Efficacy and Safety of Dipeptidyl Peptidase-4 Inhibitors in Type 2 Diabetes: Meta-Analysis

BACKGROUND: An up-to-date assessment of dipeptidyl peptidase-4 (DPP-4) inhibitors is needed to include newly available data. OBJECTIVE: To assess the efficacy and safety of DPP-4 inhibitors, including sitagliptin, saxagliptin, vildagliptin, and linagliptin, in type 2 diabetes. METHODS: We conducted a search of MEDLINE for randomized controlled trials (RCTs) of DPP-4 inhibitors in type 2 diabetes through November 2011, using the key terms sitagliptin, saxagliptin, vildagliptin, and linagliptin. We also searched for completed, but unpublished, trials at relevant web sites. RCTs were selected for meta-analysis if they (1) compared DPP-4 inhibitors with placebo or an antihyperglycemic agent; (2) had study duration of 12 or more weeks; (3) had 1 or more baseline and posttreatment efficacy and/or safety outcome; and (4) were published in English. RESULTS: In 62 evaluated articles, DPP-4 inhibitors lowered hemoglobin A(1c) (A1C) significantly more than placebo (weighted mean difference [WMD] -0.76%; 95% CI -0.83 to -0.68); however, heterogeneity was substantial (I(2) = 82%). Exclusion of Japanese trials (n = 7) resulted in a reduction of heterogeneity (I(2) = 59%). In the non-Japanese RCTs (n = 55), DPP-4 inhibitors were associated with a reduction in A1C (WMD -0.65%; 95% CI -0.71 to -0.60) but higher risk of hypoglycemia (odds ratio [OR] 1.30; 95% CI 1.00 to 1.68) compared to placebo. The 7 Japanese-specific RCTs showed a greater reduction in A1C (WMD -1.67%; 95% CI -1.89 to -1.44) and a nonsignificant increase in risk of hypoglycemia (OR 1.41; 95% CI 0.51 to 3.88) with DPP-4 inhibitors versus placebo. When comparing DPP-4 inhibitors to active comparators, the I(2) was still high after deleting Japanese studies. In these 17 active comparator trials, there was no significant difference in A1C reduction (WMD 0.04%; 95% CI -0.09 to 0.16) or risk of hypoglycemia (OR 0.60; 95% CI 0.22 to 1.61) for DPP-4 inhibitors compared to other antihyperglycemics. There were similar odds of any or serious adverse events with DPP-4 inhibitors compared to placebo, but a decreased risk compared to other antihyperglycemics. CONCLUSIONS: DPP-4 inhibitors were associated with a reduction in A1C with comparable safety profiles compared to placebo, but no significant difference in A1C compared to other hyperglycemics. Differences in efficacy and safety were observed between Japanese and non-Japanese patients.