Effects of defibrillation shock energy and timing on 3-D computer model of heart.

We present computer simulations of electrical defibrillation in a three-dimensional model of the ventricles of the heart. In this model, called HEARTSIM, the ventricles are represented by 1473 cubic elements with 3 mm sides. The action potential is described by five discrete states; absolutely refractory, three relatively refractory, and repolarized. Activation is propagated to an element's six orthogonal neighbors with the conduction velocity dependent on the refractory state of the neighbor. Delivery of several extra-stimuli with decrementing coupling intervals results in ventricular fibrillation. Following the onset of ventricular fibrillation, we simulate defibrillation using various electrode configurations, shock energies, and timings. The current density distributions in the heart model resulting from the defibrillation shocks are determined from finite element analysis of the electric fields produced by the delivery of high energy shocks. The simulations suggest that successful defibrillation shocks produce a short period of low activation followed by a complete cessation of activation for a duration of 387 +/- 162 ms. In contrast, unsuccessful shocks produce a significantly shorter period of low activation (70 +/- 12 ms) after which ventricular fibrillation resumes. HEARTSIM mimics the experimentally reported, highly variable response to near-threshold shocks--the energy for successful defibrillation varies widely (20.8 +/- 20.7 J). In addition, the success rate vs. energy curve has a sigmoidal shape that is consistent with experiments. We demonstrate that this variability in the energy requirement results from dynamic variability in the number of elements made refractory by the shock and the relative distribution of the activation pattern at the time of the shock. Further, we show that it may be possible to lower the defibrillation energy requirements by delivery of two successive low energy pulses. The most efficient timing for the second pulse corresponds to the repolarization of the elements that were excited by the first pulse. Thus, when the interval between the two pulses was 85 +/- 18 ms, the defibrillation threshold energy (DFE) is reduced by 30.7 +/- 10% with pulses of 10 ms duration, and 62.6 +/- 7.9% with pulses of 5 ms duration. Our simulations also show that there is a delicate balance of energy between the two pulses that must be reached in order to achieve energy reduction with double pulse defibrillation. In conclusion, HEARTSIM serves as a tool for studying the underlying mechanisms of the effects of DF shocks on ventricular arrhythmias, and assists in evaluation of improved strategies for shock delivery.