Development and Validation of an ECG Algorithmfor Identifying Accessory Pathway Ablation Site

in Wolff-Parkinson-White Syndrome

MAURICIO S. ARRUDA. M.D., JAMES H. McCLELLAND, M.D.,XANZHUNG WANG, M.D., KAREN J. BECKMAN. M.D.,

LAWRENCE E. WIDMAN. M.D.. PH.D., MARIO D. GONZALEZ, M.D.,HIROSHI NAKAGAWA, M.D., PH.D., RALPH LAZZARA, M.D.,

and WARREN M. JACKMAN, M.D.

From the Cardiovascular Section. Department of Medicine. University of Oklahoma Health Sciences Centerand the Department of Veterans Affairs Medical Center, Oklahoma City, Oklahoma

ECG Localization of Accessory AV Pathways. Introduction: Delta wave morphologycorrelates with the site of ventricular insertion of accessory AV pathways. Because lesions dueto radiofrequency (RF) current are small and well defined, it may allow precise localization ofaccessory pathways. The purpose of this study was to use RF catheter ahlation to develop anECG algorithm to predict accessory pathway location.

Methods and Results: An algorithm was developed by correlating a resting H-lead KCGwith the successful RF ahlation site in 135 consecutive patients with a single, anterogradelyconducting accessory pathway (Retrospective phase). This algorithm was subsequently testedprospectively in 121 consecutive patients (Prospective phase). The EC(i findings included theinitial 20 msec of the delta wave in leads I, II, aVF, and V, [classified as positive (+1, negative(-), or isoelectric (±)] and the ratio of R and S wave amplitudes in leads III and V, (classified asR > S or R < S). When tested prospectively, the ECG algorithm accurately localized the ac-cessory pathway to 1 of 10 sites around the tricuspid and mitral annuli or at suhepicardiat lo-cations within the venous system of the heart. Overall sensitivity was 90% and specificity was99%. The algorithm was particularly useful in correctly localizing anteroseptal (sensitivity75%, specificity 99%), and mid-septal (sensitivity 100%, specificity 98%) accessory pathwaysas well as pathways requiring ahlation from within ventricular venous branches or anomaliesof the coronary sinus (sensitivity 100%, specificity 100%).

Concltision: A simple ECG algorithm identifies accessory pathway ahlation site in Wolff-Parkinson-White syndrome. A truly negative delta wave in lead 11 predicts ablation within thecoronary venous system. (J Cardiovasc Electrophysiol, Vol. 9, pp. 2-12, January 1998)

Wolff-Parkinson-White syndrome, accessory pathway, electrocardiogram, radiofrequency catheterablation

Introduction

In patients with Wolff-Parkin son-White syn-drome, the morphology of the delta wave duringsinus rhythm is dependent upon the location of the

Address for correspondence: Ralph Lazzara. M.D.. Department ofMedicine. Cardiovascular Seciion, University of Oklahoma HealthSciences Center. 920 Stanton L. Young Blvd., Room 5SP3OO, Ok-lahoma City. OK 731W-3048. Fax: 405-271-2619.

Manuscript received 10 April 1997; Accepted for publication 14November 1997.

ventricular insertion of the accessory pathway,which is the site of initiation of ventricular acti-vation. A numher of investigators have corre-lated ECG patterns or algorithms for decipheringthe location of the ventricular insertion of the ac-cessoiy pathway.' - The purpose of this study wasto develop a simple, accurate algorithm to he usedin examining the 12-iead ECG during sinus rhythmto identify the location of an accessory pathwayhased upon the site of radiofrequency (RF) cath-eter ablation of accessory pathway conduction.

Arruda, et al. ECG Localization of Accessory AV Pathways 3

MethodsA

Study Population

The study population consisted of 256 consec-utive patients referred for RF catheter ablation ofa manifest accessory atiioventricukir pathway. Sub-jects with more than one anterogradely conduct-ing accessory pathway were excluded from the ret-rospective phase of this study. There were 157 menand 99 women (mean age 32 ± 15 years, range 9to 78).

Study Design

An algorithm to predict accessory pathway lo-cation was developed by conelating the preablationECG with the successful RF ablation site in 135consecutive patients with a single anterogradely con-ducting accessory pathway. The algorithm was thentested prospectively in 121 consecutive patients un-dergoing RF catheter ablation to assess its accuracyin predicting the successful ablation site. A predic-tion of the accessoiy pathway location based on tbealgorithm was made by one of the investigators(M.S.A.) prior to the electrophysiologic study. Thelocation of the ablation catheter was recorded radi-ographically in the left anterior oblique projection,at an angle in which the catheter recording the His-bundle potential points directly at the observer,and in the right anterior oblique projection, at anangle in which the catheter recording the His-bun-dle potential lays parallel to the interatrial septum.

Ahlation Site Nomenclature

Accessory pathway locations were divided intothree main regions, which were further subdividedand are illustrated in Figure IA.

(I) Septal accessory pathways were subdividedinto five regions:• Anteroseptal tricuspid annulus and right an-

teiior paraseptal (AS/RAPS), which includesaccessory pathways located up lo 10 mmanterior to the His bundle in which both theaccessory pathway and a His potential canbe recorded from the same bipolar electrode.

• Mid-.septal tricu.spid annulus (MSTA), whichincludes accessory pathways located at theseptal section of the tricuspid annulus betweenthe posteroseptal and antert^septal regions.

• Posteroseptal tiicuspid annulus (PSTA), in-cluding accessory pathways located near thecoronary sinus ostium (CSOs).

Figure 1. Schematic representation of the heart as viewedin the left anterior oblique projection. (A) Nomenclatureused to describe accessory pathway location. RA = rightanterior: RAL = right anterotaterat: /?/, = right lateral:RPL = right posterolaterat: RP = right posterior; PSTA =posteroseptal tricuspid annulus: CSOs: coronary sinus os-tium: MSTA = mid-septal tricuspid annulus: AS = an-^tero.septal: RAPS = right anterior paraseptat: MCV = mid-dle cardiac vein (coronary vein): CS = coronaiy sinus: ve-nous anomaly (coronary sinus diverticulum): PSMA ~postero.septal mitral annulus: LP = left posterior: LPL =left posterolaterat: LL = left lateral: LAL = left anterolat-ercit: HB = His bundle. (BI Acces.sory pathway locations(defined by site of successfu! catheter ablation) in the 135patients comprising the retrospective group. (C) Accessorypathway locations in the 121 patients comprising theprospective group.

4 Journal of Cardiovascular Electrophysiology Vol. 9, No. I, Jantiary 1998

• Posteroseptal mitral annulus (PSMA).• Subepicardial posteroseptal accessory path-

ways, which consist of accessory pathwaystliat nx|uired ablation fixjm within the subepi-cardial venous system (occasionally at theleft posterior region), including the middlecardiac vein and other coronary veins; or inanomalies of the coronary sinus, such as adiverticulum (Subepicardial).

(II) Right free-wall accessory pathways were sub-divided into five regions:• Right anterior (RA).• Right anterolateral (RAL).• Right lateral (RL).• Right posterolateral (RPL).• Right posterior (RP).

(Ill) Left free-wall accessory pathways were sub-divided into four regions:• Left anterolateral (LAL).• Left lateral (LL).• Left posterolateral (LPL).• Left posterior (LP).

Analysis of the ECG

A standard I2~lead resting ECG was recordedat a paper speed of 25 mm/sec, at a gain of 10mm/mV using filter band settings of 0.16 to 100Hz in each patient 1 day prior to the ablationprocedure. The six limb leads were recorded si-multaneously and the six pericardial leads wererecorded simultaneously. The onset of the deltawave in each lead was measured from the onsetof the earliest delta wave in any of the six ECGlimb leads as well as in any of the six precordialleads. In the limb leads., the onset of the delta wavewas often identified in lead I when it was positive.Otherwise, the earliest delta wave was inscribedin lead aVF.

The polarity of the delta wave was measuredwithin the initial 20 msec of the preexcitationand was classified as positive (+), negative (-),or isoelectric (±), as illustrated in Figure 2A.

Results

ECG Algorithm Development

The distribution of location of the accessory patb-way in the initial 135 patients is illustrated in Fig-ure IB. ECGs for accessoiy pathways in each re-gion were examined for distinguishing characteris-tics. No delta wave characteristics could be identifiedthat reliably differentiated between some contigu-

ous sites, including: RP and RPL; RA and RAJL;LP and LPL; and LL and LAL. For this reason,pairs of regions were grouped together in the analy-sis. An algorithni was devised tbat correctly iden-tified the acces.sory pathway location in 117 (87%)of the 135 patients. This algorithm is shown schemat-ically in Figures 3 and 4, and is described below.

Step J: If either the delta wave in lead I isnegative or isoelectric or the R wave is greater inamplitude than the S wave in lead V,, a left free-wall accessory pathway is present. If this criterionis fulfilled, lead aVF is examined. If the delta wavein lead aVF is positive, a left lateral/anterolateral(LL/LAL) accessory pathway is identified. If adelta wave in lead aVF is isoelectric or negative,the accessory pathway is located at the left poste-rior/posteroiateral (LP/LPL) region (Figs. 3 and 5).

If the criteria in leads I and V, are not ful-filled, a septal or right free-wall accessory AV path-way is identified. Proceed to step 2.

Step 2: Lead II is examined. A negative deltawave in lead II identifies the subepicardial pos-teroseptal accessory pathway (Figs. 3 and 6). If thedelta wave in lead II is isoelectric or positive, pro-ceed to step 3.

Step 3: Lead V, is examined. A negative or iso-electric delta wave in lead V, identifies a septalaccessory pathway. If this criterion is fijlfilled, leadaVF is examined. If the delta wave in lead aVF

\

R<S

Figure 2. Delta wave polarity was determined by examin-ing the initial 20 msec after earliest delta wave onset in thelimb leads as well as precordiut leuih. (A) ECG leads I, II.and 111 of a patient with an acces,sory pathway located atthe posteroseptal tricuspid annulus region. Note that it ispossible to determine delta wave polarity in all three leadsat approximately 20 msec after the onset of delta wave. (B)Determination of delta wave polarity (using the initial 20msec) in the event of changes within 40 msec.

Arruda. et al. ECG Localization of Accessory AV Pathways 5

Step 1 Left Free Wall Accessory Pathways Step 2 Subepicardial Accessory Pathways

MCVor

Venous Anomaly

Step 3 Septal Accessory Pathways

linStep 4 Right Free Wall Accessory Pathways

Figure 3. Stepwise ECG algorithm for predicting accessory pathway location. Abbreviations as tn Eigure I. See text for ex-planation.

is negative, an accessory pathway is identified,which is located at tbe posteroseptal tricuspid an-nulus or at (he coronary sinus ostium and sur-rounding region (PSTA/CSOs), If the delta wave

is isoelectric in lead aVF, the accessory pathwaymay be located close to either the po.steroseptal tri-cuspid annulus (PSTA) or the posteroseptal mi-tral annulus (PSMA) as shown in Figures 3 and 7.

-No-

Yesi

LeftFree wall

No

-No-

Subepicardial

Figure 4. Stepwise ECG algorithm for determination of accessory pathway locatton. Abbreviations as in Eigure I.

6 Journal of Cardiovascular Electrophysiology Vot. 9. No. I. January 1998

Left Anterolateral / Left Lateral

A -̂

aVL,

aVF

Left Posterolateral / Left Posterior

aVR V, A

Figure 5. Representative ECG from subjects with a teff free-watt accessory pathway. These patients are identtfied by a (-) or(±) delta wave tn tead I. or R > S tn tead V,. Delta wave polarity in lead aVf sublocates these accessory pathways.

A positive delta wave in aVE identifies a pathwaylocated within the anteroseptal/right anteriorpara.septal (AS/RAPS) or mid-septal tricuspid an-nuUis (MS) regions. These two regions are differ-entiated by examining tbe R/S ratio in lead III: R> S identifies anteroseptal/right anteriorparaseptal (AS/RAPS) accessory pathway, and R< S identifies an accessory pathway located alongtbe mid-septal tricuspid annulus (MSTA), as il-lustrated in Eigures 3 and 8.

If tbe delta wave in lead V, is positive (after hav-ing excluded patients with a left free-wall acces-sory pathway in Step 1), a right free-wall acces-sory AV pathway is identified. Proceed to step 4.

Step 4: In patients with right free-wall acces-sory pathways, examine lead aVR A positive deltawave in lead aVE identifies a righl anterior/an-terolateral accessoi-y pathway (RA/RAL). If thedelta wave in aVE is isoelectric or negative, ex-amine lead II. A positive delta wave in lead II iden-tifies a right lateral accessory pathway (RL). andan isoelectric delta wave in lead II identifies a rightposterior/posterolateral accessory pathway(RP/RPL), as illustrated in Eigures 3 and 9.

ECG Algorithm Validation

The ECG algorithm was then tested prospec-tively in 121 consecutive patients. The distribution

of accessory pathway location based upon site ofsuccessful ablation is shown in Eigure IC. The re-lationship between the predicted location (basedupon the ECG algorithm) and the actual location(based upon ablation site) is shown in Table I. Thealgorithm correctly identitied accessory pathwaylocations in 109 patients (sensitivity 90%, speci-ficity 99%. positive predictive value 93%. and neg-ative predictive value 98%), calculated as weightedaverages (weighted by number of true positives).

Discussion

Several attempts have been made to correlateelectrocardiographic findings with anatomic loca-tions of accessory AV pathways in patients withWolff-Parkinson-White syndrome.'^- ECG criteriabased upon surgical dissection of accessory path-ways have shown accuracy in identifying acces-sory pathway location.- '- The ECG algorithmdeveloped in this study differs from prior algo-rithms in its combined use of the resting (notduring atrial pacing) ECG. utilization of only fiveECG leads, localization by catheter ablation tech-niques (which may allow precise localization ofthe accessory pathway), and by prospective vali-dation of the algorithm. This algorithm is partic-ularly accurate in predicting ablation at sites nearthe AV node and His bundle with risk of AV block.

Arruda, et al. ECG Localization of Accessory AV Pathways 7

a

m

iX

XX C-< <

&sJ

— -^ r-1 ' t —

ri

such as mid-septal and anteroseptal regions, andfor predicting ablation of subepicardia! accessorypathways from the venous system., such as coro-nary veins and anomalies of the coronary sinus.This latter region has been associated witb failedablation and cardiac perforation.-' To facilitate theinterpretation of delta wave polarity, this algorithmonly uses the initial forces of preexcitation (ini-tial 20 msec) after the earlie.st delta wave onset inany of the limb leads as well as the precordial leadsto cbaracterize delta wave polaiity. This may avoidmisinterpretation when changes in polarity occurwithin 40 msec, which is not an uncommon llnd-ing (illustrated in Fig. 2B and shown in lead 11 ofFig. 7, left panel). Also, analysis of only the ini-tial 20 msec of delta wave may help to identifysome left free-wall accessory pathways producingminimal preexcitation in sinus rhythm or pathwaysexhibiting slow anterograde conduction.

ECG Criteria Based upon RF Catheter Ablation ofAccessory AV Pathways

RF catheter ablation has been used widely fortreating patients with an accessory pathway.̂ "̂̂ ^Certain ECG findings may coirelate with tbe lo-cation of accessory pathways in these patients. Webave reported previously a preliminary ECG al-gorithm for identifying posteroseptal accessorypathways'' and subsequently for identifying ac-cessoiy pathways located at 10 different sites aroundthe tricuspid and mitral annuli using only ECGleads I, II, aVF. and V,.'^ Others have used tbemaximally preexcited 12-lead ECG during sinusrhythm or atrial pacing, delta wave polarity inthe initial 40 msec of the preexcited QRS com-plexes, and frontal and horizontal delta wave axiswere used for analysis.'*' Combination of multiplevariables, such as (1) precordial transition, (2) Rand S waves amplitude ratio. (3) sum of tbe po-larities of delta wave. (4) amplitude of delta wave,(5) delta wave axis in the frontal plane, and (6)amplitude of R wave, have been used witb suc-cess,'"' but the analysis of this complex set of vari-ables is yet to be validated prospectively. In twootber studies, the ECG algorithm was testedprospectively. One used the polarity and mor-phology of QRS complexes and the highest R waveamplitude in the precordial leads."* Tbey showedan 86% accuracy in locating accessory pathwaysat nine sites around the tricuspid and mitral annu-lus. Tbe specific sensitivities and specificities werenot reported and. due to the small number of pa-tients and the fact that a broad region on the right

8 Journal of Cardiovascular Electrophysiology Vol. 9, No. i, Jatiuary 1998

Coronary Sinus Diverticulum

J1

- V .

aVR

aVL

aVF

Middle Cardiac Vein

aVR

aVL

aVF

Figure 6. Representative ECG from subfects with a subepicardial accessory pathway. These patients are identified by a (~)delta wave in lead II (after excluding stibjects with a left free-wall acces.sory pathway). Note the typical pattem of the deltawave in lead II: it is negative and joins the end of the P wave.

free wall had to be combined to a single site maycompromise the resolution of their algorithm. Theother study in which the ECG algorithm was val-idated prospectively'^ used a similar design andresolution to our preliminary reported results.'^Nevertheless, our algorithm exhibited a positivepredictive accuracy of 93% in locating accessorypatliways at 10 different sites. In addition, we coulddistinguish anteroseptal from right anterior acces-sory pathways, whereas their algorithm could notdiscriminate accessory pathways located in a high-risk region for AV block {anteroseptal) from ac-cessory pathways located in a relatively low-riskregion (right anterior).

Utility of the Algorithm for Identifying AccessoryPathways Requiring Ablation from theSubepicardial Venous System

Surgery for Wolff-Parkinson-White syndromehas shown that accessory pathways may be asso-ciated with coronary sinus venous anomalies suchas aneurysms and diverticula.-**-̂ We have reportedpreviously that some accessory pathways requireablation from within the coronary venous systemin the subepicardial posteroseptal-'̂ -̂ ^ and left free-

regions. Other investigators have also usedRF energy to ablated accessory pathways from thecoronaiy sinus, middle cardiac vein, and coronarysinus diverticulum.'**""̂

Accessory pathways associated with coronaryveins and anomalies of the coronary sinus may ac-count for catheter ablation failure at the tricuspidand mitral annulus. In our institution, approximately40% of patients refen'ed following at least one un-successful ahlation attempt for a posteroseptal ac-cessory pathway required ablation from coronaryveins or anomalies of the coronary sinus. '̂' In thisstudy, a negative delta wave in ECG lead II wasuseful in predicting ablation from coronary veinsand anomalies of coronar-y sinus in all 14 patients.It is important to rule out a brief initial isoelectricsegment of the delta wave in lead II before label-ing it as negative. This initially isoelectric deltawave may be followed by a negative componentthat is often associated with an accessory pathwaylocated at the posteroseptal tricuspid annulus, andright posterior and right posterolateral regions (Fig.2A and Fig. 7, left panel). An ECG exhibiting atrue negative delta wave in lead II will also showa negative delta wave in leads III and aVF (Fig.6). In our experience, this finding is associated with

Arruda. et al. ECG Localization of Accessory AV Pathways 9

Posteroseptal Tricuspid Annulus Posteroseptal Mitral Annulus

aVR

aVL

aVF

aVR

aVL

aVF

Figure 7. Representative ECG from subjects with a septal acces.sory pathway. These patients are identified by a (-} or (±)delta wave in lead V, (after excluding subjects with a left free-wall or subepicardial accessory pathway). Delta wave polarityin lead aVF sublocates these septal accessory pathways. See text for discussion.

Anteroseptal

aVR

aVL

aVF

Figure 8. Representative ECG from .subjects with it septal acces.sory pathway. These patients are identified by a (-) or (±)delta wave in lead V, (after excluding subjects with a left free-wall or suhepicardial accessory pathway). A < + ) delta wave po-larity in lead aVF sublocates these septal acces.soiy pathways at either the midseptat or anteroseptal tricuspid annulus re-gions. The precise region is determined by the R/S ratio in lead III. See text for discussion.

MidseptalTricuspid Annulus

III

aVR ^

aVL

aVF

• V ,

10 Jourual of Cardiovascular Electrophysiology Vol. 9. No. I, January 1998

Right AnteriorRight Anterolateral

aVR

aVL

aVF

Right Lateral

-V..

,v'i A., J /•

aVR

aVL-

aVF'

Right PosteriorRight Posterolateral

aVR •^ j

aVL .J

Figure 9. Representative ECG frotn subjects with a right free-wall accessory pathway. These patients are identified by ex-cluding subjects with a lefi free-wall, subepicardial, or septal accessory pathway. See text to sublocate these accessory path-ways at the free-wall tricuspid annulus.

a posterosepta] accessory pathway having a subepi-cardial ventricular itisertion, requiring successfulablation from the coronary venou.s system. Despitethe tact that a negative delta wave in the interiorleads is generally accepted to be associated with aposteroseptal accessory pathway, the ECG frompatients requiring successful ablation froin the en-docardium may show a negative delta wave inleads III and aVE but unlikely a negative deltawave in lead II.

Utility of the Algorithm for Identifying Anteroseptaland Mid-Septal Accessory Pathways

Surgical dissection as well as catheter ablationof accessory pathways located near tbe AV nodeand His-bundle region have been associated withprocedure failure and AV block.*̂ "•-'̂ •̂ ' This ac-counts for a signiticant proportion of the morbid-ity of RF ablation procedures.-'' Mid-.septal acces-sory pathways were first described as intermedi-ate septal accessory pathways exhibiting positivedelta wave morphology in ECG leads I, II, andaVL, and isoelectric in leads III and aVF.'̂ " In thisstudy, the initial forces of the delta wave in leadaVE were positive (Eig. 8, right panel). Others baveseparated the mid-septal region into three zones

and, in agreement to our findings, a positive deltawave in lead aVE correlated witb a mid-septal ac-cessory pathway in close proximity to tbe AV con-duction system.-' Li tbis study, 3 of 4 anterosep-tal and all 5 of 5 mid-septal accessory pathwayswere correctly localized by tbe algoritbm.

Clinical Implications

As the ECG algorithm accurately localizes ac-cessory pathways prior to ablation, it may belp thephysician advi.se the patient regarding the likeli-hood of success and complications of the proce-dure, in particular subjects with anteroseptal ormid-septal accessory pathways. The ECG algo-rithm may aid .selection of patients in whom coro-nary sinus angiography should be perfomied in or-der to delineate its anatomy, thus allowing map-ping in the coronary veins and anomalous stnjcturesof the coronary sinus.

Limitations

Multiple accessory pathways have been docu-mented in 2% to 20% of subjects.^-^ In the ret-rospective phase of our study, we only includedpatients with a single anterogradely conducting ac-

Arruda, et al. ECG Localization of Accessory AV Pathways 11

cessory pathway. During the prospective pbase,there were no patients with more than one acces-sory pathway exhibiting anterograde conduction.Because of the limited experience in using this al-gorithm in subjects with multiple pathways, wecannot comment on its utility in this setting. Ac-cessory AV pathways exhibit a slant course acrossthe tnitral or tricu.spid annulus. and delta wave mor-phology is dependent upon the site of ventricularinsertion of accessory pathways. Tbis ECG algo-rithm does not necessarily identify the site of ear-liest ventricular activation {ventricular insertion).It was developed based upon tbe site of success-ful RE catheter ablation of accessory pathways,which were detemiined primarily by recordings ofaccessory pathway activation potential. Therefore,the slant course of accessoiy pathways, whicb couldaccount for error in localization, sbould not influ-ence the accuracy of tbis ECG algorithm. Previ-ous catheter ablation may limit the usefulness ofan ECG algoritbm. The necrosis of tissue aroundthe ventricular insertion of the accessory pathwaymay affect tbe sequence of ventricular preexcita-tion, accounting for a different delta wave patternon the ECG. Nevertheless, it did not appear toaffect tbe accuracy in our study, as 48 of 121 pa-tients (40%) bad had a previous failed ablation.

References

1. RosenbiiLim FF. Hechl HH. Wilson FN. et al: Tlie po-tential variations of the thorax and the esophagus inanomalous atrioventricular excitation {Wolff-Parkin-son-White syndrome). Am Heart J l945;29:28l-326.

2. Frank R. Fontaine G, Guiraudon G, et al: Correlationentre Torientation de l'onde delta et la topographied'preexcitation dans !e syndrome de Wolff-Parkinson-White. Arch Mai Coeur 1977:70:441.

3. Gallagher JJ. Pritchett ELC, Sealy WC. et al: The pre-excitation syndromes. Proc Cardiovasc Dis 197S;20:285.

4. Iwa T, Kavi-asuji M. Misaki T, et al: Localization andinterruption of accessory conduction pathway in theWolff-Parklnson-White syndrome. J Thorac Cardio-vasc Surg 1980;80:27l.

5. Gallagher JJ, Selle JG, Sealy WC. et al: Intermediateseptal accessory pathways (IS-AP): A subset of preex-citation at risk for complete heart block/failure duringWPW surgery. {Abstract) Circulation 1986:74(SupplII):II-387.

6. Reddy GV. Schamroth L: The localization of bypasstracts in the Wolff-Parkinson-White syndrome from thesurface electrocardiogram. Am Heart J 1987; 113:984.

7. Lindsay BD. Crossen KJ, Cain ME: Concordance ofdistinguishing electrocardiographic features during si-nus rhythm with the location of accessory pathways inthe Wolff-Parkinson-White syndrome. Am J Cardiol

1987:59:1093-1102.8, Milstein S. Sharma AD. Guiraudon GM. et ai: An al-

gorithm for the eleetroeardiographie localization of ac-cessory pathways in the Wolff-Parkinson-White syn-drome. PACE 1987:10:55.5-563.

9. Willems JL. Robles de Medina EO, Bernard R, et ai:Criteria for intraventricular conduction disturbances andpre-excitation. J Am Coll Cardiol I985;5:I26I-1275,

10. Umery R, Chammil S. Wood DL, et al: Value of theresting i2-iead eiectrocardiogram and vectorcardio-gram for locating the accessory pathway in patientswith the Wolff-Parkinson-White syndrome. Br Heart J1987;58:324-332.

11. Epstein AE. Kirkiin JK. Holman WL. et al: Intennedi-ate septa] accessory pathways: Eiectrocardiographiccharacteristics, electrophysioiogic observations andtheir surgicai impiications, J Am Coll Cardiol 1991;17:i 570-1578.

12. Yuan S, IwaT, Tsubota M. et ai: Comparative study ofeight sets of ECG criteria for the iocaiization of the ac-cessory pathway in Woiff-Parkinson-White syndrome.J Electrocardiol I992;25:203-214.

13. Arruda M, Wang X. McClelland J. et al: ECG algo-rithm for predicting radiofrequency ablation site inposteroseptal accessory pathways. {Abstract) PACEI992;15(PtII):535.

14. Arruda M. Wang X, McClelland J. et al: ECG algo-rithm for predicting sites of successfui radiofrequeneyabiation of accessory pathways, {Abstract) PACEl993:l6(PtII):865.

15. Rodriguez LM. Smeets JL, de Cbillou C, et al: The 12-iead eiectrocardiogram in midseptai, anteroseptal, pos-teroscptai, and right free wall accessory pathways. AmJ Cardiol i 993:72: i 274-1280,

16. Scheinman MM. Wang YS. Van Hare GF. et al: Eiec-trocardiographic and electrophysiologic characteristicsof anterior. mid.septal and right anterior free wall ac-cessory pathways. J Am Coll Cardioi i992:20:1220-1229.

17. Fitzpatrick AP. Gonzales RP, Lesh MD, et ai: New al-gorithm for the Iocaiization of accessory atrioventricu-iar connections using a baseline electrocardiogram, JAm Coil Cardiol i994:23:iO7-i 16,

18. Xie B. Heaid SC, Bashir Y, et al: Localization of ac-cessory pathways from the i2-tead electrocardiogramusing anew algorithm. Am S Cardiol 1994;74:161-165.

19. Chiang CE. Chen SA, Teo WS. et ai: An accurate step-wise electrocardiographic algorithm for localization ofaccessory pathways in patients with Wolff-Parkin son-White syndrome from a comprehensive analysis ofdeita waves and R/S ratio during sinus rhythm. Am JCardioi i 995:76:40-46,

20. Arruda MS. Wang X. McClelland JH, et al: Negativedelta wave in lead 11 identifies posteroseptai accessorypathways requiring abiation in venous branches of thecoronary sinus. (Abstract) J Am Coil Cardiol 1994;(Suppi):224A.

21. Lorga AF. Sosa E. Scanavacca M, et al: Electrocardio-graphic identification of mid-septai accessory path-

12 Journal of Cardiovascular Electrophysiology Vol. 9, No. I. January 1998

ways in close proximity to the atrioventricular conduc-tion system. PACE i996:19(Pt n):l984-1987.

22. Iturralde P, Gomez VA. Colin L, el al: A new ECG al-gorithm for the localization of accessory pathways us-ing only the polarity of the QRS complex. J Electrocar-diol 1996;29:289-299.

23. Jackman WM. Wang X, Friday KJ. el al: Catheter ab-lation of accessory atrioventricular pathways (Wolff-Park in son-White syndrome) by radiofrequency current.N EngI J Med 1991:324:1605-1611.

24. Schulter M, Geiger M, Siebels J, et al; Catheter ablationusing radiofrequency current to cure symptomatic patientswith tachyarrhythmias related to an accessory atrioven-tricuiiir pathway. Circulation 1991; 4:1644-1661.

25. Haissaguerre M. Dartigues JF. Warin JF, et al: Electro-gram patterns predictive of successful catheter ablationof accessory pathways. Circulation 1991:84:188-202.

26. Calkins H. Langberg J, Sousa J. et al: Radiofrequencycatheter ablation of accessory atrioventricular connec-tions in 250 patients. Circulation I992;85:I337-I346.

27. Lesh MD, Van Hare GF. Schamp DJ. et al: Curativepercutaneous catheter ablation using radiofrequencyenergy for accessory pathways in all locations: Resultsin 100 consecutive patients. J Am Coll Cardiol 1992;19:1303-1307.

28. Gerlis LM. Davies MJ. Boyle R. et al: Preexcitationdue to accessory sinoventricuiar connections associ-ated with coronary sinus ancurysms. A report of twocases. BrHeartJ l985;53:314-322.

29. Ho SY, Russell G, Rowland E: Coronary venous aneu-rysms and accessory atrio-ventricular connections. BrHeart J 1988:60:348-351.

30. Robinson K. Davies MJ. Krikler DM: Type A Wolff-Parkinson-White syndrome obscured by left bundlebranch block associated with a vascular malformationof the coronary sinus. Br Heart J 1988:60:352-354.

31. Guiraudon GM. Guiraudon CM. Klein GJ. et al: Thecoronary sinus diverticulum: A pathologic entity asso-ciated with the Wolff-Parkinson-White syndrome. AmJCardioi 1988:62:733-735.

32. Stamato N. Goodwin M. Foy B: Diagnosis of coronarysinus diverticulum in Wolff-Parkinson-White syn-drome using coronary angiography. PACE 1989;12:

1589-1591.33. McGiffin DC. Maslcrson ML. Stafford WJ: Wolff-

Parkinson-White syndrome associated wiih a coronarysinus diverticulum: Ablative surgical approach. PACE1990;I3:966-969.

34. Ho S. Gupta I. Anderson RH: Aneurysm of the coro-nary sinus. Thorax 1983:38:686-689.

35. Oren J. McClelland J. Beckman K, et al: Epicardialposteroseptal accessory pathways requiring ablationfrom the middle cardiac vein. (Abstract) PACE 1992;l5(PtII):535.

36. Arruda MS. Beckman KJ, McClelland JH. et al: Coro*nary sinus anatomy and anomalies in patients with pos-teroseplal accessory pathway requiring ablation wilhina venous branch of the coronary sinus. (Abstract) J AmCollCardiot I994:224A.

37. Wang X. McClelland J, Beckman K, et al: Left free-wall accessory pathways which require ablation fromthe coronary sinus: Unique coronary sinus electrogrampattern. (Abstract) Circulation I992;86:2314.

38. Lesh MD. Van Hare G. Kao AK. et al: Radiofrequencycatheter ablation of Wolff-Parkinson-White syndromeassociated with a coronary sinus diverticulum. PACE199I;14:1479-1484.

39. Kuck KH, Schluter M, Chiladakis I: Accessory path-ways anatomically related to the coronary sinus. (Ab-stract) Circulation 1992;86:I-782.

40. Langberg JJ, Man KC. Vorperian VT, et al: Recogni-tion and catheter ablation of subepicardial accessorypathways. J Am Coll Cardiol 1993:22:1100-1104.

41. Warin JF, Haissaguerre M. D'ivernois C, et al: Cathe-ter ablation of accessory pathways: Technique and re-sults in 248 patienls. PACE 1990;13:1609-16l4.

42. Coiovita PG, Packer DL, Pressley JC. et al: Frequency,diagnosis and clinical characteristics of patients withmultiple atrioventricular accessory pathways. Am JCardiol 1987;.'>9:6OI-6O6.

43. Gallagher JJ, Sealy WC. Kasell J, et al: Multiple acces-sory pathways in patients with the preexcitation syn-drome. Circulation I976;54:571-591.

44. McClelland JH. Beckman KJ, Wang X. et al: Radiofre-quency ablalion elucidates accessory pathway anatomy.(Abstract) Circulation 199l;84:II-24.