Nonpulmonary Vein Foci: Do They Exist?DIPEN SHAH,† MICHEL HAISSAGUERRE,* PIERRE JAIS,* and MELEZE HOCINI*From the †Division of Cardiology, Hopital Cantonal de Geneve, Geneva, Switzerland, and *the Service deRythmologie, Hopital Haut Leveque, Bordeaux Pessac, France

SHAH, D., ET AL.: Nonpulmonary Vein Foci: Do They Exist? Though the majority of foci triggering atrialfibrillation (AF) have been mapped to the pulmonary veins (PV), recurrence of paroxysmal AF after isola-tion of all four pulmonary veins indicates the presence of other foci. In a series of 160 consecutive patientswho underwent PV ablation, endocardial mapping of AF and ectopy recurring after PV isolation was per-formed. Thirty-six patients (24%) had a total of 85 non-PV foci; 39 were mapped to the ostia of ablatedPVs, 30 to the posterior left atrium (LA), 5 to other parts of the LA, 5 to the right atrium (RA), 4 to thecoronary sinus (CS), and 3 to the superior vena cava (SVC) (including one persistent left SVC). Mappingwas confirmed by successful ablation. At least 16 foci could not be localized and after a follow-up of8 ± 6 months, 68% of patients were free of AF without any antiarrhythmic treatment. The occurrence ofnon-PV foci correlated with recurrence of AF, perhaps as a correlate of insufficient ostial ablation. Thesedata reinforce the requirement for more proximal disconnection of the PVs by performing ablation withinthe LA. In patients with non-PV foci that are difficult to map conventionally, the use of surface ECG data,or multielectrode contact or noncontact mapping arrays, or substrate modifying/excluding ablation maybe helpful in ablating these foci and therefore eliminating AF. (PACE 2003; 26[Pt. II]:1631–1635)

atrial fibrillation, ablation, pulmonary vein

IntroductionEndocardial mapping in patients with parox-

ysmal atrial fibrillation (AF) has shown that thegreat majority of paroxysms are initiated by ec-topics originating from the pulmonary veins (PV).1Catheter ablation using radiofrequency (RF) en-ergy delivered at the site of origin of these ec-topics has been effective in selected patients al-beit with limited success rates and high recurrencerates. Based on ‘cul-de-sac’ activation of the my-ocardium in the PVs, ostial ablation has been per-formed with the endpoint of complete isolationfrom the left atrium (LA) in order to improve suc-cess rates.2 In spite of successful PV isolation andan improvement in cure rates, a minority of pa-tients have recurrences that are triggered by non-PV foci. The following report briefly describes andanalyzes the results of mapping of non-PV foci inthese patients.

Consecutive patients referred for drug refrac-tory and symptomatic frequent paroxysmal AFwere included in this study group. Preconditionsincluded effective anticoagulation (with an INRbetween 2 and 3) for a minimum of 4 weeks pre-ceding catheter ablation, as well as the exclusionof LA thrombi by a transoesophageal echocardio-gram performed within the week preceding abla-tion. The procedure was performed after obtaininginformed consent, with the patient in the fastingstate for 6 hours.

After administering light sedation (midazo-lam 2.5 mg and nalbuphine 2.5 mg IV), femoralvenous access was obtained and multielectrodecatheters introduced into the right atrium (RA) orthe coronary sinus (CS), the LA, and the PVs as re-

quired. Two catheters were introduced into the LAthrough a transseptal puncture. A multielectrodecatheter (circular decapolar) accompanied a 4-mmtip quadripolar ablation catheter equipped with athermocouple in the LA. Intravenous heparin wasadministered in a dose of 50 units/kg and there-after boluses of 25 units/kg were administeredhourly. In some cases, an irrigated tip catheter wasused in order to deliver desired RF power levels.Bipolar endocardial signals were filtered througha bandpass of 30 to 500 Hz and recorded on a mul-tichannel polygraph; bipolar stimulation was per-formed as required at outputs at least two timesthreshold.

RF energy was delivered in a unipolar modefrom the distal tip of the ablation catheter. A powerlimit of 30 to 35 W was imposed for ablationwithin thoracic veins; a temperature limit of 50◦Cto 55◦C was imposed for the LA while 65◦C wasimposed for the RA. Maximum power in the atriawas limited to 50 W.

Mapping and Ablation Strategy

Patients with spontaneous or induced ec-topics or initiations of paroxysmal AF underwentmapping to localize the earliest activation. In casesof sustained fibrillation, cardioversion was per-formed and in cases of arrhythmias or reinitiation,mapping was repeated. Arrhythmias were mostcommonly induced by a combination of postpac-ing pauses and isoprenaline infusion. Careful map-ping was performed in order to precisely localizethe origin of ectopy outside the PV. PVs giving riseto ectopics were isolated and as further evidence ofrecurrences from unablated PVs was documented,

PACE, Vol. 26 July 2003, Part II 1631

SHAH, ET AL.

Figure 1. Diagrammatic representation of sites of successfully ablated non-PV foci in patients inwhom AF was successfully eliminated (left) versus those in whom AF recurred (right). Note thenear identical distribution and the marked clustering near PV ostia and in the posterior LA.

PVs not documented to be arrhythmogenic werealso isolated. A reversal of the activation sequencetypical of sinus rhythm (adjacent LA preceding PVactivation) was considered to indicate arrhythmiaorigin from that vein. RF energy was delivered atostial sites of earliest PV activation (documentedon the circular catheter) during sinus rhythm inorder to achieve PV isolation.2 Selective PV an-giography was performed before and after ablationto delineate anatomy, choose appropriate circularcatheter size, and assess the effects of RF delivery.The PV ostium was defined as the level of maxi-mum change in angiographic diameter.3 The end-point of PV isolation was the elimination of all PVpotentials distal to the site of ablation. Nonpul-monary venous foci were defined as being locatedin the RA, superior vena cava (SVC), CS, or in theLA on the atrial side of the level of PV isolation.RF energy was delivered to eliminate these foci atthe site of earliest activation.

After the procedure patients were monitoredtelemetrically and were maintained on thera-peutic levels of heparin until oral anticoagula-tion achieved appropriate INR levels. Follow-up included cardiology consultations, ECG,exercise stress test and 24-hour holter monitor at1, 3, 6, and 12 months. An electrophysiologicalstudy and repeat ablation was advised in case ofrecurrence.

ResultsOne hundred sixty patients including 130

males and 30 females (age 53 ± 11 years) under-went catheter ablation. Twenty-seven percent hadstructural heart disease including coronary arterydisease, dilated and hypertrophic cardiomyopa-thy, and valvular heart disease.

One, 2, 3, or 4 PVs were isolated in 6, 14, 44,and 96 patients, respectively. Five hundred thirty-

five of 540 targeted PVs were successfully isolated.Though they were difficult to precisely map be-cause of sporadic unpredictable discharge and re-peated induction of sustained AF, 85 non-PV fociwere ablated in 36 patients. Thirty-nine originatedfrom the ostia of ablated PVs including the zone be-tween ipsilateral veins, 30 from the posterior LA,5 from other parts of the LA, 5 from the RA, 4 fromthe CS, 2 from the (right) SVC, and one from apersistent left SVC. (Figs. 1–4). At least 16 focicould not be precisely localized. Three patientsdeveloped acute PV stenosis (>50% narrowing)but were asymptomatic. Two patients developed apericardial effusion and one had coronary spasm.

A single ablation was performed in 84 pa-tients while two or more ablations were performedbecause of recurrence in 76 patients. At repeatstudy, 7% of patients had arrhythmias originatingfrom previously unablated PVs; 22% had recov-ered conduction to previously isolated veins; 29%had arrhythmias from non-PV foci only, while theremaining 42% had a combination of the abovefindings. At follow-up of 8 ± 6 months, 108 pa-tients were free of AF. Of the remaining, 25 under-went linear LA ablation while 27 received only an-tiarrhythmic treatment. Twelve patients were ren-dered free of AF after additional linear ablationand 14 (5 after linear ablation) with antiarrhyth-mic treatment. Only 23% of patients free of AFafter ablation had non-PV foci as opposed to 79%of those who still had AF (Fig. 1).

DiscussionPV isolation was routinely possible and re-

sulted in the elimination of paroxysmal AF innearly 70% of this patient cohort when combinedwith the ablation of mappable triggers.2 As a resultof PV isolation, all triggers distal to the level of ab-lation were neutralized by rapid and anatomically

1632 July 2003, Part II PACE, Vol. 26

NON-PV FOCI: DO THEY EXIST?

Figure 2. Twelve-lead ECG of AF initiation by a nonvenous focus under the influence of Isoprenaline after PV isolation.Note the absence of prodromal bigeminy or isolated ectopics which is a common characteristic of PV foci.

Figure 3. An example of a triggering focus in a persistent left SVC. On the right are shown recordings from the leftsuperior pulmonary vein (LSPV) after extensive ostial ablation that failed to eliminate a discrete potential. The nearfield origin of this potential is in fact from the muscle of a persistent left SVC draining into the coronary sinus(angiogram on the left) as shown by Lasso recordings. The tracings from this circular catheter placed in the LSVCdemonstrate a sharp potential (*) coinciding with the second deflection recorded in sinus rhythm in the LSPV.

PACE, Vol. 26 July 2003, Part II 1633

SHAH, ET AL.

Figure 4. Same patient as in Figure 3. On the left, blocked as well as conducted discharges (arrows) from the LSVCmyocardium with an activation pattern reminiscent of PV ectopy. On the right, disconnection of LSVC myocardiumduring RF delivery as shown by elimination of the late spike potential ( *).

demarcated ablation. On the other hand, the spo-radic and unpredictable firing of triggers lead tosustained paroxysms of AF (Fig. 2) and limitedcatheter access to the LA, render mapping andsubsequent ablation of non-PV triggers difficult.Moreover, unlike venous foci, the only endpointof successful elimination, the disappearance ofarrhythmias, was generally unreliable because oftheir capricious and unpredictable behavior. Dis-connection of venous myocardium from the rel-evant atrium, by contrast, offered concrete evi-dence of neutralization independent of arrhythmiamanifestation. The presence of nonvenous foci ap-peared to correlate with the outcome of ablation:the majority of patients with recurrent AF had ex-tra PV foci.

The majority of these foci were located inproximity to the level of PV disconnection: in fact47% of these were situated within ∼1 cm of thesite of ablation and another 35% in the poste-rior LA. Thus, the anatomic localization of thesefoci was remarkably consistent and decreased infrequency with increasing distance from the os-tia of the PVs (Fig. 1). These findings bring intocontention our understanding of what defines thedemarcation between the PVs and the LA, andreflects insufficient proximal ablation. The idealaim of ablation of the PVs should be to discon-nect them proximal to all overtly or potentiallyarrhythmic tissue; this is straightforward for man-ifest arrhythmia but not so for eliminating po-tentially arrhythmogenic PV tissue. Clearly, theanatomy of this region is a continuum, and avail-

able histological data does not provide any usefuldistinctions other than a gradual transition fromatrium to vein. Whether this arrhythmogenic ac-tivity represents an electrophysiological manifes-tation of irritation from the border zone of ablationor whether myocardial fibers in this region (likethose within the PVs) also have an intrinsic ten-dency to arrhythmogenic firing is not clear. It hasalso recently been shown that nearly one third ofpatients with early recurrence of AF after PV abla-tion develop a spontaneous, complete, and lastingremission of their arrhythmias within the monthfollowing the ablation.4 This temporal course cor-relates well with healing related changes in theborder zone of ablation. Based on the above data, astrategy of more proximal left atrial ablation withthe reproducible endpoint of achieving PV isola-tion is logical in order to reduce recurrence rates.The effect of ostial ablation on the substrate sup-porting reentry and maintaining AF initiated byfocal triggers is unclear but may certainly be anadditional mechanism for eliminating arrhythmia.

A minority of non-PV foci originate fromother parts of the LA or from the RA. Of in-terest, other great veins opening into the heartsuch as the SVC, CS, or even the embryologicalanomaly of the persistent left SVC, all of whichpossess sleeves of striated myocardium continu-ous with the right and/or left atria, have also beenshown to be the sites of origin of foci giving riseto ectopy (Figs. 3 and 4). Data from other labo-ratories have borne this out.5 Foci within theseveins are amenable to anatomic ablation since

1634 July 2003, Part II PACE, Vol. 26

NON-PV FOCI: DO THEY EXIST?

noncircumferential ablation at key sites can dis-connect arrhythmogenic myocardium from theatria (Fig. 4). The same is probably not true of theCS given that the available histopathological datasuggest the presence of multiple strands of my-ocardium interconnecting and weaving betweenthe left and right atrium and the CS.6 Residualstrands of myocardium within the usually oblit-erated ligament of Marshall (which anomalouslypersists with a patent lumen as a left SVC draininginto the CS) have also been implicated in arrhyth-mia generation. A small vein of Marshall (distallybecoming atretic as the ligament) is angiograph-ically found in the majority of individuals, evenwithout AF. In accordance with animal data, theremay be more than just the proximal connection atits root to the CS musculature. Connections to theleft PV ostial region have been described in pre-liminary reports7 and confirmed by mapping dataobtained by cannulating this vein using thin 3 to4 Fr catheters. Though in our study we did notfind focal activity requiring epicardial ablation inthe region of the ligament of Marshall, this may

indeed have been the source of some arrhythmiain spite of successful ablation by endocardially de-livered high RF power.

Nonvenous foci are significant challenges thatneed to be overcome in order to maximize the cu-rative efficacy for AF, and an anatomic strategy ofisolating the PV more proximally is the most prac-tical solution to neutralizing them. Better map-ping strategies may be required in order to lo-cate and ablate foci that are remote from the PVostia. These include surface ECG based localiza-tion strategies8 or multielectrode arrays (contactor noncontact). The foregoing data, however, applyonly to mappable arrhythmias and even with moreproximal or left atrial isolation strategies, part ofinitiating and/or maintaining substrate, at leastin some patients (particularly those with chronicAF), is not neutralized, thus explaining the factthat success rates have not yet reached the lev-els of other supraventricular tachycardias. In thiscontext, rapid, practical, and effective substratemodifying or excluding ablation strategies must bedeveloped.

References1. Haissaguerre M, Jais P, Shah DC, et al. Spontaneous initiation of

atrial fibrillation by ectopic beats originating in the pulmonary veins.N Engl J Med 1998; 339:659–666.

2. Haissaguerre M, Shah DC, Jais P, et al. Electrophysiological break-throughs from the left atrium to the pulmonary veins. Circulation2000; 102:2463–2465.

3. Shah DC, Haı̈ssaguerre M, Jaı̈s P, et al. Curative catheter ablationof paroxysmal atrial fibrillation in 200 patients: Strategy for presen-tations ranging from sustained atrial fibrillation to no arrhythmias.PACE 2001; 24:1541–1558.

4. Oral H, Knight BP, Ozayd M, et al. Clinical significance of earlyrecurrences of atrial fibrillation after pulmonary vein isolation. J AmColl Cardiol 2002; 40:100–104.

5. Tsai CF, Tai CT, Hsieh MH, et al. Initiation of atrial fibrillation by

ectopic beats originating from the superior vena cava: Electrophys-iological characteristics and results of radiofrequency ablation. Cir-culation 2000; 102:67–74.

6. Chauvin M, Shah DC, Haı̈ssaguerre M, et al. The anatomicbasis of connections between the coronary sinus musculatureand the left atrium in humans. Circulation 2000; 101:647–652.

7. Hwang C, Chen PS. Simultaneous ablation of the left pulmonic ve-nous truncus and ligament of Marshall in patients with paroxysmalatrial fibrillation. (abstract) PACE 2002; 25:683.

8. Choi KJ, Shah DC, Hocini M, et al. T wave substraction combinedwith pacemapping for the prediction of pulmonary vein ectopysource in patients with paroxysmal atrial fibrillation. (abstract) PACE2002; 25:572.

PACE, Vol. 26 July 2003, Part II 1635