Every year 400,000 - 600,000 people in the United States die from sudden cardiac death. Sudden cardiac death is often caused by irregular electrical impulses, or arrhythmias, in the heart. Arrhythmias can be corrected through pharmacological therapies, device therapies, or both. One type of device therapy, pacemakers, are inserted in the heart to correct arrhythmias. After a period of ventricular pacing, cardiac memory is defined by changes in the T-wave that are persistent upon return to normal activation pathways. During ventricular pacing, regional stroke work in areas closest to the pacing electrode is significantly decreased. We hypothesize that the mechanical function in the region around the pacing site will continue to have altered mechanical function after cession of pacing, in effect showing a mechanical cardiac memory. To test the hypothesis, nine canine models were implanted with pacing electrodes in both the atrium and ventricle. After a forty- minute stabilization period, baseline data were obtained during atrial pacing. Cardiac memory was induced in five canine models through a two-hour period of ventricular pacing followed immediately by atrial pacing. The remaining canine models served as controls, undergoing atrial pacing for two hours. High- density mapper (HDM) was used to determine mechanical function in a region centered approximately 1 cm away from the pacing electrode. No differences in global function (tau, developed pressure, dP/dtmax, dP/dtmin) were found after two hours of ventricular pacing upon return to normal activation pathways. There was a significant decrease in regional stroke work in an area close to the electrode between baseline (5.7 ± 2.6 %), during ventricular pacing (-3.8 ± 0.9 %)(p<0.05) and after two hours of ventricular pacing upon return to normal activation pathways (2.4 ± 1.6 %)(p<0.05). Further, systolic area contraction was also significantly different between baseline (5.0 ± 6.6 %) and after two hours of ventricular pacing upon return to normal activation pathways (0.2 ± 7.4 %)(p<0.05). Diastolic twist and diastolic twist rates showed no significant differences. Finally, contractile principal strain increased by inducing cardiac memory (-2.6 ± 0.3 %) as compared to baseline (-1.1 ± 0.5 %)(p<0.05). These findings suggest there is a mechanical correlation to electrical cardiac memory.

The presented studies investigate dynamics of action potential duration (APD) tobetter understand the underlying mechanism for repolarization alternans.We recorded trans-membrane potentials (TMP) in canine endocardial muscle tissueusing standard glass microelectrode under the control of an explicit diastolic interval (DI)control pacing protocol, i.e. feedback protocol. During sequential sinusoidal DI activation,the trajectory of APD dynamics has multiple values of APD correspondent to the sameDI, i.e. restitution is a bi-modal relationship. Our results indicate that: 1) there is a delay,similar to hysteresis, of change in APD responding to change in DI, 2) and the timecourse of the delay is asymmetric for fast or slow pacing history. The alternans wasobserved during constant DI pacing, i.e. the DI preceding each APD was invariant orchanged within a limited range. This finding suggests that alternans of APD do not needthe oscillation of preceding DI, i.e. DI dependent restitution is not a necessary conditionfor the alternans. This result implies that DI independent component exists in themechanism of the alternans. Nonetheless, the amplitude of alternans was statisticallylarger during constant pacing cycle length (PCL) pacing than that during constant DIpacing, even though both PCL and DI pacing trials used similar average activation rate.These results also demonstrate the ability of the feedback protocol to analyze the memoryeffects and dissect different components in the mechanism of alternans.Two computational models, Luo-Rudy dynamics (LRD) and cardiac ventricle model(CVM) were used to study the hysteresis in restitution. By perturbing membrane current:L-type calcium current, rapid and slow potassium rectifier, and intracellular calciumtransfer rate in sarcoplasmic reticulum (SR) and using sinusoidal DI pacing sequence, weshowed that the asymmetric calcium current across the membrane and its interaction withcalcium buffer in SR during increasing and decreasing DI phase plays an important rolein the hysteresis. CVM was used to study the alternans during constant DI pacing.However CVM failed to replicate the alternans that occurred in the experiments. Thisresult implies that CVM lacks the electrophysiological kinetics related to alternans thatwas shown in our experiment.

Repolarization alternans, i.e. beat-to-beat variation of repolarization of action potential, is proposed as a predictor of life-threatening arrhythmias. Restitution relates repolarization duration with its previous relaxation time, i.e. diatstolic interval (DI), and is considered a dominant mechanism for alternans. Previously, we observed that different repolarization durations at the same DI during decelerating and accelerating pacing, i.e. restitution displays hysteresis, which is a measure of “cardiac memory”. Objective of the current study was to investigate in the pig 1) the mechanism for a previously observed hysteresis type phenomenon, where alternans, once started at higher heart rate, persists even when heart rate decreases below its initiating rate, 2) regional differences in expression of hysteresis, i.e. memory in restitution in the heart, and 3) changes in restitution and memory during manipulation of an important repolarization current, the slow delayed rectifier, IKs. Action potentials were recorded in pig ventricular tissues using microelectrodes. Regional differences were explored in endocardial and epicardial tissues from both ventricles. DIs were explicitly controlled in real time to separate restitution mechanism from non-restitution related effects. Stepwise protocols were used to explore the existence in hysteresis in alternans threshold, where DIs were held constant for each step and progressively decreased and then increased. Quantification of cardiac memory was achieved by sinusoidally changing DI protocols, which were used to investigate memory changes among myocytes from different regions of the heart and during IKs manipulation. Results show that during stepwise protocol, hysteresis in alternans still existed, which indicates that restitution is not the only mechanism underlying the hysteresis. When comparing hysteresis obtained from sinusoidally oscillatory DIs among different regions, results show memory is expressed differently with endocardium expressing the most and epicardium the least memory. This provides important implications about the location where arrhythmia would initiate. Results also show that measures for hysteresis loops obtained by sinusoidal DI protocols decreased (increased) after enhancement (attenuation) of IKs, suggesting decreased (increased) hysteresis, i.e. memory in restitution. This effect needs to be considered during drug development.

Individuals with cardiomyopathy are at higher risk to die from sudden cardiac arrest than those with non-failing (NF) hearts. This study examined the differences in electrical properties of failing and NF human hearts in terms of cardiac memory through explicit control of diastolic intervals in a sinusoidal fashion, restitution of action potential duration (APD) through standard and dynamic pacing protocols, maximum rate of depolarization and APD alternans. Recordings of transmembrane potentials were made in tissues extracted from patients with heart failure and one donor NF heart. Computational simulations were performed using the O’Hara Rudy model for generating surrogates of control data. Significant differences were seen between left ventricular (LV) tissue and NF LV tissue in tilt, and measures of memory in terms of area and thickness during the sinusoidal 400ms protocol. Minimum delay was also significantly higher in the failing LV during the sinusoidal 150ms protocol. Failing tissues showed a higher restitution slope and prolonged AP which is consistent with previous studies and is hypothesized to contribute to the increased susceptibility to unstable alternans. This study further explored how disease alters the electrical functioning of the heart and why these patients are at a higher risk of ventricular arrhythmia.