Implications of the Electrophysiologic Study Versus Electrocardiographic Monitoring Trial for Controlling Ventricular Tachycardia and Fibrillation

Implications of the Electrophysiologic Study Versus Eloctrocardiogra~hic Monitwin Trial for Controlhng \ ycardia and Fibrillation Ventricular Tac James A Retffel, MD The Electrophysiologic Study versus Eleetrocardio-alsostatistically morelikelyto achievean efficacypreraphicMonitoring(ESVEM)trialhad 2 objectives. The dictionthanany of thesodiumchannelblockingdrugs. ! rstWdstodetermine theaccuracy of noninvas-we versus Amiodaronewas notusedin ESVEM.It hasbeensuginvasivemeansof predicting theefficacy of drugtreat- gestedthattheseconclusions, whichdifferfromthoseof mentfor ventricular hzchycardia/ventricular fibrillation other,lesscontrolled, invasiveand noninvasive studies, of thePaficularefficacy crikriaused (W/VF). A secondobiective wasto determine therela- mightbe because analyses of theEStiveefficacies of 7antiarrhythmic drugsusedinthetreafi in theESVEMprotocol.Retrospective mentof ventricular tachyarrhythmias. ESVEMwas the VEMdatawereperformedusingmorerigidefficacy crifirstopportunityto compareprospectively theefficacy, teriathanwereusedin theoriginalESVEManalysis:a wasrequiredfor safety,and tolerabilityof a varietyof antiarrhythmicgreaterdegreeof ectopysuppression drugsin thesomepatientpopulation. No significant dif- Hohermonitoring,and morestringentefficacydefiniprotocol of theEPS ferencewasobservedbetweensuppression of sponta- tionswererequiredinthestimulation fromtheretrospective analysesand olher neousventricular arrhythmias on Hollermonitoring and limb.Results Inpatients suppression of inducible ventricular arrhythmias byelec- siudiessupporttheinitialESVEMconclusions. and inducible susimined ventrictrophysiologic sludy(EPS)in termsof theabilityto pre- withbothspontaneous dictthesuccess of drugtherapy.Therewasalsono di~ ulartachyarrhythmias aswellasfrequentspontaneous ventricular contractions, therapywithsatalal ference in predictiveaccuracyif patientsin the premature electrophysiologic limb showedsuppression by Hoker (guidedby eitherHohermonitoring or EPS)isa reasonof itssuperiorinitiallongmonitoringin additionto suppression by EPS.Sotalal ableinitialstrategybecause andbetteracuteandlong-termtolerability was more effectivethan the other 6 antiorrhythmictermefficacy withsodiumchannelblockingdrugs. drugs,all classI agents,in preventing deathand recur- compared (AmJCardiol1996;78(suppl4A):34-40) renceof arrhythmia.Efficacy comparedwith placebo, however,wasnotevalualed.IntheEPSlimb,sotalol was

he Electrophysiologic Study versus Electrocardiographic Monitoring (ESVEM) study conT cluded that there was no significant difference between suppression of spontaneous ventricular arrhythmias seen on Helter monitoring and suppression of inducible ventricular arrhythmias seen by electrophysiologic study (EPS) in predicting the success of antiarrhythmic drug therapy in patients with sustained ventricular arrhythmias who also have both frequent ambient ectopy and inducible ventricular tachyarrhythmia.l The interpretation and clinical applicability of these results have been questioned because of several methodologic differences between the ESVEM protocol and those used in other invasive and noninvasive studies. Four issues must be addressed to answer the questions that have been raised: ( 1) the ability of EPS and Helter monitoring to predict outcome, (2) the specifics of the testing techniques, (3) the recurrence of arrhythmias versus survival, and (4) antiarrhythmic drug selection. From the Departmentof Clinical Medicine, Columbia University College of Physicians and Surgeons, and Clinical Electrophysiology Services, Columbia-PresbyterianMedical Center, New York, New York. Address for reprints: James A. Reiffel, MD, Clinical Electrophysiolagy Services, Columbia-Presbyterian Medical Center, 161 Fort Washington Avenue, Suite 349, New York, New York 10032.

34

?31996

by Excerpta Medica,

All rights resewed.

Inc.

PREDICTIVE ABILITIES OF ELECTROPHYSIOLOGIC STUDY AND HOLTER MONITORING The therapy of ventricular arrhythmias can be

directed either by Helter monitoring, which noninvasively examines the trigger (ambient ectopy), or by EPS, which invasively examines the propensity for sustaining an arrhythmia. ESVEM was conducted to determine the accuracy of these 2 methods in predicting the efficacy of drug treatment for ventriculartachycardia/ventricularfibrillation(VT/VF). The trial design of ESVEM is shown in Figure 1.2Patients with sustainedarrhythmias,whether in the form of VT, cardiac arrest, or syncope, were randomized either to an EPS limb or a Helter monitoring(in conjunctionwith exercise testing) limb following documentation of =10 premature ventricular complexes per hour on Helter monitoringand reproduciblyinducible VT/VF at EPS prior to drug treatment.2After randomization to Helter monitoring or EPS-guided therapy, the patientsunderwentserialtestingof up to 6 antiarrhythmic drugs in a random order (Figure 1) until 1 or none was predicted to be effective.ESVEM was the first opportunity to compare prospectively the efficacy, safety, and tolerabilityof a variety of antiarrhythmicdrugs in the same patient population. 0002-9149/96/$1 Pll S0002-9149(96)00451-1

5.00

Drugs: FIGURE 1.Des@ of ESVEM trialand cmtiarrhythmic drugsused.lmipramine,mexiletine, procainamide, quin-dine, andsolalalwereusediniIiaily.1.rSe4r, imipmmine waseliminatd andpirrnenol andpmpafenone wereadded.+ = Haltermonitoring obtained forconcordance/discordanceanafysis butnotusedfordrug development; CA = cardiac arrest; S*; ~ = EM . &kQphysia!OgiC exercisetolerancetat; HM = Helter monkring;PVC= premature ventricularcontraction; VT= ventricular from CimAraardia. (Adapted .

FIGURE 2. Disposition of ESVEM pm tients fromscreening throughmndomizatian. (Adapted fromCida~ti.2)

Class 1A Imipramine Pirmenol Procainamide Quinidine

I

HM: 210 PVCs/h EPS: Inducibility Randomize

Class IB Mexiletine

EPS limb HM limb I Drug 1 I

I Drug 1 I Drug N I

Class IC/11 Propafenone Class 111/11 Sotalol

DrugN I ETT

EIT

@mE ‘I+! I’-* Screened N=2103

I

Approximately a quarter of the 2,103 patients screened for enrollment were randomized (Figure 2).2 Most patients were excluded for medical rea~ons or failure to meet enrollment criteria; some refused enrollment. Since only patients with an average of at least 10 premature ventricular complexes (PVC5) per hour dr.u@g48 hours of Helter monitoring and reproducibly inducible ventricular tachyarrhythinias at EPS were eligible for randomization, any conclusions concerning predictive accuracy derived from ESVEM might best be restricted to this patient population. For each drug, the incidence of efficacy predictions was greater in the Helter monitoring limb than in the EPS limb (Table I).1 This result is similar to findings in a much smaller Canadian trial.3 There was no significant difference, however, in the actuarial probabilities of arrhythmia recurrence or any of the death rates between the Helter monitoring and EPS groups.l The low average yearly arrhythmia death rates (5%/year overall, and approximately 3% with sotalol) indicate that drug therapy for VT/VF remains a reasonable strategy.

Randomized N=486

ESVEM has been criticized on the basis of its high arrhythmia recurrence rates (Table H). Critics have used these rates to suggest ESVEM methods were flawed. Table II shows the summary of event rates for the 296 patients who received predictions of efficacy among the 486 patients randomly assigned to the 2 study groups. The mean actuarial recurrence rates of arrhythmia were 37%, 49Y0,61Y0, and 6690 at 1, 2, 3, and 4 years, respectively. However, the 5%/year arrhythmia death rates are not significantly different from those in other VT/VF series.4 Further, there was no difference between the actuarial sudden cardiac death rate in ESVEM and that of another major contemporary trial, the Cardiac Arrest in Seattle: Conventional versus Amiodarone Drug Evaluation (CASCADE) study.5 Sudden cardiac death rates in the CASCADE study were 13% and 1970in amiodarone and conventionally treated groups, respectively, compared with 11%, 1770, 22%, and 23% in years 1; 2, 3, and 4 of ESVEM. Criticism of arrhythmia recurrence rates without consideration of the survival data is probably unwarranted, because both event rates (i.e., sudden car-

A SYMPOSIUM:

PHARMACOTHERAPY

IN CARDIAC ARRHWHMIAS

35

I TAMEI Rcwd+scsfEfficcrcyTes+ingof IndividucslDrugs

\

HM Limb (%)

EPS Limb (%)

Sotalol

56

35

43

Mexiktine

67

12

36

Quinidine

59

16

33

Pirmenol

55

19

32

Procainamide

50

26

34

Propafenone

48

14

28

Imipramirre

45

10

21

Drug

EPS=efectrophysiolqic

Overall

(%)

study; HM=Hokerm onitoting.

Adopted from NEnglJMad.’3

TABLE II Comparison of Actuarial Probabilities of Arrhythmia Recurrenceand Death for All Patients*

I Year

1

2

3

4

37 * 3 10 t 2

49 + 3 13 ~ 3

61 * 4 18 f 3

66 ~ 4 20 + 4

cardiac death (%]

11 * 2

All-cause death (%)

12 ~ 2

17 z 3 1 Er ~ 2

22 & 3 24 ~ 4

23 * d 28 ~ 4

Endpoint Arrhythmia

recurrence

Arrhyihmic

death (%)

(%)

“ A total of 296sublectswith efficacy predicted, censored ondrug discontinuation. Adapted from N

Errg/JMad.l

diac death and arrhythmia recurrence) would be expected to be high if the ESVEM methods were faulty. Reasons for the high arrhythmia recurrence rates other than methodologic are more likely and are discussed in detail elsewhere.b SPECIFICS

OF

TESTING

TECHNIQUES

A common criticism of ESVEM relates to the testing techniques that were used. It has been suggested that different results might have been obtained if different Helter monitoring or EPS criteria had been used for the definition of drug efficacy. Specifically, it has been hypothesized that more rigid efficacy criteria would have reduced the rate of arrhythmia recurrence and therefore increased predictive accuracy. Additionally, the stimulation protocol used in the EPS arm has been specifically criticized as being insufficiently aggressive because triple premature extrastimuli were used only selectively during drug testing (e.g., when required for induction of VT/VF at baseline). Another concern was that the efficacy criterion of =15 induced beats was too lenient, leading to exaggerated predictions of efficacy. These concerns can be addressed by retrospective reanalysis of the ESVEM data. The original Helter monitoring criteria for defining efficacy in ESVEM were a reduction of total premature ventricular complexes by 70%, pairs by 80%, nonsustained VT runs of 3–15 beats by 90%, and runs of VT >15 beats by 100%. A sampling of several recent trials shows that, with the exception of the Cardiac Arrhythmia Suppression Trial (CAST),7 ESVEM criteria were a little less stringent. The other trials required 100% reduction of VT of =3 beats, and most of them required a 90% reduction of pairs. 36

THE AMERICAN JOURNAL OF CARDIO1OGYC$

Accordingly, a retrospective analysis of the ESVEM datag was performed to assess the predictive accuracy of EPS versus Helter monitoring using 2 sets of more rigid Helter monitoring criteria for defining drug efficacy. The first set of new criteria was composed of a reduction of total premature ventricular complexes by 70%, pairs by 80%, and all VTS of =3 beats by 100%. The second set of new criteria was composed of a reduction of total premature ventricular complexes by 80%, pairs by 90%, and all VTS of =3 beats by 100%. The incidence of drug efficacy predictions @ the original ESVEM analysis was 45% ( 108 of 242) in the EPS group and 77% (188 of 244) in the Helter monitoring group. Using the first set of new criteria, the incidence of drug efficacy predictions was 59% ( 144 of 244), whereas with the second set of new criteria, it was 50!Z0( 122 of 244). Cumulative arrhythmia recurrence after up to 6 years of follow-up was no different for either the first set (p = 0.78) or the second set (p = 0.84) compared with patients in whom drug efficacy had been defined by noninducibility at EPS. Additionally, the frequency of nonsustained VT events per day prior to drag therapy (infrequent [1-5] versus more frequent [>5]) did not predict arrhythmia recurrence on therapy. It can be concluded from these new analyses that efficacy criteria specifying a greater degree of ectopy suppression on Helter monitoring than the original ESVEM criteria reduces the number of patients who get an efficacy prediction but does not improve the predictive accuracy of Helter monitoring. During the baseline assessment in the EPS limb of the ESVEM trial, single and double premature extrastimuli were used during sinus rhythm; single and double premature extrastimuli were used at the right ventricular apex during ventricular pacing; single and double premature extrastimuli were used at the outflow tract during ventricular pacing; and, if no sustained arrhythmia occurred, triple premature extrastimuli were introduced, first at the outflow tract and then, if needed, at the apex. The protocol required 3 drive cycle lengths. The stimulation sequence was terminated whenever a target tachyarrhythmia was reproducibly induced or on completion of all protocol steps without tachycardia induction. During drug assessment, all patients were tested with a minimum of 2 extrastimuli at the apex at all 3 drive cycles. If more aggressive stimulation was required at baseline to induce tachyarrhythmia, that same step in the stimulation protocol was reached, if necessary, during all drug testing.z Consequently, for some patients, the on-drug testing was identical to that before drug, and, for some, the on-drug testing was more aggressive. Reproducibility was required at baseline, and no patient was enrolled if a nonreproducible arrhythmia occurred. Coupling intervals were decremented by 10 msec, and each was tested twice. This protocol is different from many reported in the literature. Many laboratories now routinely use single, double, and triple premature extrastimuli to assess drug efficacy.4 Although there has been no

VOL. 78 (4A)

AUGUST

29,

1996

direct prospective comparison of outcome when drug efficacy was predicted with double versus triple premature extrastimuli, many investigators empirically support triple premature extrastimuli. Yet, Swerdlow et alg showed retrospectively that triple premature extrastimuli on-drug in patients who required only double or single premature extrastimuli to induce an arrhythmia at baseline reduced the proportion of efficacy without improving p~dictive accuracy. To address this point, new analyses’” were done among the patients with efficacy predictions in the EPS limb of ESVEM to evaluate the following: . Reproducibility of pacing protocols for induction at baseline given the nature of the stimulation sequence ● Predictive accuracy categorized by the numbers of induced beats on-drug, comparing O–2 beats with 3– 15 beats and comparing O–5 beats with 6–10 beats and >11– 15 beats . Predictive accuracy with triple premature extrastimuli versus <3 premature, extrastimuli used during on-drug EPS and with identical versus more aggressive stimulation protocols on-drug than were required for VT/VF at baseline EPS. For patients induced with single premature stimuli at baseline or with double premature extrastimuli at only 1 or 2 drive cycles, the drug study was always more aggressive (minimum of double premature extrastirnuli at the right ventricular apex at all 3 drive cycles). With respect to reproducibility of induction, in those patients whose initial VT/VF was induced by single or double premature extrastimuli, the second induction in ESVEM required triple premature extrastimuli in only 6~0 using the ESVEM stimulation protocol. This contrasts with reports such as that of Cooper et al, ” in which, using an alternative protocol, triple premature extrastimuli were needed 35?10 of the time. Historically, this experience led to the common use of triple premature extrastimuli. In ESVEM, in those patients whose initiated VT/VF was induced at the apex at baseline, only 390 required a second site for induction. Consequently, the ESVEM protocol has to be followed specifically to use or contrast the results reached. The use of 10 msec decrements, =2 trials at each coupling interval, and double premature extrastimuli at 2 sites and 3 drives prior to 3 premature extrastimuli may increase the yield and reproducibility and reduce the increments of yield from a second site or a third premature stimulus. Thus, empiric extrapolation of ESVEM results to non-ESVEM stimulation protocols should be avoided. There was no difference in actuarial arrhythmia recurrence in 6 years of follow-up for any of the new analyses requiring more stringent efficacy criteria, that is, for patients with greater reduction in the maximum number of induced beats, for patients tested with triple versus double (the minimum used) premature extrastimuli on-drug, or for patients tested with more aggressive protocols on-drug than were necessary prior to drug. It can be concluded from

these new analyses that, when using the ESVEM protocol, efficacy does not vary with the number of induced VT beats s 15; outcome in patients using triple premature extrastimuli is no different than in patients tested with double premature extrastimuli; and efficacy definitions using more premature extrastimuli or drive cycles on-drug than were required for induction at baseline do not improve predictive accuracy. COMBINED EPS AND MONITORING

HOLTER

There was no difference in the ability of Helter monitoring and EPS to predict the success of drug therapy in ESVEM. Hypothetically, event rates and long-term outcome might be expected to be better if a drug met both Helter monitoring and EPS efficacy criteria. Some of the factors that influence VT/VF (e.g., ischemia, drugs, electrolyte status, hemodynamic issues, circadian variation, and autonomic tone) can only be assessed by looking at ectopy during representative periods (e.g., by monitoring during altered autonomic tone or during periods of ischernia). To assess Helter monitoring and EPS concordance/discordance and to examine whether combined use of EPS and Helter monitoring improves prediction of drug efficacy, 24-hour Helter monitoring data were obtained from patients randomized to EPS-guided therapy at the time of the first drug trial (first Helter monitoring) and at the time of effective drug trial, in which efficacy was determined by EPS (effective Helter monitoring).” These Helter monitoring data were analyzed after completion of the ESVEM trial and were thus not used to alter drug selection or prediction. First Helter monitoring data were available in 6590of patients who had at least 1 EPS test, and effective Helter monitoring data were available in 93~0 of patients with EPS-declared efficacy. The results are reported in detail elsewhere.g Of 98 EPS-noninducible patients, 67% also had Helter monitoring suppression. The ratio of concordance/ discordance was not influenced by the antiarrhythmic agent being evaluated, and the groups were clinically similar. There was no significant (p = 0.98) outcome difference inpatients with concordant efficacy predictions between EPS and Helter monitoring (n = 66) compared with those with discordant tests, that is, efficacy predicted by EPS but efficacy criteria not met by Helter monitoring (n = 32). In addition, comparison of patients with concordant tests with patients in whom efficacy was predicted by Helter monitoring only (n = 188) revealed no outcome differences (p = 0.48). It can be concluded that there is a significant discordance between assessment by EPS and Helter monitoring criteria; requiring Helter monitoring and EPS concordance reduces the number of patients with an efficacy prediction from 45% to 2970;and simultaneous suppression of both spontaneous and inducible ventricular arrhythmias does not identify a group with better

A SYMPOSIUM:

PHARMACOTHERAPY

IN CARDIAC ARRHYTHMIAS

37

TABI.E III Comparison of Sotalol with Other Antiarrhythmic Agents in Suppressing Ventricular Tachycardia No.

Test Drug by

Patients Percentage

Not Inducible

class

with

Effective Test

[%)

Class 1A Quinidine

16 (1)

6

Pracainamide

31 (6)

19

Lidacaine

17 (4)

23

Mexiletine

11 [2]

18

Class IB

Class IC Fkainide

8 (1)

12

Encoinide

7 (1)

Recainam

6 (1)

Propofenone

6 (0)

1A 17 0

Class Ill Sotolol

.33 [15]

Amiodarone

54 [21]

Reprinted with permission from AmJ

45 39

Cardicd.14

TABLE IV krcidenceof Adverse Events Requiring Discontinuation of Study Drug During Long-termFollow-up (ESVEM Trial) Drug

No.

CV (%)

CNS (%)

GI (%)

Total (%)

Imipramine

15

7

o

7

13

Mexiletine

58

2

9

16

19

Pirmenol procainamide

27

7

0

39

3

8

0 13

31

7

Propafenone

45

11

2

4

13

Quinidine

38

8

16

26

32

Sotalol

85

1

0

7

0.012

<0.001

P Value

6 0.514

0.003

CNS = central nervoussystem;CV = cardiovascular; ESVEM = Electrophysiologic Study versus ElectrocrsrdicgraphicMonitoring; GI = gastrointestinal. Adapted from N Engl-1Med. ’3

outcome. An alternative interpretation of the ESVEM data could be that neither Helter monitoring nor EPS suppression relate to outcome, i.e., that suppression of any kind is completely irrelevant, as may have been evident in the Cardiac Arrhythmia Suppression Trial (CAST), Congestive Heart Failure Survival Trial with Antiarrhyihmic Therapy (CHFSTAT), and other ventricular premature depolarization trials. DRUG The

S~LECTION

ESVEM results indicate that sotalol should be considered a first-line drug for treatment of VT/ VF.13In the EPS limb, sotalol was statistically more likely to achieve an efficacy prediction than any of the sodium channel blockers. Although there was no difference between the drugs in achieving an efficacy prediction in the Helter monitoring group, sotalol was more effective than the other 6 antiarrhythmic drugs in preventing death and recurrence of arrhythmia following an efficacy prediction in both the EPS and Helter monitoring limbs in up to 6 years of follow-up. There were no differences in clinical characteristics between patients given sotalol and those given class I drhgs. The sotalol versus class I recur38

THE AMERICAN JOURNAL OF Cardiology@

rence and survivor benefit was greatest in patients with a left ventricular ejection fraction ,<30%. Results from other studies support ESVEM conclusions that treatment with sotalol is a reasonable initial strategy. A retrospective Veterans Affairs administration review of the results of drug testing for the suppression of inducible VT by programmed electrical stimulation showed sotalol had the highest success rate of all the antiarrhythmic agents tested (Table 111).’4The Refractory Arrhythmia Study (RAS) 15evaluated the efficacy and safety of sotalol in 481 patients with drhg-refractory sustained ventricular tachyarrhythmias, all of whom had failed at least 3 previous trials of antiarrhythmic therapy. The prediction of drug efficacy with sotalol in the EPS groups in RAS and ESVEM were identical (35%), and the Helter monitoring efficacy prediction rate was 39% in RAS. Despite these patients’ refractoriness to prior antiarrhythmic drugs, long-term sudden death rates were only around 4% (17 of 481) per year with sotalol in the RAS group of sotalol responders. Again, this is similar to ESVEM, in which the rates were around 3% per year over 4 years of follow-up. A Miinster Study16evaluated patients with clinical and inducible sustained VT/VF. In this study, sotalol suppressed 38% (151 of 396) of the patients, and sudden cardiac death event rate was 3% per year. The adverse effects of antiarrhythmic agents in the long-term segment of the ESVEM trial are shown in Table IV.13Fewer patients discontinued using sotalol than all the sodium channel blockers during dose titration except pirmenol, and most of those patients who discontinued sotalol did so for cardiac events and not for noncardiac organ toxicity. Early proarrhythmic rates for sotalol, procainamide, and quinidine were similar. Long-term, only 7% of patients discontinued sotalol in contrast to 13?%for propafenone, 19% for mexiletine, 3190 for procainamide, and 32~o for quinidine. Sotalol, therefore, appears to be both more effective and better tolerated than class I antiarrhythrnic drugs, making it an agent of choice. Notably, however, without placebo, empiric, or pure ~-blocker limbs in any of these studies, the precise advantage of sotalol and/or guided therapy remains incompletely elucidated. Gender differences should also be considered when determining the safety of drug treatment. There have been several reports suggesting that women are more prone than men to developing torsades de pointes with potassium channel blocking drugs, such as sotalol and class IA agents. Retrospective analysis of the ESVEM database supports this gender difference. Table V shows that 2 of 60 women compared with 8 of 426 men had torsades de pointes, although the difference was not statistically significant (p = 0.356). Similar trends were seen with QT prolongation (p = 0.175). The effectiveness of amiodarone could not be evaluated in ESVEM. Amiodarone was not used in ESVEM because the drug could not be randomized. Other trials suggest, however, that the safety and ef-

VOL. 78 [4A)

AUGUST

29,

1996

TAMEV Gender Differences in Incidenceof Torsades de Pointes and Prolonged QT Interval (ESVEM Imipromine

Mexiletine

118

203

11

23

Male

o

0

Female

o

0

Total receiving

Procainamide

Triol)

I

Quinidine

Sotalol

Propafenone

Pirmerd

Total

Gender

140

199

19A

98

426

88

17

35

26

11

60

12

0

1

0

0

5 2

0,

:

:

20

0

A

9

75

3

0

0

0

25

[%]

drug

Male Femakr

140 18

Torsades de pointes

prolonged

0

80

QT

Male

o

0

1

23

Female

o

0

0

0

ESVEM = ElectrophysiologicStudy verws Electrocardiographic Monitoring,

ficacy of arniodarone may be roughly comparable to that of sotalol. The CASCADE tria15 evaluated amiodarone in survivors of out-of-hospital VF not associated with anew Q-wave myocardial infarction. Therapy was randomized to empiric treatment with amiodarone versus treatment with a wide variety of sodium blockers guided by EPS, Helter monitoring, or both (conventional therapy). These tests, however, were chosen by individual physicians rather than by a randomized protocol. The primary endpoints of the study were cardiac mortality, resuscitated cardiac arrest due to documented VF, or complete syncope followed by’ a shock from an implanted automatic defibrillator, The study demonstrated that arniodarone was more effective than conventional (class I) drugs in preventing recurrence of fatal or near-fatal arrhythmias. Sotalol was not used in CASCADE. If we examine CASCADE and ESVEM together, however, the total 4-year sudden death rates for amiodarone patients in CASCADE and the subgroup of patients with VF in ESVEM were very similar. The sotalol subgroup in ESVEM had the lowest rates, although the difference was not statistically significant. Supportively, a European study in 60 patients directly comparing sotalol and amiodarone also showed identical efficacy rates.17 To understand fully the relative merits of sotalol and amiodarone, one needs to compare them with the merits of impkmtable cardioverter-defibrillators (ICDS). For this purpose, a comparison can be made of all-cause mortality at 1 and 4 years for ICDS in a large multicenter intracardiac device tria118versus sotalol in ESVEM. Such comparisons are important in analyses of the cost of devices versus the cost of drugs. Sudden death rates are reduced by 1–270 per year by defibrillators, a rate significantly better than that achieved with drug therapy, but in long-term follow-up this difference disappears. All-cause mortality was similar in the 2 studies (Table VI). If total survival is no different in patients treated with an ICD and those treated with antiarrhythmic drugs, then quality of life, frequency of VT, likelihood of early sudden cardiac death, long-term prognosis, age, and comorbid conditions should all be considered in selecting between an ICD or a drug as initial therapy. The Antiarrhythmic Versus Implantable Defibrillator (AVID) trial,19which is a larger, direct comparison of defibrillators, amiodarone, and so-

TABf.E VI Comparison of All-Cause Mortality at 1 and 4 Years far Defibrillator Versus Sotalol

I

No.

I

876 296

MuhicenterICD* ESVEMT

1Year (%] 12 12

I

4 Years (%) 1

I

31 28

‘ Dota are from Circulation.l Q T Dat.J are from N

EnglJMed.’

ESVEM = Electrophysiolcgic Study versus Elec*ocordiographic Monitoring; ICD = implantabl.scardioverterdefibrillator.

talol, is currently underway. In this multicenter evaluation, patients with VF, sustained VT with syncope, or VT with serious hemodynamic consequences (e.g., systolic blood pressure <80 mm Hg, chest pain, or near syncope) and an ejection fraction =0.40 are being randomized to receive either device or antiarrhythmic drugs. Patients randomly assigned to drugs will be further randomized to receive either sotalol or amiodarone. Amiodarone will be administered empirically, whereas use of sotalol will be guided by Helter monitoring or EPS. CONCLUSIONS The ESVEM trial

clearly showed that in patients with life-threatening, inducible ventricular tachyarrhythmias and >10 PVCS per hour, those who used aniiarrhythmic drugs predicted to be effective by Helter monitoring or EPS had a reasonably low risk of arrhythmia death. In patients who achieve predictive accuracy by EPS, requiring Helter monitoring suppression in addition does not result in a further reduction of arrhythmia recurrence. In these same patients, therapy with sotalol is a reasonable initial strategy because of its better initial and long-term efficacy and better acute and long-term tolerability. Retrospective reanalysis of the ESVEM data using more rigid efficacy criteria (a greater degree of ectopy suppression for Helter monitoring and more stringent efficacy definitions in the stimulation protocol of EPS) supports these conclusions. ESVEM Helter monitoring versus EPS results maybe applicable only to patients with the combination of spontaneous, sustained ventricular tachyarrhythrrdas; inducible, sustained ventricular tachyarrhythmias; and frequent, spontaneous PVCS. The latter 2 characteristics were found in 6590of patients with ventricular

A SYMPOSIUM:

PHARMACOTHERAPY

IN CARDIAC ARRHYTHMIAS

39

tachyarrhythmias. Patients who do not meet the ESVEM population characteristics may have arrhythmias resulting from different mechanisms than those in ESVEM. The predictive accuracy of suppression criteria in such patients needs to be tested directly. It is also possible that the ESVEM results are protocol specific.4Finally, drug efficacy results in ESVEM maybe more reasonable to extrapolate widely than those of testing methods. The class I drug effects on survival in ESVEM are directionally in line with those noted in many other subsets of patients, including postmyocardial infarction patients (CAST), patients with atrial fibrillation (the Stroke Prevention in Atrial Fibrillation Study [SPAF]), and cardiac arrest patients treated with propafenone (Cardiac Arrest Study Hamburg [CASH]), The sotalol effects in ESVEM are directionally similar to those seen with/3 blockers in multiple postmyocardial infarction and some cardiac arrest (CASH) populations and with sotalol in contrast to class I drugs in an array of VT/VF patients (RAS, Veterans’ database). One might interpret this to suggest that the methods in the ESVEM trial should be of limited concern and that the real issue in efficacy and specificity is the kind of drug used; that is, in patients with heart disease, class I drugs have specificity for higher mortality and ~ blockers have specificity for higher survival, regardless of how therapy is guided. 1. Mason JW, for the ESVEM Investigators. A comparison of electrophysiologic testing with Helter monitoring to predict arztiarrhythmic-dmg efficacy for Engl J Med 1993;329:445-451. 2. The ESVEM Investigators. Determinants of predicted efficacy of antiarrhythmic dregs in the EPS versus electrocardiographic monitoring trial. Circulation 1993;87:323-329. 3. Mitchell LB, Duff HJ, Manyari DE, Wyse G. A randomized clinical trial of the noninvasive and invasive approaches to dmg therapy of ventriculm tachycardia. N Engl J Med 1987;317:1681– 1687. 4. Reiffel JA, Reiter MJ, Freedman RA, Mamr D, Huang SKS, Hahn E, Hartz V, Mason J, for the ESVEM Investigators. Irztlrrenceof Helter monitor and ventricular tachyamhythmias. N

40

THE AMIERICAN JOURNAL OF Cardiology@

electrophysiologic study methods and efficacy criteria on the outcome of patients with ventricular tachycardia and ventricular fibrillation in the ESVEM trial, Prog Cardiovasc Dis 1996;38:359-370. 5. The CASCADE Investigators. Randomized antiurrhythmic dmg therapy in survivors of cardiac wrest (the CASCADE study). AmJ Cardiol 1993;72:280– 287. 6. Mason JW, Marcus FJ, Bigger JT, Lazzara R, Mann D, Reiffel JA, Reiter MJ. A summary and assessment of the findings and conclusions of the ESVEM trial. Prog Cardiovasc Dis 1996;38:347–358. 7. Echt DS, Liebson PR, Mitchell LB, Peters RW, Obias-Maruro D, Barker AH, Arensberg D, Baker A, Friedman L, Greene L, Hutber ML, Richardson DW, and the CAST Investigators. Mortality and morbidity in patients receiving encainide, flecainide, or placebo. N .%gl J &fed 1991;324:781-788. 8. Reiffel J, Mann D, Reiter M, Freedman R, Huang SKS, Hahn E, Hartz V, for the ESVEM Investigators. A comparison of Helter suppression criteria for declaring dmg efficacy in patients with sustained ventricular tachyarrhythmias in the ESVEM trial. (Abstr.) JAm COUCardiol 1994;23:279. 9. Swerdfow CD, Winkle RA, Mason JW. Determinants of survival in patients with ventricular tacbyarrhythmias. N Engl J Med 1983;308:1436-1442. 10.Reiffel J, Reiter M, Freedman R, Mann D, Huang SKS, Hahn E, Haztz V, for the ESVEM Investigators. Did the mrorber of premature stimuli used or the lengtJr of unsustained tachycardia induced affect the predictive accuracy of the electrophysiologic study used to guide therapy in the ESVEM trial. (Abstr,) PACE 1994;17:826. 11. Cooper MJ, Hunt LJ, Palmer KJ, et al. Quantitationof day tu day variability in mude of induction of ventricular tucbyar’rhythmiasby programmed stimulation. JAm Coil Cardiol 1988;11:101–108. 12. Reiter MJ, Mann DE, Reiffel JA, Hahn E, Hartz V, for !lre ESVEM Investigators. Significance and incidence of concordance of drug efficacy predictions by Helter monitoring and electrophysiolugic study in tJre ESVEM trial. Circulation 1995;91:1988–1995. 13. Mason JW, for the ESVEM Investigators. A comparison of seven antiarrbythrnic dregs in patients with ventricular tachyarrhytbmias. N Engl J Med 1993;329:452-458, 14. Nademanee K, Singh BN. Effects of sotalol on ventricular tachycardia and fibrillation prnduced by programmed electrical stimulation: comparison with other antiarrbythmic agents. Am J Cardiol 1990;65(suPP1) :53A-57A. 15. Kehne RF, MacNeil DJ, Zheutlin TA, Ezri MD, Nazari J, Spangenberg RB, Dmmington C, Leuken M. Safety and efficacy of oral sotalol for sustained ventricular tachyarrhythrnias refractoV to other antiarrhytbmic agents. Am J Cardiol 1993;72(suPP1):56A-66A. 16. Haverkamp W, Chen X, Rotman B, Hindricks G, Willems E, Kottkanrp H, Breithardt G, Borgreffe M. Acute and long-term efficacy of sotalol in 396 patients with malignant ventricular tachyarrhytbmias. (Abstr.) PACE 1994;17:827. 17. Arniudarune versus sotalol study group. Multicenter randomized trial of sotalol versus amiodarone for chronic mrdignant ventricular tachyarrhythnias. Eur Heart J 1989;10:685-689. 18. Lehmanrr MH, Thomas A, Jackson K, Steirmran RT, Shah M, Schuger C, Meissner MD, Mosteller R. Long-term outcome with implantable crrrdioverter defibrillatnr(ICD) therapy inamu@enter investigator-edited database. (Abstr. 0659.) Circulation 1990;82(suppl 111):165. 19. Greene HL, Antiarrhythmic drugs verausimpkurtabledefibrillators: the need for a randomized controlled study. Am Heart J 1993;127:1171-1178.

VOL. 78 (4A)

AUGUST

29,

1996