Epidemiologic data on the incidence of gross proteinuria in people with diabetes are important in medical counseling, in projecting estimates of needs and costs of health care, and for developing approaches to prevent renal complications. We performed a population-based incidence study in southern Wisconsin of insulin-taking diabetic persons diagnosed before 30 years of age. The presence of gross proteinuria (greater than or equal to 0.30 g/L) was determined by means of a reagent strip. The incidence of proteinuria in a 4-year interval was 14.4% (95% confidence interval, 11.7% to 17.0%). The relative risk of developing proteinuria after 4 years for those with glycosylated hemoglobin levels in the highest quartile compared with those in the lowest quartile was 3.0 (95% confidence interval, 1.6 to 5.3). The incidence of proteinuria was also associated with higher diastolic blood pressure, being male, taking more insulin, and having more severe retinopathy at the baseline examination. The relationship between glycosylated hemoglobin level and the incidence of proteinuria was significant even after controlling for these other risk variables. These data suggest that hyperglycemia, as measured by glycosylated hemoglobin level, is a significant risk factor for the development of gross proteinuria.