The present study was a 52-week extension of a previously published, multi-center, randomized, parallel-group study. The aim of this extension study was to compare the efficacy and tolerability of vildagliptin and metformin in drug-naïve patients with type 2 diabetes over 104 weeks. The extension population comprised 305 patients randomized to vildagliptin (100 mg daily) and 158 patients randomized to metformin (2 000 mg daily). Pioglitazone was added as rescue medication if fasting glucose was >10 mmol/l; data from patients receiving rescue medication were excluded from the primary analysis. Baseline HbA (1c) averaged 8.4+/-0.1% in patients randomized to vildagliptin and 8.8+/-0.1% in those randomized to metformin. The adjusted mean change from baseline to study endpoint was -1.0+/-0.1% in vildagliptin-treated patients and -1.5+/-0.1% in those receiving metformin (p<0.001 vs. vildagliptin). These results were similar to those reported after the 1-year core phase of the study. The adjusted mean changes in body weight from baseline to endpoint were 0.5+/-0.4 kg and -2.5+/-0.5 kg in the vildagliptin and metformin groups, respectively. One or more adverse event (AE) was reported by 82.2% of patients receiving vildagliptin and by 87.3% of those receiving metformin (p<0.001). Gastrointestinal AEs were more common in patients receiving metformin (45.6%) than in those receiving vildagliptin (25.0%, p<0.001 vs. metformin). One hypoglycemic event occurred after strenuous exercise in a single patient receiving vildagliptin (0.3%). In conclusion, both vildagliptin and metformin monotherapy provided clinically meaningful decreases in HbA (1c) over 2 years in drug-naïve patients with type 2 diabetes. Vildagliptin was weight neutral, while weight loss was observed with metformin; however, metformin was associated with significantly worse gastrointestinal tolerability.