It has recently been shown that the probability of successful defibrillation as a function of energy has a sigmoidal dose-response relationship. Determination of a defibrillation "dose-response curve" is time consuming and requires multiple defibrillation attempts. On the other hand, determination of a defibrillation threshold is achieved rapidly and would be better suited to study the effect of interventions on the ability to defibrillate patients. We assessed the relationship of defibrillation threshold to the defibrillation "dose-response curve" in twelve open chest, halothane anesthetized pigs. Ventricular fibrillation was induced electrically, and defibrillation was attempted by passing sequential pulse shocks through an indwelling catheter and plaque electrodes. Defibrillation threshold was determined by decreasing the stored voltage of the initial shock until it failed to defibrillate the heart. Five different stored voltage levels distributed around defibrillation threshold were then randomly administered, six times for each level. A "dose-response curve" was obtained for each animal. Defibrillation threshold superimposed on the "dose-response curve" at 76 +/- 7.2 percent (mean +/- SEM) defibrillation success. Energy delivered at 1.5 times average defibrillation threshold was predicted to achieve 100 percent defibrillation success for a single shock in all animals. We conclude that defibrillation threshold provides a simple and quantitative estimate of the ability to defibrillate with a predictable relationship to the "dose-response curve."