The mechanism of cycle length oscillation and its role in spontaneous termination of reentry was studied in an in vitro preparation of canine atrial tissue surrounding the tricuspid orifice. Reentry occurred around a fixed path with incomplete recovery of excitability. Among 18 experiments, there was complete concordance between the occurrence of spontaneous cycle length oscillation and spontaneous terminations; both were observed in 10 experiments and neither in the other eight (p<0.001). Local changes in conduction during oscillations resulted from the dependence of both conduction velocity and action potential duration on the preceding local diastolic interval. Interval-dependent changes in action potential duration contributed to the oscillation by altering the next diastolic interval. Because of changes in action potential duration, changes in cycle length were poorly correlated with changes in diastolic interval and, therefore, with local conduction velocity. Complex oscillations resulted from variations in conduction time at multiple sites in the circuit. Oscillations caused most spontaneous terminations. The critical event was an exceptionally long diastolic interval preceding the next-to-last cycle that accelerated local conduction (which tended to shorten the last cycle) and prolonged action potential duration and refractoriness at the site of block. Ninety-two of 99 recordings of spontaneous termination showed evidence of oscillation of conduction and refractoriness causing block. (Circulation 1988;78:1277-1287

We have characterized, in dogs, a model of inducible regular atrial tachycardia that resembles atrial flutter. The model involves creating a Y-shaped lesion comprised of an intercaval incision and a connected incision across the right atrium. It is suitable for serial studies of the effects of pacing or antiarrhythmic drugs in chronically instrumented animals studied in the awake state for at least several months. The postoperative cycle length of the induced tachycardia varies from 143 to 188 msec, depending on the size of the dog. The tachycardia cycle length was consistent for each dog, and the rhythm--once induced--was very stable until terminated by pacing. The mechanism of the tachycardia was reentry due to circus movement based on the ability to induce and terminate it by premature impulses or overdrive, the ability to reset the tachycardia by single premature stimuli, the pattern of entrainment during overdrive stimulation, and the ability to terminate the tachycardia by interrupting the conduction pathway. The window of reset determined by the range of coupling intervals of premature stimuli that were able to enter and reset the tachycardia ranged from 56 to 82 msec. There appears to be incomplete recovery of excitability by the end of the excitable gap as evidenced by the fact that even late premature impulses that enter the reentrant circuit conduct more slowly than the tachycardia impulse, and because stimulation of muscarinic receptors that shortens the duration of the action potential and refractoriness also reduces the cycle length of the tachycardia. Epicardial and endocardial activation mapping during tachycardia showed the reentrant pathway does not merely encircle the lesion, particularly over the left atrial epicardium near the intercaval lesion. Rather, the impulse appears to travel around the atrial tissue just above the tricuspid ring, including a portion that travels through the right side of the lower intraatrial septum. Thus, the model involves circus movement around an anatomic barrier through normal tissue that contains no depressed segments. During the circus movement, there is a relatively long excitable gap during which there is incomplete recovery of excitability. This model should be useful for studies of the mechanism of antiarrhythmic drug action and the responses to premature stimulation in this particular subclass of reentrant rhythms, and for comparison with the behavior and responses of other forms of reentry.

We have previously found that skeletal muscle becomes electrically inexcitable in septic patients. Work in an animal model suggests that a decrease in the available sodium current underlies the loss of electrical excitability. We examined ECGs from patients during periods of septic shock to determine whether there were any ECG abnormalities that might suggest a similar loss of excitability in cardiac tissue during sepsis. Fourteen out of 17 patients had low or significantly decreased QRS amplitudes during septic shock; 8 of 17 had long or increased QRS duration with or without bundle branch block. The mean decrease in QRS amplitude in septic patients was 41%, significantly higher than in controls where no consistent decrease in QRS amplitude was found (p < 0.01). In patients who recovered from septic shock, the QRS amplitude and the increased QRS duration both returned to normal. We conclude that there is a loss of QRS amplitude during septic shock that may be due to altered cardiac excitability.

A Y-shaped lesion in the right atrium allows induction of atrial flutter in dogs. We recorded the activation sequence during this tachycardia from 96 endocardial bipolar electrodes using intracavitary electrode arrays during 12 separate episodes in three isolated perfused hearts. In each case a reentrant impulse circulated around the tricuspid valve orifice in either a clockwise or counterclockwise direction. Cutting the pathway terminated the rhythm and prevented its reinduction. There was no discrete segment of markedly slow conduction in the reentrant circuit. The tachycardia cycle length was decreased by methacholine and increased by lidocaine. Reentry was also induced in atrial tissue around the tricuspid orifice when this structure was isolated and superfused in vitro. Tachycardia cycle lengths varied from 205 to 399 msec, depending on the circumference of the ring and temperature. Induction of tachycardia by premature stimulation depended on differences in the duration of the effective refractory period among parts of the ring. Conduction velocity was relatively uniform and was slower during tachycardias than during pacing at long cycle lengths. Analysis of the response to premature stimuli that reset the tachycardia provided evidence for incomplete recovery of excitability between depolarizations during the tachycardia. Fast-response action potentials were recorded throughout the pathway and up to six to eight cell layers deep. Histologic studies showed the supravalvular lamina, a circumferential band of fibers several cell layers below the endocardial surface, to be continuous around the tricuspid orifice. Propagation through this layer best explains the conduction velocities observed in the intact heart during flutter in this preparation. Circulation 76, No. 5, 1155No. 5, -1175No. 5, , 1987 LEWIS and his colleagues' ,2 suggested that atrial flutter may result from circus movement around naturally occurring orifices in the atria. During brief episodes of atrial flutter induced by rapid pacing in normal dogs the activation sequence usually suggested a reentrant impulse circulating around one or both venae cavae, but the pattern of activation in one dog was more consistent with circus movement in the left atrium around the mitral valve orifice. Rosenblueth and Gar- Recently we created a model of stable inducible atrial flutter in instrumented dogs by use of a more extensive right atrial lesion than that used by Rosenblueth and Garcia-Ramos. We created an intercaval lesion extending from the superior to inferior venae cavae and added a second lesion connected to the first that extended across the right atrial freewall.7 A rapid regular atrial tachycardia that resembled atrial flutter on the electrocardiogram was easily inducible by premature stimulation or rapid pacing. The rhythm was very stable but could be terminated easily by overdrive pacing or premature stimulation and satisfied criteria for identification of reentrant excitation.8' 9