Table 7.
Summary of effect on results, before and after removing outlier studies from direct meta-analyses
| Outcome | Saxagliptin 5 mg QD + metformin | Vildagliptin 50 mg BID + metformin | ||||||
|---|---|---|---|---|---|---|---|---|
| + DeFronzo et al. [89] | − DeFronzo et al. [89] | + Bosi et al. [88] | − Bosi et al. [88] | |||||
| Weighted mean difference (95% CI) | I 2 (%) | Weighted mean difference (95% CI) | I 2 (%) | Weighted mean difference (95% CI) | I 2 (%) | Weighted mean difference (95% CI) | I 2 (%) | |
| HbA1c mean change from baseline | −0.585* (−0.76 to −0.41) | 60.5a | −0.503* (−0.63 to −0.37) | 0.0 | −0.606* (−0.84 to −0.37) | 81.9a | −0.479* (−0.63 to −0.33) | 51.6a |
| Saxagliptin 5 mg QD + metformin | Vildagliptin 50 mg BID + metformin | |||||||
|---|---|---|---|---|---|---|---|---|
| + DeFronzo et al. [89] | − DeFronzo et al. [89] | + Bosi et al. [88] | − Bosi et al. [88] | |||||
| Odds ratio (95% CI) | I 2 (%) | Odds ratio (95% CI) | I 2 (%) | Odds ratio (95% CI) | I 2 (%) | Odds ratio (95% CI) | I 2 (%) | |
| HbA1c <7% | 2.272* (1.69 to 3.06) | 53.7a | 2.011* (1.62 to 2.49) | 0.0 | – | – | – | – |
BID twice daily, CI confidence interval, DPP-4 dipeptidylpeptidase-4, HbA 1c glycosylated hemoglobin, QD once daily
* Statistically significant versus metformin
a I 2 may represent moderate to considerable heterogeneity