N(α)-histogram analysis of the ventricular fibrillation ECG-signal as predictor of countershock success

Previous work has suggested that the heart in ventricular fibrillation can be defibrillated more easily if the ECG shows higher mean-frequencies or higher absolute voltages. We investigated whether information on defibrillation success is encoded in the random behaviour of the ECG-signals during fibrillation in a porcine model of cardiac arrest. To estimate the degree of randomness we used N(α)-histograms, which have been used to characterize turbulent flows or to denoise astrophysical data. We found that width of N(α)-histograms can serve as predictors for defibrillation success. The width decreases the more random the ECG-signal behaves. In our setting, the probability to find values αi>2.5 is dramatically increased by application of the vasopressor vasopressin (and conversely decreased for values αi<2.5). Even 10–12 min after fibrillation, vasopressin returns N(α)-histograms to the form they show 40 s after initiation of ventricular fibrillation.