Treatment of atrial fibrillation by silencing electrical activity in the posterior inter-pulmonary-vein atrium

Abstract

Introduction

Several investigators ^1–3 have shown that arrhythmogenic activity in the pulmonary veins (PVs) triggers and/or perpetuates episodes of atrial fibrillation (AF). Many such arrhythmogenic ectopic beats are located within or near the PV ostia. Electrical isolation performed under the guidance of a circular catheter at the PV ostia ^3–5 and circumferential encircling ⁶ by radiofrequency (RF) ablation guided by electroanatomic mapping are widely used to treat AF. ^7–9 However, the recurrence rate of AF is relatively high, particularly in persistent AF, ⁷ patients with long history of AF, ⁵ or with structural heart disease. ¹⁰ Focus firing outside PVs has been reported to be related to the recurrence of AF after PV isolation. ^{8 , 9} A macro-re-entrant rotor important for AF perpetuation was observed in the posterior wall of the left atrium (LA) in animal studies. ^{11 , 12} New ablation designs that expand the lesion from the PV ostia to the LA, especially to the posterior wall, have greatly improved clinical results. ^13–15 Isolation of the PV and posterior LA performed during open-heart surgery lowered clinical AF recurrence. These observations suggest that the posterior left atrial wall plays an important role in perpetuating AF. ^{16 , 17} Although a few cases of fatal atrio-esophageal fistulas have been reported, ^{18 , 19} RF ablation in the posterior wall of the LA was quite safe with limited energy delivery. ^{14 , 20} We sought to develop a new approach to silencing electrical activity in the posterior inter-pulmonary-vein atrium (PIA) and to evaluate the roles of the PIA in the perpetuation of different types of AF.

Methods

Patients

A total of 42 patients with symptomatic AF were involved in this study. Eighteen (42.9%) patients presented with paroxysmal, 14 (33.3%) with persistent (duration > 7 days) and 10 (23.8%) with permanent AF (duration > 6 months and sinus rhythm could not be maintained after electrical cardioversion). The clinical characteristics of patients are shown in Table  1 . All patients had been unsuccessfully treated with ≥2 antiarrhythmic drugs. Antiarrhythmic drugs were not discontinued before the ablation procedure. Class IC or III antiarrhythmic drugs and β-adrenergic blockers were administered to most patients. The study protocol was approved by the locally appointed ethics committee and informed consent was obtained from all patients before the ablation procedure.

Posterior inter-pulmonary-vein atrium electrical silence

Three RF ablation steps were required. (i) Electrical PV isolation guided by a steerable circular catheter, (ii) creation of two lines of lesions between the two superior (superior line) and inferior PVs (inferior line), and (iii) abolition of residual electrical signals between ipsilateral PVs or in the other areas within PIA if necessary. An electroanatomical mapping system (CARTO Merge ^® , Biosense Webster, Figure  1 ) was employed to achieve PIA silence.

Figure 1

Achievement of PIA silence integrated with CT images (posterior view of the LA). ( A ) CT image of the LA and the PVs. ( B ) CT image integrated with voltage map of the LA before ablation. ( C ) Voltage map of the LA after PIA silence. No electrical activity is recorded within the PIA [voltages < 0.1 mV (red, sites 1–4), which is regarded as scar in the voltage map]. Atrial electrograms are recorded outside the PIA [voltage > 0.2 mV (purple, sites 5 and 6)]. Recordings were made during coronary sinus pacing. The blue dots denote the ablation lesions around the PV ostia and the superior and inferior lines. LA, left atrium; PIA, posterior inter-pulmonary-vein atrium; PV, pulmonary vein.

Open in new tabDownload slide

All patients were treated under conscious sedation. Patients were heparinized after the catheters were placed in the LA and activated clotting time was maintained between 250 and 300 s. Pulmonary vein mapping was performed at the left atrial-PV junction with a Lasso ^® (Biosense Webster, Diamond Bar, CA, USA) steerable circular catheter. A 20-pole mapping catheter (Livewire™, Daig Co., St Jude Medical, Minnetonka, MN, USA) was positioned in the right atrium, with its distal poles in the coronary sinus and proximal poles around the tricuspid annulus. A cooling deflectable 7F quadripolar catheter was used for recording and ablation (Navistar, Biosense Webster). Cooling flow was 8–10 ml/min for PV isolation. Radiofrequency generator energy was applied in temperature-controlled mode with a cutoff at 50°C and a power limit of 30–35 W for 60–90 s at each site. Pulmonary vein isolation was achieved as described in previous reports. ^{4 , 5} Firing foci outside the PVs were also ablated if observed.

Selective PV angiography was performed by manual injection of contrast medium before and after ablation. Pulmonary vein narrowing or stenosis was reported if there was any decrease in diameter <50 or >50%. Cardiac computerized tomography was performed on all patients before the procedure and only on patients who presented symptoms indicating PV stenosis after the procedure.

Radiofrequency energy was delivered point by point to create a line of lesion between the superior or inferior PVs. Maximal temperature limit was 50°C with cooling flow 15–20 ml/min. We began energy delivery with 30 W at each target site while monitoring the on-line potentials in order to eliminate or greatly decrease the voltage of local potentials (>80%) or to achieve formation of local split potentials. Energy delivery was terminated (40–90 s) if the target effect was achieved. Energy might be increased up to 35 W if the target effect was not obtained and the location was judged empirically not to be in close proximity of the esophagus. ^{10 , 18} The power was decreased to 25 W if patients perceived pain. Ablation was performed during sinus rhythm in all patients (except for two permanent AF patients in whom electrical cardioversion failed to restore sinus rhythm). The endpoint of the procedure was PIA electrical silence. This was verified by the absence of electrical activity in the region bracketed by the PVs and the two lines of lesions (voltage < 0.1 mV, which was regarded as a scar during voltage mapping) and pacing (pulse width 2 ms, output 20 mA, at least five sites within the PIA area) without capture of the remaining atria. If any electrogram > 0.1 mV was detected within this region, ablation would be performed again at the corresponding site. The volume of the LA, the perimeter and area of the electrically silent region, and the percentage of the electrically silent area of the total LA surface were calculated by the CARTO system.

Cavotricuspid isthmus ablation was performed only in patients who had clinical documentation of atrial flutter (AFL) or occurrence of AFL during the procedure. Bi-directional isthmus block was confirmed after ablation.

Follow-up

Patients were monitored for 24–48 h after the procedure, and low-molecular heparin was given subcutaneously. Patients were discharged under oral anticoagulation. Anticoagulants were interrupted 3 months after successful elimination of AF, unless there were other risk factors. Antiarrhythmic drug therapy was continued in most patients for at least 3 months. The patients were followed-up on an ambulatory basis with clinical evaluations and electrocardiograms and 24 h Holter recordings, 3, 6, and 12 months after the procedure. Clinical recurrence of AF was defined as more than one documented episode of AF later than 1 month after the procedure (blanking period).

Statistical analysis

Continuous variables are expressed as mean ± SD and compared by unpaired Student's t -tests. Fisher's test was used to compare proportions between groups. Values of P < 0.05 were considered to be statistically significant.

Results

Electrophysiological study and ablation

The number of RF applications, volume, size, and total surface area of the LA, perimeter and area of the electrically isolated region (within PVs and lines of lesions), and mean voltage prior to ablation are presented in Table  2 . Direct-current cardioversion was performed in 29 patients. Voltage mapping was performed in 40 of 42 patients during coronary sinus pacing before ablation but not in the other two because of failure in restoring sinus rhythm by electrical cardioversion. Voltage mapping was obtained in all patients after ablation. The PV ostia and PIA region were intensively mapped. Mean numbers of sampling sites were 6, 21, 16, and 43 before ablation and 4, 20, 21, and 30 after ablation in the LA appendage, PV ostia, PIA, and the other parts of the LA, respectively. A left common PV trunk and a right middle branch PV were identified in four and two patients, respectively. Additional ablation lesions were needed to complete PIA silence between the left PVs in 21 patients and between the right PVs in 34 patients. It was necessary to ablate in the middle of the PIA region to eliminate residual electrical activities in 13 patients. Cavotricuspid isthmus block was obtained in the present or earlier procedures in 6 and 11 patients, respectively.

Procedure duration (skin to skin, including the time for direct-current cardioversion) and fluoroscopy time were 261 ± 34 and 47 ± 11 min, respectively.

Clinical results

During a follow-up period of 20 ± 4 (range 14–27) months, 14 (33.3%) of 42 patients experienced more than one episode of atrial arrhythmias (eight AF, four AFL, and two with both AF and AFL). Atrial fibrillation recurrences in different types of AF are shown in Table  3 . The recurrence rate was significantly lower in paroxysmal than that in permanent AF. In patients with paroxysmal or persistent AF who had already undergone PV isolation, recurrence rates after the PIA silencing were 10.0% (2/20). Six patients with AF recurrence (one paroxysmal, three persistent, and two permanent, including two with both AF and AFL) underwent a repeat electrophysiological study. Recovery of electrical activity was observed in the PIA and PV ostia (>2 PVs) in all patients. Electrical silence was achieved again (see also post-ablation AFL). After 5 ± 2 months, four patients were arrhythmia-free, one still has AFL episodes and one patient experienced a switch from permanent to paroxysmal AF. After the second PIA silence procedure, the success rates were 94.4, 85.7, and 60.0% in paroxysmal, persistent, and permanent AF, respectively. Antiarrhythmic drugs had been discontinued in 24 patients.

Pulmonary vein narrowing was observed in one patient. Pericardial effusion was found in one patient at the end of the procedure. No drainage was needed. Neither PV stenosis nor atrio-esophageal fistula was observed.

Comparisons between atrial fibrillation recurrence and recurrence-free patients in the posterior inter-pulmonary-vein atrium silence group

The LA volume and size in the patients with AF recurrence were larger than in the recurrence-free patients and the percentage of the silent area of the total LA surface was smaller in the patients with AF recurrence than in the recurrence-free group ( Figure  2 ). Voltages in the LA measured before ablation in the patients with AF recurrence were significantly lower than in the recurrence-free patients, independent of location ( Table  2 , Figure  2 ). No significant differences were demonstrated between recurrence and recurrence-free patients with respect to ablation applications, the total surface area of LA, the perimeter or the area of electrically silent region created by ablation ( Table  2 ).

Figure 2

Voltage maps (posterior view of the LA) from two patients with recurrence before (left) and after ablation (right). Upper panel: low voltages of atrial electrograms (mean 0.25 mV) were observed in the posterior LA. Lower panel: low percentage (27%) of the electrically silent region of the LA surface area was demonstrated because of the close neighbouring superior and inferior PVs. This patient had undergone an earlier PV isolation procedure and low voltages around the PVs are presented. Red represents regions with low voltage. Blue dots denote the ablation lesions. Abbreviations as in Figure  1 .

Open in new tabDownload slide

Post-ablation atrial flutter

Six patients experienced episodes of recurrent AFL. During the repeat procedure, typical AFL around the tricuspid annulus was demonstrated and ablated in two patients. Macro-re-entrant AFL located in the LA was identified in four patients, and in these cases re-entrant circuits around the right PVs, the mitral annulus and both were presented in one, two, and one patient, respectively. The lateral wall (between the mitral valve and the left inferior PV), the myocardium between the right PVs and close to the lower part of the right inferior PV were recognized as the isthmus, in the respective patients ( Figure  3 ). Four of these circuits were successfully ablated, except for one around the mitral annulus. Additional energy had to be delivered in the coronary sinus in three patients, with limits of 50°C and 25 W, in order to obtain conduction block in the mitral isthmus region. The PIA region was electrically silenced again at the end of the procedures.

Figure 3

Local activation time maps in left anterior oblique (left) and postero-anterior (right) views from two patients with post-ablation atrial flutter. Upper panel: a figure-of-eight macro-re-entrant atrial flutter located around the right PVs and ablation scar of the earlier inferior line. The right atrium is bystanding. Lower panel: a macro-re-entrant circuit around the mitral annulus. Two directions of wavefronts collide in the PIA. Note that in both cases recovery of conduction through the lines of lesions is demonstrated. Gray represents scar regions. Red and purple represent the earliest and latest activation in the re-entrant circuit, respectively. White arrows denote directions of wavefronts. Abbreviations as in Figure  1 .

Open in new tabDownload slide

Discussion

Roles of the posterior left atrium in atrial fibrillation

Since the recognition of focal sources of AF from PVs, ^1–3 PV isolation has frequently been employed to treat AF. ^3–5 There is considerable evidence to the effect that non-PV foci are common, including locations in the left and right atrium, the caval veins and the coronary sinus/ligament of Marshall. ^{8 , 9} Experimental animal studies have demonstrated that the highest dominant frequency during AF is usually (80%) located in the posterior LA, near or at the ostia of PVs. High-resolution video imaging has also shown that such sources corresponded to a fast vortex-like re-entry around minuscule cores. ^{11 , 12} A clinical study using complex fractionated atrial electrograms (CFAEs) as a guide for AF ablation showed that CFAEs were distributed more frequently at the interatrial septum, PVs and LA roof. Ablation in these areas produced satisfactory clinical results. ²¹ Furthermore, electrical isolation of the pulmonary venous region (posterior LA), combined with mitral isthmus ablation and excision of the LA appendage during open-heart surgery, demonstrated 92.9% (13/14) freedom from AF in lone AF patients. ¹⁶ Both experimental and clinical studies indicate that PVs and the PIA play an essential role in triggering and perpetuating AF.

We have taken a new approach to treating AF patients: so-called PIA electrical silence, based on PV isolation and on joining the two superior and inferior PVs by ablation lines. Since the area of interest was not always fully located on the anatomically posterior LA (partially to the LA roof), the nomenclature of ‘PIA’ has been adopted and defined in this study. Comparing with other techniques the electrical activities within the PIA were fully silenced in addition to PV isolation by our approach. Total freedom from atrial arrhythmias was 94.4, 85.7, and 60% after repeat procedures in paroxysmal, persistent and permanent AF, respectively, after at least 12-month follow-up. The factors related to recurrence of AF included enlargement of the LA, a small proportion of electrically silent area and low voltages in the LA. These findings indicate that the substrates for paroxysmal and persistent AF are located within the PV and PIA. In permanent AF patients, the substrates are probably located not only in these regions but also in the rest of the atria, ^{22 , 23} as was also demonstrated by our findings of low voltage in the LA. The pathological basis for this might be myocardial fibrosis after long duration of AF or organic cardiac disease.

Sanders et al . ¹⁰ have recently investigated the effect of isolation of the PV and posterior LA in chronic AF. After 21 months, 63% patients were in sinus rhythm. Although the strategy was similar, the ablation method was not the same. In that study, RF ablation was performed during AF in all patients, but in our study ablation was performed during sinus rhythm. During ongoing AF, atrial electrograms in the area of interest could be too small to interpret, and near-field electrograms from contiguous structures difficult to differentiate, which may have influenced the ablation effects. In another study, ²⁴ isolation of the posterior LA wall was obtained by continuous lesions at the anterior portions of the PVs and ablation lines of the LA roof and bottom (box isolation) without any ablation lesions at the PIA. A success rate of 95% was reported in paroxysmal AF. This technique decreased the potential complication of esophageal injury. However, the risk of conduction recovery of the PV and PIA was higher than with the compartmental ablation technique employed in our study.

Potential mechanisms related to the treatment of atrial fibrillation by posterior inter-pulmonary-vein atrium silence

Our results indicated that PIA electrical silence was effective, especially in the treatment of paroxysmal and persistent AF. Elimination of focal activities from the PVs and PIA was the main mechanism involved in curing these patients, as has been demonstrated by other investigators. ^{3–6 , 13–15} Complex fractionated atrial electrogram during AF were frequently observed close to the PV areas and the LA roof. ²¹ A recent study using endocardial video imaging and electric mapping has demonstrated that the posterior LA harboured regular, fast and highly organized activity, and that the outer limit of the maximum dominant frequency domain was where the greatest variability in propagation pattern and fractionated activity occurred. ²⁵ Most of these areas can be covered by the PIA silencing procedure. Armour et al . ²⁶ have demonstrated that numerous autonomic ganglionated plexi (GP) are located in the atrial epicardium. Of these, the superior left, posterolateral left, posterior right, and posteromedial left atrial GP are located in the posterior LA and close to the PVs. The superior right atrial GP are located in the LA roof region in front of the right superior PV. A necropsy investigation ²⁷ showed that nerve density was significantly higher in the ostia of the PVs than in their distal part and rising innervation gradients were found from the right to left and from the front to the rear of the atrium. By achieving PIA silence, most of the nerves and partial GP in the posterior LA could probably be covered and partly denervated. It has been observed that ablation of the GP and denervation during ablation procedure can improve the clinical results. ^{13 , 28}

Since the PVs and PIA anatomically flank each other and both are essential for the maintenance of AF, we might ask what kind of histo-embryological relationship these two have. An autopsy study ²⁹ has demonstrated that there was no microscopical boundary between the PVs and the LA endocardium. Each PV was surrounded by a myocardial sleeve extending from the LA. A recent study in embryos and fetuses ³⁰ has shown that the PV wall surrounded by extrapericardially differentiated myocardial cells was incorporated into the LA body. After incorporation, the composition of PVs and the smooth-walled LA body wall were histologically identical. This may explain why the PIA plays the same important role as the PVs in triggering and perpetuating AF.

Post-ablation atrial flutter

As with the other strategies of linear ablation for AF, macro-re-entrant AFL was a major sequela due to unconnected lines of lesions after myocardial recovery. Several studies ^{10 , 31} have demonstrated that re-entrant circuits around the left PVs and the mitral annulus were frequent. Our study showed that the region around the right inferior PV was often the isthmus for the re-entry, and a re-entrant circuit around the mitral valve was also observed. These parts of the myocardium are capable of recovering easily after ablation and creating a slow conduction area.

Limitations

The selection of patients in this study was based on the operator's decision, but the demographic data of the patients were comparable to those of other studies. The number of patients in the permanent AF subgroup was relatively low. No specific monitoring of the esophagus was undertaken during the procedure. Since our study commenced, a few cases of atrio-esophageal fistulas occurring as a result of ablation along the posterior LA have been reported. ^{18 , 19} No particular change was made to the ablation strategy but the power was usually limited to 30 W. Energy delivery was increased to 35 W only if this was necessary for the difficult sites. According to an expert consensus statement on catheter ablation of AF ³² and several studies ^{14 , 20} it seems safe to limit RF power to 35 W in order to minimize collateral injuries. Furthermore, not all patients underwent AF ablation with this new approach as a first procedure. This may limit the interpretation of the results. However, PV isolation is, in any case, the first step of the PIA silencing procedure. Our results show at least that this can be an alternative approach for recurrent cases after PV isolation.

Clinical implications

Our data demonstrate that the PVs and PIA play important roles in the maintenance of AF. Silencing electrical activity in the PIA is feasible, and capable of decreasing the rate of AF recurrence. Clinical recurrence after PIA silence may be linked to enlargement of the LA, low amplitude of electrogram voltage, and a small proportion of the LA being electrically silent.

Funding

This study was partly supported by the Norwegian Council of Cardiovascular Diseases and the Norwegian Foundation for Health and Rehabilitation, and part of its content was presented at Heart Rhythm 2006, Boston, USA.

Conflict of interest : none declared.

References