OBJECTIVE To assess the efficacy, safety, and tolerability of acarbose versus placebo during a 24-week treatment period in Asian type 2 diabetic patients with dietary failure. RESEARCH DESIGN AND METHODS After a 6-week screening period, 126 multiethnic Asian type 2 diabetic patients (64 men, 62 women; mean age ± SD, 53.4 ± 10 years) were randomized to receive acarbose (n = 63) or placebo (n = 63). The dosage was increased from 50 mg t.i.d. at week 0 to 100 mg t.i.d. at week 4. Patients were then followed up at weeks 10, 16, and 24. At each visit, body weight, blood pressure, and metabolic indexes were measured. At weeks 0 and 24, fasting plasma glucose and insulin were measured before and 1 h after the administration of an individually tailored breakfast. RESULTS Using the intention-to-treat analysis, there were greater reductions in (mean [95% CI]) HbA1c (−0.70 [−1.00 to −0.39] vs. −0.27% [−0.54 to 0]; P = 0.04), fasting plasma glucose (−0.37 [−0.75 to 0.02] vs. 0.41 mmol/1 [−0.08 to 0.90]; P = 0.017) and 1-h plasma glucose (−0.77 [−1.44 to −0.10] vs. 0.65 mmol/1 [−0.07 to 1.36]; P = 0.05) in the acarbose group compared with the placebo group. With acarbose treatment, 78% of patients achieved an HbA1c <8% compared with 56% in the placebo group (P = 0.003). There was a greater reduction in body weight (−1.31 [−2.46 to −0.15] vs. 0.16 kg [−3.36 to 0.10]; P = 0.02) and higher incidence of flatulence (56 vs. 37%; P = 0.032) in the acarbose than in the placebo group. Using baseline HbA1c and race as covariates, there were no significant interethnic differences in treatment responses (P = 0.232 for treatment-race interaction; P < 0.001 for treatment effect). The dropout rates were similar between the two groups (acarbose, 11 of 63; placebo, 6 of 63). There were no significant laboratory adverse events in either group. CONCLUSIONS In this multicenter study involving six ethnic groups, acarbose 100 mg t.i.d. was an effective, safe, and generally well-tolerated therapy in Asian type 2 diabetic patients with dietary failure. In some patients with troublesome gastrointestinal symptoms, a lower dosage may be necessary.