Author + information
- Received March 12, 1999
- Revision received July 2, 1999
- Accepted September 1, 1999
- Published online December 1, 1999.
- Patrick Schauerte, MDa,b,
- Benjamin J Scherlag, PhD, FACCa,b,* (, )
- Michael A Scherlag, MDa,b,
- Sunil Goli, MDa,b,
- Warren M Jackman, MD, FACCa,b and
- Ralph Lazzara, MD, FACCa,b
- ↵*Reprint requests and correspondence: Benjamin J Scherlag, Research Service 151-F, D.V.A. Medical Center, 921 NE 13thStreet, Oklahoma City, OK 73104.
To identify intravascular sites for continuous, stable parasympathetic stimulation (PS) in order to control the ventricular rate during atrial fibrillation (AF).
Ventricular rate control during AF in patients with congestive heart failure is a significant clinical problem because many drugs that slow the ventricular rate may depress ventricular function and cause hypotension. Parasympathetic stimulation can exert negative dromotropic effects without significantly affecting the ventricles.
In 22 dogs, PS was performed using rectangular stimuli (0.05 ms duration, 20 Hz) delivered through a catheter with an expandable electrode-basket at its end. The catheter was positioned either in the superior vena cava (SVC, n = 6), coronary sinus (CS, n = 10) or right pulmonary artery (RPA, n = 6). The basket was then expanded to obtain long-term catheter stability. Atrial fibrillation was induced and maintained by rapid atrial pacing.
Nonfluoroscopic (SVC) and fluoroscopic (CS/RPA) identification of effective intravascular PS sites was achieved within 3 to 10 min. The ventricular rate slowing effect during AF started and ceased immediately after on-offset of PS, respectively, and could be maintained over 20 h. In the SVC, at least a 50% increase of ventricular rate (R-R) intervals occurred at 22 ± 11 V (331 ± 139 ms to 653 ± 286 ms, p < 0.001), in the CS at 16 ± 10 V (312 ± 102 ms vs. 561 ± 172 ms, p < 0.001) and in the RPA at 18 ± 7 V (307 ± 62 ms to 681 ± 151 ms, p < 0.001). Parasympathetic stimulation did not change ventricular refractory periods.
Intravascular PS results in a significant ventricular rate slowing during AF in dogs. This may be beneficial in patients with AF and rapid ventricular response since many drugs that decrease atrioventricular conduction have negative inotropic effects which could worsen concomitant congestive heart failure.
Atrial fibrillation (AF) is the most common cardiac arrhythmia with a prevalence of 2.3% in people older than 40 years and 5.9% in those older than 65 years (1). Among other cardiac and noncardiac risk factors, congestive heart failure imposes the greatest risk of developing AF with a 4.5-fold increased risk in men and 5.9-fold increased risk in women (2). Moreover, in patients with overt congestive heart failure, the prevalence of AF can rise up to 40% (3). Ventricular rate (R-R) control during AF in patients with congestive heart failure is a significant clinical problem. The acute treatment of rapid ventricular response to AF requires the administration of drugs with an immediate depressant effect on atrioventricular (AV) conduction. However, drugs generally used for that purpose such as beta-receptor antagonists and calcium-channel antagonists may also lead to worsening of concomitant congestive heart failure and hypotension. Amiodarone and digoxin may be used for acute R-R control especially in patients with concomitant congestive heart failure; yet, for both drugs there is a latency of the onset of the negative dromotropic effect of at least 60 min (4,5).
We hypothesized that intravascular electrical stimulation of parasympathetic cardiac nerves which innervate the AV node might decrease the rapid ventricular response during AF without affecting ventricular contractility. If so, a transvenous approach to obtain stable parasympathetic nerve stimulation might offer a new opportunity for R-R control during AF, especially in patients with depressed ventricular function.
All animal studies were approved by the Research and Development Committee of the Department of Veterans Affairs Medical Center, Oklahoma City, Oklahoma. In 22 adult mongrel dogs (weight 18 to 30 kg), anesthesia was induced and maintained with intravenous sodium pentobarbital (30 mg/kg initial/50 to 100 mg for maintenance). The dogs were intubated and ventilated with room air (Harvard Apparatus Co., Natick, Massachusetts). A cannula was inserted into the left external jugular vein for fluid and drug delivery. Blood pressure was monitored through a cannula in the right femoral artery. A quadripolar catheter with 2 mm interelectrode spacing was introduced into the left carotid artery and advanced to the aortic root to record His bundle activity. A right lateral thoracotomy was performed at the fourth intercostal space. After pericardiotomy pairs of plunge wire electrodes were inserted into the right atrial appendage (RAA) and right ventricular apex for pacing and local electrogram recording. Electrocardiograms lead II and aVR were monitored continuously. All tracings were amplified and digitally recorded using a computer-based Bard Labsystem (CR Bard Inc., Billerica, Massachusetts). Electrocardiogram filter settings were 0.01 to 250 Hz, whereas bipolar electrograms were filtered at 30 to 250 Hz.
Parasympathetic stimulation (PS)
A 7-French basket catheter (Cordis Webster Corp., Diamond Bar, California) was used for PS (Fig. 1 and 2). ⇓⇓Bipolar stimulation and electrogram recording could be performed over adjacent pairs of five metal electrode arms. Periods of electrical stimulation lasting 10 s were delivered by a Grass stimulator (Astro-Med, Inc., Grass Instruments Division, West Warwick, Rhode Island) at a frequency of 20 Hz and a pulse duration of 0.05 ms. These parameters had been demonstrated to be suitable for epicardial PS while being subthreshold for excitation of myocardium (6). The stimulation voltage, which was delivered through a pair of splines of the basket catheter, was monitored on an oscilloscope (Tektronix 335, Tokyo, Japan). Parasympathetic stimulation at each of three intravascular stimulation sites [superior vena cava (SVC), coronary sinus (CS) and right pulmonary artery (RPA)] was performed during AF. The latter was induced and maintained by constant rapid pacing at a cycle length of 100 ms from the RAA. Once the effective CS, RPA or SVC site was identified, PS was started at a stimulator output voltage of 10 V and thereafter increased in 10 V steps to a stimulator output voltage of 70 V (CS, equivalent to a catheter output of 16 V), 110 V (RPA, equivalent to a catheter output of 27 V) or 150 V (SVC, equivalent to a catheter output of 39 V). Ventricular rate intervals during PS were measured in surface ECG leads II and aVR. At each stimulus strength (SST); 10 consecutive R-R intervals during PS were averaged.
Group 1: stimulation at the CS (n = 10)
For PS in the CS, the catheter was introduced through the right or left external jugular vein and advanced into the right atrium. Positioning in the os of the CS was then performed under fluoroscopic guidance. When the CS had been entered, the basket was expanded to hold the catheter in a stable position against the endovascular surface. Stimulation over each adjacent pair of electrode arms was then attempted at a catheter output voltage of 16 V. If no visible slowing of the ventricular rate during AF was observed, the basket was contracted and the catheter gently rotated or advanced further into the CS until an effective stimulation site was found. In three dogs, additional epicardial stimulation of parasympathetic ganglia was performed using a handheld bipolar stimulation electrode (electrode spacing: 5 mm). This electrode was positioned on the fat pad at the junction of the inferior cava and the left atrium.
Group 2: stimulation in the RPA (n = 6)
For placement of the catheter in the RPA, the basket catheter was introduced through a purse-string suture in the right ventricular outflow tract. Under fluoroscopic control it was then advanced into the proximal RPA. The sinus rate slowing was assessed by stimulation over each adjacent pair of electrode arms at a catheter output voltage of 27 V. If no visible ventricular rate slowing during AF occurred, the basket was contracted and withdrawn stepwise. Stimulation was repeated until a noticeable drop in ventricular rate during AF was achieved. Figure 1shows a pulmonary angiogram with the basket catheter positioned at an effective stimulation site.
Group 3: stimulation in the SVC (n = 6)
For stimulation in the SVC, the catheter was introduced through the right or left external jugular vein until a right atrial electrogram was recorded in one of the pairs of electrode arms of the basket. The catheter was then withdrawn to an SVC site at which no atrial signal was recorded. At this point the basket was expanded and the ventricular rate slowing effect during induced AF was assessed by stimulation over each adjacent electrode pair at a catheter output voltage of 39 V. If no slowing effect was observed, the catheter was withdrawn and stimulation was repeated until a slowing response was obtained. Figure 2shows an angiographic view of the SVC with the basket catheter positioned at a successful stimulation site.
Measurement of sinus rate slowing and effective refractory period (ERP) for intravascular stimulation
In those dogs in which stimulation was performed at the SVC and RPA sites, we also assessed the sinus rate slowing and atrial and ventricular effective refractory periods (ERP) and AV conduction during PS. Successive atrium to atrium (A-A) intervals during PS with a catheter output voltage of 39 V (SVC) and 38 V (RPA) were measured in the RAA electrogram. Three consecutive A-A intervals were taken for calculation of the mean A-A interval. Atrial and ventricular ERPs were determined at baseline and during PS at the RPA and SVC sites with a catheter output voltage of 39 V (SVC) and 38 V (RPA). The baseline ERPs and ERPs during PS were taken after the assessment of the PS effect on the ventricular rate during AF. For pacing and ERP measurement, 2 ms rectangular pacing stimuli at twice the diastolic pacing threshold were delivered by a Radionics stimulator (Radionics, Inc., Burlington, Massachusetts) using the extrastimulus technique (basic cycle length: 400 ms, extrastimuli: 10 ms decrements). Atrioventricular conduction was assessed by determining the cycle lengths during incremental atrial pacing, at which anterograde second degree AV block (Wenckebach type I) occurred.
Long-term stability of the slowing effect due to PS at the CS, effects of atropine and section of cervical vagosympathetic trunks
For evaluation of the long-term efficacy, continuous PS was maintained at a catheter output voltage of 25 V over 20.5 h during pacing induced AF in one dog. Every half-hour R-R intervals during PS were measured. The SST was then increased for 30 s to 39 V, and R-R intervals were measured during this period. Both values were compared with R-R intervals during a 30-s pause of PS. In six dogs, two in each study group, 2 mg of atropine was injected intravenously and PS was attempted again. In three dogs, one in each study group, both cervical vagosympathetic trunks were cut before PS.
All data are expressed as mean ± standard deviation. The voltage data refer to the voltage delivered to the heart (= catheter output voltage) as measured on the oscilloscope. In each group, repeated-measure ANOVA was used to test whether the SST affected the ventricular rate during AF. The rate slowing effect at a given SST was compared with baseline values without PS, applying a Wilcoxon matched pairs signed rank test. In the dog in which a 20.5-h measurement was performed, rate differences were compared using a Wilcoxon matched pairs signed rank test. Refractory periods were evaluated for significance by a Student paired ttest. Probability values ≤0.05 were considered significant.
Parasympathetic cardiac nerve stimulation in the CS
After introduction of the basket catheter in the CS os, identification of the CS site and the catheter splines at which a rate slowing occurred could be achieved within 5 min. Figure 3shows a representative example of PS in the proximal CS. During PS at 34 ± 8 V, there was a marked increase of the R-R intervals (shortest 1,440 ms, longest 1,976 ms). Upon termination of PS (arrow), the shortest/longest R-R intervals were 260/600 ms, respectively. The lowest voltage at which at least a 50% increase of the R-R intervals occurred during AF was 16 ± 10 V (312 ± 102 ms without PS vs. 561 ± 172 ms during PS, p < 0.001). The effect of the applied stimulus voltage on R-R slowing was assessed by a dose-response type curve as illustrated in Figure 4. Stimulus strength significantly affected ventricular rate slowing during AF (p < 0.001). The graph also shows that direct epicardial stimulation of the parasympathetic ganglia at the outer surface of the proximal CS required a much lower PS voltage level. Importantly, nerve stimulation from the epicardium during sinus rhythm with a handheld bipolar electrode effectively prolonged the atrium to His bundle (AH) interval without a change in the sinus rate but was subthreshold for stimulation of atrial myocardial tissue. By contrast, if nerve stimulation from the adjacent intravascular site was performed during sinus rhythm, this excited the atria resulting in AF at a voltage that was lower than the one which resulted in R-R slowing during AF. Parasympathetic stimulation in the CS during right ventricular pacing at a constant cycle length of 300 ms did not significantly affect peak systolic arterial blood pressure (121 ± 10 mm Hg without PS vs. 117 ± 6 mm Hg with PS; n = 3; p = NS).
Parasympathetic cardiac nerve stimulation in the RPA
Stimulation in the RPA also decreased the ventricular rate response during AF. On average it required 10 min to position the catheter within the RPA to find an effective stimulation site. The effect of PS significantly depended on the applied stimulus voltage (Fig. 4, p < 0.001). The minimal SST inducing at least a 50% prolongation of the R-R interval during AF required 18 ± 7 V (R-R interval of 307 ± 62 ms without PS vs. 681 ± 151 ms with PS, p < 0.001). This rate slowing effect was probably due to a negative dromotropic effect of PS on the AV node as indicated by a prolongation of the anterograde Wenckebach cycle length from 257 ± 40 ms without PS to 384 ± 39 ms with PS in the RPA (p = 0.01, n = 5). Stimulation in the RPA also produced a shortening of the right atrial refractory period (136 ± 13 ms without PS vs. 96 ± 22 ms during PS, n = 5, p = 0.03) whereas the right ventricular refractory period did not change significantly with PS (184 ± 25 ms with PS vs. 186 ± 26 without PS, n = 5). Moreover, a significant slowing of the sinus rate was observed during PS (1,181 ± 306 ms during PS vs. 518 ± 138 ms without PS, n = 6, p = 0.01). Parasympathetic stimulation in the RPA during right ventricular pacing at a constant cycle length of 300 ms did not significantly change peak systolic arterial blood pressure (115 ± 13 mm Hg without PS vs. 113 ± 9 mm Hg with PS; n = 3; p = NS).
Parasympathetic cardiac nerve stimulation in the SVC
Positioning of the basket catheter at an effective nerve stimulation site in the SVC was achieved within 3 min, after the introduction of the basket catheter through the jugular vein. A typical example of ventricular rate slowing during PS is shown in Figure 5. The lowest voltage at which a 50% increase of the mean ventricular cycle length during AF occurred was 22 ± 11 V (331 ± 139 ms without PS vs. 653 ± 286 ms during PS, p < 0.001). The SST during PS significantly affected the R-R during AF as illustrated in Figure 4(p < 0.001). Parasympathetic stimulus in the SVC exerted a negative dromotropic effect on the AV node: the anterograde Wenckebach cycle length increased from 263 ± 107 ms without PS to 391 ± 66 ms with PS in the SVC (p < 0.01, n = 5). During sinus rhythm, PS led to a significant increase in the average sinus cycle length (489 ± 154 ms without PS vs. 1,056 ± 355 ms with PS, n = 6, p < 0.001) and to a shortening of the right atrial refractory period (145 ± 55 ms without PS vs. 78 ± 71 ms with PS, n = 6, p = 0.03). The right ventricular refractory period did not change significantly during PS (172 ± 26 ms with PS vs. 176 ± 34 ms without PS, n = 5). In addition, PS in the SVC did not significantly change peak systolic arterial blood pressure during right ventricular pacing at a constant cycle length of 300 ms (105 ± 14 mm Hg without PS vs. 110 ± 19 mm Hg with PS; n = 3; p = NS). Intermittent stimulation of the phrenic nerve was sometimes observed during stimulation across various splines of the basket catheter. When the site and electrode pair was found at which electrical stimulation induced ventricular slowing during AF, stimulation of the phrenic nerve was never observed.
Long-term stability of the slowing effect due to PS at the CS, effects of atropine and section of cervical vagosympathetic trunks
Continuous PS in the CS was performed over 2 h in two dogs. Although the mean R-R interval during AF after 2 h of PS was shorter than at the onset of PS (644 ± 149 ms vs. 778 ± 210 ms, respectively, n = 2), it was still longer compared with the R-R intervals without PS (452 ± 114 ms, n = 2). In another dog, the stability of the rate slowing effect was assessed over a 20.5-h period as illustrated in Figure 6. The effect of R-R slowing during AF slightly diminished in the first 3 h of PS but then remained fairly constant. After 20.5 h there was still a considerable decrease of the R-R during PS as compared with no PS. The latter was determined during a 30 s period of no PS. Of note, this small loss of efficacy over 20.5 h was overcome if the PS voltage was increased from 25 to 39 V as shown in Figure 6. At all three stimulation sites, the effect of PS on the ventricular response during AF was completely abolished after intravenous injection of 2 mg of atropine. Cutting of both cervical vagosympathetic trunks did not prevent the PS related decrease of the ventricular response during AF at any of the three stimulation sites indicating that probably stimulation of efferent nerves was responsible for the rate decrease.
Effect of different impulse durations
The influence of different pulse durations of the electrical stimuli during PS in the SVC was evaluated in one dog. Nerve stimulation was performed in the SVC at 32 V, 20 Hz. Maximal rate slowing was observed at an impulse duration of 2 ms (626 ms) as compared with 341 ms at 0.05 ms and 262 ms without PS. Thus, the optimal impulse duration (2 ms) was considerably longer than the one applied in this study (0.05 ms).
In this study we described a transvenous approach for consistent and stable parasympathetic cardiac nerve stimulation in dogs at three distinctive intravascular anatomical sites. Our findings corroborate and extend the various epicardial (6–10)and endocardial (11,12)approaches described previously.
Epicardial stimulation of parasympathetic ganglia located in a fat pad near the proximal CS has been reported to affect AV nodal conduction but not to affect the sinus rate or the atrial refractory period (6,7). Likewise, Ali et al. (8)reported recently that epicardial stimulation at the proximal CS resulted in R-R slowing during atrial tachycardias and AF, which was prevented by atropine. Similarly, intravascular PS in the proximal CS caused a considerable R-R slowing during AF in this study which was completely abolished after atropine injection.
Extracardiac stimulation of a fat pad which contained parasympathetic ganglia and was located adjacent to the RPA was reported to result in mixed responses such as prolongation of AV conduction, sinus rate slowing and shortening of the atrial refractory period (9). The same observations were made in this study, when intravascular stimulation at the RPA site was performed, suggesting that similar structures have been stimulated in both studies.
These three electrophysiological effects can also be achieved during intravascular stimulation of cardiac nerves in the SVC as shown in this study. Negative dromotropic and chronotropic effects were also found by Armour et al. (10)during external stimulation of mediastinal cardiac nerves in open-chest dogs. However, besides slowing of the sinus rate and a prolongation of AV nodal conduction, stimulation of some nerves induced an increase of the supraventricular rate. The same group recently observed a sinus rate slowing during transvenous electrical stimulation of parasympathetic nerves in the right internal jugular vein or SVC using conventional electrode catheters in animals (11). In that study exact positioning of a conventional electrode catheter under fluoroscopy was crucial to obtain a sinus slowing. A significant advantage of a basket catheter as compared with a conventional electrode catheter is the ease of identification of an effective stimulation site and the stability of the catheter position at this site in the vessel over 20 h. Moreover, stimulation in the SVC with a basket catheter does not require fluoroscopy, which makes it particularly suitable for the acute treatment of rapidly conducting AF in critically ill patients.
To be of clinical value it is important to know whether the negative dromotropic effects can be maintained over a longer stimulation period. The results of our study indicate that at least for a stimulation period of 20.5 h the negative dromotropic effect in the CS can be maintained. This is noteworthy because a fade of the response during short-term PS has been reported for the negative chronotropic effect and the negative inotropic effect on the atria (13). By contrast, a decrease of the negative dromotropic response was not observed in one study (14)whereas another study reported either no decrease, an augmentation or a small fade of the negative dromotropic response during a 5 min stimulation period (15). Possible mechanisms for a fade of the parasympathetic response were elaborately discussed by Salata and Jalife (16).
Clinical application of the proposed approaches to intravascular PS requires the demonstration of similar anatomical cardiac nerve distributions and electrophysiological effects with nerve and ganglia stimulation in humans. There is evidence from recent histological studies that a large population of neural ganglia is present at the ostium of the CS and adjacent to the base of the pulmonary artery and aortic root in humans (17). More recently Chen et al. (12)were able to demonstrate intracardiac PS from the endocardial surface at the posteroseptal area of the right atrium in humans using a 4 mm tip deflectable catheter. This demonstration was made in order to confirm that slow pathway ablation which terminates AV reentrant tachycardia did not also denervate the parasympathetic input to the AV node. Thoracic cardiac nerve stimulation has been attempted in patients during coronary artery bypass operations (18). It was found that nerve excitation resulted in sinus rate slowing in some patients but led to an increase in heart rate in others. However, the study was performed during sinus rhythm and without the use of His bundle recordings. Therefore, if present, negative dromotropic effects might not have been recognized.
The stimulation method presented in this study might have several advantages as compared with the conventional treatment for AF with rapid ventricular response. First, intravascular PS decreases the R-R during AF without exerting significant negative inotropic effects on the ventricles. Thus, these approaches avoid the ventricular negative inotropic effects of many pharmacological agents that are currently used to slow AV conduction during AF, such as calcium channel antagonists and beta-adrenergic blocking agents. The proposed single catheter approach may, therefore, be used in circumstances in which AF with rapid ventricular response worsens coexisting congestive heart failure, which is a situation most commonly seen in critically ill patients.
Second, the on- and offset of the rate slowing effect occurred promptly after initiation and termination of PS, respectively. Moreover, by changing the applied stimulus voltage, the effect of rate slowing may be titrated to the specific demands in patients with dynamically changing medical conditions. By contrast, drugs—once given—exhibit their effects and side-effects depending on their half-life which is, except for the short acting agents such as esmolol, in the range of hours. Therefore, changes in the patients’ medical conditions, particularly development of or worsening of congestive heart failure, may require administration of counteracting pharmacological agents, i.e., catecholamines. This, in turn may increase the R-R during AF.
Third, in patients in whom unstable hemodynamics develop during rapidly conducting AF, intubation and cardioversion may be required. Transvenous PS may obviate the need for emergency intubation and cardioversion and its associated complications.
Parasympathetic stimulation at the CS site resulted in electrical excitation of atrial tissue. Although this may not be relevant in the setting of chronic AF, it may contribute to the perpetuation of AF in patients with recent onset AF. The use of SVC and RPA sites can provide the desired effect without atrial stimulation.
A second issue that needs to be addressed might be blood clotting within the basket. Because our dogs were heparinized, we did not observe clotting in the present experiment. Because many patients with AF receive either oral or intravenous anticoagulation, clotting within the CS might not occur in patients either. However, we cannot exclude that, without anticoagulation, formation of coagula within the basket or in the distal CS may occur.
Stimulation at the RPA and SVC site resulted in a shortening of the right atrial refractory period. We cannot rule out that this may contribute to the perpetuation of AF in patients with recent onset AF. Apart from this, a shortening of the atrial refractory periods during PS could have caused an increased atrial rate during AF, which in part may have contributed to the slowing of the R-R during PS. However, the findings of a significant prolongation of the anterograde Wenckebach cycle length during PS in the SVC and RPA are a strong evidence for a direct interference of PS with conduction in the AV node.
Finally, it also cannot be excluded that PS in the SVC may excite nerve fibers that provide the bronchiopulmonary system or the gut, possibly resulting in coughing, vomiting or chest discomfort in the unanesthetized state.
Cardiac parasympathetic nerves that project to the AV node can be stimulated intravascularly with a basket catheter in dogs. Three intravascular sites at which a substantial reduction in R-R during AF can be obtained were identified: the SVC, the ostium of the CS and the RPA. By changing the stimulation voltage, a graded response of the R-R during AF ranging from slight slowing to complete AV block was observed. If these findings translate to patients, PS may be beneficial in AF patients with rapid ventricular response and poor ventricular function in whom rate control with drugs could exacerbate already compromised ventricular contractility.
☆ Dr. Schauerte was supported by a grant from the “Friedrich-Wilhelm-Stiftung,” Germany.
- atrial fibrillation
- coronary sinus
- effective refractory period
- parasympathetic stimulation
- right atrial appendage
- right pulmonary artery
- ventricular rate
- stimulus strength
- superior vena cava
- Received March 12, 1999.
- Revision received July 2, 1999.
- Accepted September 1, 1999.
- American College of Cardiology
- Godtfredsen J.
- Prystowsky E.N.,
- Benson D.W.,
- Fuster V.,
- et al.
- Lazzara R.,
- Scherlag B.J.,
- Robinson M.J.,
- Samet P.
- Wallick D.W.,
- Martin P.J.
- Chiou C.W.,
- Eble J.N.,
- Zipes D.P.
- Armour J.A.,
- Randall W.C.,
- Sinha S.
- Chen S.A.,
- Chiang C.E.,
- Tai C.T.,
- et al.
- Martin P.,
- Levy M.N.,
- Matsuda Y.
- Loeb J.M.,
- Dalton D.P.,
- Moran J.M.
- Martin P.
- Salata J.L.,
- Jalife J.