22nd Meeting of the

European Association for the Study of Diabetes

Eye Complications Study Group (EASDec)

Dublin, Ireland 25th - 27th May 2012


I. Walraven1,2, K. van den Hurk1,3, E. van 't Riet1,3, C. D. A. Stehouwer4, G. Nijpels1,5, B. C. P. Polak1,2, J. M. Dekker1,3
1EMGO Institute for Health and Care research, VU University Medical Centre, Amsterdam, The Netherlands
2Department of Ophthalmology, VU University Medical Centre, Amsterdam, The Netherlands
3Department of Epidemiology and Biostatistics, VU University Medical Centre, Amsterdam, The Netherlands
4Department of Internal Medicine and Cardiovascular Research Institute Maastricht (CARIM), Maastricht University Medical Centre, Maastricht, The Netherlands
5Department of General Practice, VU University Medical Centre, Amsterdam, The Netherlands

INTRODUCTION. Population-based study.
PURPOSE. The aim of this study was to prospectively investigate the relationship of retinopathy with changes in left ventricular (LV) function.
METHODS. Within the Hoorn Study, a population-based cohort study of diabetes in The Netherlands, retinal photography and echocardiography were performed among 796 participants in the year 2000 (baseline). In 2008 (follow-up), 394 participants attended follow-up examinations, including echocardiography. Linear regression analyses were performed to investigate associations of retinopathy with changes in LV systolic and diastolic function. Retinopathy was graded according to the Eurodiab classification and for the present analyses defined as absence or presence of any retinopathy. Estimates of LV function included LV ejection fraction (% LV systolic function), LV mass index (g/m2.7), left atrial (LA) volume index (ml/m2) and the ratio of LV inflow (E) and early diastolic lengthening (e') velocities (LV diastolic function). All analyses were adjusted for age and gender, with subsequent analyses for baseline values of LV function and traditional risk factors of heart failure (i.e. cholesterol, HbA1c, cigarette smoking, systolic blood pressure, use of glucose-, blood pressure-, and lipid lowering medication and BMI). We tested for interaction of diabetes, hypertension and gender in the relationship between retinopathy and LV function.
RESULTS. Of the 394 participants (53 % male, mean age 69, 65% diabetes), 11.2% had retinopathy, of whom 8.1% had diabetes at baseline. At baseline, LV ejection fraction was not significantly different in participants with retinopathy compared to those without (60.2% versus 60.7%). However, the presence of baseline retinopathy was significantly associated with a 6.0% (95% CI -11.75 - -0.37) lower LV ejection fraction at follow-up. Baseline retinopathy was significantly associated with an 8.3 (95% CI 1.87-14.73) g/m2.7 higher LV mass index, but not with a significant 8-year increase of LV mass index. Adjusting for traditional cardiovascular risk factors did not alter the results, no interactions were observed.
CONCLUSIONS. In the general population in The Netherlands, the presence of retinopathy was prospectively associated with a decrease in LV ejection fraction, independent of diabetes status and traditional risk factors.


E. Berrone, E. Beltramo, M. Porta
Department of Internal Medicine, University of Turin, Turin, Italy

INTRODUCTION. Diabetic retinopathy is characterized by early pericyte drop-out and capillary degeneration. Elevated Angiotensin II (Ang II) levels and/or increased sensitivity to Ang II have been etiologically associated with major vascular diseases. The renin-angiotensin system plays a role in the modulation of vascular insulin sensitivity and endothelial function, interacting with the insulin signal and counteracting its vasodilatory effect, thus promoting vasoconstriction. The Ang II action on insulin signalling results in alteration of insulin-mediated glucose transport. We have previously demonstrated that an antagonist of the Ang II AT1 receptor, candesartan, has significant anti-apoptotic and anti-senescence effects on human retinal pericytes (HRP) grown in diabetic-like conditions.
PURPOSE. The objective of this study was to verify if candesartan influences the mechanisms regulating intracellular entry of glucose in HRP.
METHODS. Pericytes were kept at 48h alternate intervals in high (28 mM=HG) or normal (5.6 mM=NG) glucose for 8 days (intermittent HG, HGint), with or without 1 or 0.2 µM candesartan. Control cells were cultured in stable NG or HG. Glucose transporters GLUT 1, 3, 4 and insulin receptor mRNA expression were determined by RT-PCR. Intracellular glucose was measured by a commercial fluorescence ELISA kit.
RESULTS. HGint increased GLUT1 and GLUT4 (p<0.05 vs NG) and candesartan (0.2 µM) reduced the expression of GLUT1 only (p<0.05 vs HGint). Intracellular glucose levels were also increased in HGint (p<0.05 vs NG), and normalized by 1 and 0.2 µM candesartan (p<0.05 vs HGint). Insulin receptor expression was reduced in HGint (p<0.05 vs NG) and normalized by 1 µM candesartan (p<0.05 vs HGint). GLUT3 mRNA was unchanged in all experimental conditions.
CONCLUSIONS. Blockade of the Ang-II AT1 receptor by candesartan may influence intracellular glucose concentrations by modulating the expression of glucose transporter GLUT1 and insulin receptor mRNA.


E. Beltramo1, E. Berrone1, M. Lorenzi2, M. Porta1
1Department of Internal Medicine, University of Turin, Torino, Italy
2Schepens Eye Research Institute, Harvard Medical School, Boston, USA

INTRODUCTION. Human retinal pericytes (HRP) are more vulnerable to intermittent than stable high glucose exposure, differently from bovine pericytes. We have already demonstrated increased DNA fragmentation in HRP exposed to glucose fluctuations. Both caspase-dependent and independent mechanisms of apoptosis induction were shown in retinal endothelial and neural cells, differences being accounted for by species-specificity.
PURPOSE. Our aim was to discriminate among the different mechanisms by which diabetic-like conditions may trigger human pericyte death: necrosis and caspase-dependent or independent apoptosis.
METHODS. Confluent HRP were kept intermittently at 48 hrs intervals in high (28 mM)/normal (5.6 mM) glucose for 8 days (intHG), with/without the addition of equimolar concentrations of mannitol to physiological glucose to control for osmotic effects. Control cells were cultured in stable physiological (NG) or high glucose (HG). Positive control for apoptosis was obtained by stimulating cells with staurosporine. Viability, cytotoxicity and apoptosis (caspase 3/7 activity) were measured in the same wells, by means of a fluorescent/chemioluminescent assay (ApoTox-Glo™, Promega). Cytotoxicity and apoptosis data were normalized by viability (rate of death / rate of apoptosis).
RESULTS. Intermittent exposure to HG increased cytoxicity (p<0.05 vs NG), but not caspase 3/7 activity in HRP. Surprisingly, the rate of death was lower in stable HG than in NG (p<0.005). The addition of mannitol resulted in a rate of death comparable to the one obtained with HG and lower than NG and intHG (p<0.005). Stable HG exposure increased caspase 3/7 activity (p<0.05 vs NG).
CONCLUSIONS. Our conclusions, based upon these and previous results, are that intermittent high glucose exposure triggers both cytotoxic and caspase-independent apoptotic mechanisms in human retinal pericytes, while stable high glucose exposure may induce a slight increase in caspase activity, without further effects on apoptosis. Since mannitol protects these cells from cytoxicity but not from apoptosis, it is likely that the former is mediated by abrupt osmotic changes in the medium and the latter by metabolic mechanisms.


R. Simó1, A.M. Valverde2, S. Miranda2, M. Garcia-Ramírez1, A. Gonzalez-Rodríguez2, C. Hernández1
1Diabetes and Metabolism Research Unit. Vall d'Hebron Research Institute.Universitat Autònoma de Barcelona. Spain.
2Instituto de Investigaciones Biomédicas Alberto Sols CSIC/UAM: Madrid. Spain

INTRODUCTION. Case-control study
PURPOSE. Diabetic retinopathy (DR) has been classically considered to be a microcirculatory disease of the retina. However, before any microcirculatory abnormalities can be detected in ophthalmoscopic examination, retinal neurodegeneration is already present. The aim of the study was to analyze pro-apoptotic and survival signalling in the neuroretina of diabetic patients at early stages of DR.
METHODS. The retinas from 5 diabetic donors at early stages of DR were compared with the retinas from 5 non-diabetic donors closely matched by age. Glial activation was evaluated by assessment of glial fibrillar acidic protein (GFAP) by Western blot and immunofluorescence. Pro-apoptotic molecules (FasL, pro-caspase-8, active caspase-8, total Bid, truncated Bid, Bim, and active caspase -3), as well as anti-apoptotic markers (FLIP, BclxL and COX-2) were assessed by Western blot.
RESULTS. GFAP and pro-apoptotic molecules (FasL, active caspase-8, t-Bid, Bim, and active caspase-3) were significantly increased in neuroretina from diabetic patients as compared to the control neuroretinas. By contrast, no significant differences in the expression of the anti-apoptotic markers were found.
CONCLUSIONS. An imbalance between pro-apoptotic and survival signalling was found in diabetic neuroretinas. Our results reveal some key mechanistic pathways involved in the neurodegenerative process that occurs in the early stages of DR.

A. Hamid1, H.M. Wharton1, A. Mills1, J.M. Gibson1,2, M. Clarke1, P.M. Dodson1,2
1Departments of Diabetes and Ophthalmology, Heartlands Hospital, Birmingham, UK
2School of Health and Life Sciences, Aston University, Birmingham, UK

INTRODUCTION. Retrospective Study
PURPOSE. The English National Screening Programme for diabetic retinopathy (ENSPDR) states that “All people with diabetes aged 12 years and older should be offered screening for sight-threatening DR using digital photography for quality assurance purposes”.
The audit aims to assess whether the current guideline is suitable and whether diabetes duration should be taken into account when deciding at what age to start screening patients.
METHODS. Retrospective analysis of 143 randomly selected patients aged twelve years or younger who have attended diabetic retinopathy (DR) screening in the Birmingham and Black Country Screening Programme
RESULTS. 98% had Type 1 diabetes and mean visual acuity (VA) was 6/5 (6/4-6/36). 73 were under 12 with 7 the youngest age and 70 were aged 12. Both groups had mean diabetes duration of 5 years (1month-11years). For those under 12, 7/73 (9.6%) had background DR, of these mean diabetes duration was 7 years (6-8) and the youngest aged 8. In those aged 12, 5/70 (7.1%) had background DR; of these mean diabetes duration was 8 years (6-11).
In total 12 (8.4%) patients aged 12 years or under developed DR. No patients had retinopathy worse than background changes. One patient was referred to ophthalmology for VAs of 6/12 right eye and 6/18 left eye and was diagnosed with optic atrophy so returned to annual screening for DR.
CONCLUSIONS. The results show that the current guideline on when to begin screening could be amended as patients under the age of 12 were showing signs of DR. Diabetes duration may help when deciding what age to start screening younger patients as DR was not seen in those with disease duration of less than six years. We suggest screening should begin after six years of diagnosis (or age 12, whichever is earlier) for prepubertal onset type 1 diabetes.

J.J. Smith1,2
1Foresight Eyecare, Dublin, Ireland

2HSE Dublin/North East Dublin, Dublin, Ireland

INTRODUCTION. The results from imaging and quality assured grading of digital retinal images of individuals with type 2 diabetes at two points in time (4 years apart) were analysed with the aim of establishing the risk of disease progression to the point that hospital referral would be appropriate because of the risk of sight threatening diabetic retinopathy.
PURPOSE. Ireland is about to establish a national retinopathy screening program for people with diabetes. Those considering the quality of care standards need to consider the marginal benefit of frequent eye examinations when setting their quality and performance standards.
METHODS. 537 people with diabetes enrolled in the Diabetes Watch program (HSE Dublin/ North East) had 2 field photography for each eye in 2006 and had microaneurysm (MA) count undertaken using an automated image reading system. The result was an aggregate count for the two eyes. 287 of the double graded cases seen in 2006 were still attending screening in 2010 and their 2006 MA count was compared to their manual grading result in 2010. By 2010 the manual grading was a 3 step method by trained accredited graders with the original 2006 ophthalmologist acting as the third line arbitration grader. Automated MA counting was used as a first line grading tool.
RESULTS. Of a total of 208 image sets from 2006 which had an automated MA count of zero and which the manual grade was no disease 157 were screened in 2010. Of these 137/157 (87%) had no disease as their final grade in 2010. A further 16/157 (10%) had minimal background DR as their final manual grade. Only one case out of 157 identified as having no disease as the manual grade in 2006 and having a MA count as zero in 2006 had a reason for referral to the hospital eye service in 2010.This represents a 4 year risk of needing hospital referral of less than 1%.
CONCLUSIONS. The necessity for annual screening is questioned in the case of quality assured manual grading which indicates no evidence of diabetic retinopathy.

M. Porta1,2, M. Trento1, E. Sitia1, S. Albani1, E. Lamarmora1, P. Dalmasso2, F. Cavallo2
1Diabetic Retinopathy Centre, Department of Internal Medicine, University of Turin, Italy
2Department of Public Health and Microbiology, University of Turin, Italy

INTRODUCTION. Prospective analysis of results of screening in the Diabetic Retinopathy Centre.
PURPOSE. To measure the interval from negative screening to the onset of referable retinopathy, as defined by the 1990 London protocols, in patients with different clinical characteristics.
METHODS. Data of all patients with type 2 diabetes found to have no retinopathy at first screening, and for whom we had at least 2 screening available at any time, were analysed. The patients were stratified with a 2 x 2 approach into those not receiving (NIT) or receiving (IT) insulin treatment and having less or more than 10 years known duration of diabetes. In total, 2247 NIT<10yrs, 687 NIT>10yrs, 426 IT<10yrs, and 263 IT>10yrs without retinopathy at their initial screening were included.
RESULTS. The average times for 5% of the patients in the different subgroups to develop referable retinopathy were 63 (95%CI 56-72), 48 (95%CI 34-57), 49 (95%CI 33-61), and 39 (95%CI 23-49) months, respectively.
CONCLUSIONS. Diabetic retinopathy may progress to referable severity at different rates in patients with different treatment regimens and diabetes durations. Even in the absence of data on metabolic or blood pressure control, these clinical characteristics may provide useful guidance when planning re-screening appointments.

H.D.J. Jonsdottir1, Ó.P. Pálsson1, P.M. Matthíasson2, E.S. Stefánsson1,3
1Risk ehf. Reykjavík, Iceland
2Department of Economics, University of Iceland. Reykjavík, Iceland
3Department of Medicine, University of Iceland. Reykjavík, Iceland

INTRODUCTION. An economic evaluation
PURPOSE. Regular eye screening helps prevent diabetic blindness. Annual eye screening of patients with diabetes mellitus is recommended by the World Health Organization. Individualized screening based on individual risk analysis has been proposed to be more economical than current standards of screening. We compare the economics of annual, biennial (every other year) and individualized screening.
METHODS. We calculate government expenditures and total societal costs related to diabetes eye screening in Iceland. The cost resulting from three different screening models is compared over one year. Main areas of uncertainty were addressed by conducting partial and Monte Carlo sensitivity analysis.
RESULTS. Governmental cost per screening visit is €23 and social cost per screening visit is €135. Total cost accruing to publicly run heath institutions is €170,730 in annual screening, €120,675 in biennial screening of low risk patients and €69,999 in individualized screening. Total societal cost of screening is €1,005,662 in annual screening, €710,816 in biennial screening and €412,321 in individualized screening. The results of the Monte Carlo simulation were similar to model calculations adding credibility to results.
CONCLUSIONS. Individualized screening is less expensive to government and society when compared with biennial and fixed annual screening. Total societal costs are about 6 times greater than direct governmental cost. Individualized risk assessment and screening reduces all costs by more than half compared with the current standard of annual screening.


P. Massin1, T. Peto2, K. Le-Malicot3, J. Ansquer3
1Hopital Lariboisiere Paris, France
2 NIHR Biomedical Research Centre for Ophthalmology, at Moorfields Eye Hospital NHS Foundation Trust and UCL Institute of Ophthalmology

3Laboratoires Fournier SA now Abbott Daix France

INTRODUCTION. A double-masked, placebo-controlled, randomised study
PURPOSE. Fenofibrate has been shown to reduce the progression of diabetic retinopathy (DR) over 4 to 5 years and the need for photocoagulation in patients with type 2 Diabetes (T2DM).
METHODS. One hundred and ten patients with T2DM and diabetic macular edema (DME) were randomised in 20 European centres to receive fenofibric acid (FA), the active moiety of fenofibrate, as one 135 mg capsule daily or placebo (P) for one year. Evaluations included quarterly ophthalmology visits and Stratus OCT3 measurements with 7-field fundus pictures and ETDRS best corrected visual acuity at baseline and one year. Presence of DME was defined as retinal thickness >=250µm central zone (CZT) or >=300µm in inner zone(s). The primary endpoint was the change from baseline to endpoint in total macular volume (TMV) in 9 OCT subfields. Central reading of OCTs served for randomisation and ETDRS grading.
Statistical within- and between-group comparisons used intent to treat analysis of variance for subject worse eye at baseline and generalised estimated equation (GEE) for all eligible eyes.
RESULTS. Seventy-one men and 39 women aged 62 with BP 138/80mm Hg and HbA1c 7.8% were randomised to FA (57 subjects, 83 eyes) or P (53 subjects,77 eyes).All but 3 subjects in each group completed one-year treatment with unchanged BP and HbA1c and satisfactory tolerability. Only 18 patients require eye procedures for DME. Baseline mean TMV and CZT were 8.54mm3 and 350µm.FA reduced TMV at endpoint by 0.35 mm3 in worse eye and all eyes (p=0.019 and <0.001), a non-significant difference with placebo (TMV reduction by 0.11mm3 and 0.12 in worse and all eyes). Similarly, FA reduced CZT by 19µm in all eyes (p=0.018), and placebo by 12µm,a non significant between-group difference. Of interest, these changes were obtained after only 3 months. ETDRS grading and visual acuity did not significantly change.
CONCLUSIONS. In this randomised OCT study in T2DM subjects with DME, one-year fenofibric acid treatment provided a modest improvement in macular edema and stabilization of DR. Further studies are needed to explain how fenofibrate reduces progression of DR.


C. Bailey
Bristol Eye Hospital, Bristol, United Kingdom

INTRODUCTION. 24 month (M), open-label, multicenter, Phase-IIIb, extension study.
PURPOSE. Evaluate the safety and efficacy of ranibizumab 0.5 mg (RBZ) in patients (pts) with visual impairment due to diabetic macular edema (DME) in RESTORE extension study.
METHODS. 240 of 303 pts who completed RESTORE core study (Day1-M12) entered the extension study (M12-M36); all pts (RBZ/RBZ+laser/laser) were eligible to receive RBZ PRN according to pre-specified stability-based best-corrected visual acuity (BCVA) re-treatment criteria, and laser according to Early Treatment Diabetic Retinopathy Study guidelines. Here we report M24 interim analysis outcome of the incidence of adverse events (AEs), treatment exposure, changes in BCVA and visual function (VF, assessed by Visual Functioning Questionnaire [VFQ-25]). Last observation carried forward approach was used for BCVA and VF analysis.
RESULTS. 220 pts had M24 visit. The 2-year safety analysis revealed neither new AEs nor new safety risks. No cases of endophthalmitis were reported. There were 4 deaths over Day1-M24, all occurring between M12-M24 (unrelated to study drug/procedure). The mean BCVA gain of +7.9 (RBZ) and +7.1 (RBZ+laser) letters in core phase was maintained with an average of 3.9 (RBZ) and 3.5 (RBZ+laser) RBZ injections in 2nd year, with a mean BCVA change of +7.9 (RBZ) and +6.7 (RBZ+laser) letters from Day1-M24. Patients treated with laser in core phase received an average of 4.1 RBZ injections (M12-M23) with a mean BCVA improvement of +2.3 (Day1-M12) and +5.4 (Day1-M24) letters. Mean change in VF (Day1-M24) for RBZ/RBZ+laser/laser (eligible to receive RBZ as of M12) was +10.8/+10.5/+4.6 points (near activity), +5.1/+7.0/+2.8 points (distance activity) and +5.6/+5.8/+4.3 points (composite score).
CONCLUSIONS. The AEs reported in interim analysis were consistent with the published safety profile of RBZ in DME. In the 2nd year, an average of 3.8 RBZ injections over the three core treatment groups was sufficient to maintain (RBZ/RBZ+laser) or improve (laser) BCVA and VF in the RESTORE extension study. These data provide further evidence for long-term safety and efficacy of RBZ in DME.


A. Cole, C. Bailey
Bristol Eye Hospital, Bristol, United Kingdom

INTRODUCTION. FAME (Fluocinolone Acetonide in Diabetic Macular OEdema) consisted of 2 randomized, prospective, multicenter, double-masked, sham-controlled, parallel-group phase 3 trials.
PURPOSE. To determine the efficacy and safety of FAc intravitreal implants in patients with DMO and to determine a subgroup of patients experiencing the most favourable benefit/risk ratio.
METHODS. 956 patients who had at least 1 prior macular laser treatment were randomized 2:2:1 to 0.2 μg/d FAc (n=376), 0.5 μg/d FAc (n=395), or sham control (n=185). All patients were eligible for rescue laser therapy at week 6 at investigator discretion and retreatment with randomized study treatment at month 12 if retreatment criteria were met. Prespecified subgroups, including baseline DMO duration (< or ≥ median), were examined.
RESULTS. The FAME study met its primary endpoint—28.7% of 0.2 μg/d FAc-treated patients experienced a ≥ 15-letter improvement in best-corrected visual acuity (BCVA) at month 24, compared with 16.2% of control patients (P=.002). The 0.2 μg/d FAc implants were also superior to control at month 36, 28.7% vs 18.9% (P=.018). The mean excess centre point retinal thickness decreased in all groups. Most patients required only 1 FAc implant over 36 months. Among chronic DMO patients (DMO ≥ 3 years at baseline), 34.0% in the 0.2 μg/d FAc group experienced a ≥ 15-letter improvement in BCVA compared with 13.4% of controls (P<.001) at month 36. This effect was reproduced in the 2 individual FAME trials. 80.0% of phakic patients in the 0.2 µg/d FAc group and 27.3% of phakic controls had cataract surgery, and 4.8% of patients in the 0.2 µg/d FAc group required intraocular pressure (IOP)-lowering surgery vs 0.5% of controls by month 36.
CONCLUSIONS. 0.2 μg/d FAc provided rapid and sustained improvements in BCVA and retinal thickness in patients with DMO for up to 36 months. The greatest benefit was seen in patients with chronic DMO. The development of cataract was very common in the FAc-treated patients but less than 5% required incisional IOP-lowering procedures.


D.C. Jenkin1, S. Corcoran2,4, M. Ashworth3, G. Stewart4
1King's College London, United Kingdom
2NHS Diabetic Eye Screening Programme, North Middlesex University Hospital NHS Trust, United Kingdom
3Department of Primary Care & Public Health Sciences, King's College London, United Kingdom
4NHS North Central London, United Kingdom

INTRODUCTION. Cross–sectional, observational.
PURPOSE. To evaluate whether consistently high or low achievement by primary care practices on the diabetes clinical domain of the quality and outcomes framework is predictive of risk for screen detected retinopathy.
METHODS. Primary care achievement was assessed using data from the English national Quality and Outcomes Framework (QOF) between 2006 and 2010. Practices were categorised as high or low achieving if they ranked in the top or bottom quartile in all 4 years for percentage of patients with diabetes in whom the last HbA1c is 7% or less in the previous 15 months. QOF data was linked via the patient's registered GP to patient level diabetic retinal screening records. The sample was limited to those whose retinal screen showed no retinopathy in either eye at screening in 2008/9. A random effects logistic regression model was generated to examine the odds of any retinopathy at screening in 2010/11, for high and low achieving practices as compared to all other practices. Odds ratios were adjusted for age, gender, ethnicity, and area level deprivation based on patient's residential address.
RESULTS. From a total of 176 practices included, screening records for 14,296 were retrieved, of which 63.8% had no retinopathy at baseline (n = 9,121). For the HbA1c target 21 practices were high achieving and 18 low achieving. Patients in the low group have 44% higher adjusted odds of any retinopathy at screening (O.R. 1.44, 95% C.I. 1.17 to 1.77, p = 0.001), as compared with all other practices in the sample. Patients in the high group have 24% lower relative adjusted odds of any retinopathy at screening (O.R. 0.76, 95% C.I. 0.63 to 0.93, p = 0.008), as compared with all other practices in the sample.
CONCLUSIONS. Poor achievement of the primary care HbA1c target for diabetes patients may be associated with significant increase in risk of screen detected retinopathy for patients with no retinopathy at screening 2 years previously.


M.L. Rasmussen, R.B. Pedersen, U. Frydkjær-Olsen, J. Grauslund, A.K. Sjølie
Department of Ophthalmology, Odense University Hospital, Denmark

INTRODUCTION. Case-control study
PURPOSE. To identify risk factors for blindness caused by diabetic retinopathy in young patients with type 1 diabetes.
METHODS. A case-control study of 6 patients (12 eyes) who went blind and 18 age- and sex-matched controls (36 eyes).
All patients were identified in a population-based nationwide cohort study of 324 patients with early onset type 1 diabetes, who participated in a clinical baseline examination in 1995 and were followed for 15 years.
HbA1c, BMI, albumin excretion rate (AER), systolic and diastolic blood pressure were measured at baseline. Retinopathy was graded according to the EURODIAB protocol and microaneurysms (MA) were counted.
Incident blindness was registered between 1995 and 2010 in the Danish Association of the Blind -a voluntary organization for patients with a visual acuity at or below 6/60 in the best eye.
Differences were analyzed using Wilcoxon rank-sum test.
RESULTS. At baseline, median age was 21 years (range 14-23) and median duration of diabetes was 13 years (range 9-15 years) and 12 years (range 9-16), in blind and non-blind patients, respectively.
Median age at which patients became blind was 27 years (range 25-38 years) and duration of diabetes was 20 years (range 19-24 years). Baseline level of retinopathy of patients who became blind was: no retinopathy: 3 eyes, mild or moderate non-proliferative retinopathy (NPDR): 5 eyes, severe NPDR: 2 eyes and proliferative retinopathy: 2 eyes. Among controls baseline retinopathy was: no retinopathy (20/36 eyes), mild or moderate NPDR (16/36 eyes).
Patients who became blind had higher baseline MA-count (median 3.4, range 0-10) than controls (median 0.9, range 0-9), p=0.013.
Patients who went blind had higher levels of HbA1c (median 10.6% vs. 9.2%, p=0.0004), higher diastolic blood pressure (median 83mmHg vs. 75mmHg, p=0.0019), AER (median 10.2 microg/min vs. 4.3 microg/min, p=0.018) and BMI (median 26.0kg/m2 vs. 23.0kg/m2, p=0.025).
Systolic blood pressure did not differ between the groups (median 128mmHg vs. 130mmHg for the blind vs. controls, respectively).
CONCLUSIONS. Multiple systemic risk factors were identified for patients who later became blind. Even early retinopathy such as the number of microaneurysms are important indicators for blindness.


S.J . Aldington1, 2, I.M. Stratton1, 2, P.H. Scanlon1, 2
1Gloucestershire Hospitals NHS Foundation Trust, Cheltenham, United Kingdom
2English NHS Diabetic Eye Screening Programme, Gloucester, United Kingdom

INTRODUCTION. Secure web-browser accessed centralised monthly quality assurance and training support facility.
PURPOSE. Provide an on-line facility from 2009 for 86 diabetic retinopathy (DR) screening services throughout England to locally assess and monitor grader performance and to support graders in their skills development.
METHODS. Tests were available at all times through common web browsers to over 1500 graders in the English national programme. Disease-weighted sets of DR test cases, each two-field 45o retinal images, were presented in years 1 and 2 as 30 and 20 cases per month respectively. Retinopathy (R) and Maculopathy (M) levels were graded according to national protocols. Colour, red-free and green image versions, magnification and contrast/brightness tools were available. Results expressed to Users as percentage exact agreement for combined R and M with the ‘system grade'. Local service managers review more detailed results for members of their team. All printable results were anonymised. Completed Test cases roll-over into expanding Training set, accessible by Users at any time.
RESULTS. During year 1, 1301 people completed at least 1 monthly set. Following a 4-month phased implementation period, data from months 5-12 were analysed. Mean (sd) percentage scores for the 8 months were 64.3% (10.7), 61.9% (11.5), 62.5% (11.6), 68.5% (10.6), 69.7% (13.6), 66.4% (12.7), 70.5% (10.4) and 73.4% (12.9). All 8 sets were completed by 260 people with mean (sd) monthly scores ranging from 63.7% (11.0) to 73.9% (13.0). During Year 2, 1292 graders completed at least 1 set. Mean (sd) results for months 1-8 were 72.6% (9.9), 67.8% (11.6), 63.3% (10.5), 62.3% (11.2), 59.4% (11.9), 74.6% (10.7), 75.0% (9.8) and 70.6% (10.0). All 8 sets were completed by 335 people with scores ranging from 62.2% (10.8) to 76.5% (9.3). There were significant differences between the mean monthly test results (p<0.001) and between individual graders (p<0.001). To date over 425000 gradings have been completed.
CONCLUSIONS. The facility is heavily used. Programmes report the tests can identify consistently high performing staff and those needing additional training. Some test cases and months are more difficult than others.


A. Farnsworth, A. Wright, P. Dodson
Solihull and Black Country Retinal Screening Programme, Birmingham Heartlands Hospital, Birmingham, United Kingdom

INTRODUCTION. Analysis of gradable photographs where Hollenhorst was used in the grading report
PURPOSE. Our aim was to clarify the use of the term Hollenhorst plaque (HP)and to assess its application to both cholesterol emboli at bifurcations and to sheathing of segments of the retinal artery away from bifurcations.
METHODS. Review of gradable photographs of 69 patients within the Birmingham and Black Country DR screening programme 2004-2011 where HP was used in the grading report
RESULTS. Discrete, single or multiple, cholesterol emboli, were noted in 29 men (mean age 72.8yr sd7.9) and 12 women (mean age 76.0yrsd 7.3) with lesions at bifurcations in the majority. An area of sheathing over a segment of an artery was noted in 12 men (mean age 72.3 sd 10.4) and 13 women (mean age 72.1 sd 10.6) with lesions mostly away from bifurcations. In four men (mean age 72.5) lesions of both categories were noted.
In 58% patients cholesterol emboli appeared or disappeared on annual photography whereas in all patients arterial sheathing was present on initial screen and persisted unchanged.
CONCLUSIONS. HPs occur in elderly patients. The term was used for two distinct lesions, the more common being cholesterol emboli. We did not observe transition from an embolus to a plaque.


M.G. Erke1,2, G. Bertelsen1,2, T. Peto3, A.K. Sjølie4, H. Lindekleiv2, I. Njølstad2
1Department of Ophthalmology and Neurosurgery, University Hospital of North Norway, Tromsø, Norway
2Department of Community Medicine, University of Tromsø, Tromsø, Norway
3NIHR Biomedical Research Centre for Ophthalmology, at Moorfields Eye Hospital NHS Foundation Trust and UCL Institute of Ophthalmology, London, United Kingdom
4Department of Ophthalmology, Odense University Hospital, Odense, Denmark

INTRODUCTION. Population-based, cross-sectional study
PURPOSE. To examine the association between diabetes mellitus and age-related macular degeneration (AMD) in The Tromsø Eye Study.
METHODS. The Tromsø Eye Study is a sub-study of the multi-purposed Tromsø Study conducted in 2007-8 in the municipality of Tromsø, Norway. We included 2,605 Caucasians aged 65 to 87 years with gradable digital retinal photographs and available data on diabetes. Photographs were graded for AMD based on the International classification system. Predominant phenotype was determined by the more severe lesion present in either eye. Categories by increasing severity were; normal or hard drusen only, intermediate drusen (63-125μm), large drusen (>125 μm), geographic atrophy and neovascular AMD. Diabetes mellitus was defined as non-fasting blood glucose ≥11.1 mmol/L, HbA1c >6.5 % or use of anti-diabetic medication. We used logistic regression models to calculate odds ratios with 95% confidence intervals (CI) for AMD according to diabetes status, adjusted for potential confounders (age, sex, smoking status, alcohol consumption, education, body mass index, waist-to-hip ratio, blood pressure, physical activity, blood lipids).
RESULTS. Diabetic subjects were more likely to abstain from alcohol, have an inactive lifestyle, to have higher body mass index, waist-to-hip ratio, and serum blood lipids than non-diabetic subjects (all p's ≤ 0.001). Geographic atrophy or neovascular AMD was present in 12 (4.8%) of 249 diabetic subjects and 80 (3.4%) of 2,356 non-diabetic subjects. No statistically significant relationships were found between diabetes and AMD, neither in univariate nor multivariate analyses. Age- and sex-adjusted odds ratios were for intermediate drusen 0.92 (95% CI 0.68-1.25), for large drusen 0.80 (95% CI 0.56-1.14), for geographic atrophy 1.92 (95% CI 0.70-5.28), and for neovascular AMD 0.93 (95% CI 0.41- 2.13).
CONCLUSIONS. In The Tromsø Eye Study, we did not observe any statistically significant association between diabetes mellitus and AMD. This is consistent with results from other cross-sectional population-based studies.


T. Kebede1, H. Wharton1, S. Barbaccia1, B. Dass1, S. Osei-Amoako1, J.M. Gibson1,2,3, P.M. Dodson1,2,3
1Department of Diabetes and Endocrinology, Heartlands Hospital, Birmingham, United Kingdom
2Department of Ophthalmology, Heartlands Hospital, Birmingham, United Kingdom
3School of Life and Health Sciences, Aston University, United Kingdom

INTRODUCTION. Retrospective analysis
PURPOSE. To assess the prevalence of centre involving diabetic macular oedema (CIDMO) and risk factors.
METHODS. Retrospective review of patients who were screen positive for maculopathy (M1) during 2010 in East and North Birmingham. CIDMO was diagnosed by qualitative identification of definite foveal oedema on optical coherence tomography (OCT).
RESULTS. Out of a total of 15234 patients screened, 1194 (7.8%) were screen positive for M1 (64% bilateral). 137 (11.5% of M1s) were diagnosed with macular oedema after clinical assessment. OCT results were available for 123/137; 69 (56.1%) of these had CIDMO (30 bilateral) which is 0.5% of total screens and 5.8% of those screen positive for M1.
In those with CIDMO 60.9% were male and 63.8% Caucasian; 90% had Type 2 diabetes and mean diabetes duration was 20 years (SD 9.7 range 2-48). Mean HbA1c was 8.34%±1.69, with 25% having an HbA1c ≥9%. Furthermore, 62% were on insulin, 67% were on anti-hypertensive therapy, and 64% were on a cholesterol-lowering drug. 37.7% had an eGFR between 30-60% and 5.8% had eGFR <30. The only significant difference between the CIDMO and non-CIDMO group was mean age (67.83±12.26 vs. 59.69±15.82; p=0.002).
65.2% of those with CIDMO also had proliferative or pre-proliferative retinopathy in the worst eye and 68.1% had subsequently been treated with macular laser at the time of data review.
CONCLUSIONS. The results show that the prevalence of CIDMO in our diabetic population 0.5%. A significant proportion of macula oedema patients were found to have Type 2 diabetes with long disease duration, sub optimal glycaemic and hypertensive control and low eGFR. The data supports that medical and diabetic review of CIDMO patients is warranted particularly in the substantial number with poor glycaemic control and if intravitreal therapies are indicated.


W3C standards logos

site developed by Netsima Ltd »