Analysis of Intracardiac ElectrogramsShowing Monomorphic Ventricular

Tachycardia in Patients With ImplantableCardioverter-Defibrillators

Mohammad Saeed, MD, Mark S. Link, MD, Srijoy Mahapatra, MD, Majd Mouded, MD,David Tzeng, Vivian Jung, Robert Contreras, Craig Swygman, Munther Homoud, MD,

N.A. Mark Estes III, MD, and Paul J. Wang, MD

Ventricular tachycardia (VT) initiation and its relation tovarious clinical factors was studied by reviewing intra-cardiac electrograms from patients with implantablecardioverter-defibrillators. Events were divided into (1)sudden onset without preceding ventricular prematurecomplexes (VPCs), (2) extrasystolic onset with VPCs, or(3) paced, depending on the type and morphology of thelast 5 beats before initiation of VT. Prematurity index,sinus rate, cycle length, and presence of short-long-shortsequence for each episode was noted. A total of 268episodes of VT among 52 patients were analyzed. Ex-trasystolic initiation was the most frequent pattern (177;66%) followed by sudden onset (75; 28%) and paced(16; 6%). Among extrasystolic onset, 99 episodes (56%)were due to multiple VPCs and 149 episodes (84%) haddifferent VPC morphology than the subsequent VT.

Among pacing-induced VT, 13 of 16 episodes were dueto inappropriate pacing due to undersensing of prior Rwaves. Sudden-onset episodes were slower (mean cyclelength 383 6 97 ms) than extrasystolic (mean cyclelength 336 6 88 ms, p 5 0.002) and paced (mean cyclelength 313 6 85 ms, p 5 0.01) onset. Patients in thesudden-onset group had better left ventricular ejectionfraction (33 6 15%) than the extrasystolic (29 6 11%, p<0.001) and paced (28 6 14%, p <0.01) groups. Ex-trasystolic onset with multiple, late coupled VPCs wasthe most common pattern of VT initiation and was as-sociated with lower ejection fraction. Sudden-onset ini-tiation was more common with better preserved systolicfunction. Q2000 by Excerpta Medica, Inc.

(Am J Cardiol 2000;85:580–587)

Reentry, triggered activity, or abnormal automatic-ity are mechanisms thought to be responsible for

the generation of ventricular arrhythmias.1,2 The con-tribution of a particular mechanism depends on thepresence of underlying cardiac substrate and dynamicfactors such as coronary ischemia and autonomic in-fluences.3–5 In the setting of coronary artery disease orprevious myocardial infarction, reentry is more prev-alent.1,6–8 Recognition of specific electrogram pat-terns occurring at the time of ventricular tachycardia(VT) initiation can help in understanding the electro-physiologic mechanisms responsible for arrhythmiainitiation and may lead to better diagnostic and ther-apeutic interventions. Several studies analyze themechanism of onset of ventricular tachyarrhythmiasbased on retrospective evaluation of ambulatory elec-trocardiography.9–13 They are limited by the acciden-tal registration of the arrhythmia episode, the smallnumber of patients included, and because each patientis only registered once. The progress in implantablecardioverter defibrillator (ICD) technology has im-

proved the diagnostic and therapeutic efficacy of thesedevices in the management of ventricular tachyar-rhythmias. Extended recording of electrical eventssurrounding delivered and aborted device therapy notonly permits more accurate characterization of therhythm leading to device intervention, but also pro-vides documentation of electrical events immediatelypreceding the index arrhythmia.14–18 The presentstudy was undertaken to gain insight into the mecha-nism of initiation of spontaneous monomorphic VT byanalyzing stored intracardiac electrograms from pa-tients with ICDs. In particular, we tried to relate thepattern of initiation to the precise number and mor-phology of beats leading up to VT. We also sought tocorrelate the presence of various clinical features andarrhythmia characteristics such as cycle length withdifferent patterns of VT initiation.

METHODSAll episodes of VT in patients with third-genera-

tion ICD devices placed at our institution betweenOctober 1995 and January 1998 were reviewed. Intra-cardiac stored electrograms with spontaneous sus-tained VT requiring therapy with antitachycardia pac-ing or direct-current cardioversion were identified.Only events that were monomorphic and had a mini-mum of 5 beats before the onset and after terminationof the VT were selected for further analyses.

Monomorphic VT was identified by a sudden in-

From the New England Cardiac Arrhythmia Center, New EnglandMedical Center, Boston, Massachusetts. Manuscript received August19, 1999; revised manuscript received and accepted October 19,1999.

Address for reprints: Paul J. Wang, MD, New England CardiacArrhythmia Service, Division of Cardiology, Box 173, New EnglandMedical Center, 750 Washington Street, Boston, Massachusetts02111. E-mail: pwang1@lifespan.org.

580 ©2000 by Excerpta Medica, Inc. All rights reserved. 0002-9149/00/$–see front matterThe American Journal of Cardiology Vol. 85 March 1, 2000 PII S0002-9149(99)00815-2

crease in rate along with a change in electrogrammorphology from the baseline rhythm, a constant cy-cle length that did not vary.10%, and a uniformelectrogram morphology during the tachycardia. Aventricular premature complex (VPC) was identifiedas any electrogram preceding the VT with morphol-ogy different from the baseline rhythm and couplinginterval,90% of the sinus cycle length. The first beatof VT, when morphologically different from the sub-sequent tachycardia, was considered a VPC. In casesin which the first beat of VT was morphologicallysimilar to the subsequent tachycardia, the couplinginterval between the first beat and the VT was evalu-ated. That beat was considered either (1) the first beatof VT when the coupling interval was,110% of VTcycle length, or (2) a VPC if the coupling interval was.110% of the VT cycle length.

VT episodes were categorized into 3 groups de-pending on the type and morphology of the last 5 beatsbefore the initiation of VT: (1) sudden onset withoutpreceding VPC, (2) extrasystolic onset with precedingVPCs, or (3) paced beats. The initiation was charac-terized as paced or extrasystolic if paced beats orVPCs preceded the VT. If no paced beat or VPC waspresent before the VT, the initiation was labeled assudden onset. In cases of VT after paced beats, thelongest pause was compared with the bradycardialower rate interval of the ICD. When the pause wasequal to it, the paced beat was considered appropriate;when the pause was less than the lower rate interval itwas labeled as inappropriate, representing undersens-ing.

Pause-dependent initiation of VT was defined asthe presence of a short-long-short sequence immedi-ately before VT in the extrasystolic group. An intervalwas classified as short if,80% or long if.120% ofmean sinus rate, respectively. Premature depolariza-tions were recorded as similar or dissimilar to thesubsequent VT based qualitatively on electrogrammorphology. The cycle length of all beats duringbaseline rhythm and VT was measured. Mean sinusrate was taken as the average of all sinus beats beforeVT.

Prematurity index was calculated by normalizingthe coupling interval to the preceding RR interval. Thecoupling interval was defined differently dependingon the initiation category. It was taken as the intervalbetween the first beat of VT and the previous beat forsudden-onset initiation. In case of extrasystolic initi-ation, it was defined as the interval between the firstVPC and the previous beat. For initiation by a pacedbeat, the coupling interval was taken as the intervalbetween the paced beat and the previous beat.

Clinical data: Clinical information for each patientincluding age, gender, underlying heart disease, leftventricular ejection fraction, and antiarrhythmic drugusage at the time of index arrhythmia was documentedby review of clinical records.

Statistical analysis: Continuous data are expressedas mean6 SD. Means among different groups werecompared by analysis of variance. In case of nominalor ordinal data the groups were compared by using the

chi-square test. A p value,0.05 was considered sta-tistically significant. A possible relation between thevarious clinical variables and the different modes ofinitiation was evaluated by logistic linear regression.

RESULTSA total of 192 patients had third-generation ICD

devices placed between October 1995 and January1998 at our institution. In this group of patients, 555episodes of VT occurred among 67 patients from thedate of implant to April 1998. Of these, 268 episodesamong 52 patients met study criteria. Of 287 disqual-ified episodes, 118 had,5 beats before initiation ofVT, 108 had no electrogram recordings available forreview, 44 received no therapy for that particularevent, and 17 episodes were polymorphic in nature.The mean age of patients in our study was 67.66 9years. There were 42 men and 10 women in the group.The number of episodes for a single patient rangedfrom 1 to 27 episodes/patient (mean of 56 6 epi-sodes/patient). The underlying heart disease was cor-onary artery disease (40 of 52 patients), idiopathicdilated cardiomyopathy (7 patients), arrhythmogenicright ventricular dysplasia (3 patients), and idiopathicVT (2 patients). The mean left ventricular ejectionfraction for the group was 296 13%. There were 30patients taking antiarrhythmic drug therapy at the timeof index VT, whereas 12 were takingb blockers.Table I lists baseline characteristics of the study pop-ulation.

The most frequent initiation pattern was extrasys-tolic (Figure 1), observed in 177 episodes (66%),followed by 75 episodes (28%) of VT with sudden-onset initiation (Figure 2), whereas pacing-induced(Figure 3) VT occurred infrequently in only 16 epi-sodes (6%). In the extrasystolic initiation group, thepresence of$3 continuous VPCs was the most fre-quent pattern of initiation (39%) followed by a singleVPC (22%), VPC interrupted by sinus beats (21%),

TABLE I Baseline Characteristics of Patient Population (n 552)

Age (yrs) 68 6 9Men/women 42/10Disease

Coronary artery disease 40Idiopathic dilated cardiomyopathy 7Arrhythmogenic right ventricular dysplasia 3Idiopathic VT 2

Left ventricular ejection fraction (%) 29 6 13Antiarrhythmic drugs (1/2) 42/10

b Blocker 12Amiodarone 11Sotalol 6Digoxin 7Other 10

ICD manufacturerVentritex (St. Jude Medical) 32CPI (Guidant) 20

ICD recordingNear field 32Far field 20

CPI 5 Cardiac Pacemaker Incorporated.

ARRHYTHMIAS AND CONDUCTION DISTURBANCES/ANALYSIS OF INTRACARDIAC ELECTROGRAMS 581

and 2 continuous VPCs (17%). Among the extrasys-tolic initiations, VPCs in 149 episodes (84%) weredifferent in morphology from the subsequent VT. Inonly 28 of 177 episodes (16%), the premature depo-larizations preceding the VT were of the same mor-phology. Table II lists the arrhythmia characteristicsand clinical variables of the 3 main initiation catego-ries.

Seventeen patients had only 1 episode of VT,whereas 35 had$2 episodes. The reproducibility ofthe initiation sequence was quite high among patientswith $2 episodes. Among the 35 patients with mul-tiple episodes, 31 (88%) had at least 2 events with the

same initiation sequence, whereas 16 (46%) had thesame single pattern of initiation repeated during allsubsequent episodes of VT.

Six patients had 16 episodes of pacing-inducedVT. Among the pacing-induced VT, 3 episodes werepreceded by appropriately timed ventricular beats atthe lower rate cutoff. However, 13 episodes followedinappropriately paced ventricular stimuli because ofundersensing of the native R wave. Paced beats oc-curred during the compensatory pause following apremature beat in all 3 cases of VT induced by ap-propriately paced stimuli. In case of VT episodesinduced by inappropriately paced ventricular stimuli,

FIGURE 1. Sudden-onset initiation: monomorphic VT initiated by a single VPC that is similar in morphology to the VT.

FIGURE 2. Extrasystolic initiation: monomorphic VT preceded by a single VPC that is different in morphology from the VT.

582 THE AMERICAN JOURNAL OF CARDIOLOGYT VOL. 85 MARCH 1, 2000

the paced beat occurred because of failure to sense aprior R wave (premature beats in 12 of 13 and sinus beatsin 1 of 13 episodes). Inappropriate paced beats thatinduced VT had a shorter coupling interval (confidenceinterval 4326 167 ms) and prematurity index (0.660.3) than appropriately paced beats (confidence interval1,1366 118 ms, prematurity index 2.66 0.5).

The overall mean prematurity index was 0.8060.57 for all episodes. The sudden-onset group had aprematurity index of 0.786 0.2, whereas the extra-

systolic and paced group had a prematurity index of0.806 0.4 and 0.586 0.3, respectively. There was nostatistical significant difference in prematurity indexamong the 3 initiation groups before and after con-trolling for underlying heart disease, antiarrhythmicdrug usage, or degree of left ventricular dysfunction.The shortest prematurity index was associated withepisodes having a short-long-short sequence. Themean prematurity index for the events with a preced-ing short-long-short sequence was 0.656 0.1 com-

FIGURE 3. Paced initiation: a ventricular paced beat following a VPC and an undersensed sinus beat leads to initiation of VT.

TABLE II Distribution of Arrhythmia Characteristics and Clinical Variables Amongthe Three Main VT Initiation Subgroups

Initiation SequenceSudden Onset(mean 6 SD)

Extrasystolic(mean 6 SD)

Pace(mean 6 SD)

Episodes 75 177 16Arrhythmia characteristics p ,0.008

Sinus rate (ms) 765 6 141 849 6 256 778 6 175Prematurity index 0.78 6 0.2 0.80 6 0.4 0.58 6 0.3Pause dependency 0 35 4

p ,0.002VT cycle length (ms) 383 6 97 336 6 88 313 6 85

p ,0.01Clinical variables

Coronary artery disease* 20 32 5p ,0.01

Left ventricular ejection fraction* 33 6 15 29 6 11 28 6 14Drug usage p , 0.01

Class 1 antiarrhythmic 13 21 3b blocker 27 71 12Amiodarone 33 64 3Sotalol 8 14 0

p ,0.0001Digoxin 2 40 1

*Coronary artery disease and left ventricular ejection fraction are shown as distributed by patients,while rest of variables are shown as distributed by VT episodes.

ARRHYTHMIAS AND CONDUCTION DISTURBANCES/ANALYSIS OF INTRACARDIAC ELECTROGRAMS 583

pared with 0.816 0.4 for the events without such asequence (p5 0.03).

The cycle length of VT varied depending on thespecific initiation pattern, with episodes of VT follow-ing sudden onset being slower than VT followingVPC (Figure 4). The mean cycle length for VT with asudden-onset initiation pattern was 3836 97 ms com-pared with 3366 88 ms with extrasystolic initiation(p ,0.002). The mean cycle length for pacing-initi-ated episodes was 3136 85 ms, significantly less thansudden-onset initiation (p,0.01). Among the extra-systolic group, the fastest VT cycle length was of theevents with$3 continuous VPCs (3106 62 ms). TheVT cycle length following short-long-short wasshorter (3186 49 ms) than episodes without such asequence (3536 97 ms, p5 0.02). The average sinusrate before the VT varied among the 3 groups (Figure4). The mean sinus rate was slowest in the extrasys-tolic initiation group (8496 256 ms) and fastest in thegroup with sudden onset (7656 141 ms), and thedifference was statistically significant at p,0.008.The pacing-induced VT had a preceding sinus rate of778 6 175 ms, which was not significantly differentfrom sudden onset or extrasystolic initiation. Episodespreceded by a short-long-short sequence had a slowersinus rate (9576 201 ms) than episodes without ashort-long-short sequence (7996 225 ms, p,0.0002).

Clinical parameters from each patient were ana-lyzed for any possible association with the differentmodes of initiation (Table II). There was no associa-tion of gender, underlying heart disease, and use ofvarious types of antiarrhythmic drugs with the initia-tion sequence of monomorphic VT. Higher age wasassociated with a more sudden-onset pattern, whereaspatients in the younger age group had more VPC-initiated VT (p ,0.0003). The age difference re-mained statistically significant after controlling fordisease type, antiarrhythmic drug usage, and degree ofleft ventricular dysfunction. Left ventricular systolicfunction correlated with different types of initiationpattern. Patients in the sudden-onset initiation group

had a higher ejection fraction (336 15%) than pa-tients in the extrasystolic (296 11%) or paced (28613%) initiation group. The difference between thesudden-onset and extrasystolic groups, and sudden-onset and paced groups was statistically significant (p,0.001 and p,0.01, respectively). After adjustingfor other variables the difference still remained statis-tically significant.

Nine patients had 39 episodes (15%) with a short-long-short sequence (Figure 5) immediately precedingVT. The mean short interval was 6486 132 ms,whereas the mean long interval was 1,1756 223 ms.All episodes were due to a premature beat causing ashort interval followed by a sinus beat responsible forthe long interval, and then another premature beat orthe first beat of VT producing the second short inter-val. All 39 episodes of VT preceded by a short-long-short sequence occurred in patients with underlyingcoronary artery disease, whereas no pause-dependentinitiation was observed in patients without coronaryartery disease (p5 0.02). There was no correlationbetween presence of pause-dependent initiation andage, gender, or degree of left ventricular dysfunction.Episodes preceded by the short-long-short sequencehad shorter cycle length. There was a statisticallyhigher incidence of short-long-short sequences in pa-tients taking amiodarone, with 5 of 39 episodes inpatients taking amiodarone, in contrast to 9 of 227episodes in patients not taking this particular antiar-rhythmic agent (p5 0.003). Beta-blocker usage wasgreater among patients with a short-long-short se-quence (27 of 39 episodes) than among patients with-out such a sequence (71 of 229, p,0.0001) (TableIII).

DISCUSSIONMost VT episodes in our study were preceded by

VPCs. These were often multiple and different inmorphology from the subsequent VT. The couplinginterval was typically long. These findings are inagreement with the data of Marchlinski et al16 andRoelke et al19 who studied patients with coronary

FIGURE 4. Relation of prior sinus rate and VT cycle length to the different VT initiation sequences. VT episodes following sudden-onset(gray bars) initiation are slower than extrasystolic (black bars) or pace (open bars) onset. Sinus rate is faster before sudden onsetthan extrasystolic onset. *p <0.002 for difference between sudden onset and extrasystolic onset. †p <0.01 for difference betweensudden onset and paced; ‡p <0.008 for difference between sudden onset and extrasystolic.

584 THE AMERICAN JOURNAL OF CARDIOLOGYT VOL. 85 MARCH 1, 2000

artery disease and previous myocardial infarction.They found a similar incidence and prematurity ofVPCs before VT in their patient population. Meyer-feldt et al20 also looked at the mode of onset of VT ina group of patients with dilated cardiomyopathy andcoronary artery disease. In their study single prema-ture beats or couplets initiated three fourths of VTepisodes.

A significant percentage of VT episodes startedsuddenly without a preceding VPC. Other investiga-tors have classified these as VT with single VPC ofsimilar morphology.19 We chose to assign these

episodes to the extrasystolic categoryif the VPC was loosely coupled, or tothe sudden-onset category if it wastightly coupled to the subsequentVT. This classification was adoptedbecause presence of tightly coupledand morphologically similar VPCwas thought to represent a distinctelectrophysiologic mechanism. Wefound that the sudden-onset and ex-trasystolic initiation patterns tend tohave different electrical characteris-tics. Sudden-onset VT episodes hadlonger cycle length and faster sinusrate before VT, whereas extrasys-tolic-initiated episodes had shortercycle length and a slower sinus rate.These findings did not seem to bedue to differences in the electrical(e.g., coupling interval, prematurityindex) or clinical (e.g., heart dis-ease, ejection fraction, or antiarrhyth-mic drug usage) variables among

patients.It has been recognized that appropriately timed

noncompetitive ventricular pacing may potentially ini-tiate VT in patients prone to these arrhythmias.21

Recently, there have been reports of initiation of VTby inappropriate ventricular pacing in patients withICD.21–24 These cases have mostly resulted from un-dersensing of the sinus beat. In contrast, we found thatmost undersensed beats were of different morphology.Our study is the first of its kind systematically lookingat a large database of patients to document the pres-ence of ICD-induced VT from prior R-wave under-

FIGURE 5. Short-long-short sequence: a closely coupled VPC leads to a short interval followed by a compensatory pause. A mono-morphic VT follows after another short interval set up by the first beat of VT.

TABLE III Distribution of Arrhythmia Characteristics and Clinical Variables AmongSubgroups With and Without Short-Long-Short Sequence

SLS(mean 6 SD)

No SLS(mean 6 SD) p Value

Episodes 39 229Arrhythmia characteristics

VT cycle length (ms) 318 6 49 353 6 97 0.02Prematurity index 0.65 6 0.1 0.81 6 0.4 0.03Sinus rate (ms) 957 6 201 799 6 225 ,0.0002

Clinical variablesCoronary artery disease* 9 31 0.02Left ventricular ejectionfraction*

30 6 7 29.4 6 13 NS

Antiarrhythmic drugsAmiodarone 5 9 0.003Sotalol 3 19 NSClass I 1 36 NSb blocker 27 71 ,0.0001

*Coronary artery disease and left ventricular ejection fraction are shown as distributed by patients,while rest of variables are shown as distributed by VT episodes.

SLS 5 short-long-short; class I antiarrhythmic drugs included quinidine and procainamide.

ARRHYTHMIAS AND CONDUCTION DISTURBANCES/ANALYSIS OF INTRACARDIAC ELECTROGRAMS 585

sensing. It shows that this complication occurs in asmall percentage of patients and is a form of proar-rhythmia associated with the use of third-generationICD devices.

Pause-dependent initiation as evidenced by a short-long-short sequence is classically associated withpolymorphic VT.6 The short-long-short sequence, bycausing heterogeneity of refractoriness, can lead toreentry and can also cause monomorphic VT.3,25,26Inour study we observed 39 events (15%) of VT pre-ceded by a short-long-short sequence. The low inci-dence of pause-initiated VT in this group of patients isin agreement with observations made by other inves-tigators.19,27,28Studies have shown that there is a highintraindividual reproducibility of VT with the sameonset seen in up to 80% of patients with recurrentVT.20,29 Our data are consistent with the observationof other investigators, showing 88% reproducibility ofthe initiation sequence among our patients. This un-derscores patient-specific factors, namely, the pres-ence of fixed or functional anatomic pathways areimportant in the genesis of recurrent monomorphicVT.

Initiation of sudden onset appeared to be morecommon among patients with relatively preserved sys-tolic function. In contrast, patients with poor ejectionfraction (,35%) had more ventricular ectopy imme-diately preceding the VT and a higher frequency ofextrasystolic initiation. Although the mechanism re-sponsible for this correlation is not clear, this may bebecause ventricular ectopy is more frequent with in-creasing severity of underlying heart disease. Theapparent association between the degree of left ven-tricular dysfunction and specific pattern of VT initia-tion is a new finding and may have diagnostic ortherapeutic implications. Further studies are requiredto confirm this association and to elucidate its patho-physiology.

Currently recognized mechanisms of initiation ofVT include reentry, normal and abnormal automatic-ity, and triggered activity due to early or delayedafterdepolarizations.1–3 Several features of monomor-phic VT suggest reentry as the underlying mechanism,especially in the setting of chronic coronary arterydisease6–9 and right ventricular dysplasia.30 The pre-cise mechanism of ventricular arrhythmias in idio-pathic dilated cardiomyopathy is not known but reen-try and abnormal automaticity are thought to play animportant role. Although our study was an observa-tional one and did not address the mechanism of VTinitiation directly, several inferences can be made. Inour study VPCs that are often multiple, different inmorphology, and loosely coupled to the subsequenttachycardia likely initiated VT via setting up reentry.VT may also start suddenly. This may be due to aVPC arising from a site very close to the reentrantcircuit or concealed decremental conduction of thesinus beat just before the VT setting up a reentry.2

Sudden-onset initiation could also be due to a focus ofincreased automaticity or triggered activity. Theformer mechanism may be more important in patientswith coronary artery disease, whose sinus beat may

conduct slowly through diseased tissue to set up re-entry, and the first beat exiting the circuit will actuallybe the first beat of VT as evidenced by close couplingand similar morphology. Although we did not find anydifferences in initiation among the various diseasecategories, this could be related to the small samplesize of the patients without coronary artery disease inour study.

In addition to understanding the mechanism ofinitiation of VT, our findings may also have therapeu-tic implications. Bradycardia pacing devices that canprevent a short-long-short coupling interval and pro-duce rate smoothing may be capable of preventing VTepisodes whose initiation is pause dependent. Devicereprogramming or development of newer algorithmsto minimize R-wave undersensing may avoid ICD-induced VT. These findings may also be useful indeveloping algorithms to enhance specificity of futuredevices in diagnosing and treating ventricular arrhyth-mias. In addition, large databases of these events maybe helpful in answering clinical questions regardingdisease or gender specificity of ventricular tachyar-rhythmias.

Visual inspection of the intracardiac electrogram,especially that derived from rate-sensing bipole, issubjective and carries inherent limitations in interpret-ing morphology. Strict morphometric analyses of lo-cal electrograms was not performed.18,31Electrogramsthat appear similar may not originate from the sameventricular site. Also, all premature depolarizationswere assumed to be ventricular in origin, and the factthat some of them might have been supraventricularpremature depolarizations cannot be excluded.27

1. Josephson ME, Almentral JM, Buxton AE, Marchlinski FE. Mechanisms ofventricular tachycardia.Circulation 1987;75(suppl III):III-41–III-47.2. Berger MD, Waxman HL, Buxton AE, Marchlinski FE, Josephson ME.Spontaneous compared with induced onset of sustained ventricular tachycardia.Circulation 1988;78:885–892.3. Gomes JA, Alexopoulus D, Winters SL, Desmukh P, Fuster V, Suh K. The roleof silent ischemia, the arrhythmic substrate and the short-long sequence in thegenesis of sudden cardiac death.J Am Coll Cardiol1989;14:1618–1625.4. Myerberg RJ, Kessler KM, Castellanos A. Pathophysiology of sudden cardiacdeath.Pacing Clin Electrophysiol1991;14(suppl II):935.5. Myerberg RJ, Kessler KM, Castellanos A. Sudden cardiac death. Structure,function, and time dependence of risk.Circulation 1992;85(suppl I):2.6. El-Sherif N, Schlerlag BJ, Lazzara R, Hope RR. Reentrant ventricular arrhyth-mias in the late myocardial infarction period: I. Conduction characteristics in theinfarction zone.Circulation 1977;55:686–702.7. El-Sherif N, Hope PR, Scherlag BJ, Lazarra R. Reentrant ventricular arrhyth-mias in the late myocardial infarction period: II. Patterns of initiation andtermination of reentry.Circulation 1977;55:702–719.8. Mehra R, Zeiler RH, Gough WB, El-Sherif N. Reentrant ventricular arrhyth-mias in the late myocardial infarction period.Circulation 1983;67:11–24.9. de Luna BA, Coumel P, Leclerq JF. Ambulatory sudden cardiac death:mechanisms of production of fatal arrhythmias on the basis of data from 157cases.Am Heart J1989;117:151–159.10. Kempf FC, Josephson ME. Cardiac arrest recorded on ambulatory electro-cardiograms.Am J Cardiol1984;53:1577–1582.11. Pratt CM, Francis MJ, Luck JC, Wyndham CR, Miller RR, Quinones MA.Analysis of ambulatory electrocardiograms in 15 patients during spontaneousventricular fibrillation with special reference to preceding arrhythmic events.J Am Coll Cardiol1983;2:789–797.12. Milner PG, Platia EV, Reid PR, Griffith LSC. Ambulatory electrocardio-graphic recordings at the time of fatal cardiac arrest.Am J Cardiol1985;56:588–592.13. Lewis BJ, Antman EM, Graboys TB. Detailed analysis of 24 hour ambulatoryelectrocardiographic recordings during ventricular fibrillation and torsades depointes.J Am Coll Cardiol1983;2:4226–4236.

586 THE AMERICAN JOURNAL OF CARDIOLOGYT VOL. 85 MARCH 1, 2000

14. Hook BG, Marchlinski FE. Value of ventricular electrogram recordings in thediagnosis of arrhythmias precipitating electrical device shock therapy.J Am CollCardiol 1991;17:985–990.15. Marchlinski FE, Gottlieb CD, Sarter B, Finkle J, Hooke B, Callans D,Schwartzman D. ICD data storage: value in arrhythmia management.PACE1993;16:527–534.16. Marchlinski FE, Scwartzman D, Gottlieb CD, Gonz´alez-Zuelgaray J, CallansDJ. Electrical events associated with arrhythmia initiation and stimulation tech-niques for arrhythmia prevention. In Zipes DP, Jalife J, eds. Cardiac Electro-physiology. 2nd ed. Philadelphia: WB Saunders, 1995:863–877.17. Hook BG, Callans DJ, Keinman RB, Flores BT, Marchlinski FE. Implantablecardioverter-defibrillator therapy in the absence of significant symptoms: rhythmdiagnosis and management aided by stored electrogram analysis.Circulation1993;87:1897–1906.18. Auricchio A, Hartung W, Geller C, Klein H. Clinical relevance of storedelectrograms for implantable cardioverter-defibrillator (ICD) troubleshooting andunderstanding of mechanisms for ventricular tachyarrhythmias.Am J Cardiol1996;78(suppl 5A):33–41.19. Roelke M, Garan H, McGovern BA, Ruskin JN. Analysis of initiation ofspontaneous monomorphic ventricular tachycardia by stored intracardiac electro-grams.J Am Coll Cardiol1994;23:117–122.20. Meyerfeldt U, Schirdewan A, Wiedemann M, Schutt H, Zimmermann F, LuftFC, Dietz R. The mode of onset of ventricular tachycardia: a patient specificphenomenon.Eur Heart J1997;18:1956–1965.21. Nunain SO, Roelke M, Trouton T, Osswald S, Kim YH, Sosa-Surez G,Brooks DR, McGovern B, Guy M, Torchiana DF. Limitations and late compli-cations of third generation automatic cardioverter defibrillators.Circulation1995;91:2204–2213.22. Callans DJ, Hook BG, Kleiman RB, Mitra RL, Flores BT, Marchlinski FE.

Unique sensing errors in third generation implantable cardioverter-defibrillators.J Am Coll Cardiol1993;22:1135–1140.23. Reiter M J, Mann D E. Sensing and tachyarrhythmia detection problems inimplantable cardioverter defibrillators.J Cardiovasc Electrophysiol1996;6:542–558.24. Vlay LC, Vlay SC. Pacing induced ventricular fibrillation in internal cardio-verter defibrillator patients.PACE;20:132–133.25. Kay GN, Plumb VJ, Arciniegas JG, Henthorn RW, Waldo AL. Torsades depointes: the long-short initiating sequence and other clinical features: observa-tions in 32 patients.J Am Coll Cardiol1983;2:806–817.26. Cranefield PF, Aronson RS. Torsades de pointes and other pause-inducedventricular tachycarida: the short-long-short sequence and early after-depolariza-tions.PACE1998;11:670–678.27. Taylor E, Dinerman J, Riley J, Lawrence JH, Tomaselli GF, Berger RD,Calkin H. Relationship between mode of onset of VT initiation and heart diseasetype (abstr).Circulation 1997;96(suppl I):I-152.28. Wolpert C, Gunia S, Schumacher B, Lilienthal B, Schimpf R, Esmailzadeh C,Schneider C. Initiation of ventricular tachycardia: analysis of mode of onset inICD patients (abstr).Circulation 1997;96(suppl I):I-636.29. Roelke M, Garan H, McGovern BA, Ruskin JN. Intra-patient variability inthe initiation of ventricular tachycardia by analysis of stored ventricular electro-grams (abstr).J Am Coll Cardiol1993;21(suppl A):405A.30. Fontaine G, Guiraudon G, Frank R. Stimulation studies and epicardialmapping in ventricular tachycardia: study of mechanisms and selection forsurgery. In: Kulbertus HE, ed. Reentrant Arrhythmias. Lancaster, PA: MTPPress, 1977:334–35031. Hallet N, Monahan K, Casavant D, Epstein L, Josephson M. Inadequacy ofqualitative implantable cardioverter defibrillator electrogram analysis to distin-guish supraventricular from ventricular tachycardia due to electrogram changesduring normally conducted complexes.PACE1997;20:1723–1726.

ARRHYTHMIAS AND CONDUCTION DISTURBANCES/ANALYSIS OF INTRACARDIAC ELECTROGRAMS 587