OBJECTIVE To compare the effects of the alpha 1-blocker bunazosin retard and the beta 1-blocker atenolol (Uniloc) on insulin sensitivity and glucose and lipid homeostasis in patients with type-2 diabetes and hypertension. METHODS Patients with controlled type-2 diabetes (non-insulin-dependent diabetes mellitus), treated by diet or oral sulphonylurea derivatives, and with mild-to-moderate hypertension were include in a randomized, parallel group, double-blind, multicentre study. After a single-blind placebo run-in period lasting 4-6 weeks, the patients were treated either with bunazosin retard or with atenolol for a further 16 weeks including an initial dose titration period to achieve blood pressure control. Treatment involved 3, 6 or 12 mg bunazosin retard tablets or 25, 50 or 100 mg atenolol tablets, administered orally once a day and prescribed according to blood pressure response. The euglycaemic hyper-insulinaemic clamp technique was used to assess insulin sensitivity both after the placebo period and after the active treatment. A total of 95 patients was enrolled in the study (placebo phase). Forty-eight patients were withdrawn from the placebo phase, mainly due to their blood pressures being outside the required range (seated diastolic blood pressure 90-114 mmHg) and 47 patients were allocated randomly to active treatment. Of these, 23 were administered bunazosin retard and 24 atenolol. All evaluations were on an intention-to-treat basis. RESULTS Insulin sensitivity assessed as glucose utilization during the clamp was significantly higher following bunazosin retard compared with following atenolol administration (3.52 +/- 0.27 versus 2.86 +/- 0.19 units of metabolic clearance rate of glucose index, P < 0.05). The insulin level attained during clamps (infusion rate 56 mU/m2 per min) was higher (P < 0.05) following atenolol (117 +/- 5 mU/l) than it was following bunazosin retard administration (102 +/- 5) or placebo (108 +/- 3), possibly due to an impaired insulin clearance. Compared with placebo, atenolol treatment resulted in significantly increased glucosylated haemoglobin whereas bunazosin retard had no significant effect. The two drugs did not show any consistent differences in lipid profile or fibrinogen and plasminogen activator inhibitor 1 levels. During the study seven serious adverse events were reported and one was reported shortly after completion of the study. All except one were classified as not related to the study drug and five of them occurred during placebo treatment. The non-serious side effects were in general considered to be either unrelated to the test drugs or expected effects of the two respective drug classes. Both bunazosin retard and atenolol displayed acceptable safety profiles. CONCLUSIONS Bunazosin retard treatment in hypertensive non-insulin-dependent diabetes mellitus patients appears to be associated with a slightly higher insulin sensitivity than is atenolol.