The relationship between plasma C-peptide and the 6-year incidence and progression of diabetic retinopathy was examined in a population-based study in Wisconsin. Individuals with younger-onset (n = 548) and older-onset (n = 459) diabetes were included. C-peptide was measured by radioimmunoassay with Heding's M1230 antiserum. Retinopathy was determined from stereoscopic fundus photographs. Younger- and older-onset insulin-using individuals with undetectable or low plasma C-peptide (< 0.3 nmol/l) at baseline had the highest incidence and rates of progression of retinopathy, whereas older-onset individuals with C-peptides > 0.3 nmol/l had the lowest incidence and rates of progression of retinopathy. However, within each group (younger-onset using insulin, older-onset using insulin, and older-onset not using insulin), after we controlled for other characteristics associated with retinopathy, there was no relationship between higher levels of C-peptide at baseline and lower 6-year incidence or progression of retinopathy. These data suggest that glycemic control, and not C-peptide, is related to the incidence and progression of diabetic retinopathy.