The purpose of this study was to determine the most suitable treatment for Type 2 (non-insulin-dependent) diabetic patients with secondary failure to sulfonylureas (SFS). In a four-month comparative study, 36 Type 2 diabetic patients given SFS were allocated to three treatment groups: A (n = 12, M/F 6/6, HbAlc 9.1 +/- 1.6%) received 0.3 IU/Kg body weight (BW) of insulin-Zn between 10 and 11 p.m.; B (n = 12, M/F 6/6, HbAlc 9.2 +/- 1.6%) SFS plus 850 mg/day of metformin; and C (n = 12, M/F 6/6, HbAlc 9.5 +/- 2.4%) SFS plus acarbose 3 x 100 mg daily. Modifications in HbAlc, BW, blood pressure (BP), lipoprotein profile and insulin sensitivity were evaluated. HbAlc decreased in the three groups (A: 17.9 +/- 13.5%; B: 18.2 +/- 4.5%; C: 7.6 +/- 16.8%; all p < 0.05; A and B vs C = p < 0.05). BW increased in group A and decreased in the other groups. BP decreased statistically in group B. HDL-cholesterol increased (1.26 +/- 0.46 vs 1.49 +/- 0.36 mmol/L; p < 0.05) and triglyceride levels decreased (1.68 +/- 0.85 vs 1.16 +/- 0.43 mmol/L; p < 0.05) in group A. There were no significant changes in the other studied parameters. We conclude that, for Type 2 diabetic patients given SFS, both insulin and metformin plus SFS provided better glycaemic control than acarbose plus SFS. Metformin combined with SFS offered further advantages for the control of BW and BP.