We determined HBA1 (microcolumn method) and glycosylated albumin (fructosamine) in 23 healthy subjects, 35 patients with renal insufficiency without diabetes and 14 patients with diabetes mellitus and renal insufficiency. All patients with renal insufficiency required dialysis. All diabetics were of type I and had been compensated on insulin. The HBA1 in the nondiabetic patients with renal insufficiency (9.4 +/- 1.4%) was significantly raised compared to that in the control group with healthy metabolism (7.3 +/- 0.6%). Irrespective of the quality of compensation, the diabetic patients had HBA1 values of more than 11% of average. On the other hand, the concentrations of glycosylated albumin in healthy nondiabetic patients and in diabetic patients with renal insufficiency did not differ (1.3 +/- 0.5 as compared to 1.1 +/- 0.4 mmol/l) and were all in the normal range. Well-adjusted diabetics with renal insufficiency had a fructosamine concentration of 1.9 +/- 0.7 mmol/l (theoretical value for a good compensation 2.0 to 2.8). We conclude that determination of HBA1 in pronounced renal insufficiency does not provide reliable values because carbamylated hemoglobin is also registered and determination of fructosamine (which only indicates the metabolic situation in the last three weeks, however) is to be preferred in this situation.