Skip main navigation

Reduction of Atrial Defibrillation Threshold With an Interatrial Septal Electrode

Originally published 2000;102:2659–2664


    Background—The standard lead configuration for internal atrial defibrillation consists of a shock between electrodes in the right atrial appendage (RAA) and coronary sinus (CS). We tested the hypothesis that the atrial defibrillation threshold (ADFT) of this RAA→CS configuration would be lowered with use of an additional electrode at the atrial septum (SP).

    Methods and Results—Sustained atrial fibrillation was induced in 8 closed-chest sheep with burst pacing and continuous pericardial infusion of acetyl-β-methylcholine. Defibrillation electrodes were situated in the RAA, CS, pulmonary artery (PA), low right atrium (LRA), and across the SP. ADFTs of RAA→CS and 4 other lead configurations were determined in random order by use of a multiple-reversal protocol. Biphasic waveforms of 3/1-ms duration were used for all single and sequential shocks. The ADFT delivered energies for the single-shock configurations were 1.27±0.67 J for RAA→CS and 0.86±0.59 J for RAA+CS→SP; the ADFTs for the sequential-shock configurations were 0.39±0.18 J for RAA→SP/CS→SP, 1.16±0.72 J for CS→SP/RAA→SP, and 0.68±0.46 J for RAA→CS/LRA→PA. Except for CS→SP/RAA→SP versus RAA→CS and RAA→CS/LRA→PA versus RAA+CS→SP, the ADFT delivered energies of all of the configurations were significantly different from each other (P<0.05).

    Conclusions—The ADFT of the standard RAA→CS configuration is markedly reduced with an additional electrode at the atrial SP.

    Electrical cardioversion remains a first-line therapy for atrial fibrillation (AF).12 Antiarrhythmic drugs and curative surgical and ablative therapies are still of limited efficacy, and their incidence of adverse effects is not insignificant.345 Because cardioversion with intrathoracic electrodes is more efficacious than transthoracic defibrillation6 and the implantable atrial defibrillator has been shown to be feasible,7 many recent investigations have focused on minimizing the energy requirements of internal cardioversion through novel waveforms or electrode configurations.89101112 Still, the most efficient cardioversion strategy clinically used, with biphasic waveforms and lead systems in which shock fields encompass both atria, requires shock strengths that are not sufficiently low to be tolerated by most patients without sedation.12131415

    Atrial defibrillation may require a minimum potential gradient to be generated by the shock throughout the atrial myocardium, as does ventricular defibrillation throughout the ventricles. The lowest-gradient areas, usually distant from the defibrillation electrodes, are the areas from which earliest activations arise after unsuccessful shocks.1617 The region of earliest activation, and presumably lowest gradient, with a right atrial appendage–to–coronary sinus defibrillation configuration (RAA→CS) is in the posterior left atrium.18 An electrode configuration with an interatrial septal (SP) electrode approximately midway between the RAA and CS electrodes should increase the potential gradient in this region. The purpose of this study was to determine whether the atrial defibrillation threshold (ADFT) could be reduced by use of configurations with an atrial SP electrode.


    The use of experimental animals in this study was approved by the Institutional Animal Care and Use Committee at the University of Alabama at Birmingham. All studies were performed in accordance with the guidelines established in the “Position of the American Heart Association on Research Animal Use” adopted by the American Heart Association on November 11, 1984 (Circulation. April 1985).

    Of 11 adult sheep, 8 completed the experimental protocol (body mass 41±6 kg, heart mass 217±8 g). Only the data from these 8 animals are reported.

    Animal Preparation

    As a preanesthetic agent, a 1-to-1 mixture of tiletamine and zolazepam (8 to 10 mg/kg) was given intramuscularly. Approximately 10 minutes later, thiopental (2 to 6 mg/kg) was administered as a slow intravenous bolus. The animal was laid on its back on a fluoroscopy table, intubated, and placed on a volume-cycled ventilator (tidal volume 15 to 20 mL/kg) with a 4% isoflurane/oxygen mixture at a rate of 8 to 12 breaths per minute. The isoflurane concentration was later decreased to 1.5% to 3.5% to maintain a deep surgical plane of anesthesia. Ventilator settings were adjusted to maintain blood gases within normal ranges. Lactated Ringer’s solution was continuously infused with supplemental electrolytes as needed as determined by serial blood gas and chemistry analyses every 30 to 60 minutes.

    An 8F sheath was placed in the left femoral artery percutaneously for continuous arterial pressure monitoring. The animal was instrumented for lead II ECG and esophageal temperature monitoring. A heated water blanket was used to maintain body temperature at ≈37°C. Neuromuscular blockade was achieved with a 1-mg/kg succinylcholine chloride intravenous bolus followed by an intravenous drip (5 to 8 mg/min) for maintenance, depending on neuromuscular tone.

    Defibrillation Catheter Placement

    All catheters were positioned transvenously under fluoroscopic guidance. Through a jugular vein, a defibrillation lead (Perimeter 7109, Guidant Corp) with a 6-cm-long coil electrode (Figure 1) was situated in the distal CS with its tip under the left atrial appendage (Figure 2). Care was taken not to place this lead in the persistent superior vena cava, which is present in this species. A modified quadripolar catheter (Mansfield EP-Boston Scientific Corp) with a 4-cm-long coil electrode 1 cm proximal to the catheter tip (Figure 1) was positioned in the RAA through the left femoral vein (Figure 2). The bipolar tip of this catheter was used for burst pacing to induce AF. Two other catheters, each with a 4-cm-long coil electrode, were placed in the pulmonary artery (PA) and lower right atrium (LRA). The PA electrode was positioned with ≈50% of the electrode in the main PA and half in the left PA. The LRA electrode was positioned with ≈50% of the electrode in the LRA and half in the inferior vena cava.

    A custom-made 6-cm coil electrode with its distal end 3 cm from the catheter tip (Figure 1) served as the interatrial SP electrode. It was situated through a transseptal procedure. An 8F Mullins sheath was advanced into the right atrium through the right femoral vein over a 0.038-in guidewire. Then, a Brockenbrough needle replaced the guidewire and was displaced through the atrial SP, usually under left anterior oblique projection. Left atrial catheterization was confirmed by measurement of oxygen saturation of blood withdrawn through the needle and with contrast injection into the left atrium during fluoroscopy. A stiff 0.038-in guidewire was positioned in the left atrium through the sheath, and the Mullins sheath was withdrawn. Next, an 11F guide sheath was advanced over the wire into the left atrium. After withdrawal of the dilator and guidewire, the SP electrode was inserted into the left atrium through the guide sheath. The tip of the SP electrode catheter was placed against the lateral wall of the left atrial appendage. Approximately two thirds of the electrode was in the left atrium and one third in the right atrium (Figure 2). After the SP electrode was placed, 1000 U of heparin was given intravenously every hour.

    A catheter (Endotak DSP, Guidant Corp) was inserted through the other jugular vein with its defibrillation electrodes in the right ventricle and superior vena cava, and a tip electrode in the right ventricle for ventricular pacing. A 3-wire subcutaneous array was inserted over the left side of the heart. In the event of ventricular tachyarrhythmias, these electrodes were used to rescue the animal.

    Induction of Atrial Fibrillation

    To allow AF to be maintained, acetyl-β-methylcholine chloride (Sigma Chemical Co) was continuously infused into the pericardial space. The pericardial space was approached percutaneously under fluoroscopic guidance with a 3-in-long 16-gauge needle from just inferior to the right subxiphoid position with the animal turned ≈20° toward the right side. When the needle was confirmed by contrast injection to be within the pericardial space, a guidewire was gently inserted through it into the pericardial space. The pericardial location for the guidewire was confirmed by inability to move it outside the fluoroscopic image of the heart silhouette. After removal of the needle, a 6F sheath was advanced into the pericardial space over the wire. After flushing with acetyl-β-methylcholine chloride, a 4F pigtail catheter was inserted through the sheath. Typically, the catheter tip was advanced near the left atrium and the sheath was removed.

    Acetyl-β-methylcholine chloride solution (1 g/250 mL saline) was infused at a rate of 20 μL/min with a microinfuser.1920 Burst pacing used to induce AF consisted of 2-ms stimuli delivered at intervals of 30 to 80 ms. AF was defined as irregular rapid atrial activity with an irregular ventricular response on the surface ECG. Blood pressure and heart rate were recorded before and 20 minutes after acetyl-β-methylcholine chloride infusion.

    Defibrillation Waveforms and Lead Configurations

    Once AF was maintained for >10 minutes, the defibrillation protocol was begun. Briefly, a monophasic, truncated-exponential waveform was produced by a programmable defibrillator (HVS-02, Ventritex, Inc) with a discharge capacitance of 150 μF. This monophasic waveform was divided into a biphasic waveform by a set of high-voltage, cross-point switches; for sequential shocks, 2 biphasic, truncated-exponential waveforms were created with the use of an additional set of cross-point switches.9 Each biphasic waveform had a first-phase duration of 3 ms and a second-phase duration of 1 ms. The interval between each phase of the biphasic waveforms and between the 2 biphasic waveforms of sequential shocks was 0.02 ms.9 All phases of the waveforms exhibited decaying voltage from a single capacitor, in which the trailing-edge voltage of each preceding phase was equal to the leading-edge voltage of the succeeding phase.

    In each animal, the ADFTs of 5 test configurations (Table 1) were determined. Three configurations that used the SP electrode were named A1, A2, and A3. The 2 others were named B and C. The order of determining the ADFTs was randomized as follows. The order among A, B, and C was initially randomized, and then the order of A1, A2, and A3 (within A) was randomized. ADFTs of the SP electrode configurations were measured consecutively to obviate the need for repositioning this electrode. During ADFT testing, passive electrodes not delivering any shock for that configuration were removed. Shock delivery was synchronized to right ventricular pacing at a cycle length of 250 to 400 ms, depending on the intrinsic ventricular rate during AF. Shocks were delivered 20 ms after the eighth pacing pulse.

    Each ADFT was determined by a multiple-reversal method with an initial commended peak voltage of 100 V and step sizes of 40/20/10 V.9 If the initial shock failed, the next and subsequent shock voltages were increased by 40 V until a shock succeeded. After the first shock that successfully terminated AF, voltages of subsequent shocks were decreased by 20 V until a shock failed. Then shock voltages were increased by 10 V until a shock succeeded again. Conversely, if the initial shock succeeded, subsequent shock voltages were decreased by 40 V until a shock failed. Then shock voltages were increased by 20 V until a shock succeeded, after which shock voltages were decreased by 10 V until a shock failed again. The last successful shock of the third reversal was deemed the ADFT shock for that configuration. The ADFT characteristics for each configuration were derived from this ADFT shock. All test shocks were applied after AF was sustained for 1 minute. When a test shock failed, a rescue shock of 200 to 300 V was given. A 1- to 2-minute period of sinus rhythm was allowed before the next AF induction.

    Data Acquisition

    For each test shock, the leading-edge voltage and current were recorded, and the impedance and delivered energy were computed by a waveform analyzer (DATA 6100, Data Precision).9 The stored energy for each shock was calculated as CV2/2, in which C is the discharge capacitance (150×106 F) and V is the commended peak voltage of the shock. The ADFT leading-edge voltage, current, impedance, and delivered and stored energies were derived from the ADFT shock.

    Postmortem Examination

    After completion of the experimental protocol, euthanasia was induced with an intravenous bolus of potassium chloride. The chest was opened, and the location of the electrodes of the last test configuration was confirmed by palpation through the heart walls. The heart was then removed and weighed.

    Statistical Analysis

    Results are expressed as mean±SD. The overall effects of the 5 test configurations on each ADFT characteristic and the impedance differences of current pathways were tested by repeated-measures ANOVA. Differences between the 5 configurations were tested by Duncan’s multiple range test.21 A value of P<0.05 was considered significant.


    Reproducible, sustained AF was induced in all animals. The ventricular rates in sinus rhythm before and 20 minutes after administration of acetyl-β-methylcholine chloride were similar (112±8 versus 109±11 bpm, respectively, P=NS). The drug significantly lowered the sinus-rhythm systolic and diastolic blood pressure, however (107±8 versus 84±7 and 83±4 versus 62±5 mm Hg, respectively). During AF, the ventricular rate was 131±36 (range 80 to 178) bpm. In 1 sheep with an unusually long sinus recovery time, temporary postshock atrial pacing was performed.

    ADFT Leading-Edge Voltages

    The ADFT leading-edge voltage of configuration A2 was significantly lower than each of the other 4 configurations (Figure 3). The ADFT leading-edge voltage of configuration A1 was significantly lower than that of configurations A3 and B, but not C. The ADFT leading-edge voltage of configuration B was significantly higher than each of the other 4 configurations.

    ADFT Leading-Edge Currents

    The ADFT leading-edge current of configuration A2 was significantly lower than that of A1, A3, and B and trended lower than that of C (Figure 4). The ADFT leading-edge current of configuration C was significantly lower than that of configurations A1, A3, and B. The ADFT leading-edge current of configuration A1 was significantly higher than each of the other configurations.

    Configuration Impedances

    Impedances of the same current pathway in different test configurations were not significantly different. The impedance of the RAA+CS→SP current pathway was lower than that of all of the other pathways tested (Table 2). The impedance of CS→SP was significantly lower than that of RAA→SP, and the impedances of CS→SP and RAA→SP were both significantly lower than that of RAA→CS. The impedance of LRA→PA was significantly lower than that of RAA→CS.

    ADFT Shock Energy

    Configuration A2 had a significantly lower ADFT delivered energy than each of the other 4 test configurations (Figure 5). Configuration C had significantly lower ADFT energy than that of configurations A3 and B and trended lower than that of A1. The ADFT delivered energy of configuration C was 50±9% lower than that of configuration B. Configuration A1 had significantly lower ADFT energy than configurations A3 and B (by 18±52% and 37±15%, respectively). The difference in ADFT delivered energy between configurations A3 and B was not significant. The ADFT delivered energies for each animal are shown in Figure 5B. On each animal, the ADFT delivered energy of configuration A2 was lower than that of configuration B. The configuration A2 ADFT delivered energies of all 8 animals were lower than the configuration B mean ADFT delivered energy, as well.

    The ADFT stored energies of configurations A1, A2, A3, B, and C were 1.32±0.88, 0.55±0.24, 1.66±0.94, 2.42±1.18, and 0.98±0.61 J, respectively. Except for A1 versus A3, A1 versus C, and A2 versus C, the ADFT stored energies for all of the test configurations were significantly different for each other.


    This study demonstrates that, in a sheep model of sustained AF, atrial defibrillation configurations that used an additional electrode at the interatrial SP required significantly lower energy than the standard RAA→CS configuration. The ADFT delivered energy of RAA+CS→SP was 0.86±0.59 J, 37±15% lower than that of RAA→CS (1.27±0.67 J), and the ADFT delivered energy of the sequential-shock configuration RAA→SP/CS→SP was 0.39±0.18 J, 68±8% lower than that of RAA→CS.

    Previous experimental and clinical studies found the most efficient simple single-shock configuration to be RAA→CS.2223 This is the most common configuration for in-hospital internal cardioversion and is the standard configuration used by the first stand-alone implantable atrial defibrillator.7 With this RAA→CS configuration, ADFTs in humans without a significant history of AF are 1.5 to 2.5 J and are significantly greater in patients with chronic AF.124252627 Unfortunately, shocks of this magnitude are uncomfortable to unsedated patients; although a “discomfort threshold” has not been clearly defined with this or any other shock configuration, it appears that shocks need to be well below 1 J to be well tolerated in most patients.12131415

    Cooper and colleagues9 have reported ADFTs <1 J in a sheep model using a sequential, 2-current-pathway configuration. In this study, the RAA→distal CS/PA→proximal CS ADFT (0.36±0.13 J) was significantly lower than that of the standard single-shock RAA→distal CS configuration (1.29±0.26 J). In humans, a related sequential-shock configuration, RAA→distal CS/left subclavian vein→proximal CS, was tested and found to have lower ADFTs than RAA→distal CS (2.0±0.4 versus 5.1±1.8 J), as well.27

    In the present study, the ADFT of RAA→CS/LRA→PA is similar to the RAA→distal CS/PA→proximal CS configuration studied by Cooper and colleagues. We used an electrode in the LRA near the ostium of the CS instead of one in the proximal CS because our CS catheter lacked an electrode at the CS ostium. In the present study, the RAA→CS/LRA→PA ADFT (0.68±0.46 J) was 50±9% lower than that of the RAA→CS configuration (1.27±0.67 J). In the study by Cooper et al, the RAA→distal CS/PA→proximal CS configuration exhibited an ADFT energy 74% lower than that of RAA→distal CS. Thus, the RAA→CS/LRA→PA configuration did not exhibit quite the relative reduction in ADFT compared with the control RAA→CS configuration as the analogous configuration tested in the study by Cooper et al.11 Still, the RAA→SP/CS→SP ADFT, the most efficacious sequential configuration tested with the SP electrode in this study, was not only 68±8% lower than the standard RAA→CS configuration but also 36±15% lower than the RAA→CS/LRA→PA configuration.

    In our model, reversing the order of the sequential shocks from RAA→SP/CS→SP to CS→SP/RAA→SP greatly increased the ADFT. The impedance of the RAA→SP pathway was 46 Ω, whereas that of the CS→SP pathway was 35 Ω (Table 2), which might be attributed to the facts that (1) the RAA electrode (4 cm) was shorter than the CS electrode (6 cm) and (2) the distance between the SP and CS electrodes was shorter than that between the SP and RAA electrodes. When the larger first shock of the sequential stimulus was delivered to the pathway with higher impedance, probably encompassing a larger mass of atrial tissue, and the smaller second shock was delivered to the pathway with lower impedance, the ADFT was lower. This is consistent with the concept that defibrillation requires a minimum potential gradient throughout the entire atrial myocardium.

    In the present study, the atrial SP electrode was placed through a transseptal procedure with most of the electrode situated near the high left atrial SP and the smaller portion near the low right atrial SP. This electrode was between the RAA and CS electrodes near the posterior left atrial wall, where the earliest activations after unsuccessful RAA→CS shocks have been reported to appear.18 This SP electrode was far from the sinus and atrioventricular nodes and remained firmly fixed in location throughout each study.

    Study Limitations

    The study was performed in sheep with acute AF maintained with acetyl-β-methylcholine chloride. Therefore, the results cannot be extrapolated directly to patients with AF. Furthermore, the SP electrode used in this study is not currently clinically acceptable, because it carries the risk of thrombus and embolism and may be problematic to extract.

    Clinical Implications

    As the population ages and the incidence of AF concomitantly rises, the clinical application of atrial cardioversion becomes broader. One of the prime detriments to internal atrial defibrillation is the fact that the intensity of shocks required to successfully cardiovert patients is uncomfortable. If a wholly right-sided SP electrode can be developed that does not damage the sinus and atrioventricular nodes, these data suggest a means for reducing atrial defibrillation energy requirements.

          Figure 1.

    Figure 1. Catheter-based coil electrodes for test configurations. Top, 5F defibrillation lead with 6-cm-long coil electrode used within CS. Middle, Modified quadripolar catheter with 4-cm-long coil electrode ending 1 cm proximal to tip; used within RAA, PA, and LRA. Bottom, Custom-built catheter with a 6-cm coil electrode ending 3 cm proximal to tip for use across atrial SP.

          Figure 2.

    Figure 2. Fluoroscopic images of positions of defibrillation electrodes. Left anterior oblique ≈30° (A) and right anterior oblique ≈30° (B) projections show electrodes in RAA, CS, right ventricle, and across atrial SP.

          Figure 3.

    Figure 3. ADFT leading-edge voltages. Bars give mean and SD voltages. Percentages are mean±SD by which A2 is lower than other configurations. Leading-edge voltage of configuration A2 is significantly lower than each of others.

          Figure 4.

    Figure 4. ADFT leading-edge currents. Bars give mean and SD currents. Percentages are mean±SD by which A2 is lower than other configurations. Leading-edge current of configuration A2 is significantly lower than that of configurations A1, A3, and B, and trended lower than C.

          Figure 5.

    Figure 5. ADFT delivered energies. A, Mean ADFT delivered energies. Bars give mean and SD energies. Percentages are mean±SD by which A2 is lower than other configurations. Delivered energy of configuration A2 is significantly lower than each of others. B, Animal-to-animal comparison of ADFT delivered energies of all configurations.

    Table 1. Test Configurations

    First ShockSecond ShockBiphasic Waveform

    1First-phase anode and cathode.

    Table 2. Impedances of the Current Pathways

    ConfigurationCurrent PathwaysImpedance, Ω
    SingleRAA+CS→SP25 ±3
    FirstCS→SP35 ±4
    FirstRAA→CS54 ±7

    This study was supported in part by National Institutes of Health grant HL-42760, the Whitaker Foundation, and Guidant Corp.


    Correspondence to Raymond E. Ideker, MD, PhD, Volker Hall B140, 1670 University Blvd, Birmingham, AL 35294-0019. E-mail


    • 1 Murgatroyd FD, Slade AKB, Sopher SM, et al. Efficacy and tolerability of transvenous low energy cardioversion of paroxysmal atrial fibrillation in humans. J Am Coll Cardiol.1995; 25:1347–1353.CrossrefMedlineGoogle Scholar
    • 2 Lévy S, Ricard P, Gueunoun M, et al. Low-energy cardioversion of spontaneous atrial fibrillation: immediate and long-term results. Circulation.1997; 96:253–259.CrossrefMedlineGoogle Scholar
    • 3 Nattel S, Hadjis T, Talajic M. The treatment of atrial fibrillation: an evaluation of drug therapy, electrical modalities and therapeutic considerations. Drugs.1994; 48:345–371.CrossrefMedlineGoogle Scholar
    • 4 Coplen SE, Antman EM, Berlin JA. Efficacy and safety of quinidine therapy for maintenance of sinus rhythm after cardioversion: a meta-analysis of randomized control trials. Circulation.1990; 82:1106–1116.CrossrefMedlineGoogle Scholar
    • 5 Cox JL, Boineau JP, Schuessier RB, et al. Five-year experience with the MAZE procedure for atrial fibrillation. Ann Thorac Surg.1993; 56:814–824.CrossrefMedlineGoogle Scholar
    • 6 Alt E, Ammer R, Schmitt C, et al. A comparison of treatment of atrial fibrillation with low-energy intracardiac cardioversion and conventional external cardioversion. Eur Heart J.1997; 18:1796–1804.CrossrefMedlineGoogle Scholar
    • 7 Wellens HJ, Lau CP, Luderitz B, et al. Atrioverter: an implantable device for the treatment of atrial fibrillation. Circulation.1998; 98:1651–1656.CrossrefMedlineGoogle Scholar
    • 8 Cooper RAS, Wallenius ST, Smith WM, et al. The effect of phase separation on biphasic waveform defibrillation. Pacing Clin Electrophysiol.1993; 16:471–482.CrossrefMedlineGoogle Scholar
    • 9 Cooper RAS, Smith WM, Ideker RE. Internal cardioversion of atrial fibrillation: marked reduction in defibrillation threshold with dual current pathways. Circulation.1997; 96:2693–2700.CrossrefMedlineGoogle Scholar
    • 10 Krum D, Hare J, Mughal K, et al. Optimization of shocking lead configuration for transvenous atrial defibrillation. J Cardiovasc Electrophysiol.1998; 9:998–1003.CrossrefMedlineGoogle Scholar
    • 11 Tomassoni G, Newby KH, Kearney MM, et al. Testing different biphasic waveforms and capacitances: effect on atrial defibrillation threshold and pain perception. J Am Coll Cardiol.1996; 28:695–699.CrossrefMedlineGoogle Scholar
    • 12 Harbinson MT, Allen JD, Imam Z, et al. Rounded biphasic waveform reduces energy requirements for transvenous catheter cardioversion of atrial fibrillation and flutter. Pacing Clin Electrophysiol.1997; 20:226–229.CrossrefMedlineGoogle Scholar
    • 13 Jung J, Heisel A, Fires R, et al. Tolerability of internal low-energy shock strengths currently needed for endocardial atrial cardioversion. Am J Cardiol.1997; 80:1489–1490.CrossrefMedlineGoogle Scholar
    • 14 Zipes DP, Jackman WM, Heger JJ, et al. Clinical transvenous cardioversion of recurrent life-threatening ventricular tachyarrhythmias: low energy synchronized cardioversion of ventricular tachycardia and termination of ventricular fibrillation in patients using a catheter electrode. Am Heart J.1982; 103:789–794.CrossrefMedlineGoogle Scholar
    • 15 Nathan AW, Bexton RS, Spurrell RAJ, et al. Internal transvenous low energy cardioversion for the treatment of cardiac arrhythmias. Br Heart J.1984; 52:377–384.CrossrefMedlineGoogle Scholar
    • 16 Wharton JM, Wolf PD, Smith WM, et al. Cardiac potential and potential gradient fields generated by single, combined, and sequential shocks during ventricular defibrillation. Circulation.1992; 85:1510–1523.CrossrefMedlineGoogle Scholar
    • 17 Shibata N, Chen PS, Dixon EG, et al. Epicardial activation following unsuccessful defibrillation shocks in dogs. Am J Physiol.1988; 255:H902–H909.MedlineGoogle Scholar
    • 18 Cooper RAS, KenKnight BH. The role of waveforms and lead configurations for internal atrial cardioversion. HerzschrElektrophys.1998; 9:1–7.Google Scholar
    • 19 Fotuhi PC, Cooper RA, Sreenan CM, et al. Can early timed internal atrial defibrillation shocks reduce the atrial defibrillation threshold? Pacing Clin Electrophysiol.1999; 22:1179–1185.CrossrefMedlineGoogle Scholar
    • 20 Sih HJ, Berbari EJ, Zipes DP. Epicardial maps of atrial fibrillation after linear ablation lesions. J Cardiovasc Electrophysiol.1997; 8:1046–1054.CrossrefMedlineGoogle Scholar
    • 21 Montgomery DC. Design and analysis of experiments. 3rd ed. New York, NY: John Wiley and Sons; 1991:72–76.Google Scholar
    • 22 Cooper RA, Alferness CA, Smith WM, et al. Internal cardioversion of atrial fibrillation in sheep. Circulation.1993; 87:1673–1686.CrossrefMedlineGoogle Scholar
    • 23 Saksena S, Prakash A, Mangeon L, et al. Clinical efficacy and safety of atrial defibrillation using biphasic shocks and current nonthoracotomy endocardial lead configurations. Am J Cardiol.1995; 76:913–921.CrossrefMedlineGoogle Scholar
    • 24 Heisel A, Jung J, Neuzner J, et al. Low-energy transvenous cardioversion of atrial fibrillation using a single atrial lead system. J Cardiovasc Electrophysiol.1997; 8:607–614.CrossrefMedlineGoogle Scholar
    • 25 Lau CP, Lok NS. A comparison of transvenous atrial defibrillation of acute and chronic atrial fibrillation and the effect of intravenous sotalol on human atrial defibrillation threshold. Pacing Clin Electrophysiol.1997; 20:2442–2452.CrossrefMedlineGoogle Scholar
    • 26 Lévy S, Ricard P, Lau C, et al. Multicenter low energy transvenous atrial defibrillation (XAD) trial results in different subsets of atrial fibrillation. J Am Coll Cardiol.1997; 29:750–755.CrossrefMedlineGoogle Scholar
    • 27 Cooper RA, Plumb VJ, Epstein AE, et al. Marked reduction in internal atrial defibrillation thresholds with dual-current pathways and sequential shocks in humans. Circulation.1998; 97:2527–2535.CrossrefMedlineGoogle Scholar