We investigated how fasting or postprandial insulin levels were altered by treatment with acarbose or sulfonylureas. Plasma glucose and serum insulin, C-peptide, and proinsulin levels were measured before as well as 1 and 2 h after breakfast in 23 patients with non-insulin-dependent diabetes mellitus and 17 patients with impaired glucose tolerance. In the diabetic patients, 12 weeks of acarbose therapy decreased the postprandial levels of glucose (1 h: -60.0%; 2 h: -67.6%), insulin (1 h: -67.5%; 2 h: -72.2%) and proinsulin (1 h: -55.2%; 2 h: -46.7%), and proinsulin (1 h: -20.9%; 2 h: -57.5%). In contrast, sulfonylurea treatment increased postprandial insulin and proinsulin levels. Since increased in the serum insulin or proinsulin levels are associated with a higher risk of cardiovascular disease, the present findings suggest that the acarbose-induced reduction of the postprandial serum insulin or proinsulin responses to food intake might be useful for preventing vascular complications in patients with diabetes.