OBJECTIVE: To compare long-term efficacy and safety of initial combination therapy with nateglinide/metformin versus glyburide/metformin. RESEARCH DESIGN AND METHODS: We conducted a randomized, multicenter, double-masked, 2-year study of 428 drug-naïve patients with type 2 diabetes. Patients received 120 mg a.c. nateglinide or 1.25 mg q.d. glyburide plus 500 mg q.d. open-label metformin for the initial 4 weeks. During a subsequent 12-week titration period, glyburide and metformin were increased by 1.25- and 500-mg increments to maximum daily doses of 10 and 2,000 mg, respectively, if biweekly fasting plasma glucose (FPG) > or = 6.7 mmol/l. Nateglinide was not titrated. Blinding was maintained by use of matching placebo for nateglinide and glyburide. An 88-week monitoring period followed, during which HbA1c (A1C), FPG, and postprandial glucose excursions (PPGEs) during an oral glucose tolerance test were measured. RESULTS: In nateglinide/metformin-treated patients, mean A1C was 8.4% at baseline and 6.9% at week 104. In glyburide/metformin-treated patients, mean A1C was 8.3% at baseline and 6.8% at week 104 (P < 0.0001 vs. baseline for both treatments, NS between treatments). The deltaPPGE averaged -96 +/- 19 (P < 0.0001) and -57 +/- 22 mmol.l(-1).min(-1) (P < 0.05) in patients receiving nateglinide/metformin and glyburide/metformin, respectively, whereas deltaFPG was -1.6 +/- 0.2 (P < 0.0001) and -2.4 +/- 0.2 mmol/l (P < 0.0001) in patients receiving nateglinide/metformin and glyburide/metformin, respectively (P < 0.01 between groups). Thus, the two treatments achieved similar efficacy with differential effects on FPG versus PPGE. Hypoglycemia occurred in 8.2 and 17.7% of patients receiving nateglinide/metformin and glyburide/metformin, respectively. CONCLUSIONS: Similar good glycemic control can be maintained for 2 years with either treatment regimen, but nateglinide/metformin may represent a safer approach to initial combination therapy.