Twenty-three RCTs were included: 10 of pioglitazone (n=3,305) and 13 of rosiglitazone (n=5,245).
Blood glucose level.
Pioglitazone, when used as monotherapy, significantly reduced HbA1c levels by -0.99% (95% CI: -1.32, -0.66) at a dose of 30 mg/day and by -1.21% (95% CI: -1.79, -0.62) at 45 mg/day, compared with placebo. When used in combination with other antidiabetic agents, pioglitazone (30 and 45 mg/day) significantly reduced HbA1c levels by -1.16% (95% CI: -1.41, -0.90) and -1.56% (95% CI: -1.96, -1.16), respectively, in comparison with placebo.
Rosiglitazone, when used as monotherapy, significantly reduced HbA1c levels by -0.90% (95% CI: -1.42, -0.38) at a dose of 4 mg/day and by -1.50% (95% CI: -1.75, -1.24) at 8 mg/day. When used in combination with other antidiabetic agents, rosiglitazone (4 and 8 mg/day) significantly reduced HbA1c levels by -1.05% (95% CI: -1.19, -0.90) and -1.26% (95% CI: -1.48, -1.04), respectively, in comparison with placebo.
FBG.
Pioglitazone, when used as monotherapy, significantly reduced FBG concentrations at doses of 30 and 45 mg/day, compared with placebo. When used in combination with other antidiabetic agents, pioglitazone (30 and 45 mg/day) significantly reduced FBG concentrations in comparison with placebo.
Similarly, rosiglitazone, when used as monotherapy, significantly reduced FBG concentrations at a dose of 4 mg/day. When used in combination with other antidiabetic agents, rosiglitazone (4 and 8 mg/day) significantly reduced FBG concentrations in comparison with placebo.
Lipids.
Rosiglitazone, when used as monotherapy or in combination with other antihyperglycaemic agents, significantly increased LDL-C level (15 mg/dL, 0.39 mmol/L). No statistically significant difference was shown for the effect of pioglitazone on LDL-C level (-0.4 mg/dL, -0.01 mmol/L) compared with placebo.
Treatment with rosiglitazone had no statistically significant effect on the fasting plasma triglyceride concentration. Pioglitazone was shown to significantly decrease fasting plasma triglyceride level (-40 mg/dL, 0.45 mmol/L).
Both rosiglitazone and pioglitazone were found to significantly increase the HDL-C concentration, by 2.7 mg/dL (0.07 mmol/L) and 4.55 mg/dL (0.12 mmol/L), respectively.
A significant increase in total cholesterol was shown with rosiglitazone when compared with placebo (21.3 mg/dL, 0.55 mmol/L). Treatment with pioglitazone had no statistically significant effect on total cholesterol level (-0.1 mg/dL, -0.003 mmol/L).
Blood-pressure.
No significant differences were shown between rosiglitazone and placebo in changes in systolic or diastolic blood-pressure.
Weight.
The average weight gain within 6 months of treatment initiation with thiazolidinediones was 2.7 kg (95% CI: 1.8, 3.7). Significant heterogeneity was found, but sensitivity analyses indicated that drug grouping did not explain this heterogeneity. An average weight gain of 0.73 kg (95% CI: 0.23, 1.23) without statistical heterogeneity was demonstrated in the Japanese trials, compared with an average weight gain of 3.3 kg (95% CI: 2.5, 4.2) with significant statistical heterogeneity in the non-Japanese trials.