OBJECTIVE- To assess the long-term safety and effectiveness of a titrated dose of the alpha-glucosidase inhibitor miglitol (BAY m 1099) in Hispanic NIDDM patients. RESEARCH DESIGN AND METHODS- A 1-year double-blind randomized placebo-controlled study in which diet-treated or diet plus sulfonylurea-treated Hispanic NIDDM patients received either placebo (n = 131) or miglitol in doses of 50, 100, 150, or 200 mg t.i.d. (n = 254), up-titrated and down-titrated based on tolerability. Efficacy parameters included changes from baseline in HbA(1c), fasting and 2-h postprandial plasma glucose and serum insulin, fasting serum lipids, and urinary albumin-to-creatinine ratio (ACR). Safety assessments consisted primarily of tabulation of adverse events and intercurrent illnesses, and of periodic laboratory determinations. RESULTS- Reductions from baseline in HbA(1c) levels at the 6-month (primary efficacy) endpoint were significantly greater by 0.83% in the miglitol group than in the placebo group. HbA(1c) reductions in the miglitol treatment group significantly exceeded those in the placebo group by 0.63, 0.73, and 0.92% at 3, 9, and 12 months of treatment, respectively. Reductions in 120-min postprandial glucose and insulin levels mere significantly greater in the miglitol group than in the placebo group at all postbaseline visits. There was little difference between treatments for changes in fasting insulin or lipid levels. Miglitol-associated reductions versus placebo in Easting plasma glucose (P = 0.0587 at 6 months) and in ACR (P = 0.0541 at 1 year) were nearly statistically significant. These efficacy results were not notably different between the 6-month endpoint, at which time the mean miglitol dose was 100 mg t.i.d., and the 1-year visit, when the mean miglitol dose was 149 mg t.i.d. Notable adverse events seen significantly more often in the miglitol group than in the placebo group were flatulence and diarrhea (or soft stools). The incidence of these gastrointestinal adverse events appeared to be dose dependent. CONCLUSIONS- Miglitol treatment of non-insulin-requiring Hispanic NIDDM patients at doses from 50 to 200 mg t.i.d. produced statistically and clinically significant reductions of HbA(1c), primarily associated with reduction of glucose and insulin levels in the postprandial period, which were sustained over a year of treatment. Adverse events related to the drugs mechanism of action were common, but generally well tolerated. Doses above 100 mg t.i.d. were not associated with notably enhanced efficacy in most patients.