Diabetic Retinopathy: More Patients, Less Laser A longitudinal population-based study in Tayside, Scotland JAMES H. VALLANCE, MBCHB, BSC, MRCOPHTH1 PETER J. WILSON, MBCHB, BSC, MRCSED1 GRAHAM P. LEESE, MD, FRCP2,3 RITCHIE MCALPINE, BSC4 CAROLINE J. MACEWEN, MD, FRCS, FFSEM FRCOPHTH1 JOHN D. ELLIS, MPH, PHD, FRCOPHTH1 OBJECTIVE — We aim to correlate the incidence of diabetic retinopathy and maculopathy requiring laser treatment with the control of risk factors in the diabetic population of Tayside, Scotland, for the years 2001–2006. RESEARCH DESIGN AND METHODS — Retinal laser treatment, retinal screening, and diabetes care databases were linked for calendar years 2001–2006. Primary end points were the numbers of patients undergoing first or any laser treatment for diabetic retinopathy or maculopathy. Mean A1C and blood pressure and retinal screening rates were followed over the study period. RESULTS — Over 6 years, the number of patients with diabetes in Tayside increased from 9,694 to 15,207 (57% increase). The number of patients receiving laser treatment decreased from 222 to 138 and first laser treatments decreased from 100 (1.03% of diabetic population) to 56 (0.37%). The number of patients with type 2 diabetes treated for maculopathy decreased from 180 in 2001 to 103 in 2006 (43% reduction, P � 0.03). Mean A1C decreased for type 1 and type 2 diabetic populations (P � 0.01) and a reduction in blood pressure was observed in type 2 diabetic patients (P � 0.01). The number of patients attending annual digital photographic retinopathy screening increased from 3,012 to 11,932. CONCLUSIONS — Laser treatment for diabetic maculopathy in type 2 diabetic patients has decreased in Tayside over a six-year period, despite an increased prevalence of diabetes and increased screening effort. We propose that earlier identification of type 2 diabetes and improved risk factor control has reduced the incidence of maculopathy severe enough to require laser treatment. Diabetes Care 31:1126–1131, 2008 A number of recent studies have re- ported a lower incidence and preva- lence of severe diabetic retinopathy and maculopathy (1–5). Reduction in blindness in patients with diabetes has also been reported, but this observation is not universal (6 – 8). The use of blindness as an end point for studies of diabetic eye disease is often rendered imprecise by reliance on incomplete blindness registration data and by difficulty in attributing visual loss to di- abetic retinal disease (9). The majority of visual impairment in patients with diabe- tes is not due to diabetic retinopathy (10), and accordingly the incidence of retinop- athy requiring therapeutic intervention (laser) is a more accurate reflection of in- cident diabetic retinal disease provided population and treatment records are complete. National Health Service (NHS) Tay- side serves a predominantly Caucasian rural and urban population, which in- creased from 338,750 in 2001 to 391,639 in 2006 (11). A retinal screening program has been in place since 1990, using digital photography since 2000 (12). In 2003, Scotland introduced a national screen- ing program (13) using annual single- field digital photography with staged mydriasis, a standardized grading sys- tem (14), trained screeners, and rigor- ous quality assurance (15). Tayside also benefits from an established national di- abetes database (16 –18). Laser treat- ments take place at a single site within the region and are recorded on a single database using the same unique patient identifier, allowing easy case linkage studies. Using these data sources, we describe trends in laser utilization, ret- inal screening, and the control of reti- nopathy risk factors in Tayside for the years 2001–2006. RESEARCH DESIGN AND METHODS — We performed a histor- ical cohort study of retinal laser in Tay- side, Scotland. The data sources used in this study were databases of regional laser treatment, retinal screening (“Eyestore”), and the national diabetes register (Scot- tish Care Information–Diabetes Collabo- ration [SCI-DC]) for the complete calendar years 2001–2006. Retinal laser within Tayside is re- corded on a custom-designed database, including treatment given and date. The primary end points for this study ob- tained from this dataset were first laser treatments for diabetic retinopathy or maculopathy and number of patients re- ceiving any laser for diabetic retinopathy or maculopathy per annum. The SCI-DC database uses hierarchi- cal multiple data source captures to create a real-time national diabetes register. In- dependent data sources (e.g., community prescribing, regional biochemistry data- base) are integrated using custom- designed software (16). The health board regions are clearly demarcated and there- fore can be accurately constrained to the Tayside population (16,18). Population risk factors for laser extracted from ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● From the 1Department of Ophthalmology, Ninewells Hospital and Medical School, Dundee, U.K.; the 2University Department of Medicine, Ninewells Hospital and Medical School, Dundee, U.K.; the 3Diabetes Centre, Ninewells Hospital and Medical School, Dundee, U.K.; and the 4Medicines Monitoring Unit, Nine- wells Hospital and Medical School, Dundee, U.K. Corresponding author: Dr. John Ellis, MPH, PhD, FRCOphth, Department of Ophthalmology, Ninewells, Dundee, U.K. DD1 9SY. E-mail: john.ellis@nhs.net. Received for publication 1 August 2007 and accepted in revised form 3 March 2008. Published ahead of print at http://care.diabetesjournals.org on 17 March 2008. DOI: 10.2337/dc07-1498. Abbreviations: NHS, National Health Service; SCI-DC, Scottish Care Information–Diabetes Collabora- tion. © 2008 by the American Diabetes Association. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. E p i d e m i o l o g y / H e a l t h S e r v i c e s R e s e a r c h O R I G I N A L A R T I C L E 1126 DIABETES CARE, VOLUME 31, NUMBER 6, JUNE 2008 SCI-DC were A1C, duration of diabetes, and blood pressure. BMI, cholesterol, and method of diabetic treatment were also extracted. Eyestore contains all information from digital retinal screening performed in Tayside including date of screening and grading outcome (12). Data drawn from Eyestore were total number of screening events and number of events at which re- ferable retinopathy or maculopathy were identified for each year. Referable reti- nopathy and maculopathy were as de- fined in the national screening framework (14). For the purposes of this study, a screening event resulting in treatment was defined as one that occurred no more than 6 months before laser. This defini- tion was used to state with reasonable confidence that screening had identified treatable pathology, not merely referable pathology. This assumption could not be made for laser occurring �6 months after screening, since this could well encom- pass new pathology arising during oph- thalmic clinic follow-up. The three databases were checked for internal validity (Modulus 11 algorithm, identification and exclusion of noninci- dent laser events), and external cross- references between the databases were made. Where discrepancies were identi- fied, arbitration was sought from bio- chemical and clinic attendance records. In addition to the data described above, unique patient identifiers were obtained and matched between the relevant data- bases, before anonymization of the data by a third party. The opinion of the local medical re- search ethics committee was sought. They indicated that Caldicott Guardian ap- proval alone was required. This was ob- tained, and the principles of data protection were adhered to throughout this study. Statistical analysis The administration of SCI-DC changed after the first 2 years of the study period. As a result, with the exception of disease duration, only means of variables were available for 2001 and 2002. Neverthe- less, the large sample sizes meant that this was an acceptable representation of the group. To demonstrate trends in these variables, weighted linear regression was performed, using (N/SD2) to calculate weight (SPSS, Chicago, IL). Since accu- rate measures of N and SD were not avail- able for 2001 and 2002, the weights were estimated allowing for low patient num- bers and high SDs. The robustness of this technique was tested and validated through comparison with the full dataset for duration of disease. Statistical analysis was performed under the supervision of the statistician for NHS Tayside. RESULTS — From 2001–2006 the number of registered diabetic patients in- creased from 9,694 to 15,207 (57% increase). The number of first laser treat- ments per annum fell from 100 to 56 (44% decrease), and the total number of patients receiving laser fell from 222 to 138 (38% decrease). The number of pa- tients undergoing digital retinal photo- graphic screening annually rose from 3,012 in 2001 to 12,035 in 2005. A total of 55,103 retinal screening events were performed (47,864 patients), 1,884 (3.4%) of which identified referable ret- inopathy. However, of patients referred Figure 1— The relationship between the known diabetic population of Tayside, digital retinal photographic retinopathy screening, and progression to laser treatment for the years 2001–2006. On the primary axis: F � total number of patients with diabetes and f � patients undergoing digital retinal photographic retinopathy screening in that year. On the secondary axis: E � patients graded as having referable retinopathy at screening as defined by national guidelines, � � all patients undergoing any form of laser treatment in that year, and ‚ � number of patients undergoing laser within 6 months of a screening event detecting referable disease. Vallance and Associates DIABETES CARE, VOLUME 31, NUMBER 6, JUNE 2008 1127 Table 1—Number of patients with diabetes, number of patients undergoing digital retinal photographic screening, and number of patients undergoing first or any laser treatment in the years 2001–2006 2001 2002 2003 2004 2005 2006 Patients with diabetes† 9,694 11,216 11,932 13,582 14,811 15,207 Prevalence of diabetes in Tayside (%) 2.5 2.9 3.1 3.5 3.8 3.9 Patients undergoing digital retinal photographic screening† 3,012 3,238 6,216 10,294 12,035 11,932 Patients with referable retinopathy from photography 189 149 262 425 302 343 Percentage of patients screened with referable retinopathy (%)† 6.3 4.6 4.2 4.1 2.5 2.9 All patients receiving laser for diabetes 222 201 202 252 199 138 % of all patients with diabetes receiving laser† 2.3 1.8 1.7 1.9 1.3 0.9 Patients receiving first laser for diabetes 100 73 87 105 82 56 % of all patients with diabetes receiving first laser* 1.0 0.7 0.7 0.8 0.6 0.4 Laser �6 months after screening 22 9 11 52 55 35 Laser within 6 months of screening as a percentage of patients screened (%) 0.7 0.3 0.2 0.5 0.5 0.3 Data are n or %. *P � 0.01, †P � 0.05 over the study period. Table 2—Number of patients with type 2 diabetes, number of type 2 diabetic patients undergoing digital retinal photographic screening, and number of patients undergoing first or any laser treatments in the years 2001–2006 correlated with the type 2 diabetic population mean risk factors and hypoglycemic treatment 2001 2002 2003 2004 2005 2006 Type 2 diabetic patients* 8,593 9,935 10,594 12,112 13,352 13,660 Patients undergoing digital retinal photographic screening* 4,979 6,339 6,706 8,933 10,676 10,619 Patients with referable retinopathy from photography† 295 342 409 495 476 441 Patients with referable retinopathy as a percentage of all patients screened (%)† 5.9 5.4 6.1 5.5 4.5 4.2 Laser All laser Macular† 180 168 163 174 147 103 Macular as % of all patients* 2.11 1.69 1.54 1.44 1.1 0.75 Panretinal 58 79 95 86 70 51 Panretinal as % of all patients 0.67 0.8 0.9 0.71 0.52 0.37 First laser Macular 77 45 49 69 65 38 Macular as % of all patients 0.9 0.45 0.46 0.57 0.49 0.28 Panretinal 6 13 13 16 9 15 Panretinal as % of all patients 0.07 0.13 0.12 0.13 0.07 0.11 Risk factors Mean A1C (%)* 7.9 7.6 7.4 7.4 7.5 7.4 Mean systolic blood pressure (mmHg)* 142 141 141 141 138 137 Mean diastolic blood pressure (mmHg)* 79 78 77 76 75 75 Mean age (years) 66.5 66.8 66.9 66.3 66.4 66.6 Mean duration of diabetes (years)* 7.7 7.5 7.5 7.3 7.3 7.4 Mean BMI (kg/m2)* 30.0 30.1 30.3 30.5 30.7 30.9 Mean total cholesterol (mmol/l)* 5.0 4.9 4.8 4.6 4.4 4.3 Treatment Insulin only (%)* 16.0 15.0 15.0 13.6 12.3 11.8 Insulin and oral hypoglycemics (%)† 0.7 1.1 0.9 1.5 3.3 4.4 Oral hypoglycemics (%) 54 55.8 53.7 53.8 50.6 52.0 Diet only (%)† 26 26.1 27.8 28.3 30.7 29 Not known (%) 3.3 2.0 3.3 2.8 2.9 2.8 Data are n, %, or mean. *P � 0.01, †P � 0.05 over the study period. Diabetic retinopathy 1128 DIABETES CARE, VOLUME 31, NUMBER 6, JUNE 2008 to ophthalmology, only 184 (9.8%) proceeded to laser intervention within the following 6 months (Fig. 1, Table 1). Between 2001 and 2006, the number of patients with type 2 diabetes rose from 8,936 to 13,660 (53% increase, Table 2). The most frequently performed treatment was macular laser in type 2 diabetic pa- tients. A total of 180 type 2 diabetic pa- tients (2.1% of the type 2 diabetic population) received macular treatment in 2001 and 103 (0.75% of the type 2 diabetic population) in 2006, a 43% de- crease (P � 0.03). Type 2 panretinal treat- ments peaked in 2004 with 95 patients receiving treatment, falling back to 51 pa- tients in 2006 (Fig. 2A) with no statisti- cally significant trend over the 6-year period as a whole. The type 1 diabetic population grew from 1,158 to 1,547 (34% increase, Table 3) over the same period. Macular treat- ments in type 1 diabetic patients similarly peaked in 2004 at 42 patients, falling to 12 in 2006 (Fig. 2B). Type 1 diabetic pa- tients undergoing panretinal treatment fell from 44 in 2001 to 29 in 2006. This was a significant reduction when viewed as a percentage of the type 1 diabetic pop- ulation (P � 0.01, Table 3). In type 1 diabetic patients, an in- crease in systolic blood pressure was ob- served during the study period (P � 0.01), whereas for type 2 diabetic patients, mean systolic blood pressure fell by 5 mmHg (P � 0.01) and diastolic blood pressure fell by 4 mmHg (P � 0.01). Mean A1C fell from 9.1% to 8.8% in type 1 diabetic patients (P � 0.01) and from 7.9% to 7.4% in type 2 diabetic patients (P � 0.01). Mean duration of diabetes decreased from 7.7 to 7.4 years in type 2 diabetic patients (P � 0.01). Mean BMI rose for both type 1 and type 2 diabetic populations and mean cholesterol decreased in both groups (Tables 2 and 3). The percentage of type 2 diabetic pa- tients whose only treatment was dietary advice increased from 26% in 2001 to 29% in 2006 (P � 0.01). There was no significant change in the proportion of type 2 diabetic patients using insulin (16.7% in 2001, 16.2% in 2006, P � 0.45). CONCLUSIONS — In the Tayside population, the absolute number of pa- tients with type 2 diabetes requiring laser treatment for maculopathy fell by 43% between 2001 and 2006. When taken as a proportion of all patients with type 2 di- abetes, this represented a threefold de- crease in those requiring treatment. The number of type 2 diabetic patients requir- ing panretinal photocoagulation and the number of type 1 diabetic patients requir- ing either macular or panretinal laser de- creased, but not enough to achieve statistical significance. Over the same pe- riod, the prevalence of diabetes and dia- betic retinopathy screening effort have both increased. Why was there no con- comitant increase in individuals with ret- inopathy or maculopathy severe enough to require laser treatment? One potential explanation would be a change in the criteria for laser treatment. During the period of the study, there have been no changes in national or local guidelines for the use of laser in diabetic eye disease and no local changes in per- sonnel or practice (19,20). No patients received intravitreal treatment over this period, and indications for surgical prac- tice were unaltered. We failed to identify any patients with disease severity (e.g., persistent vitreous hemorrhage, trac- tional retinal detachment) sufficient to re- quire immediate surgery without first attempting argon laser treatment. How- ever, it is difficult to exclude an unan- nounced change in practice in the application of macular photocoagulation in patients with “good” visual acuity. Pop- ulations in which screening has been es- tablished report a lower incidence and Figure 2—Trends in laser treatment during 2001–2006 for patients with type 2 diabetes (A) and patients with type 1 diabetes (B). E, All patients treated with macular laser; �, all patients receiving first macular laser treatment; F, all patients treated with panretinal laser; f, all patients receiving first panretinal laser. Vallance and Associates DIABETES CARE, VOLUME 31, NUMBER 6, JUNE 2008 1129 prevalence of diabetic visual loss (3,21), but it is difficult to separate the beneficial effect of screening from the effect of better general diabetic disease management, since the two factors frequently coexist. After 2003, digital retinal photography became almost the sole means of screen- ing in Tayside for patients with diabetes. There was a small prevalence screen effect with laser activity peaking in 2004 before falling over the final 2 years of the study. This effect was particularly marked for type 1 maculopathy and it is possible that decreases in 2005 and 2006 could be a result of earlier identification of pathology under the new annual screening system. In contrast, in the type 2 diabetic popula- tion, only a small peak in incident macu- lopathy treatment was seen in 2004, and comprehensive digital retinal photogra- phy screening had little impact on the overall trend of decreasing laser treat- ment. This may be due to relatively ade- quate screening in Tayside before comprehensive digital photography. In areas where historically there have been fewer resources, the impact of the na- tional screening program has been greater, with a more sizeable initial surge of patients with previously unrecognized sight-threatening retinopathy requiring laser treatment (22,23). A reduction in the mean disease du- ration and the proportion of type 2 dia- betic patients treated with insulin and/or oral hypoglycemic agents suggests that patients are being diagnosed with diabe- tes earlier, reducing the period of subclin- ical dysglycemia. This will increase the prevalence of individuals with clinical type 2 diabetes and might be predicted to translate into a drop in the proportion of patients requiring laser treatment. How- ever, we observed a reduction in the ab- solute numbers of type 2 diabetic patients requiring treatment for maculopathy and not simply a drop in the proportion, indi- cating this is an inadequate sole explana- tion for the trends observed. Another consequence of earlier iden- tification of type 2 diabetic patients is the possibility that patients are receiving treatment early enough in the disease pro- cess to avoid the development of sight- threatening maculopathy. Mean diastolic blood pressure in our type 2 diabetic pop- ulation decreased by 4 mmHg over 6 years to a final mean population blood pressure of 137/75 mmHg. The U.K. Pro- spective Diabetes Study Group (24) com- pared tight control of blood pressure (mean 144/82 mmHg) with less tight con- trol (mean 154/87 mmHg) in type 2 dia- betic patients and showed a 34% risk reduction for progression of retinopathy by two or more steps over 7.5 years. Fur- thermore, there was a 47% risk reduction for loss of three or more lines of Early Treatment of Diabetic Retinopathy Study visual acuity and a 35% reduction in in- dividuals undergoing laser treatment over this period. Since diabetic maculopathy is the main cause of visual impairment in type 2 diabetes, this reduction in visual loss suggests tight blood pressure control reduces the risk of maculopathy. In our study, the mean diastolic blood pressure achieved for the entire population is 7 mmHg lower than the U.K. Prospective Diabetes Study tight control group. A statistically significant fall in A1C was also observed. The 2006 population mean of 7.4% is comparable with the tight control group of newly diagnosed type 2 Table 3—Number of patients with type 1 diabetes, number of type 1 diabetic patients undergoing digital retinal photographic screening, and number of patients undergoing first or any laser treatments in the years 2001–2006 correlated with type 1 diabetes population mean risk factors 2001 2002 2003 2004 2005 2006 Number of patients* 1,158 1,281 1,338 1,470 1,548 1,547 Patients undergoing digital retinal photographic screening* 676 817 847 1,080 1,238 1,128 Patients with referable retinopathy from photography 92 97 154 173 218 145 Patients with referable retinopathy as a percentage of all patients screened (%) 13.6 11.9 18.2 16.0 17.6 12.9 Laser Any laser Macular 30 22 21 42 17 12 Macular as % of all patients 2.59 1.72 1.57 2.86 1.1 0.78 Panretinal 44 38 42 43 37 29 Panretinal as % of all patients* 3.80 2.97 3.14 2.93 2.39 1.87 First laser Macular 11 7 10 13 5 4 Macular as % of all patients 0.95 0.55 0.75 0.88 0.32 0.26 Panretinal 11 9 15 15 9 8 Panretinal as % of all patients 0.95 0.55 0.75 0.88 0.32 0.26 Risk factors Mean A1C (%)* 9.1 8.9 8.8 8.9 8.9 8.8 Mean systolic blood pressure (mmHg)* 129 129 132 132 132 132 Mean diastolic blood pressure (mmHg) 76 75 75 75 75 75 Mean age (years)* 35.9 36.4 36.7 36.7 37.9 38.2 Mean duration of diabetes (years) 17.5 17.5 17.4 17.1 17.4 17.7 Mean BMI (kg/m2)* 25.4 25.6 25.7 25.0 26.6 26.6 Mean total cholesterol (mmol/l)* 5.1 5.0 4.9 4.9 4.6 4.6 Data are n, %, or mean. *P � 0.01 over the study period. Diabetic retinopathy 1130 DIABETES CARE, VOLUME 31, NUMBER 6, JUNE 2008 diabetic patients reported in the U.K. Pro- spective Diabetes Study 33 (25). This group had a 29% lower risk of retinal photocoagulation over 10 years from di- agnosis when compared with individuals receiving “conventional” treatment (mean A1C 7.9%). Mean total cholesterol also decreased significantly over the study pe- riod, and although plasma lipids have not been conclusively proven to influence the course of diabetic retinopathy or macu- lopathy, the Fenofibrate Intervention and Event Lowering in Diabetes (FIELD) study demonstrated a reduction in laser treatment in those treated with fenofi- brate (26). In conclusion, the incidence of macu- lopathy requiring laser treatment in type 2 diabetic patients in Tayside has de- creased over the last 6 years despite in- creased prevalence of type 2 diabetes and increased screening effort. The national screening program contributed a greater number of patients receiving first laser, but did not alter the overall trend to less laser treatment for this group. We suggest that earlier diagnosis and improved man- agement of the risk factors for diabetic maculopathy is reducing the incidence of maculopathy severe enough to require la- ser treatment in type 2 diabetes. Acknowledgments — This study was funded by the Wellcome Trust, Speed Pollock Memo- rial Trust. References 1. Knudsen L, Lervang H-H, Lundbye- Christiansen S, Gorst-Rasmussen A: The North Jutland Count Diabetic Retinopa- thy Study: population and characteristics. Br J Ophthalmol 90:1404 –1409, 2006 2. LeClaire T, Palta M, Zhang H, Allen C, Lien R, D’Alessio D: Lower-than-expected prevalence and severity of retinopathy in an incident cohort followed during the first 4 –14 years of type 1 diabetes: The Wisconsin Diabetes Registry Study. Am J Epidemiol 164:143–150, 2006 3. Olafsdottir E, Andersson D, Stefansson E: Visual acuity in a population with regular screening for type 2 diabetes and eye dis- ease. Acta Ophth Scand 85:40 – 45, 2007 4. Younis N, Broadbent DM, Harding SP, Vora JP: Incidence of sight-threatening retinopathy in type 1 diabetes in a system- atic screening programme. Diabet Med 20: 758 –765, 2003 5. Younis N, Broadbent DM, Harding SP, Vora JP: Incidence of sight-threatening retinopathy in type 2 diabetes in the Liv- erpool Diabetic Eye Study: a cohort study. Lancet 361:195–200, 2003 6. Trautner C, Haarert B, Giani G, Berger M: Incidence of blindness in southern Ger- many between 1990 and 1998. Diabetolo- gia 44:147–150, 2001 7. Backlund LB, Algvere PV, Rosenqvist U: New blindness in diabetes reduced by more than one-third in Stockholm County. Diabet Med 14:732–740, 1997 8. Porta M, Tomalino MG, Santoro F, Ghigo LD, Cairo M, Aimone M, Pietragalla GB, Passera P, Montanaro M, Molinatti GM: Diabetic retinopathy as a cause for blind- ness in the province of Turin, north-west Italy in 1967–1991. Diabet Med 12:355– 361, 1995 9. Barry RJ, Murray PI: Unregistered visual impairment: is registration a failing sys- tem? Br J Ophthalmol 98:995–998, 2005 10. Rhatigan MC, Leese GP, Ellis J, Ellingford A, Morris AD, Newton RW, Roxburgh ST: Blindness in patients with diabetes who have been screened for eye disease. Eye 13:166 –169, 1999 11. General Register Office for Scotland: Mid- 2006 Population Estimates Scotland. July 2007 at http//www.gro-scotland.gov.uk 12. Leese GP, Morris AD, Swaminathan K, Petrie JR, Sinharay R, Ellingford A, Taylor A, Jung RT, Newton RW, Ellis JD: Imple- mentation of national diabetes retinal screening programme is associated with a lower proportion of patients referred to ophthalmology. Diabet Med 22:1112– 1115, 2005 13. Scottish Executive: Diabetic Retinopathy Screening Services in Scotland: Recom- mendations for Implementation (DRSIG). Edinburgh: Scottish Executive 2003. Available from http://www.diabetesin scotland.org/diabetes 14. Diabetic Retinopathy screening services in Scotland: recommendations for imple- mentation. Report by the diabetic screen- ing implementation group, June 2006 (annexe E). At www.scotland.gov.uk/ resource/doc/46930/0013818.pdf 15. NHS Quality Improvement Scotland. Clinical Standards in Diabetic retinopathy screening. March 2004. Available at http://www.nhshealthequality.org 16. Morris AD, Boyle DI, MacAlpine R, Emslie-Smith A, Jung RT, Newton RW, MacDonald TM: The Diabetes Research and Audit in Tayside Scotland (DARTS) study: electronic record linkage to create a district diabetes register. BMJ 315:524 – 528, 1997 17. Scottish Diabetes Core Dataset NHS Scotland 2003 at www.scotland.gov.uk/ Resource/Doc/47021/0013911.pdf 18. Morris AD, Boyle DIR, McMahon AD, Greene SA, MacDonald TM, Newton RW: Adherence to insulin treatment, glycae- mic control, and ketoacidosis in insulin- dependent diabetes mellitus. Lancet 350: 1505–1510, 1997 19. Guidelines for diabetic retinopathy 1997. Royal college of ophthalmologists, Lon- don 1997 20. Guidelines for diabetic retinopathy 2005. Royal college of ophthalmologists, Lon- don 2005. Available at www.rcophth. ac.uk/docs/scientific/publication 21. Stefansson E, Bek T, Porta M, Larsen N, Kristinson JK, Agardth E: Screening and prevention of diabetic blindness. Acta Ophthalmol Scand 78:374 –385, 2000 22. Scanlon PH, Carter S, Foy C, Ratiram D, Harney B: An evaluation of the change in activity and workload arising from dia- betic ophthalmology referrals following the introduction of a community based digital retinal photographic screening programme. Br J Ophthalmol 89:971–975, 2005 23. Leese G, Morris AD, Petrie J, Jung RT, Ellis JD: Diabetic retinopathy screening programmes and reducing ophthalmolo- gists’ workload-authors reply. Diabet Med 23:449 – 450, 2006 24. Tight blood pressure control and risk of microvascular and microvascular compli- cations in type 2 diabetes: UKPDS 38: UKPDS Group. BMJ 317:703–713, 1998 25. U.K. Prospective Diabetes Study (UK- PDS) 33: Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 di- abetes (UKPDS 33): UKPDS Group. Lan- cet 352:837– 853, 1998 26. Keech AC, Mitchell P, Summanen PA, O’Day J, Davis TM, Moffitt MS, Taskinen MR, Simes RJ, Tse D, Williamson E, Mer- rifield A, Laatikainen LT, d’Emden MN, Crimet DC, O’Connell RL, Colman PG: Effect of fenofibrate on the need for laser treatment for diabetic retinopathy (FIELD study): a randomised controlled trial. Lancet 370:1687–1697, 2007 Vallance and Associates DIABETES CARE, VOLUME 31, NUMBER 6, JUNE 2008 1131