Radiofrequency Catheter Ablation of the Atria Reduces Inducibility and Duration of Atrial Fibrillation in Dogs
Background The purpose of this study was to prevent induction of sustained atrial fibrillation (AF) by radiofrequency catheter ablation (RFCA) of the atria in an open-chest canine model.
Methods and Results In dogs randomized to acute studies, RFCA of the atria was performed after reproducible induction of sustained AF (lasting >30 minutes) with burst stimulation or premature atrial pacing and perpetuation by low level cervical vagal stimulation or IV infusion of methacholine. Additionally, in four dogs, the long-term effectiveness of RFCA was assessed 7 to 21 days after ablation. Continuous discrete transmural lesions were produced with radiofrequency energy pulses (20 to 40 W for 60 seconds) delivered to five atrial epicardial sites and endovascularly to the coronary sinus wall. RFCA electrically isolated regions of the atria that became dissociated from the nonisolated parts. Atrial RFCA markedly attenuated vagally induced shortening of effective refractory period (ERP) at both isolated and nonisolated test sites located in the left and right atria (P<.001, n=5). RFCA rendered noninducible sustained AF maintained by cervical vagal stimulation. The dose-response curve relating the dose of methacholine required to maintain AF was shifted down and to the right. AF was only inducible with high doses of methacholine. Atrial RFCA reduced the maximal sinus rate and prolonged the corrected sinus-node recovery time (P<.001, n=6). However, RFCA did not affect atrial contractile function, AV-nodal ERP, or AV-nodal or His-Purkinje conduction times. In dogs in the chronic group, normal sinus rhythm and normal AV conduction were preserved and AF was only inducible with a high dose of methacholine. No atrial perforations resulted.
Conclusions RFCA in open-chest dogs produces partial vagal denervation and reduces the inducibility of AF.
Atrial fibrillation (AF) is one of the most prevalent cardiac arrhythmias; it affects ≈1.0% of the population and is associated with significant clinical and public health implications.1 2 3 4 5 6 7 Ideal treatment should abolish the detrimental effects of AF by restoring sinus rhythm with a synchronous atrial kick8 and reducing the risk of thromboembolism. Pharmacological therapy may not be successful for maintaining sinus rhythm and may have proarrhythmic and other significant side effects.9 10 11 12 If pharmacological therapy fails, catheter ablation of the His bundle with pacemaker implantation is often performed to control the ventricular response rate, but the risk of thromboembolism and compromised hemodynamics remains.13 14
In selected cases, AF is treated surgically. One of the most promising surgical methods to treat AF is the maze procedure developed by Cox et al.15 The surgical-maze procedure divides the atrial myocardium to force atrial activation over a surgically determined route. The mass of tissue to be electrically activated in this atrial maze presumably is too small for multiple wavelets of reentry to sustain. The maze procedure restores normal sinus rhythm and AV synchrony with demonstrable atrial contractile function and reduces the risk of thromboembolism.
In the present study, we tested the effectiveness of epicardial application of radiofrequency energy to prevent the induction of AF in an open-chest canine model. The impetus to perform the present study was results from the surgical-maze procedure.
Surgical Preparation for Acute Study
For groups 1 through 4 and 6, 27 mongrel dogs of either sex weighing 25 to 35 kg each were anesthetized with pentobarbital 30 mg/kg IV. Additional doses were injected as needed to maintain anesthesia during the study. Dogs were ventilated with room air by use of a cuffed endotracheal tube and a constant volume–cycled respirator (model 607, Harvard Apparatus). A fluid-filled cannula was placed in the left femoral artery and connected to a transducer (Statham p-23Db, Gould) to monitor arterial blood pressure, and the left femoral vein was cannulated to infuse normal saline at 100 to 200 mL/h to replace spontaneous fluid losses. ECG lead II was monitored throughout the study. A His-bundle electrogram was recorded in all dogs with a 7F bipolar electrode catheter (USCI) introduced through the right carotid artery and advanced in a retrograde manner into the noncoronary cusp of the aortic valve. The signals were amplified, filtered between 30 and 1000 Hz, and recorded simultaneously on a recorder (model 2800, Gould Brush).
The chest was opened through a median sternotomy. The pericardium was opened and sewn to the chest wall to cradle the heart. Bipolar stainless-steel Teflon-coated wire electrodes were sutured to the right ventricular free wall to record the ventricular electrogram. An array of six stainless-steel electrodes embedded in an acrylic base with an interelectrode distance of 3 mm was sutured in the right atrial epicardium 1 to 2 cm inferior to the sinus node to record bipolar atrial electrograms and to pace the atrium. In 13 dogs, five plunge electrodes made from Teflon-coated wires, insulated except for their tips, were inserted in the right and left lateral atrial myocardium to a depth of 3 to 4 mm. These electrodes served as the cathodes for unipolar stimulation to determine the atrial refractory period. An anodal electrode was placed in the abdominal wall. A thermistor (model 430, series 400, Yellow Springs Instruments) was used to monitor epicardial temperature, which was maintained between 36° and 38°C by use of an operating-room table lamp.
Surgical Preparation for Chronic Study
For group 5, 4 mongrel dogs were premedicated with antibiotics (cefazoline 500 mg IM), heparinized with heparin 3000 U IV, and anesthetized initially with sodium thiopental 30 mg/kg. Anesthesia was maintained by ventilation with 1% to 3% isoflurane gas mixed with oxygen at a rate of 1.0 to 2.0 L/min. By use of a sterile surgical technique, a right and left lateral intercostal thoracotomy was performed, the ribs and lungs were retracted, and the pericardium was incised. Bilateral pericardial windows were created to expose the atria. Serum electrolytes, pH, base excess, and blood gases were monitored during the procedure and were repeated daily for 3 days after surgery. Bipolar stainless-steel wire electrodes were sutured to the atrial epicardium to pace the atria. Before radiofrequency catheter ablation (RFCA) of the atria (described below), all dogs underwent open-chest electrophysiological study to obtain baseline data. After atrial RFCA, the chest was closed in layers and negative pressure was reestablished in the pleural cavity. Antibiotics and heparin were administered for 5 days after surgery, and analgesics were given as needed.
Radiofrequency Energy Application
After reproducible induction of sustained AF (duration >30 minutes) with atrial burst pacing (cycle length, 50 ms) or pacing with multiple atrial extrastimuli at various coupling intervals and perpetuation with either cervical vagal stimulation or infusion of methacholine, radiofrequency energy was delivered through the epicardium to the posterolateral right atrium, the anteromedial side of the superior vena cava, both the right and left atrial appendages, and the transverse sinus (Fig 1⇓, top) by means of 6F or 7F catheters with hexapolar or octapolar multielectrode (interelectrode distance, 3.4 mm; 3-mm rings with a 3-mm tip electrode) steerable tips (EP Technologies). The catheter tips were positioned manually on the epicardial surface of the atria under direct visualization to ensure optimal tissue contact and energy delivery. Continuous unmodulated radiofrequency current of 300 to 750 kHz with power output settings of 20 to 40 W was delivered from a radiofrequency generator (EP Technologies) by means of the multielectrode catheter tips at each ablation site. The radiofrequency generator measured root-mean-square voltage and current and actual power output by use of the average value of the product of voltage and current. This reflected the effective heating power that was delivered to the tissue from the steerable ablation catheter tip. Impedance was calculated as measured voltage divided by measured current. Duration of radiofrequency power output was programmable, and the number of radiofrequency applications was counted automatically. Lesions were produced in the epicardium of the atria at power-output settings of 30 to 40 W, with a duration of 60 seconds. Initially, an excessive rise in impedance often limited radiofrequency power delivery during ablation. Since impedance rises abruptly when temperatures at the electrode-tissue interface exceed 100°C,16 17 18 19 excessive heat formation was prevented by flushing the catheter tip with small amounts of saline during epicardial radiofrequency energy delivery. Cooling the tip reduced the excessive rise of impedance and allowed delivery of larger amounts of radiofrequency power.20 After epicardial atrial RFCA was performed, a 6F catheter with a hexapolar multielectrode steerable tip (EP Technologies) was introduced through the right external jugular vein and advanced into the distal coronary sinus (Fig 1⇓, bottom). The power output setting for the coronary sinus lesion was 10 to 20 W, depending on the impedance level, and had a duration of 30 seconds. If the measured tissue impedance was <50 Ω or >300 Ω, radiofrequency current flow was shut off with the built-in automatic shutoff feature of the radiofrequency unit. The ablation catheters (EP Technologies) simultaneously delivered the radiofrequency energy in a bipolar mode through the three tip-electrode pairs to produce the coronary sinus lesion. Epicardial atrial lesions were produced by simultaneously delivering radiofrequency energy through the four tip-electrode pairs. The energy was thus divided over the tip-electrode pairs and was not delivered over only the distal electrode of the catheter. Since all energy deliveries were bipolar, we did not use a backplate.
Measurement of Effective Refractory Period
The atrial effective refractory period (ERP) was determined at each electrode site by the extrastimulus technique with a programmable stimulator (Krannert Medical Engineering) and a constant current isolator. Each atrial test site was driven with a 2-ms rectangular stimulus that was twice the diastolic threshold and which was measured during each intervention. A train of eight stimuli (S1) was followed by a late premature stimulus (S2) that produced a propagated response. The S1-S1 interval was 280 to 300 ms and was kept constant throughout the experiment. The atrial responses to S1 and S2 were recorded in the lead II ECG and from the bipolar atrial electrode and were displayed on a storage oscilloscope. The S1-S2 interval was shortened in 1-ms steps until S2 failed to produce a propagated response. The ERP was defined as the longest S1-S2 interval at which S2 failed to produce a propagated response. The ERP was determined at least twice at each electrode site. If the values differed by >1 ms, the data were discarded and the determination was repeated. Since sustained AF may affect atrial refractoriness, vagally induced ERP shortening was measured at baseline, after a period of vagal stimulation, and after resolution of vagally induced sustained AF before performing RFCA. The atrial ERP alterations measured after resolution of vagally induced sustained AF before ablation were used as control data to compare with the ERP measurements that were obtained in the same fashion after RFCA. Atrial ERPs were measured at both isolated and nonisolated atrial sites.
AV-nodal ERP was determined by pacing the high right atrium with a train of eight stimuli (S1-S1 interval, 400 ms) followed by a late extrastimulus (S2) that produced a ventricular response. The S1-S2 interval initially was shortened by 5-ms steps and was shortened by 1-ms steps near the AV-nodal ERP. The longest S1-S2 interval at which S2 failed to produce a propagated ventricular response was considered to be the AV-nodal ERP.
Evaluation of Sinus-Node Automaticity and AV Conduction
Maximal heart rate (MHR) was determined by infusing graded doses of isoproterenol (0.9 to 9.7 μg/min). Sinus-node recovery time (SNRT) was determined by pacing the atria with a programmable stimulator (Krannert Medical Engineering, Indianapolis, Ind) for 30 seconds at progressively shorter drive-cycle lengths (400, 350, 300, and 250 ms) with a unipolar electrode placed ≈1 cm inferior to the sinus node. The longest time interval from the last paced atrial depolarization to the first spontaneous sinus cycle was recorded as the SNRT. Corrected SNRT (CSNRT) was determined by subtracting the mean sinus-cycle length (SCL) measured before the beginning of atrial pacing from the SNRT.
To investigate AV conduction, conduction time in the AV node (AH interval) was measured from the earliest onset of rapid right atrial activity recorded on the His-bundle electrogram to the onset of the His-bundle potential during constant atrial pacing at a cycle length of 400 ms. His-Purkinje conduction time (HV interval) was measured from the onset of His-bundle deflection to the beginning of the first ventricular depolarization recorded on any lead during constant right atrial pacing at a cycle length of 400 ms. Five consecutive AH and HV intervals were measured and averaged to obtain final values.
The dogs assigned to the acute studies (groups 1 through 4 and 6) were all autonomically decentralized. Since group 5 consisted of dogs for survival study, autonomic decentralization was not performed in these dogs. In the dogs in the acute groups (groups 1 through 4 and 6), the heart was autonomically decentralized by isolating, double-ligating, and transecting both cervical vagi and ansae subclaviae as exited from the stellate ganglia. In dogs assigned to group 1, both cervical vagi were stimulated through two Teflon-coated wire electrodes embedded in the cardiac end of each vagal nerve. Rectangular 4-ms pulses were delivered at a frequency of 8 to 20 Hz by use of separate constant-current isolators driven by a programmable stimulator (model SD-88, Grass Instrument Co). The current strength for low-level vagal stimulation was 0.05 mA greater than that required to produce sinus bradycardia (rate <75 beats per minute). The current strength for high-level vagal stimulation was 0.05 mA greater than that required to produce sinus arrest (>2 seconds) for the right vagus and complete AV block for the left vagus. The conditions of neural stimulation were kept constant in each experiment.
Histopathology of the Radiofrequency Lesions
All dogs were killed at the end of each study in the anesthetized state by excision of the heart. After the cardiac cavities were washed with saline, myocardial tissue was preserved in neutralized formaldehyde. After fixation, transmural tissue samples were taken from normal atrial tissue and tissue encompassing ablation sites. Sections were taken from the center of the lesion, fixed, and stained with hematoxylin and eosin or Masson’s trichrome stain for histopathological examination. For each ablation site, assessment of lesion size, transmural extent of necrosis, degree of inflammation, and fibrosis were made from dogs in both the acute and chronic groups.
Group 1: Effect of RFCA on Inducibility of AF and on Refractory-Period Response to Vagal Stimulation
In five dogs (group 1), reproducible sustained AF lasting >30 minutes was induced with rapid atrial pacing (pacing cycle length, 50 ms) or multiple premature atrial stimuli at various coupling intervals21 22 23 24 25 and was maintained by low-level (right vagus: 2 V, 4-ms duration, 8 Hz; left vagus: 3 V, 4-ms duration, 8 Hz) cervical vagal stimulation. The stimulation parameters were kept constant during each experiment. The procedure for atrial RFCA (described above) was performed, and the inducibility of AF was retested as in the baseline setting. Before and 30 minutes after completion of the RFCA, atrial ERPs were determined at isolated atrial sites, ie, atrial appendages and five different nonisolated left and right atrial sites in the baseline setting and during low-level and high-level (right vagus: 5 V, 4-ms duration, 20 Hz; left vagus: 7 V, 4-ms duration, 20 Hz) cervical vagal stimulation.
Group 2: Effect of RFCA on Inducibility and Duration of AF During Administration of Low and High Doses of Methacholine
In seven dogs (group 2), the effects of administration of a low dose (1.5 μg/kg per minute IV) and a high dose (3.6 μg/kg per minute IV) of methacholine on the atrial ERP were examined to determine whether the atrial myocardium was normally responsive after RFCA. Inducibility of AF was tested with atrial burst pacing (cycle length, 50 ms) or extrastimulus pacing during administration of low and high doses of methacholine before and after RFCA. Atrial ERP measurements were obtained during continuous infusions of low and high doses of methacholine before and after RFCA.
Group 3: Effect of RFCA on the Dose of Methacholine Required to Induce Sustained AF
In five dogs (group 3), we determined the relation between the dose of methacholine (0.2 to 6.5 μg/kg per minute IV) required to promote sustained AF and the inducibility and duration of AF before and after performing RFCA.
Group 4: Effect of RFCA on Sinus-Node Automaticity and AV Conduction
In six dogs (group 4), we assessed the effects of RFCA on sinus-node automaticity and AV-nodal conduction by determining SNRT, CSNRT, AH and HV intervals, and AV-nodal ERP before and after ablation.
Group 5: Long-term Effectiveness of RFCA
In four dogs in the chronic group (group 5), short-term AF was induced with rapid atrial pacing (pacing cycle length, 50 ms) or multiple premature atrial stimuli at various coupling intervals and perpetuated with a low dose of methacholine (1.5 μg/kg per minute). Since these dogs were kept alive, we minimized the amount of methacholine infused by stopping methacholine infusion after inducing AF for >10 minutes. Sterile open-chest atrial RFCA was performed after reproducible induction of sustained AF. Seven to 21 days later, all dogs underwent open-chest electrophysiological testing, in which the inducibility and duration of AF were retested during infusion of graded doses of methacholine (0.2 to 6.5 μg/kg per minute). Multiple atrial electrograms, a His-bundle electrogram, and an ECG of surface lead II were recorded during normal sinus rhythm and during pacing from isolated and nonisolated sites on the atria.
Group 6: Effect of RFCA on Atrial Contractile Function and Intra-atrial Conduction
In four dogs (group 6), cross-sectional and pulsed-wave Doppler echocardiographic examinations were performed before and after RFCA to assess the transvalvular flow-velocity spectra across the mitral and tricuspid valves before and after RFCA with a 2.5 MHz transducer and a Toshiba (Sonolayer SSA 270A) or Hewlett Packard (Sonos 1500) cardiac ultrasound imaging system. The sample volume was placed on the ventricular side at or near the tips of the mitral and tricuspid valve leaflets; small adjustments were made to record optimal transvalvular Doppler flow-velocity profiles in the apical four-chamber view. Echocardiographic recordings and an ECG of surface lead II were recorded on videotape, and video prints were obtained of optimal recordings. Diastolic flow across the mitral and tricuspid valves normally exhibits two peaks at slower sinus rates. However, the second peak of the flow-velocity curve (A wave), which corresponds to atrial contraction, is fused with the first peak (E wave) during normal sinus rate (120 to 130 beats per minute) in dogs. The sinus rate was slowed by 20% to 30% by vagal stimulation before and after atrial RFCA during echocardiographic assessment of the presence of atrial contraction (A wave).
To assess conduction across the radiofrequency lesions after RFCA, epicardial atrial electrograms were recorded from multiple sites on the right and left atrial appendages, low and high right atria, and left atrial free wall during sinus rhythm and pacing from the isolated regions, ie, the atrial appendages.
Data are presented as mean±SD, except for the duration of AF, which is presented as median (minimum, maximum). Comparisons within groups were done by paired t test, whereas a group t test was used for comparisons between groups. For comparisons among more than two groups, ANOVA was used. If ANOVA was significant, then the least significant difference test for multiple comparisons was used to define where the differences occurred. Since the duration of AF was not distributed normally, comparisons before and after ablation were made by the Wilcoxon signed rank test. The methacholine dose-response curves before and after ablation were compared by repeated-measures ANOVA. The interaction term from this analysis indicates whether the dose-response curves are similar before and after ablation.
Radiofrequency Energy Application
In group 1 dogs, the radiofrequency energy output was much higher than in dogs in the other experimental groups (Table 1⇓). Because the incidence of excessive rise in impedance (>300 Ω) in the first seconds of energy delivery was much greater in the group 1 dogs, and because RFCA studies with atrial tissue samples showed that a lower radiofrequency energy output setting and a longer exposure time (30 to 40 W, 60 seconds) reduced the incidence of excessive impedance rise and still created transmural lesions, radiofrequency energy output was reduced in subsequent studies in the other groups. This difference is not reflected in Table 1⇓, because exposure times <5 seconds were not included in the data analysis. Although the mean impedance level in groups 2 through 6 was higher with the lower radiofrequency energy output setting, the excessive rise in impedance (>300 Ω) was much less than in group 1 dogs. Fig 2⇓ shows a representative recording of actual power output during coronary sinus ablation (Fig 2⇓, top) and during epicardial atrial ablation (Fig 2⇓, bottom). The RFCA procedure, ie, positioning the catheters and delivering the radiofrequency energy pulses, took ≈30 to 40 minutes.
Effect of RFCA on Inducibility of AF and Refractory Period Response to Vagal Stimulation: Group 1
Fig 3⇓ shows that atrial RFCA attenuated vagally induced shortening of ERP at test sites in the right and left atria during both low-level and high-level cervical vagal stimulations. Test sites were both isolated, ie, atrial appendages, and nonisolated, ie, free wall, portions of the atria. The magnitude of decrease in ERP shortening at two left and three right atrial sites was most pronounced with high-level vagal stimulation (right atrium ΔERP, 30.8±8.7 ms before ablation versus 6.3±2.8 ms after ablation; P=.004, n=15 sites; left atrium ΔERP, 31.4±8.2 ms before ablation versus 4.2±1.6 ms after ablation; P=.001, n=10 sites). In contrast, shortening of the ERP elicited by infusion of a low dose or a high dose of methacholine remained unchanged after ablation at these atrial test sites (P=.25 for the right side and P=.19 for the left side by repeated-measures ANOVA for ablation effect). Bilateral high-level cervical vagal stimulation that caused sinus arrest or second degree or complete AV block before RFCA was markedly attenuated after ablation.
After epicardial atrial RFCA was performed, pacing induced atrial flutter in three dogs. After epicardial atrial ablations, endovascular application of radiofrequency energy through a hexapolar electrode configuration to the distal coronary sinus wall rendered atrial flutter noninducible in these dogs. Sustained AF became noninducible after ablation in all group 1 dogs. Only nonsustained AF was induced (duration, 5.6±4.1 minutes) after >10 minutes of concomitant high-level vagal stimulation and rapid atrial pacing at three times the diastolic threshold level.
Effects of RFCA on the Dose of Methacholine Necessary to Induce AF: Groups 2, 3, and 5
In group 2 dogs, sustained AF became noninducible after RFCA during infusion of a low dose (1.5 μg/kg per minute) of methacholine that induced sustained AF before ablation. AF was still inducible during infusion of a high dose (3.6 μg/kg per minute) of methacholine infusion.
To better quantify that relation, dose-response curves were obtained in group 3 and group 5 dogs (Fig 4⇓). In these dogs, RFCA shifted the dose-response curve relating the dose of methacholine necessary to maintain AF down and to the right. At the lowest doses of methacholine necessary to induce sustained AF, AF duration was significantly reduced after ablation in dogs in both the acute and chronic groups (P=.003, n=9). The inducibility and duration of AF after ablation were attenuated more markedly in group 5 (chronic) dogs compared with dogs in groups 2 and 3 (acute). However, at higher doses of methacholine (≥3.6 μg/kg per minute for group 2 and ≥6.0 μg/kg per minute for group 5), AF duration remained unchanged after ablation (Fig 4⇓).
One dog in group 2 had spontaneous paroxysms (>3 per hour) of AF with a mean duration of 8±3 minutes before ablation. After RFCA, no further paroxysms occurred. AF was also not inducible by infusion of a low dose of methacholine after ablation in this dog.
Effects of RFCA on Sinus-Node Function and AV Conduction: Group 4
Fig 5⇓ shows the effects of the RFCA on sinus-node function in group 4 dogs; spontaneous SCL remained unchanged, but the MHR was attenuated after ablation (ΔMHR, 17.6±7.7; P<.001, n=6), whereas SNRT and CSNRT were significantly prolonged after ablation (ΔSNRT, 74±22 ms; ΔCSNRT, 40.7±14.5 ms; P<.001; n=6). AV-nodal ERP and AH and HV intervals were not significantly affected (Table 2⇓).
Long-term Effectiveness of Atrial RFCA: Group 5
Nonsustained AF was inducible in all group 5 dogs (n=4) in the control state (duration, 11±32 seconds), and sustained AF (duration, >10 minutes) was inducible with infusion of low doses of methacholine (1.5 to 2.0 μg/kg per minute) before ablation. Seven to 21 days after RFCA in all 4 dogs, AF was only inducible with high doses of methacholine (≥6.0 μg/kg per minute, see Fig 4⇑). Sinus rhythm was preserved, and the PR interval remained unchanged 7 to 21 days after ablation. His-bundle recordings during the acute follow-up electrophysiological study 7 to 21 days after ablation showed normal AV conduction. In 1 dog, sustained atrial flutter (cycle length, 315 ms) with 1:1 AV conduction was inducible with rapid atrial pacing. Histopathologic examination of the coronary sinus wall in this dog showed a nontransmural lesion.
Effect of RFCA on Intra-atrial Conduction and Atrial Contractile Function: Group 6
Pulsed-wave Doppler ultrasound imaging study (group 6, n=4) of the flow across the mitral and tricuspid valves showed that atrial contraction (A wave) was still present in all 4 dogs after RFCA (Fig 6⇓). Right and left AV synchrony was preserved after ablation. Since the atrial contribution (A wave) during diastolic filling of the ventricles depends on many factors, eg, preload and afterload, we did not assess quantitative changes in atrial contractile function. No mitral or tricuspid valve dysfunction was observed after RFCA.
Epicardial atrial electrograms after ablation from multiple atrial sites during sinus rhythm and pacing from the isolated portions of the atria, ie, atrial appendages, showed no conduction across the lesions. The electrical activity recorded from the atrial appendages was dissociated from the rest of the atria (Fig 7⇓).
Histopathology of the Radiofrequency Lesions
Grossly, the ablated areas were well-circumscribed contiguous zones of discoloration that tended to be larger on the epicardial side than on the endocardial side (Table 3⇓). There was no hemopericardium, perforation, or rupture. Microscopically, the lesions consisted of distinct zones of coagulative necrosis bordered by zones of inflammation and fibroblast proliferation extending from adjacent normal tissues (Fig 8A⇓). In the centers of the lesions, the necrotic myocardial fibers remained “ghosts,” lacking nuclei and cross-striations (Fig 8B⇓). The zones of proliferating fibrous tissue at the borders of the lesions contained numerous mononuclear inflammatory cells and small blood vessels (Fig 8C⇓). Macroscopically, 1 to 2 cm of the distal part of the lesions showed a well-circumscribed area of discoloration extending from the coronary sinus wall to the surrounding fat tissue in the AV groove. Microscopically, the coronary sinus and surrounding tissues also exhibited coagulative necrosis and a thrombus was attached to the wall (Fig 8D⇓). In dogs that were killed immediately after treatment, no lesions were visible.
Development of the RFCA Procedure
The RFCA procedure we used is based on atrial lesions produced by the surgical-maze procedure,26 but this procedure does not replicate the exact pattern or number of these surgical incisions. The impetus was to abolish the induction of AF in dogs while eliminating the need for cardiopulmonary bypass or extensive atriotomies and ultimately to develop a closed-chest approach.
Our initial attempts focused on replication of all surgical-maze lesions with RFCA. Attempting to replicate the surgical lesions in close proximity to the sinus node affected the blood supply and the functional integrity of this important structure, causing sinus arrest or severe sinus bradycardia in some dogs. The surgical atrial incisions are sharp, narrow lines, whereas the epicardial atrial radiofrequency lesions are wider and may affect vascularization and innervation of the surrounding tissue to a greater extent. Consequently, we eliminated the lesions that produced sinus arrest but still could show that AF was noninducible at low levels of vagal stimulation and infusion of low doses of methacholine. This prompted us to develop an effective pattern of atrial lines of conduction block with a minimum of radiofrequency lesions. We did this by assessing inducibility of AF after performing RFCA at each lesion site. The resultant RFCA procedure (Fig 1⇑) may still reduce the mass of atrial myocardium that must be activated sufficiently to prevent occurrence of sustained multiple wavelets of reentry and thus abolish the inducibility of AF.
We demonstrated that our RFCA procedure eliminated induction of AF in dogs with atrial pacing and perpetuation of AF by use of either low-level cervical vagal stimulation or infusion of low doses of methacholine. The inducibility and duration of AF during high-level vagal stimulation and infusion of high doses of methacholine were not affected significantly by RFCA. It is likely that a high dose of methacholine would induce AF owing to microreentry in virtually any amount of atrial tissue.
The chronic studies indicate that the procedure is safe and effective for long-term prevention of induction of AF. Despite the fact that the radiofrequency lesions were transmural, atrial-wall rupture or coronary sinus perforation was not observed 7 to 21 days after ablation (Fig 8A⇑ through 8D). Furthermore, we showed that sinus rhythm and AV-nodal conduction were preserved and that inducibility of AF at low levels of vagal stimulation and methacholine infusion was abolished. In one dog, however, atrial flutter was inducible with rapid atrial pacing. Histopathologic examination of this dog showed that radiofrequency energy delivery to the coronary sinus did not produce a transmural lesion, which corroborates the results of the surgical-maze procedure by suggesting that transmural ablation of the coronary sinus wall is essential to preventing induction of atrial flutter after ablation.26
Randall et al27 28 showed with selective intrapericardial autonomic denervation of restricted portions of the canine heart, by use of either phenol application or a surgical ablation technique, that the vagal nerve fibers project to the heart by traversing along the superior vena cava, over the right atrium, and over the junction between the inferior vena cava and the inferior atrial surface. As anticipated from their experiments, we showed that the RFCA procedure attenuated the ERP shortening of nonisolated portions of both right and left atrial myocardium induced by efferent vagal nerve stimulation but did not affect the ERP response to intravenous methacholine. Since the radiofrequency ablations created contiguous transmural lesions, the vagal fibers projecting to the isolated regions, ie, the atrial appendages, were also interrupted, and, subsequently, the vagally induced atrial ERP shortening in isolated regions was totally abolished after RFCA. These data indicate that prejunctional vagal nerve fibers projecting to the atria are interrupted by radiofrequency ablations without affecting postjunctional responsiveness of the atrial myocardium to methacholine. The duration of the atrial ERP is an important factor for the vulnerability to AF.21 22 23 24 25 Since vagotomy prolongs the ERP of the atria or at least blunts its shortening, vagal denervation induced by RFCA may play a role in the prevention of AF induction. The surgical-maze procedure could produce a similar result. Intravenous infusion of methacholine substantially reduces the atrial ERP and potentiates the inducibility of AF. Importantly, however, RFCA still made it more difficult for methacholine to induce AF, thus indicating that, in addition to vagal interruption, the multiple lines of interatrial conduction block probably directly affect the atria.
Sinus-Node Function, AV Conduction, and Hemodynamics
The surgical-maze procedure maintains sinus-node function with preserved AV conduction.15 26 30 However, McLoughlin et al29 have demonstrated that the surgical-maze procedure causes acute sinus-node dysfunction that may improve spontaneously over the months following the procedure. RFCA prolongs the sinus-node recovery time and the corrected sinus-node recovery time and decreases the maximal heart rate significantly, corroborating this previous study by suggesting that interventions close to the sinus node or its blood supply cause acute sinus-node dysfunction. We did not assess sinus-node function in the chronic animal models, and future studies are needed to establish whether this RFCA procedure affects sinus-node function over the long term. AV-nodal and His-Purkinje conduction times were not affected significantly by RFCA in dogs in the acute and chronic groups. Pulsed-wave Doppler echocardiography showed that both right and left atrial transport function was preserved after RFCA (Fig 6⇑), indicating that ablation did not significantly affect atrial contractile function.
Potential Advantages of RFCA Compared With Surgery
In centers experienced with the technique, the surgical-maze procedure has a success rate of 89% for permanently eliminating AF without the need for antiarrhythmic medication and with restoration of sinus initiated rhythm, AV synchrony, and preservation of atrial booster-pump function.30 The surgical-maze procedure requires an intricate pattern of atrial incisions and extensive suturing to reestablish atrial integrity. The surgery is time consuming and requires prolonged cardiopulmonary bypass. Cox et al30 have reported a 3.9% rate of occurrence of neurological events (transient ischemic attacks or stroke) during or after surgery. These ischemic neurological events may be related to prolonged cardiopulmonary bypass and small air emboli during atriotomies. If the open-chest RFCA procedure abolished induction of AF in humans, it would eliminate the need for atriotomies and cardiopulmonary bypass. Naturally, if atrial RFCA can be performed endocardially, surgery can be eliminated altogether. That is the ultimate goal.
Consideration of the Experimental Model and Limitations of the Study
In the present study, we examined the effects of the RFCA procedure on the inducibility and duration of AF and on the parasympathetic innervation of the atria in a canine model. As in many previous studies, we used premature atrial pacing at different coupling intervals and burst-pacing atrial stimulation to induce AF. In contrast to the animal model used by Cox et al26 for the surgical-maze procedure, either electrical cervical vagal stimulation or infusion of methacholine was used to perpetuate AF. Despite differences between the animal models, the results of the present study show that the RFCA maze procedure reduced the ability to induce AF in the dog. Prevention of induction of AF may depend on the atrial lines of conduction block and vagal denervation produced by the RFCA procedure. Although atrial contractile function is preserved, we do not know whether RFCA increases the risk of thromboembolism, particularly in the atrial appendages. Paroxysmal AF induced by enhanced vagal tone is rare in humans, and, thus, it is possible that the observations of this phenomenon in the present canine model may not be applicable. This procedure also needs to be tested in dogs with self-perpetuating AF induced by chronic rapid atrial pacing or by some other means. It is important to remember that the dimensions of the canine atria are smaller than those of the human atria, which may conceivably affect the number of lesions required to prevent multiple wavelets of reentry from sustaining in patients. Recently, however, Shyu et al31 demonstrated that chronic AF caused by mitral valve disease can be eliminated surgically by creating a limited number of lesions that compartmentalize the enlarged atria. The results of this latter “compartment operation” corroborate our findings by suggesting that creation of a limited number of lines of interatrial conduction block can effectively abolish induction of AF.
This work was supported in part by the Herman C. Krannert Fund; by grants HL-42370 and HL-07182 from the NIH, US Public Health Service; by the American Heart Association, Indiana Affiliate, Inc; and by EP Technologies. The authors thank Douglas Segar, MD, for the echocardiographic studies, Naomi Fineberg, PhD, for statistical analysis of the data, and Jacob Rohleder for technical assistance.
Presented in part at the 15th Annual Scientific Sessions of the North American Society of Pacing and Electrophysiology, Nashville, Tenn.
- Received August 18, 1994.
- Revision received November 1, 1994.
- Accepted November 14, 1994.
- Copyright © 1995 by American Heart Association
Onundarson PT, Thorgeirsson G, Jonmundsson E, Sigfusson N, Hardarson T. Chronic atrial fibrillation: epidemiologic features and 14 year follow-up: a case control study. Eur Heart J. 1987;8:521-527.
Boston Area Anticoagulation Trial in Atrial Fibrillation Investigators. The effect of low dose warfarin on the risk of stroke in patients with nonrheumatic atrial fibrillation. N Engl J Med. 1990; 323:1505-1511.
Lodder J, Dennis MS, Van Raak L, Jones LN, Warlow CP. Cooperative study on the value of long-term anticoagulation in patients with stroke and nonrheumatic atrial fibrillation. Br Med J. 1988;296:1435-1438.
Killip T, Baer RA. Hemodynamic effects after reversion from atrial fibrillation to sinus rhythm by precordial shock. J Clin Invest. 1966;45:657-671.
Horowitz LN, Speilman SR, Greenspan AM, Mintz GS, Morganroth J, Brown R, Brady PM, Kay HR. Use of amiodarone in the treatment of persistent and paroxysmal atrial fibrillation resistant to quinidine therapy. J Am Coll Cardiol. 1985;6: 1402-1407.
Sodermark T, Jonsson B, Olsson A, Oro L, Wallin H, Edhag O, Sjogren A, Danielsson M, Rosenhamer G. Effect of quinidine on maintaining sinus rhythm after cardioversion of atrial fibrillation or flutter: a multicenter study from Stockholm. Br Heart J. 1975;37: 486-492.
Coplan S, Antman EM, Berlin JA, Hewitt P, Chalmers TC. Efficacy and safety of quinidine therapy for maintenance of sinus rhythm after cardioversion: a meta-analysis of randomized controlled trials. Circulation. 1990;82:1106-1116.
Langberg JJ, Chin MC, Rosenqvist M. Catheter ablation of the atrioventricular junction using radiofrequency energy. Circulation. 1989;80:1527-1535.
Cox JL, Boineau JP, Schuessler RB, Ferguson TB, Lindsay BD, Cain ME, Corr PB, Kater RM, Lappas DG. Review of surgery for atrial fibrillation. J Cardiovasc Electrophysiol. 1991;2:541-561.
Borggrefe M, Hindricks G, Haverkamp W, Budde T, Breidthardt G. Radiofequency ablation. In: Zipes DP, Jalife J, eds. Cardiac Electrophysiology: From Cell to Bedside. Philadelphia, Pa: WB Saunders Co; 1990:997-1004.
Haines DE. Determinants of lesion size during radiofrequency catheter ablation: the role of electrode-tissue contact pressure and duration of energy delivery. J Cardiovasc Electrophysiol. 1991;2:509-515.
Wittkampf FHM, Hauer RNW, Robles de Medina EO. Control of radiofrequency lesion size by power regulation. Circulation. 1989;80:962-968.
Haines DE, Verow AF. The impedance rise during radiofrequency ablation in vivo is prevented by maintaining an electrode tip temperature below the boiling point. Circulation. 1989;80(suppl II):II-41. Abstract.
Sykes C, Riley R, Pomeranz M, Rowan R, Imran M. Cooled tip ablation results in increased radiofrequency power delivery and lesion size. PACE Pacing Clin Electrophysiol. 1994;17(part II): 782-167.
Moe GK. On the multiple wavelet hypothesis of atrial fibrillation. Arch Int Pharmacodyn Ther. 1962;140:183-188.
Allessie MA, Bonke FIM, Schopman FJG. Circus movement in rabbit atrial muscle as a mechanism of tachycardia, III: the ‘leading circle’ concept: a new model of circus movement in cardiac tissue without the involvement of an anatomical obstacle. Circ Res. 1977;41:9-18.
Allessie MA, Lammers WJEP, Bonke FIM, et al. Experimental evaluation of Moe’s multiple wavelet hypothesis of atrial fibrillation. In: Zipes DP, Jalife J, eds. Cardiac Arrhythmias. Orlando, Fla: Grune and Stratton; 1985:265-276.
Rensma PL, Allessie MA, Lammers WJEP, Bonke FIM, Schalij MJ. The length of the excitation wave as an index for the susceptibility to reentrant atrial arrhythmias. Circ Res. 1988;62:395-410.
Moe GK, Abildskov JA. Atrial fibrillation as a self-sustaining arrhythmia independent of focal discharge. Am Heart J. 1959;58: 59-70.
Randall WC, Thomas JX, Barber MJ, Rinkema LE. Selective denervation of the heart. Am J Physiol. 1983;244:H607-H613.