Aims/hypothesis. The short-term efficacy, safety and tolerability of rosiglitazone were compared with placebo in patients with Type II (non-insulin-dependent) diabetes mellitus in a dose-ranging study.¶Methods. After a 2-week placebo run-in phase, 303 patients were randomly assigned to 8 weeks of treatment with twice-daily placebo or 2, 4 or 6 mg of rosiglitazone.¶Results. All rosiglitazone doses significantly reduced fasting plasma glucose compared with baseline. All rosiglitazone treatment groups showed significantly reduced peak postprandial glucose concentrations compared with baseline (p < 0.001) and with placebo (p < 0.0001) and reduced postprandial glucose excursion, without an increase in the area under the postprandial insulin concentration-time curve. Rosiglitazone at 4 and 6 mg twice daily prevented the increase in HbA_{1 c} observed in the placebo group. C peptide and serum insulin concentrations were significantly reduced from baseline in all rosiglitazone treatment groups. In all rosiglitazone treatment groups, non-esterified fatty acids decreased significantly (p < 0.0001) and triglycerides did not change. Although total LDL and HDL cholesterol increased significantly in the rosiglitazone treatment groups, total cholesterol/HDL ratios did not change significantly. The proportion of patients with one or more adverse event was similar in all four treatment groups. No patient showed evidence of hepatotoxicity.¶Conclusion/interpretation. Rosiglitazone given twice daily significantly reduced fasting and postprandial glucose concentrations, C peptide, insulin and non-esterified fatty acids in Type II diabetic patients. The glucose-lowering effect of the 4-mg twice-daily dose of rosiglitazone was similar to that of 6-mg twice daily, suggesting that 4 mg twice daily should be the maximum clinical dose. [Diabetologia (2000) 43: 278–284]