Current concepts for the explanation of endothelial dysfunction and accelerated atherosclerosis in uremia propose a reduced vascular bioavailability of nitric oxide (NO). The aim of the present study was to test the contributions of NO and NO/prostacyclin (PGI(2))-independent mechanisms to both baseline vascular tone and agonist-induced endothelium-dependent vasodilation in patients on hemodialysis (HD). In 10 HD patients and eight matched healthy control subjects, forearm blood flow (FBF) was measured at rest and during intrabrachial infusions of norepinephrine (NE; endothelium-independent vasoconstrictor, 60, 120, and 240 pmol/min) and N-monomethyl-L-arginine (blocker of NO synthases, 16 micromol/min). After inhibition of cyclo-oxygenase by ibuprofen (1200 mg orally), endothelium-dependent and -independent vasodilation was assessed by infusion of acetylcholine (ACh; 1, 5, 10, 50, 100, and 300 nmol/min) and sodium-nitroprusside (2.5, 5, and 10 microg/min). NO/PGI(2)-independent vasodilation was tested by equal infusions of ACh during NO clamp. N-monomethyl-L-arginine reduced resting FBF to a comparable degree in both groups. Vascular responses to ACh were reduced in HD (P = 0.003 versus control by ANOVA), whereas those to sodium nitroprusside were mainly at control level. Infusion of ACh during NO clamp caused a similar increment of FBF in both groups. NO-mediated vasodilation as calculated by the difference between ACh-induced responses without and with NO clamp was substantially impaired in HD (P < 0.001) compared with control. In HD patients, baseline NO-mediated arteriolar tone is at control level. This study provides first evidence that endothelial dysfunction of uremic patients as shown by reduced agonist-induced endothelium-dependent vasodilation is attributable to reduced stimulation of NO, whereas the NO/PGI(2)-resistant portion of ACh-mediated vasodilation is unaffected.