OBJECTIVE - To evaluate the safety and effectiveness of triple therapy using Insulin, metformin, and thiazolidinedione following a course of dual therapy using insulin and metformin or insulin and a thiazolidinedione in type 2 diabetes. RESEARCH DESIGN AND METHODS - Twenty-eight type 2 diabetic subjects using insulin monothcrapy (baseline HbA(1c) level 8.5%) who had been randomly assigned to insulin (INS) and inetfortnin (MET) (INS + M ET, It = 14) or INS and the Lhiazoliclinedione troghLazone (TGZ) (INS + TGZ, n = 14) (dual therapy) for 4 months were given INS, MET, and TGZ (triple therapy: INS + MET, add TGZ, or INS + TGZ, add MET) for another 4 months. The INS close was not increased. RESULTS - HbA(1c) levels decreased in both groups during dual therapy and improved further during triple therapy (INS+ MET 7.0 +/- 0.8, INS + TGZ 6.2 +/- 0.8, P < 0.0001 INS + MET, acid TGZ 6.1 +/- 04%, P < 0.001 INS + TGZ, add MET 5.8 +/- 0.6%, P < 0.05, and INS + TGZ vs. INS + MET, P = 0.02). Significant reductions In total daily insulin close occurred in the INS + TGZ (- 14.1 units, P < 0.0001), INS + TGZ add MET (- 13.7 units, P < 0.01), and the INS + MET add TGZ aroups (- 17.3 units, P < 0.003), but not in the INS + MET group (-3.2 units) (INS + TCZ vs. INS + MET P < 0.05). Subjects in the INS + TGZ group experienced significant weight gain (4.4 +/- 2.7 kg, P < 0.0005). No weight gain occurred in the INS + MET, INS + MET add TGZ, and INS + TGZ add MET groups. CONCLUSIONS - Triple therapy using INS, MET, and TGZ resulted in lower HbA(1c) levels and total daily insulin dose than during dual therapy. The use of triple therapy resulted in 100% of subjects achieving an HbA(1c) <7.0%, while decreasing the dose of INS. Weight gain was avoided when MET therapy preceded the addition of TGV therapy. The addition of TGZ resulted in the greatesL redUCLIOnS in HbA, levels and insulin close, Triple Lherapy using INS, MET, anct a thiazolidinedione (such as TGZ) can be a safe and effective treatment in type 2 diabetes.