22nd Meeting of the

European Association for the Study of Diabetes

Eye Complications Study Group (EASDec)

Dublin, Ireland 25th - 27th May 2012


A. Diandomba2, L. Assumpta1, A. Eballe1, E. Emilienne1, E. Sobngwi1, P. Bensaid2
1Faculty of Medicine and Biomedical Sciences University of Yaounde, Faculty of Medicine and Pharmaceutical Sciences, University of Douala, Project Management and Prevention of diabetic retinopathy, Yaounde, and Military Hospital No. 1 Region in Yaounde, Cameroon.
2Eye Centre Ophtalmology Sans Frontiere, Garoua Lagdo, Cameroon

INTRODUCTION. Retrospective analysis of prevention and treatment of diabetic retinopathy (DR) in the Yaounde Hospital, Cameroon, between November 2007 to December 2008.
PURPOSE. To analyze the characteristics of diabetic retinopathy in Black Africans of Cameroon.
METHODS. All patients with diabetes attending during the study interval received and ophthalmic examination including fluorescein angiography. A standard questionnaire was used to gather information on age, sex, duration and type of diabetes, medical history, visual acuity and finally results of angiography analysed by a trained ophthalmologist.
RESULTS. Of the 472 patients with diabetes, DR was absent in 274 (58%) and present in 198 (42%). resulting in total of 393 eyes. Three eyes were enucleated prior to the study. Of those with DR, 113 were men (57%). The age range of those affected was 39-87 years with a wide range in duration of diabetes (1-34 years) and 95.5% of them were not on insulin therapy. All eyes with clinical diagnosis of DR had lesions in angiography.
Of patients with DR, 64.6% had bilateral non-proliferative DR, 26.8% had bilateral proliferative DR, while 7.1% nonproliferative DR in one eye and proliferative in the other. The three patients with one eye already enuclated all had proliferative DR in the only eye. Half of those with proliferative DR had vitreous hemorrhage and/or retinal detachment. Nearly all of these patients (94.3%) had Type 2 diabetes with greater than 15 years duration. Based on angiography results, 10.6% of the patients had macular edema.
CONCLUSIONS. The growth of diabetic retinopathy is as important as that of diabetes in Africa. The black African population must be better informed and more aware of importance of prevention of diabetes and its complications. Any patients with diabetes must have access to complete eye examination by trained health to prevent visual loss and introduce timely treatment. In every African country a national program to fight against diabetic blindness must be established, using appropriate equipment and trained personnel.


A. Hirose1, S. Kitano1, Y. Uchigata2
1Department of Diabetic Ophthalmology, Diabetes Centre, Tokyo Women's Medical University, Japan
2Department of Medicine, Diabetes Centre, Tokyo Women's Medical University, Japan

INTRODUCTION. Retrospective study.
PURPOSE. To evaluate a predictive index for retinopathy development using HbA1c values in a long-term follow-up study of younger onset type 1 diabetes.
METHODS. To study the influence of hyperglycemia on retinopathy development as accurately as possible, we selected younger onset diabetics (to minimize the influence of hypertension and dyslipidemia) with continuous HbA1c data substantially just after the onset of diabetes (to minimize the influence of metabolic memory). Major inclusion criteria were: (1) type 1 diabetes diagnosed before the age of 30 years; (2) first visit to the Diabetes Centre between 1988 and 1990 within 12 months after onset; (3) follow-up during 20 years after onset with HbA1c values measured in the Centre. Each patient underwent a principally annual fundus examination through dilated pupils by Centre ophthalmologists. Retinopathy for a given year was defined as positive by a level of 15/<15 or more severe on the final Retinopathy Severity scale of the Early Treatment Diabetic Retinopathy Study. The first year of retinopathy development was established when two consecutive retinopathy-positive years were detected. Patients without a 20th year fundus examination were excluded. HbA1c (Japan Diabetes Society value≈ National Glycohemoglobin Standardization Program value-0.4 (%)) was measured by high performance liquid chromatography. A current year's excess HbA1c value was calculated by the year's mean HbA1c minus a hypothetical normal HbA1c threshold 6.0(%). Then the Xth year's predictive index (PI) was calculated by cumulating the yearly excess HbA1c values from the 1st to Xth year.
RESULTS. Of the 15 patients (6 males and 9 females), 5 developed retinopathy with PI values of 36.3~43.1 during the 20-year follow-up, while 10 did not, with maximum PI values of 3.4~32.5.
CONCLUSIONS. This index could be valuable in predicting the development of retinopathy in younger onset type 1 diabetes.


N. Bilkhu1, H. Wharton1, M. Clarke1, A. Syed1, P. Dodson1,2
1Departments of Diabetes and Ophthalmology, Heartlands Hospital, Birmingham, United Kingdom
2School of Life and Health Sciences, Aston University, Birmingham, United Kingdom

INTRODUCTION. Retrospective analysis.
PURPOSE. To determine whether visual acuity (VA) recording is reliable and reproducible between Screening and Hospital Eye Service (HES). This is particularly important to those with referable maculopathy defined by macular haemorrhages/microaneurysms with VA ≥6/12, in the English Diabetic Retinopathy Screening Programme.
METHODS. A retrospective analysis was performed of East and North Birmingham patients with diabetes, that were referred to Ophthalmology for the first time from diabetic retinopathy screening for maculopathy over a 15 month period. VA was recorded using Snellen at Screening and HES appointments.
RESULTS. Data on 193 eyes of 140 patients were analysed. From Screening to the First HES appointment 35 eyes (19%) of 193 eyes showed an improvement in visual acuity, where 12% of eyes improved by one line, 6% improved by two lines, and 1% improved by three lines. Of these 193 eyes, there were 24 eyes (13%) of twenty-two patients that were referred with haemorrhages/microaneurysms within 1 disc-diameter of fovea with VA ≥6/12, of which 7 eyes (30%) of seven patients improved to ≤6/9 at the first HES appointment.
CONCLUSIONS. Visual acuity recorded during the screening process and HES shows an improved variability of 1 line or more using Snellen recording. This is important particularly for those patients with referable maculopathy, due to haemorrhages/microaneurysms in macula with VA ≥6/12, in which a third of patients show, that a repeat of VA's would not necessitate eye clinic referral. Therefore, training for a more accurate VA testing at Screening would be beneficial to prevent unnecessary eye clinic referrals. Optical coherence tomography can also be used to determine if a patient needs a referral to the eye clinic.


C.J. Clark1, I. Rostom1, A. Ashurst2, W. Gatling2, B.N. Matthews1,2
1The Royal Bournemouth and Christchurch Hospital NHS Foundation Trust, United Kingdom
2Dorset Diabetes Eye Screening Programme, United Kingdom

INTRODUCTION. Retrospective analysis of retinopathy screening grade for patients on the Dorset Diabetes Eye Screening Programme 2010-11 aged 90 and over.
PURPOSE. To ascertain whether retinopathy screening in this age group is necessary.
METHODS. The Orion database was analysed for screening grade and method of screening in a population of patients aged 90 and over at the time of screening in the year April 1st 2010-March 31st 2011. Additional information was collected for the group with a referable retinopathy grade (R2,R3,M1)to confirm diagnosis and treatment. Also data from laser coding was analysed to confirm whether any patients >90 years had been lasered for retinopathy.
RESULTS. 27,794 diabetic patients were screened in this year. 337 (1.2%) were aged >90years. 202 (60%) female; 135 (40%) male. Age range was 90-102 years. 237 (70%) were gradable by digital imaging. 100 (30%) were graded by slitlamp biomicroscopy compared to 8% of the overall population. 87 patients (25.8%) had background retinopathy (R1M0).13 patients (3.85%) had a referable retinopathy grade (R2,R3 or M1). Of these 10 had an M1 grade due to VA<6/12 with red dot in macula and 90% of those graded M1 had other pathology eg AMD, CRVO. One patient had stable treated diabetic maculopathy, another R3 patient had collaterals due to old CRVO, one patient was initially diagnosed and treated for diabetic maculopathy but subsequently developed wet AMD. One additional patient had been lasered for diabetic maculopathy but had been graded incorrectly as M0.
CONCLUSIONS. The prevalence of sight threatening retinopathy in this age group for this screening year was very low. 0.59% had diabetic maculopathy. None had proliferative retinopathy. Screening this age group digitally is often difficult and 30% require slitlamp examination which is more time consuming and costly. Further evidence is required but this data suggests screening in this age group is not necessary or justifiable.


C. Tadros1,2,3, N. Kontodimopoulos2, E. Feretis1, S. Kabanarou1, T. Peto3
1Hellenic Red Cross Hospital, Athens, Greece
2Hellenic Open University, Patra, Greece
3NIHR Biomedical Research Centre for Ophthalmology at Moorfields Eye Hospital NHS Foundation Trust and UCL Institute of Ophthalmology, London, United Kingdom

INTRODUCTION. Diabetic retinopathy (DR) and maculopathy (DMac) can have a negative impact on the physical, emotional and social well-being of patients and their families. Patient fear blindness the most and this allows ophthalmologists to put prevention and management of DR and DMac in the forefront of providing better quality of life.
PURPOSE. We have conducted a study on Health-Related Quality of Life (HRQoL) amongst Greek patients who had DR and DMac.
METHODS. The study enrolled 85 patients with DR/DMac who completed HRQoL tests using three different types of measurements: the NEI VFQ-25, the SF-12 and the EQ-15D. Patients were recruited while attending their scheduled appointment at the Diabetic Eye Clinic of the Hellenic Red Cross Hospital in Athens, Greece.
RESULTS. The response rate was 86.73%. There were 51 males, the average age was 57.7±16.4 years old, the average duration of diabetes was 10.9±6.6 years and 75.3% of patients had type II diabetes. The mean value of the composite VFQ-25 score of participants was 86.4 ± 13.6. Patients with proliferative DR and existing DMac scored lower on all VFQ-25 subscales compared to patients with non-proliferative DR and those without DMac, respectively, especially in the dimensions “General Health”(34.2±24.9 versus 74.3±15.7 and 25.1±18.7 versus 69.8±19.2) and “Role Difficulties”(33.7±28.3 versus 87.7±25.2 and 23.6±23.1 versus 79.9±26.7). In the SF-12 questionnaire the mean value in the Physical Component Scale (CS) was 45.6±11.3 and in the Mental CS was 45.2±16.8 while values in general Greek population are 49.4 and 48.9, respectively. In EQ-15D the mean value was significantly lower, 0.761, while published values in Greek diabetic patients without DR are 0.813. There was clinically significant correlation between the composite VFQ25 score and the 15D score as well as between the Physical and Mental CS of SF-12 and the 15D score.
CONCLUSIONS. This study identified that Greek DR/DMac patients report poorer HRQoL when compared to general population. Further studies are needed to explore the causes and potential interventions so the ophthalmic community can act on these in order to improve quality of life for these patients.


P. Douglas, C. Mallon, M. Purvis, J. Grattan, P M. Hart
Northern Ireland Diabetic Retinopathy Screening Programme, Royal Victoria Hospital, UK

INTRODUCTION. Database search, analysis of demographic data & retrospective image review
PURPOSE. To establish the trend in PQI over 5 years; to understand the risk factors for PQI; to identify areas for improvement.
METHODS. The annual PQI rate between 2006-2011 was identified. Demographic and clinical data on those with PQI in either/both eyes in 2010-11 was extracted from the Programme Database. In a 1:5 sample a detailed review of image/photographer/patient characteristics was undertaken.
RESULTS. PQI rate ranged from 3.0% to 4.5%, meeting the QA Standard of <5% in all 5 years.
In 2010-11 the total number of patients screened by the Programme was 37484 and in 4% referral was advised. 40% of these referrals were due to PQI; PQI rate was 3.4%. Mean age of PQI patients was 73.4 years (range 12-98, SD 12.3). Mean duration of diabetes was 10.3 years (range 0-68, SD 9.7). PQI rate for the right eye was 36%, left 40% (no significant difference), both 24%. 94% of PQI subjects were aged >50 at screening and 99% of these had undergone pharmacological dilatation. The PQI rate in those screened ≤50 years (not routinely dilated) was 1% and 4% in those >50 years (routinely dilated). 62% of patients with PQI required referral. Reasons for non-referral included being already ophthalmologically managed/known long-standing pathology in the subject eye. The detailed review on 288 patients showed: 90% had pharmacological dilatation; 95% (358/376 eyes) had anterior segment images (AS) captured. Of the 28 patients (39 eyes) not dilated, small pupil was identified in 18% (2% total sample reviewed) and compliance problems in 13%. Most showed evidence of media opacities on AS imaging. In the sample, 64% (184/288) were referred; 1% (3/288) referred urgently for suspected neovascularisation, demonstrating that even a poor image can potentially detect proliferative diabetic retinopathy and thus given a maximum retinopathy grade.
CONCLUSIONS. As we and others have previously shown, the prime risk factor for PQI is age at screening. Most PQI are patient-dependent, not operator/process dependent and have had dilating drops instilled. Optimal quality anterior segments are useful.


E. Bandurska-Stankiewicz1, W. Matuszewski1, D. Wiatr-Bykowska1, J. Pieczynski2, A. Wroblewska3,

J. Rutkowska1
11 Endocrinology, Diabetology and Internal Diseases Clinic, Chair of Internal Diseases, University of Warmia and Mazury in Olsztyn, Poland
22 Ophthalmic Ward, Provincial Specialist Hospital in Olsztyn, Poland
33 Endocrinology and Diabetology Ward, Provincial Specialist Hospital in Olsztyn, Poland

INTRODUCTION. Prospective study.
PURPOSE. Evaluation of the incidence of diabetic retinopathy (DR) in diabetic patients acc. to results in the Province of Warmia and Mazury.
METHODS. The screening programme for DR enrolled 1005 patients (2010 eyes) with Type 1 (DM1) and Type 2 (DM 2) diabetes, a random sample of patients from diabetes clinics and family doctor practices. The evaluation of the fundus was performed based on 50 degree colour fundus photography taken with a Topcon TRC NW8 fundus camera. The severity of DR has been assessed using the ETDRS scale.
RESULTS. A colour photograph of the fundus has been obtained from 1005 patients (2010 eyes) aged 55,2±15,6; 267 had DM1 (26,6%) at the age of 37,9±13,6 years with diabetes duration of 12,6±9,9 years. 738 (73,4%) suffered from DM2 with the average age of 61,4±10,9 and average duration of 10,4±8,1 years. DR was diagnosed in 20,4%, and macular oedema in 4,8 %. In DM1 patients, DR was present in 25,1% and macular oedema in 6,7 %. In DM2 patients, DR was present in 18,7% and macular oedema in 4,1%. The age of DM1 patients with eye disease was 44±13,3 years with 20,3±8,1 years of duration. The age of the same changes in DM2 patients was 62,2±9,6 with 15,4±8,8 years of duration.
CONCLUSIONS. Diabetic retinopathy has been diagnosed in 25,1% of DM1 and 18,7% of DM2 patients. Owing to the high number of the tested patients, the results can be extrapolated to the whole Polish population. Colour fundus photography is a simple and easy screening test for diabetic retinopathy when conducted by trained medical staff.


R. Svancarova1, T. Sosna1,2, R. Koznarova2, M. Netukova3, F. Saudek2, M. Adamec2
1 Department of Ophthalmology, Thomayer Hospital, Czech Republic
2 Diabetes Centre, Institute for Clinical and Experimental Medicine, Czech Republic
3 Ophthalmology Clinic, Faculty Hospital Kralovske Vinohrady, Czech Republic

INTRODUCTION. Prospective comparative study.
PURPOSE. To evaluate a difference in the effect of combined pancreas and kidney transplantation and isolated pancreas transplantation on the course of diabetic retinopathy.
METHODS. In our study we evaluated 98 patients of type 1 diabetes and 1 patient of secondary diabetes with a complete data of regular ophthalmological follow-up for an average of 86,4 months. Patients were divided into two groups: group 1) recipients of combined pancreas and kidney transplantation (SPK – simultaneous pancreas kidney transplantation; n=78, follow-up 122 ± 60 months), group 2) recipients of isolated pancreas graft (PTA – pancreas transplantation alone; n=21, follow-up 51 ± 47 months. Measurement of best corrected visual acuity (BCVA), evaluation of anterior segment and fundoscopy were performed before transplantation and at the end of follow-up period. Because of mostly very advanced fundoscopic findings of diabetic retinopathy quantitative evaluation of the lesions wasn't possible. Therefore changes of the fundus were rated semiquantitatively as clinically nonsignificant (stabilization), clinically significantly deteriorated or improved with the aid of photodocumentation.
RESULTS. BCVA at the end of follow-up was improved in 31,2% of cases, remained stable in 49,3% and was deteriorated in 19,5% in the SPK group. In the PTA group we noted amelioration in BCVA in 23,8%, stabilisation in 47,6% and deterioration in 28,6%. Fundoscopic findings were improved in 0%, stabilized in 84 % and got worse 16 % in the SPK group. The respective numbers for the PTA group was 7,7%, 77,3% and 15%. There was no statistically significant difference in the change of BCVA and in the course of diabetic retinopathy after transplantation between the two groups. Laser coagulation was performed in 87,2% of patients in the SPK group and in 42,9% in the PTA group prior to transplantation. The need for additional laser coagulation after operation was 21,8% in the SPK group and 19% in the PTA group.
CONCLUSIONS. Even though we have expected better results in the PTA group than in the SPK group, we have not proved a statistically significant difference in the effect of SPK and PTA transplantation on the course of diabetic retinopathy. We are fully aware of insufficient number of examined patients, especially in the PTA group, thus the study is to be continued with more patients and longer follow-up periods.


J. W. Miah1, H.M. Wharton1, P.M. Dodson1,2, M. Clarke1, J.M. Gibson1,2
1Department of Diabetes and Ophthalmology, Heartlands Hospital, Birmingham, United Kingdom
2School of Health and Life Sciences, Aston University, Birmingham, United Kingdom

INTRODUCTION. Retrospective Analysis
PURPOSE. To evaluate the morphology and location of optic disc haemorrhages (ODH) identified at diabetic retinopathy (DR) screening to establish whether particular ODH are predictive of ocular disease (e.g. glaucoma).
METHODS. Retrospective analysis of 77 patients who presented with ODH at DR screening in the Birmingham and Black Country screening programme between June 2009-March 2010. Mean age was 71 years (range 39-89). Cup/disc ratio (CDR), location and morphology of the haemorrhage were recorded. The outcome of the referral and the status of the ODH were followed up a year later.
RESULTS. Of the 77 referred, 34 patients were unassessed for possible glaucoma. Of the 43 patients that were assessed in the hospital eye service for glaucoma, 11 (26%) were diagnosed with glaucoma. These glaucoma patients mostly presented with flame haemorrhages (64%) and blot haemorrhages (36%). Haemorrhages tended to adjoin the margin of the OD (64%), and were more commonly flame shaped (64%). They less commonly occurred in the optic disc itself (36%), and were all blot shaped. The OD Cup/disc ratio (CDR) of the patients with glaucoma (n=11) ranged from 0.33-0.57. It is interesting to note the highest CDR was 0.68 in the 77 patients referred. 32 patients were confirmed as not having glaucoma. 24 (75%) of these patients presented with an ODH adjoining the margin, of which 20 (83%) were flame, and 4 blot (17%) shaped. Only 8 (25%) presented with an ODH in the OD, of which 6 (75%) were blot shaped. One year follow up of the 77 referred cases revealed that the ODH resolved in 45 (57%) patients while 10 (13%) still had an ODH present. 15 (21%) were still under ophthalmology hence digital retinal photos were not available for assessment. Six patients (8%) (age range 71-91 years) died within the year, and one lost to follow up.
CONCLUSIONS. The results suggest that a significant number of patients with ODH have glaucoma and that the differing morphology of the haemorrhage is not a major predictor i.e. blot versus flame shaped, adjoining or in the optic disc. The cup/disc ratio did not predict glaucoma either.


S. Pathai1, M. Rajan2, H. Weiss1, L. D'Costa1, A. Palaniraj3, C. Gilbert1, T. Peto4
1International Centre for Eye Health, London School of Hygiene & Tropical Medicine, London, United Kingdom
2Rajan Eye Care Hospital, Chennai, India
3Regional Institute of Ophthalmology & Govt Ophthalmic Hospital, Chennai, India
4 NIHR Biomedical Research Centre for Ophthalmology, at Moorfields Eye Hospital NHS Foundation Trust and UCL Institute of Ophthalmology

INTRODUCTION. Cross-sectional validity study
PURPOSE. To compare images obtained from the Appasamy ADF50 mydriatic fundus camera with an industry reference standard in terms of image quality and grading in patients with diabetic retinopathy (DR).
METHODS. 56 participants (112 eyes) underwent 7-field mydriatic digital photography with both cameras in random order. 224 image sets from the two cameras were graded independently for retinopathy levels and image quality. Grading levels attained with the ADF50 were compared to the reference standard. Sensitivity, specificity and levels of agreement were also determined.
RESULTS. Good overall correspondence was obtained between the two cameras regarding retinopathy levels (κ=0.77) and macular oedema (κ=0.79). The sensitivity of the Appasamy camera to correctly identify ‘no' versus ‘any' fundus changes by DR was 90.1%, and specificity 97.6%. Image quality was comparable for small pupil size (≤3mm).
CONCLUSIONS. These findings indicate that the ADF50 camera may be used reliably in similar settings, and as a screening device for diabetic retinopathy.


M. Foster, S. Wright, F. Stirling, L. MacDuff, PM. Hart
Diabetic Retinopathy Screening Programme, Royal Victoria Hospital, Belfast, United Kingdom

INTRODUCTION. Review of macular grades (M) given to patients screened in 2010-11, who had previous sight threatening diabetic Maculopathy (STDM) treated by laser photocoagulation (PLTM). Detailed and explanatory re-grading of PLTM eyes ignoring other non-macular details.
PURPOSE. This programme has agreed that patients who have undergone laser photocoagulation for STDM should be retained within an eye clinic, not routinely screened. The rationale includes: STDR already diagnosed, a modified retinal appearance making grading difficult and no evidence to support this modality for post treatment monitoring. In practice patients with PLTM are sometimes seen within the screening programme (e.g.) delays in hospital reviews, GP concern, patient concern. Using our grading protocol, M0, M1 are non referable Maculopathy grades; M2 (possible Maculopathy) and M3 (probable Maculopathy) are referable. Graders may however allocate M8, an un-gradable macula.
METHODS. This exercise was a retrospective data base review of all the grades allocated to PLTM cases (2010-2011). In depth re-grading was then carried and findings compared. Previous retinal images and relevant clinical correspondence was also utilized.
RESULTS. PTLM was recorded in 187 of 37484 (0.5%) patients screened in 2010-11. M8 was allocated in 63 (33.7%); M3 in 10.7%; M2 in 22.0%; M1/M0 (non referable) in 33.6%. Referable = 66.4%. Had no grading taken place, ophthalmology follow-up would have been advised in 100% PLTMs. On grading review of M8 cases, 5 (2.7%) retained M8 grade; 6 became M3, 6 M2, and 37 became M1/M0 (i.e.) non referable. Referable = 41.7%. There was no significant change in other grades allocated. Previous images were deemed very important.
Analysis of change in M allocated, showed that a modifiable M8 was often given when patients already attended an ophthalmologist OR when other referable retinopathy in either eye was identified. Thus graders reviewed and assessed the value of the time spent grading a difficult macula, especially considering the final outcome and/or further management was not influenced. Difficulties with PLTM cases: time consuming; lack of relevant previous comparative images; identification of laser spots and altered lipid; modified macular morphology.
CONCLUSIONS. Further work is required to ensure confident management of PLTM cases. Some patients could probably attend screening, if post discharge images were captured.


S E. Johnston, P G. McElduff, C A . McKenna, R S . Robinson, H. I. Wilson, P. M. A . Hart
Diabetic Retinopathy Screening Programme N. Ireland, Belfast HSC Trust, Belfast, United Kingdom

INTRODUCTION. Retrospective re-audit of the expedited report pathway, on all urgent referrals advised between 2009 and 2011.
PURPOSE. To assess if current QA standards are being met, what progress had been made since last audit (06-07) and identify any areas for improvement.
METHODS. For all urgent reports (09-11) the following were recorded - diagnosis, pathway from screening to generation of report and pathway timescales.
The results were analyzed and findings compared with those of the 06-07 audit and with agreed QA standards.
RESULTS. 217 expedited reports were sent, 0.30% of total number screened and 6.32% of all referrals. In 06-07, 54 expedited reports were sent, 0.3% of total screened and 5% of total referrals. In 125 cases (57.6% urgent referrals) Proliferative Diabetic Retinopathy was the reason for the urgent referral. (87% in 2006-7). The difference in this % probably reflects the introduction of a dedicated ARMD service with a fast track pathway. The reason for 22.12% of urgent reports in 2009-11 was a suspected treatable CNVM. Of the PDR cases 103 (82.4%) were flagged as ‘urgent' at time of image capture (78.7% in 06-07). Of the 22 not flagged as ‘urgent' at image capture, 9 were already attending an ophthalmologist. In 67.28% reports were issued within two weeks image capture (57.41% in 06-07). In the period 2010-2011 alone this was 88%.
CONCLUSIONS. Within this expanding service the proportion of expedited reports remains fairly constant and the variations of diagnosis on the reports reflects changes in local ophthalmology services. Both identification of PDR at time of image capture and timely reporting continue to improve. It remains difficult to meet locally set standards however these are currently being reviewed.


SJ. Wilson, C.A. McKenna, J. Moore, P.M. Hart
Diabetic Retinopathy Screening Programme, N. Ireland, Belfast HSC Trust, Belfast, United Kingdom

INTRODUCTION. Retrospective review

PURPOSE. The aims were firstly, to identify the number of patients referred with suspected glaucomatous disc/s (SGD), secondly, to review current referral protocol for SGD (C:D ≥ 0.5) and finally to provide information which may be useful in consideration of local protocols.
METHODS. During 2009 (12 months), for each referral with SGD, the following were recorded: patient ID; feedback from ophthalmology assessment and review of optic disc features present on DRSP images.
RESULTS. 103 patients (0.003% of total screened) were referred with SGD. 30 were excluded (8 did not attend for assessment, 9 died prior to assessment, 11 were not referred by their GP and feedback was not available for 2) giving N=73. 43 patients (58.9%) were confirmed as having glaucoma, 11 patients (15.1%) remain under review as glaucoma suspects and 2 patients had optic atrophy.17 patients (23.3%) had a diagnosis of no glaucoma of which 4 had large discs and hence physiological cupping. 5 had asymmetrically increased C:D ratio because one disc was larger than the other.1 patient had anomalous discs. Of those diagnosed with glaucoma, 39/43 (91%) had C:D ratio ≥ 0.5 and were referred according to protocol. 2 had C:D ratio < 0.5 with a notch and disc asymmetry. 2 had C:D ratio < 0.5 with a notch. 27/43 had a C:D ratio ≥ 0.7.
CONCLUSIONS. During the sample period, 0.06% of all referrals from DRSP were for SGD. Although it is not possible to accurately measure disc size from digital images, disc size should be considered. Current practice would have advised referral in 91% of those confirmed as having glaucoma. If the protocol had been C:D ratio to ≥0.7 this would have excluded 16/43 (37%) of those subsequently diagnosed with glaucoma, and if protocol the protocol had been C:D ratio ≥ 0.7 or another feature (notch or disc asymmetry ≥ 0.2) 100% of patients would have been referred. This review provides information that may be useful for local discussion.


S. K. Golar1, H . Wharton1, M. Clarke1, J.M. Gibson1,2, P.M. Dodson1,2
1Departments of Ophthalmology and Diabetes, Heartlands Hospital, Birmingham, United Kingdom
2School of Life and Health Sciences, Aston University, Birmingham, United Kingdom

INTRODUCTION. Retrospective study
PURPOSE. To assess whether there are changes in foveal thickness (FT) and total macular volume (TMV) in pregnancy in diabetic subjects.
METHODS. The audit consisted of pregnant women with diabetes, with no maculopathy, who completed their antenatal care at Birmingham Heartlands Hospital. The Zeiss Stratus Optical coherence tomography (OCT) was performed on patients attending diabetic retinopathy (DR) screening at intervals throughout their pregnancy. To be included in the audit patients' had to have at least one OCT scan during their pregnancy.
RESULTS. Altogether there were 8 type 1 and 22 type 2 patients with mean diabetes duration of 6 years (range 1-20). Mean gestation of DR screening with OCT during the first trimester was 9.7 weeks (6-13) (n=22). The mean and standard deviation for FT for the right was 179.1µm ±21.49 and for left eye was 187.3µm ±23.55. The mean TMV was right 6.43µm ±0.35 and left 6.50µm ±0.39. The mean gestation of DR screening with OCT during the second trimester was 23.4 weeks (18-26) (n=25). The mean FT for the right was 191.4µm ±22.70 and the left 195.6µm ±24.77. The mean TMV was right was 6.74µm ±0.45 and left was 6.91µm ±0.35. The gestation of DR screening with OCT during the third trimester was 31.1 weeks (27-36) (n=15). The mean FT for the right was 181.5µm ±24.84 and for the left 193.1µm ±28.55. The mean TMV was right 6.80µm ±0.40 and left was 6.84µm ±0.31. There were no significant differences in FT over the three trimesters. The TMV showed a significant difference when comparing the first and second trimesters (p=<0.05). However, there was no significant statistical difference in TMV in the second and third trimesters. None of the patients showed any macula oedema on the OCT.
CONCLUSIONS. The results suggest there is no significant change in foveal thickness in pregnancy in diabetic subjects. There was a significant statistical difference in total macular volume in the second trimester, however this would not be clinically significant. This is an important observation proven by the OCT which has not been previously studied.


F. Picton1, B. Hamill2, T. Peto1, K.A. Muldrew1, MJ. Quinn1 on behalf of NetwORC UK
1Reading Centre, Moorfields Eye Hospital, London, United Kingdom
2Reading Centre, Centre for Vision and Vascular Science, Queens University Belfast, Belfast, United Kingdom

INTRODUCTION. Observational study, with inter-observer variability.
PURPOSE. To compare the reproducibility of field thickening on stereo colour fundus images, and of the maximum retinal thickness measurements on Stratus OCT images in participants with Diabetic Retinopathy.
METHODS. A set of 10 participant images were randomly selected from a Diabetic Macular Oedema Study undertaken by a reading centre network. From each participant image set an F2 colour stereo pair and an OCT scan were selected. Measurements for the EDTRS Retinal Thickening were made on the colour images. Two aspects of Retinal Thickening were graded (1) presence or absence of retinal thickening due to oedema (2) the maximum thickness, assessed subjectively on the colour images and by manual drawing on the OCT image.
All images were independently graded by 6 accredited graders of the NetwORC UK and 1 clinician from the 3 sites of this Reading Centre Network of the UK.
RESULTS. In 9 out of 10 cases there was 100% agreement on the presence or absence of retinal thickening on the colour images. In the one other set, 2 graders graded thickening absent and 5 graded thickening present.
There was less agreement, however, in the subjective measurement of thickening height on the colours. Of the 7 options for thickening height, in 5 cases there was agreement within 3 category levels and in the other 5 cases there was agreement within 4 category levels. In the OCT maximum retinal height measurements there was good agreement amongst graders with reproducibility of 14%.
CONCLUSIONS. Good agreement and reproducibility was found in this exercise. Grading the colour images for qualitative assessment shows better agreement than quantitative assessment of retinal thickening. However results show that reproducible quantitative measurements can be achieved by grading the OCT images.


M.G. Ratzlaff1, 2, A. Murphy2, K. O'Connor2, C. O'Brien2, J. Loughman3, P. Connell2
1University College Dublin School of Medicine & Medical Science, Dublin, Ireland
2Mater Misericordiae University Hospital, Dublin, Ireland
3Dublin Institute of Technology, Dublin, Ireland

INTRODUCTION. An observational, cross-sectional audit study.
PURPOSE. To compare macular pigment optical density (MPOD) in diabetic and nondiabetic patients by using heterochromatic flicker photometry and to investigate the correlation of MPOD with diabetes phenotype, glycosylated hemoglobin (HbA1C), serum lipid levels, and levels of dietary carotenoids.
METHODS. Patients were recruited from diabetic retinopathy (DR) clinics at the Mater Misericordiae University Hospital. Sixty-two patients were divided into controls (unremarkable ocular history, n=30) and patients with diabetes (type 1 or 2 diabetes with no or nonproliferative DR, n=32) (Note: data collection is ongoing and the diabetic group will be separated into those with no DR and those with nonproliferative DR). MPOD was measured with a macular densitometer and compared between groups (analysis of variance). The correlation of diabetes phenotype (type, duration, level of retinopathy, BMI), HbA1C, serum lipid (HDL, LDL, total cholesterol, and triglycerides) levels, and levels of dietary lutein and zeaxanthin (using the LZQ Dietary Questionnaire) with MPOD are determined for each group (linear regression analysis).
RESULTS. Mean +/- SD ages between controls (53.8 +/- 15.8 years) and diabetics (58.9 +/- 14.9 years) were similar (P=0.199). Mean +/- SD MPOD comparing a point 7 degrees with 0 degrees from the fovea was significantly lower in patients with diabetes (0.29 +/- 0.19 density units [DU]) than in controls (0.45 +/- 0.38 DU) (P<0.05).
CONCLUSIONS. To our knowledge, this is the first study to demonstrate reduced MPOD in patients with diabetes patients when compared with nondiabetic patients using heterochromatic flicker photometry. Further studies are needed, and these may show correlations between MPOD and phenotype of diabetes, HbA1C, serum lipid levels, and level of dietary carotenoids.


U.E. Bautina, D.V. Lipatov, D.S. Atarshchikov, T.A. Chistyakov, A.G. Kuzmin, N.P. Goncharov,

M.V. Shestakova
Federal State Endocrinology Research Centre,

INTRODUCTION. Prospective cohort study in humans
PURPOSE. The intravitreal anti-VEGF agents were shown to be effective for DME. But this treatment hasn't been investigated thoroughly in DME and proliferative DR simultaneously. The aim was to estimate the effectiveness of the anti-VEGF treatment for DME in combination with proliferative DR.
METHODS. Three intravitreal injections of the anti-VEGF (ranibizumab) monthly were performed in groups of the twenty patients with DME and proliferative DR. Ophthalmic examination, optic coherent tomography (OCT) with central retinal thickness (CRT) measurement, colour fundus photography and fluorescein angiography were done before the treatment, between the injections and at 4 weeks after the last injection. The effectiveness of the treatment was checked by OCT exam, colour fundus photography, visual acuity and requirements of laser photocoagulation of the retina.
RESULTS. Preoperatively the central retinal thickness (CRT) was 632±156 µm, visual acuity was 0.7 (the measurement was performed by the Golovin-Sivtsev table from a 5-m distance). Only 2 intravitreal injections (II) were performed in groups of 6 patients. The final report will have done in the march 2012. Preliminary data shown that average reduction of the CRT was 45±18 µm, the visual acuity enhanced up to 0,8-0,9 letters. It was observed the regression of the neovascularization after first II in 75% of cases. No adverse effects have been observed during the study.
CONCLUSIONS. Anti-VEGF agents might be effective in treatment DME and proliferative DR simultaneously.


M Treacy, L O Toole, Y Delaney, A Murphy, K O Connor, P Connell, C O Brien

Department of Ophthalmology, Mater Misericordiae University Hospital, Dublin, Ireland

INTRODUCTION. A prospective study of 800 consecutive patients within a Diabetic Screening Programme
PURPOSE. To identify optic discs suspicious for glaucoma within our Diabetic Screening Programme
METHODS. Patients with suspicious optic discs were identified on an opportunistic basis. We then graded the optic discs using a ten point scoring system. Those patients who were noted to have a high risk score were subsequently referred to the glaucoma services for review.
RESULTS. 35 patients were identified with suspicious discs and referred to the glaucoma services. Six cases were subsequently diagnosed with chronic open angle glaucoma. One case was determined to have low tension glaucoma and one case was diagnosed with ocular hypertension. Twenty six patients were given follow up appointments with the glaucoma services. Two patients were discharged back to the Diabetic Screening Programme.
CONCLUSIONS. The purpose of a Diabetic Screening Programme is to detect sight threatening retinopathy, not to detect concomitant ocular disease. However it is important that evident optic nerve pathology is not overlooked. This scoring system allowed graders to identify suspicious discs and make appropriate referrals. By alerting our graders to glaucomatous disc changes, detection of sight threatening glaucoma can occur within a Diabetic Screening Programme.


S.M.G. Beare1,2, S.J. Wilson1,2, P.M. Hart1
1Northern Ireland Diabetic Retinopathy Screening Programme , Belfast HSC Trust, Belfast, United Kingdom
2Macular Service, Belfast HSC Trust, Belfast, United Kingdom

INTRODUCTION. Retrospective review of feedback to the screening programme from a dedicated macular service.
PURPOSE. Within the screening programme, ARMD2 is used to code the appearance of a suspected treatable CNVM i.e. macular blood not thought to be diabetes related and/or grey/green lesion suggestive of a CNVM. This internal quality assurance exercise sought to: identify the number referred with ARMD2, ascertain the Positive Predictive Value (PPV) of ARMD2 for a treatable CNVM and review the current ARMD2 protocol.
METHODS. During May to November 2009, all cases referred with ARMD2 were identified and details recorded on; patient ID, macular features on digital image, outcome of referral to the macular service and retrospective review of DRSP images.
RESULTS. 23 eyes were referred with ARMD2 (0.11% of total screened in that period). This represented 42% of all urgent referrals in that period of time. 15 of these (65.2%) had a CNVM confirmed at the macular service, and in 14 intravitreal anti-VEGF treatment was advised. One patient was diagnosed with a small CNVM but was not treated (VA was 6/6, and patient asymptomatic). 5 of those referred had dry Age Related Macular Degeneration, 1 a macular hole, and 2 eyes adult vitelliform disease. Review of screening images showed 15 eyes had macular blood, 14 of these had a CNVM confirmed, one had a macular hole. Blood was absent in the screening images of 8 ARMD2 referrals. 5 of these were diagnosed with dry macular degeneration, 2 with adult vitelliform disease and 1 with a small CNVM, untreated.
CONCLUSIONS. The management of other pathologies identified in passing within any screening programme can be a challenge. Treatment centres for CNVM patients stress the importance of prompt referral for suspected cases of this serious sight-threatening disease. The positive predictive value of ARMD2 from a DRSP for a CNVM was found to be 0.65. The presence of blood in this small sample was the strongest predictor of a treatable CNVM.


L. Falck, M. Birch Viese, N. Larsen
Øjenlægernes Hus, Copenhagen, Denmark

INTRODUCTION. A pilot study was made to get an overview of the level of diabetic retinopathy among known diabetic patients in Nuuk, Greenland, based on digital retinal imaging.
PURPOSE. The prevalence of diabetes in Greenland is increasing. Our purpose of this study was to investigate the level of diabetic retinopathy (DR) among known diabetic patients in Nuuk, Greenland,
METHODS. 185 known diabetic patients from Nuuk were screened at the local hospital in the period from 2006 - 2011. If one patient were screened more times the latest images were used. The digital fundus images were obtained with a nonmydriatic camera (Zeiss Nonmy VisionCam 45 degrees). The images were stored and graded for DR levels using VisioPac (Zeiss) and EyeCare (Steno Diabetes Centre, Denmark).
RESULTS. 135 (73%) patients had no DR, 28 (15%) patients had mild DR, 22 (12%) patient had moderate DR and none had severe DR. The duration of known diabetes was less than 10 years for 128 (69%) patients. 175 (95%) patients had type 2 diabetes and 10 (5%) patients had type 1 diabetes.
CONCLUSIONS. The level of diabetic retinopathy is still very low, despite the increasing prevalence of diabetes in Greenland. The best predictor of DR is the duration of diabetes. Since the predominant length of diabetes among the patients was less than a decade, a low level of DR was expected. However, as the duration and prevalence of diabetes increases in time, the level of DR in Greenland is likely to change. Thus implementation of regular screening, treatment in appropriate time and surveillance of diabetic retinopathy is highly recommended in Greenland.

T. Peto1,2,3 on behalf of NetwORC UK
1NIHR Biomedical Research Centre for Ophthalmology, at Moorfields Eye Hospital NHS Foundation Trust and UCL Institute of Ophthalmology, London, United Kingdom
2Queen's University, Belfast, United Kingdom
3St Paul's Eye Unit, University of Liverpool, Liverpool, United Kingdom

INTRODUCTION. Concordance grading
PURPOSE. NetwORC UK consists of three Reading Centres (RC): Moorfields Eye Hospital's in London, Queen's University in Belfast and St Paul's Eye Unit in Liverpool, all in the UK. Image reading for large studies are distributed amongst graders of the RCs using a common platform for image analysis and data entry. Graders are often involved in training and quality control for national and international programmes therefore keeping quality and repeatability at a high level is important. The purpose of this study was to determine the concordance between graders using the UK National Screening Committee (NSC) grading criteria.
METHODS. In total, 400 eyes were graded by 8 graders of the NetwORC UK. The images used covered the whole spectrum of diabetic retinopathy (DR) and maculopathy (DMac) from eyes with no disease visible on the images to proliferative disease. Graders also were asked to make a decision on potential referral of the subject. Finally, they were asked to time their grading.
RESULTS. Exact agreement and kappa values showed high level of concordance (62.5%-88%, p<0.05 for all, respectively). There was high level of agreement both for DR and DMac amongst graders ranging between 62.1%-97.9% (p<0.05 for all categories). For action to be taken based on the images only the agreement ranged between 70.0%-94.5%. (p<0.05). Those trained to NSC standards tended to have higher level of agreement with the system grade. On average, 1.5 minutes grading time was needed per eye, slightly longer for eyes without diabetic changes than those with.
CONCLUSIONS. The graders had good agreement with the system grade and performed well on the referral criteria following their training at NetwORC UK. Shortening grading times on normal eyes,
could potentially have a large impact on manpower needed in screening and grading for DR and DMac.

S. Doherty1, D. Quinlan1,2, S. Mc Hugh1,3, I. Perry1,3, G. O'Keefe1,4, M. James1,5, E. Keane1,6
1Association of Optometrists Ireland, Greenmount House, Harolds Cross Road, Dublin, Ireland
2Diabetes in General Practice, Lancaster Hall, 6 Little Hanover St, Cork, Ireland
3Department of Epidemiology and Public Health UCC, College Road, Cork, Ireland
4General Manager HSE South Lee, Cork, Ireland
5Department of Ophthalmology, Cork University Hospital, Cork, Ireland
6Director of Public Health for Cork and Kerry HSE, Sarsfield House, Wilton, Cork, Ireland

INTRODUCTION. This study used a cross-sectional design to evaluate community based diabetic retinopathy (DR) screening. 32 GP practices invited 3598 patients to attend local optometrists for screening. One invitation letter was issued. The results and outcomes were compared to the quality assurance standards outlined in the framework for development of a diabetic retinopathy screening programme in Ireland drafted by the EAG National Retinopathy Screening Committee.
PURPOSE. 1. To assess the provision of community- based DR screening using existing optometry/ophthalmology services in terms of structure, process and outcomes. 2. To compare the performance of the initiative to the objectives and standards proposed in the national framework document, in terms of uptake, quality and timeliness of the screening procedure. 3. To make recommendations to the development and delivery of a national quality assured diabetic retinopathy screening programme in Ireland
METHODS. The participants included patients, GP practices, Optometrists and Community Ophthalmologists. Patients with type 1 and type 2 diabetes, registered with the participating GP practices in Cork Kerry region (32) were eligible to attend (N=3598). Excluded were patients with pre-diabetes, patients under 18 and those who did not have perception of light in either eye. Patients underwent dilation and digital-retinal photography. Images were uploaded to custom designed feature based grading software provided by Medical Diagnostic Treatment Solutions. All images were both primary and secondary graded. Arbitration level grading was carried out where there was disagreement between primary and secondary grading.
RESULTS. Uptake was 49% resulting in 1763 patients screened. Prevalence of any DR was 25.6% with a low prevalence (<1%, n=11) of Proliferative DR. Six of these patients had no record of previous ophthalmology appointment. Overall, 3% had Pre Proliferative DR. Of the whole sample, 4.4% had clinically significant Maculopathy, of these 54% had mild DR, 35% had severe DR and the rest were PDR. Six percent of the images were ungradable.
CONCLUSIONS. Results closely followed those expected. Uptake was lower than desired however some patients already had ophthalmology appointment and decided not to avail of free screening. Initiative addressed the need for community based screening which can be enhanced using existing Optometry/Ophthalmology resources.


F. Ibrahim1, S. O'Connor2, A. McNabb2, A. Dunne2, H. Nolan1, M.T. Cahill1
1Research Foundation, Royal Victoria Eye and Ear Hospital, Dublin, Ireland
2Centric Health, Rockford Medical Campus, Ballaly, Dublin, Ireland

INTRODUCTION. Hospital-based prospective cohort study
PURPOSE. To asses sensitivity and specificity of automated grading and determine role of OCT in a screening programme
METHODS. Patients underwent digital photographic screening; image sets using Digital Healthcare OptoMize software and automatically using Medalytics iGrading system Patients with pre-proliferative DR (grade R2), proliferative DR (grade R3) and definite DMO(grade M1) were referred to a Hospital eye clinic, as were patients cataract, abnormal discs, retinal vascular disease and neovascular AMD. Patients with possible DMO (grade M1), AMD and choroidal naevus were referred to an Ophthalmic Photographic Diabetic Review (OPDR) clinicwhere a second set of 45° fundus photographs were taken as was a macular SD-OCT. This allowed decisions to be made regarding the need for referral to a Hospital clinic during the OPDR visit.
RESULTS. 452 patients were screened. 79 patients (17.50%) were referred for diabetic eye disease; 40 patients had sight threatening DR. 20 patients (4.4%) had proliferative DR and 20 patients had DMO (4.4%). 174 (38.0%) patients had other eye diseases, the most common were AMD (18.4%), cataract (11.1%) and glaucoma (4.6%). Using SD-OCT in the OPDR reduced the number of patients referred to the hospital clinic for DMO by 51% and the number of patietns referred for AMD by 86%. Manual grading was more sensitive and specific than automated grading , which was not calibrated to detect non-diabetic eye disease.
CONCLUSIONS. Prevalence rates of stDR were similar in this study to those reported in other studies of unscreened diabetic populations. Accurately detecting DME (grade R1M1) and dealing with the workload of non-diabetic eye disease referrals, particularly assessment of cataract, glaucoma and AMD, are the two main challenges when screening for DR. The use of automated grading software, agreed protocols for assessing non-diabetic eye disease detected on graded photographs and SD-OCT in OPDR clinics are three innovative methods that could help meet these challenges when planning a population based DR screening service


S.N. Dhubhghaill1, M. O'Rourke2, N. Collins2, S. Murphy1, F. Hayes1, N. Horgan1,2
1St. Vincent's University Hospital, Dublin, Ireland
2Royal Victoria Eye and Ear Hospital, Dublin, Ireland

INTRODUCTION. The screening service was established as a prospective interim project in 2010 to provide diabetic screening until a national program is established. All participants were new referrals to the ophthalmology services for diabetic assessment.
PURPOSE. To assess the prevalence of diabetic retinopathy and maculopathy in a cohort of newly referred diabetic patients. Secondary goals were generate a demographic profile for the service, generation of non-diabetic eye referrals and the burden of ophthalmic follow-up assessment and treatment resulting from this hospital based screening program.
METHODS. Participants were assessed for associated risk factors including hypertension, smoking, duration and type of diabetes and recent HbA1C measurements recorded. Imaging consisted of 2 fundal photographs per eye which were graded by a consultant ophthalmologist. The images were graded on the standard grading criteria.
RESULTS. A total of 247 attended for screening from June 2009 breaking down to 12.6% type 1 and 87.4% type 2. The average age at referral was 54.63 yrs [SD 15.53] of which 38.4% were female.
On analysis of risk factors 13.87% were smokers and 86.12% were non-smokers. Hypertension was noted in 60.82% and renal impairment in 5.71%. The median year of diabetic diagnosis was 2010 and average HbA1C at referral 7.35% [SD 1.609]. There was no evidence of retinopathy in 88.26% and these were given screening follow up in 1 year. Patients eligible for 6 monthly screening amounted to 0.4%. A total of 17.34% required referral for ophthalmic assessment. Of the referrals to ophthalmology clinic non diabetic eye conditions [76.74%] exceeded diabetic complications diabetic complications [23.25%].
CONCLUSIONS. In an 18-month pilot diabetic screening project, 247 patients were screened.
17% of patients screened required ophthalmic follow-up, of which non-diabetic disease outnumbered diabetic disease by 3 to 1. The rate of diabetic retinopathy detection requiring ophthalmic specialist referral was 4.04%. The low rate of diabetic complications may be accounted for by the early stage of referral after the initial diagnosis of diabetes.


  « previous page    


W3C standards logos

site developed by Netsima Ltd »