Background: Proliferative diabetic retinopathy (PDR), a prevalent late-stage complication of diabetes, is associated with severe visual loss. The objectives of this report were to estimate the incidence of and risk factors for PDR using routinely collected data from a clinical information system at University Hospital Nottingham for insulin- and non-insulin-dependent (insulin-treated and non-insulin-treated) diabetes separately. We also attempted to assess the risk of blindness in these diabetes clinic attenders.
Methods: During a mean (standard deviation (SD)) follow-up period of 5.1 (2.9) (range 0.5-12.4) years, 3482 diabetic patients (1915 male and 1567 female) from three outpatient clinics at University Hospital, Nottingham were examined. The mean (SD) age of the participants was 49.3 (17.9) years with a mean duration of diabetes of 7.1 (8.7) years at registration.
Results: Among the 3482 patients who attended the clinic at least twice in the period 1979-1992, and who were free of PDR at registration, the overall incidence of PDR was 16.2 per 1000 person-years, based on 17,618 person-years of follow-up. The incidence rate of PDR was nearly three times as high among patients with non-proliferative diabetic retinopathy (NPDR) as in those without any retinopathy (42.1 vs. 15.0 per 1000 person-years). Based on a Cox's Proportional Hazards Model, significant independent predictors of PDR recorded at baseline were glycosylated haemoglobin (HbA1), systolic blood pressure, and longer duration of diabetes for patients without PDR or any retinopathy among insulin-dependent patients. Longer duration of diabetes was the only independent predictor of PDR for patients without PDR or any retinopathy in both insulin- and non-insulin-treated non-insulin-dependent patients. These clinic-based data clearly indicate the higher risk of PDR in non-insulin-dependent patients. Gender, age, BMI, creatinine, proteinuria and cigarette smoking, had no significant independent association with PDR when other covariates were considered in all groups. The risk of blindness was greater among those with PDR than those with NPDR in all three types of diabetes, but was substantial even for those without retinopathy.
Conclusion: These data are of value in identifying those diabetes clinic attenders who may be most at risk.