Aim The aim of this study is to assess the efficacy and safety of vildagliptin 50?mg bid as add-on therapy to insulin in type 2 diabetes mellitus (T2DM). Methods This is a multicentre, double-blind, placebo-controlled, parallel group, clinical trial in T2DM patients inadequately controlled by stable insulin therapy, with or without metformin. Patients received treatment with vildagliptin 50?mg bid or placebo for 24?weeks. Results In all, 449 patients were randomized to vildagliptin (n?=?228) or placebo (n?=?221). After 24?weeks, the difference in adjusted mean change in haemoglobin A1c (HbA1c) between vildagliptin and placebo was -0.7 +/- 0.1% (p?<?0.001) in the overall study population, -0.6 +/- 0.1% (p?<?0.001) in the subgroup also receiving metformin and -0.8 +/- 0.2% (p?<?0.001) in the subgroup without metformin. Vildagliptin therapy was well tolerated and had a similarly low incidence of hypoglycaemia compared with placebo (8.4 vs. 7.2%, p?=?0.66) in spite of improved glycaemic control, and was not associated with weight gain. (+0.1 vs. -0.4?kg). Conclusions Vildagliptin 50?mg bid added to insulin significantly reduced HbA1c in patients with T2DM inadequately controlled by insulin, with or without metformin. Vildagliptin was well tolerated, with a safety profile similar to placebo. These results were achieved without weight gain or an increase in hypoglycaemia incidence or severity in spite of improved glycaemic control.