Advanced glycation end (AGE)-products, a complex and heterogeneous group of compounds, have been implicated in diabetes-related long-term complications. Up to the present, only few data exist about serum levels of the AGE-proteins N- epsilon -carboxymethyllysine (CML) and pentosidine in selection-free populations of patients with type 1 and insulin-treated type 2 diabetes mellitus. In the present 10-year, population-based trial of patients with insulin-treated diabetes mellitus, serum CML and pentosidine levels were examined in correlation to the patients' quality of diabetes control and the prevalence of diabetes-related long-term complications. Jena's St. Vincent Trial (JEVIN) was started in 1989/1990. At this time, a centralised diabetes care system existed. After the baseline examination of 190 patients (83% of the target population) with insulin-treated diabetes mellitus, follow-up examinations were performed in 1994/1995 and 1999/2000. In 1994/1995, the CML concentration in patients with type 1/type 2 diabetes mellitus was 1096.47+/-405.50/1136.43+/-405.24 ng/ml. In 1999/2000, it was significantly lower (727.49+/-342.91 ng/ml, P=.033/743.76+/-312.47 ng/ml, P<.0001). The same tendency showed the AGE-protein pentosidine (type 1: 1994/1995 203.18+/-118.88 vs. 1999/2000 156.59+/-104.84 pmol/ml [P=.029], type 2: 1994/1995 189.72+/-67.66 vs. 1999/2000 151.54+/-127.73 pmol/ml [P=.020]). Parallel to the decrease in the mean concentration of the AGE-products CML and pentosidine mean HbA1c improved and the prevalence of diabetic long-term complications (retino-, neuro-, and nephropathy) remained comparable 1999/2000-1989/1990. Comparing the data of 1999/2000 with those from 1994/1995, there was not only a substantial improvement in patients' quality of diabetes control but also a decrease in the concentration of AGE-products. In patients with diabetes mellitus, the AGE-products seem to be mainly influenced by the quality of diabetes control. However, the most important parameter reflecting the risk for development and progression of diabetes-related long-term complications seems not to be the AGE-products, but patients' HbA1c.