We developed an isolated limb preparation to evaluate the direct vasoactive properties of cardioactive drugs on the forearm vasculature in patients with congestive heart failure. Using this model, we infused milrinone in subsystemic doses (1, 10, and 20 μg/min per 100 ml forearm volume [FAV]) into the brachial artery of 13 patients with moderate-to-severe congestive heart failure. We monitored forearm hemodynamics, systemic hemodynamics, and milrinone plasma concentration from both the forearm venous effluent and pulmonary artery. This preparation enabled us to assess the direct forearm vascular response to milrinone. Compared with baseline forearm blood flow (2.46 – 1.37 ml/min/100 ml FAV), the three doses of milrinone resulted in increases in forearm blood flow to 2.66 ± 1.43, 4.21 ± 1.79, and 6.73 ± 3.69 ml/min/100 ml FAV. This was associated with a reduction of forearm vascular resistance from the baseline value of 52 + 38 U to 47 ± 36, 25 + 13, and 17 ± 10 U. The p value for the difference in response of flow and resistance after the 10 and 20 gg doses vs that at baseline was .05. This forearm vasodilatation occurred without change in systemic hemodynamics or therapeutic milrinone plasma concentrations in the pulmonary artery. In five patients, we compared the response to intra-arterial milrinone with that of nitroprusside. At a dose of 10 μg/min/ 100 ml FAV, the response to nitroprusside (7.20 ± 3.24 ml/min/100 ml FAV) was greater than that to milrinone (4.65 2.18 ml/min/100 ml FAV) (p < .05). When milrinone was administered by a systemic intravenous route, the magnitude of forearm vasodilatation was not as great as that with intra-arterial milrinone, suggesting different sensitivities of vasodilatation in alternate vascular beds or the influence of baroreceptor autoregulation. Thus, this study identifies direct vasodilator properties of milrinone that are independent of its inotropic activity.