We first compared glomerular charge selectivity index in two matched groups of Type 1 (insulin-dependent) diabetic patients with micro- and normoalbuminuria respectively, and secondly, investigated prospectively in a randomized clinical trial, the influence of improved metabolic control on selectivity index in diabetic patients with microalbuminuria. In Study 1, 27 patients with microalbuminuria (albumin excretion > or = 15 micrograms/min in at least two out of three overnight urine samples) were matched (age, diabetes duration, mean 1-year HbA1c, gender) with normoalbuminuria patients (n = 24), and in Study 2, 23 microalbuminuric patients were randomly allocated to either intensive (continuous subcutaneous insulin infusion) or conventional treatment. Glomerular charge selectivity index was measured as IgG/IgG4 selectivity index, i.e. total IgG/IgG4 clearance ratio in timed overnight urine samples. The microalbuminuric patients had a significantly reduced selectivity index compared to the normoalbuminuric patients: 1.20 (0.92-1.40) vs 1.68 (1.22-2.21), median and 95% confidence interval (p < 0.01). In Study 2, the HbA1c improved in the intensive-treatment group compared to the conventional-treatment group: at 2, 6 and 12 months the difference in mean percentage HbA1c between the groups was 1.1, 1.2 and 1.4, respectively (p < 0.01). A sharp 50% increment in IgG/IgG4 selectivity index was seen in the intensive-treatment group during the first 6 months (p < 0.05 compared to the conventional group). We conclude that adolescents and young adults in an early stage of diabetic nephropathy have reduced glomerular charge selectivity, which may be improved by reducing the mean blood glucose level.