An international, prospective, double-blind trial compared the long-term therapeutic value of glimepiride with glibenclamide in patients with Type 2 diabetes mellitus. Patients stabilised on glibenclamide were randomised to 1 mg glimepiride (524 patients) or 2.5 mg glibenclamide (520 patients). The treatment groups were comparable at baseline with respect to age (60.2 years), body mass index (26.5 kg/m2), duration of diabetes (5.0 years) and fasting blood glucose levels (163 mg/dl [9.0 mmol/l]). Doses were increased stepwise, up to 8 mg for glimepiride (once-daily) and 20 mg for glibenclamide (> 10 mg as divided dose), until metabolic control (fasting blood glucose < or = 150 mg/dl [8.3 mmol/l]), or maximum dose was achieved. After one year of treatment, patients entered a long-term follow-up study. Primary endpoints for evaluation of metabolic control, mean glycated haemoglobin and mean fasting blood glucose, were 8.4% and 174 mg/dl (9.7 mmol/l) for glimepiride and 8.3% and 168 mg/dl (9.3 mmol/l) for glibenclamide. Differences between treatment groups were not considered clinically relevant (95% confidence intervals (-0.05, 0.19%) for glycated haemoglobin and (2, 11 mg/dl) [0.1, 0.6 mmol/l] for fasting blood glucose). Statistically significant lower fasting insulin and C-peptide values were observed in glimepiride patients compared with glibenclamide (differences: insulin, -0.92 microU/ml [p = 0.04]; C-peptide, -0.14 ng/ml [p = 0.03]). Both treatment groups showed an equivalent safety profile. Adverse events were consistent with the nature of the diabetic patient population studied. Fewer hypoglycaemic reactions occurred with glimepiride than with glibenclamide (105 versus 150 episodes). The long-term follow-up (457 patients) confirmed that glimepiride (1-8 mg) once daily provides equivalent metabolic control to a higher dosage (2.5-20.0 mg) of glibenclamide. Both treatments were well tolerated.