Right and Left Atrial Activation DuringExternal Direct-Current CardioversionShocks Delivered for Termination of

Atrial Fibrillation in HumansAtul Prakash, MD, Sanjeev Saksena, MBBS, MD, Ryszard B. Krol, MD, PhD, and

George Philip, MS

We examined the regional electrophysiologic effects ofsuccessful and unsuccessful direct-current cardioversionshocks on different right and left atrial regions in pa-tients with sustained atrial fibrillation (AF). Patients withsustained AF undergoing external cardioversion under-went simultaneous mapping of the right and left atria.Electrogram changes after shock delivery, regional atrialactivation, and effects of shock intensity were analyzed.Twenty-two patients with sustained AF received 52shocks (mean 2.4/patient, 22 successful and 30 unsuc-cessful). The efficacy of 50, 100, 200, and 300 J was18%, 39%, 100%, and 100%, respectively. In all 22successful shocks, there was virtually simultaneous ter-mination of electrical activity in all right and left atrialregions mapped. Unsuccessful shocks resulted in a sig-nificant increase in mean atrial cycle length at lateralright atrium, superior left atrium, and proximal, mid,and distal coronary sinus (p 5 0.01), but not at the

interatrial septum (p >0.2), which often disappearedbefore the next shock. This cycle length prolongationwas accompanied by reduction in fragmented and cha-otic electrograms (p <0.03) and emergence of discreteelectrograms at all right and left atrial regions thatpersisted until the next shock. The changes in electro-gram morphology failed to alter the surface electrocar-diographic appearance of AF. There was no correlationbetween the shock intensity and the magnitude of theseeffects. We conclude that termination of AF with externalcardioversion shocks is associated with the widespreadextinction of regional atrial wave fronts. Unsuccessfulshocks are associated with a temporary slowing of atrialactivation at all regions except at the interatrial septumand emergence of organized and/or rapidly propagat-ing wave fronts. Q2001 by Excerpta Medica, Inc.

(Am J Cardiol 2001;87:1080–1088)

Both external direct-current cardioversion and cath-eter defibrillation have been used to restore sinus

rhythm in atrial fibrillation (AF), with efficacy ratesranging from 60% to 100%.1,2 The mechanism ofdefibrillation is believed to be related to depolarizationand/or hyperpolarization of a critical mass of myocar-dium at different atrial regions, preventing propaga-tion of reentrant wave fronts in AF. However, humanvalidation studies are lacking. In addition, little isknown regarding atrial regions critical to successfulcardioversion or defibrillation. Unlike mapping ofventricular defibrillation, the electrophysiologic ef-fects of cardioversion shocks on different right andleft atrial regions has, hitherto, not been studied inhumans.3 Furthermore, the impact of unsuccessfulcardioversion shocks in the atria is not known. Theseelectrophysiologic effects are especially importantwhen we attempt to understand the mechanism(s)underlying successful AF termination, and clinicallyhighly relevant to the development of effective atrial

defibrillation protocols for use with implantable andexternal defibrillation devices.4 In this prospectivestudy, we evaluated the electrophysiologic effects ofexternal direct-current cardioversion shocks in differ-ent right and left atrial regions. Specifically, we ex-amined the effect of these shocks on AF electro-graphic cycle length and atrial electrographic mor-phology in multiple right and left atrial regionssimultaneously. We also analyzed the temporal rela-tion of these regional effects to shock delivery and theeffect of increasing shock intensity.

METHODSPatient selection: Consecutive patients with sponta-

neous AF requiring electrical cardioversion for resto-ration of sinus rhythm during clinically indicated elec-trophysiologic studies were included in this study.Patients were undergoing electrophysiologic studywith biatrial mapping for evaluation of tachycardiasand their mechanisms, symptoms, associated electricaldisorders such as conduction system disease, and hadspontaneous AF. Electrical cardioversion was clini-cally indicated for termination of a symptomatic epi-sode of AF. Written informed consent was obtainedfrom the patient for the procedure. Patients with atrialflutter at the time of cardioversion were excluded fromthe study. All patients had either undergone anticoag-ulation for a minimum of 3 weeks or had had trans-

From the Arrhythmia & Pacemaker Service, Cardiovascular Institute-Atlantic Health System, Passaic; and the Electrophysiology ResearchFoundation, Millburn, New Jersey. Manuscript received August 18,2000; revised manusript received and accepted November 27,2000.

Address for reprints: Sanjeev Saksena, MD, Cardiovascular Insti-tute, Atlantic Health System (Passaic), 55 Essex Street, Suite 3-2,Millburn, New Jersey 07041. E-mail: cmenj@aol.com.

1080 ©2001 by Excerpta Medica, Inc. All rights reserved. 0002-9149/01/$–see front matterThe American Journal of Cardiology Vol. 87 May 1, 2001 PII S0002-9149(01)01465-5

esophageal echocardiography and intravenous heparintherapy before the cardioversion protocol. Patientswith atrial thrombi or significant spontaneous echocontrast on transesophageal echocardiography werenot included to undertake a longer period of antico-agulation.

Regional endocardial mapping: The methods of re-gional endocardial contact catheter mapping in ourlaboratory has been previously reported.5,6 To sum-marize, 4 to 6 multipolar catheters were positioned inthe right and left atrium. A 7Fr duodecapolar (1-mm

band electrodes with 3-mm interelectrode distance,Cordis Webster, Inc., Baldwin Park, California) cath-eter was positioned in the right atrium (Figure 1). Two6Fr decapolar catheters (2-mm electrodes with 5-mminterelectrode distance, Daig Corp., St. Paul, Minne-sota) were positioned in the coronary sinus to recordfrom the inferior left atrium, and in the left pulmonaryartery to record the superior left atrium. Three to 5close bipolar electrograms were obtained in each atrialregion. The right atrial regions mapped were the lat-eral right atrium, interatrial septum, His bundle loca-tion, and coronary sinus ostium. Left atrial recordingswere obtained epicardially via the coronary sinus andleft pulmonary artery and directly endocardially via apatent foramen ovale or retrogradely from the leftventricle using decapolar catheters. Multiple record-ings were obtained in the superior left atrium at septaland lateral locations and at the proximal, mid-, anddistal coronary sinus. Simultaneous 12-lead electro-cardiograms were available before and after the deliv-ered shock(s). Multiple bipolar recordings were ob-tained in each region and stored digitally on a CardioLab system (Prucka Engineering, Inc., Houston, Tex-as). Electrograms were amplified and filtered between30 and 100 Hz.

Cardioversion protocol: All patients were in sus-tained AF at the time of the cardioversion shockattempt. Cardioversion was performed after all thecatheters were in position and the recorded atrial ac-tivation was consistent with a diagnosis of AF. Pa-tients were anesthetized with intravenous midazolamand/or propofol. External direct-current R-wave syn-chronous shocks were delivered using a predefinedstep-up defibrillation protocol using a conventional

FIGURE 1. Fluoroscopic location of catheter electrodes duringright and left atrial mapping of cardioversion shocks. AO 5 aor-ta; CS 5 coronary sinus; HB 5 His bundle; HRA 5 high rightatrium; IAS 5 interatrial septum; LA 5 left atrium; LLPA 5 leftlower pulmonary artery; LLRA 5 low lateral right atrium;LSPV 5 left superior pulmonary vein os; LV 5 left ventricle;RA 5 right atrium.

TABLE 1 Patient Characteristics, Shock Outcome, and Arrhythmia Cycle Length by Cardiac Disease

Pt. No.Age(yrs)

AFDuration

(mo)Heart

DiseaseShocks

Delivered

50 JPreshockCL (ms)

50 JPostshockCL (ms)

100 JPreshockCL (ms)

100 JPostshockCL (ms)

200/300 JPreshockCL (ms)

200/300 JPostshockCL (ms)

11 56 1 CAD 1 252 SR - - - -21 62 2 CAD 1 176 SR - - - -8 68 6 CAD 3 181 160 181 195 178 SR4 70 1 CAD 2 157 255 161 SR - -

22 72 3 CAD 1 154 SR - - - -14 78 8 CAD 2 161 179 165 SR - -1 56 1 IDC 3 163 172 157 190 166 SR

18 32 1 O 3 148 164 152 166 150 SR17 48 3 O 3 182 222 202 224 222 SR6 51 3 O 2 199 234 213 SR - -

10 63 5 O 3 211 234 210 232 220 SR3 74 1 O 3 195 165 190 214 194 SR9 78 1 O 2 227 260 234 SR - -

20 46 3 SH 3 191 210 188 214 194 SR5 67 3 SH 2 162 171 161 SR - -

19 68 3 SH 3 172 176 178 202 182 SR7 72 3 SH 2 187 243 205 SR - -

16 76 3 SH 1 178 SR - - - -12 79 9 SH 3 154 167 161 174 164 SR15 80 4 SH 2 155 181 161 SR - -2 86 3 SH 3 185 202 174 225 182 SR

13 74 4 VHD 3 180 206 195 217 190 SR

CAD 5 coronary artery disease; CL 5 mean arrhythmia cycle length; IDC 5 idiopathic dilated cardiomyopathy; O 5 none; SH 5 systemic hypertension; SR 5

sinus rhythm; VHD 5 valvular heart disease.

ARRHYTHMIAS AND CONDUCTION DISTURBANCES/EFFECTS OF EXTERNAL DIRECT–CURRENT CARDIOVERSION 1081

damped sine wave defibrillator (Hewlett-Packard,CodeMaster XL, model M1722B). Shocks were de-livered via cutaneous patch electrodes (Hewlett-Pack-ard defibrillation pads M3501A, Andover, Massachu-setts), which were positioned anterior to the sternumat the fourth left intercostal space and posteriorly overthe spine. The initial shock energy was 50 J.5 If this

was unsuccessful a shock of 100 J was then delivered.The third shock was either 200 or 300 J (high energy).Repeat shocks were delivered within 90 to 120 sec-onds later, although this was not always constant.

Definitions and study analysis: Shock outcome wasdeemed successful if there was restoration of sinusrhythm. Successful shocks were further subdivided

FIGURE 2. Panel A, regional atrial map with surface electrocardiographic leads and intracardiac recordings illustrating bipolar re-gional local atrial electrograms before and after a successful 100-J shock. Arrows show superior and inferior or medial to lateral ac-tivation sequence. Discrete and fractionated electrograms are seen before shock delivery. Note that there is simultaneous cessation ofatrial activation at all the atrial regions mapped. No atrial region shows continuation of electrical activity with restoration of sinusrhythm, with noise or motion artifacts being present on the left lower pulmonary artery recording. Panel B, efficacy of direct-currentcardioversion shocks at increasing energy levels. Patients with induced atrial flutter as a result of a previously ineffective shock wereexcluded from this analysis. CSd 5 distal coronary sinus; CSp 5 proximal coronary sinus; LRA 5 lateral right atrium; SLAd 5 supe-rior lateral left atrium; SLAp 5 superior proximal left atrium; other abbreviations as in Figure 1.

1082 THE AMERICAN JOURNAL OF CARDIOLOGYT VOL. 87 MAY 1, 2001

into type 1 termination if sinus rhythm ensued imme-diately after the shock, or type 2 if the termination waswithin 5 seconds of shock delivery. Electrogram anal-ysis was performed for mean cycle length and mor-phology on the 10 AF cycles before and the 10 cyclesafter an unsuccessful shock delivery. Based on mor-phology, the atrial electrograms were classified asdiscrete, fractionated, or chaotic electrical activity.5–7

Measurements were obtained at paper speed of 200mm/s using the initial high-frequency deflection of theelectrogram. Chaotic electrograms were not used formeasurement. In a region with multiple recordings,the midregional bipolar recording was preferentiallyused for analysis:atrial fibrillation: defined on thebasis of 12-lead surface electrocardiography by theabsence of P waves with irregular fibrillatory waves.350 beats/min;discrete electrograms:local bipolaratrial electrograms of constant morphology and regu-lar cycle length with reproducible isoelectric intervalsbetween successive atrial electrograms;fragmentedelectrograms:prolongation of duration of local bipo-lar atrial electrogram.50 ms with or without multiplehigh-frequency deflections;chaotic electrical activity:local bipolar atrial electrogram with variable morphol-ogy with indistinct or absent isoelectric interval be-tween electrograms;stable activation sequence:sameactivation sequence between and within each atrialregion of the first and tenth AF beat after an unsuc-cessful cardioversion shock with a variable AA cyclelength and electrogram morphology; andatrial flutter:discernable flutter waves with rates from 200 to 350beats/min on the surface electrocardiogram with astable activation sequence, no AA variability in cyclelength, and stable electrogram morphology.

RESULTSPatient group: Twenty-two patients (15 men, mean

age 686 16 years) were included in the study. Sixteenpatients had structural heart disease and 6 had coro-nary artery disease; 8 patients had hypertensive heart

disease, 1 patient had valvular heartdisease, and 1 had dilated cardio-myopathy. Six patients had lone AF(Table 1). AF duration varied from1 to 6 months. Their mean left atrialdiameter was 376 8 mm, whereasthe mean left ventricular ejectionfraction was 446 12%. All patientshad persistent or permanent AF thatwas sustained at the time of shockdelivery and required clinically in-dicated cardioversion. No patientarrived in the laboratory in sinusrhythm. The mean duration of thesustained episode that was cardio-verted using this protocol was3.3 6 7.0 months.

Shock efficacy and regional map-ping of successful shocks: A total of52 shocks were delivered to 22 pa-tients (mean 2.4 shocks/patient).Twenty-two shocks were success-

ful, whereas 30 shocks were unsuccessful. Twenty ofthe 22 successful shocks resulted in type 1 termina-tion, whereas in 2 patients, type 2 termination wasobserved. In all 22 successful shocks (types 1 and 2),there was virtually simultaneous termination of elec-trical activity in all atrial regions (Figure 2A). Type 1and 2 terminations were not associated with discor-dance at the time of electrical activity in$1 atrialregion. In 4 of 22 patients (18%), 50-J shocks weresuccessful. In the remaining 18 patients, 100-J shockswere successful in 7 patients (39%). The remaining 12patients had either a 200-J (6 patients) or a 300-Jshock (6 patients) delivered, which were uniformallysuccessful in all 12 patients (100%). Figure 2B showsthe percent efficacy of the delivered intermediate en-ergy (50 and 100 J) and high-energy (200 or 300 J)shocks.

Regional mapping of unsuccessful shocks: EFFECTSON LOCAL ATRIAL ELECTROGRAM CYCLE LENGTH: Themain finding was an increase in the mean local atrialelectrogram cycle length after unsuccessful shock de-livery in all right and left atrial regions, except at theinteratrial septum (Figure 3). However, there was nosignificant difference in the magnitude of this increaseamong the different regions. Thirteen of 18 patientshad a mean cycle length increase of$10 ms after a50-J shock, and all 11 patients exposed to 100-Jshocks had this change. Six patients (46%) with aprolongation at 50 J had AF termination at 100 Jcompared with 1 of 5 patients (20%) without thisfinding. Figure 4 shows regional atrial endocardialmapping of an unsuccessful 50-J shock. Note theimmediate increase in electrogram cycle length(shown for 1 of the 10 measured cycles after theshock) at all sites except in the interatrial septum,which shows a trend toward regional organization.These effects, however, are transient with restorationof original cycle length in the later recordings. MeanAF cycle lengths for individual patients are listed inTable 1.

FIGURE 3. Mean cycle length of local bipolar atrial electrogram before (Pre-Shock) andafter (Post-Shock) delivery of unsuccessful shocks at different right and left atrial re-gions. Note the significant increase in the mean cycle length after shock delivery in allthe mapped atrial regions except the interatrial septum (IAS). DCS 5 distal coronarysinus; MCS 5 midcoronary sinus; PCS 5 proximal coronary sinus; other abbreviationsas in Figures 1 and 2.

ARRHYTHMIAS AND CONDUCTION DISTURBANCES/EFFECTS OF EXTERNAL DIRECT–CURRENT CARDIOVERSION 1083

In a few shocks, there was acceleration or a de-crease in local atrial cycle length. Acceleration in$2atrial regions was seen in 4 shocks in 4 patients. Twoof these patients subsequently had slowing with de-livery of a shock of a higher energy; 1 patient hadslowing with a shock at lower energy, whereas an-other patient had subsequent slowing at the samedelivered energy during a second AF episode. Threeof these 4 patients had a change in the regional acti-vation sequence with acceleration. Thus, unsuccessfulshocks invariably modified local atrial electrical ac-tivity for rate, sequence, or both.

The concordant and discordant behavior of localatrial electrical activity in different atrial regions wasalso analyzed. Slowing after shock delivery at$2regions without simultaneous acceleration at any otherregion was seen with 20 shocks (67%) in 13 patients(59%). Slowing at$2 regions, accompanied by si-multaneous acceleration at another atrial region, wasseen with 5 shocks (17%) in 5 patients (27%). The siteof acceleration was the interatrial septum in 4 patientsand distal coronary sinus in 1 patient. Figure 5 is atypical example of intracardiac recordings during de-livery of a 100-J unsuccessful shock in such a patient.The shock was unsuccessful in restoring sinus rhythm,but there was an increase in the local atrial electro-gram cycle length at all right and left atrial regions.

The duration of unsuccessful shock effects on localatrial electrogram cycle length was also assessed. Inpatients with$2 unsuccessful shocks during the pro-tocol, the mean cycle length before the first shock wascompared with the mean cycle length before shockdelivery for the next 2 successive shocks. The pre-shock cycle length for the first shock was similar tothe parameter for the second shock, and was also notsignificantly different from the parameter for the thirdshock (Figure 6). Mean time interval between the firstand second shock and between the second and thirdshock was 1506 27 and 1026 43 seconds, respec-tively.

EFFECTS OF UNSUCCESSFUL SHOCKS ON LOCAL ATRIALELECTROGRAM MORPHOLOGY:Our main finding was ashift toward discrete electrogram morphologies fromfractionated or chaotic electrograms in virtually allatrial regions mapped. Figure 7A and 7B illustrate thedifference in local atrial electrogram morphologiesbefore and after unsuccessful shock delivery at differ-ent right and left atrial regions. Note there is anincrease in discrete atrial electrogram morphology,with a simultaneous reduction in chaotic and fraction-ated atrial electrogram activity in many right and leftatrial regions. Furthermore, unlike the effect of slow-ing on atrial electrograms, this effect on the electro-gram morphologies was also seen at the interatrialseptum.

FIGURE 4. Regional right and left atrial mapping during delivery of a 50-J external direct-current cardioversion shock. The measuredatrial cycle after shock delivery was greater at all mapped atrial sites (except at the interatrial septum) than that before shock deliv-ery. The electrogram cycle length for 1 of the 10 measured cycles is shown just above the bipolar recording at each mapped site.Abbreviations as in Figures 1 and 2.

1084 THE AMERICAN JOURNAL OF CARDIOLOGYT VOL. 87 MAY 1, 2001

The shift to more discrete electrogram morpholo-gies was seen at$1 atrial region with a 50-J shock in14 of 18 (78%) patients. In the remaining 4 patients,there was no change at any atrial region in 2 patients,

and a change from a discrete to afractionated or chaotic electrogrammorphology in 2 patients. Withshocks of 100 J, the shift to discreteelectrogram morphology was seenin 8 of 12 patients (72%). Therewas no change in local electrogrammorphology in 2 patients, whereasin 1 patient there was a reverse shifttoward a more fractionated or cha-otic electrogram morphology. Onepatient had a shift toward discreteelectrograms in 1 atrial region and ashift toward fractionated electro-grams in another atrial region.There was no statistically signifi-cant difference between 50 and 100J with respect to effects on localatrial electrogram morphology.

The persistence of these effectson electrogram morphology wasanalyzed. Of the 14 patients dem-onstrating a shift to a more discrete

local electrogram morphology with a 50-J shock, thisdiscrete morphology persisted up to the delivery of thenext 100-J shock in 11 of 14 patients (78%). In theremaining 3 patients, the morphology reverted back to

FIGURE 5. Regional right and left atrial mapping of a 100-J synchronous external direct-current cardioversion shock. Arrows showsuperior and inferior or medial to lateral sequential bipolar recording in each region. Before shock delivery, the different atrial re-gions show varying cycle lengths and incoordinate activation sequences. There is an increase in cycle length at all the mapped re-gions and appearance of an organized atrial wave front. The magnitude of increase in cycle length is similar for the atrial regionsmapped. LLPAd 5 left lower pulmonary artery distal recording for superior left atrial activation; LLPAp 5 left lower pulmonary proxi-mal artery recording for superior left atrial activation; other abbreviations as in Figures 1 and 2.

FIGURE 6. Mean local atrial cycle length in different right and left atrial regions for 10cycles preceding cardioversion shock delivery before the first, second, and third shocksin patients receiving repetitive shocks. Note that there is no difference in the mean cy-cle length before shock delivery for successive shocks, indicating that the increase inmean atrial cycle length as a result of shock delivery did not persist until the deliveryof the next shock. Abbreviations as in Figures 1 to 3.

ARRHYTHMIAS AND CONDUCTION DISTURBANCES/EFFECTS OF EXTERNAL DIRECT–CURRENT CARDIOVERSION 1085

the fractionated or chaotic electrical activity presentbefore the delivery of the first shock. Among the 8patients who had the same finding with a 100-J shock,these changes persisted until delivery of the nextshock in 6 patients (86%). In no patient did a reversechange toward a fractionated or chaotic electrogrammorphology persist until the delivery of the nextshock, suggesting spontaneous change may operate inselected patients.

Figure 8 is an intracardiac recording during anunsuccessful 50-J cardioversion shock delivery. Thereis a marked decrease in electrogram duration in thelateral right atrium, interatrial septum, coronary sinus,

and left atrial recordings. In the cor-onary sinus, chaotic fractionatedactivity that obscured most of theisoelectric line is replaced by dis-crete but fractionated electrogramswith restoration of isoelectric peri-ods.

EFFECT ON ACTIVATION SE-QUENCE: Twenty-one of the 22 pa-tients had an unstable and variableright and left atrial activation se-quence before first cardioversionshock delivery. In 18 patients withan unsuccessful shock of 50 J, theglobal atrial activation sequencebecame stable and repetitive in 7patients (39%). In 2 of these 7 pa-tients, the arrhythmia became type1 atrial flutter. In 4 of these 7 pa-tients, the stable activation se-quence during AF persisted untilthe next shock of 100 J. The re-maining patient reverted back to anunstable activation sequence. A100-J shock resulted in a further 2patients with an unstable activationsequence reverting to a stable se-quence with the arrhythmia becom-ing sustained atrial flutter (type 1clockwise in 1 patient and type 2atrial flutter in 1 patient). This wasterminated by the next shock.

Site of initial activation after un-successful shock: The site of earliestactivation for the first AF beat afteran unsuccessful shock varied frompatient to patient. This was the in-teratrial septum in 17 shocks, cristaterminalis in 5 shocks, proximalcoronary sinus in 2 shocks, andmid- or distal coronary sinus in 3shocks (p,0.05 interatrial septumvs other sites). In patients with 2unsuccessful shocks, the site of ear-liest activation was the same in 7patients and different in 5 patients.

DISCUSSIONMechanisms of efficacy and ter-

mination of atrial fibrillation with successful shocks: The39% efficacy for the restoration of sinus rhythm with100-J cardioversion shocks (intermediate energy), in-creasing to 100% with energies$200 J is similar tothat reported in other studies.8,9 Termination of AFwith successful shocks was associated with extinctionof atrial electrical activity at virtually all the mappedsites without regional differences. This could implythat the widespread extinction of electrical wavefronts was a prerequisite for the termination of AFwith the external cardioversion energies used in thisstudy. These findings would be consistent with theconcept that a critical mass of myocardium must show

FIGURE 7. Panel A, comparison of the frequency of discrete, fractionated, and chaoticlocal bipolar atrial electrograms at different right atrial regions before and after car-dioversion shock delivery. Both the lateral right atrium (RA) and the interatrial septumshow a significant increase in discrete electrograms with a concomitant decrease infractionated electrogram activity. Chaotic recordings are also reduced after cardiover-sion shocks. Panel B, comparison of the frequency of discrete, fractionated, and chaoticelectrograms at 2 different left atrial sites before and after cardioversion shock deliv-ery. As seen in right atrial regions, there is an increase in discrete electrograms with asimultaneous decrease in fractionated and chaotic electrograms at both the midcoro-nary sinus and distal coronary sinus recordings after shock delivery.

1086 THE AMERICAN JOURNAL OF CARDIOLOGYT VOL. 87 MAY 1, 2001

wave front interruption by shock delivery for success-ful defibrillation.

Electrophysiologic effects and mechanisms of failureof unsuccessful cardioversion shocks: Unsuccessful car-dioversion shocks of 50 or 100 J usually elicitedslowing in multiple right and left atrial activationwave fronts, producing a shift toward more discreteatrial electrogram activity at almost all mapped re-gions. There may be induction of conduction delay/block or an increase in atrial refractory periods. Al-ternatively, there could be fewer activation wavefronts around the recording sites. The absence of thiseffect at the interatrial septum implies that this regionwas more impervious to the effects of anteroposteriorexternal shock vector used in the study, and may alsopotentially be a site for continuation or even reinitia-tion of AF. The slowing of the atrial cycle length by$10 ms was associated with a higher possibility of AFtermination at the next energy level (46% vs 20%, p,0.2). This may suggest increased vulnerability to AFtermination due to the previously mentioned mecha-nisms.

The conversion of chaotic electrograms to electro-grams of a more discrete variety may represent one ofseveral potential mechanisms. Unmasking of rapidlypropagating wave fronts from rotors involved in ini-tiation or sustenance of AF, e.g., focal atrial tachycar-

dia from pulmonary veins or atypical atrial flutter,may explain this observation. Conduction block indead end pathways, not essential to the sustenance ofAF, may also occur. Fewer activation wave frontsactivating larger areas in a more cohesive fashion maypotentially result in the electrograms appearing morediscrete. Our bipolar electrogram findings after unsuc-cessful shocks correspond to a change from type 4 toa type 1 or 2 AF (Wells et al11), with concordanteffects on electrogram morphology and local atrialcycle length.10

Konings et al10 suggested that discrete electro-grams imply rapidly conducting activation wavefronts, whereas split or fragmented potentials mayindicate slow conduction at pivot points and lines ofblock in the atrium. Our data could imply that unsuc-cessful shocks impact vulnerable regions of slow con-duction or wave front pivot points. These data couldalso imply that organized activation wave fronts dueto rotors may exist in AF and are unmasked at unsuc-cessful shock energies due to the disappearance ofslow conduction. Jalife et al12 suggested that orga-nized rotors exist during established AF. We reportedorganized atrial wave fronts at onset and terminationof human AF.5,6 Our data also suggest that some atrialregions such as the septum and coronary sinus areparticularly prone to fibrillatory conduction.5,6 Our

FIGURE 8. Regional right and left atrial mapping before and after delivery of a synchronous 50-J external cardioversion shock. Thereis a change in electrogram (EGM) morphology from fractionated to discrete at the lateral right atrium. At the interatrial septum (IAS),there is a reduction in local bipolar atrial electrogram duration and reduction in the number of its component deflections. At the prox-imal coronary sinus (CSp), the chaotic activity is replaced by discrete and fractionated electrograms with a clear isoelectric line be-tween each successive electrogram. The duration of the atrial electrograms is shown above each recording site. Abbreviations as inFigures 1 and 2.

ARRHYTHMIAS AND CONDUCTION DISTURBANCES/EFFECTS OF EXTERNAL DIRECT–CURRENT CARDIOVERSION 1087

data do not preclude induction of new tachycardias byshock delivery. Thus, the appearance of typical andatypical flutter in 4 patients could be explained eitherby unmasking of existing rotors or induction of newarrhythmias. However, the transformation of an unsta-ble activation sequence to a stable activation sequenceas seen in some patients in the study substantiates theview that unsuccessful shocks electrophysiologicallymodify the AF episode despite nondiscernable surfaceelectrocardiographic findings. Similar effects havebeen seen with antiarrhythmic drug use in AF and inlinear ablation in experimental studies.13,14

Study limitations: Saturation of the recording am-plifiers was seen for up to 200 ms after shock delivery.For successful shocks, some atrial sites could havebeen activated and gone unrecognized during thisperiod. However, despite this limitation, our data im-ply that .1 repetitive wave front did not occur anddoes not change the interpretation of the data. Shocksof intermediate and lower energies could have resultedin further regional differences both for successful andunsuccessful shocks.

1. Desilva RA, Graboys TTS, Podrid PJ, Lown B. Cardioversion and defibrilla-tion. Am Heart J1980;100:881–895.2. Levy S, Lauribe P, Dolla E, Kou W, Kadish AH, Calkins H, Pagannelli F,Moyal C, Bremondy M, Schork A, Shyr Y, Das S, Shea M, Gupta N, Morady F.A randomized comparison of external and internal cardioversion of chronic atrialfibrillation. Circulation 1992;86:1415–1420.

3. Saksena S, Pantopoulos D, Hussain SM, Gielchinsky I. Mechanisms ofventricular tachycardia termination and acceleration during transvenous cardio-version as determined by cardiac mapping in man.Am Heart J1987;113:1495–1506.4. Wellens HJJ, Lau C-P, Luderitz B, Akhtar M, Waldo AL, Camm AJ, Tim-mermans C, Tse H-F, Jung W, Jordaens L, Ayers G, for the METRIX Investi-gators. Atrioverter: an implantable device for the treatment of atrial fibrillation.Circulation 1998;98:1651–1656.5. Saksena S, Giorgberidze I, Mehra R, Hill M, Prakash A, Krol RB, Mathew P.Electrophysiology and endocardial mapping of induced atrial fibrillation in pa-tients with spontaneous atrial fibrillation.Am J Cardiol1999;83:187–193.6. Saksena S, Prakash A, Krol RB, Shankar A. Regional endocardial mapping ofspontaneous and induced atrial fibrillation in patients with heart disease andrefractory atrial fibrillation.Am J Cardiol1999;84:880–889.7. Prakash A, Delfaut P, Krol RB, Saksena S. Regional right and left atrialactivation patterns during single- and dual-site atrial pacing in patients with atrialfibrillation. Am J Cardiol1998;82:1197–1204.8. Lown B. Electrical reversion of cardiac arrhythmias.Br Heart J1967;29:469–489.9. Kerber RE. Energy requirements for defibrillation (abstr).Circulation 1986;74(suppl IV):IV-117.10. Konings KTS, Smeets JLRM, Penn OC, Wellens HJJ, Allessie MA. Config-uration of unipolar atrial electrograms during electrically induced atrial fibrilla-tion in humans.Circulation 1997;95:1231–1241.11. Wells JL, Karp RB, Kouchoukos NT, MacLean WAH, James TN, Waldo AL.Characterization of atrial fibrillation in man: studies following open heart surgery.Pacing Clin Electrophysiol1978;1:426–438.12. Jalife J, Berenfeld O, Skanes A, Mandapati R. Mechanisms of atrial fibril-lation: mother rotor or multiple daughter wavelets or both.J Cardiovasc Elec-trophysiol1998;9:S2–S12.13. Rensma PL, Allessie MA, Lammers WJAP, Bonke FI, Schlaij A. Length ofexcitation wave and susceptibility to reentrant atrial arrhythmias in normalconscious dogs.Circ Res1988;62:395–410.14. Oritz J, Niwano S, Abe H, Rudy Y, Johnson NJ, Waldo AL. Mapping ofconversion of atrial flutter to atrial fibrillation and atrial fibrillation to atrial flutter.Insights into mechanisms.Circ Res1994;74:882–894.

1088 THE AMERICAN JOURNAL OF CARDIOLOGYT VOL. 87 MAY 1, 2001