Radiofrequency Catheter Ablation of the Atria Eliminates Pacing-Induced Sustained Atrial Fibrillation and Reduces Connexin 43 in Dogs
Background We assessed the effects of radiofrequency catheter ablation (RFCA) of the atrial epicardium on pacing-induced sustained atrial fibrillation (AF) and the expression and distribution of the intercellular gap junction protein connexin 43 (Cx43) in dogs.
Methods and Results In 12 mongrel dogs, after creation of complete AV block and implantation of a ventricular inhibited pacemaker, a high-rate pulse generator (20 to 30 Hz to induce AF) was implanted in the neck, connected to a right atrial endocardial pacing lead, and used to pace the atrium for 10 to 14 weeks. In group 1 (n=9 dogs), corrected sinus node recovery time (CSNRT), P-wave duration, 24-hour Holter ECG, maximal heart rate (MHR) in response to isoproterenol, and intrinsic heart rate (IHR) after atropine (0.04 mg/kg) and propranolol were measured before and after atrial pacing and RFCA. Group 2 dogs were used to assess the effect of chronic AF alone on Cx43 expression and distribution. All group 1 dogs developed sustained (>24 hours) AF. Right-sided RFCA of the atria eliminated the sustained AF in 5 dogs, but both right and left atrial RFCA was required to abolish sustained AF in the other 4 dogs. After RFCA restored sinus rhythm, CSNRT and P-wave duration were prolonged and MHR and IHR were decreased. Chronic rapid atrial pacing (group 2) increased the expression of Cx43, which was absent in ablated areas and markedly depressed in viable atrial myocytes near the ablation zones (group 1).
Conclusions Rapid atrial pacing for long time periods induced sustained AF that can be eliminated by linear right and left atrial lesions created with RFCA, with preservation of sinus rhythm and atrial contractile function. Chronic AF increased the expression and distribution of gap junction protein Cx43, which became reduced in ablated and nearby nonablated areas.
Atrial fibrillation remains a major challenge1 2 that, in selected cases, has been treated successfully by surgical compartmentalization of the atria to create isolated regions in which the amount of atrial myocardium is insufficient to sustain multiple wavelets of reentry.3 4 5 6 7 8 We have shown previously that creation of a limited number of atrial lesions by use of an RFCA procedure reduced the inducibility and duration of AF induced by cervical vagal stimulation or intravenous infusion of low doses of methacholine.9 Because vagally induced AF may not be common10 11 and the results of this study not generalizable to large numbers of patients with AF,12 13 we sought to test this ablation technique on sustained AF induced by chronic rapid atrial pacing.14 We also determined the changes in the expression and distribution of the intercellular gap junction protein Cx43 in ablated and nonablated regions of the atrial myocardium,15 16 17 18 19 on the premise that such changes would affect conduction and might be partly responsible for the effectiveness of the RFCA technique.
Twelve conditioned mongrel dogs of either sex weighing 20 to 30 kg were initially sedated with a cuffed endotracheal tube and ventilated with 2% to 3% isoflurane mixed with oxygen at a rate of 1.5 to 3.0 L/min. We produced AV junction ablation with RFCA and then implanted a ventricular demand pacemaker as described previously.20 In 9 (group 1) of the 12 dogs, we also assessed sinus node automaticity by measuring SNRT, MHR, IHR, intra-atrial conduction from P-wave duration and PA interval, and atrial ERP as described previously.20 The remaining 3 (group 2) dogs were used to study the effect of chronic rapid atrial pacing on Cx43 expression.
Holter Recording to Determine the Inducibility and Duration of AF (Groups 1 and 2)
At the EPS-1, the inducibility and duration of AF were assessed by burst pacing or extrastimulus pacing of the atria. Because the duration of induced AF was short (<2 minutes) at this first study, a 24-hour Holter recording was not obtained. After 2 to 6 weeks of rapid atrial pacing, the atrial pulse generator was turned off and 24-hour Holter recordings using two-channel tape recorders (Marquette Electronics) and two bipolar leads were obtained in all 12 dogs.
Rapid pacing was resumed, and after 10 to 14 weeks, the atrial pulse generator was turned off again and 24-hour Holter recordings were repeated in 12 dogs.
Echocardiography (Group 1)
Cross-sectional and Doppler echocardiographic examinations were performed to assess the right and left atrial dimensions and transvalvular flow-velocity spectra across the mitral and tricuspid valves with a 2.5-MHz transducer and a Toshiba (Sonolayer SSA 270A) or Hewlett Packard (Sonos 1500) cardiac ultrasound imaging system. Echocardiographic images were obtained in the apical four-chamber, parasternal long-axis, and parasternal short-axis planes. Atrial size was assessed by calculation of the atrial area by planimetry in the apical four-chamber view. The measurements of atrial area were obtained in triplicate and averaged for each dog. The transvalvular flow was assessed by placement of the Doppler sample volume on the ventricular side at or near the tips of the mitral and tricuspid valve leaflets in the apical four-chamber view.
Macroscopic Histopathology of the RF Lesions (Group 1)
All dogs were killed in the anesthetized state at the end of the electrophysiological follow-up study after ablation by excision of the heart. After the cardiac cavities were washed with saline, the hearts were preserved in neutralized formaldehyde. Macroscopic anatomic examination of the lesions was performed. For each lesion, data on lesion length, width, and transmural extent were obtained.
Histology and Immunohistochemistry Techniques
Right atrial tissue was harvested from anesthetized open-chest dogs and immersion-fixed in 10% neutral buffered zinc formalin (Anatech). Tissue blocks were embedded in paraffin and sectioned at 9 to 10 μm, and alternate sections were stained with HE. Photomicrographs were made with a Leitz FU microscope. Occasionally, elastin and connective tissue elements were stained with a combined Gomori’s aldehyde fuchsin with a trichrome counterstain.21 Unstained sections were deparaffinized and processed for localization of Cx43 and the surface membrane.22 23 24 Sarcolemmal staining involved binding of fluorochrome-conjugated WGA by a modification of the method of Dolber et al.17 Sections were incubated 60 minutes at 22°C to 24°C with 25 μg/mL of rhodamine-labeled WGA (Vector Laboratories) in PBS (10 mmol/L phosphate buffer at pH 7.4 in 2.7 mmol/L KCl, 137 mmol/L NaCl) to which 1 mmol/L CaCl2 was added. Sections were washed several times with PBS to eliminate unbound WGA. For immunostaining of Cx43, sections were incubated with 0.1% NP40 in PBS for 30 minutes, then incubated 12 to 15 hours overnight with affinity-purified anti-Cx43 antibody at a 1:500 dilution. Preparation and purification of the antibody to residues 367 to 379 of Cx43 have been described previously.15 Bound antibody was identified by goat anti-rabbit IgG coupled to FITC. Slides were rinsed three times in PBS-BSA, then FITC-conjugated secondary antibody (Boehringer Mannheim) was added at 1:200 dilution for 1 hour. After two washes with PBS-BSA, slides were covered with 3% N-propyl gallate in 90% glycerol and examined.
A Bio-Rad MRC-600 confocal microscope was used to localize different fluorescein-labeled antibodies bound to atrial cells. Immunostained cells were scanned with a Kr/Ar laser at 488 and/or 568 nm, and optical sections were made at 1-μm increments with a stepper motor. To increase the signal-to-noise ratio, images were obtained from five to seven scans with Kalman averaging. Optical sections at 1-μm steps were collected from the entire 9-μm thickness of the sample. Commercial software (Bio-Rad COMOS system) was used to merge all serial sections into one composite image. Colocalization of Cx43 and WGA within the same focal plane was determined by merging dual-channel scans of identical fields. Images were viewed on-screen and printed on a Sony videoprinter or by import of digitized data into a graphical routine (Aldus Photoshop).
Experimental Protocols (Group 1)
EPS-1 was performed after the animals were allowed to recover from surgery for 3 days. Anesthesia was induced with sodium thiopental (30 mg/kg) and maintained with 2% to 3% isoflurane mixed with 2 L/min oxygen. A 6F quadripolar catheter was placed percutaneously in the high right atrium under fluoroscopic guidance. Lead II ECG, bipolar high right atrial electrograms, and arterial blood pressure were recorded continuously during the study and stored on a videotape for later analysis of the data with a computer software program (Lab View, National Instruments). P-wave duration and PA interval measurements were obtained as described earlier.20 Inducibility and duration of AF were assessed by premature atrial pacing or rapid atrial pacing. The atrial ERP was determined at a high right atrial site close to the atrial pacing lead insertion site. The same atrial ERP determination site with the atrial lead insertion site as an anatomic landmark was used during the subsequent electrophysiological follow-up studies in each animal.
After atrial ERP measurements were obtained, the catheter tip was positioned in the sinus node region under fluoroscopic guidance. SNRT was measured and MHR determined after isoproterenol infusion. The dogs were then allowed to recover for 40 to 50 minutes. Muscarinic- and β-receptor blockade were then induced by intravenous injection of atropine and propranolol to determine the IHR. The SNRT determinations were repeated during autonomic blockade. At the end of the study, the atrial pulse generator was turned on and the atria were paced (rate, 20 Hz; 4.5-ms impulse width and three times diastolic threshold) for 2 to 6 weeks.
After 2 to 6 weeks of high-rate atrial pacing, the atrial pulse generator was turned off and the duration of AF was assessed by 24-hour Holter ECG monitoring in 12 dogs. EPS-2 was performed after spontaneous termination of pacing-induced sustained AF. The dogs were in sinus rhythm for 6 to 12 hours before data on sinus node function and atrial refractoriness were obtained. EPS-2 was performed under general anesthesia, and atrial ERPs, SNRTs, and IHR and MHR data were obtained in the same fashion as during EPS-1. After completion of EPS-2, rapid atrial pacing was reinitiated and continued for 10 to 20 weeks to induce sustained AF.
RFCA of the Atria
The atrial pulse generator was turned off after 10 to 14 weeks of rapid atrial pacing, and a 24-hour Holter ECG was obtained to assess the duration of pacing-induced sustained AF. Rapid atrial pacing was then reinitiated and continued while the dogs underwent RFCA of the right atrium and the coronary sinus wall as described below. Immediately after completion of the right-sided ablation procedure, rapid atrial pacing was stopped, and the duration of pacing-induced AF was reassessed. If AF lasted for >15 minutes, the ablation procedure was considered unsuccessful, and rapid atrial pacing was reinitiated and continued until the next left atrial RFCA procedure carried out 1 week later. If AF stopped, three attempts using premature and burst right atrial pacing were tried to reinduce the AF. If pacing-induced AF lasted for <15 minutes, the ablation procedure was considered successful. In these dogs, rapid atrial pacing was reinitiated and continued until the next electrophysiological follow-up study. All dogs were allowed to recover from the surgical intervention for 1 week, after which rapid pacing was again stopped.
The procedure for RFCA of the atria was performed in the following steps. The dogs were premedicated with antibiotics (cefazoline 1 g IM), given heparin 2500 U IV, and anesthetized with sodium thiopental 30 mg/kg IV. Anesthesia was maintained with 2% to 3% isoflurane gas mixed with oxygen at a rate of 1 to 2 L/min. A right intercostal thoracotomy was performed, the ribs and lungs were retracted, and the pericardium was incised. A pericardial window was created to expose the right atrium. The epicardial ablations were then produced by manual positioning of a hexapolar ablation catheter on the epicardial surface of the atria under direct visualization to ensure optimal tissue contact and energy delivery. We chose this approach rather than an endocardial approach to be sure that all lesions were transmural and that if failure to terminate AF occurred, it was not due to technique. RF energy was delivered (30 to 40 W for 60 seconds, EP Technologies) in a bipolar fashion simultaneously between adjacent tip electrode pairs. Excessive impedance rise was prevented by cooling the ablation electrodes with small amounts of normal saline during the RF energy delivery. RF energy was delivered in this fashion epicardially to the posterolateral right atrium (Fig 1E⇓), the anterolateral side of the SVC (Fig 1D⇓), around the right atrial appendage (Fig 1A⇓), and around the right-sided portion of the transverse sinus (Fig 1C⇓). A 6F catheter with a hexapolar multielectrode steerable tip was then introduced percutaneously through the right external jugular vein, and the distal tip was advanced into the middle part of the coronary sinus. RF energy was delivered in a unipolar fashion between the catheter tip electrode and a dispersive patch electrode placed on the lateral chest wall. RF energy pulses were delivered (15 to 20 W for 30 seconds) to the coronary sinus wall sequentially through the most distal tip and the second distal electrodes, respectively. The tip of the ablation catheter was then pulled back to the ostium of the coronary sinus, and RF pulses were delivered to the posterior and posteroseptal part of the right atrium sequentially through the six electrodes of the ablation catheter. If the animals remained in AF 1 week after the right-sided ablations were performed, the left-sided ablations, ie, around the left atrial appendage (Fig 1B⇓) and the left-sided portion of the transverse sinus (Fig 1C⇓), were produced (Fig 1⇓). The number of lesions was kept to the minimum necessary to abolish pacing-induced sustained AF. Therefore, in some dogs, only lesions in the right atrium were sufficient, whereas in others, RFCA of both atria was necessary to eliminate pacing-induced sustained AF.
EPS-3 was performed 1 week after recovery from the RFCA procedure. The dogs were anesthetized with sodium pentobarbital, and anesthesia was maintained with additional amounts of sodium pentobarbital. The dogs were intubated and ventilated with room air with a volume-cycled respirator (Harvard Apparatus). The chest was opened via a median sternotomy, and the heart was suspended in a pericardial cradle. A fluid-filled cannula placed in the femoral artery was connected to a Statham P23 ID transducer (Gould) to monitor arterial blood pressure, and a femoral venous cannula was used to infuse normal saline at 100 to 200 mL/h to replace spontaneous losses. The dog was placed on a heating pad, and the thoracotomy was covered by a plastic sheet. An operating room lamp was used to maintain epicardial temperature between 37°C and 39°C. Bipolar stainless steel plunge electrodes made from Teflon-coated wires, insulated except for their tips, were inserted in the right and left atrial myocardium. The recording electrodes were placed in the high right atrium near the sinus node region, low right atrium, right atrial appendage, left atrial free wall, and left atrial appendage. To assess conduction across the RF lesions after RFCA, epicardial electrograms from these latter recording sites were obtained during normal sinus rhythm and pacing from the isolated regions, ie, atrial appendages. A bipolar pacing electrode was inserted near the insertion site of the atrial pacemaker lead to determine the atrial ERPs. Atrial ERPs, SNRTs, and IHR and MHR data were obtained in the same fashion as during EPS-1 and EPS-2.
The data are expressed as mean±SD, except for the duration of AF, which is presented as median (minimum, maximum) and the data in Fig 10⇓, which are mean±SEM. For statistical analysis, two-way repeated-measures ANOVA was used, with each animal serving as its own control. Because the duration of AF was not distributed normally, comparisons before and after chronic pacing were made by Wilcoxon signed-rank test. A value of P<.05 was set as statistically significant.
Effect of the RFCA Procedure on Chronic Rapid Atrial Pacing–Induced Sustained AF (Group 1)
Electrical activity recorded from the high and low right atrium had a median fibrillation cycle length of 143±5.3 ms during induced AF before initiation of chronic rapid atrial pacing (Fig 2⇓). Pacing-induced AF lasted <30 seconds. After 2 weeks of rapid atrial pacing, the median AFCL shortened to 107±4.9 ms and the median duration of AF increased to 2.1 hours. After 10 to 14 weeks of rapid atrial pacing, the median AFCLs were 98±3.7 ms in the high and low right atrium (Fig 2⇓). Twenty-four-hour Holter ECG recordings obtained after 10 to 14 weeks of rapid atrial pacing showed that spontaneous AF after rapid atrial pacing was turned off was sustained, ie, >24 hours.
Pacing-induced sustained AF was converted to sinus rhythm in <15 minutes with only right-sided RFCA in 5 animals, ie, RF lesion extending from the lateral side of the SVC to the IVC, the medial side of the SVC to the right atrial appendage, and around the right atrial appendage and RF lesion in the coronary sinus (Fig 1⇑, Table 1⇓). The difference in AFCL preablation between the 4 dogs (median AFCL, 94.8±1.2 ms) requiring right and left atrial RFCA compared with the 5 dogs (median AFCL, 98.2±4.1 ms) requiring only right atrial RFCA for successful termination of pacing-induced AF was not statistically significant (P=NS). We did not perform left heart catheterization in these dogs; therefore, measurements of the difference in AFCL and ERP between the left and right atria could not be obtained. Right-sided RFCA of the atrium acutely reduced the duration of AF after rapid atrial pacing was turned off to 559±212 seconds in 5 dogs that converted to sinus rhythm. Immediately after right-sided RFCA and before conversion to sinus rhythm, median AFCL was 127±2.6 ms and sustained AF (>15 minutes) was not inducible after at least three attempts with premature or burst pacing of the atria in these 5 dogs (Table 1⇓). Fig 3⇓ shows a representative histogram example (dog 1) of the increase in right atrial AFCL documented immediately and 1 week after ablation. In these 5 animals, sustained AF was not inducible with burst pacing of the atria, despite maintenance of rapid atrial pacing 1 week after right-sided RFCA, and the bipolar atrial electrograms recorded from the same right atrial sites after rapid atrial pacing was turned off had a median AFCL of 134±3.6 ms 1 week after ablation.
In the remaining 4 animals, after completion of the right-sided RFCA, pacing-induced AF did not convert to sinus rhythm in <15 minutes after rapid atrial pacing was turned off. The median AFCL just after right-sided RFCA was shorter in these dogs (median AFCL, 121±4.3 ms) than in the 5 dogs that converted to sinus rhythm in <15 minutes after right-sided RFCA only (median AFCL, 127±2.6 ms, P=.02). The ablation procedure was considered unsuccessful in these 4 dogs because the AF lasted >15 minutes after right-sided RFCA, and rapid atrial pacing was reinitiated and continued until the left atrial RFCA session 1 week after right atrial RFCA.
Left-sided RFCA of the atria, ie, RFCA lesions around the left atrial appendage and in the sinus transversus (Fig 1⇑), reduced the median duration of pacing-induced AF to 259 seconds. Rapid atrial pacing was continued in these 4 animals until the next EP study 1 week after left-sided RFCA. Sustained AF was no longer inducible with extrastimulus or burst pacing of the atria after at least three attempts. In 2 dogs, atrial flutter/tachycardia with a cycle length of 215±17 ms was inducible and terminated with burst pacing. Repeating the RFCA lesion from the SVC to the IVC in 1 dog and from the left atrial appendage to the sinus transversus in the other dog abolished the inducibility of this atrial tachyarrhythmia. Fig 4⇓ shows a representative histogram example (dog 7) in which AFCL prolongs just after and 1 week after (not shown) right-sided RFCA (median AFCLs, 124 and 135 ms, respectively) and further prolongs immediately after and then 1 week after left-sided RFCA (median AFCLs, 139 ms and 146 ms, respectively). Table 1⇑ summarizes the AF characteristics of each dog before ablation and after ablation.
In the 4 dogs requiring both right- and left-sided RFCA, pacing-induced AF did not terminate after the right-sided thoracotomy and right-sided RFCA. These dogs served as their own controls to exclude the possible antifibrillatory effects of thoracotomy and manipulation of the heart or the anesthesia.
Effect of Pacing-Induced Sustained AF and the RFCA Procedure on Atrial Refractoriness and Intra-Atrial Conduction (Group 1)
The atrial ERP was significantly shorter after 2 to 6 weeks of pacing-induced AF compared with baseline, ie, before initiation of chronic rapid atrial pacing. The curve relating the basic drive cycle length and the duration of atrial ERP showed shortening of atrial ERP with shortening of the basic drive cycle lengths at baseline. However, this latter curve was shifted downward at all basic drive cycle lengths after 2 to 6 weeks of pacing-induced AF (Fig 5⇓, left). The duration of atrial ERPs became independent of basic drive cycle length after 2 to 6 weeks of rapid atrial pacing. After 10 to 14 weeks of rapid atrial pacing, sustained AF (duration >24 hours) was very easily inducible; therefore, we could not obtain any ERP data after 10 to 14 weeks of rapid atrial pacing. Atrial ERPs measured 1 to 2 weeks after the RFCA procedure showed normal cycle length dependency, ie, ERP shortening with shorter basic drive cycle lengths (Fig 5⇓, left).
The P-wave duration and PA interval prolonged after 2 to 6 weeks of rapid atrial pacing and further prolonged after the RFCA procedure (Fig 5⇑, right, P<.05). Bipolar atrial electrograms recorded after ablation from multiple sites in the right and left atria during sinus rhythm and pacing from the isolated portions of the atria showed no conduction across the RFCA lesions. Fig 6⇓ shows that the electrical activity recorded from the atrial appendages was dissociated because of block from the rest of the atria.
Effect of Pacing-Induced Sustained AF and the RFCA Procedure on Atrial Dimension and Hemodynamics (Group 1)
After 10 to 14 weeks of rapid atrial pacing and before RFCA, the left atrial area increased from 8.4±1.4 to 13.1±2.1 cm2 (56%) and the right atrial area from 5.2±1.3 to 8.0±1.6 cm2 (54%, P<.001). Left and right atrial areas were significantly larger in the group of dogs requiring right- and left-sided RFCA to terminate pacing-induced sustained AF than in the 5 dogs requiring only right-sided RFCA (left atrial area, 15.7±2.1 versus 11.6±1.7 cm2 and right atrial area, 7.4±1.4 versus 8.7±1.5 cm2, P<.01). Because the position of the heart in the thoracic cavity changed after the RFCA procedure, we did not measure left atrial and right atrial dimensions after ablation. Pulsed-wave Doppler echocardiographic recordings of the flow across the mitral and tricuspid valves showed that at physiological PR intervals, atrial contraction (A wave) during diastolic filling of the ventricles was still present after the RFCA procedure. Because atrial contribution to the diastolic filling of the ventricles depends on many factors, eg, preload and afterload, no quantitative data on changes in atrial contractile function were obtained.
Effect of Rapid Atrial Pacing–Induced AF and RFCA of the Atria on Sinus Node Function (Group 1)
Spontaneous sinus cycle length increased after rapid atrial pacing for 2 to 6 weeks. Atropine and propranolol administration further enhanced this latter difference (Fig 7⇓, left). The CSNRTs prolonged just after rapid atrial pacing for 2 to 6 weeks (Fig 7⇓, middle). RFCA of the atria further increased the spontaneous sinus cycle length before and after induction of pharmacological autonomic blockade (Fig 7⇓, left). RFCA of the atria further prolonged the CSNRTs compared with values obtained before ablation (Fig 7⇓, middle). Pharmacological autonomic blockade with atropine and propranolol further increased these latter differences in sinus cycle length and CSNRTs (Fig 7⇓, left and middle). MHR during infusion of isoproterenol was decreased after the RFCA procedure (Fig 7⇓, right).
Macroscopic Histopathology of the RF Lesions (Group 1)
The changes noted were very similar to our previous findings.9 In brief, macroscopically, the ablated areas were well-circumscribed contiguous zones of discoloration that tended to be larger on the epicardial side than on the endocardial side. The midportion of the coronary sinus showed a well-circumscribed, contiguous area of discoloration extending from the coronary sinus wall to the surrounding fat tissue in the AV groove. There was no hemopericardium, perforation, or rupture. Table 2⇓ summarizes the anatomic data on the RF lesions.
Histology and Immunohistochemistry (Groups 1 and 2)
In normal canine right atrial tissue, rhodamine-conjugated WGA produced fluorescent labeling of the entire circumference of atrial myocytes (Fig 8⇓). Longitudinal sections yielded roughly rectangular fluorescent profiles (Fig 8A⇓ and 8C⇓) in which the surface membrane and interior regions were readily distinguished. Immunostaining of the same sections for Cx43 showed colocalization with surface-bound WGA for most of the immunoreactive Cx43. The majority of Cx43 was identified in end-to-end junctions typical of intercalated disks (Fig 8C⇓), but some was also found along the lateral sides of the myocytes (Fig 8D⇓). A very similar distribution of immunofluorescent staining of Cx43 in normal dog atrium has been reported previously by us and others.15 16
We studied the effects of chronic atrial pacing alone on Cx43 protein expression (group 2). AF was induced by rapid atrial pacing, as described in the “Methods” section, but no RF ablation was performed in these 3 dogs. The atrial samples from the right atria were processed and stained with Cx43 antibody as well as WGA. The results demonstrated that Cx43 protein expression was increased in pacing-induced AF dogs compared with normal sinus rhythm (Fig 9B⇓). Fig 10⇓ displays the quantitative differences of Cx43 expression between normal sinus rhythm, pacing-induced chronic AF, the ablated area, and the area ≈1 to 2 cm away from the ablated center.
After ablation of the atrium, well-circumscribed regions of discolored myocardium were grossly visible, extending from the location of the coronary sinus to the surrounding fat pad in the AV groove. No hemopericardium, perforation, or rupture of the atrium was observed. HE-stained sections showed myocyte loss and replacement with fibrous tissue throughout the entire thickness of the ablated regions. Bound WGA no longer was noted over the entire circumference of the myocytes. Instead, WGA labeling showed irregular staining in a disorganized pattern (Fig 9⇑). In the border zone adjacent to the ablation, the WGA staining of the surface membrane showed discontinuities in the fluorescent cell outline and distortions of the shape of the atrial cells. Immunoreactive Cx43 was nearly absent in the fibrotic regions that had been ablated and was significantly decreased in the border zone as well (Figs 9⇑ and 10⇑). Frequently, islands of apparently normal atrial myocytes (several hundred micrometers to 1 mm in diameter) were found in the midst of the fibrous tissue (Fig 11A⇓). Although the islands of atrial myocytes appeared normal by HE, WGA staining incompletely outlined the cell surface. Discontinuities of the WGA staining were widespread. Costaining for Cx43 revealed loss of Cx43 in the fibrotic zones (Fig 11C⇓). Immunoreactive Cx43 was abundant at the surface of the clusters of surviving myocytes but no longer appeared in the linear arrays typical of intercalated disks. Instead, Cx43 was found in a globular pattern densely interspersed among the surviving atrial myocytes (Fig 11D⇓).
In this animal model of AF produced by sustained rapid atrial pacing,14 20 25 compartmentalizing the right atrium or both atria by RFCA abolishes pacing-induced sustained AF and preserves sinus-initiated rhythm and atrial contractile function. Pacing-induced sustained AF alone prolonged sinus cycle length and CSNRT and decreased IHR and MHR, as we have shown previously.20 After chronic rapid atrial pacing, the duration of atrial ERPs was significantly shortened and rendered independent of the basic drive cycle length. One week after the RFCA procedure, the indices of sinus node function, ie, sinus cycle length and CSNRT, further prolonged both before and after induction of pharmacological autonomic blockade, and MHR in response to isoproterenol infusion decreased further after ablation. At this time, however, the duration of atrial ERPs reversed to baseline and showed normal cycle length dependency. The gap junction protein Cx43 distribution and density in atrial myocardium was increased by chronic rapid atrial pacing before RFCA and was absent at ablated areas and significantly reduced in viable areas close to the ablated sites after ablation.
Effect of RFCA on the Inducibility of Pacing-Induced Chronic AF
In this study, sustained AF (AF duration >24 hours) was produced by chronic rapid atrial pacing for 10 to 14 weeks. Wijffels et al14 showed that in goats, intermittent burst atrial pacing markedly shortened the atrial ERP and wavelength during AF without significantly increasing atrial dimensions so that multiple reentrant wavelets could propagate through the atria, favoring perpetuation of AF. The rationale for long, linear, contiguous lesions in the atria to treat AF9 follows the work of Cox et al.3 4 5 Compartmentalizing the atria creates interatrial lines of conduction block and reduces the amount of atrial myocardium to be electrically activated at any one time so that the isolated atrial regions cannot support the number of reentrant wavelets required to perpetuate AF, thereby increasing the statistical probability of simultaneous extinction of the wavelets.
In the present study, only right-sided RFCA eliminated pacing-induced AF in 5 dogs. Although in the remaining 4 dogs, the right atrial electrograms were more organized after right-sided RFCA, left-sided lesions were necessary to abolish pacing-induced AF. Both the right and left atrial areas were larger and the median AFCL was shorter in the dogs requiring both right and left atrial RFCA compared with those requiring only right-sided RFCA to terminate pacing-induced AF. Haissaguerre et al12 used only right-sided RFCA to effectively treat a patient with paroxysmal AF with predominantly organized electrical activity during AF. They suggested that the right atrium played an important role in the genesis of their patient’s AF. Conversely, Swartz et al13 reported that only right-sided RFCA was insufficient to terminate chronic AF in their patients. Subsequent observations by Haissaguerre et al29 are consistent with that observation.
The AFCL was still relatively short immediately after ablation and prolonged to baseline values 1 week after ablation (Table 1⇑). Because the atria were in a vulnerable period, ie, short AFCL and short refractory period immediately after ablation, the cutoff point for successful ablation was arbitrarily set at <15 minutes of pacing-induced AF immediately after ablation.
Effect of RFCA on Sinus Node Function, Atrial Electrophysiology, and Hemodynamics
We have previously reported20 that long-term rapid atrial pacing–induced sustained AF impaired sinus node function. In the present study, 10 to 14 weeks of rapid atrial pacing produced sustained AF (duration >24 hours) in all dogs. The RFCA procedure was then performed. After the RFCA procedure, sinus cycle length and CSNRTs were prolonged and MHR in response to isoproterenol infusion decreased. However, we did not restore sinus rhythm and test sinus node function and intra-atrial conduction before the RFCA procedure, as we did in our previous study20 ; therefore, the sinus node dysfunction and prolonged intra-atrial conduction after the RFCA procedure may be due to both the effects of pacing-induced sustained AF and the RFCA lesions. Atrial ERPs returned to the values measured before ablation and showed normal cycle length dependency 1 to 2 weeks after ablation, at a time when sinus node function was still depressed. We have shown previously that the RFCA procedure causes efferent vagal denervation of the atria.9 Because vagal denervation prolongs atrial refractoriness or at least attenuates its shortening, it is possible that in the present study, parasympathectomy contributed in part to the increase in atrial refractoriness after the RFCA procedure.
Histology and Immunohistochemistry
The receptor sugar for WGA is N-acetylglucosamine, with preferential binding to multimeric forms (dimers, trimers) of this sugar. Consequently, WGA can bind oligosaccharides common to those found in membrane glycoproteins. Incubation of sections of atrial myocardium with rhodamine-conjugated WGA produced a fluorescent labeling of the entire circumference of normal atrial myocytes. Fluorescent WGA also stained the matrix in the interstitium around the myocytes as well as vascular structures like capillaries.17 Costaining of Cx43 and WGA-bound lectins enabled us to identify Cx43 at the surface membrane versus the interior regions of the myocytes. In normal atrium, the vast majority (but not all) of the immunoreactive Cx43 was found at the cell surface (as demarcated by fluorescent WGA binding). WGA staining also helped to identify the orientation of atrial myocytes (longitudinal versus transverse axis) to aid in assignment of Cx43 localization in end-to-end junctions versus those along the lateral sides of the myocytes.
We have found that pacing-induced AF increased Cx43 protein expression (group 2). It is not clear whether the increase was caused by an increment in Cx43 mRNA transcription or by a mechanism in which the metabolism of the protein was changed. An increase in intracellular Ca2+ concentration could reduce gap junction conductance and cause an increase in Cx43 expression. Further, we have found that the side-to-side junctions seemed to be increased in number in the fibrillating atria, suggesting a change in the distribution of gap junctions. Whether these changes contribute to the occurrence of AF is unknown, but it has been demonstrated that conductance between neighboring cells is significantly enhanced by overexpression of gap junction protein in Xenopus oocytes.26 27 Although other factors may play a role in the development of sustained AF, the distribution and functional changes in gap junction channels may be an important factor to be considered. One could postulate that the increase in gap junction protein expression contributes to developing multiple wavelets of reentry. Although an increase in gap junction formation should enhance atrial conduction, we have found evidence that it may be depressed, on the basis of the prolonged P-wave duration and PA interval. We have found this to occur in another study20 as well, whereas Wijffels et al14 found no change in conduction over Bachmann’s bundle. An increase in Cx43 is not incompatible with this observation because gap junctions, though increased, may not be normal. Further, an increase in gap junctions and/or a change in distribution may impair orderly impulse propagation and prolong conduction times. These possibilities are presently being explored by mapping techniques in our laboratory.28
Seven days after RF ablation of the atrium, WGA staining was disorganized both within the fibrotic tissue replacing the ablated myocytes and in the border zone surrounding the site of ablation, most likely because of loss of glycoproteins within these zones so that WGA could no longer bind to the atrial sarcolemma. Not surprisingly, immunoreactive Cx43 was lost in the fibrotic regions, as might be expected for a tissue that represents a barrier to cardiac conduction. However, Cx43 was also markedly reduced in the adjacent atrial myocardium surrounding the region of ablation. Although immunolocalization of connexins does not indicate the actual changes in conduction, the reduction of Cx43 in the border zone suggests that the effects of RF ablation on gap junctional transmission extend beyond the localized area of myocyte injury. Interestingly, surviving clusters of atrial cells within the fibrotic zone had a significant increase in the amount of Cx43 expression, but it was accompanied by redistribution to intercellular loci in a haphazard globular pattern. Such a disorganized pattern of Cx43 distribution might be expected to redirect any electrical impulses that might be transmitted to these surviving cells. Perhaps broader areas of block produced by the ablation lesion may extend beyond the regions directly contiguous with catheter contact due to Cx43 remodeling 1 to 2 cm away.
Selected Abbreviations and Acronyms
|CSNRT||=||corrected sinus node recovery time|
|EPS-1||=||first electrophysiological study before atrial ablation|
|EPS-2||=||second electrophysiological study before atrial ablation|
|EPS-3||=||electrophysiological follow-up study after atrial ablation|
|ERP||=||effective refractory period|
|HE||=||hematoxylin and eosin|
|IHR||=||intrinsic heart rate|
|IVC||=||inferior vena cava|
|MHR||=||maximal heart rate|
|RFCA||=||radiofrequency catheter ablation|
|SNRT||=||sinus node recovery time|
|SVC||=||superior vena cava|
|WGA||=||wheat germ agglutinin|
This study was supported in part by the Herman C. Krannert Fund and by grant HL-52323 from the National Heart, Lung, and Blood Institute of the National Institutes of Health. Dr Elvan was a recipient of a research fellowship grant (R95006) from the Dutch Heart Foundation, Den Haag, Netherlands. The authors thank Dr Paul C. Dolber for the advice on WGA staining, Dr Edwin Duffin of Medtronic, Inc, for help in obtaining the pacemakers, and Dzung Nguyen for technical assistance.
- Received October 2, 1996.
- Revision received February 28, 1997.
- Accepted March 5, 1997.
- Copyright © 1997 by American Heart Association
Crijns H, Van Gelder I, Lie K. Benefits and risks of antiarrhythmic drug therapy after DC electrical cardioversion of atrial fibrillation or flutter. Eur Heart J. 1994;15(suppl A):17-21.
Ferguson TJ, Cox J. Surgery for atrial fibrillation. In: Zipes D, Jalife J, eds. Cardiac Electrophysiology: From Cell to Bedside. Philadelphia, Pa: WB Saunders Co; 1995:1563-1576.
Van Hemel N, Defauw J, Kingma H. Long-term results of the corridor operation for atrial fibrillation. Br Heart J. 1994;71:170-176.
Elvan A, Pride H, Eble J, Zipes D. Radiofrequency catheter ablation of the atria reduces the inducibility and duration of atrial fibrillation in dogs. Circulation. 1995;91:2235-2244.
Coumel P, Attuel P, Lavallee J, Flammeng D, Leclercq J, Slama R. Syndrome d’arythmie auriculaire d’origine vagale. Arch Mal Coeur. 1978;71:645-656.
Coumel P. Paroxysmal atrial fibrillation: a disorder of autonomic tone? Eur Heart J. 1994;15(suppl A):9-16.
Swartz JF, Pellersels G, Silvers J, Patten L, Cervantes D. Catheter-based curative approach to atrial fibrillation in humans. Circulation. 1994;90(suppl I):I-335. Abstract.
Wijffels M, Kirchhof C, Dorland R, Allessie M. Atrial fibrillation begets atrial fibrillation: a study in awake chronically instrumented goats. Circulation. 1995;92:1954-1968.
Pressler M, Munster P, Huang X-D. Gap junction distribution in the heart: functional relevance. In: Jalife J, Zipes D, eds. Cardiac Electrophysiology: From Cell to Bedside. Philadelphia, Pa: WB Saunders; 1995:144-151.
ten Velde I, de Jonge B, Verheick E, van Kempen M, Analbers L, Gros D, Jongsma H. Spatial distribution of connexin 43, the major cardiac gap junction protein, visualizes the cellular network for impulse propagation from sinoatrial node to atrium. Circ Res. 1995;76:802-811.
Luke R, Saffitz J. Remodeling of ventricular conduction pathways in healed canine infarct border zones. J Clin Invest. 1991;87:1594-1602.
Davis L, Kanter H, Beyer E, Saffitz J. Distinct gap junction protein phenotypes in cardiac tissues with disparate conduction properties. J Am Coll Cardiol. 1994;24:1123-1132.
Elvan A, Wylie K, Zipes DP. Pacing-induced chronic atrial fibrillation impairs sinus node function in dogs. Circulation. 1996;94:2953-2960.
Morillo C, Klein G, Jones D, Guiraudon C. Chronic rapid atrial pacing: structural, functional, and electrophysiological characteristics of a new model of sustained atrial fibrillation. Circulation. 1995;91:1588-1595.
Dahl G, Werner R. Differential closure of mRNA-induced as compared to endogenous cell-to-cell channels in pairs of Xenopus oocytes. Biophys J. 1986;49:201a. Abstract.
Dahl G, Werner R, Levine E. Paired oocytes: an expression system for cell-to-cell channels. In: Herzberg E, Johnson R, eds. Gap Junctions. New York, NY: Alan R Liss; 1988:183-197.
Sih HJ, Chiou CW, Berbari EJ, Zipes DP. Maps of electrically induced atrial fibrillation after linear ablation lesions. Circulation. 1996;94(suppl I):I-555. Abstract.