OBJECTIVES: A meta-analysis of Pioglitazone and Rosiglitazone in patients with type 2 diabetes was conducted to evaluate the effects of each drug on glycemic control, lipids, blood pressure, and weight. In addition, this study was aimed to identify if there is an ethnic difference in the effects between Asian and Caucasian patients by subgroup analysis. METHODS: Among the randomized controlled trials of Pioglitazone or Rosiglitazone which had been published before February 2008, 63 randomized controlled trials that were eligible for the inclusion/exclusion criteria were collected. For quantitative meta-analysis, the weighted pooled effect sizes and their 95% confidence intervals were calculated. RESULTS: Glucose lowering effect was higher in Rosiglitazone, whereas  Pioglitazone produced a more favorable lipid profile. Both thiazolidinediones demonstrated similar increases in body weight. The effect on blood pressure was slightly higher in Pioglitazone. Regarding the effects of Pioglitazone, the pooled effect sizes of Pioglitazone 15mg were very similar for two different ethnic groups in all parameters. When the effects of Pioglitazone in Caucasian patients group at dose of 30mg and Asian patients group at dose of 15mg were compared, most of the overall effects sizes were higher in Caucasian patients group than that of Asian patients group. In case of Rosiglitazone, the pooled effect sizes of Asian patients group administered 4mg were higher than Caucasian patients group administered the same dose. Furthermore, in comparing the effect sizes of Asian patients group administered 4mg and Caucasian patients group administered 8mg, the effect sizes were rather similar than they administered the same dose. CONCLUSIONS: The effects of Pioglitazone on the cardiovascular risk factors are almost the same in different ethnic groups, whereas those of Rosiglitazone varies with ethnicity. In this regards, well-designed head-to-head comparative trials as well as long-term cardiovascular outcome studies should be conducted in order to accurately determine the various effects of the two thiazolidinediones on different ethnic groups.