Objective: The study objective was to examine the influence of glycaemic control and ethnic variations on the incidence and progression of diabetic retinopathy (DR).Design, subjects and setting: Eight hundred and ninety-two persons with type 1 diabetes mellitus, and 1 998 persons with type 2 diabetes mellitus, who were enrolled in a private diabetes mellitus management programme in South Africa, participated in the study. Survival analyses were conducted to assess the relationship between the risk factors and the incidence of DR and referable DR, and the progression of DR.Outcome measures: Cumulative incidence of diabetic retinopathy and referable diabetic retinopathyResults: The seven-year cumulative incidence of DR and referable DR was 536 and 50 cases per 1 000 persons with type 1 diabetes mellitus without DR at baseline, and 351 and 47 cases per 1 000 persons with type 2 diabetes mellitus. The seven-year cumulative incidence of referable DR was 332 cases per 1 000 persons with type 1 diabetes mellitus with background DR at baseline, and 360 cases with type 2 diabetes mellitus, representing a seven- and eightfold increase compared to no DR at baseline. After controlling for known risk factors for DR, a high baseline haemoglobin A1c (HbA1c) and non-Caucasian ethnicity were associated with the incidence of referable DR in patients with type 1 and type 2 diabetes mellitus.Conclusion: It was revealed in the first study to report on the incidence and progression of DR in South Africa that a high baseline HbA1c, ethnicity, and the presence of background DR increased the risk of the development of referable DR.