European Association for the Study of Diabetic Eye Complications

Annual Meeting Coimbra 27-29th May 2005

Paper Sections 1 - 20



Astakhov Y.S., Shadrichev F.E., Sivas Y.Y. State Medical University named after I.P.Pavlov, St Petersburg, Russia

PURPOSE: To study risk factors of diabetic retinopathy (DR) progression during 3 years after the switch to insulin therapy (IT) in patients with diabetes Type 2.

METHODS: 215 diabetic patients (40-79 years old, 82,3% - female, duration of diabetes – up to 34 years, median body mass index (BMI) 27,1 kg/m2) with indications for IT were examined before the switch to insulin and each 12-month after, during 3 years. The study included laboratory tests, binocular ophthalmoscopy and seven-field stereo fundus photography. Retinopathy was graded according to the ETDRS scale. To assess the relationship between retinopathy and covariants, logistic regression analysis (SPSS 10.1 for Windows, 2000) was used.

RESULTS: During the observation period of 3 years, 73,6% of patients progressed in the retinopathy scale; 11,4% progressed more than 2 levels. The highest progression of DR was registered in the 1st year of IT – 36,4%, 22,6% in the 2nd and 23,6% in the 3rd. Using stepwise logistic regression analysis we found that progression of retinopathy during the first year of IT was related to systolic hypertension (P=0,004) initial degree of DR (P=0,006), HbA1c level (P=0,005), cholesterol level (P=0,008) and the duration of diabetes (P=0,01). During the second year, DR progression correlated mostly to BMI level (P=0,01) and during the 3rd year – to arterial hypertension (P=0,004), cholesterol level (P=0,007) and BMI level (P=0,02).

CONCLUSIONS: The risk factors of progression of diabetic retinopathy in Type 2 diabetic patients during 3 years after the institution of insulin treatment include: systolic hypertension, severity of DR at the beginning of the research, HbA1c level, cholesterol level, the duration of diabetes and BMI.


Monica Lövestam-Adrian; C Hansson-Lundblad; O Torffvit.
Ophthalmology Department,University Hospital, Lund, Sweden

Purpose. To study associations between diabetic retinopathy and development of stroke, myocardial infarction and death.

Methods. During a 10yr observation period, 363 Type 2 diabetic patients (diagnosis>30 yrs) attending an outpatient clinic were studied regarding the prevalence and incidence of retinopathy and associated risk factors (HbA1c, blood pressure, albuminuria, creatinine, age, sex and duration) in relation to the development of myocardial infarction, stroke and death. The degree of retinopathy was classified as no retinopathy, background or sight-threatening retinopathy, i.e. clinically significant macular edema, severe nonproliferative or proliferative retinopathy.

Results. During the study period 62 patients had had mycocardial infarction, 54 stroke and 99 patients died. Patients with sight-threatening retinopathy at baseline (n=41) had 2.2 (P<0.01) times higher risk for death than patients with no or background retinopathy, even when controlled for medical riskfactors (Cox regression analysis). Patients with no retinopathy at baseline (n=226) who developed no or background retinopathy (n=187) had 5.4 times higher risk for death (P=0.007), (adjusted for medical riskfactors), compared to patients who developed sight-threatening retinopathy (n=39) and patients who developed no or background retinopathy had higher incidence of myocardial infarction (P=0.049) (disappeared when adjusted for medical riskfactors), than patients who developed sight-threatening retinopathy. More patients who developed sight-threatening retinopathy were treated with ACE inhibitors for hypertension than patients who did not (41% vs.24%; P=0.030).

Conclusion. In spite of higher levels of medical riskfactors, patients who developed sight-threatening retinopathy had lower risk for myocardial infarction or death compared to patients with no or background retinopathy at follow-up. ACE inhibitors might contribute to the longer survival.


A ERGINAY (1), P MASSIN (1), N ROBERT (1), AC BOSSU (1), C MORENO (1), S HUGEROT (1), M CAHANE (2), J TICHET (2,3), I MERCAT-CAUDAL (2,3), JJ ROBERT (2), G REACH (2), A GAUDRIC (1) (1) Ophthalmology Department, Lariboisière Hospital, University Paris 7, Paris, France (2) Young Diabetics Association (AJD), France (3) Inter Regional Institut for the Health (IRSA), France

Purpose: To estimate the prevalence of diabetic retinopathy (DR) in children and adolescents with type 1 diabetes.

Methods: 504 children and adolescents (250 boys et 254 girls), aged 10-18 years (mean13+/-2years), were screened for DR using nonmydriatic photography, during their stay in a holiday camp for young diabetics. Five 45° digital colour images of 5 overlapping fields were taken for each eye: one image was centred on the macula, including the optic disc and one each on the nasal, temporal, superior and inferior fields. The images, recorded on compact discs were sent to the Lariboisiere Hospital reading center for independent grading by two ophthalmologists.

Results: All images were of excellent quality. Non proliferative mild DR was diagnosed in 22 children (4.4%) (10 boys, 12 girls). Mean diabetes duration was 4.8 ± 3.4 years. Mean HbA1C was 8.4 ± 1.5%. Prevalence of DR was highly correlated to the age of the chidren (p=10-5) with a prevalence of 1% before 14 years, 5.2% between 14 to 15 and 17.7% between 16 to 18, to the duration of diabetes (p=0,02) and non-significantly to glycemic control (p=0.08). When the 22 children with DR were compared with 66 diabetic children matched for age, the only factor correlated with the presence of DR was the duration of diabetes.

Conclusion: Fundus photography is a sensitive and effective tool to detect DR in young type 1 diabetic patients. Early detection of DR in adolescents is very important, as it allows identification of patients with a high-risk of progression towards severe stages of DR and justifys more frequent ophthalmological examinations for them.


Manon vonHecke1,2, JM. Dekker 2, C.D.A. Stehouwer 3, B.C.P. Polak 1, J Fuller 4, AK Sjolie 5, A Kofinis 6, R Rottiers 7, M Porta 8, N Chaturvedi 9 and the Eurodiab Prospective Complications Study Group 1 Dept. of Ophthalmology and 2 Institute for Research in Extramural Medicine, VU University Medical Center, Amsterdam, the Netherlands; 3 Dept. of Internal Medicine, Academic Hospital Maastricht, Maastricht, the Netherlands; 4 Dept. of Epidemiology and Public-Health, Royal Free and University College London Medical School, London, UK; 5 Dept. of Ophthalmology, Odense University Hospital, Odense, Denmark; 6 Diabetic Centre, Hippokration Hospital, Athens, Greece; 7 Dept. of Endocrinology, University Hospital Gent, Gent, Belgium; 8 Dept. of Internal Medicine, University of Turin, Turin, Italy; 9 International Centre for Circulatory Health & National Heart & Lung Institute, Imperial College London, London, UK.

Purpose To study the relationship of non-proliferative and proliferative retinopathy with all-cause mortality and cardiovascular disease incidence in type 1 diabetic patients and additionally the role of cardiovascular risk factors in these associations.

Methods This prospective study included 2237 type 1 diabetic patients from 31 centers in 16 European countries at baseline, aged 15-60 years, who were examined for retinopathy, by taking two-field 45º fundus photographs, which were centrally graded. Mortality and cardiovascular morbidity follow-up was assessed 6-8 years after baseline examination according to a standardized protocol.

Results After 7.9 years of follow-up, 64 patients had died and 128 patients had incident cardiovascular disease. The age- and sex-adjusted hazard ratios of all-cause mortality were 1.45 (95% CI: 0.71-2.96) and 4.16 (1.96- 8.84) in patients with non-proliferative and proliferative retinopathy at baseline, respectively. Adjustments for cardiovascular risk factors completely obliterated the association with non-proliferative retinopathy, whereas the association with proliferative retinopathy remained 2-fold increased, although non-significant. The age- and sex- adjusted hazard ratios of incident cardiovascular disease were 1.73 (1.15-2.60) and 2.05 (1.22-3.45) in patients with non-proliferative and proliferative retinopathy, respectively. After adjustments for cardiovascular risk factors both associations were attenuated and lost statistical significance.

Conclusions This study shows that type 1 diabetic patients with non-proliferative or proliferative retinopathy have an increased risk for all-cause mortality and incident cardiovascular disease. The presence of cardiovascular risk factors explained the associations to a large extent, except for the associations with proliferative retinopathy, which suggests that other shared mechanisms may be involved.


John Conrath, Ali Erginay, Amélie Lecleire-Collet, Pascale Massin Ophthalmology Department, Hôpital Lariboisière, Paris

Purpose: To compare the effect of the JPEG and JPEG2000 compression algorithms on the detection of diabetic retinopathy lesions.

Methods: 45 color fundus photographs obtained with a digital nonmydriatic retinal fundus camera were saved in the uncompressed TIFF format. These images were graded jointly by two retina specialists at a one month interval (to assess intragrader variability and establish a gold standard) for: soft exudates, hard exudates, clinically significant macular edema, newvessels, intraretinal microvascular abnormalities (IRMA) and retinal hemorrhages and/or microaneurysms (HMA). They were compressed at ratios 1:11, 1:22, 1:31 and 1:47 by both algorithms and 1:53 by the JPEG algorithm. Then, all 405 compressed images were placed in random order and graded again jointly by the two retina specialists. Subjective image quality was graded and sensitivity, specificity, positive and negative predictive values and the kappa statistic were calculated for all parameters at all compression ratios.

Results: The 1:11 ratio showed no evident effect on image quality and kappa values were high (0.94-1). Subjective image degradation became important at the 1:47 ratio for both algorithms with artefacts (“blocking” for JPEG and “rice-grain” for JPEG2000) observed. At high compression levels, IRMA and HMA detection were most affected, with JPEG2000 performing slightly better than JPEG.

Conclusions: The 1:11 compression ratio does not alter detection performance for diabetic retinopathy lesions, using either JPEG or JPEG2000 algorithms. The 1:22 ratio is still acceptable with both performing similarly; higher compression ratios show slightly better results with JPEG2000 compression, but may be insufficient for screening purposes.


G.J.M. Tangelder, G.L. Van der Heijde, M.S.U. Suttorp-Schulten, B.C.P. Polak, P.J. Ringens. VU University Medical Center,Amsterdam , The Netherlands

Purpose: To evaluate measurement precision of retinal thickness assessment with Optical Coherence Tomography (OCT) and the Retinal Thickness Analyzer (RTA) in healthy volunteers and diabetic patients with and without macular edema. To this end retinal thickness measurement precision was expressed as 95% limits of agreement (LA95%).

Methods: Retinal mapping of the posterior pole was performed in 28 eyes of 14 healthy volunteers, 49 eyes of 25 patients without macular edema and 17 eyes of 9 patients with macular edema. Each subject was measured 3 times with both instruments within one hour on the same day. The acquired retinal thickness maps of each instrument were divided into 9 ETDRS-like areas and the average retinal thickness was calculated for each area. The intra-individual standard deviation of 3 measurements was used to calculate the LA95% for each area as measured with both instruments.

Results: The LA95% of the OCT was significantly smaller for all areas and for all groups in comparison to the RTA, except for foveal thickness (p = 0.07) and superior perifoveal thickness (p = 0.06) in diabetes patients with macular edema. The average difference in measurement precision was 25.8 µm in healthy volunteers, 20.4 µm in diabetes patients without macular edema and 21.1 µm in diabetes patients with macular edema.

Conclusion: Measurement precision of the OCT is greater than the RTA, making it a more suitable instrument for retinal thickness analysis.


Nik M. Mustapha, Bahaedin M. Ben-Mahmud, Giovanni E. Mann, Eva M. Kohner, Rakesh Chibber Cardiovascular Division, GKT School of Biomedical Sciences, King's College London, Guy's Campus, London SE1 1UL.

Purpose: One of the earliest changes in the development of retinopathy in diabetes is accelerated apoptosis of pericytes and the formation of acellular capillaries. The mechanism by which glucose causes apoptosis of pericytes is still unclear, but recent evidence suggests the involvement of oxidative stress. It has been proposed that reactive oxygen species (ROS) production is enhanced by glucose through various mechanisms such as glycation, polyol pathway and increased release from mitochondria. In contrast, increasing evidence suggests that NADPH oxidase is the most important source of ROS production in blood vessels. The aim of the present study was to examine if NADPH oxidase-derived ROS contributes to the loss of pericytes in early diabetic retinopathy.

Methods: Cultured bovine retinal capillary pericytes (BRP) were exposed to normal (5.6mM) and high (25mM) glucose. Western blot and immuno-gold staining were used to localise NADPH oxidase. Apoptosis was detected by measuring caspase-3 activity and DNA fragmentation. Intracellular glucose concentration was measured using Amplex Red™ glucose assay kit. Protein oxidation was determined by Oxyblot™ protein oxidation detection kit.

Results: Western blot analysis demonstrated for the first time that pericytes express the important subunits of NADPH oxidase, gp91-phox and p47-phox. Ultrastructural analysis using immuno-gold staining demonstrated that gp91-phox is mostly localised in the cytosol. The expression of gp91-phox, NADPH oxidase activity and the level of intracellular protein oxidation were increased in BRP exposed to high glucose (25mM). The level of apoptosis was also increased in BRP maintained in high glucose. The addition of apocynin (500µM), an inhibitor of NADPH oxidase assembly but not MitoQ 10 nitrate (1 µM), a mitochondrial-targeted antioxidant, significantly reversed glucose-induced caspase-3 activity (77.8 5 ± 5.9 % vs. 121.5 ± 7.3 % in high glucose without apocynin).

Conclusion: Our present results suggest that NADPH oxidase plays a key role in glucose-induced apoptosis of retinal pericytes in early diabetic retinopathy. Acknowledgements: Funded by Malaysian Government


KH Jensen*, A Hessellund*, C Aalkjær#, J.D.C. Lambert#, and T Bek*. *Dept. of Ophthalmology, Aarhus University Hospital,#Dept of Physiology, Aarhus University, Denmark

Purpose: Retinal hyperperfusion due to deficient tone in retinal arterioles is a basic phenomenon involved in the pathophysiology of diabetic maculopathy. A precondition for understanding this pathological hyperperfusion is to have an understanding of tone regulation in retinal arterioles. This understanding can be obtained by in vitro studies of retinal arterioles. However, previous studies have shown that the presence of perivascular tissue in the in vitro setup causes the retinal arterioles to loose responsiveness to vasoconstrictors. Therefore, in order to interpret studies of retinal vascular tone regulation, it is important to elucidate the physiological role of this phenomenon.

Methods: Porcine retinal arterioles (diameter ~ 150 µm, length ~ 1.8 mm) with a 2 mm patch of adjacent tissue on each side of the vessels were mounted in a small vessel wire myograph for isometric tone measurements. Dose-response experiments were carried out with the vasoconstrictor U46619 in seven concentration steps between 1.0E-9 M and 3.0E-6 M before and after removal of the perivascular retina, and it was tested whether the response was affected by the glutamate receptor subtype agonist NMDA (100 µM, n=7), The NMDA antagonist DL-APV (50 µM, n=32), and the non-selective COX inhibitor Ibuprofen (10 µM, n=16).

Results: The presence of retinal tissue impaired the contracting effect of U46619. This impairment was eliminated by blocking of the glutamate NMDA (p<0.01) receptor and inhibition of COX (p<0.01), but could be reversed by NMDA (p<0.01).

Conclusion: The contractile effect of U46619 in retinal arterioles is decreased by a pathway that involves the glutamate NMDA receptor subtype and COX. In vitro studies of tone regulation in retinal arterioles should take into account that the results may be affected by substances released from the surrounding retina. This has implications for the elucidation of the background for early hyperperfusion in diabetic maculopathy.


E Beltramo, E Berrone, S Giunti, G Gruden, P Cavallo-Perin, M Porta. Department of Internal Medicine, University of Turin, Turin, Italy.

Purpose: Pericyte loss is one of the first events of diabetic retinopathy. We and others have previously showed that high glucose concentrations can induce apoptosis in cultured retinal pericytes. Both systemic and capillary hypertension are also believed to be important in the onset and progression of diabetic retinopathy (DR). The haemodynamic insult of retinal capillary hypertension can be mimicked by exposing retinal pericytes to mechanical stretch in vitro. Therefore, we investigated the effect of stretch combined with high glucose on both pericyte proliferation/apoptosis and morphology.

Methods: Bovine retinal pericytes (BRP), cultured in either normal (5.6 mmol/L) or high (28 mmol/L) glucose concentrations in six-well flexible-base plates (Flex I), were exposed or not for 48/72 hours to mechanical stress (10% elongation, 60 cycles/min). Cell replication was determined by cell count and BrdU incorporation; DNA fragmentation was measured by ELISA to assess cell apoptosis; cell morphology and cytoskeleton distribution were evaluated by FITC-conjugated-phalloidin.

Results: Reduction of cell proliferation and increase in apoptosis were confirmed in the presence of high glucose alone. When cells were subjected to stretch, proliferation was significantly reduced in both normal and high glucose in comparison with non-stretched controls, at 48 as well as at 72 h; apoptosis was increased in both normal and high glucose at 72h. In both cases, an important synergic effect of hyperglycaemia combined with stretch was shown. Cell morphology showed great modifications of the cytoskeleton in all experimental conditions; in particular, cells subjected to stretch showed a clear elongation and traslocation of actin fibers.

Conclusions: Our results show that stretch, alone or combined with high glucose, reduces cell proliferation, increases apoptosis and induces morphological changes in pericyte cytoskeleton. Further elucidations of the molecular mechanisms at the basis of reduced proliferation of pericytes subjected to high glucose and stretch could be useful to investigate the effects of combined hyperglycaemia and hypertension in the pathogenesis of DR.


Elena Berrone, Elena Beltramo, Massimo Porta. Dept. Internal Medicine, University of Turin, Turin, Italy.

Purpose: High glucose induces pathological alterations in small and large vessels, possibly through increased formation of AGE, activation of aldose reductase and protein kinase C, and increased flux through the hexosamine pathway. We showed that thiamine (T) and its lipophilic analogue benfotiamine (BT), acting as a coenzyme for transketolase (TK), are able to correct the increased apoptosis due to high glucose in endothelial cells and retinal pericytes. Moreover, it has been demonstrated that treating diabetic rats with high doses of BT can prevent diabetic retinopathy, through the inhibition of these pathways. An increase in hexokinase (HK) expression may play a role in the regulation of the levels of glucose in the vascular cells. The aim of this study was to verify if T and BT can modify intracellular glucose levels, TK activity and HK expression in bovine retinal pericytes (BRP) cultured in high ambient glucose.

Methods: BRP were cultured for 7 days in normal (5.6 mmol/l) or high (28 mmol/l) glucose concentrations, with or without T or BT (50 or 100 µmol/l). Intracellular glucose concentration was determined by a fluorescent ELISA method, HK mRNA expression by relative quantitative RT-PCR, while TK activity was measured spectrophotometrically by decrease of NADH.

Results: Intracellular glucose levels, as well as HK mRNA, were increased in BRP cultured in high glucose concentrations, in comparison with those grown in normal glucose, while TK activity was increased only in the combined presence of high glucose and T or BT. Both T and BT, at either concentration, when added to high glucose, were able to reduce significantly intracellular glucose concentrations and HK mRNA expression.

Conclusions: T and BT reduce HK mRNA expression and intracellular glucose concentrations, while increasing TK activity, in retinal pericytes cultured in high glucose. The activation of TK by T and BT could alleviate the hyperglycaemic stress in the cells, diverting the excess of intracellular glucose and its reactive metabolites towards the pentose phosphate pathway.


AF Ambrósio (1,2),EC Leal(1,3),A Maninnavan (4),C Aveleira (1,A Serra (1),A Castilho (1),T. Terasaki(6),K. Hosoya (7),M Cotter(5),JV. Forrester(3) 1Center Ophthalmology Coimbra, IBILI, Faculty Medicine, University Coimbra, Portugal, 2Center Neuroscience Coimbra, Portugal; 3Ophthalmology, 4Biomedical Physics & Bioengineering, 5Biomedical Sci, Sch Med Sci, University Aberdeen, UK, 6Grad Sch of Pharm Sci, Tohoku Univ, Sendai, Japan, 7Fac Pharm Sci, Med and Pharm Univ, Toyama, Japan.

Purpose: Inflammation is believed to be involved in the development of diabetic retinopathy (DR). Some evidences showed that inflammatory mediators, such as nitric oxide (NO), are involved in blood-retinal barrier (BRB) breakdown. We aimed to study the role of NO in leukocyte adhesion to retinal vessels and in BRB breakdown in DR.

Methods: Leukocytes, from spleens of normal mice, were labelled with calcein-AM and transferred to normal, diabetic, or diabetic-treated (with NO synthase inhibitors, L-NAME or aminoguanidine) mice. After Evans Blue infusion, flat mounted retinas were examined for leukocyte adhesion and vessel leakage under confocal microscope. In vivo leukocyte adhesion was assessed by scanning laser ophthalmoscope (SLO). Endothelial cells (TR-iBRB2) were incubated with high glucose or NOC-18. We quantified the levels of adhesion molecules (ICAM-1, V-CAM-1) and tight junction proteins (ZO-1, occludin) by flow cytometry and western blotting, respectively. In vitro leukocyte adhesion was assessed by fluorescence assay.

Results: BRB permeability and leukocyte adhesion were increased in diabetic mice. L-NAME, but not aminoguanidine, prevented the diabetes-induced leukocyte adhesion. Also, high glucose and NOC-18 increased the adhesion of leukocytes to endothelial cells, and NOC-18 increased ICAM-1 levels. VCAM-1 was not detected in TR-iBRB2. Occludin expression was not altered by high glucose or NOC-18, but high glucose decreased ZO-1 levels, in endothelial cells.

Conclusions: These results show the involvement of NO in leukocyte adhesion to retinal vessels in diabetic mice, also supported by in vitro studies. BRB permeability is also increased in diabetic mice and NO blockade decrease BRB permeability. However, NO did not alter thigh junction protein levels in endothelial cells. (Support: FCT-POCTI/CBO/38545/2001; SFRH/BD/9686/2002 and FEDER)


BL Justesen(1), P Mistry (2), SA Thom(2,N Witt(2) DM Köhler(3),AD Hughes (2) and AK Sjølie (1) 1,Department of Ophthalmology, Odense University Hospital, Denmark; 2International Centre for Circulatory Health, St. Mary's Hospital, London, UK; 3Department of Anaesthesiology and Intensive Care, Odense University Hospital, Denmark.

Purpose: The geometric arrangement of the vasculature determines the efficiency of blood flow and nutrient delivery, and deviations from an optimum design may confer fundamental disadvantages to the circulation. A previous study has shown, that by breathing 3 test gasses (different oxygen concentrations and hypercapnia), patients with hypertension have decreased vascular reactivity compared to normotensives. This method has not previously been used in relation to diabetes. The purpose of this paper is to apply the method for analysing geometry of retinal vessels to investigate the vascular reactivity in people with compared to people without diabetes.

Methods: The 61 participants, 29 with and 32 without diabetes inhaled 3 test gasses: hypoxia, hypercapnia and hyperoxia. The diameter of arteries and veins was measured from digital retinal images taken before and after inhaling the gasses using image analysis software.

Results: There was no significant difference between diameter of the arteries (p=0.38) and veins (p=0.16) in people with and without diabetes before they inhaled the test gasses. Hypoxia caused an increase in diameter for arteries of 1.73 % for people with diabetes and 2.94 % for people without. The increase in veins was 1.54 % and 2.92 % respectively, although this difference was not statistically significant. Hypercapnia only caused minor changes, overall a small increase in diameter. Hyperoxia caused a significantly different decrease in diameter for arteries, of 3.95 % and 7.75 % for people with and without diabetes, respectively (p=0.01). In the veins the difference was not statistically significant, 9.53 % and 11.44 % respectively.

Conclusions: When exposed to hyperoxia, arteries of people with diabetes do not contract as much as arteries in people without diabetes. This may be caused by an impaired auto regulation in the vessels of people with diabetes, leading to a lesser response to the stimuli of the test gasses. The geometric analysis gives a non- invasive method for testing the reactivity of the retinal vessels and make comparisons between people with and without diabetes.


P Jeppesen, ST.Knudsen, P Løgstrup,CE Mogensen,O Schmitz, T Bek Dept. of Ophthalmology, Aarhus University Hospital, Aarhus, Denmark

Purpose: To study the influence of acute hyperglycaemia on the myogenic response in large retinal arterioles from healthy persons.

Method: Nine healthy persons were examined in a double-blinded randomized study in the morning and in the afternoon on two days separated with at least one month. On all study days the blood glucose was controlled using clamp technique with infusion of somatostatin, insulin and glucose to either a level of normoglycaemia (5 mmol/l) or hyperglycaemia (15 mmol/l). In five of the patients normoglycaemia was maintained during the first study day, whereas on the second study day normoglycaemia in the morning was followed by hyperglycaemia in the afternoon. The remaining four patients were examined with the glycaemia pattern in the reverse order. On all study days of all patients the following were measured during both the morning and the afternoon: retinal arteriolar diameter as measured by a retinal vessel analyzer (RVA) before and during isometric exercise induced by lifting a hand weight of 2 kg, mean arterial blood pressure, intraocular pressure, retinal thickness using OCT, estimation of the systolic and diastolic pressure in the ophthalmic artery and ocular pulse amplitude using an ophthalmodynamometer (SmartLens). The measurements were compared using paired t-tests.

Results: Diameter change during isometric exercise, mean arterial blood pressure, intraocular pressure, retinal thickness using OCT, estimation of the systolic and diastolic pressure in the ophthalmic artery and ocular pulse amplitude did not differ significantly between normoglycaemia and hyperglycaemia.

Conclusion: Acute elevation of blood glucose does not change the myogenic response in retinal arterioles. This suggests that glucose does not have a direct effect of on the regulation of retinal arteriolar smooth muscle tone.


C Hernández,E Carrasco,R Casamitjana,R Deulofeu,J García-Arumí and R Simó. Diabetes Unit. Hospital Vall d'Hebron ,Spain

Purpose: There is growing evidence to indicate that somatostatin (SS) could be added to the list of natural antiangiogenic factors that exist in the vitreous fluid. In addition, a deficit of intravitreous SS-like immunoreactivity (SLI) has been found in diabetic patients with proliferative diabetic retinopathy (PDR). In the present study we have determined the main molecular variants of SS (SS-14 and SS-28) in the vitreous fluid of non-diabetic control subjects and diabetic patients with PDR in order to evaluate: a) the predominant molecular variant of SS that exists in the vitreous fluid, and b) the main SS molecular form accounting for the reduction of SLI observed in the vitreous fluid of diabetic patients with PDR. Finally, the contribution of cortistatin (CST), a neuropeptide with strong structural similarities to SS, to SLI and its levels in vitreous and plasma in both non- diabetic and diabetic patients has also been measured.

Methods: Plasma and vitreous fluid from 22 diabetic patients with PDR and 22 non-diabetic control subjects were analyzed.SS-14, SS-28 and cortistatin (CST) were measured by RIA. The antiserum employed to SS-14 determination recognized 100% of SS-14 but also cross-reacted a 31.3% for SS-28 and 30.8% for CST. By contrast, the antiserums against either SS-28 or CST were 100% specific. Consequently,SS-28 and CST were assessed by RIA, but separation by HPLC was required to measure SS-14.

Results: The predominant molecular form of SS within the vitreous fluid was SS-28 (5 fold higher than SS-14 in control subjects and 3 fold higher in PDR patients). CST significantly contributed to SLI and its intravitreous levels (pg/ml) were higher than those detected in plasma (non-diabetic controls: 146 [102-837] vs. 77 [24-32]; p=0.01. PDR patients: 186 [87-998] vs. 62 [24-472]; p=0.01). Intravitreous SS14 was similar in both PDR subjects and the control group (p=0.87). By contrast, SS-28 concentration (pg/ml) was lower in PDR patients than in non-diabetic control subjects (350±32 vs. 595±66; p=0.004).

Conclusions: 1) SS-28 is the main molecular variant in the vitreous fluid. 2) The intravitreous SLI deficit detected in PDR patients is mainly due to SS-28. 3) CST is abundant in the vitreous fluid and significantly contributes to SLI. These findings could open up new strategies for PDR treatment.


AK Sjølie, Dept. Ophthalmology, Odense Univ. Hospital, Odense , Denmark

Purpose: The primary objective of the DIRECT Programme is to examine primary and secondary prevention of diabetic retinopathy when blocking the angiotensin II type 1-receptors with candesartan in normoalbuminuric, normotensive Type 1, and secondary prevention only in normoalbuminuric, normotensive or treated hypertensive Type 2 diabetic patients. The secondary objectives include examining the pooled incidence of proliferative retinopathy in all three populations combined.

Methods: Standardised investigations for patients at enrolment included blood pressure measurement, analysis of HbA1c and serum lipids, and a detailed ophthalmological examination. Retinopathy outcomes are assessed yearly. Retinopathy is graded centrally, based on seven-field stereo photographs using the ETDRS protocol. When the results of retinal gradings were available, randomisation to either candesartan or placebo was performed. A sample size re-assessment was carried out when approximately 70% of the patients had been randomised in the DIRECT Programme.

Results: 1421 and 1905 patients are evaluated in the primary and secondary prevention studies in Type 1 patients, respectively and 1905 patients are evaluated in the secondary prevention study in Type 2 patients. Of the Type 2 patients 37% were on insulin and 63% were treated for hypertension. HbA1c showed mean values of 8.1, 8.5 and 8.2% for the Type 1 primary, Type 1 secondary and Type 2 secondary prevention studies, respectively with no significant regional differences. In the Type 1 secondary prevention study 49% of the patients had level 2 in at least one eye, and 9% level 43-47. In Type 2 patients, 17% had level 43-47 and the remainder less severe retinopathy.Significant correlation (Spearman) with retinopathy levels was found with duration of diabetes (p<0.0001), HbA1c (p<0.0001), and systolic blood pressure (p=0.003) in both Type 1 and Type 2 patients. Diastolic blood pressure was significantly correlated with retinopathy level in Type 1 patients (p=0.001).

Conclusions: RAS blockade has been shown superior to other antihypertensive therapy in slowing progression of renal disease in diabetic patients however questions remain regarding diabetic retinopathy. Adjusting the sample sizes resulting from the updated power calculation was considered necessary for the possibility for the DIRECT Programme to be conclusive.


L Duarte, J Figueira, S Nunes, R Bernardes, J Cunha-Vaz Center of Ophthalmology of Coimbra, IBILI, Faculty of Medicine, University of Coimbra, Portugal

Purpose: We carried out a prospective randomised two-year follow up study to compare the effect of stepwise intensive treatment of diabetes versus standard treatment on the progression of preclinical retinopathy.

Methods: Forty type 2 diabetes patients were allocated in this open, parallel trial to either standard (n=20) or intensive (n=20) treatment. Thirty-five patients completed the two-year period of follow-up. The age at inclusion was 55 ±6 years (mean ± SD) and a mean known duration of their diabetes of 9 ±5 years. Intensive treatment was a stepwise implementation of behaviour modification and pharmacological therapy. The primary endpoints were: number of new red dots, alteration of the Blood-Retinal Barrier (BRB) and capillary closure. Other endpoints were measurements of retinal thickness and retinal capillary blood flow.

Results: There was retinopathy progression of one level of Wisconsin grading in 32 of the 35 eyes. No differences in retinopathy progression were found between the intensive and nonintensive treatment groups, or between the groups that showed HbA1C values =or< than 7,1, or > than 7,1 throughout the 2-year follow-up period. Overall, the alteration of the BRB remained stable whereas there were significant increases in retinal thickness (p<0,001) and retinal capillary blood flow (p=0,042). Capillary closure developed in 5 of the 35 eyes. Two groups of eyes showed 2 well defined patterns of retinopathy progression. Ten eyes showing higher levels of leakage (>10% over mean +2SD) had higher increases of retinal thickness (p=0,03) and higher values of retinal capillary blood flow (p=0,02). The five eyes that developed capillary closure showed low levels of leakage, thickness and capillary blood flow.

Conclusions: Intensified multifactorial intervention in patients with type 2 diabetes and preclinical retinopathy showed no clear influence on retinopathy progression. Retinopathy progression appears to follow different patterns in different patients independently of metabolic control. Two different patterns of progression were identified in this study, one associated with marked alterations of the BRB and the other with the development of capillary closure.


Aniz Girach, Matthew Sheetz, Lloyd P Aiello, Matthew Davis, Roy Milton, Kuolung Hu, Louis Vignati. LILLY RESEARCH CENTRE, WINDLESHAM,SURREY,UK

Purpose:Loss of visual acuity often occurs when diabetic macular edema (DME) involves the center of the macula. Since VEGF induced retinal vascular permeability may be linked to PKCß; activation, inhibition of PKCß; using the selective inhibitor RBX, may have a beneficial effect on DME and associated visual loss in patients with diabetes.

Methods: Additional analysis was performed on a) eyes of 128 patients from the PKC-DRS trial with any severity of DME and diabetic retinopathy (DR) level of 35-53E (ETDRS retinopathy scale) and on b) eyes of 343 patients from the PKC-DMES trial with DME > 300 microns from the center of the macula and DR level 35-53E. Grading of stereoscopic fundus photographs was used to assess DME. Ocular characteristics were compared among treatment groups by categorical tests or ANOVA. All analyses adhered to the intent to treat principle.

Results: At all post-baseline visits where both VA and DME were assessed, eyes from the PKC-DRS trial with DME not at the center of the macula had mean ETDRS visual acuity from 80-84 letters. Eyes with DME at the center of the macula had mean visual acuities of 60 letters (placebo) and 71 letters (RBX) (p<0.001). In the PKC-DMES trial, 32% of eyes of placebo patients with DME >500 mm from the center of the macula at baseline eventually developed DME at the center, whereas 44% of eyes with DME <500 mm at baseline eventually developed center involvement. Center DME development in RBX treated eyes was 21% and 43%, respectively (p=0.001 for distance, 0.039 for therapy from the logistic regression).

Conclusion: Diabetic macular edema at the center of the macula is associated with visual loss. RBX treatment may mitigate both DME progression to the center of the macula and the associated visual loss.

SPONTANEOUS EVOLUTION OF DIABETIC MACULAR EDEMA P Massin , I Zundane, A Erginay, R Benosman, M Laloi-Michelin, C Boutron, S Azancot, E Vicaut , PJ Guillausseau, A Gaudric Ophthalmology Department /Hôpital Lariboisière ,Pari, France

PURPOSE: Macular edema (ME) is the main cause of visual impairment in diabetic patients. Few data on its spontaneous evolution in the short term are available, due to the absence, until recently, of an objective method of assessing macular thickness. Optical Coherence Tomography (OCT) now allows sensitive reproducible measurements of macular thickness. The aim of our study was therefore to investigate the spontaneous evolution of diabetic ME in the short tem using OCT and its possible correlation with changes in glycemia and blood pressure.

METHODS: 23 diabetic patients (4 type 1 and 19 type 2), aged 29 to 75 years, with clinically significant diabetic ME involving the center of the macula, were included. Visual acuity using the ETDRS chart, and macular thickness using OCT were measured every two weeks, for 3 months. On one occasion, they were also measured 4 times on the same day. Glycaemia and blood pressure were measured at each visit, and were also recorded continuously for 24 hours on one day.

RESULTS: Mean central macular thickness was 384±117 (256-668) µm at baseline and the mean ETDRS score was 63±12. There was a significant correlation between macular thickness and visual acuity (p=0.0007). The mean variation in central macular thickness was 88±71 µm (range, 21-356) during the 3-month follow-up period, and 32.±20µm (8-87) when measured 4 times during the same day. The mean coefficients of variation of these two measurements were 9.2 (5.6-12.5) and 4.04 (3-5) respectively. Sixteen patients exhibited, at least once, a 10% variation relative to their median macular thickness during the 3-month follow-up period, and 3 patients did so relative to this thickness during the same day. No correlation was observed between changes in macular thickness and those in glycaemia and blood pressure.

CONCLUSION: This study demonstrates that ME fluctuates spontaneously with time. This spontaneous variability and its quantification are important for the design of clinical protocols for diabetic ME treatment using OCT to determine the efficacy of therapeutic agents.


R Tadayoni, C Loeuillet, J Oliary, A Gaudric, P Massin Ophthalmology and Pharmacy dept, Lariboisiere University Hospital,Paris, France.

Purpose: To establish the dose of triamcinolone actually injected intravitreally according to the injection modality.

Methods: The vial of triamcinolone was shaken for 1 minute and its content was then dropped into a little cup. Next, either 0.1 ml of the suspension was taken 10 seconds, 1 minute and 2 minutes later into a 1ml syringe and injected immediately via a 30 g needle, or 0.5 ml was taken, left in the syringe for 1 minute, adjusted to 0.1 ml and injected into a light cup as it is done during an intravitreous injection. The solvent was evaporated and the dose of triamcinolone measured. The dose was measured 10 times with each of 4 modalities. As controls, 10 cups were weighed twice when empty, and once after injection of 0.1 ml of water to measure the volume actually injected.

Results: In theory, the triamcinolone dose was 4mg in 0.1ml. The volume actually injected was 0.92±0.16 ml. The triamcinolone doses actually injected differed significantly for each of the 4 injection modalities (4.2±0.6, 3.2±0.4, 2.8±0.5 and 6.0±1.6 mg respectively, p=0.0001). For, the mean dose injected, the difference between the 2 operators was largest for the fourth method (respective differences for each modality: 0.5, 0.1, 0.1 and 2.1 mg).

Conclusions: The actual dose of intravitreously injected triamcinolone may differ significantly with small changes in the preparation and injection modalities. Here, a delay of only 1 minute in the aspiration of the suspension reduced the dose by 1 mg. Some techniques also seem more reproducible than others: thus, the fourth method resulted in the highest standard deviation for the dose injected and the largest difference between operators.


Christian A. Frederiksen, Peter Jeppesen, Søren T.Knudsen, Per Løgstrup, Carl Erik Mogensen, Toke Bek Dept. of Ophthalmology, Aarhus University Hospital, Aarhus, Denmark

Purpose: To compare the response in diameter of retinal arterioles and retinal thickness in patients with different stages of diabetic maculopathy during a rise in arterial blood pressure.

Methods: Nineteen normal persons (group A) and fifty-seven type 2 diabetic patients were studied. The patients consisted of three groups matched for diabetes duration and for age and gender with controls: Group (B) no retinopathy. Group (C) 1-4 microaneurysms. Group (D) hard exudates in the macular area not requiring laser treatment. The diameter changes of a large retinal arteriole were measured continuously using the Retinal Vessel Analyzer (RVA, Imedos, Germany) before, during and after a rise in blood pressure induced by isometric exercise when lifting a hand weight in one arm. Blood pressure was measured on the other arm using the cuff technique. After each examination retinal thickness was measured using optical coherence tomography.

Results: Isometric exercise induced an average increase in blood pressure of 21.2±1.2 mmHg with no difference between the studied groups. Retinal arterial autoregulation showed a significant decrease with increasing degree of diabetic maculopathy (p=0.01, Friedmann ANOVA). The diameter changes were -0.70%±0.48 in group A and -1.15%±0.44 in group B indicating preserved autoregulation, but 0.41%±0.49 in group C and 0.54%±0.44 in group D indicating loss of autoregulation. When comparing regions without hard exudates in the maculopathy group, with groups A, B & C, retinal thickness showed mean values of 256±56;m±4.6 (group A), 249&#956;m±2.9 (group B), 251±56;m±4.1 (group C), 264±56;m±4.6 (group D). The retinal thickness in group D with diabetic maculopathy differed significantly from the other groups (p=0.02, Friedmann ANOVA).

Conclusions: Retinal arterial autoregulation is impaired in diabetic patients with any retinopathy, whereas retinal thickness increases with maculopathy. The thickening of the retina appears generalized, and not only in areas with hard exudates.


W3C standards logos

site developed by Netsima Ltd »