Catheter Ablation of Atrial FibrillationIntroduction Atrial ﬁbrillation (AF) is the most common cardiac arrhythmia seen in clinical practice and results in signiﬁcant morbidity,

Catheter Ablation of Atrial Fibrillation

Chi Keong Ching MD, Dimpi Patel MD, Andrea Natale MD

Cleveland Clinic, Cleveland, OH, USA

Atrial fibrillation is the most common arrhythmia encountered in clinical practice andcan result in significant morbidity and mortality. Catheter ablation has become a feasibletherapeutic option for the management of this complex and challenging arrhythmia. In thisarticle, we have discussed the mechanism of atrial fibrillation, different imaging modalitiesused for atrial fibrillation ablation, different ablation strategies targeting non-pulmonaryveins, complications associated with atrial fibrillation ablation and its management,alternative energy sources for ablation and new antiarrhythmic drugs.(J Arrhythmia 2007; 23: 85–101)

Key words: Atrial fibrillation, Catheter ablation, Treatment, Novel therapies

Introduction

Atrial fibrillation (AF) is the most commoncardiac arrhythmia seen in clinical practice andresults in significant morbidity, mortality and health-care costs. Catheter ablation has emerged as animportant treatment strategy for the eradication ofAF. Centers that are extensively experienced inperforming AF ablation report long-term successrates ranging from 75–90%. This paper will reviewthe mechanism of initiation and perpetuation of AF,the role of imaging in catheter ablation, variouscatheter ablation strategies, complications of AFablation, alternative energy sources used for ablationand new antiarrhythmic drugs (AADs).

Mechanism of Atrial Fibrillation

AF is characterized by the presence of multipleactivation waves within the atria. The mechanisms

by which these activation waves occur have beenhotly debated for many years. The focal sourcehypothesis states at a single rapidly firing focus actsas a driver, causing the rest of the atria to go into AF.Studies in the 1950s demonstrated that local injec-tion of aconitine into the atria could result in arapidly firing focus that drives the atria into AF.1) Asustained reentry2) can similar act as a driver andgenerates AF. On the other hand, multiple wavelethypothesis states that heterogeneity in atrial repola-rization is responsible for the generation of multiplewavelets that sustain AF. If true, the multiplewavelets hypothesis suggests that AF is self-sustain-ing and does not need a driver. Central to thesehypotheses are thoracic veins, which have beenshown to be highly arrhythmogenic. In 1998,Haissaguerre et al.3) reported that repetitive rapiddischarges from the pulmonary veins (PVs) wereresponsible for triggering AF in a series of 45patients with symptomatic paroxysmal AF. Radio-frequency ablation of these PVs foci successfully

Address for correspondance: Andrea Natale MD, Head, Section of Cardiac Electrophysiology and Pacing Director, Center for Atrial

Fibrillation, Department of Cardiovascular Medicine, Cleveland Clinic, 9500 Euclid Avenue F15, Cleveland, OH 44195, USA. Tel: 216

444-4293 Fax: 216 445-2753 E-mail: [email protected]

85

Ching CK Catheter ablation of atrial fibrillation

Review Article

eliminated AF in up to 62% with a median follow-upperiod of 7 months.

Humans have a significant amount of left atrialmyocardial tissue cardiac muscle which extends intoPVs.4) The myocardial architecture in normal humanPVs is also highly variable.5) In addition, aging mayresult in fibrosis at the PV antral (PVA) junctions,which further produces greater anisotropy in theseveins. Several investigators6,7) have demonstratedthat in animal models, PVs are capable of generat-ing independent electrical activity. Perez-Lugoneset al.8) examined PV tissue from five autopsies out ofwhich four had a history of AF and demonstratedthat under electron microscopy, four out of five hadmorphological features consistent with that ofPurkinje cells in the PVs. It is conceivable then thattriggered activity and/or automaticity may have arole in initiating and sustaining AF.

In addition, we have come to realize over theyears that other thoracic veins such as the superiorvena cava (SVC), the ligament and/or vein ofMarshall, the coronary sinus musculature are alsohighly arrhythmogenic and may play a role in theinitiation and maintenance of AF.

Imaging in AF ablationMost AF ablation requires LA instrumentation via

transseptal puncture and up to 3 punctures is madedepending on the strategy adopted.9) In most cases,fluoroscopy provides sufficient information for safetransseptal punctures. However, fluoroscopy may belimited in the presence of interatrial septal varia-tions, atrial or aortic root dilatation, prior LAinstrumentation and the need to introduce up to 3sheaths into the LA. Transesophageal echocardiog-raphy (TEE) allows identification of the fossa ovalisand its surrounding structures, provides real timeassessment of the procedure, and visualizes tentingof the fossa ovalis, the subsequent entry andadvancement of the sheath into the LA. However,the TEE probes obstructs fluoroscopic field and isutilized in only in-anesthetized patient, therebypreventing periodic neurological assessment of thepatient. Intracardiac echocardiography (ICE) pro-vides similar information as TEE but it can be usedin a conscious patient and does not obstruct fluoro-scopy. In addition, phased array transducers allowvisualization of LA structures and a slightly poste-riorly directed transseptal puncture can be safelyperformed to facilitate placement of mapping andablation catheters at various PV antrum regions.

The goal of present day ablation strategies is toelectrically isolate the PVs from the rest of the LA.Importantly, ablation performed at the PV antrum-

LA junction decreases the incidence of PV stenosis.However, PV anatomy in humans is highly varia-ble.10) Furthermore, variation in orifice size, orienta-tion and diameter occurs during the cardiac cycleand is further accentuated by respiration.11) Giventhis marked variability, information on the numberand anatomy of the PV is essential in performingsuccessful and safe ablation. This can be performedpre-procedurally using computerized tomography(CT) or magnetic resonance imaging (MRI). Fur-thermore, these CT or MRI images can be fullyintegrated with Carto (Biosense Webster, DiamondBar, CA, USA) or side by side with NavX (EnSite,St. Jude Medical, St Paul, MN, USA) 3D mappingsystem. The clinical usefulness of these 3D mappingsystems during ablation will be dependent on theaccuracy of image integration.

Venography, TEE and ICE can be performed toprovide real time imaging of the PVs. Wood et al.compared imaging of PV ostial anatomy with CT,TEE, ICE and venography in 24 patients undergoingAF ablation.12) CT and ICE were superior to TEE inidentifying the number of PV ostia. Ostial diameterwas overestimated by venography and underestimat-ed by TEE. Both CT and ICE yielded similar ostialdiameters. This discrepancy could be due to theelliptical shape of the ostia whereby measurement inone plane will either under or over estimate its truediameter. Recently, multislice CT scanning is supe-rior to ICE in demonstrating the oval shape of theostia.13)

ICE is able provide real time assessment of thePVs and the PVA junction. Titration of energydelivery during RF ablation by the development ofmicrobubbles can be performed with this imagingutility, thereby providing real time information onlesion formation only in non-open irrigated catheterablation systems. Both ICE and TEE are also usefulfor monitoring complications within the LA. Itsability to provide real time imaging identifies LAappendage, clot formation and the detection ofpericardial tamponade. Recently, ICE allows earlydetection of small thrombi formation on the trans-septal sheath14) that allows appropriate sheath man-agement to avoid inadvertent dislodgement and apotential catastrophic thromboembolism.

Catheter Ablation of Atrial Fibrillation

Catheter ablation for the treatment of AF hasevolved over the years. The concept of PV isolation(PVI), was initially limited to the PVs, and has beenmodified by The Cleveland Clinic to involve theLA myocardium surrounding the PV to achieve

J Arrhythmia Vol 23 No 2 2007

86

complete electrical isolation of the PV antra as aprocedural endpoint, namely the LA-PV antrumisolation (PVAI).15) In addition to circular mappingcatheter guided PVAI, we electrically isolate furtherAF trigger sites, modify the substrate for AFmaintenance, and are possibly modulating cardiacautonomic tone at various LA, right atrial (RA) sitesand within the coronary sinus to achieve a higher AFfree success rate.

Under light sedation to facilitate periodic neuro-logical assessment, ICE (10 Fr phased array, Sie-mens AG Inc., Malvern, PA, USA) guided doubletransseptal punctures are performed using Mullinsand SRO sheaths at the midposterior septum,through which a 10 pole, 20mm diameter circularmapping catheter and an open irrigated ablationcatheter are introduced respectively. In patients withsevere LA enlargement, a J-curve ablation catheteris preferred to the usual F-curve ablation catheter.Heparin boluses are given prior to the first andsecond transseptal puncture and subsequent heparininfusion is started to achieve an activated clottingtime > 350 s.

Pulmonary vein antrum isolationAn ablation strategy that only isolates the PV with

spontaneous or provoked AF triggers has beenassociated with higher recurrence rates.16) Thesefindings led to the current strategy of empiricalPVAI for all 4 PVs without provocative measures,in an attempt to reduce AF recurrence (Figure 1).Isolation of the PVs is usually performed in thefollowing order: left superior, left inferior, rightsuperior and right inferior. The circular mappingcatheter is placed at the PV antrum-LA junctiondefined by ICE. RF ablation is performed at sites

where PV potentials are recorded in the mappingcatheter. With an open irrigated catheter system,a temperature of 45�C and a maximum power of50W is used. As the PV antrum is a large structure,multiple movements of the circular mapping catheterare required to identify potentials at various regionsof the respective PV antra. An esophageal temper-ature probe is inserted prior to ablation and to titrateenergy delivery during ablation of the posterior wallof the LA. Energy delivery is halted temporary whenthe esophageal temperature exceeds 38.5�C. Usingthis approach, we are fortunate not to have any casesof atrio-esophageal fistula thus far. To reducefluoroscopy, ICE imaging, ablation artifact on thecircular mapping catheter and 3D mapping systems(CARTO, NavX) are used as surrogate markers ofcatheter position. The end point for PVAI is toeliminate potentials at the respective PV antra. Entryblock is verified when no PV potentials can berecorded at the antrum or within the PV by thecircular mapping catheter. Occasionally, electricaldissociation of the PV from the LA confirms exitblock. At the end of the procedure, all four PV antraare extensively remapped to identify remnant orrecovered potentials for ablation to achieve completeelectrical isolation.

Non pulmonary vein ablationThe incidence of non-PV triggers giving rise to

AF has been reported to be approximately 20–32%17)

(Figure 3). The SVC is a common site for non-PVtriggers in patients with paroxysmal AF. In humans,cardiac excitation has been registered up to 5 cmabove the anatomical junction of the SVC and RA,defined at the base of the RA appendage.18) Itcontains atrial myocytes capable of exhibiting

A B

Figure 1 Anatomic definition of pulmonary vein ostium and antrum using 3D CT imaging (A) and ICE (B).Note that the right and left pulmonary vein antra include the posterior wall of the left atrium. LSPV: left superior

pulmonary vein; RSPV: right superior pulmonary vein.


87

enhance automaticity and afterdepolarization.19) In aseries of 240 patients who had a total of 359 ectopicfoci initiating AF, 20% originated from non-PVsites.20) Of which the SVC accounted for 37% ofthese non-PV sites either exclusively or otherwise.RF ablation of the ectopic SVC sites eliminatedparoxysmal AF and up to 86.7% of patients withexclusive SVC foci were free of AF without AADs.In our series of 407 patients who underwent catheterablation for AF, SVC triggers were identified in 12%of the initial 190 patients in the setting of intra-venous isoprotenerol up to 20 ug/min.21) In view ofthis finding, empirical SVC isolation was performedin the rest of the 217 patients. 16% of these patientshad recurrence of AF during long-term follow-up.Of these, 25 patients underwent a repeat procedureand 5 patients were identified to have SVC triggers.Of note 4 out of these 5 patients did not have aprior SVC isolation. At our centre, empirical SVCisolation is performed after PVAI. The cathetersand sheaths are pulled into the RA. The SVC ismapped with the circular mapping catheter at thelower border of the pulmonary as identified on ICE.Isolation of the SVC is then performed at this level.High voltage pacing at the lateral portion of the SVCis performed to check for phrenic nerve stimulationas well. If phrenic nerve stimulation is present, thesesites are precluded from ablation. Otherwise com-plete isolation of the SVC is performed.

The crista terminalis (CT), a structure in the rightatrium, has been identified as a source of non-PVtriggers. These triggers appear to be catecholaminesensitive and have been implicated in the initiationof AF following cardioversion in patients withchronic AF, which was observed in 37.5% of ourseries of 48 patients,22) with majority of foci arisingfrom the mid to superior aspect of the CT. In a smallseries of 5 patients who underwent PVI and SVCisolation, non-contact mapping was used to identifyseveral areas around the crista terminalis, whichhave spontaneous activity that initiate AF. Subse-quent ablation of these sites abolished the CT free ofectopy-initiating AF.23)

The CS musculature is also an origin of focalatrial tachycardias. Repetitive rapid discharges fromthe CS musculature may be a potential driver for AFand these triggers have been documented in 2patients who underwent prior PVI and substratemodification with linear lesions for chronic AF.24)

Electrical disconnection from the adjoining atrialmyocardium permanently eliminated drivers main-taining AF. Furthermore, electrical isolation of theCS from the LA has been shown to decrease AFinducibility in patients with paroxysmal AF.25)

The ligament of Marshall (LOM) is a vestigialstructure of the left primitive veins. It runs from theCS superiorly on the posterior wall of the LA to theregion near the orifice of the left superior PV. Insome cases, this LOM connects to a small auricularvein that drains into the CS. This vein is called thevein of Marshall (VOM). The muscle extensions ofthe LOM is complex, with multiple myocardial tractinsertions into the LA free wall and coronary sinus(CS).26) These myocardial tracts exhibit electricalactivity and may display enhanced automaticityduring isoprotenerol infusion.27) Furthermore, theseectopic potentials have been recorded and ablated inthe vicinity of the left superior PV in patients withnon-PV triggers. The VOM has also been cannulatedto guide mapping and ablation of LOM ectopicactivity initiating AF.28)

The persistent left SVC (LSVC) may generaterepetitive rapid discharges that initiate AF (Figure 2).Hsu et al. described the role of persistent LSVC asthe arrhythmogenic source of AF in 5 patients afterPVI.29) The LSVC was electrically connected to thelateral LA and via the CS to the RA. The CS-LSVCand LA-LSVC connections were mapped with the

Figure 2 CT angiography with 3D reconstruction of theheart.LAO view with caudal angulation showing the relation of the

persistent LSVC to the left atrium, appendage, and pulmonary

veins. Also note the hugely dilated coronary sinus in which the

LSVC as well as the other tributaries drain. TA = Tricuspid

annulus, MA = Mitral annulus, RAA/LAA = right/left atrial

appendage.


88

circular mapping catheter and ablation was per-formed to achieve electrical isolation of the LSVC asconfirmed the LSVC and LA in 4 out of 5 patients.These 4 patients remained in sinus rhythm as a meanfollow up period of 15� 10 months without the needof any antiarrhythmic drugs. We described a seriesof 6 patients with LSVC, 4 of which developedrecurrence of AF post PVAI.30) The PVs of these 4patients were found to be electrically isolated duringa subsequent repeat procedure. Spontaneous atrialectopy was observed in 3 out of 4 patients, of whichAF was initiated in one patient. Circular mappingcatheter guided RF ablation electrically isolated theLSVC-LA-CS connections. All patients remained insinus rhythm without any antiarrhythmic drugs atmean follow-up period of 13� 7 months.

The LA appendage as a source of ectopy for themaintenance of AF has been described in a singlecase report.31) Circumferential ablation of the ostiumof the LA appendage eliminated AF resulting incomplete electrical isolation as evidenced by dis-sociated trigger potentials originating from the LAappendage was achieved. At 5 months follow-upthis patient remained in sinus rhythm without anythromboembolic events while on anticoagulation.

The inferior vena cava (IVC) has rarely beendocumented to have AF triggers. A survey of theliterature identified 3 cases. In two cases, ectopicactivity was mapped to the posterolateral aspect ofthe IVC ostium and these were successfully ablat-ed.32) In a separate report, non-PV triggers after PVIwas mapped to a site deep in the IVC. Focal ablation

guided by a circular mapping catheter eliminated thetrigger and subsequent IVC isolation was performedat the IVC-RA junction.33)

The sinus node region as a non-PV trigger for AFfollowing surgical Maze procedure in patients withchronic AF has been reported. The PVs were foundto be isolated while triggers in the sinus node regionwere identified. Following ablation linear lesionsisolation the triggers, AF was eliminated. Unfortu-nately, the patient had AF recurrence 3 weeks postprocedure.34)

Substrate Modification

A survey of the literature has identified strategiesin addition to PVI to improve long term success rateswithout concomitant antiarrhythmic drugs adminis-tration. These strategies include extensive or limitedlinear lesions, ablation of complex fractionated atrialelectrograms, regions that display dominant frequen-cy during atrial fibrillation and ablation of autonomicganglionated plexuses. The following section aims toreview the rationale of these strategies, its applica-tion and clinical outcome.

Linear ablationBiatrial surgical Maze procedure pioneered by

Cox et al. was effective at segmenting the atria toprevent the occurrence of reentrant activity. Sub-sequently, modification in this surgical procedurewas able to achieve up to 95% success rate at longterm follow-up with equivocal results betweenpatients with or without structural heart disease andin paroxysmal or persistent AF.35) This surgicaltechnique forms the basis for percutaneous catheterbased techniques in substrate modification.36) Linearlesions were placed in both RA and LA. Althougheffective, this percutaneous catheter based Mazestrategy was limited by long procedural time, highthromboembolic risk and potential proarrhythmiadue to incomplete lesion continuity between ana-tomical obstacles. Adjunctive linear ablation follow-ing PVI has been assessed in patients with eitherparoxysmal or persistent. AF. Haissaguerre andcolleagues37) performed linear ablation with an openirrigated catheter system to join the superior PVs,which then connect to the anterior mitral annulus inaddition to a cavotricuspid isthmus ablation. Com-plete linear block was achieved in 14 patients (58%)of which 9 (64%) remained AF free without anti-arrhythmic drugs at 28� 4 months of follow up. Ofthose who did not achieve complete linear block,only 30% remained AF free. Subsequently, the sameinvestigators assessed the efficacy of mitral isthmus

Figure 3 The predominant non-PV trigger sites responsiblefor initiation of AF are indicated in this segmented 3D contrast

cardiac CT of a patient presenting for AF ablation.LAA = left atrial appendage; LAPW = left atrial posterior wall;

LOM = ligament of Marshall; CS = coronary sinus; SVC =

superior vena cava; SN = sinus node; CT = crista terminalis;

IVC = inferior vena cava. The persistent left SVC is also a

potential trigger site.


89

ablation in patients with paroxysmal AF followingPVI. Mitral isthmus conduction block was achievedin 88% of the patients, but this approach oftenrequired ablation within the CS. This adjunctiveablation strategy safely increased success rate to79%.38) The strategy of adjunctive LA roof linearlesions and mitral isthmus ablation in addition towide area circumferential PV ablation has been alsodemonstrated to effective.39) Ernst et al. created oneof four linear ablation configurations with a 4mm tipcatheter using the CARTO electroanatomic mappingsystem. The long-term success rate in patients withcompleted linear lesions reached 76%.40)

The Cleveland Clinic performs SVC isolation asan adjunct approach to ICE-guided PVAI. This hasbeen shown to be safe and effective strategy intreating patients with all paroxysmal, persistent orpermanent AF.41) This approach has accounted for along-term success rate of 85.4% in paroxysmal AFand 74.9% in persistent/permanent AF.42) A repeatprocedure increases the success rate to 94.6% and92.9% respectively (Figure 4).

Complex fractionated atrial electrogram ablationComplex fractionated atrial electrogram (CFAE)

during AF as a substrate modifying strategy has beenstudied. These CFAEs have first been identifiedduring surgical epicardial mapping as sites exhibit-ing slow conduction, functional conduction blockand with pivot points.43) These complex electricalactivities exhibit short cycle lengths and heteroge-neous temporal and spatial patterns.44) Mapping and

ablation of CFAEs has been performed in patientswith PAF and chronic AF.45) CFAEs were defined asfractionated atrial electrograms (EGMs) with two ormore deflections and atrial EGMs with cycle length<120msec (averaged over 10 seconds). The CAR-TO system was used for electroanatomic correlation,dividing the RA and LA into nine distinct areas.Based on CFAE distributions three types weredescribed: Type I CFAEs were found in only onearea and focal ablation terminated AF; Type II intwo areas; Type III required three or more areas ofablation (PV was one area regardless of how manyPVs expressed CFAEs). CFAE ablation terminatedAF in 95% patients (28% required concomitantibutilide). Type III distribution was more commonlyidentified in patients with chronic AF. The interatrialseptum was the most common site for CFAEs, whichwere not present in the appendages. At the 1-yearfollow-up, 91% patients were free of arrhythmiasand symptoms, with 20% requiring a second proce-dure.

Ablation targeting dominant frequency sites dur-ing atrial fibrillation

Previous investigators have observed that admin-istration of antiarrhythmic drugs can decrease thefrequency of fibrillatory waves in the surface ECG.A low dominant frequency (DF) associated with ahigh degree of organization predicts a higher successrate for electric cardioversion of AF. Such observa-tions were seen in intracardiac recordings of isolatedsheep hearts under acetylcholine infusion. In addi-

Figure 4 Pulmonary venous (PV) isolation at their antrum aspect is shown in this segmented 3D contrast cardiac CT of another

patient presenting for AF ablation.Panel A illustrates the placement of RF lesions at the PVs antrum and at the SVC. Note the lesions’ continuity along the posterior wall and

at the anterior aspect of the LA roof. Panel B: PV antra (dotted circles) seen from the superior view. Also, note the close proximity of the

SVC to the right superior PV (RSPV) and the circumferential lesion created for SVC isolation. LSPV = left superior PV; LIPV = left

inferior PV; RIPV = right inferior PV; RAA = right atrial appendage.


90

tion, Jalife et al. provided evidence for spatiotem-poral organization due to the presence of AF. Theyalso reported a variable DF between the intracardiacmapping sites and sites with the highest DF areregarded as source of AF. In a series of 32patients,46) frequency mapping identifies localizedsites of high frequency activity during AF in humanswith different distributions in paroxysmal and per-manent AF. Patients with paroxysmal AF frequentlyharbored DF within the PVs. Ablation of these DFsites slowed and terminated 87% of patients withparoxysmal AF but none in those with permanentAF. Spectral analysis and frequency mapping canidentify high-frequency activity potentially correlat-ing with rotors responsible for the maintenance ofAF.47) The efficacy of this novel strategy for AFablation remains to be shown prospectively usingreal-time spectral analysis.

Ablation of autonomic ganglionated plexusesLocalization and ablation of autonomic ganglio-

nated plexuses (GP) in the LA have been evaluatedin animal and clinical studies (Figure 5). The GPs,present in epicardial fat pads, were identified byhigh-frequency stimulation (HFS—at PCL 50msec,12V, PW 1–10msec) at the PV and PVA (endocar-dial and epicardial sites). A positive response to HFSwas indicated by induction of AF and/or increasedvagal tone leading to bradycardia or AV block. HFSidentified GPs at four distinct locations (Figure 3):anterior to the right PVs, inferior to the RIPV (rightinferior pulmonary vein), superior and medial to theLSPV, and inferior to the LIPV. For GPs ablation,a median of 2–6 RF applications was required toeliminate the vagal response to HFS. Interestingly,GPs ablation abolished PV firing in 95% of patients,but did not affect sustained AF inducibility eitherprior to or following PVAI.

In the 33 patients undergoing this combinedapproach, all patients with PAF and 81% withpersistent were AF free at a median follow-up of 5months.48) The same investigators had observed thatGPs are commonly located within three major CFAEareas of the LA.49) They have also shown that GPstimulation can automatically induce conversion ofPV focal firing to AF and that this response iseliminated by injecting neuronal blockers into the fatpad.50) Experimental and clinical findings suggestthat LA autonomic denervation targeting GPs maydownregulate PV firing and suppress AF inducibil-ity. In our clinical experience with HFS, the GP sitesidentified by vagal response were tagged and blindedto the operator performing ablation. Following ourstandard PVAI procedure without targeting the GPs,

vagal response was abolished in all patients. Thissuggested nonintentional GP modulation duringPVAI. Interestingly, in another group of patientspresenting with recurrent AF following PVAI, HFSfailed to produce vagal response in all patients.51)

We have also assessed the long-term effect of GPsablation in a canine model and found that thedenervation effects achieved by ablation had dis-appeared on repeat evaluation at 4 weeks postabla-tion.52)

Optimizing the cure of atrial fibrillationThe reported successes with various ablation

strategies underlie the complex pathogenesis of AF.Different centers have utilized various combinationablation strategies to further improve the successrate. Linear LA lesions across areas of CFAE as anadjunct to LA circumferential ablation have beenstudied in a series of 100 patients with paroxysmalAF.53) Following LA circumferential ablation toencircle the left and right-sided pulmonary veinsduring AF, with additional ablation lines in theposterior left atrium and mitral isthmus with an 8-mm-tip catheter were performed. After completingthis lesion set AF was still present or induced in 60%of patients. These patients were randomized to eitherno ablation (Group 2, 30 patients) or further addi-tional ablation lines (Group 3, 30 patients) along theleft atrial septum, roof, and/or anterior wall targetingfractionated electrograms. At 6-month follow-up,67% of patients in group 2 were free of AF withoutdrug therapy compared with 86% of patients ingroup 3 (p ¼ 0:05).

Figure 5 The four regions containing the ganglionatedplexuses (GP) are illustrated in a segmented 3D contrast

cardiac CT.The GPs are identified by induction of AF and/or increased vagal

tone leading to bradycardia or AV block during high-frequency

stimulation. ARGP = anterior right GP; IRGP = inferior right

GP; SLGP = superior left GP; ILGP = inferior left GP.


91

The same investigators randomized 80 patientswith chronic AF to either LA circumferentialablation (n ¼ 40) or nonencircling linear ablation(n ¼ 40).54) In LA circumferential ablation, the PVswere encircled, with additional lines made in themitral isthmus and posterior wall or roof. In non-encircling linear ablation, 4� 1 ablation lines werecreated targeting areas of complex electrograms withthe objective of voltage abatement. Linear lesionswere created in the roof, anterior wall, septum,mitral isthmus and posterior annulus. At 9� 4

months follow up, 68% of patients who underwentLA circumferential ablation remained free of AFwithout the need of antiarrhythmic drugs comparedto 60% of patients who underwent nonencirclinglinear ablation.

A stepwise ablation curative approach with highincidence of AF termination has been evaluated in60 patients with long lasting persistent AF.55) Arandomized ablation sequence including PVI, SVC,and CS isolation and atrial ablation at sites exhibit-ing continuous electrical activity or CFAE termi-nated AF in 53% of patients. The addition of linearlesions at the LA roof, mitral isthmus, and cavo-tricuspid isthmus accounted for AF termination in atotal of 83% of patients. Termination of AF to eitheratrial tachycardias or sinus rhythm was typicallypreceded by lengthening of AF cycle length. Only13% of patients converted directly to sinus rhythm,while the majority developed atrial tachycardias.Activation and mapping entrainment was subse-quently performed to identify its mechanism. Focalatrial tachycardias were localized to the PVs, LAA,and CS and these sites were subsequently ablated.All focal atrial tachycardias were interrupted with upto 2 RF applications. Macroreentrant atrial tachy-cardias utilized the mitral isthmus, LA roof or thecavotricuspid isthmus and linear lesions were createdat these sites. The LAA, CS, and PV-LA junctionscontained the initiators of AF in most patients.Lesser critical LA structures included the roof,septum, posterior wall, and mitral isthmus. However,24 (40%) patients developed atrial tachycardia by 3months out of which 22 were sustained AT. Of these,23 underwent repeat procedure of 55 ATs. 4 patientsrequired 3 procedures in total. In these 23 patients,focal AT was localized to similar sites as identifiedby the first procedure. Similarly, macroreentrant ATsutilized the gaps or recovered conduction of previouslinear ablation at the mitral isthmus, LA roof orcavotricuspid isthmus. With this approach, 95% ofpatients remained in sinus rhythm at 11� 6 monthsof follow-up with majority of them without takingany antiarrhythmic drugs.

At the Cleveland Clinic, AF nest ablation guidedby real-time spectral mapping in SR as an adjunctiveapproach to ICE-guided PVAI and SVC isolation hasbeen evaluated in a randomized. Real-time spectralmapping using fast Fourier transform (FFT) in SRhas identified atrial sites with unusually highfrequencies, namely, fibrillar myocardium or AFnest56) at the LA appendage, left intraatrial septum,inferior posterior LA along the mitral annulus,SVC/RA junction, crista terminalis and the CS os(Figure 6). The AF nests may exhibit anisotropicconduction and short refractoriness. This highlyresonant tissue may harbor sites expressing CFAEduring AF. Ablation of atrial tissue identified as AFnests as a sole approach without intentional PVI, hasmaintained SR in 94.1% patients at 9:9� 5 monthsfollow-up. However, 41.1% remained on previouslyineffective AAD. Importantly, full lesion thicknessis not required to normalize the spectrum and maybe less likely to create a substrate for macroreentrantATs while sparing viable atrial myocardium. RFdelivery for 20–30 seconds typically abolishes thehigh-frequency potentials and normalizes the spec-trum of the local bipolar EGM. This approach maybe preferable to targeting continuous electricalactivity or CFAE sites that may be unrecognizeddue to their temporal variability. We have beenevaluating the adjunctive role of empirical ablationof typical AF nest regions in addition to ICE guidedapproach to PVAI and SVC isolation in a pro-spective randomized study. Initial results of 157patients with persistent or paroxysmal AF werepromising. We randomized these patients to eitherPVAI + SVC isolation + empirical ablation oftypical AF nest regions (group 1) or conventionalPVAI and SVC isolation (Group 2). The ablationstrategy for patients in Group 1 was safe and resultedin significant reduction of AF recurrences as com-pared to group 2. We eagerly await completion ofthis study.

Complications of Catheter Ablation of AtrialFibrillation

Extensive ablation within the atria coupled withinstrumentation with various catheters in the settingof aggressive anticoagulation during AF ablation setsthe stage for potential complications. A variety ofthese complications have been reported and some ofthem have been fatal. While the majority occurredduring or within 48 hours of the procedure, some areknown to have an insidious course. This sectionreviews these complications with an objective tosuggestive preventive and definitive measures.


92

Cerebrovascular eventsThe incidence of thromboembolism in AF ablation

ranges from 0.5 to 2.8%.57) A worldwide surveyreported periprocedural deaths due to massivecerebral thromboembolism in two patients out of8745 patients.58) Prevention remains the best strategyin reducing thromboembolic events during AFablation. At our centre, warfarin therapy is notdiscontinued prior to ablation. If patients withpersistent AF arrive to the EP laboratory with anon-therapeutic INR, a TEE is performed prior to theablation procedure to exclude any intra-atrial clots.Prior to the first transseptal puncture, a bolus ofheparin of about 100mg/kg is given and an infusionis started to maintain an ACT of 400 to 450 s. Bothtransseptal sheaths are continuously perfused withheparinized saline and meticulously flushed whencatheters are exchanged. In addition, we minimizechar formation during lesion creation by titratingpower delivery to prevent abrupt impedance rise.

ICE is routinely used for early detection of intra-cardiac thrombi. Despite these measures, we reporta cerebral thromboembolic rate of 0.6% in 3060patients who underwent PVAI at our centre (unpub-lished data). Only two deaths were reported, onewithin one month and the second death occurred 4years after the procedure. The rest of the patientsexperienced full neurological recovery.

PV stenosisThe incidence of PV stenosis, defined as >70%

narrowing, was reported to occur in 3.4%59) ofpatients. The pathogenesis of PV stenoses is due toan initial ablative insult that precipitates a healingreaction culminating in an endovascular contractionand proliferation of elastic lamina/intima. Theincidence of PV stenosis has decreased due thecurrent technique of PVAI that targets antra far fromPV ostia. Clinical diagnosis of PV stenosis is oftendifficult and a high index of suspicion is required.

Figure 6Panel A: The typical locations exhibiting fibrillar myocardium (AF nest) during real-time spectral mapping in sinus rhythm,

following PVs and SVC isolations, are shown in a segmented 3D contrast cardiac CT of a patient presenting for AF ablation. Note

in Panel B, the frequency spectra of two consecutive bipolar electrograms recorded from the CS (arrow in Panel C). The high-

frequency components characteristic of AF nests are indicated by the arrow in Panel B. Radiofrequency delivery up to 35W for

20 seconds eliminated the fibrillar pattern (shown in Panel D).


93

These patients present a variety of respiratorysymptoms, including cough (89%), hemoptysis(63%), dyspnea (58%), pleuritic chest pain (58%)and wheezing (42%).60) However up to a third ofpatients with severe PV stenosis remained asympto-matic while symptomatic patients were erroneouslydiagnosed as having pneumonia, new onset asthma,pulmonary embolism and/or lung cancer. At present,the gold standard for diagnosis of PV stenosis isspiral computed tomographic angiogram and/ormagnetic resonance. In a series of 23 patients withsevere PV stenosis, balloon angioplasty alone orwith stenting immediately relieved symptoms.However, in-stent or in-segment stenosis occurredin 61% of patients during a 17-month follow-up.61)

Neumann et al.62) reported the long-term outcome ofprimary angioplasty versus primary angioplasty withPV stenting for symptomatic severe PV stenosis in12 patients with a total of 15 PV stenoses. PVstenting resulted in no restenosis over the 12-monthfollow-up while those who underwent PV angio-plasty had a high recurrence rate of 73%.

We recently described 18 patients who progressedto total PV occlusion, defined as �95% stenosis orcomplete loss of patency of the PV on CT scan.63) Of1,780 patients who underwent AF RF ablation duringthe period of 1999 to 2004 at our institution, 16(0.9%) were found to have pulmonary vein occlusion(PVO) of at least 1 PV, whereas 2 additional patientswere referred from other institutions for furtherevaluation and management of their occluded PVs.The clinical presentation of these patients wasvariable; 4 (22.2%) patients were totally asympto-matic throughout their course (defined as grade 0), 2(11.1%) had mild symptoms of dry cough or dyspneaon moderate/severe exertion (NYHA class I;grade 1), 4 (22.2%) had moderate symptoms ofdyspnea on mild exertion (NYHA class II/III),together with persistent cough (grade 2). Three(16.7%) patients had severe symptoms as severedyspnea (NYHA class III/IV) with hemoptysis,fever, or pleuritic chest pain (grade 4) at their initialpresentation, whereas 5 other patients (27.8%). CTscan identified a total of 24 occluded veins in 18patients, with a mean of 1.4 veins per patient. Ofthese 24 occluded veins, the left superior PV had thegreatest incidence of occlusion (54.2%), followedby the left inferior PV (29.2%), right inferior PV(8.3%), and the right superior PV (8.3%). Byconsidering the total venous drainage of eachindependent lung, the percentage of stenosis of theaffected veins on the same side were added togetherto yield an average cumulative stenosis of thevascular cross-sectional area draining the affected

lung (cumulative stenosis index [CSI]). CSI = sumof percent stenosis of the unilateral veins divided bythe total number of ipsilateral veins. We concludedthat patients with a CSI � 75% and a relative lungperfusion of �25% based on lung perfusion scanhave the greatest risk of severe symptoms and lungdiseases. In these patients, early and, when required,repeated PV intervention should be considered forrestoration of pulmonary flow and prevention ofassociated lung disease.

Esophageal injury and fistulaThe posterior wall of the LA is adjacent to the

esophagus and RF ablation procedures may substan-tially elevate the temperature within the esophageallumen (Figure 7). Atrial-esophageal fistulas mayresult from this thermal injury leading to tissuenecrosis. Esophageal damage, perforation, and atrio-esophageal fistulas were first described after poste-rior left atrial radiofrequency ablation performedduring open-heart surgery.64) Subsequently, threepatients with this complication after percutaneousradiofrequency catheter ablation from experiencedcenters were published in 2004.65,66) We reporteda series of 9 patients who had this devastatingcomplication via an anonymous volunteer reportingmethod.67) Clinical presentations occurred withinabout 2 weeks of the ablation procedures (mean, 12days [range, 10 to 16 days]). Patients presented withgeneral malaise, leukocytosis, and persistent fever ofundetermined origin. All patients developed septicshock and cardiovascular collapse. Eight patientshad neurologic findings consistent with multipleembolic strokes. Of these, 2 had intravascular airon computed tomography. Six patients had echocar-diograpic features of endocarditis. Two patientspresented with symptoms consistent with transientangina associated with ST-segment elevation onelectrocardiography. Three patients reported sub-stantial gastrointestinal bleeding, but five patientshad occult bleeding documented by fecal testing.Unfortunately, all 9 patients died and autopsyconfirmed the presence of atrioesophgeal fistula.

Although atrioesophageal fistula formation isapparently rare, it seems to almost always be fatal.Therefore, evaluation and management of patientspresenting with this potential complication mustfocus on rapid diagnosis and triage. Fever, malaise,leukocytosis, dysphagia, and neurologic symptomsin patients with a recent catheter ablation procedureshould raise suspicion of atrioesophageal fistula. Thediagnosis must be rapidly confirmed by noninvasiveimaging techniques, including transthoracic echo-cardiogram, magnetic resonance imaging, or com-


94

puted tomographic scan, especially when combinedwith water-soluble contrast.68) Any form of esoph-ageal instrumentation should be avoided in thesepatients. Once diagnosed, prompt surgical interven-tion is required. Recently, a case of successfultemporary esophageal stenting to seal off theesophageal perforation following LA catheter abla-tion was reported.69)

We studied the influence of LA ablation bymeasuring intraluminal temperature via an esopha-geal thermometer in 81 patients who underwentPVAI.70) The esophageal intraluminal temperaturewas significantly higher when ablation was per-formed at the posterior LA, adjacent to the esoph-agus. This was also associated with microbubblesgeneration detected by ICE. We concluded thattitration of RF delivery with ICE in non openirrigated catheter systems and/or continuously mon-itoring esophageal temperature may be an option toenhance safety for posterior LA wall ablation. Sinceutilizing this strategy, we report no atrioesophagealfistulas. Alternative strategies in visualizing theesophagus during AF ablations include ingestion ofGastrografin dye, tagging the esophagus on 3Delectroanatomic mapping system and merging 3Drecreated image of the esophagus with an electro-anatomic shell of the LA.

Pericardial effusion and cardiac tamponadePericardial effusion is not uncommonly observed

in patients undergoing AF ablation. However, thereported incidence of cardiac tamponade is 1.22%(Worldwide survey of AF ablation). The commonmechanisms include cardiac perforation by variouscatheters or steam pops. We identified 26 (7.4%) out

of 352 patients who had steam pops over a three-month period during RF ablation for AF.71) Therewere 44 steam pops at these locales: left interatrialseptum (36.4%), PVA area (20.5%), right atrium(18.2%), left atrial roof (11.4%), left atrial posteriorwall (6.8%), mitral annulus (4.5%) and left atrialanterior wall (2.3%). Majority (>90%) of popsoccurred with the catheter orientated perpendicularlyto the tissue. Seventy-five percent of these popsoccurred with a power � 35W, temperature � 39Cand impedance � 90 ohms. There was not an ob-servable decrease in impedance prior to steam popsand no acute procedural complications. Preventivestrategies include titrating RF energy to preventsudden impedance rise during ablation. In openirrigated catheter systems, we do not observe suchrise in impedance prior to the occurrence of steampops. Titrating RF energy to prevent sudden rise incatheter tip temperature may be a more practicaloption.

Organized atrial tachyarrhythmiasLeft atrial tachycardia (AT) following segmental

or wide area circumferential PVI and LA circum-ferential ablation with additional linear ablation hasbeen variably reported as ranging from 2.5% to 27%.The mechanism of these ATs has been described asmacroreentry AT due to reentry related to gaps in thelinear ablation line. These gaps have been identifiedas a slow conduction isthmus in the reentry circuitprimarily located at the LA posterior wall or mitralisthmus. These macroreentry ATs are often sympto-matic necessitating a repeat ablation. On the otherhand, focal ATs typically originate from PV ostialsegments that have partially recovered conduction

Figure 7 Three-dimensional computed tomographic reconstruction of the left atrium in relation to the

esophagus.


95

to the LA. Of note, centers that adopt a strategy ofwide area circumferential PV ablation and linear LAablations72,73) appear to experience an increasedprevalence of AT. This may imply that the inabilityto create contiguous linear lesions and/or recoveryof partial conduction along such lesions can beproarrhythmic. We adopt the strategy of PVAIwithout empirical linear lesions in the LA in aneffort to minimize the incidence of iatrogenic AT inour centre.

Phrenic nerve injuryPhrenic nerve injury (PNI) is a rare complication

that may occur as a result of different catheterablation procedures. The course of the right phrenicnerve is lateral and slightly posterior to the SVC andRA, and descends anteriorly to the right superiorPV near the right superior PV-LA orifice alongsideof the pericardium, anterior to the lung hilum beforereflecting into the right diaphragm.74) Sanchez-Quintana et al.75) showed that the minimum distancebetween the right phrenic nerve and the rightsuperior PV or SVC was shorter than that betweenthe right phrenic and right superior PV-LA or SVC-RA junction.

We studied 17 patients who sustained PNIfollowing catheter ablation with a variety of energysources including RF ablation (13 patients), cryoa-blation (1 patient), ultrasound balloon ablation (2patients) and laser balloon ablation (1 patient).76)

Two patients with RF ablation-induced transientright phrenic nerve paresis were detected on fluoro-scopy during energy delivery. Diaphragmatic move-ment spontaneously recovered upon cessation of RFdelivery. The temporal correlation with onset andoffset of energy delivery and the observed transientnerve dysfunction may suggest a current mediatedelectromagnetic effect. RF currents can alter theaction potential duration and amplitude of the nerve,thus producing a conduction block before injurytakes place.77) Symptoms of PNI are variabledepending on patient’s preexisting lung condition.Fifty percent of our patients were asymptomaticwhile the majority of the remaining patients had mildsymptoms and did not need special therapy. Onlytwo patients developed lung complications thatrequired respiratory treatment. In our experience,all patients who suffered from persistent PNI dueto catheter ablation clinically recovered within amedian time of 6 months.

Early detection of transient PNI by fluoroscopycould result in fewer cases of persistent PNI. Whilehigh-output pacing prior to ablation in certain casesis a valuable tool to prevent phrenic nerve damage,

the most effective way is to deliver energies awayfrom the phrenic nerve. Ablation at the level of thePV antrum or the SVC-RA junction when perform-ing isolation of these veins for AF could represent apotential strategy to minimize the risk of PNI. Anovel method of protecting the phrenic nerve duringepicardial catheter ablation by using a ballooncatheter in the pericardial space to mechanicallyseparate the left phrenic nerve from the ablationcatheter has been recently reported.78)

Other complications of extracardiac damage in-cluded four patients in whom acute pyloric spasmand gastric hypomotility were observed followingposterior LA lesions.79)

Energy Sources for AF Ablation

Radiofrequency ablation is widely used in catheterablation of atrial fibrillation. The placement ofablation lesions may be tedious and time consumingespecially in patients with persistent AF. Contactwith the atrial tissue is paramount to effective energydelivery for creation of transmural lesion. As such,non-transmural lesions in linear ablations are proar-rhythmic to macroreentrant atrial tachycardias. Therisk of thrombus formation occurs especially whenthe electrode-tissue interface exceeds 80�C. WithRF ablation there is a risk of cardiac perforationassociated with the use of high power in the LA.Such limitations create an impetus to developtechnologies that use alternative forms of energydelivery. Any technology that can create uniformand transmural lesions, decreases the need of contactto deliver sufficient energy, reduces procedure timeand improves safety (Table 1).

Ultrasound energyUltrasound propagates sound waves at 2 to

20MHz to cause oscillation of molecules and atomsfor heat generation. Ultrasound energy decreasesproportionally with distance, unlike that of RFenergy that decreases by the square of the distance.Perhaps, the main advantage is its ability to becollimated as it passes through a fluid media. Thisforms the basis of a High-Intensity Focused Ultra-sound Balloon Catheter (HIFU-BC) for PVAI inpatients with AF. A feasibility study including 27patients with AF has just been reported.80) The 14 FrHIFU-BC has been developed to simplify PVAI withlower risk of thrombus formation, PV stenosis andleft atrial perforation.81) It uses a parabolic water-carbon dioxide balloon interface to focus a 20, 25or 30mm diameter ring of ultrasound 2 to 6mmforward of the balloon to produce a circumferential


96

transmural LA lesions with each application. TheHIFU-BC successfully isolated all PV antra in 74%of patients. At 12 months follow-up, 57% of patientswere free of symptomatic atrial arrhythmias. Therewere no PV stenosis or thromboembolic complica-tions. One patient sustained pulmonary hemorrhagedue to perforation of the lingular branch of the leftsuperior PV during manipulation of the guidewire.There was a case of phrenic nerve palsy in a patientwith a usually large right superior PV. The mainadvantages were low risk of thromboembolism sinceno direct heating of blood occurs to form thrombus.Internal irrigation of the balloon also minimizesformation of thrombus on its surface. Completecircumferential contact is not required for lesionplacement. The limitations include the need toposition the HIFU-BC over a guidewire necessitatinginstrumentation of individual PVs, non-deflectableand difficulty in completely sonication of the leftPVs as the balloon often extended over the LAappendage ridge, leaving a small segment non-isolated. Improvement in the design incorporating adeflectable curve may produce better results.

CryoablationCryoablation causes direct cell injury and induces

vascular mediated tissue injury. The advantagesinclude minimal endocardial disruption with preser-vation of underlying tissue architecture, less plateletactivation, lower thrombogenicity and stable adhe-sion of the catheter to the tissue during freezing.Recent clinical studies82) have shown that trans-venous cryoablation is a safe and feasible method forPV isolation with acute procedural success compa-rable to conventional RF ablation. Although thelong-term success rate appeared to be lower,83) nopatients developed PV stenosis. Further refinementof catheter designs to improve/increase tissue con-tact surface area or delivery of lower temperaturemay improve its efficacy.

Microwave energyMicrowave generates heat by inducing oscillation

of dipoles in a medium such as water. Hence, tissuewith high water content will allow better energytransfer and heat generation. This may be useful inscarred myocardium, as viable myocardium will be

Table 1 Features of different energy sources for catheter ablation of atrial fibrillation.

Radiofrequency Cryothermy Laser Microwave Ultrasound

Clinical Experience +++ ++ + + +

Potential for High Lower Low Low/medium Lowendocardialdisruption

Thrombogenicity High Low High Medium Medium

Mapping capability Limited Yes No No No

Ability to Create Requires optimal Requires Requires Optimal Requiredtransmural lesions contact optimal contact optimal contact and optimal

contact orientation Orientationpreferred

Lesion size ++ ++ +++ +++ ++

Linear lesion ++ + ++++ ++++ +

Perforation rate Low Low High Low Low

Procedure time Short Long Short Medium Short

Injury to adjacent structures:

Phrenic nerve ++ �/+ +/� ++ +++

Esophagus +++ � +/� + +++

Coronary artery +++ � � � �PV stenosis

Other advantages +++ Lack of Combined with Combined withelectromagnetic fiberoptic ultrasoundinterference catheter for system for

direct lesion direct lesionvisualization visualization

+ = Positive; � = negative


97

preferentially heated, creating deeper lesions withless surface heating and subsequent endocardialdisruption or coagulation formation. Unlike RFablation in which lesion expansion was minimalafter 60 s, the depth of microwave lesion size couldstill expand in depth.84) This may potentially damageadjacent structures such as the esophagus, phrenicnerve and coronary arteries. Although tissue contactwith the catheter is of less importance, a parallelantenna orientation to the tissue is required foreffective heating.85) Microwave ablation in thesurgical treatment of AF has been proven to be safeand effective with similar clinical efficacy compara-ble with conventional RF ablation.86) However,currently there is limited data on transvenous micro-wave catheter ablation and it remains an investiga-tional product.

Laser energyLaser (light amplification by stimulated emission

of radiation) emits photons in a monochromaticcoherent beam. Absorption of laser energy createsheat by induction of oscillation in water molecules.It could potentially create deeper lesions withreduced coagulum formation. In experimental mod-el, lesions created are sharply demarcated withoutepicardial crater or endocardial thrombus formation.Laser energy delivery requires a fiberoptic systemthat can double up as an angioscopy. Diode laser isa low energy laser with a wavelength of 980 nmesc.Using a linear diffuser, linear ablation can be de-livered. Recently, a laser balloon catheter has beendesigned to allow PV isolation by delivering circularbeams of energy at the PV antra. However, clinicalexperience in humans is scarce and the use of laserenergy in a fiberoptic balloon catheter ablation forPV isolation remains an investigational tool.

New Antiarrhythmic Agents for Treatmentof AF

DronedaroneDronedarone is a new synthetic non-iodinated

derivative of amiodarone undergoing phase III clin-ical trials. Similar to amiodarone, dronedaronepossesses in vitro electrophysiologic characteristicsof all four classes of antiarrhythmic action. Specif-ically, it blocks sodium channels at rapid pacingrates, prolongs cardiac action potentials and refrac-toriness, and possesses Caþþ antagonistic properties.Additionally, dronedarone shows a non-competitiveantiadrenergic action. Dronedarone at the dose of400mg twice daily was effective in preventing bothsymptomatic and asymptomatic recurrences of atrial

fibrillation or atrial flutter and had a safety profilesimilar to that of placebo.87) The added advantagewas an absence of thyroid and lung side effects.Moreover, there was no proarrhythmic effects ortorsades des pointes seen and its effect on QTprolongation is at most modest.88) However, in theANDROMEDA trial, patients with recent class IIIor IV congestive heart failure and left ventricularejection fraction < 35% had an excess mortalitywith dronedarone therapy vs placebo, leading toearly discontinuation of the trial. Hence, furtherexperimental studies and long-term clinical trials arerequired to provide additional evidence of efficacyand safety of dronedarone.

AzimilideAzimilide is another investigational Class III

antiarrhythmic agent. It blocks the slow delayedrectifier potassium current (IKs) in addition to therapid delayed rectifier outward potassium current(IKr). Initial studies showed promising results inprolongation in the time to the first recurrence of AFor atrial flutter compared with placebo with accept-able risk of torsade de pointes (0.9%) and earlyreversible severe neutropenia (0.2%).89) However,recent trial showed otherwise. 658 patients withsymptomatic persistent atrial fibrillation wererandomized to either placebo, azimilide or sotaloland all received planned cardioversion.90) Mediantime to AF recurrence was 14 days for azimilide, 12days for placebo, and 28 days for sotalol. Of note, 5patients on azimilide developed torsades de pointesand marked QTc prolongation occurred in 3.5% ofpatients on sotalol and 7.6% on azimilide resulting ininterruption of the study. These investigators con-cluded that azimilide is slightly inferior to sotalolwith at modest antiarrhythmic effect, thus limiting itsadministration for the treatment of AF.

Future Directions

Curative AF ablation requires significant expertisein operator skills, various ablation tools and imagingto deliver safe and effective therapy. It will be idealto pursue tools to shorten procedure time whiledelivering effective lesions safely, thereby avoidingpotentially life threatening or limiting complications.Remote magnetic navigation (Stereotaxis, Inc., St.Louis, MO, USA) and the robotic steerable guidecatheter (SGC, Hansen Medical, Inc., Palo Alto, CA,USA) are two technologies that could greatlyfacilitate movement of intracardiac catheter. Re-search on different energy delivery systems isongoing and we are hopeful of an ablation tool that


98

creates larger area of lesions in a controlled fashion.Ablation strategies incorporating a combination ofPV isolation and various forms of substrate mod-ification for treatment of persistent AF are underinvestigation.

Acknowledgements

Figures 1 to 6: Reproduced with permission from Heart

Rhythm

Figure 7: Reproduced with permission from Annuls of Internal

Medicine, American College of Physicians

Table 1: Reproduced with permission from Journal of Cardi-

ovascular Electrophysiology

References

1) Prinzmetal M, Corday E, Brill IC, et al: Mechanism ofauricular arrhythmias. Circulation 1950; 1: 241–245

2) Jalife J, Berenfeld O, Skanes A, et al: Mechanism ofatrial fibrillation: Mother rotors or multiple daughter

wavelets, or both? J Cardiovasc Electrophysiol 1998; 9:S2–S12

3) Haissaguerre M, Jais P, Shah DC, et al: Spontaneousinitiation of atrial fibrillation by ectopic beats originating

in the pulmonary veins. N Engl J Med 1998; 339: 659–666

4) Moubarak JB, Rozwadowski JV, Strzalka CT, et al:Pulmonary veins-left atrial junction: Anatomic and

histological study. Pacing Clin Electrophysiol 2000; 23:1836–1838

5) Ho SY, Cabrera JA, Tran VH, et al: Architecture of thepulmonary veins: Relevance to radiofrequency ablation.

Heart 2001; 86: 265–2706) Cheung DW: Pulmonary vein as an ectopic focus in

digitalis-induced arrhythmia. Nature 1981; 294: 582–5847) Chen YJ, Chen SA, Chang MS, et al: Arrhythmogenic

activity of cardiac muscle in pulmonary veins of the dog:

Implication for the genesis of atrial fibrillation. Cardio-vasc Res 2000; 48: 265–273

8) Perez-Lugones A, McMahon JT, Natale A, et al:Evidence of specialized conduction cells in human

pulmonary veins of patients with atrial fibrillation. JCardiovasc Electrophysiol 2003; 14: 803–809

9) Ouyang F, Bansch D, Ernst S, et al: Complete isolationof left atrium surrounding the pulmonary veins: New

insights from the double-Lasso technique in paroxysmalatrial fibrillation. Circulation 2004; 110: 2090–2096

10) Kato R, Lickfett L, Meininger GR, et al: Pulmonary veinanatomy in patients undergoing catheter ablation of atrial

fibrillation. Circulation 2003; 107: 2004–201011) Lickfett L, Dickfeld T, Kato R, et al: Changes of

pulmonary vein orifice size and location throughout thecardiac cycle: Dynamic analysis using magnetic reso-

nance cine imaging. J Cardiovasc Electrophysiol 2005;16: 582–588

12) Wood MA, Wittkampf FH, Henry D, et al: A comparisonof pulmonary vein ostial anatomy by computerized

tomography, echocardiography and venography in pa-tients with atrial fibrillation having radiofrequency

catheter ablation. Am J Cardiol 2004; 93: 49–53

13) Jongbloed MR, Bax JJ, Lamb HJ, et al: Multislicecomputed tomography versus intracardiac echocardiog-

raphy to evaluate the pulmonary veins before radio-frequency catheter ablation of atrial fibrillation: A head-

to head comparison. J Am Coll Cardiol 2005; 45: 343–350

14) Maleki K, Mohammadi R, Hart D, et al: Intracardiac

ultrasound detection of thrombus on transseptal sheath:Incidence, treatment and prevention. J Cardiovasc

Electrophysiol 2005; 16: 561–56515) Kanj M, Wazni O, Natale A: How to do circular mapping

catheter-guided pulmonary vein antrum isolation: TheCleveland Clinic approach. Heart Rhythm 2006; 3(7):

866–86916) Gerstenfeld E, Callans D, Dixit S, et al: Incidence and

location of focal atrial fibrillation triggers in patientsundergoing repeat pulmonary vein isolation: Implica-

tions for ablation strategies. J Cardiovasc Electrophysiol2003; 14: 685–690

17) Arruda M, Natale A: The Adjunctive role of nonpulmo-nary venous ablation in the cure of atrial fibrillation.

18) Spach MS, Barr RC, Jewett PH: Spread of excitationfrom the atrium into thoracic veins in human beings and

dogs. Am J Cardiol 1972; 30: 844–85419) Chen YJ, Chen YC, Yeh HI, et al: Electrophysiology and

arrhythmogenic activity of single cardiomyocytes fromcanine superior vena cava. Circulation 2002; 105: 2679–

268520) Lin WS, Tai CT, Hsieh MH, et al: Catheter ablation of

paroxysmal atrial fibrillation initiated by non-pulmonaryvein ectopy. Circulation 2003; 107: 3176–3183

21) Marrouche N, Saliba W, Natale A, et al: Feasibility andcomplications of electrical disconnection of the superior

vena cava guided by circular mapping in patients withsymptomatic atrial fibrillation. Pacing Clin Electrophy-

siol 2003; 26: 75022) Natale A, Pisano E, Beheiry S, et al: Ablation for right

and left atrial premature beats following cardioversionin patients with chronic atrial fibrillation refractory to

antiarrhythmic drugs. Am J Cardiol 2000; 85: 1372–1375

23) Lin YJ, Tai CT, Liu TY, et al: Electrophysiologicalmechanisms and catheter ablation of complex atrial

arrhythmias from crista terminalis: Insight from three-dimensional noncontact mapping. Pacing Clin Electro-

physiol 2004; 27: 1231–123924) Sanders P, Jais P, Hocini M, et al: Electrical disconnec-

tion of the coronary sinus by radiofrequency catheterablation to isolate a trigger of atrial fibrillation. J

Cardiovasc Electrophysiol 2004; 15: 364–36825) Oral H, Ozaydin M, Chugh A, et al: Role of the coronary

sinus in maintenance of atrial fibrillation. J CardiovascElectrophysiol 2003; 14: 1329–1336

26) Kim DT, Lai AC, Hwang C, et al: The ligament ofMarshall: A structural analysis in human hearts with

implications for atrial rrhythmias. J Am Coll Cardiol2000; 36: 1324–1327

27) Doshi RN, Wu TJ, Yashima M, et al: Relation betweenligament of Marshall and adrenergic atrial tachyarrhyth-

mia. Circulation 1999; 100: 876–88328) Hwang C, Wu TJ, Doshi RN, et al: Vein of Marshall

cannulation for the analysis of electrical activity in


99

patients with focal atrial fibrillation. Circulation 2000;101: 1503–1508

29) Hsu LF, Jais P, Keane D, Wharton JM, et al: Atrialfibrillation originating from the persistent left superior

vena cava. Circulation 2004; 109: 828–83230) Elayi C, Fahmy T, Natale A, et al: Left superior vena

cava isolation in patients undergoing pulmonary vein

antrum isolation: impact on atrial fibrillation recurrence.Heart Rhythm 2006; 3(9): 1019–1023

31) Takahashi Y, Sanders P, Haissaguerre M, et al: Dis-connection of the left atrial appendage for elimination of

foci maintaining atrial fibrillation. J Cardiovasc Electro-physiol 2005; 16: 917–919

32) Mansour M, Ruskin R, Keane D: Initiation of atrialfibrillation by ectopic beats originating from the ostium

of the inferior vena cava. J Cardiovasc Electrophysiol2002; 13: 1292–1295

33) Scavee C, Jais P, Haissaquerre M, et al: The inferior venacava: An exceptional source of atrial fibrillation. J

Cardiovasc Electrophysiol 2003; 14: 659–66234) Takahashi Y, Sanders P, Haissaguerre M, et al: Sinus

node region as an ultimate source driving chronic atrialfibrillation. J Cardiovasc Electrophysiol 2005; 16: 1023

35) Cox JL: Cardiac surgery for arrhythmias. J CardiovascElectrophysiol 2004; 15: 260–262

36) Swartz JF, Pellersels G, Silvers J, et al: A catheter basedcurative approach to atrial fibrillation in humans.

Circulation 1994; 90 (Suppl I): I-33537) Sanders P, Jais P, Hocini M, et al: Electrophysiologic

and clinical consequences of linear catheter ablation totransect the anterior left atrium in patients with atrial

fibrillation. Heart Rhythm 2004; 1: 176–18438) Jais P, Hocini M, Hsu LF, et al: Technique and results of

linear ablation at the mitral isthmus. Circulation 2004;110: 2996–3002

39) Pappone C, Manguso F, Vicedomini G, et al: Preventionof iatrogenic atrial tachycardia after ablation of atrial

fibrillation: A prospective randomized study comparingcircumferential pulmonary vein ablation with a modified

approach. Circulation 2004; 110: 3036–3042

40) Ernst S, Ouyang F, Lober F, et al: Catheter-inducedlinear lesions in the left atrium in patients with atrial

fibrillation: An electroanatomic study. J Am Coll Cardiol2003; 42: 1271–1282

41) Marrouche N, Saliba W, Schweikert R, et al: Feasibilityand complications of electrical disconnection of the

superior vena cava guided by circular mapping inpatients with symptomatic atrial fibrillation. Pacing Clin

Electrophysiol 2003; 26: 75042) Bhargava M, Marrouche N, Martin D, et al: Chronic cure

rate after pulmonary vein isolation in patients withnonparoxysmal atrial fibrillation: Impact of a second

ablation. J Am Coll Cardiol 2004; 43: 133A43) Konings KT, Smeets JL, Penn OC, et al: Configuration

of unipolar atrial electrograms during electrically in-duced atrial fibrillation in humans. Circulation 1997; 95:

1231–124144) Jais P, Haissaguerre M, Shah DC, et al: Regional

disparities of endocardial atrial activation in paroxysmalatrial fibrillation. Pacing Clin Electrophysiol 1996; 19

(11 Pt 2): 1998–2003

45) Nademanee K, McKenzie J, Kosar E, Schwab M, et al: Anew approach for catheter ablation of atrial fibrillation:

Mapping of the electrophysiologic substrate. J Am CollCardiol 2004; 43: 2044–2053

46) Sanders P, Berenfeld O, Hocini M, et al: Spectralanalysis identifies sites of high frequency activity

maintaining atrial fibrillation in humans. Circulation

2005; 112: 789–79747) Sarmast F, Kolli A, Zaitsev A, et al: Cholinergic atrial

fibrillation: I (K, ACh) gradients determine unequal left/right atrial frequencies and rotor dynamics. Cardiovasc

Res 2003; 59: 863–87348) Scherlag BJ, Nakagawa H, Jackman WM, et al:

Electrical stimulation to identify neural elements on theheart: Their role in atrial fibrillation. J Interv Cardiol

Electrophysiol 2005; 13 (Suppl 1): 37–4249) Nakagawa H, Jackman W, Scherlag B, Yokoyama K, et

al: Relationship of complex fractionated atrial electro-grams during atrial fibrillation to the location of cardiac

automatic ganglionated plexi in patients with atrialfibrillation. Circulation 2005; 112: II-746

50) Scherlag BJ, Yamanashi W, Patel U, et al: Autonomi-cally induced conversion of pulmonary vein focal firing

into atrial fibrillation. J Am Coll Cardiol 2005; 45: 1878–1886

51) Verma A, Saliba W, Lakkireddy D, et al: Vagalresponses induced by endocardial left atrial parasympa-

thetic ganglion stimulation (PGS) pre and post intra-cardiac echocardiography (ICE)-guided pulmonary vein

antrum isolation (PVAI) for atrial fibrillation (AF).Circulation 2005; 112: 17: II-746

52) Seil O, Zhang Y, Nassir FM, et al: Epicardial fat padablation does not have long-term effects on atrial

fibrillation inducibility. Circulation 2005; 112: 17: II-18953) Oral H, Chugh A, Lemola K, et al: Noninducibility of

atrial fibrillation as an end point of left atrial circum-ferential ablation for paroxysmal atrial fibrillation: A

randomized study. Circulation 2004; 110: 2797–280154) Oral H, Chugh A, Good E, et al: Randomized compar-

ison of encircling and nonencircling left atrial ablation

for chronic atrial fibrillation. Heart Rhythm 2005; 2:1165–1172

55) Haissaguerre M, Sanders P, Hocini M, et al: Catheterablation of long-lasting persistent atrial fibrillation:

Critical structures for termination. J Cardiovasc Electro-physiol 2005; 16: 1125–1137

56) Pachon MJC, Pachon MEI, Pachon MJC, et al: A newtreatment for atrial fibrillation based on spectral analysis

to guide the catheter RF-ablation. Europace 2004; 6:590–601

57) Iesaka Y: Complications of catheter ablation of atrialfibrillation: cause, prevention and management. J Car-

diovasc Electrophysiol. 2006 Dec; 17 Suppl 3: S50–S5558) Cappato R, Calkins H, Chen SA, et al: Worldwide survey

on the methods, efficacy, and safety of catheter ablationfor human atrial fibrillation. Circulation 2005; 111:

1100–110559) Saad EB, Rossillo A, Saad CP, et al: Pulmonary vein

stenosis after radiofrequency ablation of atrial fibrilla-tion: functional characterization, evolution, and influence

of the ablation strategy. Circulation 2003; 108: 3102–


100

310760) Ren J, Marchlinski FE, Callans DJ, et al: Increased

intensity of anticoagulation may reduce risk of thrombusduring atrial fibrillation ablation procedures in patients

with spontaneous echo contrast. J Cardiovasc Electro-physiol 2005; 16: 474–477

61) Packer DL, Keelan P, Munger TM, et al: Clinical

presentation, investigation, and management of pulmo-nary vein stenosis complicating ablation for atrial

fibrillation. Circulation 2005; 111: 546–55462) Neumann T, Sperzel J, Dill T, et al: Percutaneous

pulmonary vein stenting for the treatment of severestenosis after pulmonary vein isolation. J Cardiovasc

Electrophysiol 2005; 16: 1180–118863) Di Biase L, Tamer SF, Wazni OM, et al: Pulmonary vein

total occlusion following catheter ablation for atrialfibrillation: Clinical implications after long-term follow-

up. JACC 2006; 48(12): 2493–249964) Gillinov AM, Pettersson G, Rice TW: Esophageal injury

during radiofrequency ablation for atrial fibrillation. JThorac Cardiovasc Surg 2001; 122: 1239–1240

65) Scanavacca MI, D’Avila A, Parga J, et al: Left atrial-esophageal fistula following radiofrequency catheter

ablation of atrial fibrillation. J Cardiovasc Electrophysiol2004; 15: 960–962

66) Pappone C, Oral H, Santinelli V, et al: Atrio-esophagealfistula as a complication of percutaneous transcatheter

ablation of atrial fibrillation. Circulation 2004; 109:2724–2726

67) Cummings J, Schweikert R, Walid S, et al: BriefCommunication: Atrioesophageal fistulas after radiofre-

quency ablation. Ann Intern Med 2006; 144(8): 572–57468) Doll N, Borger M, Fabricius A, et al: Esophageal

perforation during left atrial ablation: is the risk toohigh? J Thorac Cardiovasc Surg 2003; 125: 836–842

69) Bunch TJ, Nelson J, Foley T, et al: Temporaryesophageal stenting allows healing of esophageal perfo-

rations following atrial fibrillation ablation procedure. JCardiovasc Electrophysiol 2006; 17: 435–439

70) Cummings JE, Schweikert RA, Saliba WI, et al: Assess-

ment of temperature, proximity, and course of theesophagus during radiofrequency ablation within the left

atrium. Circulation 2005; 112: 459–46471) Ching CK, Elayi C, Natale A, et al: Steams pops in real

life: Confessions of open irrigated catheter. (submitted asan abstract to HRS 07)

72) Mesas CE, Pappone C, Lang CCE, et al: Left atrialtachycardia after circumferential pulmonary vein abla-

tion for atrial fibrillation. J Am Coll Cardiol 2004; 44:1071–1079

73) Pappone C, Manguso F, Vicedomini G, et al: Preventionof iatrogenic atrial tachycardia after ablation of atrial

fibrillation: A prospective randomized study comparingcircumferential pulmonary vein ablation with a modified

approach. Circulation 2004; 110: 3036–304274) Fell SC: Surgical anatomy of the diaphragm and the

phrenic nerve. Chest Surg Clin N Am 1998; 8: 281–29475) Sanchez-Quintana D, Cabrera JA, Climent V, et al: How

close are the phrenic nerves to cardiac structures?Implications for cardiac interventionalists. J Cardiovasc

Electrophysiol 2005; 16: 309–313

76) Bai R, Patel D, Di Biase L, et al: Phrenic nerve injuryafter catheter ablation: should we worry about this

complication? J Cardiovasc Electrophysiol 2006; 17(9):944–948

77) Bunch TJ, Bruce GK, Mahapatra S, et al: Mechanisms ofphrenic nerve injury during radiofrequency ablation at

the pulmonary vein orifice. J Cardiovasc Electrophysiol

2005; 16: 1318–132578) Buch E, Vaseghi M, Cesario DA, et al: A novel method

for preventing phrenic nerve injury during catheterablation. Heart Rhythm 2007; 4(1): 95–98

79) Shah D, Dumonceau JM, Burri H, et al: Acute pyloricspasm and gastric hypomotility: an extracardiac adverse

effect of percutaneous radiofrequency ablation for atrialfibrillation. J Am Coll Cardiol 2005; 46: 327–330

80) Nakagawa H, Antz M, Wong T, et al: Initial ExperienceUsing a Forward Directed, High-Intensity Focused

Ultrasound Balloon Catheter for Pulmonary Vein An-trum Isolation in Patients with Atrial Fibrillation. J

Cardiovasc Electrophysiol 2007 Jan 9 (Epub ahead ofprint)

81) Nakagawa H, Aoyama H, Pitha JV, et al: Pre-clinicalcanine testing of a novel high intensity, forward-focused

ultrasound balloon catheter for pulmonary vein isolation.Pacing Clin Electrophysiol 2003; 26: 954A. Abstract.

82) Tse HF, Reek S, Timmermans C, et al: Pulmonary veinisolation using transvenous catheter cryoablation for

treatment of atrial fibrillation without risk of pulmonaryvein stenosis. J Am Coll Cardiol 2003; 42: 752–758

83) Verma A, Natale A: Should atrial fibrillation ablation beconsidered first-line therapy for some patients? Why

atrial fibrillation ablation should be considered first-linetherapy for some patients. Circulation 2005; 112: 1214–

122284) Liem LB, Berube D: Transcatheter microwave ablation.

In: Liem LB, Downar E, ed. Progress in CatheterAblation. Clinical Application of New Mapping and

Ablation Technology. Doordrecht, The Netherlands:Kluwer Academic Publishers, 2002, 425–442

85) Whayne JG, Nath S, Haines DE: Microwave catheter

ablation of myocardium in vitro. Assessment of thecharacteristics of tissue heating and injury. Circulation

1994; 89: 2390–239586) Wisser W, Khazen C, Deviatko E, et al: Microwave and

radiofrequency ablation yield similar success rates fortreatment of chronic atrial fibrillation. Eur J Cardiothorac

Surg 2004; 25: 1011–101787) Zareba KM: Dronederone:Anewantiarrhythmic agent.

Drugs of Today 2006; 42: 75–8688) Touboul P, Brugada J, Capucci A, et al: Dronedarone for

prevention of atrial fibrillation: A dose-ranging study.Eur Heart J 2003; 24: 1481–1487

89) Connolly SJ, Schnell DJ, Page RL, et al: Dose-responserelations of azimilide in the management of symptomat-

ic, recurrent, atrial fibrillation. Am J Cardiol 2001; 88:974–979

90) Lombardi F, Borggrefe M, Ruzyllo W, et al: Azimilidevs. placebo and sotalol for persistent atrial fibrillation:

the A-COMET-II (Azimilide-CardiOversion Maint-Enance Trial-II) trial. Eur Heart J 2006; 27(18): 2224–

2231


101

Documents

Catheter Ablation of Atrial FibrillationIntroduction Atrial ﬁbrillation (AF) is the most common cardiac arrhythmia seen in clinical practice and results in signiﬁcant morbidity,