1067

Spontaneous Pulmonary Vein Firing in Man: Relationship toTachycardia-Pause Early Afterdepolarizations and Triggered

Arrhythmia in Canine Pulmonary Veins In VitroEUGENE PATTERSON, PH.D., WARREN M. JACKMAN, M.D., KAREN J. BECKMAN, M.D.,

RALPH LAZZARA, M.D., DEBORAH LOCKWOOD, B.M.B.CH., M.A.,BENJAMIN J. SCHERLAG, PH.D., RICHARD WU, M.D.,∗ and SUNNY PO, PH.D., M.D.

From the University of Oklahoma Health Sciences Center, OU Medical Center; DVA Medical Center, Oklahoma City, Oklahoma, USA;and ∗The University of Texas Southwestern Medical Center, Dallas, TX, USA

Triggered Firing in Pulmonary Veins. Introduction: Rapid firing originating within pulmonaryveins (PVs) initiates atrial fibrillation (AF). The following studies were performed to evaluate spontaneousPV firing in patients with AF to distinguish focal versus reentrant mechanisms.

Methods: Intracardiac recordings were obtained in 18 patients demonstrating paroxysmal AF. Micro-electrode (ME) recordings were obtained from superfused canine PV sleeves (N = 48).

Results: Spontaneous PV firing (566 ± 16 bpm; 127 ± 6 ms cycle length) giving rise to AF (52 episodes)was observed. Tachycardia-pause initiation was present in 132 of 200 episodes of rapid PV firing and 34 of 52AF episodes. The pause cycle length preceding PV firing was 1,039 ± 86 ms following tachycardia (420 ± 40ms cycle length). The remaining episodes were initiated following a 702 ± 32 ms pause during sinus rhythm(588 ± 63 ms). Spontaneous firing recorded with a multipolar mapping catheter did not detect electri-cal activity bridging the diastolic interval between the initial ectopic and preceding post-pause sinus beat.Tachycardia-pause initiated PV firing (138 ± 7 ms coupling interval) in patients correlated with tachycardia-pause enhanced isometric force, early afterdepolarization (EAD) amplitude, and triggered firing withincanine PVs. Rapid firing (1,172 ± 134 bpm; 51 ± 8 ms cycle length) following an abbreviated couplinginterval (69 ± 12 ms) was initiated in 13 of 18 canine PVs following tachycardia-pause pacing during nore-pinephrine + acetylcholine superfusion. Stimulation selectively activating local autonomic nerve terminalsfacilitated tachycardia-pause triggered firing in canine PVs (5 of 15 vs 0 of 15; P < 0.05).

Conclusions: The studies demonstrate (1) tachycardia-pause initiation of rapid, short-coupled PV firingin AF patients and (2) tachycardia-pause facilitation of isometric force, EAD formation, and autonomic-dependent triggered firing within canine PVs, suggestive of a common arrhythmia mechanism. (J CardiovascElectrophysiol, Vol. 18, pp. 1067-1075, October 2007)

pulmonary veins, atrial fibrillation, calcium transient triggering, arrhythmia, cardiac electrophysiology

Introduction

Rapid focal excitation originating within left atrial my-ocardial extensions into the pulmonary veins (PVs) initiatesatrial fibrillation (AF) in the majority of patients present-ing for arrhythmia ablation.1 Short periods of focal firingmay be observed at rates ≥ 500/min as the initiating eventfor AF.1,2 Coupling intervals for PV extrasystoles are ab-breviated (<150 ms) and refractory periods within the PVmyocardial sleeves as brief as 60 ms have been observed.3

Localized reentry of atrial impulses entering the PV hasbeen proposed as a mechanism for the “focal” firing that ini-tiates AF in man. It has been demonstrated that the structuraland electrophysiological properties of the PV are favorable

The research was supported in part by Grants-in-Aid from the AmericanHeart Association, Heartland Affiliate.

Address for correspondence: Eugene Patterson, Ph.D., 6E 103 ET CardiacArrhythmia Research Institute, 1200 Everett Drive, Oklahoma City, OK73104, USA. Fax: 405-271-7455; E-mail: eugene-patterson@ouhsc.edu

Manuscript received 27 February 2007; Revised manuscript received 15 May2007; Accepted for publication 21 May 2007.

doi: 10.1111/j.1540-8167.2007.00909.x

to reentry; reentry within the PV can be readily induced byprogrammed stimulation (short-coupled extrasystoles) anddemonstrated during AF.4-6 However, the role of reentry inthe spontaneous focal firing that initiates AF during sinusrhythm is unclear.

An arrhythmia mechanism producing very rapid focalfiring within isolated superfused canine PVs during auto-nomic nerve stimulation7 or combined acetylcholine + nore-pinephrine administration8 has been described and has beentermed “calcium transient triggering.” In this mechanism,a high cytosolic Ca2+ produced by enhanced calcium tran-sient is maintained during terminal repolarization of an ab-breviated action potential. The enhanced Ca+2

i occurs atmembrane voltages negative to the equilibrium potential forsodium–calcium exchange (NCX), activating inward NCXcurrent, generating an early afterdepolarization (EAD), andre-excitation of myocardium. Inherent to this mechanismare requirements for both an enhancement of the calciumtransient by sympathetic stimulation and/or pacing, and anabbreviated action potential dependent upon both the in-trinsic electrophysiologic properties of the tissue as well asacetylcholine-induced shortening of the action potential.7,8

The brief action potentials of the PVs provide a favorable pre-condition facilitating localized reentry9,10 as well as calciumtransient triggering.7,8

1068 Journal of Cardiovascular Electrophysiology Vol. 18, No. 10, October 2007

The purpose of the present studies was to examine initi-ation patterns of PV firing in patients with spontaneous andpacing-induced PV firing in an attempt to distinguish a fo-cal from reentrant mechanism for the beats that originatedwithin the PVs. Direct recordings of spontaneous PV firingwere obtained using a multipolar electrode basket catheter.The results are directly compared with the initiation patternsof EAD formation and the demonstrated focal firing observedin isolated superfused canine PV preparations.

Methods

Patients

The electrophysiologic records were examined for 15 con-secutive patients (1) undergoing electrophysiologic study andcatheter ablation of paroxysmal or persistent AF and (2)demonstrating spontaneous firing from one or more PVs. Thepatients were studied under general anesthesia (propofol sup-plemented with nitrous oxide and/or desflurane in some pa-tients) while also receiving cisatracurium. Electrode catheterswere placed to record His bundle (HB), right atrial, and coro-nary sinus (CS) activation. The right atrium was paced fromthe appendage. Left atrial activation (including PV sleeveactivation during focal firing) was recorded using a circularmapping catheter with 10 bipolar recording sites (Lasso�)catheter, Biosense Webster (Diamond Bar, CA, USA) anda mapping catheter with a 3.5-mm-tip inserted through atransseptal sheath. Electrical recordings were stored digitallyand recovered using a Bard� electrophysiology system. Ap-proval for the human studies was obtained from the Institu-tional Review Board at the University of Oklahoma HealthSciences Center. In three additional patients, a 31 mm 64-pole (32 bipolar electrode recording sites) (Constellation�,Boston Scientific, Natick, MA, USA) was introduced into thePV of originating focal firing producing AF.

Canine PV Preparation

IACUC approval (OUHSC and DVAMC) was obtainedfor the animal portions of the study. Male dogs (N = 32)were anesthetized with intravenous sodium pentobarbital.The heart was excised and placed into Tyrode solution con-taining (mM): NaCl, 130; KCl, 4.0; MgCl2, 1.0; NaHCO3,20; NaH2PO4, 1.0; glucose, 5.5; and CaCl2, 1.35, bubbledwith 95% oxygen: 5% CO2 (pH 7.40–7.45) at 36–37◦. Prepa-rations containing the PV antrum, the visible PV myocardialsleeve, and 3–5 mm of the PV distal to the visible sleeve wereexcised and dissected free of residual adipose and visceral tis-sues. The preparation was cut lengthwise, opened, and pinnedfor superfusion of the endocardial surface (20 mL/min).

Up to three bipolar electrograms (0.10 mm diameterTeflon� coated silver wires, 1 mm apart) and an intracel-lular glass microelectrode (ME; 3 M KCl, 10–30 M�) wererecorded on a Gould Windograf recorder. The preparationwas paced at 2–3 × diastolic threshold using 2 ms durationstimuli.

Isometric Force

The preparation was superfused with a solution contain-ing (mM): NaCl, 135; KCl, 4.5; MgCl2, 2.5, glucose, 5.0;HEPES, 5.0; Na2HPO4, 1.0; aspartic acid, 1.0; CaCl2, 2.5,pH adjusted to 7.45 with 5 mM NaOH. The atrial end wasmounted in a clamp containing a bipolar pacing electrode and

the distal vein attached to a Grass FT-01 force transducer.Resting tension was adjusted to achieve maximal isometricforce, recorded with a Grass polygraph.

Statistics

Data are expressed as the mean ± the standard error of themean. Differences between groups were determined usingStudent’s t-test for paired or unpaired data or one-way anal-ysis of variance followed by Neuman–Keuls test, as appro-priate. Differences in arrhythmia incidence were determinedusing chi-square analysis.

Results

Initiation Patterns for AF in Patients

Fifteen consecutive patients (13 male and two female,mean age = 55 ± 3 years) with documented paroxysmalor persistent AF demonstrating recurrent spontaneous fir-ing from the PVs were evaluated. Spontaneous PV firing(162 episodes lasting >3 seconds) at a 127 ± 6 ms cyclelength (566 ± 16 bpm) gave rise to 34 episodes of AF witha right atrial rate of 412 ± 17 bpm and a ventricular rateof 109 ± 6 bpm. Spontaneous PV firing giving rise to AF(580 ± 15 bpm) was significantly faster than nonsustained PVfiring (502 ± 23 bpm; P < 0.008) with a shorter PV–PV cou-pling interval (141 ± 13 ms) than for nonsustained PV firing(153 ± 12 ms; P = 0.002).

A clear tachycardia-pause pattern of initiation (the mean ofthe preceding three atrial cycle lengths less than half the pausecycle length) was observed in 102 of 162 (63%) episodes ofrapid PV firing and in 17 of 34 (50%) episodes of PV firinggiving rise to AF. The atrial cycle length (pause) initiatingPV firing was 1,039 ± 86 ms following an atrial cycle lengthof 420 ± 40 ms. The coupling interval of the initial PV de-polarization was 143 ± 7 ms and originated within the rightsuperior (52%), left superior (36%), or left inferior (12%) PV(LIPV).

The remaining episodes of rapid PV firing (60/162, 37%)were initiated during sinus rhythm at a mean cycle length of619 ± 67 ms. Although the atrial cycle length of the beat ini-tiating rapid PV firing (715 ± 32 ms) was significantly longerthan the mean atrial cycle length of the 10 sinus beats preced-ing PV firing (P=0.02), the rhythm preceding arrhythmia ini-tiation failed to demonstrate the prominent tachycardia-pauserelationship of the atrial rhythm described in the precedingtwo paragraphs. The coupling interval for the spontaneousPV beats initiating AF during sinus rhythm was not differ-ent (141 ± 13 ms) from the coupling interval of an initialbeat of PV firing initiating AF following a post-tachycardiapause (143 ± 7 ms). The mean cycle length for PV firinginitiated during sinus rhythm (114 ± 10 ms; 528 ± 33 bpm)was not different from the cycle length of PV firing rate ob-served following spontaneous tachycardia-pause initiated PVfiring (103 ± 11 ms; 584 ± 17 bpm; P = 0.10). Eight pa-tients demonstrated only tachycardia-pause-dependent firing,three patients demonstrated PV firing only during a relativelyconstant sinus rate, and four patients demonstrated both pat-terns of arrhythmia initiation. In addition to occurring afterspontaneous sinus tachycardia-pause internals in 12 patients,PV firing was observed following termination of rapid atrialpacing in all 15 patients.

Patterson et al. Triggered Firing in Pulmonary Veins 1069

An example of tachycardia-pause-induced PV firing ob-served following the termination of rapid atrial pacing isshown in Figure 1. Right atrial pacing was performed at acycle length of 575 ms for 16 beats. A pause of 1,380 ms fol-lows the cessation of atrial pacing, ending with a sinus beat.The sinus beat initiates rapid firing from the right superiorPV (RSPV; <90 ms cycle length; > 700 bpm) subsequent toa short (130 ms) coupling interval. AF is then initiated by ashort period of rapid RSPV firing.

Recordings Utilizing a Multipolar Basket Catheter in Man

An additional 68 episodes of PV tachycardia giving rise to18 episodes of AF were recorded from three patients, usinga multielectrode basket catheter (32 bipolar recording sites)(Fig. 2). No electrical activation within the PV sleeve wasrecorded at any of the 32 bipolar recording sites during the110 ± 23 ms period linking the sinus beat following a sinuspause and the initial spontaneous PV depolarization givingrise to rapid PV firing. PV activation for the first ectopic beatspread from the earliest site of activation deep within the PVsleeve, at least 12 mm from the PV–left atrium junction asestimated by angiography and electrode position. The earli-est activation of ectopic beats arose from the same site forsuccessive beats of nonsustained firing/AF and for at least theinitial beats of more sustained firing, giving rise to prolongedAF.

Examples of recordings from the multielectrode basketcatheter are shown in Figure 2 for a nonsustained PV fir-ing/AF episodes (left hand panel) and for an episode of PVfiring initiating AF (right hand panel). The bottom panel ofFigure 2 shows a three-lead surface electrocardiogram (ECG)demonstrating the spontaneous tachycardia-pause rhythmgiving rise to the nonsustained PV firing shown in the upperleft hand panel. Both episodes (nonsustained and sustained

Figure 1. Initiation of PV firing following cessation of atrial pacing. Electrical recordings are shown from a patient undergoing a clinical electrophysiologicstudy. Lead I, II, III, and V1 EKG recordings are shown. A Lasso� recording electrode with nine bipolar electrode recording sites was placed into the rightsuperior PV (RSPV). Activation in the region of the His bundle (HB) was recorded. The right atrial appendage (RAA) was used as a site for atrial stimulationand as a bipolar recording site. Proximal and distal recordings were obtained from the left inferior PV (LIPV) and the coronary sinus (CS). Right atrialpacing was performed at a cycle length of 575 ms for 16 beats. Note the separate early atrial (A) and late pulmonary vein (PV) potentials in the RSPV andLIPV recordings. An atrial pause of 1,380 ms follows the cessation of atrial pacing. Firing from the RSPV, initiated following a short (130 ms) couplinginterval at a rate >700 bpm, is observed to initiate AF.

PV firing/AF) originated following a prolonged <1,200 mspause. In each example, activation of the initial ectopic beatoriginated from 1–2 cm deep within the PV sleeve, at least 70ms after relatively rapid and uniform activation of the sinus oratrial beat had ceased within the PV sleeve. Earliest activationfor each beat shown in Figures 3 and 4 was observed from thesame site or an immediately adjacent bipolar recording site ata cycle length of 133 ± 12 ms (453 ± 45 bpm). Focal firingwas always initiated at least 1–2 cm inside the PV sleeve,as identified by the recording site of the multipolar electrode(electrodes 1–2 or 3–4) and confirmed by angiography. Thelocalized focal firing driving the initial rhythm giving riseto AF as well as slow conduction and block within the PVscan be observed in the activation maps shown in Figures 3(nonsustained PV firing/AF) and 4 (PV firing giving rise tosustained AF). Slow conduction and block arising from thefocal firing were consistently observed to give rise to slowconduction, localized conduction block, and reentry from theleft atrium into the PV, but were not responsible for the sus-tained firing arising from the PV.

Isolated Superfused Canine PV Preparation

In the presence of both acetylcholine (10−7 M) and nore-pinephrine (3.2 × 10−8 M), pacing at cycle length of 143–200 ms, followed by a prolonged pause produces protractedepisodes of rapid firing within isolated superfused canine PVsleeve preparations (Fig. 5A). The arrhythmia incidence in-creases with (1) both the rate and the duration of a pacingtrain and (2) the pause duration following the pacing train(Table 1). Pacing trains at cycle lengths longer than 200 ms,train durations less than 2 seconds, and pacing pauses lessthan 1 second fail to produce triggered arrhythmia.

Local autonomic nerve stimulation was also used to fa-cilitate tachycardia-pause induction of triggered arrhythmia

1070 Journal of Cardiovascular Electrophysiology Vol. 18, No. 10, October 2007

Figure 2. Initiation of nonsustained and sustained PV firing in humans—recordings using a 32 bipolar electrode basket catheter—Constellation�. In theupper left hand panel (A) is a recording of spontaneous nonsustained PV firing beginning with the first post-pause sinus beat. Eight splines containing fourbipolar electrodes each were inserted into the PV. The distal bipolar electrodes are labeled 1-2 for splines A–E. Moving proximally along each spline arebipolar electrode recording sites 3-4, 56, and 78. The same recordings are shown for sustained PV firing observed 90 seconds later in the same patient in theupper right hand panel (B). In each example, right firing originates from site E 1-2 (marked with asterisk). Activation maps from these recordings are shownin Figures 3 and 4. The tachycardia-pause nature of the initiation of sustained PV firing in the left hand panel can be ascertained from the three lead surfaceelectrocardiogram (ECG) recordings shown in the lower panel of the figure (C).

within isolated superfused canine PV preparations. Short du-ration stimulus trains (50 ms) of rapid (100 Hz), short duration(0.1 ms) impulses at 150 V were coupled to each pacing stim-ulus. The stimulus train was applied both during a 20-beatpacing train at a 167–200 ms basic pacing train, as well asduring the first beat following a 2- or 4-second duration pause.The stimulus train shortened action potential duration at 50%and 90% of repolarization of the post-pause beat (28 ± 5 and68 ± 14 ms post-stimulus train, respectively vs 48 ± 9 and146 ± 12 ms, pre-stimulation) (P < 0.01) and facilitated trig-gered arrhythmia formation. Triggered arrhythmia was ob-served in five of 15 preparations with local autonomic nervestimulation and was not observed in all in the absence of astimulus train (P = 0.042). An example of autonomic stimula-tion and tachycardia-pause-dependent-triggered rhythm for-mation is shown in panel B of Figure 5.

Either local autonomic nerve stimulation or combinedacetylcholine and norepinephrine administration was neededfor triggered arrhythmia formation. Triggered arrhythmiaswere not observed with tachycardia-pause pacing intervalswith acetylcholine administration alone, norepinephrine ad-ministration alone, or under control superfusion conditions.With local autonomic nerve stimulation, atropine adminis-tration (3.2 × 10−8 M) prevented action potential shortening(−3 ± 8 and −9 ± 3 ms, APD50 and APD90, respectively,P < 0.05 vs pre-drug) and triggered arrhythmia formation (0of 5 vs 5 of 5 pre-drug, P < 0.01). Atenolol (3.2 × 10−8 M)prevented EAD formation and prevented triggered arrhyth-mia formation (0 of 5 vs 5 of 5 pre-drug, P < 0.01) withoutpreventing action potential shortening with local autonomicnerve stimulation.

Ryanodine pre-treatment (10 µM for 10 minutes)prevented the initiation of focal firing associated withtachycardia-pause pacing in the presence of both acetyl-choline (10−7 M) and norepinephrine (3.2 × 10−8 M; 0 of5 vs 5 of 5 pre-ryanodine treatment; P < 0.02). A transientsuppression of tachycardia-pause arrhythmia induction wasalso observed with a transient increase in Ca+2

o from 1.35 to5 mM (0 of 5 vs 5 of 5 pre-treatment; P < 0.01).

Tachycardia-Pause-Dependent Early AfterdepolarizationFormation and Increased Isometric Force Development inIsolated Superfused Canine Pulmonary Veins

Under control conditions, tachycardia-pause pacing si-multaneously increased both isometric force developmentand EAD formation within the isolated superfused caninePV sleeve preparation. Thus, the same variables facilitatingarrhythmia formation: (1) an increase in the pacing train ratepreceding a pause, (2) an increase in the duration of the pac-ing pause following a pacing train, and (3) an increase in theduration of the pacing train preceding a pause, increased iso-metric force and facilitated EAD formation (Fig. 6). In thepresent study, EAD formation is quantified as the durationof the action potential at 90% of repolarization, because theterminal PV action potential is frequently slurred and maylack an inflection point clearly defining the beginning of anEAD.

An example of increased EAD formation is shown inFigure 5C. The ME and isometric force recordings were ob-tained in separate experiments and superimposed. At a pacedcycle length of 1,000 ms (1.0 Hz), APD90 is 138 ms with

Patterson et al. Triggered Firing in Pulmonary Veins 1071

Figure 3. Activation maps during a sinus beat and for three subsequent beats arising from the human PV (from left hand panel of Figure 2). In the upperleft hand panel, rapid and homogeneous activation of the LSPV vein from the PV ostium to the deep vein is shown for an atrial beat of sinus origin. Theapproximate anatomic dimension of the map is 3 × 4 cm2. The vein is opened longitudinally, with the upper and lower edges contiguous in vivo. Timing forthe isochronous lines of the sinus beat is based from the earliest recorded atrial activation on the surface ECG. Timing for the three subsequent ectopic beatsis continuous, beginning with the end of PV activation for the sinus beat. Activation for the first ectopic beat (upper right hand panel; 135 ms) originates fromsite E 1-2, deep within the PV, 140 ms after the latest activation observed within the PV for the previous sinus beat. Slow conduction is observed away fromthe site of earliest activation with local area of conduction block. Localized reentry from the left atrium is observed at sites A 7-8, B 7-8, C 7-8, and H 7-8following initial conduction block. Activation at sites F 5-6, G 7-8, A 7-8, B 7-8, and C 7-8 occur after initiation of the second ectopic beat at E 1-2 (lower lefthand panel). The stippled areas represent complete conduction block for a given beat. Earliest activation for beat 2 (238 ms; lower left hand panel) is againobserved at E 1-2, with conduction out of the vein along a narrow channel followed by delayed reentry from the left atrium at B-G 7-8. Reentrant activationfrom the left atrium from beat 2 occurs after initiation of beat 3 (345 ms; lower right hand panel) from the same site as beats 1 and 2. Activation from beat 3exits the vein along a narrow channel (A) and reenters the PV along B-G 7-8.

a contractile force of 170 dynes (not shown). Following a1-second pause subsequent to a 100 beat pacing train at200 ms cycle length, APD90 prolongs to 222 ms with a con-tractile force of 215 dynes.

Compared with control determinations, norepinephrine(3.2 × 10−8 M) further increases isometric force and APD90

TABLE 1

Arrhythmia Incidence in Isolated Superfused Canine PV Sleeves—10−7 M Acetylcholine + 3.2 × 10−8 M Norepinephrine—Effects of Pacing Rate, PacingTrain Duration, and Pause Duration

Effect of pacing rate (4-second train duration; 2-second pause)

N = 12 240 bpm 300 bpm 360 bpm 420 bpmIncidence of triggered arrhythmia 0/12 2/12 5/12 7/12P vs 4.0 Hz — P = NS P = 0.04 P = 0.005

Effect of pacing train duration (6.0 Hz; 2-second pause)

N = 12 1 second 2 second 4 secondIncidence of triggered arrhythmia 0/12 4/12 5/12P vs 1-second train duration — P = NS P = 0.04

Effect of pause duration (6.0 Hz; 2-second duration train)

N = 12 500 ms 1 second 2 second 4 secondIncidence of triggered arrhythmia 0/12 1/12 4/12 7/12P vs 500 ms pause duration — P = NS P = NS P = 0.005

prolongation associated with (1) an increase in the pacingtrain rate preceding a pacing pause, (2) an increase in theduration of the pacing pause following a pacing train, and(3) an increase in the duration of the pacing train preceding apacing pause (Fig. 6). The addition of acetylcholine (10−7 M)significantly shortened APD90 at a basic cycle length of

1072 Journal of Cardiovascular Electrophysiology Vol. 18, No. 10, October 2007

Figure 4. Activation maps of the initial 4 beats of sustained firing from the human PV (from Fig. 2, right hand panel). Activation of an initial ectopic beatfollowing a sinus beat is shown in the upper left hand panel (E 1-2). Beat 2 (upper right hand panel) arises from the same site in the deep PV, exits the PV tothe left atrium (F-H 7-8), and reenters A-D 7-8. Beat 3 arises again from E 1-2 and conducts to the left atrium H 7-8 and A-C 7-8 with conduction block atE-F 5-6 and D-G 7-8 (lower left hand panel). Beat 4 still originates at E 1-2, conducting out of the vein around the same area of block as observed in beat 3,but reentering the site of block from the left atrium producing an area of 2:1 conduction block (beat 4).

1,000 ms (128 ± 12 to –70 ± 7 ms, P < 0.001). The increasein isometric force development observed with norepinephrineadministration is incompletely reversed and EAD formationis not prevented by acetylcholine. The same parameters in-crease both isometric force development and EAD formationunder control conditions: (1) an increase in the pacing trainrate preceding a pacing pause, (2) an increase in the dura-tion of the pacing pause following a pacing train, and (3) anincrease in the duration of the pacing train preceding a pac-ing pause, also increase EAD formation in the presence ofnorepinephrine + acetylcholine (Fig. 6). As described previ-ously, in the presence of norepinephrine + acetylcholine, asufficiently rapid and prolonged pacing train, followed by apacing pause exceeding 1 second, elicited triggered arrhyth-mia (1,172 ± 134 bpm; 51 ± 8 ms cycle length) followingan abbreviated coupling interval (69 ± 12 ms) within 13 of18 preparations, 72%.

EAD Formation—A Role for the CalciumTransient—Effect of Ryanodine Administration

Ryanodine (10 µM for 10 minutes before the reintroduc-tion of normal Tyrode solution) abolished isometric forcedevelopment in the superfused canine PV preparation andsuppressed pause-dependent APD90 prolongation associatedwith tachycardia-pause pacing (Fig. 5). Baseline APD90 dur-ing pacing at 1.0 Hz was not significantly altered by ryanodineadministration.

Discussion

Proposed Mechanism for Rapid Focal Firing

In isolated PV tissue preparations, calcium transient trig-gering has been identified as a mechanism generating veryrapid discharges under conditions in which repolarization isabbreviated by activation of IkAch and the calcium transient isaugmented by β-adrenergic receptor stimulation. EAD for-mation and triggering are observed in the latter phase of anabbreviated action potential.7,8 It is proposed that inwardINCX is activated by increased [Ca2+]i coexistent with anaugmented and/or delayed calcium transient observed at amembrane voltage negative to the INCX equilibrium voltage,a direct consequence of accelerated repolarization. EAD andtriggered firing can be suppressed by a transient elevation in[Ca2+]o, reducing forward INCX, and by ryanodine, eliminat-ing the calcium transient.7,8 Calcium transient triggering hasalso been proposed as a mechanism generating some singleatrial beats during AF11 and for the reinitiation of AF by asingle atrial beat observed subsequent to a prolonged pausefollowing termination of AF.12

In the present experiments, we use isometric force devel-opment as a surrogate marker of [Ca2+]i. Although isometricforce development is an indirect marker and temporally lagsbehind [Ca2+]i, the measurement of isometric force providesimportant information concerning [Ca2+]i and is obtainedwithout the use of agents binding internal calcium. By bind-ing [Ca2+]i, the interventions used to directly measure the

Patterson et al. Triggered Firing in Pulmonary Veins 1073

Figure 5. Initiation of triggered firing within the isolated superfused canine right superior PV (RSPV) preparation. In the upper panel (A), a 20-beat pacingtrain at 6 Hz is followed by a 2-second pacing pause in the presence of norepinephrine (NE; 3.2 × 10−8 M) + acetylcholine (Ach; 10−7 M). Bipolar electroderecordings are shown from the deep vein (VEIN) and from the PV sleeve (SLEEVE). The microelectrode (ME) recording is from the RSPV sleeve. Rapid firing(1,220 bpm; 49 ms cycle length) is initiated from the PV sleeve from a stimulated beat following the pacing pause. In the lower panel (B), the last two beatsare shown from a 10-beat pacing train at a cycle length of 180 ms (CONTROL; upper tracings of panel B). Both a ME and bipolar electrode recording fromthe PV sleeve are shown. Following a 2-second pause, the stimulated beat demonstrates early afterdepolarization (EAD) formation. In the lower panel of B,the same pacing train is shown during stimulation with a 50 ms duration train of 0.1 ms duration stimuli at 100 Hz and 150 V, timed to the pacing stimulus.Action potential duration is shortened, but two triggered beats are observed subsequent to a single stimulated beat following a 2-second duration pacingpause. In the right hand panel (C), a superimposed isometric force development and a ME recording from a PV sleeve are shown under control conditions(CONTROL) and during ryanodine administration (RYANODINE). Ryanodine pretreatment suppressed both isometric force development and EAD formationfollowing a 1-second pause subsequent to a 10-beat pacing train at 5.0 Hz.

calcium transient are thereby capable of altering [Ca2+]i, thevery parameter they are utilized to measure.

The EAD occurring in the setting of rapid repolarizationhave been termed “late phase EAD”12 to distinguish themfrom EAD generated in the condition of delayed repolariza-tion. EAD generation with prolonged repolarization is notdependent upon an initial enhanced calcium transient, but re-quires reactivation of ICaL during a plateau phase13 or a sec-ondary spontaneous calcium release from the sarcoplasmicreticulum during phase 3 of the action potential.14 Becausethe calcium transient interrupts rapid repolarization in an ab-breviated action potential, the coupling of the first triggeredbeat in this mechanism is brief and the rapid firing is ob-served.7,8 The short coupling intervals reported during focalfiring for spontaneous ectopic beats originating within thePVs in man, as well as the very rapid firing rates of foci orig-inating within the PVs fit the mechanism of calcium transienttriggering.

We have previously reported that rapid pacing trains, fol-lowed by a pause, augment isometric force development andEAD formation,8 without examining the role of pacing rate,pacing train duration, and pause duration upon the incidence

of triggered arrhythmia. The observations may be directlyrelated to a feature termed rest potentiation, in which the cal-cium transient and isometric force development in atrial my-ocardium are augmented following a prolonged pause.15 Restpotentiation is believed to result from (1) an increased sar-coplasmic reticulum calcium content accumulated through-out the pause interval and (2) a greater fractional releaseof calcium from the sarcoplasmic reticulum. The magni-tude and temporal response of rest potentiation are species-dependent, but are significant in both canine and humanmyocardium.15

Absence of Evidence for Reentry as a Basis for Rapid PVFiring in Man

The initial spontaneous ectopic beat originating withinthe PV in man demonstrates concentric activation from alocalized focal site, as seen in the present study.16 Activa-tion conforms to the activation pattern observed with highdensity optical mapping in canine PVs during arrhythmiatriggered with tachycardia-pause pacing in the presence ofacetylcholine + norepinephrine.8 Although mapping with

1074 Journal of Cardiovascular Electrophysiology Vol. 18, No. 10, October 2007

Figure 6. The relationship between pacing rate, pause duration, and train duration for rapid pacing trains followed by a pacing pause enhancement ofcontractility and EAD FORMATION. In the upper and lower left hand panel, EAD formation (expressed as � APD90) and contractility (isometric force),respectively, are shown for different pacing train rates (10-beat trains) followed by a 2-second pacing pause for control (◦), 3.2 × 10−8 M norepinephrine(•), and 3.2 × 10−8 M norepinephrine + 10−7 M acetylcholine (�) treatment groups. The middle panels examine different pause durations for a 10-beattrain at 5.0 Hz for the same treatment groups. The right hand panels examine different train durations for a 5.0 Hz pacing train, followed by a 2-secondpacing pause for the same treatment groups. Significant differences versus control APD90 or isometric force are marked as follows: ∗P < 0.05; +P < 0.01.Although acetylcholine partially reverses the increase in isometric force and EAD formation provided by norepinephrine, EAD formation is not suppressedand arrhythmia formation is enhanced.

multipolar electrode catheters in human PVs has a muchlower spatial resolution (1 cm) than optical mapping in iso-lated canine PV tissue preparations (200 µm), and reentrycircuits with dimensions less than 1 cm could have escapeddetection in any single map, no electrical signals that couldrepresent all or part of the diastolic segment of a reentry cir-cuit linking a post-pause sinus beat to the ectopic beat wereobserved. There was a period of electrical quiescence last-ing 50–140 ms throughout the PV in each of 234 recordingsdemonstrating rapid PV firing. If one or more of the elec-trodes were located over a portion of the reentrant circuit,electrograms generated by these elements would have beenobserved in the quiescent interval. “Microreentry” with di-mensions on the order of a few millimeters or less can beproduced only with extremely low conduction velocities. Noevidence of conduction slowing or block was observed withpost-pause sinus activation of the human PV. Sinus activationwas uniform and rapid (0.2–1.0 M/sec) as shown in Figure 3(upper left hand panel).

Reentry within the PV-atrum junction can be demon-strated, instigated by rapid pacing, short-coupled extrasys-tole, or AF. Under these conditions, activation engages het-erogeneously refractory tissues causing conduction slowing

and block, essential components forming a reentrant cir-cuit.9,10 Reentrant activation within the canine and humanPVs may well be induced by calcium transient triggering andrapid focal firing. Other investigators have also used multipo-lar catheter electrodes (32 bipolar sites) (Constellation�) torecord from human PVs. Kumagai et al.5 used the multipolarrecording catheter to demonstrate complex patterns of acti-vation and block produced within PV sleeves during rapidpacing, while Arentz et al.16 utilized the catheter to map theorigin and nature of spontaneous PV ectopy initiating AF,similar to the use of the same catheter in the present studies.Although complex patterns of activation, block, and reentryfrom the left atrium can be documented to occur as a con-sequence of rapid atrial rates, short-coupled extrasystole, orrapid pacing, spontaneous and stable reentry circuits cannotbe demonstrated as a mechanism initiating rapid PV firing. Inthe present studies, localized reentry from the left atrium wasobserved as a consequence of slow conduction and localizedconduction block produced by rapid focal firing from the PV,with reactivation of the PVs terminated by the next focal beatarising within the PV sleeve. Reentry was never observedspontaneously or immediately pursuant to a prolonged sinuspause.

Patterson et al. Triggered Firing in Pulmonary Veins 1075

Other mechanisms have been proposed as an electro-physiologic basis for focal firing within the PVs. Bothdelayed afterdepolarizations8,17,18 and enhanced automatic-ity6,9 have been demonstrated in PVs and are enhanced withβ-adrenergic receptor stimulation. The rate of firing providedby these arrhythmia mechanisms has not approached the rapidrates achieved by calcium-transient triggering in canine PVpreparations (1,200 bpm) or the rapid ectopic rates observedin man (300–600 bpm).

Other Observations Concerning PV Firing in Man

In a previous study by Lu et al.19 77 patients with doc-umented frequent episodes of paroxysmal AF were exam-ined. Spontaneous atrial arrhythmia was initiated from boththe superior vena cava (7%) and the PVs (93%). Two pat-terns of arrhythmia initiation were described: (1) episodespreceded by cycle length oscillation (53%) (the cycle lengthwas prolonged or shortened by 20% or more than the pre-ceding beat) and (2) episodes initiated by a single prema-ture beat preceding a relatively constant cycle length (47%).Tachycardia-pause initiation would be included within theauthors’ definition of cycle length oscillations, although theincidence of tachycardia-pause induction sequences remainsunreported. Tachycardia-pause pacing sequences, bigemi-nal rhythms, and an early extrasystole (cycle length oscil-lations) can each increase the magnitude of the calciumtransient/isometric force development,8,15 a prerequisite forcalcium-transient triggering. Our results are consistent withthese earlier observations and provide additional observa-tions consistent with triggered firing and calcium-transienttriggering as a mechanism producing rapid PV firing inman.

Conclusions

Pacing rate and pause-dependent triggering of focal ac-tivity within human PVs as a mechanism precipitating AFis supported by bipolar electrode recordings and activationmapping failing to demonstrate mid-diastolic electrical ac-tivity linking a sinus-paced beat with the subsequent ectopicbeat originating within the PV. The same tachycardia-pausepacing interventions enabling in vitro-triggered rhythm for-mation in isolated superfused canine PV preparations triggerrapid firing from PV sleeves in man, suggesting a commonarrhythmia mechanism.

References

1. Haissaguerre M, Jais P, Shah DC, Takahashi A, Hocini M, Quiniou G,Garrigue S, Le Mouroux A, Le Metayer P, Clementy J: Spontaneous ini-tiation of atrial fibrillation by ectopic beats originating in the pulmonaryveins. N Engl J Med 1998;339:659-666.

2. Hsiesh MH, Chen SA, Tai CT, Tsai CF, Prakash VS, Yu WC, Liu CC,Ding YA, Chang MS: Double multielectrode mapping catheters facili-tate radiofrequency catheter ablation of focal atrial fibrillation originat-

ing from pulmonary veins. J Cardiovasc Electrophysiol 1999;10:136-144.

3. Jais P, Hocini M, Macle L, Choi KJ, Deisenhofer I, WeerasooniyaR, Shah DC, Garrigue S, Raybaud F, Scavee C, Le Metayer P,Clementy J, Haissaguerre M: Distinctive electrophysiological proper-ties of pulmonary veins in patients with atrial fibrillation. Circulation2002;106:2479-2485.

4. Hamabe A, Okuyama Y, Miyachi Y, Zhou S, Pak H-N, KaragueuzianHS, Fishbein MC, Chen P-S: Correlation between anatomy and electri-cal activation in canine pulmonary veins. Circulation 2003;107:1550-1555.

5. Kumagai K, Ogawa M, Noguchi H, Yasuda T, Nakashima H, SakuK: Electrophysiologic properties of pulmonary veins assessed usinga multielectrode basket catheter. J Am Coll Cardiol 2004;43:2281-2289.

6. Arora R, Verheule S, Luis S, Navarrete A, Katari V, Wilson E, Vaz D,Olgin JE: Arrhythmogenic substrate of the pulmonary veins assessedby high-resolution optical mapping. Circulation 2003;107:1816-1821.

7. Patterson E, Po SS, Scherlag BJ, Lazzara R: Triggered firing in pul-monary veins initiated by in vitro autonomic nerve stimulation. HeartRhythm 2005;2:624-631.

8. Patterson E, Lazzara R, Szabo B, Liu H, Tang D, Li YH, Scherlag BJ,Po S: Na-Ca exchange initiated by the Ca2+ transient: A trigger forarrhythmia formation in tissues with an abbreviated action potential. JAm Coll Cardiol 2006;47:1196-1206.

9. Po SS, Li Y, Tang D, Liu H, Geng N, Jackman WM, Scherlag BJ,Lazzara R, Patterson E: Rapid and stable reentry within the pulmonaryvein as a mechanism initiating paroxysmal atrial fibrillation. J Am CollCardiol 2005;45:1871-1877.

10. Kalifa J, Jalife J, Zaitsev AV, Bagwe S, Warren M, Moreno J, BerenfeldO, Nattel S: Intra-atrial pressure increases rate and organization of wavesemanating from the superior pulmonary veins during atrial fibrillation.Circulation 2003;108:665-671.

11. Zhou S, Chang C-M, Wu T-J, Miyauchi Y, Okuyama Y, Park AM,Hamabe A, Omichi C, Hayashi H, Brodsky LA, Mandel WJ, Ting C-T,Fishbein MC, Karagueuzian HS, Chen P-S: Nonreentrant focal activa-tions in pulmonary veins in canine model of sustained atrial fibrillation.Am J Physiol 2002; 283:H1244-H1252.

12. Burashnikov A, Antzelevitch C: Reinduction of atrial fibrillation im-mediately after termination of the arrhythmia is mediated by latephase 3 early afterdepolarization-induced triggered activity. Circula-tion 2003;107:2355-2360.

13. January CT, Riddle JM: Early afterdepolarizations: Mechanism ofinduction and block: A role for L-type Ca2+ current. Circ Res1989;64:977-990.

14. Szabo B, Sweidan R, Rajagopalan CV, Lazzara R: Role of Na+: Ca2+exchange current in Cs+-induced early afterdepolarizations in Purkinjefibers. J Cardiovasc Electrophysiol 1994;5:933-944.

15. Bers DM: Excitation Contraction Coupling and Cardiac ContractileForce, 2nd edition. Kluwer Academic Publishers, Natick, MA, 2001,p. 269.

16. Arentz T, Haegeli L, Sanders P, Weber R, Neumann FJ, Kalusche D,Haissaguerre M: High-density mapping of spontaneous pulmonary veinactivity initiating atrial fibrillation in humans. J Cardiovasc Electrophys-iol 2007;18:31-38.

17. Chen YJ, Chen SA, Chang MS, Lin CI: Arrhythmogenic activity of car-diac muscle in pulmonary veins of the dog: Implication for the genesisof atrial fibrillation. Cardiovasc Res 2000;48:265-273.

18. Chen Y-J, Chen S, Chen Y-C, Yeh HI, Chan P, Chang MS, Lin CI:Effects of rapid atrial pacing on the arrhythmogenic activity of singlecardiomyocytes from pulmonary veins: Implication in initiation of atrialfibrillation. Circulation 2001;104:2849-2854.

19. Lu TM, Tai CT, Hsieh MH, Tsai CF, Lin YK, Yu WC, Tsao HM, LeeSH, Ding YA, Chang MS, Chen SA: Electrophysiologic characteristicsin initiation of paroxysmal atrial fibrillation from a focal area. J AmColl Cardiol 2001;37:1658-1664.