259 Intra-Atrial Reentrant Tachycardia After Palliation of Congenital Heart Disease: Characterization of Multiple Macroreentrant Circuits Using Fluoroscopically Based Three-Dimensional Endocardial Mapping JOHN K. TRIEDMAN, M.D., KATHY J. JENKINS. M.D.. STEVEN D. COLAN. M.D.. J. PHILIP SAUL. M.D.. and EDWARD P. WALSH. M.D. From the Department of Cardiology. Children's Hospital, and the Department of Pediatrics, Harvard Medical School. Boston. Massachusetls Catheter Mapping of IART. introduction: The anatomic substrate of intra-atrial reen- trant tachycardiu (lART) following congenital heart surgery is poorly understood, but is pre- sumed to be different than common atrial flutter. Methods and Results: To study the mechanisms of IART, we used a new technique for high- densitv endocardial mapping using recordings from a multipolar basket recording catheter (25 bipolar pairs). For each recording, biplane Huorograpbic reference points were digitized to ob- tain tbe spatial locations of electrode pairs, and activation times were calculated using temporal reference points from the surface ECG. Using custom software, data were combined to create three-dimensional atrial activation sequence maps, wbich were displayed as animated se- quences. Using this technique, recordings were made in induced and/or spontaneous IART in 8 patients following congenital heart surgery (5 Fontan, 2 tetralogy of Fallot repair, 1 ventricular septal defect repair), and in 3 patients witb normal intraeardiac anatomy (I witb type I atrial flutter). Ten discrete IART activation sequences were recorded; 2 patients bad 2 sequences each. IART maps were constructed using a median of 108 electrode positions (range 27 to 197) from a median of 6 recordings/sequence (range 3 to 11). Sinus or paced atrial rhythms were also recorded, and maps were created in a similar fashion. Visual analysis of activation sequences of sinus and paced rhythm were anatomically concordant with known mechanisms of atrial activa- tion. I.\RT sequences revealed diverse mechanisms; only 1 IART cireuit was similar to that as- sociated with common atrial flutter. Activation wavefront emergence from presumed zones of slow conduction, lines of conduction block, and apparent bystander activatiim were observed. Conclusions: Higb-density atrial activation sequence maps demonstrate tbat IART following congenital beart surgery utilizes diverse circuits and is distinct from common atrial flutter. Tbe tecbnique used to create these three-dimensi(mal activation sequences may improve under- standing of these complex atrial arrhythmias and assist in the development of ablative thera- pies. (J Cardiovasc Electrophysiol. Vol. 8, pp. 259-270. March 1997) atrial Jlutter, intra-afrial reentrant tachycardia, endocardial mapping This work wa.s supported in part by N!H Grant K08-HL0.1I98 to Dr. Trieilman. Presented in part at the Scientific Sessions of the American Col- lege of Cardiology. OrlaiiiJo, Florida. March 1996. Address for correspondence: John K. Triedman. M.D., Depart- mcni of Cardiology. Children's Hospital, 300 Longwood Avc. Boston. MA 02115. Fax: 617-355-7513; E-mail: triedman@car- dio.tch.harvard.cdii Manuscripl received 9 August 1996: Accepted tor publication 30 October 1996. Introduction Modem cardiac surgical techniques have im- proved outcomes tor palients with severe congen- ital heiirt disea.se so that many suivive into adult- hood. In general, the duration and quality of these patients" lives has increased, but some suffer de- bilitating and potentially lethal cardiac unhythmias as sequelae.'^ Of these iirrhythinias. intra-atrial neen- trant tachycardia (IART) is prevalent and often
Transcript
259
Intra-Atrial Reentrant Tachycardia After Palliation ofCongenital Heart Disease: Characterization of Multiple
Macroreentrant Circuits Using Fluoroscopically BasedThree-Dimensional Endocardial Mapping
JOHN K. TRIEDMAN, M.D., KATHY J. JENKINS. M.D..STEVEN D. COLAN. M.D.. J. PHILIP SAUL. M.D..
and EDWARD P. WALSH. M.D.
From the Department of Cardiology. Children's Hospital, and the Department of Pediatrics,Harvard Medical School. Boston. Massachusetls
Catheter Mapping of IART. introduction: The anatomic substrate of intra-atrial reen-trant tachycardiu (lART) following congenital heart surgery is poorly understood, but is pre-sumed to be different than common atrial flutter.
Methods and Results: To study the mechanisms of IART, we used a new technique for high-densitv endocardial mapping using recordings from a multipolar basket recording catheter (25bipolar pairs). For each recording, biplane Huorograpbic reference points were digitized to ob-tain tbe spatial locations of electrode pairs, and activation times were calculated using temporalreference points from the surface ECG. Using custom software, data were combined to createthree-dimensional atrial activation sequence maps, wbich were displayed as animated se-quences. Using this technique, recordings were made in induced and/or spontaneous IART in 8patients following congenital heart surgery (5 Fontan, 2 tetralogy of Fallot repair, 1 ventricularseptal defect repair), and in 3 patients witb normal intraeardiac anatomy (I witb type I atrialflutter). Ten discrete IART activation sequences were recorded; 2 patients bad 2 sequenceseach. IART maps were constructed using a median of 108 electrode positions (range 27 to 197)from a median of 6 recordings/sequence (range 3 to 11). Sinus or paced atrial rhythms were alsorecorded, and maps were created in a similar fashion. Visual analysis of activation sequences ofsinus and paced rhythm were anatomically concordant with known mechanisms of atrial activa-tion. I.\RT sequences revealed diverse mechanisms; only 1 IART cireuit was similar to that as-sociated with common atrial flutter. Activation wavefront emergence from presumed zones ofslow conduction, lines of conduction block, and apparent bystander activatiim were observed.
Conclusions: Higb-density atrial activation sequence maps demonstrate tbat IART followingcongenital beart surgery utilizes diverse circuits and is distinct from common atrial flutter. Tbetecbnique used to create these three-dimensi(mal activation sequences may improve under-standing of these complex atrial arrhythmias and assist in the development of ablative thera-pies. (J Cardiovasc Electrophysiol. Vol. 8, pp. 259-270. March 1997)
This work wa.s supported in part by N!H Grant K08-HL0.1I98 toDr. Trieilman.
Presented in part at the Scientific Sessions of the American Col-lege of Cardiology. OrlaiiiJo, Florida. March 1996.
Address for correspondence: John K. Triedman. M.D., Depart-mcni of Cardiology. Children's Hospital, 300 Longwood Avc.Boston. MA 02115. Fax: 617-355-7513; E-mail: [email protected]
Manuscripl received 9 August 1996: Accepted tor publication 30October 1996.
Introduction
Modem cardiac surgical techniques have im-proved outcomes tor palients with severe congen-ital heiirt disea.se so that many suivive into adult-hood. In general, the duration and quality of thesepatients" lives has increased, but some suffer de-bilitating and potentially lethal cardiac unhythmiasas sequelae.'^ Of these iirrhythinias. intra-atrial neen-trant tachycardia (IART) is prevalent and often
260 Journal of Cardiovascular Electrophysiology Vol. 8, No. 3, March 1997
refractory to standard antiarrhythmic therapies.''•'Preliminary reports indicate that radiofrequency(RF) ablation may have a role in the managementof postojjerative IART.̂ "' also referred lo as "'inci-sional" tachycardia and atypical atrial Huttei- in thesereports. However, acute and chronic success ratesare lower than for type I atrial Outter. which is usu-ally encountered in anatomically normal hearts.
Experimental intraoperative mapping and en-trainment pacing studies performed in animals andhumans have demonstrated that the electrophysio-logic mechanism underlying boih type I atrial flut-ter and lART is atrial muscle macroreenuy with anexcitable gap."'"' However, the specific elec-troanatomic arrhythmia substrates of lART in hu-mans remain poorly chiiracterized. Vaiiations in cy-cle length and ECG morphology and tbe frequentobservation of multiple distinct clinical tachycar-dias in a single patient suggest tliat these substratesare diverse."^ Current electrophysiologic techniquesoften fail to identity the electroanatomic substrateof IART with precision in this complex group ofpatients: activation sequences may not be uniquelydetemiined by standard intracardiac recording tech-niques, and tluoroscopy may yield inadequately de-tailed anatomic information. Thus, when attemptsat RF ablative therapy are unsuccessful, it is oftennot possibie to detennine whether failure was dueto inadequate mapping of the arrhythmia circuit,inadequate RF lesion pnxluction, or both.
In this repori, preliminary results of a novelmapping technique based on signals obtained witha muitielectr(xle basket catheter are presented. Thistechnique was used to explore the electroanatomicsubstrate of IART. High-density activation se-quences of both automatic and reentrant atrialrhythms were created in patients with and withoutsurgically palliated congenital heart disease, by su-
perimposition of multiple recordings of catheterposition and electrical activity. These activation se-quences were tben displayed in three dimensions,conelated with surgieai and congenital anatomy,and interpreted with respect to the known elec-tropbysiologic properties of IART.
Methods
Study Enrollment
Thirteen patients undergoing catheteri/ation forhemodynamic and/or electniphysioiogie indicationswith symptomatic supraventricular arrhythmiaswere studied. All patienis were included in trialsevaluating use of the Webster-Jenkins basket cath-eter under IDE protocols approved by the Foodand Drug Administration and the Children's Hos-pital IRB."'^ Infonned consent was obtained. Nopatient had a significant intra-atrial shunt or throm-bus detected at precatheterization echocardiogram.
Cardiac Catheterization
Catheterization was perfonned under generalanesthesia with femoral venous access. Anticoag-ulation was obtained by administration of i(X) U/kgheparin, with additional doses given to maintainactivated clotting time > 2(K) seconds. After esti-mation of right atrial dimension was made by an-giography, a 10-French Mullins sheath was placedat tbe junction of the atrium and the superior venacava and used to guide a 5()-elecirode (25 bipolarpairs) catheter (Webster-Jenkins Basket Catheter.Cordis-Webster. Baldwin Park. CA. USA) into theright atrium (Fig. 1). Endocardial contact wasimproved by expanding the basket with a coaxialpuller mechanism attached to the catheter tip.
Figure LAP and lateral views of the multielectrode basket catheter in human right atrium.
Triedman. et al. Catheter Mapping of IART 261
Multiple recordings were obtained of the pri-mary encountered rhythm, with the basket rotated,advanced, and retracted in the atrium to sampleas much of tbe endocardial surface as possible.Each recording was documented with cinefluo-rograms taken in AP and lateral views to estab-lish catheter electrode locations. Subsequently,encountered rhythms were similarly documented.These consisted of atrially paced rhythms (via atransesophageal electrode or a basket electrodepair) and/or induced atrial reentrant tachycardias.Total mapping time was limited to 60 minutes,after which the catheter was removed by relax-ing the puller mechanism and readvancing thesheath over the basket to collapse it.
Measurement of Electrogram Timing
Because the study laboratory was limited to16 channels of amplifier capacity, electrogramswere displayed and recorded in a rapid sequentialmanner from 2 spokes (10 electrode pairs) and 4surface ECG leads at a time, using a computerizedelectrophysiology signal processing and displaysystem (Prucka Engineering, Houston. TX. USA).Signals were bandpass filtered from 40 to 4(X) Hz.Atrial activation times were manually determinedfor each electixide pair bearing a signal using elec-tronic calipers alter selection of a constant fiducialpoint (e.g.. onset of P wave, pacing artifact, esoph-ageal electrognim). Total right atria! activation timeswere calculated as the duration in msec from theearliest to the latest atrial electrograms recordedfrom all electrode pairs during a given automaticatrial rhythm.
Measurement of Electrode Location
For studies in which several discrete record-ings of a stable rhythm of interest were available,reconstruction of the activation sequence in threedimensions was achieved by utilizing the spatiallocations of the electrode pairs recorded by cine-fluorography. Geometric algorithms for determi-nation of intracardiac spatial location using bi-plane fiuoroscopy have been developed exten-sively for application in radiotberapy"* and havebeen validated for the location of intracardiaccatheters by Hauer et al.,'*̂ who demonstrated anachievable accuracy of 0.5 ± 3.1 mm in relationto fixed anatomic stmctures. In this study, a sim-plification of that technique excluding calibrationto an external grid and compensation for pin-
cusbion distortion was possible, because spatiallocation was only determined relative to othermapped points, and because tbe catheter stnictureitself provided a fixed scale for calibration of APand lateral fluoroscopic views. Using both man-ual and computer-assisted (Adobe Photoshop v3.!.Optimas v5.1a) techniques, locations of individ-ual pairs and fixed radiopaque intrathoracic ref-erence points (e.g.. surgical clips and wires) wereobtained from the orthogonal views. In caseswhere electrode positions were poorly reproducedor obscured by radiopaque structures., knowledgeof catheter design was used to interpolate elec-trode pair location when possible. An 1 X Y Z ]-coordinate vector was generated for each elec-trode pair. Spatial locations of pairs from multi-ple recordings were aligned by subtraction of the[ X Y Z I coordinates of the reference markers.
Creation and Display of Activation Sequence Maps
For a given rhythm, the activation time foreach recording electrode pair was matched withits [ X Y Z ] coordinates. The resultant matrixdefined a ragged shell of points, the surface ofwhich was smoothed to a variable degree usingan algorithm described in Appendix A. Activa-tion sequences were represented as an animatedseries of time-slices of the smooUied matrix, witlithe electrode pairs most recently activated high-lighted in color. In this report, these activation se-quences are presented in series of 12 still fraines;animations of these activation sequences can beobtained using an ftp protocol (Appendix B). Alldata processing was performed using custom soft-ware (MATLAB v4.2c.l).
Statistics
Numerical data are presented as group medianand range. Comparisons of total right atrial acti-vation times were made using Wilcoxon's ranksum test.
Results
Patient Characteristics
Patient demographics, anatomic and electro-physiologic diagnoses, and prior cardiac surgicalprocedures involving atiial surgery are summarizedin Table 1. Nine of 13 patients studied had priorcardiac surgeries.
262 Journal ut'Cardiovascular Elfctrophysiologj' Vol. 8, No. 3. March 1997
TABLE IPatient Characteristics
Patient IL)Age
(years) Anatomic l>x Prior Surgery .\rrhytluniii DxK WR.B.K.R.M.G.B.B.D.C.K.D.S.S.D.L.R.H.CR.N.S.C.K.
EAT = ectopic alrial tachyL-ardia; LVtransposition of the great arteries; TOF
• left venlricle: RV = right veniricle: SVT = supravcniricular tUL-hyLardia; TGA =tetralogy of Fallot; VSD = ventricular septal detect.
Electrogram Recordings
A summary of data used to create atrial endo-cardial activation sequences is provided in Table2. A total of 155 complete recordings of basketcatheter electrograms and biplane cinetluorogramswere obtained. Tbese were used to map 10 IARTcircuits and 12 examples of sinus or atrially pacedrhythm in 8 patients with congenital heart disease,as well as 3 tnaps of sinus rhythm and I map oftype I atrial flutter in 3 patients without structuralheiui disease. In I patient with and I patient with-out structural heait disease, total right atrial acti-vation time during sinus rhythm was determined,but no rhythms were mapped due to inadequatecinefluorography. A median of 6 (range 3 to II)complete recordings were made of each mapped
rhythm, resulting in a median of 108 (range 27 to197) electrode pairs were mapped in each rhythm.
Activation Sequence Mapping in Sinus Rhythm,Atrial Flutter, and IART
Figures 2 and 3 illustrate the display techniqueused to present the activation sequence data. Re-construction of activation sequences are shownof sinus rhythm in a patient with an anatomicallynormal heart (Fig. 2A) and type I atrial flutter inthe same patient (Fig. 2B). IART in a patient withrepaired congenital heart disease is presented influoroscopic overlay in Figure 3. Figtire 4 schemat-ically depicts the identified macrorecntrant activa-tion sequetices in the patient with type I atrial flut-
TABLK 2Rlectrophysiologic Data Act|iiiroti Durinj; Mapping
PatientID
Sinus Rhvlhm Alrial Mutter and lAKT
RhvthmKAAT(msec)
No. ofRecords
No. Pts.Mapped
PatientID Rhvthm
CL(msec)
No. ofRecords
No. Pis.Mapped
K.W.K.K.R.B.M.G.B.B.D.CK DS.S.S.S.R.H.C R .N.S.C.K.
AFI. = lype I alrial nmicr: CL - c>i:lc lonyth; IART -•-= intra-airiiil rcontraiil iach>cardia: RAAT -= righl alrial iiciivationtime; SR = sinus rhythm; Pu K.W. and B.B. had detenniiiatioii ut RAAT unly. lahlc does nol include three maps madeduring atrially paced rhythm.
Triedman. et al. Catheter Mapping of IART 263
ter and 5 of the lARTs mapped from patients withrepaired congenital heart disease. In these draw-ings, the atrial endocatdial surface is shown froma posterior point of view, with the sinus venosusopened from superior vena cava to inferior venacava (IVC). Congenital and surgical mtxiificationsare drawn in their approximate kx:ations. based onprior echocardiographic description of congenitalanatomy, surgical reports, and observed activationsequences.
Mapping of atrial flutter using this techniquelevealed a mechanism consistent with prior un-derstanding of this rhythm, i.e., conduction in acounterclockwise direction involving the isthmusof atrial tissue between tricuspid annulus and IVC,as well as rapid left-to right activation of theentire posterior right atrial wall, which appearedto terminate in the lateral right atrium (Figs. 2Band 4A). Macroreentrant sequences identified forIART circuits wete diverse and different frotii thatobserved for type I atrial flutter, but appeared tobe constrained by known anatomic and/or surgi-cal boundaries. Patient N.S. (Fig. 4C) had anactivation sequence that rotated clockwise aboutan obstacle in the lateral right atrium, and ap-peai'ed to utilize a corridor of tissue between cristaterminalis and a presumed site of right atriotomyin a cephalocaudal direction. Patients R.H. (Fig.4D) and D.C. (not shown) had circuits that alsoappeared to use protected corridors of atrial ac-tivation along the lateral wall of the right atrium.Intermittent diastolic activation was observed inthis area over a long period of relative atrial elec-trical silence, again presumably between cristaterminalis and a right atiiotomy scar, and suc-cessful ablation of IART was subsequently per-formed in that area in both patients. PatientD.L. (Fig. 4E) had a circuit that revolved in theopposite direction, but that also may have utilizedslowly conducting tissue in the lateral wall of theright atriuin. This patient aiso displayed 2:1 en-trance block into the indicated area (high pos-terolateral right atrium) during tachycardia. Pa-tients S.S. (Figs. 3 and 4F) and K.D. (not shown)both showed rapid activation cephalad on the pos-terior right atrial surface and caudad on the an-terior surface, presumably making use of large,surgically created central obstacles and withoutclear evidence of a discrete area of slow and/orconcealed conduction. Patient S.S. subsequentlyunderwent cardiac transplantation for preexistinghemodynamic indications; correlation of activa-tion sequence mapping with posl-transplant ex-amination of the right atrium of patient S.S. (Fig.
5) suggests that a repaired atrial septal defect(ASD) was tbe central obstacle.
Sinus Rhythm Mapping in Patients with IART
Compared to 4 patients with nonnal intracardiacanatomy, total right atrial activation time during si-nus rhythm was prolonged in 9 patients with IARTafter congenital heiut surgery (medians: 92.5 vs 234msec, P = O.CK)55, Wilcoxon's rank sum test: Fig.6). In several patients who were status ptist Fontan,activation sequence mapping of sinus and/or atri-ally paced rhythms tevealed prolonged and cir-cuitous routes of right atrial activation, wiih evi-detice of conduction block often noted in the an-terior right atrium. In the case of patient D.C. (notshown), ati'iiU activation in sinus rhythm was nearlyidentical to that observed in IART. the only differ-ence being an apparent change in the direction ofactivatitm of the lateral right atrial wall from cra-niocaudal to caudtx:ninial. This suggests that markedprolongation of total right atrial activation time mayitself be a predisposing factor for initiation of 1ARTF.
Discussion
The current repori details our preliminary clin-ical investigations of the electroanatomic substratesof IART and atrial tlutter using a novel, higli-dcn-sity catheter mapping technique. Correlation of theobserved activation sequences with known con-genital, surgical, and pathologic anatomy demon-strate the participation of specific anatomic andsurgically created obstacles to conduction in thegenesis of diverse macrotieenuimt iurhytlimia mech-anisms. Although the isthmus between tricuspidvalve and IVC appeitred to play a role in the ar-rhythmia circuit in some patients with congenitalheart disease and postoperative IART. IART mech-anisms were otherwise dissimilar to that of type Iatrial flutter. No comtnon anatt>mic isthmus uti-hzed by all IART circuits was identified.
Examination of atrial activation in sinus rhythmand/or atriaJ pacing in several cases revealed acti-vation sequences that were prolonged and ciivuitous,demonstrating lines of conduction block and the po-tential for "dead end" activation of conidoi-s of tis-sue. This finding is concordant with prior t^jxirtsof prolonged atrial activation times in patients afterFontati operatioTi as measiiiied fivm liigh right attiumto His activation.'" In IART. the significant elec-troanatomic feiitures observed were viiriable, includingprotected zones of slow atrial conduction that weremanifest as periods of relative electrical "silence" or
264 Journal of Cardiovascular Electrophysiology Vol. 8, No. 3. March 1997
Fiyure 3.
Triedman. el at. Catheter Mapping of IART 265
CL 200 ms
E.
CL 220 ms CL 340 ms
CL 250 ms
F.
CL 375 ms J CL 230 ms
Figure 4. Schematic diagrams of 6 of 11 mapped reentrant circuits viewed from the posterior aspect of the right atrium,which has been opened from superior vena cava to inferior vena cava (IVC) behind the crista terminalis. indicated as shadedcrescent. Surgical alterations to anatomy are located based on review of operative notes and comparison with observed acti-vation sequences. Darkly shaded ureas indicate zones of slowed conduction, and dashed circles indicate surgical patches. (A)Anatomically nonnal heart with type I atrial flutter; (B) ventricular septal defect repair with IART. which is slower but simi-lar to type I atriiil flutter with re.ipect to dependence of anatomic circuit on the isthmus hetween tricuspid valve and IVC: (C)tetralogy of Fallot repair with IART appearing to utilize lower end of crista terminalis and either or both corridors defined bycrista terminalis. right am'oroniy line, and tricuspid annulus: (O) Fonian with IART that utilizes a slow corridor of protectedatrial tissue in the lateral wall of the right atrium; (El) Fontan with IART with similar mechanism operating in opposite direc-tion; 2:1 entrance block (dashed arrow) is noted into high posterolateral right atrium; (F) Fonian with IART that is definedby central electrical gap a.'i.sociated with the atrial septal defect patch, (ftp filenames: [Aj mgafl.mov. IBj ckart.mov. IC}nsarl.mov. [Dj rhart.mov. I El diart.mov. [F} ssart.tnov)
Figure 2. (A) Right atrial activation sequence of sinus rhythm (SR) in a patient with an anatomically normal heart (KHpoints/6 recordings). Each frame shows a projection of endocardial points recorded in the atrium projected in an RAO35°/cranial 10° view; the time in milliseconds relative to the start of each sequence is indicated at the top of each frame. Acti-vation proceeds from high in the pnsterolateral right atrium (-8 lo 10 msec) inferomedially toward the atrial .septum (2S to 46msec), with the perinodal region activated around H2 msec. Within euch frame, re J points have been active within prior 0 to 5msec of time indicated, and yellow points within prior 6 to 10 msec of time indicated, (ftp filename: mgsr.mov) (B) Activationsequence of atrial Jlutter in the same patient (149 polnts/7 recordings). The .same view and conventions are used as in panelA. Activation sweeps mediolaterally across the posterior right atrium and stops at the lateral right atrial border (12 to HOmsec). Subsequently, a slower activation front funnels into the isthmus between the tricuspid annulus and inferior vena cava(UI msec) and crawls up toward the perinodal region {14H to 199 msec), (ftp filename: mgafl.mov)
Figure 3. Activation sequence of/ART of cycle length 277 msec in a patient with a Fontan (149 points/7 recordings) superim-po.ted on a lateral right atriogram. Clockwise rotation of the activation wavefront is noted around a central gap. which corre-sponds to the location of an atrial septal defect patch (see Figs. 4F and 5). Within each frame, violet dots show the distribu-tion of all mapped points, red stars have been active within the prior 0 to 8 msec of time indicated, and yellow stars withinprior 9 to 16 msec of time indicated. White crosses visible at the posterior sternal surface are alignment markers, (ftp file-name: ssarl.mov)
Figure 5. Explanted heart examined after transplantationfor nonarrhythmic indications. The locations of a circularatrial septal defect patch (ASD), an atriotomy scar (*). andthe crista terminalis (CT) are indicated in the line drawingon the hottoni. The ASD patch corresptmds to the centralgap ob.sened in Figures S and 4F. with the path of the acti-vation wavefront as indicated. The relative positions of thecrista terminalis and the atriotomy scar create a narrowcorridor, which may also participate in some IART circuits(.see Fig.s. 4C and 4DI. An apparently successful RF abla-tion of a clinically distinct IART circuit (not mapped withcurrent technique) had been performed 7 months before ex-plantation (identified hy review of prior cinefluorograms).hut no lesi(ms associated with that intervention are visible.MV = mitral valve.
focal activation of the lateral right atrial wall, andnonconductive centnil obstacles with continuous elec-trical activation. This diversity of macroreentrantmechanisms of IART may affect the relative effica-cies of electrt>physiologiciUly and iinatomically con-ceived techniques for ablative therapy.
Relation of Findings to Prior Data on IART
Tlic mochunisin of human type I atrial flutter,commonly seen in adults with normal cardiacanatomy, has been characterized as a counter-
ckx'kwi.se mtation of right alrial activatit>n involvingthe myoctirdial isthmus between the tricuspid valveand the IVC.-' -- As shown in carefully mappedanimal models of atrial reentry.-'-^ this mechanismdepends on both anatomic and functional obsta-cles to conduction, and it is facilitated by slow con-duction through the tricuspid valve-IVC isthmus.
In ct)ntiasl. the anatomic substrate of IART isvariable and poorly defined, and the impoiianccof protected zones is less well understood. Al-though these arrhythmias may be encountered inpatients with anatomically normal hearts,-'' theyare usually seen in patients who have had exten-sive atrial surgery of several varieties, includingrepairs of ASD. tetralogy of Fallot. tbe Mustardand Senning atrial switch procedures, and Fontanoperations for palliation of a functionally singleventricle.'^ Multiple aberrant P wave morpholo-gies witb variably prolonged cycle lengths and iso-electiic pcricKis during alrial diastole aiv more com-monly seen than classic "flutter waves'" and sug-gest that IART mechanisms are diflVrcni ihanthat of type I atrial flutter.'"'"-**
Animal mcxlels have begun lo clarity the acuteeffects of surgery that contribute to IART. In par-ticular, the impoitancc of slow condtiction acrosssuture lines"-'' and the sulcus terminalis'--^" hasbeen demonstrated. Chronic factors thai may alsopredispose this heterogeneous patient group to de-velopment of IART include: (1) the surgical pro-cedures themselves, which may create new dis-continuities in the atrial surface; (2) focal atrialscarring and conduction block associated withsuture lines; (3) atrial fibrosis and thickenitig causedby pericardial inflammation and/or abnormal atrialwall stress; (4) abnormal atrial anatorny and size;and (5) changes in atrial refractotiness asscK'iatedwith sinus node dysfutiction and prolonged duta-tion of atrial activation.-" Given that the specificcontribution of each of these factors will vary be-tween patienls. it is likely Ihat tbe diverse electro-physiologic manifestations of IART represent bothdifferences in the electroanalomic substrates andvariable utilization of a given substrate.
Experimentally, Schoels et al." correlated acti-vation sequence maps of IART with atrial activityon tbe surface ECG. Compared to flutter wavetachycardias with no discemable istjelectric inter-val on surface ECG. tachycardias with a discreteP wave and an isoelectric interval also had longercycle lengths and spent more time in protectedzones of slow conduction. Although both types offlutter were due to circus movemenl. tbe authorsconcluded that a large central anatornic obstacle
Triedman. et al. Catheter Mapping of IART 267
ooE 200
NoCHD CHD
NoCHD o Fontan ATOF
Figure 6. Right atrial activation times during sinusrhythm in 13 patients. CHD = congenital heart disease;TOF = tetralogy ofFallot.
may reduce or obviate requirement for a discrete"slow conduction zone" by increasing the totaltransit titne of the circuit. The data in this articlesupport these mechanistns in human postoperativeIART. and. in the case of patient S.S.. surgical andanatomic boundaries that couid support both typesof IART mechanisms were identified pathologi-cally. Consideration of total right atriiU activationtime as a potential facilitator of IART additionallysuggests that the eifect of a central obstacle to con-duction can be duplicated by extensive noncon-ductive or poorly conductive boundaries.
Implications for RF Ablative Therapy of IART
Initial attempts to treat typ)e I atrial flutter byRF ablation used electrophysiologic approaches toidentify the slow conduction zone and select sitesfor ablation.^-" Because this flutter always utilizesthe IVC-tricu.spid isthmus, anatomic approachesusing linear lesions bridging that area have beensuccessful.̂ •*-'<'
Using both electrophysiologic and anatomicmapping criteria and placing either single or lin-
ear RF lesions, the acute success rates in these se-ries have tanged frotn 70% to 90%,^-i"-' but IARTrecurrence and need for reahlation or resumptionof medical thet^py have also been t̂ eported. Kalmanet al.'* have detnonstrated that succes.stui sites forablation of IART are associated wilh the phe-nomenon of concealed entrainment. but have notassessed the specificity ot this finding, in that studyand a report by Baker et al.'" in which site selec-tion was more anatomically based, a significantlimitation was the need to itifer the sj-iecitlc loca-tion atid identity of zones of conduction blixk notcaused by large, fluoroscopically evident ob.-̂ tacles.Using .standard mapping and entrainment tech-niques, it is problematic to verify either the globalactivation pattern or the anatomic boundaries of aspecific IART circuit. Thus, the ptoblein of"false positive" site selection is difficult to resolvebecause a given ablation failure may result froman error in mapping, inadequate size, width or depthof the lesion generated, or both.
In the current study, anatomic mapping of theIART circuits provides additional and qualitativelydifferent information about potential vulnerablesites for ablation than entrainment pacing. Addi-tionally, although it is based on tiuoroscopicmeasurement of spatial location, creation of inte-grated, high-density maps allows for more specificidentification of lines of conduction block that arenot exclusively related to large atiatomic features.Narrowed conduction zones were observed in thelateral right alrium. presumably between the siteof a right atriotomy and the crista terminaUs, in-terolaterally between the tricuspid annulus and apresumed atriotomy site, and in the isthmus be-tween the tricuspid valve or an ASD patch and tbeIVC. These areas are anatomically concordant wilhsites that have been previously reported as suc-cessful sites for acute termination of IART withRF ablation.^-'"
Although functional conduction block and slowconduction appear to play important roles in theonset and maintenatice of both atrial tlutter andIART, most clinical atrial reentrant circuits appearto have an excitable gap'*̂ and thus depend, at leastin part, on areas of fixed anatomic conductionblock. Electroanatoinic mechanisms ot" IART otherthan those reported here are likely to be eneoun-tered as more patients are studied. Accumulationof a comprehensive knowledge of specific IARTmorphologies and their relation to the size andlocation of barriers to conduction in the postoper-ative right atrium will define an upper limit on thenumber of unique circuits associated with IART.
268 Journal of Cardiovascular Electrophysiology Vol. 8, No. 3, March 1997
TTiis information may facilitate the developmentof anatomic approaches to IART for use in bothcatheter-based and surgical ablative techniques, ashas been the case for development of curative sur-gical therapies lor atrial fibrillation.''^'^
Current Technique and Limitations of Three-Dimensional Eluoroscopic Mapping
Quantitative ventricular activation sequence map-ping has been used as an iidjunct to surgical resec-tion of both ventricular tachyeardia^^'"^ and atrialflutter.'- Although the number of endocardial sig-nals that can be simultaneously recorded from theheart at catheterization has increased with the in-ti'cxiuction of multielectrode iurays such as basketcatheters.'^*' the potential for therapeutic clinical ap-plications of high-density endcxardial mapping datain the catheterization lab remains limited because:(1) the number of electrodes that can be deployedin the heait remains low relative to the desired den-sity of coverage; and (2) difficulties exist in the sys-tematic quantification of the spatial locations ofmapped points. Tlie mapping technique used in thisstudy addi^sses these limitations. Activation time ismapped to the location of each electrode pair de-termined in three dimensions by digital analysis ofelectrode position in orthogonal fiuoroscopic views.A high density of electRxie coverage is obtained bysuperimposition of multiple recordings obtained ina stable cardiac rhythm of interest. This is achievedby recording signals with the catheter in a numberof positions and referencing the position of eachelectrtxle pair in each recording to a measurementgrid that is stable with rcs{x;ct to the patient's tho-rax. Adviuitages of this approach include: (1) thepossibility of recording a theoretically unlimitednumber of endocardial activation points, resultingin realization of high-resolution activation .sequencemaps; (2) the intrinsic relation of these maps to theOuoroscopic data from which they were derived;and (3) the ability to analyze the activation sequencemaps in three dimensions.
Given current understanding of the mechanismsof IART, the results obtained using this mappingtechnique are credible. Additionally, demonstra-tions of activation sequences obtained in patientswith normal anatomy and well-characterized acti-vation patterns (e.g., sinus rhythm and atrial fiut-ter) are likely to further validate this technique ina qualitative manner. However, although this tech-nique is implicitly compared to surgical mappingmethodologies that have used high-density plaqueelectrodes to characterize atria! reentrant arrhyth-
mias," •"' experimental studies to validate tempo-ral concordance between this new mapping tech-nique and the "gold standard" of intraoperativeelectrode array mapping have yet to be pertbnned.Similarly, although the spatial accuracy of elec-trode pair identification is determined by fluoro-scopic resolution, the presence of cardiac and re-spiratory mechanical motion iind the use of smooth-ing algorithms to reduce the effect of the resultantspatial "noise" introduce as yet unquantified dis-tortion of the reconstructed endocardial surface.Finally, the variable number of points that may beobtained and used to make a composite arrhyth-mia map, detennined by atrial size and anatomyand the quality of endocardial electrograms thatmay be recordable, will also have an impact onthe accuracy of local determination of activationand possibly on the ability of an entire map touniquely identify a reentrant mechanism. Animalexperimental studies specifically targeted at ad-dressing these issues are in progress.
Conclusions
Diverse electroanatomic substrates of IART ap-peaj- to utilize both naturally occurring anatomicand surgically created zones of conduction block.Only one IART mechanism observed in a patientwith postoperative congenital heart disease wasanatomically similar to that observed for type Iatrial flutter, in a patient status post ventricular sep-tal defect repair. The presence of postoperativeIART was associated with prolonged right atrialactivation during sinus rhythm, which may allowfor the development of IART.
It is possible to construct useful, high-densityatrial activation sequence maps in three dimensionsby correlation of data obtained using standard elec-trophysiologic and fluoroscopic recording tech-niques. The ability to create such maps may im-prove the understanding of complex macroreen-trant iirrhythmias iind help further development ofablative therapies.
Acknowledgment: The authors would like to thank Mr. DavidWilloughby for creating the digital video files used for presenta-tion and distribution.
Appendix A
Smoothing Algorithm Used for Creation of Maps
To reduce the artifactual effect of mechanicalmotion on the reconstructed image of the atrial en-
Triedman. et ai Catheter Mapping of IART 269
docimjial surface, the foilowing algorithm for spher-ical smoothing was implemented:
(1) A zero-mean spatial array was created bysubtracting the mean | X Y Z ] coordi-nate for the entire array from each point ofthe data array I X Y Z 1;
(2) The zero-mean array was converted from[ X Y Z ] coordinates to a spherical coor-dinate system { [ p 0 ip \ coordinates);
(3) A user-specified smoothing factor indicat-ing the fractional surface area of a sphereover which smoothing should occur (0.04to 0.08) was converted into a reference arcmea.si]red in steradians;
(4) For each point, p was reset to its meanvalue for all rays intersecting the aic spec-ified by the smoothing factor and centeredon the ray defmed by 0 and .p;
(5) The smoothed [ p 9 (p J-coordinate arraywas reconverted to [ X Y Z ] coordinatesto facilitate display.
This algorithm smooths the spatial data usinga spherical template. A smoothing factor of 1.0results in all points being mapped to a perfectsphere with a radius equal to the mean radius ofall points, while a smoothing factor of 0 resultsin no smoothing.
Appendix B
Access to Animated Files
Stand-alone computer animation tiles (Macro-media Director v4.0) and QuickTime video filesof the examples presented will be available forreview by interested readers for 6 months afterthe publication date of this report. TTiey can beobtained using anonymous file transfer protocolat: ftp.tch.harvard.edu under directory —/re-search/tried m an/I ART. Filenames aie given witheach figure and in README.TXT. System re-quirements: Windows 3.x or higher and Quick-time for Windows. Recommended Pentium mi-croprocessor with 16 MB RAM.
References
!. Bink-Boelkens MT, Metizelaar KJ. Eygelaar AA: Ar-rhythmias after repair of secundum atrial seplal defect:The influence of surgical tnodificalion. Am Heart J19H8;115:629-633.
2. Rods-Hesselink J. Perlrolh MG. McGhie J, et al:Atrial arrhythmias in adults after repair of Tetralogyof Fallot: Correlations with clinical, exercise and
3. Flinn CJ. Wolff GS. Dick M: Caidiac rhylhm after theMustard operation for complete transposiiion of thegreat arteries. N Engl J Med 1984:310:1635-1642.
4. Gelatt M. Hamilton RM. McCrindle BW. el :il: Riskfactors for alrial lachydysrhythmias after Fontan opera-tion. J Am Coll Cardiol 1994:24:1735-1741.
5. Fernandes SM, Mayer JE Jr. BunieU JT. et al: Throm-bosis and thromboembolism in the adult after Fontansurgery, (Abstract) J Am Coll Cardiol I996:27:44A.
6. Weindling SN. Saul JP. Triedman JK. ct al: Recurrentintra-atriai reentrant tachycardia following congenitalheart disease surgery: The search for an optimum ihcr-apy. (Abstract) Circulation !995;92:l-765.
7. Rhodes LA, Walsh HP. Gamble WJ. et al: Benefits ;indpotential risks of atrial antitachycardia pacing after re-pair of congenital heart disease. PACE I995;l8(PtI): 1005-1016.
8. Kalman JK. Van Hare GF. Olgin JE. et al: Ablation of"incisional" reentrant atrial tachycardia complicationsurgery for congenital heart disease. Circulation1996:93:502-512.
9. Triedman JK. Jenkins KJ. Saul JP. et al: Right atrialmapping in humans using a niultielectrode basket calh-eler. (Abstract) PACE 1995;18:8OO.
10. Baker BM. Lindsay BL. Frazier DW. et al: Sites ofsuccessful radiofrequency catheter ablation of atypicalatrial flutter. (Abstract) Circulation 1995:92:1-83.
11. Boineau JP. Schuessler RB. Cain MB, et al: Activationmapping during normal atrial rhythms and atriai flutter.In Zipes DP, Jalife J. eds: Cardiac Elcctrophysioloj'v:From Celt to Bedside. WB Saunders. Philadelphia.1990. pp. 537-547.
12. Yamauchi S, Schuessler RB, Kawamoto T, el al: Useof intraoperative mapping to optimize surgical ablationofattHal flutter. Ann Thorac Surg 1993:56:337-342.
13. Boyden PA: Activation sequence during atrial flutter indogs with surgically induced right atrial enlargement:I. Observations during sustained rhyihms. Circ Re.sI988;62:596-608.
14. Frame LH. Page P, Boyden PA. et al: Circus move-meni in Ihe canine atrium during experimental atrialflutter and during reentry in vitro. Circulation 1987;76:1155-1175.
15. Stambler BS, Wood MA, Ellenbogen KA: Pharmaco-logic alterations in human Type I atrial flutter cyclelength and monophasic action potential duratit)n: Evi-dence of a fuiiy excitable gap in the reentrant circuit- JAm Coll Cardiol 1996:27:453-461.
16. Muller GL Deal BJ. Strasburger JF. et al: Electrocar-diographic features of atrial tachycardias after opera-tion for congenital heart disease. Am J Cardiol 1993;71:122-124.
17. Jenkins KJ. Walsh EP. Colan SD. ei al: Muliipolar en-docardial mapping of the right atrium during cardiaccatheierization: Description of a new technique. J AmColl Cardiol I993;22:l 105-11 K).
18. Metz CE, Fencil LE: Determination of three-dimen-
270 Journal of Cardiovascular Electrophysiology Vol. 8. No. 3. March 1997
sional structure in biplane radiography without priorknowledge of the relationship between the two views:Theory. Med Phys 1989;16:45-S1,
19. Hauer RNW, Heethaar RM, deZwart MT, ct al: Endo-cardial catheter mapping: Validation of a cineradi-ographic method for accurate localization of left ven-tricular sites. Circulation 1986;74:862.
20. Kurer CC, Tanner CS, Vetter VL: Electrophysiologicfindings after Fontan repair of functional single ventri-cle. J Am Coll Cardiol I991;I7:I74-181.
21. Olshansky B. Okumura K. Hess PG. ct iil: Demonstra-tion of an area of slow conduction in human atrial llut-ter. J Am Coll Cardiol 1990; 16:1639-164S.
22. Olgin JE, Kalman JM. Eitzpatrick AP. et al: Role ofright atrial endocardial structures as barriers to conduc-tion during human Type 1 atrial flutter: Activation andentrainment mapping guided by intracardiac echocardi-ography. Circulation 1995;92:1839-1848,
23. Frame LH. Page RL. Hoffman BE: Atrial reentryaround an anatomic barrier with a partially refractoryexcitable gap. Circ Res I986;58:495-511.
24. Isber N, Restivo M. Gough WB, et al: Circus move-ment atrial flutter in the canine sterile pericarditismodel: Cryothermal termination from the epicardialsite of the slow zone of the reentrant circuit. Circula-tion I993;87:1649-I66O.
25. Schoels W. Kuebler W. Yang H. et al: A unified func-tional/anatomic substrate for circus movement atrialflutter: Activation and refractory patterns in the canineright atrial enlargement model. J Am Coll Cardiol1993;21:73-84,
26. Haines DE. DiMarco JP: Sustained intraatrial reentranttachycardia: Clinical electrocardiographic and electro-physiologic characteristics and long-term follow-up. JAm Coll Cardioi 1990; 15:1345-1354,
27. Chen SA. Chiang CE. Yang CJ. et al: Radiofrequencycatheter ablation of sustained intra-atrial reentranttachycardia in adult patients. Circulation 1993:88:578-587.
28. Triedman JK. Saul JP. Weindling SN. et al: Radiofre-quency ablation of intraatrial reentrant tachycardia fol-lowing surgical palliation of congenital heart disease.Circulation 1995:91:707-714.
29. Rodefeld MD. Bromberg Bl, Huddleston CB, et at:Atrial flutter after lateral tunnel construction in themodified Fontan operation: A canine model. J ThoracCardiovasc Surg 1996;! 11:514-526.
30 . Cox JL. Canavan TE. Schuessler RB. et al: The surgi-cal treatment of atrial fibrillation. II. Intraoperativeelectrophysiologic mapping and description of theelectrophysiologic basis of atrial llutter and atrial fi-brillation. J Thorac Cardiovasc Surg 1991;IO1:4O6-426.
31. Schoels W, Offner B. Braehmann J. et al: Circusmovement atrial flutter in the canine sterile pericarditismodel: Relation of characteristics of the surface elec-trocardiogram and conduction properties of the reen-trant pathway. J Am Coll Cardiol 1994:2.^:799-808,
32. Touboul P. Saoudi N. Atallah G. et al: Direct entrain-ment guided ablation of the slow conduction zone(SCZ) in human type I atrial flutter. Am J CardiolI989:64:79J-82J.
33. Eeld GK. Fleck RP. Chen PS. et al: Radiofrequencycatheter ablation for the treatment of human type !atrial flutter: Identification of a critical zone in thereentrant circuit by endocardial mapping techniques.Circulation 1992:86:1233-1240.
34. Kirkorian G. Moncada E. Chevalier P. et al: Radiofre-quency ablation of atrial flutter: Efficacy of an anatom-ically guided approach. Circulation 1994:90:2804-2814.
35. Steinberg JS, Prasher S. Zelenkofske S. et al: Radiofre-quency catheter ablation of atriai flutter: Proceduralsuccess and long-term outcome. Am Heart J 1995:130:85-92.
36. Poty H. Saoudi N. Abdc! Azi?, A. ct al: Radiofrequencycatheter ahlation of type I atrial flutter: Prediction oflate success by electrophysiological criteria. Circula-tion 1995:92:1389-1392.
37. McCarthy PM, Castle LW. Maloney JD. et al: Initialexperience with the maze procedure for atriai fibrilla-tion. J Thorac Cardiovasc Surg 1993;I05:1077-1087.
38. Mickleborough LL, Usui A, Downar E, et al: Transa-trial balloon technique for activation mapping duringoperations for recurrent ventricular tachycardia, J Tho-rac Cardiovasc Surg 1990:99:227-233.
39. Ostermeyer J, Borggrefe M. Breithardt G. et al: Directoperations for the management of life-threatening isch-emic ventricular tachycardia. J Thorac CardiovascSurg 1987:94:923-928.
40. Eldar M. Eitzpatrick AP. Ohad D, et al: Percutaneousmultielectrode endocardial mapping during ventriculartachycardia in the swine model. Circulation 1996:94:1125-1130,
