We investigated whether short-term changes in serum insulin would effect a reduction of arterial pressure in subjects with therapy-resistant essential hypertension. Six patients were examined twice with a 3 week's interval in a single-blind cross-over design with euglycemic insulin clamps (A and B). A reduction of endogenous serum insulin was achieved by continuous infusion of 50 microgram octreoid (a somatostatin analogue) per hour. During clamp A low dose insulin infusion (5 mU/m2/min) was given, whereas during clamp B insulin was infused at a rate of 60 mU/m2/min. Preceding each clamp a standard drug therapy was given for one week (50 mg atenolol+ 30 mg furosemide per day). During clamp A plasma insulin was reduced from 21.4 +/- 7.5 to 10.8 +/- 1.2 mU/l (p < 0.01) whereas plasma insulin rose during clamp B from 20.0 +/- 7.5 to 99.0 +/- 17.2 mU/l (p < 0.001). The mean arterial blood pressure did not decrease during clamp A (low dose insulin infusion). There was an increased natriuresis during the high-insulin clamp (70 vs. 38 mmol, p = 0.13), but no difference in arterial pressure between the clamps. The results do not support the notion that high insulin levels contribute to hypertension in therapy resistant hypertensive patients by any direct and immediate mechanism.