This study examines the effect of rosiglitazone on urinary albumin excretion (UAE) in patients with type II diabetes. Urinary albumin: creatinine ratio (ACR) was measured in a 52-week, open-label, cardiac safety study comparing rosiglitazone and glyburide. Patients were randomised to treatment with rosiglitazone 4 mg b.i.d. or glyburide. ACR was measured at baseline and after 28 and 52 weeks of treatment. Statistically significant reductions from baseline in ACR were observed in both treatment groups at week 28. By week 52, only the rosiglitazone group showed a significant reduction from baseline. Similar results were observed for the overall study population and for the subset of patients with baseline microalbuminuria. For patients with microalbuminuria at baseline, reductions in ACR did not correlate strongly with reductions in glycosylated haemoglobin, or fasting plasma glucose, but showed strong correlation with changes in mean 24-h systolic and diastolic blood pressure for rosiglitazone-treated patients (deltaACR vs deltamean 24-h systolic blood pressure, r=0.875; deltaACR vs deltamean 24-h diastolic blood pressure, r=0.755; P < 0.05 for both). No such correlation was observed for glyburide-treated patients. In conclusion, rosiglitazone treatment was associated with a decrease in urinary albumin excretion. These findings suggest a potential beneficial effect of rosiglitazone in the treatment or prevention of renal and vascular complications of type II diabetes.