It has been suggested that cocaine acts directly in the brain to enhance central sympathetic outflow. However, some studies suggested that the cardiovascular effects of cocaine are related to a peripheral action. To characterize further the site of cocaine's cardiovascular effect, we compared the hemodynamic effects of cocaine (2 mg/kg, i.v. bolus) with those observed after administration of an equimolar dose (2.62 mg/kg, i.v. bolus) of cocaine methiodide, a quaternary derivative of cocaine that does not penetrate the blood-brain barrier, by using sufentanil-sedated dogs. Cocaine produced significant (p < 0.05) increases in heart rate (+37+/-11 beats/min), mean arterial pressure (+55+/-11 mm Hg), left ventricular end-diastolic pressure (+5.3+/-1.0 mm Hg), and cardiac output (+2.4+/-0.9 L/min). Cocaine methiodide produced increases in heart rate (+57+/-11 beats/min), mean arterial pressure (+45+/-11 mm Hg), left ventricular end-diastolic pressure (+3.4+/-1.0 mm Hg), and cardiac output (1.1+/-0.9 L/min), which were not significantly different from those observed with cocaine. Because opiate sedation potentially might have attenuated central sympathetic outflow, we further confirmed the qualitative similarity of the actions of cocaine and cocaine methiodide on heart rate and blood pressure in unsedated, conscious dogs. Our data suggest that the cardiovascular effects of cocaine result primarily from a peripheral site of action.