How many steps of progression of diabetic retinopathy are meaningful? The Wisconsin epidemiologic study of diabetic retinopathy.

Abstract

To determine whether a 1-step or more or 2-step or more progression on the Early Treatment Diabetic Retinopathy Study retinopathy severity scale over a 4-year period is meaningful in predicting the subsequent incidence of proliferative diabetic retinopathy (PDR) and clinically significant macular edema (CSME) over the following 6 years. Population-based study of diabetic persons with 10 years of follow-up. Eleven-county area in southern Wisconsin. There were 1025 persons with diabetes who had fundus photographs at baseline and at 4- and 10-year follow-up examinations. Incidence of PDR or CSME between the 4- and 10-year follow-up examinations as determined by masked grading of color stereoscopic fundus photographs of 7 standard fields. In a univariate analysis, those with 1 or more steps of progression (n = 551) over the first 4 years of the study were significantly (P<.0001) more likely to develop PDR over the next 6 years than those with no progression (n = 474) (26% vs 4%) (relative risk, 5.85; 95% confidence interval, 4.05-8.47). Similarly, those with 2 or more (n = 364) (33%) or 3 or more (n = 231) (41%) steps of progression over the first 4 years of the study were significantly (P<.0001) more likely to develop PDR over the next 6 years than those with lesser progression (n = 661 [7%] and n = 794 [9%], respectively) (relative risk, 5.10; 95% confidence interval, 3.83-6.80; and relative risk, 4.61; 95% confidence interval, 3.57-5.99, respectively). Similar associations were apparent at every level of retinopathy, duration of diabetes, and glycosylated hemoglobin, and by type of diabetes at baseline. There were also associations between retinopathy progression and incidence of CSME. It seems that 1 or more or 2 or more steps of progression of retinopathy over a 4-year period strongly predict the development of PDR over the next 6 years. Therefore, using these end points of progression would result in the need for fewer subjects or shorter follow-up in some clinical trials.