BACKGROUND: Diabetic retinopathy (DR) is a leading cause of blindness in the working-age population in UK. Glycaemia control is determinant in the natural history of eye disorders but even under well-controlled diabetes there is increased risk of DR.
OBJECTIVES: To estimate the DR incidence rate (IR) and identify risk factors for incident DR in type II diabetes in the context of current management of diabetes in UK primary care.
METHODS: We conducted a case-control analysis nested in a cohort of newly diagnosed type II diabetes patients aged 1–84 years identified in The Health Improvement Network (THIN) database between 2000 and 2007. We followed patients until DR code (N = 7,735), age 85 years, death, or 31/12/2008. DR diagnosis was confirmed by general practitioners in 72% of instances. Cases were all patients with DR and controls were a random sample of study cohort (N = 9,395). Adjusted odds ratios (OR; 95% CI) were estimated for life-style and medical factors, and hypoglycemic drugs.
RESULTS: The DR IR was 23.8 per 1,000 person-years (95% CI: 23.17–24.43). The risk of DR increased with diabetes duration, although 21% of cases were identified at first diabetes diagnosis. DR increased slightly with systolic BP greater than or equal to 140 mmHg (OR 1.14; 1.07–1.21), high alcohol consumption (OR 1.34; 1.11–1.61), glycated haemoglobin (HbA1c) (OR 9–10% 1.14; 1.00–1.31; OR 10–11% 1.25; 1.07–1.45;OR greater than or equal to 11% 1.21; 1.07–1.37) and retinal venous occlusion (OR 2.47; 1.67–3.66). History of glaucoma was associated with an OR of 0.71 (0.60–0.84). Patients with HDL greater than or equal to 1.55 mmol/L (OR 0.88; 0.80–0.98), triglycerides greater than or equal to 1.7 mmol/L (OR 0.89; 0.83–0.96) or current smokers (OR 0.89; 0.81–0.97) presented a slightly reduced risk. Proteinuria, BMI, cholesterol or LDL were not associated with diagnosis of DR. Use of metformin (OR 1.21; 1.13– 1.30), sulphonylureas (OR 1.25; 1.15–1.36) and insulin (OR 1.50; 1.30–1.74) showed an increased risk of DR.
CONCLUSIONS: A proportion of DR was found at first diabetes diagnosis suggesting late diabetes diagnosis. Multiple factors appeared to be associated with DR, including high HbA1c, systolic BP, alcohol consumption and use of hypoglycemic drugs. High HDL appeared to confer a reduced risk.
