%, and 45% of the effect of age per year. Lens power and ocular refractive error were not affected by DM type 1.

CONCLUSIONS. The results of this study show that DM type 1 has a major impact on lens biometry. The decrease in equivalent refractive index of the lens appeared to compensate for the profound increase in lens convexity in patients with DM type 1, resulting in no significant changes in lens power or ocular refraction with the duration of DM.


K.H. Holm, M.L.A. Lovestam-Adrian

Institute of Ophthalmology, Lund - Sweden

PURPOSE. To evaluate the influence of hard exudates on macular function in patients with diabetic retinopathy.

METHODS . 37 eyes from 27 diabetic patients, aged 14-57 years, (QUERY: Are ranges correct through abstract?) diabetes duration 12.59 years, not previously treated with photocoagulation, underwent fundus photography, multifocal electroretinography (mfERG) and optical coherent tomography (OCT).

Hard exudates were graded from fundus photography with superimposed OCT and a hexagonal pattern (mfERG) by two independent retinal specialists, masked to mfERG and OCT results. We defined three groups; A=eyes with exudates in the analyzed zone and elsewhere, B=eyes with no exudates in the analyzed zone but elsewhere, and C=eyes with no exudates (n=15), which were controls. We compared the mfERG responses from four defined areas in the macula, between the three groups. The average thickness (m) assessed with OCT was measured in the corresponding zones and compared between the groups.

RESULTS . The implicit time was significantly prolonged in group A compared to group C in all four defined areas, 31.9msec vs. 29.1; p=0.045, 30.0 vs.29.1; p= 0.029, 33.4 vs. 30.3; p=0.024 and 31.3 vs. 29.3; p=0.016, respectively. A significantly prolonged implicit time was also seen in group B compared to group C in two areas, 33.3 vs. 29.1; p=0.007 and 30.7 vs. 29.3; p=0.032, respectively. Amplitudes differed between group A and C only in one area, 14.2 vs. 21.1; p=0.038, and between group B and C in one area, 13.7 vs. 20.1; p=0.045. Macular thickness assessed with OCT was similar between the groups.
CONCLUSIONS. In patients with diabetic retinopathy areas with hard exudates demonstrate a significantly prolonged implicit time assessed with mfERG. Our results indicate that hard exudates have an influence on macular function, irrespective of macular thickness.


E. Ganteris-Gerritsen, T. Missotten, J.P. Martinez

The Rotterdam Eye Hospital, Rotterdam - The Netherlands

PURPOSE. To evaluate the efficacy of intravitreal injection of Bevacizumab (Avastin) for the treatment of persistent diabetic macular edema.

METHODS. 75 patients with persistent, therapy resistant diabetic macular edema were treated with a 0.05-mL intravitreal injection containing 1.25 mg of bevacizumab. Visual acuity and retinal thickness measurement by optical coherence tomography (OCT) were evaluated at base line and follow-up.

RESULTS . At baseline, mean visual acuity was 0.25±0.18 Snellen letters. In the first 6 months the mean rate of injections was 31, in the following 6 months 11 injection were administered (n= 24). After 6 months the mean visual acuity increased with 1 line Snellen to 0.36±0.21 (p=0.048). After 1 year there was no significant improvement of visual acuity in comparison to baseline visual acuity (p=0.26). Injections after
6 months did not improve or decrease the visual acuity (0.050.1 Snellen) (p=0.48)

CONCLUSIONS. In patients with persistent diabetic macular edema not responding to photocoagulation (steroid injections) a small but significant improvement of visual acuity was observed 6 months after treatment with intravitreal injections of Avastin.


I. Klaassen1, J.M. Hughes1, C.G. Schalkwijk2, C.J.F. van Noorden1, R.O. Schlingemann1

1AMC, Amsterdam - The Netherlands
2Maastricht University, Maastricht - The Netherlands

PURPOSE. Disruption of the blood-retina barrier (BRB) is an early phenomenon in preclinical diabetic retinopathy (PCDR). Two pathways may be affected, the paracellular pathway involving endothelial cell tight junctions and transcellular pathway mediated by endocytotic vesicles (caveolae). The relative contribution of both pathways to vascular permeability in PCDR is unknown.

METHODS. Transcription levels of genes coding for proteins involved in endothelial cell tight junctions and vesicles were evaluated in intact rat retinas and cultured bovine retinal endothelial cells (BRECs) and pericytes (BRPCs). These were compared with transcription levels in rat retinas at 6 and 12 weeks after streptozotocin (STZ)-induced diabetes and in BRECs exposed to high levels of glucose or VEGF, using real-time quantitative RT-PCR. Protein expression was investigated in rat retinas by immunohistochemistry.
RESULTS. Paracellular transport-related mRNA and protein were specifically expressed by retinal endothelial cells, whereas vesicle transport-related mRNA and proteins were present in various retinal cell types, including endothelial cells. Expression of selected endothelial cell junction-related genes and particularly that of Ocln and Cldn5 was reduced in the diabetic retina and in BRECs but not BRPCs after exposure to high levels of glucose or VEGF. Expression of the majority of vesicular transport-related genes was upregulated after STZ-induced diabetes. Cav1 immunostaining was mainly vascular in vivo, but transcription levels were not upregulated in vitro. The endothelium-specific vesicle protein PV-1 was induced in diabetic retina and in BRECs by high levels of glucose and VEGF in vitro.

CONCLUSIONS. The alterations in gene expression indicate that diabetes-induced BRB permeability occurs not only by induction of paracellular leakage, but also by vesicle-mediated transport.


E. Beltramo, E. Berrone, S. Tarallo, M. Porta

University of Torino, Torino - Italy

PURPOSE/BACKGROUND. Thickening of basement membrane and pericyte loss are well-know early events in the pathogenesis of diabetic retinopathy. We have demonstrated that human retinal pericyte (HRP) apoptosis is strongly affected by matrix produced in high glucose (Beltramo E et al, 2007). The aim of this study was to verify if thiamine and benfotiamine are able to reverse this phenomenon, when added to high glucose in producing extracellular matrices (ECMs).

METHODS. Conditioned ECMs were obtained by growing human endothelial cells (HUVECs) for 7 days in culture media containing physiological (5.6 mmol/L) or high (28 mmol/L) D-glucose with or without 50 and 100 mol/L thiamine and benfotiamine. Cells were then lysed and ECM fixed by NH4OH. HRP were cultured in physiological glucose on these conditioned ECMs. Pericyte apoptosis was evaluated measuring DNA fragmentation by ELISA; the expression of two molecules known to be involved in glucose-mediated apoptosis, Bax with a pro-apoptotic and Bcl-2 with a pro-proliferating function, was measured by RT-PCR and their concentration by ELISA.

RESULTS. Apoptosis in HRP was greatly enhanced by high glucose-conditioned matrix in comparison with physiological glucose ECMs (1.230.48 abs units vs 0.060.02, p=0.002) and this was confirmed by Bax and Bcl-2 mRNA expression (Bax: 193.843.0% of physiological glucose, p=0.008; Bcl-2: 50.313.0%, p=0.001) and protein concentration (Bax: 36.189.8 vs 17.6613.00 ng/mg total proteins, p=0.001; Bcl-2: 0.660.22 vs 1.290.31 ng/mg total proteins, p=0.002). Both thiamine and benfotiamine were able to reverse this effect:
DNA fragmentation in HRP cultured on ECMs produced by HUVECs in high glucose plus thiamine and benfotiamine at both concentrations was fully normalized, as well as Bax and Bcl-2 expression and concentration.

CONCLUSIONS. Thiamine and benfotiamine are able to counteract the damaging effects of high glucose, probably reducing the glycation of matrix proteins, which, in turn, can lead to damaging interactions between pericytes and their surrounding microenvironment and, finally, to the onset of diabetic retinopathy.
REFERENCES. Beltramo E, Nizheradze K, Berrone E, Tarallo S, Porta M. High glucose conditioned matrix enhances apoptosis of human retinal pericytes. Eur J Ophthalmol 2007; 17: 466 (abstr).


E. Beltramo, E. Berrone, S. Tarallo, M. Porta

University of Torino, Torino - Italy

PURPOSE/BACKGROUND. To investigate the pathophysiological roles of pericytes, animal cells have been used so far for in vitro studies. We have previously demonstrated that human retinal pericyte (WT-HRP) are more vulnerable to fluctuating glucose concentrations than bovine retinal pericytes (Bel-
tramo E et al, 2006). Our aims were to establish an immortalized human retinal pericyte cell line and to compare its behaviour with that of wild-type human pericytes, after pulsed exposure to high glucose concentrations.

METHODS. WT-HRP were immortalized by electroporation with a plasmid vector containing the Bmi-1 oncogene, that induces telomerase activity. These cells (Bmi-HRP) were characterized for typical pericyte morphology and marker expression, and evaluated for senescence-associated-galactosidase activity. Subsequently, Bmi-HRP were grown in physiological/high glucose for 7 days, and then in physiological glucose for another 24, 48 or 72h. HRP (wild-type and immortalized) were also kept intermittently at 48h intervals in high/normal glucose for 8 days. DNA fragmentation, Bcl-2 and Bax mRNA expression and protein concentration, as markers of glucose-induced apoptosis, were determined.

RESULTS. The oncogene Bmi-1 transfection resulted in the establishment of a permanent pericyte cell line (Bmi-HRP) which showed telomerase activity and facilitated propagation, maintaining the same morphology and typical pericyte markers. Apoptosis in Bmi-HRP increased within 24h of physiological glucose re-entry, but not with continuous exposure to high glucose, as already shown in WT-HRP. Bcl-2/Bax expression/concentration results were consistent with DNA fragmentation in HRP. Intermittent exposure to high glucose increased apoptosis in both type of HRP.

CONCLUSIONS. Bmi-HRP have the structural and functional properties characteristic of WT-HRP, suggesting a role for this cell line in studying retinal pathophysiology. Moreover, Bmi-HRP behave like WT-HRP when exposed to fluctuating glucose levels, reinforcing the hypothesis that daily blood
glucose fluctuations play a major role in the development of diabetic retinopathy.


M. Lorenzi, G. Feke, F. Berisha, J. Kolodjaschna, L. Pitler

Harvard Medical School, Boston - USA

PURPOSE. We seek to identify early retinal abnormalities that can function as risk markers for diabetic retinal microangiopathy, and thus as informative surrogate endpoints to test drugs for prevention.

METHODS. To focus on the abnormalities that are not prevented by the best efforts with current means of controlling hyperglycemia, we studied patients with type 1 diabetes (age 31±9 years, 16F/11M, diabetes duration 12±6 years) on intensive insulin treatment and free of retinopathy. We evaluated retinal hemodynamics (diameter, blood speed, and blood flow in the superior temporal retinal artery) by laser Doppler, foveal thickness by optical coherence tomography, and macular sensitivity by microperimetry (i) at steady-state in comparison to matched nondiabetic controls; (ii) at steady-state in longitudinal observations, and (iii) in response to a stimulus –a postural change from sitting to reclining, which produces an increased blood pressure at the entrance to the retinal circulation.

RESULTS. The retinal measurements at steady-state were similar in the diabetic and control group. Diabetic patients showed over 12 months a decrease in HbA1c from 7.6 1.0 to 7.0 0.7% (p=0.002), and an increase in foveal thickness (from 217±22 to 221±20 m, p=0.01). In response to reclining, only 9 of the 15 diabetic patients tested to date showed a normal decrease in arterial diameter (-6.4±3.8%); the remaining 6
showed no change or an increase. All patterns were repeatable. The two subgroups of diabetic patients differed only with respect to the longitudinal behavior of retinal arterial diameter: unchanged over 12 months in the subgroup with normal vasoconstriction but increased (from 119±8 to 125±11m, p=0.05) in the subgroup without vasoconstriction.

CONCLUSIONS. Defective retinal vasoconstriction in response to reclining precedes clinical microangiopathy in a subgroup of well-controlled type 1 diabetic patients. Insofar as the abnormality is qualitative, reproducible, and occurs in response to a vessel-intrinsic stimulus and in association with the widening of arterial diameter, it is a candidate risk marker for retinal microangiopathy. Studies of mechanisms and reversibility are in progress to determine suitability as a surrogate endpoint.


A. Lecleire-Collet, A. Erginay, T. Meas, P.J. Guillausseau, P. Massin

Lariboisière Hospital, Paris - France

PURPOSE. The pathogenesis of diabetic retinopathy is complex and still remains incompletely understood. The diabetic retinopathy is usually considered as a retinal microangiopathy. However, an increasing number of studies suggest that the impairment of retinal neural cells may precede the earliest signs of retinal vascular impairment. The aim of the study was to assess in a controlled study the retinal functional abnormalities in diabetic patients without clinical appearance of retinopathy.

METHODS. Twenty-eight patients with diabetes (11 patients with type 1 diabetes, 17 patients with type 2 diabetes) without diabetic retinopathy on fundus examination, and 28 age and sex-matched healthy volunteers were included. Each subject underwent a complete ophthalmologic examination, a colour vision test (Lanthony D-15 desatured colour test), a contrast sensitivity test (Peli-Robson test), a pattern-ERG, a
full-field ERG, and a multifocal ERG, according to the ISCEV guidelines. The comparison of the results between the 2 groups were assessed using the Student-t test. p< 0.05 were considered as the level of significance.

RESULTS. In diabetic patients, the total colour difference score at the D-15 desatured colour test was significantly higher than in healthy volunteers. The implicit times of the oscillatory potentials were significantly longer and the oscillatory potential amplitudes were significantly lower in diabetic patients
as compared with the healthy volunteers. The implicit times of the N35 and N95 waves at the pattern-ERG were significantly longer, and the amplitudes of the P50 and N95 waves were significantly lower; the b-wave implicit time at the rod ERG and the a and b- wave implicit times at the single-flash cone ERG were significantly longer in diabetic patients as compared with the healthy volunteers.

CONCLUSIONS. Retinal functional abnormalities occurred in diabetic patients before the clinical appearance of diabetic retinopathy. The most early impaired retinal cells seem to be the amacrine and bipolar cells, the ganglion cells and cone photoreceptors cells. These results suggest that an early
neurodegenerative process, which could precede the vascular impairment, may occur in diabetic retinopathy


F.D. Verbraak1, H.W. van Dijk1, P.H.B. Kok1, M. Abramoff 2

1Academic Medical Center, Amsterdam -The Netherlands
2University of Iowa, Department of Ophthalmology, Iowa USA

PURPOSE. A comparison of thickness measurements of segmented optical coherence tomography (OCT) derived topographic maps in patients with diabetes mellitus (DM) and no or minimal diabetic retinopathy (DR) versus healthy controls.

METHODS. Ninety nine patients, 59 with type 1 DM, and 40 with type 2 DM, with no or minimal DR underwent full ophthalmic examination, fundus photography and OCT (StratusOCT, Model 3000, Carl Zeiss Meditec, Dublin, CA, USA, software version 4.0.1) using the fast macular thickness protocol. Following au-
tomated segmentation mean thickness was calculated for 5 layers: A/ Retinal Nerve Fibre Layer (RNFL), B/ Ganglioncell layer(GCL) + Inner Plexiform Layer (IPL), C/ Inner Nuclear Layer + Outer Plexiform Layer, D/ Outer Nuclear Layer + Inner Segments (photoreceptor), E/ Outer Segments (photoreceptor) in the ETDRS defined regions of the macula and compared to 100 age and sex matched normal controls.

RESULTS. In diabetic patients both the RNFL and the GCL + IPL were significantly thinner in the pericentral area (area 2 to 5 ETDRS region) compared to the controls. The largest averaged difference occurred at 20 A-scans from the fovea, with a mean of 6 micron (p < .05). All other areas, and all other
layers did not show a significant difference.

CONCLUSIONS. The previously described decreased total retinal thickness in the pericentral area in diabetic patients with no or minimal retinopathy is due to a selective loss of thickness in the inner retinal layer (RNFL and the GCL + IPL), and supports the concept of early DR as a neurodegenerative disease.


W3C standards logos

site developed by Netsima Ltd »