This study evaluated ventricular fibrillation mean frequency and amplitude to predict defibrillation success in a porcine cardiopulmonary resuscitation (CPR) model using repeated administration of vasopressin or epinephrine. After 4 min of cardiac arrest and 3 min of CPR, 10 pigs were randomly assigned to receive either vasopressin (early vasopressin: 0.4, 0.4, and 0.8 units/kg, respectively, n = 5) or epinephrine (early epinephrine: 45, 45, and 200 microg/kg, respectively, n = 5). Another 11 animals were randomly allocated after 4 min of cardiac arrest and 8 min of CPR to receive every 5 min either vasopressin (late vasopressin: 0.4 and 0. 8 units/kg, respectively, n = 5) or epinephrine (late epinephrine: 45 and 200 microg/kg, n = 6). Ventricular fibrillation mean frequency and amplitude on defibrillation were significantly higher in the vasopressin groups than in the epinephrine groups, respectively. In vasopressin versus epinephrine animals, mean frequency immediately before defibrillation was 9.6 +/- 1.5 Hz vs 7. 0 +/- 0.7 Hz (P < 0.001), mean amplitude was 0.65 +/- 0.26 mV vs 0. 21 +/- 0.14 mV (P < 0.001, and coronary perfusion pressure was 27 +/- 9 mm Hg vs 8 +/- 4 mm Hg (P < 0.00001), respectively. In contrast to no epinephrine animals, all vasopressin animals were successfully defibrillated and survived 1 h (P < 0.05). Mean fibrillation frequency and amplitude predicted successful defibrillation and may serve as noninvasive markers to monitor continuing CPR efforts. Furthermore, vasopressin was superior to epinephrine in maintaining these variables above a threshold necessary for successful defibrillation.