Voltage gradients in transvenous and subcutaneous defibrillation and their risk of myocardial damage

Transvenous implantable cardioverter‐defibrillator (ICD) shocks have been associated with cardiac biomarker elevations and are thought in some cases to contribute to adverse clinical outcomes and mortality, possibly from myocardium exposed to excessive shock voltage gradients. Currently, there are only limited data for comparison with subcutaneous ICDs. We sought to compare ventricular myocardium voltage gradients resulting from transvenous (TV) and subcutaneous defibrillator (S‐ICD) shocks to assess their risk of myocardial damage.


| INTRODUCTION
Transvenous (TV) ICD shocks have been associated with cardiac biomarker elevations and adverse clinical outcomes, including significant mortality increase detected with both appropriate and inappropriate shocks. 1,2 It is not clear whether subcutaneous ICD (S-ICD) shocks are similarly associated with adverse clinical outcomes. Subcutaneous defibrillators, with their extra-thoracic electrodes, expose the heart to less mechanical and electrical stress than transvenous defibrillators, and this could make S-ICD shocks potentially less damaging than TV-ICD shocks. Indeed, Weigand and colleagues found that Troponin and CK-MB elevation with S-ICD implantation including defibrillation testing in 43 patients were significantly lower than a TV-ICD cohort that had no defibrillation testing. 3 Localized myocardial regions of very high electrical gradients produced by ICD shocks may cause myocardial damage. The degree of reversibility depends on the magnitude of the gradients, with higher gradients causing irreversible damage. The relative proximity of the RV coil of a TV-ICD to the heart muscle may predispose ICDs to this type of electrical damage, compared to the S-ICD coil which is much further from the myocardium. The possible damage to noncardiac tissues (such as pectoralis and intercostal muscles in proximity to the S-ICD coil, and serratus anterior and latissimus dorsi muscles in proximity to the generator) from the higher output of an S-ICD and coil proximity to skeletal muscles is noted, though its clinical significance may be less, due to their lower physiological importance as compared to the heart.
Animal studies have demonstrated less cardiac damage with S-ICD shocks than TV-ICD shocks. 4,5 A randomized controlled human study by Semmler et al. 1 clearly established the association of TV-ICD shocks to an increment of injury-related cardiac enzyme elevation. To further understand the mechanisms by which shocks could contribute to myocardial injury and overall mortality, and to examine potential differences between TV-ICDs and S-ICDs, we performed computer modeling of ventricular myocardial gradients resulting from TV-ICD and S-ICD shocks in various configurations.

| METHODS
The numeric data that support the findings of this study are available from the corresponding author upon reasonable request. The computer model was built from thoracic MRI images of a 63-year-old male with ischemic disease. The source deidentified MRI data set was licensed from an academic research center (Bakken Medical Instrumentation and Device Laboratory, University of Minnesota), so no ethical approval was required. The computer model and the methods for defibrillation simulation have been described previously, 6 though now an upgraded simulation software is used (Comsol version 5.6). To summarize here, images were taken at end-diastole to simulate cardiac arrest, from neck to abdomen. Organs and tissues were segmented and converted to a finite element model with resistive properties that electrically represent the thorax. The original heart was segmented and then manually enlarged, to represent a dilated cardiomyopathy (DCM) heart. The left ventricular end-diastolic diameter (LVEDD) was 70 mm, and the right and left ventricular free wall thicknesses were 4 and 10 mm, respectively.  The energy levels used were those of maximum device output: 80 and 35 Joules for the subcutaneous and transvenous systems, respectively. Energy calculations assumed a conventional biphasic defibrillation waveform with 50% tilt (the percentage drop of the truncated exponential as it decays from its peak initial amplitude to its final one).
When solved numerically, the computer model contains the voltage gradient solution (i.e., the electric field) in units of Volts/cm, in the whole thorax. As a main outcome of this study, for the devices compared, we report the volume (in cubic cm) of ventricular myocardium that has a particular voltage gradient magnitude. For example, a certain device at maximum output energy may yield 10 cc of myocardium having more than 40 V/cm gradient. This gives an indication of the extent of electrical stress to which the myocardium is subjected. We defined a voltage gradient to be high when it was above 100 V/cm. This threshold was based on the work of Jones et al., 12 who subjected a culture of beating chicken myocardial cells to varying voltage gradients. They found reversible pauses, tachyarrhythmia, and contracture that started at about 80-120 V/cm, with recovery durations that were amplitude dependent.
With respect to data integrity, the first author (AB) had full access to all the data in the study and takes responsibility for its integrity and analysis. respectively). It can be seen that the myocardium in the S-ICD case has no gradients above 40 V/cm, while the transvenous systems have greater myocardial exposure, particularly the dual coil system (RV septal coil + SVC coil). Table 1 gives the results in tabular form. The dual coil system has 120 cubic mm of myocardium at a very high gradient of more than 400 V/cm (a gradient at which irreversible electroporation of myocardial tissue has been reported to occur), while there is no significant exposure at this level for the other TV-ICD configurations or for the S-ICD.

| DISCUSSION
TV-ICDs defibrillate the heart via a shock coil in the right ventricular (RV) cavity, in close proximity to the RV and septal myocardium. To produce an adequate electrical field spanning all regions of the LV and RV, the TV-ICD shock produces very high and potentially damaging electrical fields near the RV coil (often including the RV and F I G U R E 3 Volume of ventricular myocardium having at least the given voltage gradients. The transvenous systems have about 5-25 cc of myocardium above 40 V/cm, while the subcutaneous implantable cardioverterdefibrillator (S-ICD) has nearly zero. In other words, with an S-ICD, the myocardium has no gradients above 40 V/cm.

C E N T R A L I L L U S T R A T I O N
Voltage gradients in the heart, at mid-height level. Cutaway view with ventricular blood removed. Coils with close proximity to the myocardium yield higher gradients. The dual coil transvenous system (TV septal+SVC) had the highest exposure, with a concentration in the basal septal areas near the right atrium. The papillary muscle in the RV is subject to high gradients from coils in the RV. ventricular septum), with progressively lower gradients further from the coil. In contrast, the S-ICD coil is much further from the closest myocardium: this results in a generally higher absolute energy defibrillation threshold than a TV-ICD, but also with much more uniform electrical fields for the S-ICD spanning the RV and LV, with much less local myocardial exposure to very high gradients. However, the S-ICD exposes noncardiac tissue in close proximity to the coil to high electrical gradients. This is consistent with a study in swine, which demonstrated much higher Creatinine Kinase Muscle Isoenzyme (CK-MM) elevation (skeletal muscle), but much lower Troponin I elevation (cardiac muscle), from S-ICD compared to TV-ICD shocks. 5 The modeling results presented here clearly show that coil proximity to myocardial tissue is related to higher local voltage gradients. When the ICD coil electrode was not close to the ventricular myocardium, either by a mid-cavity TV-ICD or much more distant with an S-ICD, high gradients ( > 100 V/cm) in the myocardium were markedly reduced. T A B L E 1 Volume of myocardium above gradient (cc).  F I G U R E 4 Ventricular myocardium subject to more than 30 V/ cm (left column), and 100 V/cm (right column), for the four configurations tested. Only tissue above 30 V/cm (left) and 100 V/cm (right) are shown in color. For transvenous systems, the septum bears higher gradients, while for the SICD, it is the anterior RV-free wall. SICD, subcutaneous implantable cardioverter-defibrillator.

| Dual coils and high gradients
Dual coil TV-ICD leads are implanted less often than single coil TV-ICD leads in the current era, but dual coil leads were used more frequently in the past when published studies documenting worse clinical outcomes after TV-ICD shocks (both appropriate and inappropriate shocks) were undertaken. [13][14][15] Many patients still have dual coil leads in use predominantly related to earlier implants, although dual coil TV-ICD implants are still performed. A metaanalysis reported lower DFTs, but higher all-cause mortality, with dual coil compared to single coil TV-ICDs 16 ; both findings could be explained by the higher myocardial voltage gradients with dual vs single coil TV-ICDs modeled in our study. The dual coil TV system including an SVC coil demonstrated particularly high voltage gradients in our modeling. A principal cause for this, (besides the proximity of the distal coil to the septal myocardium), is the manner by which an ICD shock waveform adapts to lower impedances. In our models, adding an SVC coil reduced the impedance of the TV system from about 61 to 36 ohms (with corresponding higher myocardial voltage gradients), effects that are consistent with clinical reports. 17,18 Since the ICD voltage per prescribed energy dose is the same, a higher current results from the lower impedance, per Ohm's law, (current = voltage/resistance). The injection of such elevated current causes higher voltage gradients in the tissues.
Further details on the topic of ICD waveform response to impedance variation,-with tilt and energy constancy as dose controllers,-are critically reviewed by Kroll et al. 19

| High gradients and myocardial damage
Our findings of higher voltage gradients with TV-ICDs compared with S-ICDs are consistent with studies of cardiac biomarkers comparing both kinds of devices. 3,4 Beyond the reversible effects that Jones et al. 12 described at gradients of about 100 V/cm, a possible mechanism for mortality with TV-ICD shocks is that stunned or blocked myocardium regions -even if reversible-could facilitate tachyarrhythmic reentrant circuits, just like infarct scar does. Furthermore, it is possible that pulseless electrical activity (PEA) could be related to myocardial dysfunction induced by high voltage gradients even if the gradients are not high enough to cause irreversible damage. A study that examined recordings from 320 patient deaths showed that the most common mechanism of sudden death with a TV-ICD present was VT/VF treated with an appropriate shock followed by electromechanical dissociation. 20 Regarding higher gradient magnitudes, 400 V/cm is typically cited as a threshold for consistent irreversible damage to myocardium. 21,22 In our study, the dual coil system with a septal coil position in the RV had 120 cubic mm of myocardium exposed to more than 400 V/cm, while the other tested device configurations had no such very high gradient exposure. The clinical significance of such small volume of irreversibly damaged tissue is presumably dependent on its location relative to important conduction structures, or perhaps on the creation of additional scar or electrically impaired regions. There is likely to be progressively more irreversible damage as myocardial F I G U R E 5 Voltage gradients illustrated at 10 Volts per line, at mid-heart level, at maximum output. The transvenous systems have a high concentration of isopotential lines in the heart near the RV coil, while the S-ICD has a smaller and more uniform gradient over the myocardium. RV, right ventricle; SICD, subcutaneous implantable cardioverter-defibrillator. gradients increase from the range of 100 V/cm to 400 V/cm, and our modeling demonstrates that gradients in this range are likely to occur in much larger regions of myocardium with TV-ICD shocks compared to S-ICD shocks.

| CLINICAL IMPLICATIONS
Most studies comparing TV-ICDs and S-ICDS have focused on rates of procedural and lead-related complications as well as rates of unsuccessful and inappropriate shocks. There is also the possibility, however, that harm from ICD shocks could vary between TV-ICDs and S-ICDs, with this computer modeling study demonstrating much higher maximal myocardial voltage exposures with TV-ICDs. This is consistent with animal studies demonstrating greater cardiac enzyme elevation after TV-ICD compared to S-ICD shocks. 4,5 This raises the intriguing possibility that S-ICD shocks, as a result of their more homogenous electrical fields across the myocardium, may be inherently less cardiotoxic than TV-ICD shocks. Future clinical studies will clearly be needed to substantiate these findings.

| LIMITATIONS
This is a computer modeling study using tissue properties obtained from prior basic science studies. Ours is not a study of actual tissue damage (e.g., using histology and pathology observations from clinical or preclinical experiments), so we present no actual measured data.
Nevertheless, computer models can be informative in guiding clinical practice, as our prior study of subcutaneous defibrillation efficacy suggested. 6 Voltage gradient thresholds for irreversible injury were taken as 400 V/cm, derived from in vitro studies 21 and conventional assumptions. That threshold may vary with device waveform, so actual volumes of in-vivo damaged myocardium can vary from those we report.
Variations in placement of S-ICD and TV-ICD leads and generators, and patient-specific anatomic differences, could affect the myocardial gradients resulting from ICD shocks compared to the typical DCM anatomy and lead/device placements modeled in this study.

| CONCLUSIONS
Our models suggest that S-ICD shocks produce more uniform voltage gradients in the myocardium, with less exposure to potentially damaging electrical fields, compared to TV-ICDs. Dual coil TV leads yield higher gradients, as does closer proximity of the shock coil to myocardium. These observations are consistent with clinical and animal studies, and may represent an important difference between TV-ICD and S-ICD shocks.

ACKNOWLEDGMENTS
The study was funded by Boston Scientific Corp.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.