Acarbose, 300 mg/day, was administered over one month in a cross-over trial to 18 hyperglycemic patients aged 41-66 years with non-insulin-dependent diabetes mellitus (NIDDM). All showed "normal" or exaggerated insulin release after a glucose challenge and remained in poor control (random glucose levels greater than or equal to 13 mmol/l) despite involvement in a diabetes intervention programme and prior use of oral hypoglycemic agents. During the one month treatment with Acarbose, fasting glucose and % HbAl concentrations were not different from those observed during placebo therapy. Furthermore, glucose tolerance was unchanged by Acarbose treatment. Glucose concentrations after a 1.6 MJ test meal were reduced by Acarbose from peak values of 17.3 +/- 1.0 to 15.0 +/- 1.1 mmol/l and were associated with lower post-prandial C-peptide (CPR) and insulin responses. Nevertheless, daily insulin production, as assessed by CPR excretion rates and plasma CPR and insulin concentrations, was not reduced by Acarbose. In fact, fasting plasma insulin and CPR levels were significantly higher during Acarbose then placebo therapy. Acarbose (100-400 mg/day) was continued for six months in 12 of these patients. During treatment, post-prandial glucose levels remained lower but monthly MBG values, determined by self-measurement of blood glucose, were unchanged except for small reductions in the 4th and 5th treatment months. % HbAl levels did not change. These data show that Acarbose treatment of a defined group of patients with poorly controlled NIDDM: resulted in small but sustained reductions of post-prandial glucose levels but without improving glucose tolerance, and reduced the circulating concentrations of insulin and CPR postprandially, but overall did not reduce daily production.