Aim: To compare the effect of vildagliptin versus metformin on pancreatic beta-cell function among newly diagnosed, non-obese Asian Indians with type 2 diabetes (T2D) in 2 years.

Methods: Randomized, multicentric, controlled, parallel-group study of 2 years treatment with vildagliptin or metformin in drug-naïve T2D persons (baseline HbA1c - 7.0 - 8.5% and body mass index (BMI) < 25kg/m2). Plasma glucose, insulin, HbA1c and C-peptide were measured at intervals of 6 months for 2 years. We randomised 203 (104 men and 99 women) T2D persons in two groups. Group 1 was treated with metformin, 500mg - 2000mg (n = 100), and group 2, with vildagliptin 100mg (n = 103).

Results: In the metformin group 36% and in the vildagliptin group 27.2 % required only monotherapy for a median duration of 48 weeks (25 - 70 weeks). At the end of the study, changes in the area under the curve of insulin to glucose were non-significant in both groups. There were no changes in the postprandial C-peptide. Reduction in insulin resistance (HOMA-IR) was similar in both groups. Insulin sensitivity (Matsuda Index) increased significantly at 6 and 12 months with metformin but no change was seen in the vildagliptin group. The HbA1c level decreased significantly in both groups until 18 months. The fasting plasma glucose decreased significantly till 12 months in both groups. Postprandial plasma glucose decreased only with vildagliptin till 12 months. Occurrences of adverse events were similar in both groups. Higher C-peptide and lower HbA1c values at the baseline predicted better glycemic control with vildagliptin. Lower baseline HbA1c predicted good glycemic control with metformin.

Conclusion: Both metformin and vildagliptin had similar protective effects on the beta-cell function, reduced insulin resistance and glycemic levels. Improved insulin sensitivity was seen with metformin.


K. Satheesh: None. C. Snehalatha: None. A. Nanditha: None. A. Raghavan: None. R. Vinitha: None. P. Susairaj: None. A. Ramachandran: None.


India Diabetes Research Foundation

Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered. More information is available at