Overdrive pacing of spiral waves in a model of human ventricular tissue

High-voltage electrical defibrillation remains the only reliable method of quickly controlling life-threatening cardiac arrhythmias. This paper is devoted to studying an alternative approach, low-voltage cardioversion (LVC), which is based on ideas from non-linear dynamics and aims to remove sources of cardiac arrhythmias by applying high-frequency stimulation to cardiac tissue. We perform a detailed in-silico study of the elimination of arrhythmias caused by rotating spiral waves in a TP06 model of human cardiac tissue. We consider three parameter sets with slopes of the APD restitution curve of 0.7, 1.1 and 1.4, and we study LVC at the baseline and under the blocking of INa and ICaL and under the application of the drugs verapamil and amiodarone. We show that pacing can remove spiral waves; however, its efficiency can be substantially reduced by dynamic instabilities. We classify these instabilities and show that the blocking of INa and the application of amiodarone increase the efficiency of the method, while the blocking of ICaL and the application of verapamil decrease the efficiency. We discuss the mechanisms and the possible clinical applications resulting from our study.


Supplementary
Here we show results for a straightforward stimulation with a constant period. We made two series of simulations with a TP06 model and isotropic medium. The first series used a stimulation period of 228 ms, which is about 0.96 of the spiral wave period. The second series used a stimulation period of 235 ms or about 0.99 of the spiral wave period. We varied the start of the pacing from 1900 ms to 2125 ms with a step of 25 ms and time of S2 stimulus (which changed the spiral wave position) from 315 ms to 360 ms with a step of 5 ms in both series. Totally, each series consisted of 100 simulations. The pacing was given from 2 s to 30 s and the result was assessed at 31 s, that is, after 1 s of the silence of the electrode. Our statistics are shown in Supplementary Table S 4. Note that the row 'Spiral removed' shows results for spiral elimination due to both superseding by overdrive pacing and annihilation with other spirals. The outcome 'Still a single spiral' in many cases just indicates that 30 s was not sufficient for spiral removal, thus we have much more such cases for the stimulation period of 235 ms when the induced drift speed was much lower. New spirals occurred here at the electrode as a result of simulation at the vulnerable phase.
The results for stimulation with a constant period are provided here just as a benchmark, as such straightforward stimulation protocol is used neither in clinical nor in most numerical settings.
Supplementary The pacing with a period of 228 ms was successful only in a few cases because the long electrode caused new spiral waves to appear via a mechanism similar to the one used in the S1S2 protocol. When the pacing phase was appropriate, new spiral waves did not emerge, and we saw the growth of the electrode-controlled area in the domain toward the spiral wave core and then the induced drift. After the original spiral wave was superseded, the pacing of the 'empty' domain caused two new spiral waves to appear; therefore, a timely stop of the pacing is also important, at least in our experiments.
The stimulation with a period of 235 ms removed the spiral wave only accidentally when it annihilated with a new spiral wave generated at the electrode. The complete failure of LVC with that period was due to the very slow induced drift of the original spiral. In a few cases of the 56 cases, when still a single spiral wave persisted, a pacing duration of up to 1 min could move the spiral wave to the domain boundary.
In addition, the pacing with a constant period was ineffective in many cases with changed slopes and ionic currents due to the unreadiness of the cells to be stimulated faster than the spiral wave period. Therefore, pacing with a period essentially lower than T sw was done using a slow decreasing of the period.