Forearm vascular responses to arginine vasopressin (AVP) infused into a brachial artery in a wide range of infusion rates (0.05-2.0 ng/kg per min) were examined in 20 young healthy volunteers. Intraarterial AVP at lower doses (0.05 and 0.1 ng/kg per min) caused forearm vasoconstriction, whereas AVP at a dose of 0.2 ng/kg per min or higher caused forearm vasodilatation. The maximal forearm vasoconstriction was induced at the venous plasma AVP level of 76.3 +/- 8.8 pg/ml. Forearm vasodilatation was associated with the venous plasma AVP level of 369 +/- 43 pg/ml or higher. Forearm vasodilatation was the result of the direct effect of AVP since forearm blood flow and vascular resistance in the contralateral arm did not change. We attempted to explore the mechanisms involved in AVP-induced direct vasodilatation. The treatment with indomethacin, 75 mg/d for 3 d, did not alter AVP-induced forearm vasodilatation. In contrast, intraarterial infusion of isoosmolar CaCl2 totally prevented AVP-induced forearm vasodilatation. Intra-arterial CaCl2 also markedly attenuated forearm vasodilatation induced by intraarterial sodium nitroprusside, but did not alter forearm vasodilatation induced by intraarterial isoproterenol. These results indicate that the direct vascular effects of intra-arterial AVP on the forearm vessels are biphasic, causing vasoconstriction at lower doses and vasodilatation at higher doses. The direct vasodilatation induced by intraarterial AVP at higher doses is not mediated by prostaglandins but may involve cGMP-related mechanisms.