To determine if multiple shocks adversely affect the success of later shocks compared with early shocks, we analyzed the success rates of initial shocks (defibrillation attempts 1-5), first half shocks (defibrillation attempts 1-20) and second half shocks (defibrillation attempts 21-40) in a canine model. Epicardial patches were placed on the right and left ventricle in 28 dogs. Ventricular fibrillation was induced by a 60-Hz shock. After 30 seconds, defibrillation was attempted using 7, 12, 13, or 18 joules with either a uniphasic or biphasic rectangular waveform. The uniphasic waveform was 5 msec in duration; the biphasic waveform was 10 msec, with the lagging 5-msec pulse one-half the amplitude of the leading 5-msec pulse. For uniphasic shocks, the right ventricular patch was positive; for biphasic shocks, the right ventricular patch was positive during the leading 5 msec of the shock and negative during the lagging milliseconds. A total of 960 fibrillation episodes were evaluated; no dog was involved in more than 40 fibrillation episodes. The success rates of defibrillation attempts 1-5, defibrillation attempts 1-20, and defibrillation attempts 21-40 were similar at 12, 13, and 18 joules. This information supports the continued use of up to 40 fibrillation trials in canine cardiac defibrillation. However, at 7 joules defibrillation attempts 21-40 were more successful than defibrillation attempts 1-5, and 1-20. With our methodology, these data are consistent with the hypothesis that low energy shocks create a "sensitizing" effect on cardiac tissue, allowing more successful defibrillation with repeated shocks.