Additional Lead Improves Defibrillation Efficacy With an Abdominal ‘Hot Can’ Electrode System
Background Although the left prepectoral site is preferred for “hot can” placement, this site is unavailable in some patients. We evaluated the influence of electrode location on defibrillation thresholds with alternative hot can and transvenous lead configurations.
Methods and Results Three interrelated studies were performed. In group 1, the importance of hot can location was investigated by pairing a right ventricular lead to five different hot can placement sites in seven pigs. The defibrillation energies for right pectoral, left pectoral, left subaxillary, and right and left abdominal hot can sites were 20.3±2.7,* 15.9±3.8, 14.9±2.5, 32.0±3.4,* and 30.0±3.4 J,* respectively (*P<.005 versus left pectoral and left subaxillary sites). In group 2, the value of a three-electrode configuration with an abdominal hot can placement was investigated by adding a subclavian vein lead to the pectoral or abdominal hot can configurations in seven pigs. The defibrillation energies for left pectoral and abdominal sites were 18.6±4.2 and 29.0±5.8 J (P=.0001), respectively. The addition of a right or left subclavian vein lead with an abdominal hot can reduced the threshold to 19.3±4.2* or 18.8±3.2,* respectively (*P=.0001 versus abdominal site). In group 3, the contribution of the abdominal hot can electrode to the three-electrode configuration was tested by a comparison with two purely transvenous two-electrode configurations in six pigs. The defibrillation energy (19.9±3.2 J) for the abdominal hot can with a subclavian vein lead was lower than the transvenous lead configurations with a subclavian vein (29.0±2.5 J, P=.0001) or a superior vena cava lead (30.7±3.7 J, P=.0001). The right ventricular lead was the sole cathode during the first phase of the biphasic shock in all experiments.
Conclusions Defibrillation energy depends on the hot can placement site. The addition of a subclavian vein lead with an abdominal hot can improves defibrillation efficacy to the level of the pectoral placement and is better than a purely transvenous lead configuration.
The ICD has documented its clinical efficacy in the prevention of sudden cardiac death.1–6 Transvenous lead configurations have been developed to simplify implantation of defibrillators and now constitute accepted therapy for malignant ventricular arrhythmias.7–12 “Hot can” systems with shocks delivered between a single RV electrode and the ICD can produce a reduction in DFTs.13–15 In most clinical cases, the implantation site of the hot can electrode is the left pectoral area.13–16 In some cases, however, such as when a patient has a pacemaker implanted in the left pectoral area or if there is left subclavian vein thrombosis or obstruction, it is impossible or inconvenient to implant the hot can electrode at such a site.17 Although the exact left pectoral location of the hot can electrode affects defibrillation efficacy,18,19 alternative sites to left pectoral implantation have not been systematically explored. In addition, the optimal defibrillation lead configuration and location are unknown when a hot can electrode is implanted in an abdominal location.
The purpose of this study was (1) in group 1, to evaluate the best geometry for defibrillation with a hot can system using a single RV lead; (2) in group 2, to determine whether the addition of a transvenous lead located in the subclavian vein improves DFTs when an abdominal hot can implantation site is used; and (3) in group 3, to measure the contribution of the abdominal hot can electrode to the reduction of DFTs in a three-electrode configuration.
Experimental Preparation and Instrumentation
Pigs were anesthetized with sodium pentobarbital (30 mg/kg IV), ketamine (20 mg/kg IM), and morphine (2 mg/kg IM); intubated with a cuffed endotracheal tube; and ventilated with room air and oxygen though a Drager SAV (North American Drager). Infusion of sodium pentobarbital (1 to 2 mg/kg) and pancuronium bromide (0.1 to 0.2 mg/kg) was repeated as necessary to maintain anesthesia and muscle relaxation, respectively. An intra-arterial line was inserted via a cutdown to the left carotid artery for continuous systemic blood pressure monitoring as well as blood gas sampling. An intravenous line was also inserted via a cutdown in the left internal jugular vein for continuous infusion of normal saline solution. Surface ECGs in four leads (I, II, aVR, and V2) and the intracardiac electrogram from the RV apex were displayed continuously on a computer monitor and stored on a computer-based, physiological recording system (EP Laboratory, Quinton Electrophysiology Corp). Arterial blood samples were drawn every 30 to 60 minutes to determine pH; Po2; Pco2; base excess; total CO2 content; and bicarbonate, sodium, potassium, and calcium concentrations. Body temperature was continuously monitored with a rectal thermistor and maintained between 36.0°C and 37.0°C by heating the table with a warm circulating water blanket (Sarns Inc). A unipolar defibrillation lead (model 497, Intermedics Inc) with a 5-cm-long electrode with an 8.3F diameter (surface area of 4.4 cm2) was inserted via cutdown of the right internal jugular vein and advanced into the RV apex under fluoroscopy.
The external waveform amplifier and defibrillator (Angeion Research Defibrillation System, ARD-9000, Angeion Corp) was used. The ARD-9000 operates as a high-voltage, linear amplifier using software-generated waveforms. The device samples the current and voltage every 0.1 ms and adjusts the waveform voltage to mimic a capacitive discharge. The continuous voltage adjustment responds to the impedance changes that occur as a function of voltage changing during a discharge.20 Preliminary tests in a saline bath demonstrated that the voltage waveforms generated by this device have <3% variation from waveforms generated by a true capacitor discharge in the range of 100 to 750 V. The ARD-9000 continuously calculates the instantaneous impedance by dividing the delivered voltage by the current. The median of all these values is then used as the overall shock impedance. The ARD-9000 measures the current and voltage every 0.1 ms. The average current was calculated as the sum of all sampled current values divided by the number of samples.
VF was induced with 60-Hz alternating current (15 V) for 4 seconds through the RV defibrillation lead. After sustained VF lasting 10 seconds from initiation of the alternating current, a biphasic defibrillation waveform (135-μF capacitance, 65% tilt in each phase) was delivered. If VF was not terminated by this waveform, a rescue shock (monophasic 5-ms square waveform, 20 to 50 J) was delivered to terminate the VF. A recovery period of at least 3 minutes was allowed between each episode of VF. VF was not reinitiated until the heart rate and blood pressure stabilized and returned to preshock levels.
Determination of DFT
Threshold testing randomized the sequence of the implantation sites (group 1) and defibrillation lead configurations (groups 2 and 3). The location of the RV lead was fixed after the initial lead configuration was validated. This required a single shock success at ≤35 J for the abdominal hot can–to–RV lead configuration or the purely transvenous lead configuration and at ≤25 J for all other lead configurations. After validation, DFT was determined by a “down-up, down-up” technique21,22 until three reversals of defibrillation success were completed. After the initial successful shock, the stored energy was decremented by 1 J after each successful defibrillation. When a shock failed to defibrillate, the next trial was performed by incrementing the stored energy by 1 J. This process was repeated until three reversals in decrement or increment occurred. The final shock was always a success after the last failure. The DFT was defined as the average of the shock energy, voltage, or ampere values obtained with all trials starting from the successful shock before the first defibrillation failure until the last successful defibrillation.
After determination of the DFT for the lead configurations, preparation stability was ensured by repeating the DFT measurements for the first of the randomized lead configurations tested in that animal. If the stored energy at DFT was within 2 J of the first DFT, then the preparation was deemed stable, and the experiment was included in the study. If there was a change of >2 J between the first DFT energy and this last DFT energy, the preparation was not considered stable, and the data were rejected.
The use of experimental animals in this study was approved by the Animal Research Committee of the Cleveland Clinic Foundation and conformed to the position of the American Heart Association on Research Animal Use.
The mean and SD of all parameters were calculated. Repeated-measures one-way ANOVA was used to compare stored energy, delivered energy, peak voltage, peak current, average current, median impedance, and pulse width among the lead systems. Pairwise comparisons of each defibrillation parameter were performed by the method of contrasts (Fisher’s least significant difference test)23 for the post hoc analysis. The null hypothesis was rejected for P<.05.
Group 1 Defibrillation Protocol
This protocol tested the impact on DFTs of the placement of the 55-cm3 titanium hot can electrode (Angeion Corp) with a conductive surface area of 92 cm2 implanted in the five different test locations (Fig 1⇓). The RV defibrillation lead and the hot can electrode were connected to the ARD-9000 (Angeion Corp). DFT parameters were evaluated at five different hot can electrode implantation sites: right pectoral, left pectoral, left subaxillary, right upper abdominal, and left upper abdominal, in random order in each pig (detailed in Fig 1⇓).
Group 1 Results
Complete DFT data sets were obtained from seven pigs (34±3 kg). The DFT parameters and pulse width for each hot can implantation site are given in detail in Table 1⇓.
Fig 2⇓ shows the stored energy at DFT in five different hot can implantation sites. The stored energies at DFT for the abdominal implantation sites were significantly higher than for the pectoral implantation sites (P=.0001 for each pectoral site versus each abdominal site). There was no difference in stored energy between the right abdominal and left abdominal sites (P=.1588). Conversely, the stored energies at DFT for both the left pectoral and left subaxillary sites were ≈25% lower than for the right pectoral site and were statistically significant (P=.0042 and P=.0007 versus right site, respectively). There was no difference in stored energy between the left pectoral and left subaxillary sites (P=.4711). Fig 3⇓ shows the individual experimental stored energy data at DFT for the right and left pectoral sites. Except for one experiment, the stored energy at DFT for the left pectoral site was lower than for the right pectoral site.
The impedance at DFT for the abdominal implantation sites was significantly higher than for the pectoral implantation sites. In the abdominal implantation sites, there were no differences in the impedance between the right and left sites (P=.4376). Conversely, for the pectoral implantation sites, the impedance at DFT for the left pectoral site was lower than for the left subaxillary site (P=.0041).
Group 2 Defibrillation Protocol
This protocol tested the impact on DFTs of the addition of a lead in the right or left subclavian vein when an abdominal hot can electrode (Angeion Corp) implantation site was used. The hot can electrode (see group 1 for a complete description) was implanted in the left pectoral and the left upper abdominal sites. A unipolar defibrillation lead (model 497, Intermedics Inc) with a 5-cm-long electrode and an 8.3F diameter (surface area of 4.4 cm2) was advanced to either the right or left subclavian vein under fluoroscopy (Fig 4⇓). The RV apex, either the right or left subclavian vein lead, and the hot can electrode were connected to the ARD-9000. DFT parameters were evaluated with four different defibrillation lead configurations, which included control experiments with pectoral or abdominal hot can electrodes and the abdominal hot can electrode with a right or left subclavian vein lead in random order in each pig (detailed in Fig 4⇓).
Group 2 Results
Complete DFT data sets were obtained from seven pigs (26±2 kg). The DFT parameters and pulse width for each electrode configuration are given in detail in Table 2⇓.
The stored energy at DFT for the abdominal control configuration was ≈55% higher than the pectoral control configuration (P=.0001). However, the addition of a subclavian vein lead to the abdominal control configuration reduced the stored energy at DFT and performed as well as the pectoral control configuration (Fig 5⇓). There was no difference in stored energy between the right and left subclavian vein experiment configurations (P=.6784). Fig 6⇓ shows the individual experimental stored energy data at DFT for the pectoral control and left subclavian vein experiment configurations. There was no significant difference in stored energy between the two lead configurations (P=.887). Fig 7⇓ shows the stored energy data at DFT in individual experiments for the abdominal control and left subclavian vein experiment configurations. An additional left subclavian vein lead with the abdominal placement of the hot can electrode significantly reduced the stored energy at DFT in all experiments (P=.0001).
Although the impedance at DFT for the abdominal control was significantly higher than for the pectoral control (P=.0001), an additional subclavian vein lead with the abdominal control configuration reduced the impedance. There was no difference in impedance between the right subclavian vein experiment and left subclavian vein experiment configurations (P=.9842).
Group 3 Defibrillation Protocol
This protocol tested the value of using a hot can electrode for abdominal implantation. The hot can electrode (see group 1 for a complete description) was implanted in the left upper abdominal sites. A unipolar defibrillation lead (model 497, Intermedics Inc) with a 5-cm-long electrode and an 8.3F diameter (surface area of 4.4 cm2) was advanced to either the superior vena cava or left subclavian vein under fluoroscopy (Fig 8⇓). The RV apex and either the superior vena cava or left subclavian vein lead or the abdominal hot can electrode were connected to the ARD-9000. DFT parameters were evaluated at three different defibrillation lead configurations, which included the abdominal hot can electrode with a left subclavian vein lead configuration (three-electrode abdominal hot can) and two purely transvenous lead configurations (subclavian vein to RV and superior vena cava to RV) in random order in each pig (detailed in Fig 8⇓).
Group 3 Results
Complete DFT data sets were obtained from six pigs (38±6 kg). The DFT parameters and pulse width for each electrode configuration are given in detail in Table 3⇓.
Fig 9⇓ shows the mean stored energy at DFT in each defibrillation lead configuration. The stored energy at DFT for the three-electrode abdominal hot can configuration was 31% and 35% lower than the subclavian vein–to–RV lead (P=.0001) and the superior vena cava–to–RV lead configurations (P=.0001). There was no significant difference in stored energy between the two purely transvenous lead configurations (P=.2475). An additional subclavian vein lead with the abdominal placement of the hot can electrode configuration (three-electrode abdominal hot can) significantly reduced the stored energy at DFT in all experiments compared with both purely transvenous lead configurations (“cold can”).
The impedance at DFT for both cold can configurations using either the subclavian vein or superior vena cava was 66.6 or 53.4 Ω, respectively. However, the abdominal hot can three-electrode configuration reduced the impedance to 44.7 Ω (P=.0001).
The results of these experiments suggest that the implantation site of the hot can electrode plays a significant role in determining defibrillation efficacy. Although the abdominal implantation site of the hot can in such a lead configuration was inferior to pectorally placed ICDs, the addition of a subclavian vein lead with an abdominal implantation performed as well as the pectoral placement.
Previous studies using two epicardial patches have reported that the patch position that directed current through a large portion of the LV myocardium, particularly the interventricular septum, was predictive of defibrillation efficacy.24–27 In addition, with implantation of two transvenous leads as the defibrillation lead configuration, positioning the proximal defibrillation lead in the subclavian innominate vein decreased DFT energy requirements compared with positioning it in the superior vena cava.28 Similarly, with a single RV lead unipolar system, the location of the hot can electrode affects defibrillation efficacy.18,19 Thus, the electrode location plays an important role and influences defibrillation efficacy.
Although the preferred implantation site of the hot can electrode is the left pectoral area, it may be impossible or inconvenient to implant the hot can electrode in such an area in some cases. Schofield et al17 reported that DFT energies were satisfactory with a hot can electrode implanted in the right pectoral side even though the vector between the cathode and anode passed predominantly through the RV. In our study, the DFT energy for the right pectoral implantation site was 30% higher than the left pectoral site. In a preliminary study, Seidl et al18 examined the differences in biphasic shock DFT energy achieved at hot can ICD implantation using four different left pectoral locations, including high pectoral medial position, high pectoral lateral position, low pectoral medial position, and low pectoral lateral position. Their results suggested that the low pectoral medial position was best. In another preliminary study, Min et al19 investigated the effect of hot can position on DFT energy in a human thorax model by finite-element analysis. This model indicated that locating the hot can electrode beneath the clavicle in a submuscular position would lower DFT energy. In our study, there was no significant difference in DFT energy between the left pectoral and left subaxillary sites. However, three of seven pigs had lower DFT energies for the left subaxillary site than for the left pectoral site. In a nonthoracotomy subcutaneous patch electrode system, Saksena et al29 reported that the use of an axillary patch generated lower DFT energy compared with pectoral and apical locations. The mechanism by which the axillary site produces a lower DFT energy may be related to the axillary site vector, which is more perpendicular to the ventricular septum, thus passing more current through the LV septum.25
In addition, our study indicated that the DFT energy for the abdominal hot can sites was approximately twice that of the pectoral hot can sites. The abdominal hot can–to–RV lead configuration may reduce the current delivery through the myocardium because of an increased impedance and altered geometry. Thus, changing the hot can electrode position may change the DFT energy by altering the impedance and field distribution.
A left pectoral hot can electrode system incorporating a single transvenous RV lead is clinically acceptable in most patients.13–16 In a few reports, despite lowering pathway resistance, the addition of a third defibrillation lead in the superior vena cava30 or coronary sinus31 position did not generally affect the DFT energy of an RV lead–to–left pectoral hot can electrode configuration. In contrast, a few clinical reports showed that the addition of a superior vena cava lead to a hot can–RV lead configuration improved defibrillation efficacy with lower impedance.32,33 This was supported by finite-element analysis.33 Thus, previous studies investigating the defibrillation effect of an additional lead with the left pectoral implantation site of a hot can electrode have reported mixed findings.
In our study, however, the addition of a defibrillation lead in the right or left subclavian vein improved defibrillation efficacy for abdominal hot can implants. In fact, this additional lead decreased the stored energy at DFT by one third (Fig 5⇑) and lowered impedance by 28% compared with the hot can electrode alone. Furthermore, the DFT energy for an addition of a subclavian vein lead with an abdominal hot can–to–RV lead system was lower than those for a left pectoral hot can electrode–to–RV lead system in three (43%) of the seven pigs (Fig 6⇑). This increase in defibrillation efficacy with the addition of a subclavian vein lead is probably due to both an improved current vector through the LV and a reduction of shock impedance.
The impedance of the defibrillation pathway is an important determination of defibrillation efficacy.34–37 Reduction of impedance directly results in higher peak and average current emanating from the RV electrode at the same voltage. Shocks into high-impedance systems result in weaker current density fields than shocks into low-impedance systems for a constant voltage source. If the intramyocardial electrical field strength produced by the shock into the high-impedance system is above a critical minimum value,38,39 defibrillation may occur. Thus, the intramyocardial current field strength produced by the shock appears to be a critical parameter that determines whether a particular shock fails. Therefore, reductions in impedance should correspond to reductions in energy requirements for defibrillation. However, not all reductions of impedance necessarily result in reduced DFT. For example, impedance is reduced 20% when the left subclavian vein electrode is moved to the superior vena cava (see Table 3⇑). However, DFT energy is not affected. Thus, there are some methods of reducing impedance that help and others that do not help with defibrillation. If the heart can be viewed as a resistor, then reduction of impedance in series to the heart would improve the voltage gradient with the heart. However, reductions in impedance that are parallel to the heart would shunt current away from the heart and reduce the voltage gradient in the heart.
In patients in whom left pectoral implantation is impossible, it will often be possible to implant a hot can electrode in the right pectoral site. However, even though the size of defibrillators continues to decrease, it still precludes pectoral implantation for some patients. In these cases, the device must be implanted in an abdominal site. Therefore, the reduction in DFT energy with an additional lead may be important in such cases. Other reasons, such as previous infection, pacemakers, and physician choice, may require an increase in safety margins for patients undergoing abdominal implantation of a hot can electrode ICD.
The pig heart is similar anatomically to the human heart. Hence, the pig is commonly used by many investigators as a VF model. However, the anatomy of the abdomen and thorax in the pig is not exactly the same as that of the human. Therefore, the findings of this study may not be completely consistent with the clinical situation.
Although the addition of a subclavian vein lead with the abdominal hot can electrode placement can significantly increase defibrillation efficacy, the long lead tunneling procedure from the thoracic venous entry site to the abdominal hot can electrode pocket may be an aggravating factor for infection and lead complications.40
The major findings in this study are as follows. (1) The hot can electrode implantation site plays a important role in the determination of defibrillation efficacy. The DFT energies for the left pectoral and left subaxillary implantation sites were virtually the same. However, the DFT energy increased 30% for the right pectoral implantation and 100% for the abdominal implantation. (2) The addition of a subclavian vein lead with the abdominal hot can implantation improves defibrillation efficacy, making DFT energy levels comparable to that of the left pectoral implants. (3) Furthermore, this defibrillation lead configuration reduces DFT energy by 30% compared with a purely transvenous lead configuration (cold can).
Selected Abbreviations and Acronyms
|ICD||=||implantable cardioverter defibrillator|
This work was supported in part by a grant from the Angeion Corporation. We would like to thank Donald G. Hills for his technical assistance in the care of the animals before and during the experiments.
- Received May 27, 1997.
- Revision received August 25, 1997.
- Accepted September 12, 1997.
- Copyright © 1997 by American Heart Association
Echt DS, Armstrong K, Schmidt P, Oyer PE, Stinson EB, Winkle RA. Clinical experience, complications, and survival in 70 patients with the automatic implantable cardioverter/defibrillator. Circulation. 1985;71:289–296.
Kelly PA, Cannom DS, Garan H, Mirabal GS, Harthorne JW, Hurvitz RJ, Vlahakes GJ, Jacobs ML, Ilvento JP, Buckley MJ, Ruskin JN. The automatic implantable cardioverter-defibrillator: efficacy, complications and survival in patients with malignant ventricular arrhythmias. J Am Coll Cardiol. 1988;11:1278–1286.
Tchou PJ, Kadri N, Anderson J, Caceres JA, Jazayeri M, Akhtar M. Automatic implantable cardioverter defibrillators and survival of patients with left ventricular dysfunction and malignant ventricular arrhythmias. Ann Intern Med. 1988;109:529–534.
Yee R, Klein GJ, Leitch JW, Guiraudon GM, Guiraudon CM, Jones DL, Norris C. A permanent transvenous lead system for an implantable pacemaker cardioverter-defibrillator: nonthoracotomy approach to implantation. Circulation. 1992;85:196–204.
Sweeney M, Ruskin JN. Mortality benefits and the implantable cardioverter-defibrillator. Circulation. 1994;89:1851–1858.
Bardy GH, Johnson G, Poole JE, Dolack GL, Kudenchuk PJ, Kelso D, Mitchell R, Mehra R, Hofer B. A simplified, single-lead unipolar transvenous cardioversion-defibrillation system. Circulation. 1993;88:543–547.
Schofield IJ, Bennett DH. Right sided pectoral implantation of an ‘active can’ transvenous implantable cardioverter-defibrillator with single right ventricular lead. Br Heart J. 1995;74:204.
Seidl K, Hauer B, Schwick N, Werling C, Isgro F, Senges J. Is the pectoral implantation site of the ‘active can’ cardioverter/defibrillator device of importance for the defibrillation threshold? J Am Coll Cardiol. 1996;27:348A. Abstract.
Min X, Mehra R. Finite element analysis of effect of can electrode position for Can-RV. Pacing Clin Electrophysiol. 1996;19:655. Abstract.
Yamanouchi Y, Mowrey KA, Nadzam GR, Hills DG, Kroll MW, Brewer JE, Donohoo AM, Wilkoff BL, Tchou PJ. Large change in voltage at phase reversal improves biphasic defibrillation thresholds: parallel-series mode switching. Circulation. 1996;94:1768–1773.
Kirk RE. A priori nonorthogonal contrasts. In: Kirk RE ed. Experimental Design: Procedures for the Behavioral Sciences. Belmont, Calif: Brooks/Cole Publishing Co; 1982:106–110.
Saksena S, DeGroot P, Krol RB, Raju R, Mathew P, Mehra R. Low-energy endocardial defibrillation using an axillary or a pectoral thoracic electrode location. Circulation. 1993;88:2655–2660.
Bardy GH, Dolack GL, Kudenchuk PJ, Poole JE, Mehra R, Johnson G. Prospective, randomized comparison in humans of a unipolar defibrillation system with that using an additional superior vena cava electrode. Circulation. 1994;89:1090–1093.
Kudenchuk PJ, Bardy GH, Dolack GL, Poole JE, Mehra R, Johnson G. Efficacy of a single-lead unipolar transvenous defibrillator compared with a system employing an additional coronary sinus electrode: a prospective, randomized study. Circulation. 1994;89:2641–2644.
Jung W, Frazier DW, Saksena S, Stanton MS, Thakur RK, DeGroot PJ. Paired randomized comparison of defibrillation threshold with Can-RV and Can+SVC-RV lead systems. Circulation. 1996;94(suppl I):I-438. Abstract.
Min X, DeGroot PJ, Mehra R. Finite element analysis of Can+SVC and Can-RV electrode system for ventricular defibrillation and comparisons with the clinical data. Circulation. 1996;94(suppl I):I-439. Abstract.
Kerber RE, Kouba C, Martins J, Kelly K, Low R, Hoyt R, Ferguson D, Bailey L, Bennett P, Charbonnier F. Advance prediction of transthoracic impedance in human defibrillation and cardioversion: importance of impedance in determining the success of low-energy shocks. Circulation. 1984;70:303–308.
Chen PS, Wolf PD, Claydon FJ, Dixon EG, Vidaillet HJ Jr, Danieley ND, Pilkington TC, Ideker RE. The potential gradient field created by epicardial defibrillation electrodes in dogs. Circulation. 1986;74:626–636.
Lerman BB, Halperin HR, Tsitlik JE, Brin K, Clark CW, Deale OC. Relationship between canine transthoracic impedance and defibrillation threshold: evidence for current-based defibrillation. J Clin Invest. 1987;80:797–803.
Kerber RE, Martins JB, Kienzle MG, Constantin L, Olshansky B, Hopson R, Charbonnier F. Energy, current, and success in defibrillation and cardioversion: clinical studies using an automated impedance-based method of energy adjustment. Circulation. 1988;77:1038–1046.
Wharton JM, Wolf PD, Smith WM, Chen PS, Frazier DW, Yabe S, Danieley N, Ideker RE. Cardiac potential and potential gradient fields generated by single, combined, and sequential shocks during ventricular defibrillation. Circulation. 1992;85:1510–1523.
Zhou X, Daubert JP, Wolf PD, Smith WM, Ideker RE. Epicardial mapping of ventricular defibrillation with monophasic and biphasic shocks in dogs. Circ Res. 1993;72:145–160.