To determine whether adrenergic drugs administered during CPR alter the distribution of artificial cardiac output, the authors measured regional blood flow and cardiac output using radioactive microspheres in 12 dogs. Ventricular fibrillation was induced electrically and CPR was immediately begun with a mechanical chest compressor and ventilator (Thumper) at 60 compressions/min, with a ventilation: compression ratio of 1:5, a compression duration of 0.5 sec, and a ventilation pressure of 20 cm H2O. Compression force was sufficient to develop 40--50 mm Hg peak intraesophageal pressure. After 30 sec of CPR, either 0.9% saline vehicle or 50 micrograms/kg of epinephrine, phenylephrine, or isoproterenol was administered through a central venous catheter. One min later, microspheres were injected into the left ventricle. After 250 sec of CPR, the ventricles were defibrillated electrically. Between each drug injection, 20-min recovery periods were interposed. Each dog received all three drugs and saline according to a predetermined sequence. After saline, epinephrine, phenylephrine, and isoproterenol treatment, respective, cardiac output averaged 392, 319, 255, and 475 ml/min; brain blood flow averaged 37, 54, 29, and 28 ml/min; coronary blood flow averaged 25, 79, 26, and 15 ml/min; and kidney blood flow averaged 44, 4, 16, and 29 ml/min. Epinephrine improved blood flow to the brain, probably because of its alpha-adrenergic activity. Epinephrine improved blood flow to the heart during CPR much more than the other agents, probably because of its combined alpha- and beta-adrenergic activity. This effect may explain its superiority in restoring circulation after prolonged arrest and resuscitation. Isoproterenol should not be used in CPR because it shunts blood away from vital organs.