Author + information
- Received August 12, 1998
- Revision received March 6, 1999
- Accepted April 21, 1999
- Published online August 1, 1999.
- Michael Cooklin, MD, MRCPa,
- Mary R Olsovsky, MDa,
- Randall G Brockman, MDa,
- Stephen R Shorofsky, MD, PhD, FACCa and
- Michael R Gold, MD, PhD, FACCa,* ()
- ↵*Reprint requests and correspondence: Dr. Michael R. Gold, Division of Cardiology, N3W77, University of Maryland Medical System, 22 South Greene Street, Baltimore, Maryland 21201
The purpose of this study was to compare transvenous atrial defibrillation thresholds with lead configurations consisting of an active left pectoral electrode and either single or dual transvenous coils.
Low atrial defibrillation thresholds are achieved using complex lead systems including coils in the coronary sinus. However, the efficacy of more simple ventricular defibrillation leads with active pectoral pulse generators to defibrillate atrial fibrillation (AF) is unknown.
This study was a prospective, randomized assessment of shock configuration on atrial defibrillation thresholds in 32 patients. The lead system was a dual coil Endotak DSP lead with a left pectoral pulse generator emulator. Shocks were delivered either between the right ventricular coil and an active can in common with the proximal atrial coil (triad) or between the atrial coil and active can (transatrial).
Delivered energy at defibrillation threshold was 7.1 ± 6.0 J in the transatrial configuration and 4.0 ± 4.2 J in the triad configuration (p < 0.005). Moreover, a low threshold (≤3 J) was observed in 69% of subjects in the triad configuration but only 47% in the transatrial configuration. Peak voltage and shock impedance were also lowered significantly in the triad configuration. Left atrial size was the only clinical predictor of the defibrillation theshold (r = 0.57, p < 0.002).
These results indicate that low atrial defibrillation thresholds can be achieved using a single-pass transvenous ventricular defibrillation lead with a conventional ventricular defibrillation pathway. These data support the development of the combined atrial and ventricular defibrillator system.
The implantable cardioverter defibrillator (ICD) is now primary therapy for many patients with life-threatening ventricular arrhythmias (1–3). Clinically significant atrial fibrillation (AF) is frequently observed in patients with ICDs; indirect data suggest an incidence of at least 20%, although the exact figure is unknown and may be higher (4). The presence of AF in general is associated with an increased incidence of thromboembolic events including stroke, adverse effects on cardiac function and increased mortality (5–7). In addition, many patients are severely symptomatic even if the AF is not hemodynamically compromising (8).
An implantable atrial defibrillator is currently undergoing clinical evaluation (9). On the basis of experimental and clinical studies, low atrial defibrillation thresholds have been demonstrated using complex lead systems including coils in the coronary sinus (10–12). Given the high incidence of AF in patients with ICDs, the possibility of using these devices for atrial defibrillation is intriguing. Because standard active pectoral ICD lead systems deliver current with a vector through the heart including the atria, we postulated that an active pectoral, single-pass lead would be an effective means to achieve atrial defibrillation, and further, that the efficacy of this system would be dependent on the shocking pathway. Accordingly, the present study was a prospective randomized evaluation of active can lead configuration on atrial defibrillation thresholds in patients undergoing ICD implantation. In addition to measuring the efficacy of atrial defibrillation, the clinical predictors of a low defibrillation threshold were assessed.
Thirty-four consecutive patients undergoing initial ICD implant for standard indications were evaluated. By protocol, all patients were in sinus rhythm at the time of implant, and all implants were left-sided. Written informed consent was obtained from each patient, and the study was approved by the Institutional Review Board of the University of Maryland.
Defibrillation lead system implantation
Each patient received an Endotak DSP defibrillation lead (Model 0125), which was positioned at the right ventricular apex under fluoroscopic guidance via either the left cephalic, axillary or subclavian vein. This lead consists of two defibrillating coils, a proximal coil at the right atrial/superior vena caval junction and a second distal coil in the right ventricle. In patients undergoing dual-chamber ICD implant, a separate active fixation atrial lead (Model 4269) was placed in the right atrium. In the remaining patients, a temporary quadripolar pacing catheter was advanced (using the same venous access as the ICD lead) to the lateral right atrium for induction of AF. For acute defibrillation testing, a pulse generator emulator (Model 6967) with a surface area of 78.4 cm2was placed in the left subcutaneous prepectoral pocket. All implanted lead components were manufactured by Cardiac Pacemakers Inc. (Guidant, St. Paul, Minnesota).
Atrial defibrillation testing
Measurement of the atrial defibrillation threshold was performed under conscious sedation with midazolam and fentanyl. Atrial fibrillation was induced with either high output ramp pacing or with alternating current stimulation. When AF had been sustained for >1 min, defibrillation testing was performed. The R-wave synchronized defibrillation shocks were delivered with an external defibrillator (ECD Model 2815, Cardiac Pacemakers), which delivers fixed 60/50 tilt biphasic shocks through a 125 μF capacitance. A step-up protocol was employed, starting at 0.5 J and increasing (1, 2, 3, 5, 8, 10, 15, 20 J) until there was restoration of sinus rhythm. The atrial defibrillation threshold was defined as the lowest energy shock to terminate atrial fibrillation. A threshold ≤3 J was defined prospectively as low energy to allow comparison with reduced output atrial defibrillators (9).
Two different shocking pathways were evaluated in each patient, with the order of testing randomized (Fig. 1). In the triad configuration, the distal right ventricular coil is the anode for the first phase of the biphasic shocks and the proximal right atrial coil and emulator are connected electrically as the cathode (13). The second shocking pathway was the transatrial configuration in which the proximal coil is the anode and the emulator is the cathode. The right ventricular coil was excluded from the system in this configuration.
The following variables were analyzed as possible predictors of atrial defibrillation threshold: age, gender, body surface area, past history of AF, amiodarone use in the previous three months, left ventricular ejection fraction and left atrial size as assessed by echocardiography. The relationships between continuous variables and atrial defibrillation threshold were assessed by linear regression using a least-squares algorithm. For discrete variables, unpaired ttests were performed. Multivariate analysis was then done on all variables with p < 0.10 by stepwise multiple logistic regression with the triad atrial defibrillation threshold as the dependent variable (StatMost 3.5, DataMost, Salt Lake City, Utah).
Analysis of the clinical characteristics of subjects with high atrial defibrillation thresholds was also performed. Patients were grouped by the prospectively defined cutoff of 3 J. Comparison of the high and low threshold groups were made using unpaired ttests for continuous variables and the Fisher exact test for discrete variables. Electrical parameters between the two lead configurations were compared using paired ttests. All data are presented as mean ± SD, and a p value <0.05 was considered significant.
There were 34 consecutive patients enrolled in this study, and sustained AF could be induced in 32 of them who formed the study group. The subjects were 81% male with a mean age of 65 ± 10 years and a mean ejection fraction of 0.30 ± 0.14. Coronary artery disease with ischemic cardiomyopathy was the primary structural heart disease in 28 patients, 1 had idiopathic dilated cardiomyopathy and 3 patients had primary electrical disease with no known structural heart disease. At the time of implantation, 6 patients (19%) were receiving amiodarone. No patients were receiving other type I or type III antiarrhythmic drug therapy.
All patients tolerated the induction and termination of AF without complications. Specifically, there were no embolic events or ventricular arrhythmias induced with atrial defibrillation testing. Moreover, all patients were amnestic to the shocks following implantation. A mean of 8.9 ± 3.9 atrial defibrillation shocks were delivered per patient.
A summary of the electrical parameters measured at the atrial defibrillation threshold for the two shocking pathways is shown in Table 1. Atrial defibrillation energy requirements were reduced from 7.1 ± 6.0 J in the transatrial configuration to 4.0 ± 4.2 J in the triad configuration (p < 0.005). Thus, by delivering shocks between the ventricular coil and the atrial coil connected to the emulator, there was a 44% reduction of defibrillation thresholds compared with the shocking pathway between the right atrium and left pectoral emulator (Fig. 1).
Peak voltage at the atrial defibrillation threshold was reduced 26% in the triad configuration (p < 0.005). As expected, the shock impedance was also lowered significantly (18%) with the three-electrode, dual-coil triad configuration compared with the two-electrode, single-coil transatrial configuration. However, peak current did not differ significantly between the two configurations. This indicates that the reduction of impedance in the triad configuration is sufficient to account for the lowering of defibrillation threshold voltage and energy (13).
Histograms of the distributions of atrial defibrillation thresholds are shown in Figure 2. We prospectively defined defibrillation with ≤3 J as a low threshold, because this is the approximate energy requirement needed to implant a stand-alone atrial defibrillator with limited output capability (9). Low-energy thresholds were achieved in 69% (22 of 32) of subjects in the triad configuration, but only 47% (15 of 32) in the transatrial configuration (p = 0.06). The triad and optimal atrial defibrillation threshold histograms are very similar. In fact, the transatrial defibrillation threshold was more than 1 J lower than the triad threshold in only 3 patients (9%), whereas the triad defibrillation threshold was more than 1 J lower than the transatrial in 17 patients (53%, p < 0.01). This suggests that the ability to program the shocking pathway in this patient population would offer little benefit.
The clinical characteristics of the patient population are shown in Table 2. The patients are grouped by the triad defibrillation threshold. Of the seven clinical variables assessed, only left atrial size differed significantly between the groups with high and low thresholds. Similarly, multivariate analysis, with left atrial size and age as independent variables (see Methods section), identified left atrial size as the only independent predictor of the triad atrial defibrillation threshold. A scatter plot of the relationship between defibrillation threshold and left atrial size is shown in Figure 3.
The major finding of this study was the high efficacy of a single-pass, dual-coil transvenous ventricular defibrillator lead to terminate induced AF. Defibrillation energy requirements were lower in the triad configuration compared with the transatrial configuration. Only left atrial size, among those clinical parameters evaluated, showed a significant correlation with the atrial defibrillation threshold.
Comparison with previous studies
Internal cardioversion of atrial defibrillation was first reported nearly 30 years ago, and since then, several animal and human studies have confirmed the ability of transvenous shocks to defibrillate AF (10–12,14,15). The shocking lead configuration, and more recently the defibrillation waveform, have been shown to be important determinants of defibrillation thresholds (16,17). In general, the lowest thresholds are observed with complex lead systems including defibrillation coils in the coronary sinus, although these thresholds are not sufficiently low to prevent pain in nonsedated subjects. Moreover, the long-term safety and ability to extract these leads because of infection or malfunction are unknown. Consequently, a simple single lead that is easy to implant should be advantageous for chronic systems.
There are limited data on the ability of standard ventricular defibrillator lead systems to defibrillate AF. We previously showed that properly timed low-energy shocks could induce AF, but defibrillation required shock energies ≥3 J (18). Two lead systems were evaluated in that study, a right atrial coil and left ventricular patch configuration and a dual-coil transvenous lead. The relationship between the induction and termination of atrial fibrillation was consistent with the upper limit of vulnerability hypothesis, suggesting that the use of low-energy shocks in sinus rhythm to induce AF may be useful to predict atrial defibrillation thresholds. However, atrial defibrillation thresholds were not evaluated systematically in that study. Both human and animal studies have demonstrated the ability to achieve atrial defibrillation with a dual-coil transvenous lead (19,20).
The potential for atrial defibrillation shocks to induce ventricular fibrillation remains a major concern (20–22), although to date this has not been reported in clinical trials with the atrial defibrillator. However, reduced output atrial defibrillators have only undergone evaluation in patients without structural heart disease, because backup ventricular defibrillation capabilities are not present in these devices. It is noteworthy, therefore, that no ventricular arrhythmias were induced with atrial defibrillation testing in the present study in which all but three patients had a reduced ejection fraction. Although the induction of ventricular arrhythmias is still undesirable with atrial defibrillation shocks, the consequences are likely less severe with ICD systems capable of ventricular defibrillation.
Clinical predictors of defibrillation threshold
Left atrial size was the only clinical predictor of a high atrial defibrillation threshold. This is analogous to clinical predictors of ventricular defibrillation thresholds, in which left ventricular size or mass has been shown to predict high biphasic thresholds (23,24). Studies of transthoracic atrial defibrillation have identified left atrial size and duration of AF as predictors of high thresholds or unsuccessful cases (25,26). The duration of AF is not a factor for defibrillation with implantable systems because shocks are delivered acutely after the onset of the arrhythmia.
Our study must be interpreted in the face of certain methodologic limitations. First, defibrillation thresholds of induced, but not spontaneous, AF were measured. Therefore, it is possible that the lead configurations tested would have a lower efficacy for the termination of spontaneous AF. However, previous studies have shown that defibrillation thresholds of induced AF are similar to spontaneous AF (27,28). Moreover, at least 1 min of AF was required before testing was begun, thus minimizing the chance that AF terminated spontaneously rather than by the delivered shock. Second, atrial defibrillation thresholds were measured in two configurations, so it is possible that the optimal shocking pathway for atrial defibrillation with this single-pass lead was not identified. Finally, only a single biphasic waveform was evaluated. Recently, it was shown that a large capacitance waveform markedly reduced thresholds (29). Therefore, it is possible that lower thresholds could be obtained with other waveforms or capacitances and this lead system.
This study demonstrated low atrial defibrillation thresholds using a single-pass transvenous ventricular defibrillation lead with a conventional ventricular defibrillation pathway. Given the simplicity of implantation and chronic stability of integrated ventricular defibrillation leads, this single-lead, active pectoral configuration should be considered as an alternative to the more complex multiple-lead systems now in clinical trials. These data support the development of the combined atrial and ventricular ICD system, although further studies are needed to identify optimal defibrillation pathways and waveforms.
☆ This study was supported by a grant from Cardiac Pacemakers, St Paul, Minnesota.
- atrial fibrillation
- implantable cardioverter defibrillator
- Received August 12, 1998.
- Revision received March 6, 1999.
- Accepted April 21, 1999.
- American College of Cardiology
- Winkle R.A,
- Mead R.H,
- Ruder M.A,
- et al.
- Wolf P.A,
- Abbott R.D,
- Kannel W.B
- Middlekauff H.R,
- Stevenson W.G,
- Stevenson L.W
- Lévy S,
- Ricard P,
- Lau C.-P,
- et al.
- Murgatroyd F.D,
- Slade A.K,
- Sopher S.M,
- Rowland E,
- Ward D.E,
- Camm A.J
- Powell A.C,
- Garan H,
- McGovern B.A,
- Fallon J.T,
- Krishnan S.C,
- Ruskin J.N
- Cooper R.A,
- Alferness C.A,
- Smith W.M,
- Ideker R.E
- Cooper R.A,
- Johnson E.E,
- Wharton J.M
- Ayers G,
- Alferness C,
- Ilina M,
- et al.
- Raitt M.H,
- Johnson G,
- Dolack G.L,
- Poole J.E,
- Kudenchuck P.J,
- Bardy G.H
- Cooper R.A.S,
- Plumb V.J,
- Epstein A.E,
- Neal Kay G,
- Ideker R.E