Ventricular fibrillation (VF) voltage was previously identified as a predictor of the success of cardiopulmonary resuscitation. In the present study we investigated the mechanism by which VF voltage predicts the success of cardiac resuscitation in a well-established rodent model of cardiac arrest. After 4 minutes of untreated VF, precordial compression was initiated and maintained for 6 minutes. Increases in coronary perfusion pressure during precordial compression were associated with concomitant increases in VF voltage (r = 0.61, p = 0.013). Significantly greater coronary perfusion pressure (24 vs 17 mm Hg) and VF voltage (0.17 vs 0.12 mV) were observed in resuscitated animals. To obviate electrical artifacts produced by precordial compression, boluses of oxygenated blood were injected into the ascending aorta in another 5 animals as an alternative method of cardiac resuscitation. This restored myocardial perfusion before defibrillation. Increases in VF voltage from 0.04 mV to 0.47 mV during aortic infusions were again correlated with coronary perfusion pressure (r = 0.62, p < 0.01) and predicted the success of cardiac resuscitation. Greater VF voltages after initiation of cardiac resuscitation were associated with increases in myocardial creatine phosphate, from 0.23 to 0.70 mmol/kg wet weight, and significant decreases in lactate content, from 22.8 to 13.9 mmol/kg wet weight. Increases in creatine phosphate were highly correlated with increases in VF voltage (r = 0.99, p < 0.01). Accordingly, increases in VF voltage during cardiac resuscitation reflect increases in myocardial perfusion and favorable changes in myocardial energy metabolism. As such, VF voltage, like coronary perfusion pressure, serves as a quantitative predictor of the success of cardiopulmonary resuscitation.