Efficacy and tolerability of rosiglitazone and pioglitazone in drug-naïve Japanese patients with type 2 diabetes mellitus: a double-blind, 28 weeks’ treatment, comparative study

OBJECTIVE: A 28-week, randomized, placebo-controlled study was performed to evaluate efficacy and tolerability of rosiglitazone in Japanese type 2 diabetes patients. RESEARCH AND DESIGN METHODS: 373 patients were randomized to rosiglitazone (4-8 mg/day), pioglitazone (15-45 mg/day) or placebo. Agents were titrated to maximum doses at fixed time points in a pre-defined manner. Primary endpoints were superiority of each active treatment compared to placebo in HbA(1c) at week 16, and non-inferiority between active agents in HbA(1c) at week 28, based on a -0.45% margin. RESULTS: At week 16, improvements versus placebo were observed with rosiglitazone 4 mg/day (-0.96%, p < 0.001) and pioglitazone 30 mg/day (-1.26%, p < 0.001). At week 28, rosiglitazone and pioglitazone were associated with significant changes from baseline of -0.94% and -1.35%, respectively and rosiglitazone produced statistically and clinically significant improvement versus placebo (-1.29%, CI: -1.62, -0.97). Pioglitazone also showed significant improvement versus placebo (-1.64%, CI: -1.96, -1.31). Non-inferiority of rosiglitazone (4-8 mg/day) to pioglitazone (30-45 mg/day) was not demonstrated (treatment-difference: -0.41%, 95% CI: -0.64, -0.18). More patients treated with pioglitazone were withdrawn from the study by adverse events compared with rosiglitazone (14 vs. 4, p = 0.015). Pioglitazone was associated with higher incidences of adverse events relating to edema and weight gain compared with rosiglitazone (edema: 25.2 vs. 11.3%, weight gain: 9.4 vs. 4.4%). There were no reports of ischemic heart disease or congestive heart failure in any treatment group. CONCLUSION: Although non-inferiority to pioglitazone up to 45 mg in efficacy was not shown, rosiglitazone was confirmed to have clinically meaningful efficacy over placebo and fewer fluid-related events than pioglitazone. The study is registered on ClinicalTrials.gov as protocol NCT00297063.