INTRODUCTION: The efficacy and safety of taspoglutide, a long-acting human glucagon-like peptide-1 analog, were compared with sitagliptin or placebo, as adjunct to metformin, in patients with inadequately controlled type 2 diabetes. METHODS: In this randomized, double-blind, double-dummy, parallel-group trial, patients were randomized to taspoglutide 10 mg once weekly (QW), 20 mg QW, 100 mg sitagliptin once daily (QD), or placebo for 24 weeks, followed by 28-week short-term and 104-week long-term extension periods. The primary endpoint was change in glycosylated hemoglobin (HbA(1c)) after 24 weeks. RESULTS: In this study, 666 patients (baseline HbA(1c), 7.96% [SD, 0.87]; fasting plasma glucose, 9.61 mmol/L [2.56]; body weight, 92.4 kg [19.3]) were randomized to taspoglutide 10 mg QW (n = 190), 20 mg QW (n = 198), 100 mg sitagliptin QD (n = 185), or placebo (n = 93) for 24 weeks. After 24 weeks, least squares mean (SE) HbA(1c) reductions were greater with taspoglutide 10 mg (-1.23% [0.06]) and 20 mg (-1.30% [0.06]) versus sitagliptin (-0.89% [0.06]) or placebo (-0.10% [0.08]). Mean treatment differences with taspoglutide 10 mg and 20 mg were -0.34 (95% confidence intervals [CI]: -0.49, -0.19) and -0.41 (-0.56, -0.26) versus sitagliptin; and -1.13 (-1.31, -0.95) and -1.20 (-1.38, -1.02) versus placebo. Weight loss was greater with taspoglutide 10 mg (-1.8 kg [0.3]) and 20 mg (-2.6 kg [0.3]) than sitagliptin (-0.9 kg [0.3]) or placebo (-0.5 kg [0.4]). Effects on HbA(1c) and weight loss continued through 52 weeks of treatment. No cases of severe hypoglycemia occurred with any active treatment. Gastrointestinal adverse events, and allergic and injection-site reactions were higher in the taspoglutide groups, causing higher discontinuation rates. Anti-taspoglutide antibodies were confirmed in 46% of patients. CONCLUSION: Taspoglutide demonstrated better efficacy on glycemic control and weight loss than sitagliptin, but a high incidence of adverse events led to high discontinuation rates. The safety profile of taspoglutide in this trial was similar to other trials in the clinical program, and led to the discontinuation of dosing.