In type 2 diabetes with "secondary failure of sulfonylurea therapy" good metabolic control can seldom be achieved by insulin therapy even with high insulin doses. Hyperinsulinemia however is a possible risk factor of cardiovascular disease in type 2 diabetes. Maintaining the effects of sulfonylurea action insulin should be added in as small amounts as possible to avoid hyperinsulinemia and to ameliorate hyperglycemia. 16 type 2 diabetics with "secondary failure" were treated either with insulin alone (group A; n = 8) or with 3.5 mg b.i.d. glibenclamide plus small amounts of intermediate insulin (group B; n = 8) in a randomised order. After the inpatient period outpatient control was performed monthly up to six months, later on four times a year up to two years. Both groups were comparable with regard to age, duration of diabetes, body weight and metabolic control. The daily insulin dose was 14 +/- 2 IU (means +/- SEM) after one month and 19 +/- 2 IU after two years in group B. In contrast 30 +/- 3 IU and 43 +/- 5 IU respectively were needed in group A (p less than 0.001). All patients B were treated with one daily injection, all patients A needed two injections. Resulting in nearly identical metabolic control in group A basal insulin levels exceeded those in group B after two years significantly (28.6 +/- 3.7 vs. 18.6 +/- 1.6 mcU/ml; p less than 0.01).(ABSTRACT TRUNCATED AT 250 WORDS)