This study compares the effects of methoxamine, a pure alpha 1-agonist, and epinephrine on cerebral and myocardial blood flow, central hemodynamics, and survival in a randomized placebo-controlled fashion during prolonged ventricular fibrillation (VF) in a canine model. Twenty-four anesthetized and ventilated adult mongrel dogs were instrumented for regional blood flow determinations using radio-labeled microspheres. The dogs were randomized to receive either 20 mg of methoxamine as a single intravenous bolus or repeated boluses of 0.02 mg/kg of epinephrine, 0.2 mg/kg of epinephrine, or normal saline solution placebo beginning at three minutes following induction of VF and initiation of closed chest cardiac massage (CCCM). Organ blood flow measurements were determined during normal sinus rhythm and after five and 20 minutes of VF. All six dogs receiving methoxamine were successfully resuscitated in contrast to only one in each of the epinephrine-treated groups and none of the dogs receiving placebo (p less than .01). Although epinephrine was associated with significantly higher blood pressures than placebo during cardiopulmonary resuscitation (CPR), blood pressures achieved with methoxamine were significantly higher than those observed in the other three treatment groups (p less than .001). Cerebral blood flow was significantly higher with both methoxamine and high-dose epinephrine (p less than .05). Mean left and right ventricular myocardial flows were highest with methoxamine but this did not achieve statistical significance. In contrast, organ flows measured in the animals receiving the lowest dose of epinephrine were not significantly higher than those associated with placebo. Cardiac output after 20 minutes of CPR was significantly lower with high-dose epinephrine than with methoxamine or placebo (p less than .05). Our results suggest that methoxamine significantly improves regional cerebral blood flow and survival during CPR and although high-dose epinephrine is associated with comparable improvements in regional cerebral blood flow, this treatment is associated with deterioration in central hemodynamics during prolonged VF and does not enhance survival.