Can the electrophysiologic study predict treatment outcome in patients with sustained ventricular tachyarrhythmias unrelated to coronary artery disease?

ELSEVIER SCIENCE IRELAND

International

Journal of Cardiology 45 (1994) 53-68

Can the electrophysiologic study predict treatment outcome in patients with sustained ventricular tachyarrhythmias unrelated to coronary artery disease? Lloyd M. Davis *, Mark J. Cooper, Norman Sadick, Karen Byth, John B. Uther, David A.B. Richards, David L. Ross Cardiology Unit. Westmead Hospital, Hawkesbury Road, W&mead,

N.S. W. 2145, Australia

(Received 15 December 1993; accepted 3 February 1994)

Abstract Sustained ventricular tachyarrhythmias unrelated to coronary artery disease are uncommon. Currently there are no clear guidelines to aid selection of the most appropriate treatment strategy. Therefore, factors potentially predictive of arrhythmia recurrence and death and the ability of the electrophysiologic study to predict treatment outcome in patients with spontaneous sustained ventricular tachyarrhythmias unrelated to coronary artery disease were examined in 41 medically treated patients followed for a median of 25 (range l-76) months. Examined factors were: syncope associated with the spontaneous arrhythmia, the morphology and cycle length of the presenting arrhythmia, underlying ventricular function, cardiac pathology, and the results of drug assessment at electrophysiologic study. Random variability in the ease of arrhythmia induction at electrophysiologic study was measured for the group as a whole and was allowed for in prediction of an effective drug response. The 95% confidence intervals for variability in the ease of repeat arrhythmia induction at the same study were 5 1 extrastimulus and for variability in the ease of repeat arrhythmia inductions at different studies were s 2 extrastimuli. Poisson regression models were used for data analysis. Arrhythmia recurrence was most likely in: (1) patients on treatment not predicted to be anti-arrhythmic at electrophysiologic study; (2) patients whose treatment was not assessable at electrophysiologic study because the arrhythmia was not reliably inducible; (3) patients with impaired ventricular function; and (4) re-entered patients whose arrhythmia had recurred on previously allocated therapy. The risk of arrhythmia recurrence decreased with time from hospital assessment. All five deaths occurred in patients with impaired ventricular function. Conclusions: drug efficacy should be tested at electrophysiologic study in patients with reproducibly inducible clinical arrhythmias. Treatment not proven to be anti-arrhythmic at electrophysiologic study is usually ineffective. Patients with ventricular dysfunction are at highest risk of death from arrhythmia recurrence and should be considered for an implantable defibrillator, arrhythmia surgery, or heart transplantation if drug treatment is not predicted to be effective or is not assessable at electrophysiologic study. Key words: Arrhythmia;

Electrophysiologic

study; Ventricular

tachycardia

* Corresponding author, at the following address: Departments of Paediatrics and Hematology, Room 1050, Children’s Hospital, I Children’s Place, St. Louis, Missouri, USA 63110. 0167-5273/94/$07.00 0 1994 Elsevier Science Ireland SSDI 0167-5273(94)02022-B

Ltd. All rights reserved.

54

1. Introduction Sustained ventricular tachycardia or fibrillation in patients with normal coronary arteries is uncommon [ 1,2]. Most reports on such patients have been based on small numbers of patients [3-111. Although Holter monitoring and drug testing by programmed stimulation have been used to predict response to therapy and help guide clinical decision making in patients with sustained ventricular arrhythmias unrelated to coronary artery disease, the efficacy of these tests for predicting long-term outcome is disputed [ 12,131. Previous studies have not controlled for the presence of potential confounding variables that also may determine patient outcome such as ventricular function [6,7,11, 14,151, anti-arrhythmic drugs, the cycle length and morphology of the ventricular tachyarrhythmia [14], the presence or absence of syncope associated with the presenting arrhythmia and the underlying cardiac pathology. Previous studies also have not allowed for spontaneous variability in the ease of arrhythmia induction when assessing response to drug therapy at electrophysiologic study. Consequently there are no clear criteria to guide physicians and patients in choosing the most appropriate form of therapy. In this study we aimed to identify which factors predict long-term outcome in patients with a history of sustained ventricular tachyarrhythmias and normal coronary arteries and assessed the value of antiarrhythmic drug testing by programmed stimulation for predicting the outcome of drug therapy.

L.M.

Davis et ~1. /Int.

J. Curdid.

45 (1994)

53-68

an arrhythmia lasting ten or more seconds in duration. Coronary artery disease was considered to be present if there was a 50”/0 or more stenosis of a coronary artery. Patients were excluded if they were older than 75 years at initial presentation or if the ventricular tachyarrhythmia was related to the long QT syndrome or an acute ischaemic or metabolic event. 2.2. Assessment lar function

of coronary anatomy

and ventricu-

Cardiac catheterization was performed in all but three patients to confirm normal coronary anatomy and to assess left ventricular function in 38 cases and right ventricular function in eight cases. Three patients did not undergo cardiac catheterization and were assumed to have normal coronary anatomy. Two were aged 18 and 34 years and had no prior history of infarction, a normal resting electrocardiogram and normal ventricular function on gated heart blood pool scan and echocardiography. The other patient was a 44-year-old male with no history of ischaemia, and a normal resting and exercise electrocardiogram. Ventricular function also was assessed by radionuclide labeled gated heart pool scan in 39 patients and in 18 cases by 2D-echocardiography. Ventricular dysfunction was defined as the presence of any abnormality in wall motion assessed at either ventriculography, radionuclide scan or ‘D-echocardiogram. 2.3. Electrophysiologic

assessment

2. Methods 2.1. Patient selection We reviewed all 41 medically treated patients managed at Westmead Hospital between 1981 and 1989 with a history of documented spontaneous sustained ventricular tachycardia or fibrillation unrelated to coronary artery disease. Patients seen after 1989. when automatic implantable defibrillators became readily available, were excluded from the study to prevent the effects of insertion of automatic defibrillators from influencing the assessment of drug treatment. Sustained ventricular tachycardia or fibrillation in this study refers to

All patients underwent electrophysiologic study using a standard protocol. This involved insertion of two quadpolar catheters to the high right atria1 appendage and the right ventricular apex and a tripolar catheter to the His bundle region. Ventricular electrical stability was assessed by programmed stimulation at a single site, the apex of the right ventricle, at twice diastolic threshold. Extrastimuli were introduced after a stable eightbeat drive train as close to 600 ms in cycle length as possible. There was a 3-s pause between successive drive trains. Delivery of extrastimuli was commenced at an initial coupling interval of 300 ms, which was then decreased in 10 ms steps until

ventricular refractoriness was encountered or a sustained ventricular tachyarrhythmia was induced. Each extrastimulus coupling interval was repeated three times if no tachycardia was induced before decrementing by 10 ms. When the right ventricular effective refractory period was reached, the extrastimulus coupling interval was increased by IO-ms increments until stable ventricular capture and a further extrastimulus was introduced. The end point of stimulation in the baseline state was either the induction of sustained ventricular tachycardia or fibrillation, or refractoriness of a fourth extrastimuius. If the clinical arrhythmia was not induced by this protocol, programmed stimulation was repeated with a shorter drive train cycle length during infusion of isoproterenol (S- 10 ~~rnin) which invariably shortened the sinus cycle length to at least 500 ms. At the baseline study, the full stimulation protocol (i.e. beginning again with a single extrastimulus) was repeated three times in I9 (49%) cases and twice in five (13%) cases to assess the reproducibility and the variability in ease of ventricular ta~hyarrhythmia induction in the absence of anti-arrhythmic drugs. Isoprenaline infusion was used for the repeat inductions in the five patients who had required isoprenaIine for the initial inductions of their clinical arrhythmias, In one case reproducibility of arrhythmia induction also was assessed by programmed electrical stimulation at the right ventricular outflow tract. The reproducibility and variability in ease of arrhythmia induction were not assessed at the same study in 15 (38%) cases. These cases were studied early in our experience before the full importance of variability in arrhythmia induction was realized and allowed for. Wowever all these patients had further off-drug arrhythmia inductions carried out as a baseline at subsequent intravenous drug studies that were included in our assessment of the between-study variability in ease of arrhythmia induction. 2.4. Assessment a~re~r~duc~~ility md ~ar~ab~~~ty in ease of induction oj‘ the c&&al arr~yt~~lia a&efectrop~ysio~o~icstudy Reproducible arrhythmia induction was defined as the ability to induce the clinical arrhythmia or

monomurphic ventricular tachycardia in more than one of the off-drug inductions using the above protocol. Arrhythmia inductions were regarded as non-reproducible if the clinical arrhythmia was initiated on fewer than two occasions or if a ventricular ta~hyarrhythmia was not always initiated by the stimulation protocol. In patients whose clinical arrhythmia was reproducibly inducible, ‘within study’ variability in the ease of arrhythmia induction was measured by the change in the number of extrastimuli required for repeat arrhythmia induction at the same study in the absence of anti-arrhythmia drugs. Similarly, between-study variability was calcufated by comparing the number of stimuli required at the first baseline induction with the number of extrastimuJi required at further baseline inductions at different studies on different days. Using this information, we defined an anti-arrhythmia effect as the requirement for a larger number of extrastimuli after drug administration than the 95Y6 confidence limit for repeat arrhythmia induction in the absence of an anti-arrhythmic agent. Drug therapy was then classified as predicted antiarrhythmic or not anti-arrhythmic using the above de~nition based on within-study variability and the results of intravenous drug testing at electrophysiologic study. In a similar way, based on the 95% confidence limits for between-study variability in ease of arrhythmia induction, the results of oral drug testing also could be classified as anti-arrhythmic or not. However, because there was much greater spontaneous variability in the number of extrastimuli required to induce the clinical arrhythmia in the absence of anti-arrhythmic drugs between studies on different days, oral drug testing was a very insensitive method for predicting drug efficacy. Only four patients were predicted to be effective based on the results of oral drug testing. This number was too small for meaningful analysis. Therefore this study concentrated on the results of intravenous drug testing. Therapy was classified as not assessable if the spontaneous clinical arrhythmia had not been reproducibly inducible at electrophysiologic study. 2.9. ~t~t~~t~cui analysis The statistical software packages SPIDA and

56

GLIM 3.77 were used to analyze the survival data [ 16,171. The following variables were considered: (1) ventricular function, (2) assessment of drug therapy at electrophysiologic study, (3) drug therapy, (4) underlying cardiac pathology, (5) arrhythmia morphology and cycle length, and (6) the presence or absence of syncope associated with the presenting arrhythmia. Because amiodarone takes 6-8 weeks to manifest its chronic electrophysiologic effects [ 181, the results of intravenous amiodarone tests were excluded in the analysis of the outcome of intravenous drug testing at electrophysiologic study. Poisson errors and the log link function in GLIM 3.77 were used to estimate relative risks of arrhythmia recurrence and death. For practical purposes these estimates are very similar to those obtained using Cox regression models with time-dependent covariates. In the Poisson regression model, the probability of arrhythmia recurrence or death in any given month is in effect the outcome variable. The relationship between this variable and those described earlier, together with duration of follow-up and the number of previous arrhythmia recurrences was examined. Because the data were considered on a month by month basis it was not necessary to remove patients from the study when their treatment status changed or when they experienced an arrhythmia recurrence. In this way the full data set could be used and the power of the study increased compared with the traditional method of analysis. Traditional data analyses (Log Rank tests and Kaplan Meier curves) lack statistical power in groups of this size and were not used.

L.M.

Davis et ul. /ht.

J. Curdiol.

45

( 1994)

53-68

chi-squared test was used to compare the variability in the ease of induction of sustained ventricular arrhythmias in patients without coronary artery disease, with that in patients with sustained ventricular tachyarrhythmias due to coronary artery disease which we have reported previously [ 191. Fisher’s exact test was used to examine factors associated with patient death. 3. Results 3.1. Patient characteristics The 41 patients studied had a mean age of 46 f 17 years (median 44 years, range 14-72 years). Eleven patients were female and 30 were male. Follow-up was for a mean of 28 + 21 months (median 25 months, range l-76 months). Of the 41 patients studied, 23 suffered syncope associated with their presenting arrhythmia and 18 did not (Table 1). 3.2. Presenting arrhythmia The presenting arrhythmia was ventricular tachycardia in 34 cases and ventricular fibrillation in seven cases. The morphology of the ventricular tachycardia was right bundle branch block in seven cases, left bundle branch block in 12 cases, indeterminate bundle branch block in three cases. In 12 cases, ventricular tachycardia morphology was not accurately assessable because the only documentation was on a rhythm strip. The cycle lengths of the presenting arrhythmias are shown in Table 1.

2.10. Other analyses 2.5. Drug assessment at electrophysiologic study One-way analysis of variance was used to calculate the standard errors in the number of extrastimuli required for repeat inductions of ventricular tachycardia or fibrillation at the same study and at different studies in patients with reproducibly inducible clinical arrhythmias. These standard errors were used to calculate the 95% confidence intervals for the difference in number of extrastimuli required for a subsequent induction of a ventricular tachyarrhythmia at the same study and at a different study done on another day. The

Serial drug testing was performed during subsequent studies in the patients with reproducibly inducible clinical arrhythmias. This involved a baseline induction of ventricular tachycardia or fibrillation using the above protocol, including isoproterenol if it had been necessary in the original electrophysiologic study, then infusion of an ‘anti-arrhythmic’ agent and then repeat stimulation after the drug, starting again at the beginning of the above same protocol. The end point of stim-

L.M. Davis et al. /hr. Table 1 Patient characteristics

J. Cardiol. 45 (1994) 53-68

compared

57

to outcome Arrhythmia

recurrences

Initial therapy

Altered

Patients

Additional trials of therapy

Patients having an arrhythmia recurrence

Total no. of arrhythmia recurrences

therapy

Deaths

Patients having an arrhythmia recurrence

Intravenous drug assessment

Anti-arrhythmic Not anti-arrhythmic Not assessable Not tested Assessable but allocated amiodarone

8 10 12 3 8

0

0

0

8

1

3

10 4 2

20 11 5

0

Oral drug assessment 0

0

I

5 9

18 3

4 0

8 12

2

Syncope 23 No syncope 18 Cycle length of the presenting arrhythmia <200 ms 7 20 l-250 ms 6 251-350 ms 19 > 350 ms 9

Anti-arrhythmic Not anti-arrhythmic Not tested Ventricular

I 2

4

I 9 13

12 2

4 2

12 9

0

I 9

0

7

4

2 20

2 3

6 7

12 12

4 6

18 25

3 2

I 8 14 20

0

4 21 5

1 I

function

Normal function Right ventricular dysfunction only Left ventricular dysfunction only Biventricular dysfunction

0

Symptoms of the presenting arrhythmia

Morphology

1 4 9 10

0

4

of the presenting arrhythmia

Left bundle branch block Right bundle branch block Indeterminate bundle branch block Ventricular fibrillation Not classifiable (on rhythm strip

12 7 3

2 5 1

I 8 2

7 I2

4

14 7

3 3

3 4

1

13 14 9

2 5

2 5

0

I

8 5

3 2

10 5

2 0

0

10

4

2 2 24

0

0

2 1

1

only) Cardiac pathology

Apparently normal heart Idiopathic dilated cardiomyopathy Alcoholic cardiomyopathy Congenital heart disease Other

5 3 13

I 5

ulation after drug administration was the induction of sustained ventricular tachycardia or fibrillation, or refractoriness of a seventh extrastimulus. Seven extrastimuli were the maximum

2

our stimulator could deliver. This upper limit to programmed stimulation after drug administration was chosen to ensure that true drug effects outside the range of normal variability in the number of

58

extrastimuli required for induction of ventricular tachycardia in the baseline state could be measured. One to three drugs were tested in any given patient. The drugs used for testing changed during the 9 years of the study as evidence became available of the relative inefficacy of some of the tested agents in similar patients with ventricular tachycardia based on coronary artery disease. The intravenous drugs tested were sotalol 1.5 mg/kg in 22 patients, amiodarone 10 mg/kg in 19 patients, metoprolol 0.3 mg/kg in seven patients, mexiletine 5 mg/kg in three patients, quinidine 10 mg/kg in two patients and verapamil 10 mg/kg in one patient. A drug free period of at least 6 days was allowed to elapse between drug tests. Intravenous amiodarone was always tested last. 2.6. Treatment

allocation

For treatment allocation during the course of the study, anti-arrhythmic was prospectively defined as the requirement for two or more extrastimuli after drug administration above that needed in the baseline (off-drug) state for the induction of a sustained ventricular tachyarrhythmia. The drug predicted to be most anti-arrhythmic at electrophysiologic study was commenced as oral therapy. If no drug was predicted effective, the patient was allocated to one of the tested drugs according to physician preference. 2.7. Assessment

of oral drug therap?

Patients returned between 7 days and 6 weeks after starting drug therapy for a further electrophysiologic study, during oral drug therapy, at which two further inductions of sustained ventricular tachycardia or fibrillation were performed using the same end points as after intravenous drug testing. 2.8. Follow-up Follow-up was by telephone contact with the patient, their relatives and attending physicians. Telephone contact was attempted at regular 6month intervals after initial hospital assessment. Arrhythmia recurrence was defined as an elec-

trocardiographically documented episode of ventricular tachycardia or fibrillation lasting at least 10 s or resulting in hemodynamic collapse requiring cardioversion. Sudden death was treated as an arrhythmia recurrence and was defined as instantaneous death occurring without other obvious cause. 3.3. Ventricular function Ventricular function as assessed by ventriculography, or gated heart blood pool scan, or echocardiogram was normal in 18 (44%) of the 41 patients. Three patients (7%) had only right ventricular dysfunction, eight (20%) had only left ventricular dysfunction, and 12 (29%) had biventricular dysfunction. The mean left ventricular ejection fraction in those with normal ventricular function was 63 f 5%. in those with right ventricular dysfunction was 60 f 7%, in those with left ventricular dysfunction was 39 f 15% and in those with biventricular dysfunction was 39 f 19%. 3.4. Underlying etiology Of the 41 patients studied, 13 had normal ventricular function and no detectable underlying etiology. Nine patients had idiopathic dilated cardiomyopathy, live patients had congenital or valvular heart disease, four patients had a history of ethanol abuse, three patients had a right ventricular cardiomyopathy and two patients had mitral valve prolapse. In one case each, the patient had hypertrophic cardiomyopathy, sarcoidosis, polyarteritis nodosa, rheumatoid arthritis and myotonia dystrophica. 3.5. Electrophysiologic

studies

Electrophysiologic studies were performed in all cases. The clinical arrhythmia was not reproducibly inducible in 12 cases. In the 29 patients in whom the clinical arrhythmia was reproducibly inducible, 24 in the off-drug state and five during induction arrhythmia infusion, isoprenaline required 5 2 extrastimuli in 54% of the 113 arrhythmia inductions, three extrastimuli in 36%

L.M. Davis et al. /ht.

of inductions ductions.

59

J. Cardiol. 45 (1994) 53-68

and four extrastimuli

in 10% of in-

3.6. Variability in ease of induction of the clinical arrhythmia

Using a one-way analysis of variance, the standard error of the number of extrastimuli required for repeat arrhythmia induction in the off-drug state at the same study was 0.408. Therefore, the 95% confidence intervals for the difference in the number of extrastimuli required for two sequential ventricular tachyarrhythmia inductions at the same study were 1.96 times the square root of 2, times the standard error, i.e. ho.94 rounded to f 1 extrastimuli. Consequently, if ventricular tachycardia or fibrillation was reproducibly inducible, there was a less than 5% chance that at a subsequent induction at the same electrophysiologic study an increment of two or more extrastimuli above that required at the initial induction would have been needed to reinduce ventricular tachycardia or fibrillation in the absence of antiarrhythmic drugs. Similarly, the standard error of the number of extrastimuli required for repeat arrhythmia induction in the off-drug state at different studies was 0.77 1. The 95% confidence intervals for the difference in the number of extrastimuli required to reinduce a ventricular tachyarrhythmia were + 2.1 rounded to f 2 extrastimuli. Therefore, if ventricular tachycardia or fibrillation was reproducibly inducible, there was a less than 5% chance that an increment of three or more extrastimuli over that used in the initial study

Table 2 Comparison physiologic Coronary disease Absent Present b

of spontaneous study in patients artery

would have been required for a subsequent induction of a sustained ventricular tachyarrhythmia at a different electrophysiologic study on a different day. On the basis of this logic, explained in detail in our previous studies of patients with ventricular tachycardia and coronary artery disease [19] we defined drug therapy as anti-arrhythmic if two or more extrastimuli above that needed in the baseline state were required for arrhythmia induction after intravenous drug administration. Similarly, for oral drug studies, therapy was predicted to be anti-arrhythmic if three or more extrastimuli above that needed in the baseline state were required for arrhythmia induction. Therapy was categorized as ‘not predicted to be effective’ if the change in the number of extrastimuli required after drug administration to induce a sustained ventricular arrhythmia compared to that in the off-drug state had a greater than 5% probability of occurring by chance. Therapy was classified as ‘not assessable by electrophysiologic study’ if the clinical arrhythmia was not reproducibly inducible (12 patients). Three patients elected to undergo only oral drug testing and were classified as not tested at intravenous study. Within-study variability in the ease of arrhythmia induction was similar to that in patients with ventricular tachycardia related to coronary artery disease. However, between-study variability in the number of extrastimuli required for repeat arrhythmia induction was significantly greater than that described in similar patients with coronary artery disease (P = 0.02 x2 test, Tables 2 and 3) 1191.

variability in the ease of induction ot- sustained with and without coronary artery disease

No. of additional

extrastimuli

ventricular

tachyarrhythmias

at the same electro-

required”

S-1

0

+I

r2

N

5 (13%) 12 ( 15%)

32 (82%) 61(780/u)

2 (5%) 5 (6%)

0 0

39 78

x2 corrected for continuity = 0.24, P = 0.88; “For re-induction of sustained VTiVF at the same electrophysiologic study compared bData for this group of patients was obtained from Cooper et al. [19].

to that at the initial induction.

60

L.M. Davis et al. /ht.

Table 3 Comparison of the spontaneous variability in ease of induction days in patients with and without coronary artery disease Coronary disease

artery

Absent Presentb

No. of additional

extrastimuli

of sustained

s-1

0

+I

22

N

10 (22%) 22 (28%)

18 (39%) 40 (51%)

1 I (24%) 15 (19’Y”)

I ( 15%)

46 78

3.7. Drug therapy

Drug status at EPS

to electrophysiologic Initially Sotalol

allotted

Amiodarone

Treatment assessable at EPS Predicted effective 1 (N=8)

Predicted ineffective (N = 10)

Not tested/refused drug assessment (N=3) Amiodarone (N = 8)

Treatment (N=12)

status during therapy

I

2

day of sustained

8

VTiVF compared

to that at the initial

the study

Others

4

6

I (I’%,)

Drug status at EPS

I

for studies done on different

testing. No patient was given a drug that made ventricular tachycardia or fibrillation significantly easier to induce at electrophysiologic study. Twelve (29%) of the 41 patients had nonreproducible clinical arrhythmias preventing meaningful drug assessment at electrophysiologic study. Drug treatment was changed by attending

Drug treatments used during the course of the study are shown in Table 4. Eight (20%) of the 41 patients were treated with drugs predicted to be anti-arrhythmic at an electrophysiologic study. Ten patients initially received treatment not predicted to be anti-arrhythmic at intravenous

according

tachyarrhythmias

45 (1994) 53-6X

required”

Continuity corrected x2 = 10.2, P = 0.02; “For re-induction at a different electrophysiologic study on a different induction. bData for this group of patients was obtained from Cooper et al. [25].

Table 4 Drug therapy

ventricular

J. Car&l.

Subsequent

treatment

Sotalol

Amiodarone

Predicted effective Predicted ineffective Not tested Amiodarone

trials Others

No therapy

3

I I -

Predicted effective Predicted ineffective Not tested Amiodarone

I

1

-

Not tested

I

Predicted effective Predicted ineffective Not tested

2

1

I

I

I

not assessable ut EPS 5

I

6

2

3

5

L. M. Davis et al. /ht.

B

Alive

4

Arrhythmia free

0.6 -

.tj

61

J. Cardiof. 45 (1994) 53-68

1 R

0.4

IE s1 0

0.2

ct 0.0

I

0

20

*

’

I

x

’

40

I

60

’

=

,

80

-

’

I

100

Months Fig. 1. Arrhythmia recurrence and death during follow-up. In this life table patients were censored at the time of their first arrhythmia recurrence. The numbers attached t the graph refer to the number of patients in the study at the time of follow-up.

venous testing. Drug levels were not obtained at the time of the spontaneous arrhythmia recurrences. The number of arrhythmia recurrences in patients stratified by ventricular function, arrhythmia morphology and cycle length, cardiac pathology and the presence or absence of syncope associated with the presenting arrhythmia are detailed in Table 1. The incidence of arrhythmia recurrence was no higher or lower for any one drug compared to any other drug, including amiodarone. Four of 11 trials of amiodarone therapy were complicated by an arrhythmia recurrence compared to eight of the 43 other drug trials. Interestingly, only four of the 11 trials of no drug therapy led to an arrhythmia recurrence. Death occurred in five patients (all sudden), all while on allocated therapy. In two of these cases, allocated treatment had not been assessable at electrophysiologic study. In the remaining three cases therapy had not been predicted to be antiarrhythmic at intravenous drug testing. 3.9. Results of Poisson regression analysis

physicians or patients during the course of followup in 19 cases, and in some on several occasions (Table 4). Drug therapy was ceased for variable periods in ten of these cases. 3.8. Arrhythmia recurrence Recurrence of sustained ventricular tachycardia or fibrillation occurred in 16 patients (Fig. l), including multiple recurrences in seven patients. Of these, seven patients experienced an arrhythmia recurrence after they had either ceased (two patients) or changed their hospital-allocated treatment (five patients). Arrhythmia recurrence occurred on hospital-allocated therapy in 13 patients; four (31%) in the 12 patients who were unsuitable for drug testing by programmed electrical stimulation, three (23%) in the eight patients with reproducibly inducible arrhythmias who were allocated amiodarone and four (31%) in the 13 patients on therapy not predicted to be effective at programmed stimulation (Table 1). No patient experienced an arrhythmia recurrence while on therapy predicted to be anti-arrhythmic at intra-

With this analysis, both the duration of followup and the number of previous arrhythmia recurrences were significantly associated with the risk of arrhythmia recurrence. These factors were best considered as ordered categorical variables in the model. Arrhythmia recurrences were categorized as none, one, two, and three or more recurrences. Duration of follow-up was divided into three groups: O-3, 4-12, and > 12 months after initial assessment, respectively. The number of arrhythmia recurrences had a relative risk for further arrhythmia recurrence of 3.1 (95% confidence intervals 2.5-4.0). The duration of follow-up carried a relative risk of 0.5 (95% confidence intervals 0.3-0.8) after adjustment for the number of arrhythmia recurrences. In other words, the longer the time from initial hospital assessment the lower was the chance of an arrhythmia recurrence. Table 5 sets out the relative risks for the other examined variables adjusted for the effect of time and the number of arrhythmia recurrences. The only variables signScantly associated with risk of arrhythmia recurrence after this adjustment were ventricular function, drug assessment at electro-

Table 5 Relative risk of arrhythmia on amiodarone therapy

recurrence

Characteristic

for various

patient

characteristics

Relative risk”

determined

by Poisson

Contidence intervals”

df

Intravenous drug assessment at EPS

9.1 7.2 1 1.9

Syncope No syncope

I I.1

P

2

8.0

0.02

1

3.1

0.08

I

0.1

0.75

4

11.5

0.02

7.4

0.01

7.7

0.10

0.6-2.1

morphology

Left bundle branch block Right bundle branch block Ventricular fibrillation Not classifiable (rhythm strip only) Indeterminate bundle branch block

1 0.7 0.3 0.1 0.3

0.3-1.7 0.1-1.3 0.0-0.6 0.0-2.4

cycle length

< 250 ms

I

> 250 ms

4.9

1.2-20.’

Cardiac pathology No macroscopic

abnormality Idiopathic dilated cardiomyopathy Alcoholic cardiomyopathy Congenital heart disease Other “Relative risks and confidence assessment.

not

0.9-3.9

Symptoms

Arrhythmia

in patients

1.2-70.0 1.0-50.5

function

Normal ventricular function Abnormal ventricular function

Arrhythmia

x2

analysis

I

Anti-arrhythmic Not anti-arrhythmic Not assessable Ventricular

regression

1 2.0 0.3 1.8 2.0 intervals

have been adjusted

physiologic study (after exclusion of the results of intravenous amiodarone testing) and the cycle length of the presenting arrhythmia. Arrhythmia recurrence tended to be more likely the longer the cycle length of the presenting arrhythmia. There was no difference in risk of arrhythmia recurrence amongst the different drugs used in this study. The best fitting multivariate model for prediction of the risk of arrhythmia recurrence in patients in whom the results of intravenous amiodarone tests are excluded is shown in Table 6 and included the following variables: (1) the number of arrhythmia recurrences suffered, (2) the duration of follow-up, (3) whether treatment was predicted to be antiarrhythmic or not or was not assessable at electrophysiologic study and (4) whether ventricular function was normal or not (Table 6). Fitting of

0.8-4.7 0.0-2.4 0.4-8.6 0.8-4.7 for the number

of arrhythmia

recurrences

suffered

and time since

the model to the actual data was not improved by inclusion of any of the other examined variables. 3.10. Mortality The calculated mortality rate in the first year of follow-up by Poisson model methods was 5”/u (95% confidence limits = 2-13%). Using the same methods the arrhythmia recurrence rate at 1 year of follow-up was 21% (95% confidence limits: 12-37%). Regression methods were not used to analyze mortality because of the small number of deaths. However there was a trend for those with abnormal ventricular function to be at greater risk of death. All five deaths occurred in patients with ventricular dysfunction (P < 0.05). Over the first year of follow-up the mortality rate in those with

63

L.M. Davis et ul. /ht. J. Cardiol. 45 (1994) 53-68 Table 6 Best model for prediction of the risk of arrhythmia recurrence in patients with sustained ventricular tachyarrhythmias unrelated to coronary artery disease Factor

Patients not on amiodarone therapy Relative risk

95% Conftdence limits

tachycardia induction was allowed for. However, anti-arrhythmic drug treatment not predicted to be effective at electrophysiologic study, or not able to be tested at electrophysiologic study because the clinical arrhythmia was non-inducible or not reliably inducible, was associated with a high arrhythmia recurrence rate. 4.1. Drug assessment

No arrhythmia recurrences” No of arrhythmia recurrences Time since hospital assessment Treatment not predicted to be anti-arrhythmic at EPS Treatment not assessable at electrophysiologic study Abnormal ventricular function

1.0 3.2 0.7

2.3-4.3 0.4-l .3

10.9

1.3-90.9

14.1

1.9-105.7

3.8

1.7-8.5

Results of drug assessment at electrophysiologic study and ventricular function are adjusted for the effects of each other, time since assessment and the number of arrhythmia recurrences. Arrhythmia recurrence was an ordered categorical variable, categories consisting of 0. I. 2, and 3 or more recurrences. Time since hospital assessment also was considered as an ordered categorical variable, categories consisting of O-3 months. 4- 12 months and > I2 months. The calculated relative risks for these ordered categorical variables was multiplicative. Symptoms, the cycle length and the morphology of the presenting arrhythmia, drugs used and underlying cardiac pathology did not add to the predictive power of the model. “After O-3 months of hospital assessment, on treatment predicted to be anti-arrhythmic after testing at EPS, and normal ventricular function.

impaired left ventricular function was 8% (95% confidence limits = 3-21%). There was no detectable relation between the risk of death and any of the other examined variables (Fisher’s exact test). 4. Discussion This study demonstrates that the long-term prognosis of patients with ventricular tachycardia not related to coronary artery disease is generally good and, even in the presence of impaired ventricular function, is better than that of patients with ventricular tachycardia based on coronary artery disease. The study showed that the electrophysiologic study could be used to select effective longterm treatment if baseline variability in ventricular

at electrophysiologic

stud4

Previous studies have provided conflicting evaluations of the value of drug testing at electrophysiologic study for predicting treatment outcome in patients with sustained ventricular tachyarrhythmias unrelated to coronary artery disease. Some studies found that electrophysiologic drug testing can predict treatment outcome [6,20-221. Others reached the opposite conclusion [7,12,23]. Milner et al., in a study of 18 patients with dilated cardiomyopathy and sustained ventricular tachyarrhythmias, found no difference in outcome between those on therapy predicted effective and those on therapy predicted to be ineffective, despite using up to three extrastimuli in their stimulation protocol [7]. Five of their nine patients on therapy predicted to be effective had an arrhythmia recurrence, compared to two of the four patients on therapy predicted to be ineffective. Five of their patients did not have inducible arrhythmias at their baseline electrophysiologic study. Deal et al. reported similar results in 12 patients with inducible sustained ventricular tachyarrhythmias using a stimulation protocol with up to two or three extrastimuli [23]. Their study also used the results of Holter studies to select treatment. During 13 trials of predicted effective therapy there were three arrhythmia recurrences. compared to only one arrhythmia recurrence during four trials of predicted ineffective therapy. Roy et al. studied 85 patients with ventricular tachyarrhythmias of whom 34 patients had ventricular tachycardia unrelated to coronary artery disease 1121. Their study also failed to demonstrate that electrophysiologic drug testing predicted outcome of treatment despite using up to three extrastimuli during programmed stimulation. Six of

L.M.

64

their 41 patients on treatment predicted to be effective had an arrhythmia recurrence compared to five of the 27 patients on therapy predicted to be ineffective (P = 0.6). There are several potential reasons why these three studies failed to demonstrate that electrophysiologic testing could predict treatment outcome and why the positive results of other studies are disputed. None of the above studies, including those that found drug testing by programmed stimulation could predict treatment outcome, have assessed the reproducibility of arrhythmia induction or allowed for baseline variability in ventricular tachycardia induction in defining anti-arrhythmic drug effects [6,12,15,20-231. In addition these studies have used non-aggressive stimulation protocols with a maximum of only two or three ventricular extrastimuli, usually without the addition of isoproterenol [4,6,7,20.21,24]. Furthermore multiple regression analysis was almost never used to control for the effects of other potentially predictive variables. Such an analysis has been performed in only one of 15 studies to date [22]. Previous studies also have lacked statistical power reducing the chances of finding a convincing relationship between the results of programmed stimulation and outcome of drug therapy. These problems are discussed in detail below. 4.2. Variability

in the ease of arrhythmia

induction

Failure to allow for normal variability in ease of induction of ventricular tachyarrhythmias when interpreting drug tests is a likely reason why some previous studies failed to demonstrate that drug testing at electrophysiologic study predicted treatment outcome [7,12,23]. Cooper et al. have demonstrated the presence of significant spontaneous variability in the number of extrastimuli required for repeat arrhythmia induction in the absence of anti-arrhythmic drugs [ 19,251. Spontaneous variability in ease of arrhythmia induction should be allowed for when considering whether a drug is predicted to be anti-arrhythmic at programmed electrical stimulation. In our analysis, treatment was defined as antiarrhythmic only when the number of additional extrastimuli required for arrhythmia induction

Duvis el al. /ht.

J. C‘ordiol.

4.5

( 1994)

53-68

after drug administration had increased outside the 95% confidence limit for baseline variability in the absence of anti-arrhythmic drugs. This approach requires the use of an aggressive stimulation protocol. Most (90%) inductions of the patients’ clinical arrhythmias in the absence of anti-arrhythmic drugs require at least three extrastimuli and an additional 10% require four extrastimuli. Previous studies that used only two or three ventricular extrastimuli and defined an anti-arrhythmic effect as failure to induce ventricular tachycardia are likely to have made false predictions of drug efficacy in lo-50% of cases, because ventricular arrhythmias became noninducible with their stimulation protocols due to spontaneous variation in the ease of arrhythmia induction unrelated to the administered drug. If a placebo had been substituted for the drugs in these tests, it also would have been predicted to be antiarrhythmic in these studies.

4.3. Reproducibility

of arrhythmia

induction

Ventricular tachycardia must be reliably inducible before electrophysiologic study can be used to test anti-arrhythmic drug effect. This problem has been almost universally ignored in previous studies that assess the value of drug testing with programmed electrical stimulation. Only three of the 15 previously published studies of patients with ventricular tachycardia unrelated to coronary artery disease performed more than one off-drug arrhythmia induction [2,3,26]. In our study, the clinical arrhythmia was induced on one occasion in 32 (78%) patients but was not reproducibly induced on repeat attempts despite using an aggressive stimulation protocol, in three (7%) of the 41 patients, thereby preventing drug assessment at electrophysiologic study. One of these three patients had multiple arrhythmia recurrences. Inclusion of these patients with those in whom treatment is predicted to be anti-arrhythmic at electrophysiologic study would obscure any ability of programmed stimulation to predict treatment outcome. Serial electrophysiologic study should not be used in patients whose clinical arrhythmia is not reproducibly inducible.

L.M. Davis et al. /ht.

J. Cardiol. 45 (1994) 53-68

4.4. Lack of statistical

65

power

Previous studies also probably lacked enough statistical power to consistently demonstrate the ability of drug testing by programmed stimulation to predict treatment outcome. Conventional analyses in which patients are censored at their first arrhythmia recurrence or on change of drug therapy require more than 220 patients to have an 80% chance of detecting a significant (P = 0.05) difference in outcome between predicted effective versus predicted ineffective therapy. Previous studies have been based on significantly fewer patients than this. Such large numbers are difficult to obtain in a reasonable time unless the trial involves small numbers of patients from a large number of centers (a theoretically unsound structure for a multicentre trial). 4.5. Potential limitations of programmed tion for assessing drug treatment

stimula-

ulation protocol and the clinical arrhythmia was initiated more than once. Occasional inductions of ventricular fibrillation were not a reason for exclusion of patients from the reproducible group. Ventricular fibrillation can occur even in normal patients as a by-product of an aggressive stimulation protocol and should be expected to complicate a percentage of arrhythmia inductions. A more rigid definition of reproducibility would exclude the patients in whom ventricular Iibrillation was occasionally and inadvertently induced. Such a definition would not change the extent of the variability in ease of arrhythmia induction observed in our patients. Seven (24%)) of the 29 patients in our study whose clinical arrhythmias were reproducibly inducible had ventricular fibrillation induced on one or more occasions and would be excluded from further testing if this more rigid definition of reproducible arrhythmia induction was used. 4.6. Oral drug testing at electrophysiologic

A low rate of reproducible arrhythmia induction has been suggested as a significant limitation in the application of programmed stimulation for drug testing [15]. Our data do not support this contention. Arrhythmia induction was sufficiently reproducible to permit drug testing in 70% (29) of our 41 patients. Previous studies have determined the incidence of reproducible induction of sustained clinical ventricular tachyarrhythmias by programmed stimulation in only a small number of patients (27 in total) [2,3,26]. They found reproducible arrhythmia induction in 18-43% of cases [2,3,26]. These percentages are lower than in this study and reflect the less aggressive stimulation protocols used (maximum of only two extrastimuli). Higher arrhythmia induction rates (up to 72%), similar to our findings, have been reported using more aggressive stimulation protocols but these studies have not commented on the reproducibility of arrhythmia induction [23]. Some may argue that our definition of reproducible arrhythmia induction incorporates some patients whose clinical arrhythmias are not always inducible. In our study, arrhythmia inductions were regarded as reproducible if a ventricular tachyarrhythmia was always initiated by our stim-

stud?

Our study suggests that oral drug testing at electrophysiologic study is not clinically useful. The large random variability in ease of ventricular tachyarrhythmia induction in the absence of antiarrhythmic drugs in studies done on a different day makes any anti-arrhythmic drug effect virtually impossible to demonstrate. Drug efficacy is demonstrable with programmed stimulation at oral testing only if arrhythmia induction is retarded by an additional three extrastimuli to that needed in the baseline study. We could demonstrate such an anti-arrhythmic effect in only four of 29 assessable patients at oral testing whereas eight patients were predicted to be on anti-arrhythmic therapy with intravenous drug testing. Therefore oral drug studies are likely to be less useful for predicting long-term efficacy of anti-arrhythmic drugs. 4.7. Ventricular function Our study demonstrated an association between ventricular function and risk of arrhythmia recurrence in patients with sustained ventricular tachyarrhythmias unrelated to coronary artery disease. Similar observations have been made in

L.M. Davis 1’1ul. /ht.

66

patients with heart failure from any cause [27], in patients with dilated cardiomyopathy [28,29] and in patients with a history of sustained ventricular tachycardia based on coronary artery disease [30].

4.8. Previous arrhythmia

recurrence

While it appears logical that those who already . . have had an arrhythmia recurrence on therapy will be more likely to have further recurrences, even if therapy is changed, this idea has not been examined previously. This is because most statistical methods of analysis censor patients after the first arrhythmia recurrence. This question was able to be studied with the statistical methods used in this paper. We found that arrhythmia recurrence was an independent predictor of further arrhythmia recurrence(s). 4.9. Mortality

Our study showed that patients with ventricular tachycardia not related to coronary artery disease have a good prognosis. The mortality rate was 5% (95% confidence intervals = 2-13%) in the first year after hospital assessment. This is a better outcome than in similar patients with ventricular tachycardia based on coronary artery disease where the yearly sudden-death rate is 11% in our experience. The risk of an arrhythmic death in our study was related mainly to underlying ventricular function. Those with normal ventricular function had no mortality, whereas those with impaired left ventricular function had a mortality rate of 8% (95% confidence intervals = 3-21%) in the first year. Previous studies in patients with normal ventricular function and ventricular tachyarrhythmias unrelated to coronary artery disease also have a zero mortality [2,3,8,10,11,15,21,23,31-341 whereas studies including patients with ventricular dysfunction have mortality rates of up to 34% [6,7,22]. The association between mortality and arrhythmia morphology described by Buxton and colleagues was not confirmed in our study [ 141 but we cannot discount it because of the small number of deaths in our study.

4.10. Limitations

J. Cardiol. 45 (1994) 53-68

of study

The relative rarity of sustained ventricular tachyarrhythmias in patients without coronary artery disease limited this study to a modest number of patients with a variety of underlying disease processes which may have different outcomes. Although cardiac pathology was not found to predict outcome in this study. studies with larger numbers of patients with only one disease process are needed to confirm these results. More sophisticated analyses of variability in ventricular tachycardia induction and methods of determining true drug effect at electrophysiologic study have been performed by our group in patients with ventricular tachycardia based on coronary artery disease [19]. We did not have enough data in the current study to produce probability matrices of that type. Drug levels were not obtained at the time of arrhythmia recurrence. Patient compliance as a potential factor in treatment outcome was not examined. 5. Conclusions Patients with documented sustained ventricular tachyarrhythmias unrelated to coronary artery disease are at increased risk of an arrhythmia recurrence if: (1) they are on therapy that is not predicted to be anti-arrhythmic at electrophysiologic study, or (2) their drug therapy is not assessable at electrophysiologic study because their clinical arrhythmia was not reproducibly inducible, or (3) they have abnormal ventricular function, or (4) they have already had an arrhythmia recurrence. Inability to reproducibly induce the clinical arrhythmia prevents drug assessment at electrophysiologic study in a significant proportion (29%) of patients with ventricular tachyarrhythmias unrelated to coronary artery disease. Empirical therapy in this group is often unsuccessful. In this study four (33%) of the 12 patients in this group had an arrhythmia recurrence on their initially allotted treatment. New approaches to treatment are needed in this group of patients. In patients whose clinical arrhythmias are reproducibly inducible, the random variability in

L.M. Davis et al. /ht.

67

J. Cardiol. 45 (1994) 53-68

the ease of arrhythmia induction should be considered when determining drug efficacy by proOur study electrical stimulation. grammed suggests that arrhythmia induction needs to be retarded at the same study by two or more extrastimuli to be 95% confident of an anti-arrhythmic drug effect. Most (90%) inductions of patients’ clinical arrhythmias in the absence of antiarrhythmic drugs require at least three extrastimuli and an additional 10% require four extrastimuli. Thus determination of a true anti-arrhythmic drug effect in these patients requires a stimulation protocol that includes at least four extrastimuli for testing re-induction after the drug. Patients in our study had a relatively low mortality rate (5% in the first year). Arrhythmic death appeared to be almost exclusively associated with abnormal ventricular function. Therefore these patients should be considered for either an automatic implantable defibrillator or arrhythmia surgery if drug therapy is not predicted effective at electrophysiologic study. 6. Acknowledgments This study was supported by National Heart Foundation Grant No. PM100 (National Heart Foundation of Australia, P.O. Box 2, Woden, A.C.T, Australia, 2606). We wish to thank Anita Watts and Vicki Eipper for their help with data collection.

6

7

8

9

IO

11

12

13

14

15

7. References 16 Ruskin JN, DiMarco JP, Garan H. Out of hospital cardiac arrest - electrophysiologic observations and selection of long-term anti-arrhythmic therapy. N Engl J Med 1980; 303: 607-613. Rahilly CT, Prystowsky EN, Zipes DP, Naccarelli GV. Jackman WM, Heger JJ. Clinical and electrophysiologic findings in patients with repetitive monomorphic ventricular tachycardia and otherwise normal electrocardiogram. Am J Cardiol 1982; 50: 459-468. Pedersen DH, Zipes DP. Foster PR, Troup PJ. Ventricular tachycardia and ventricular fibrillation in a young population. Circulation 1979; 605: 988-997. Poll DS. Marchlinski FE, Buxton AE, Josephson ME. Usefulness of programmed stimulation in idiopathic dilated cardiomyopathy. Am J Cardiol 1986; 58: 992-997. Rapaport E. Prevention of recurrent sudden death. N Engl J Med 1982; 306: 1359-1360.

17

18

19

20

Poll DS, Marchlinski FE, Buxton AE, Doherty JU, Waxman HL, Josephson ME. Sustained ventricular tachycardia in patients with idiopathic dilated cardiomyopathy: electrophysiologic testing and lack of response to drug therapy. Circulation 1984; 70.3: 45 I-456. DiMarco JP, Mimer PC, Lerman BB. Electrophysiological evaluation of sustained ventricular tachyarrhythmias in idiopathic cardiomyopathy. PACE 1988; 11: 562-568. Blomstrom-Lundqvist C, Sabel K. Olsson SB. A longterm follow-up of 15 patients with arrhytlimogenic right ventricular dysplasia. Br Heart J 1987; 58: 477-488. Marcus FI, Fontaine GH, Guiraudon G, Frank R, Laurenceau JL, Malergue C et al. Right ventricular dysplasia: A report of 24 adult cases. Circulation 1982; 65.2: 384-398. Belhassen B. Shapira I, Shoshani D. Paredes A, Miller H, Laniado S. Idiopathic ventricular fibrillation: inducibility and beneficial effects of class I anti-arrhythmic agents. Circulation 1987; 75.4: 809-816. Lemery R, Brugada P, Bella PD. Dugernier T, Wellens HJJ. Ventricular fibrillation in six adults without overt heart disease. J Am Coil Cardiol 1989; 13,4: 91 I-916. Roy D, Waxman HL, Kienzle MC, Buxton AE, Marchlinski FE, Josephson ME. Clinical characteristics and long-term follow-up in 119 survivors of cardiac arrest: relation to inducibility at electrophysiologic testing. Am J Cardiol 1983; 52: 969-974. Friedman L, Yusuf S. Does therapy directed by programmed electrical stimulation provide a satisfactory clinical response? Circulation 1986; 73, suppl II: 59-66. Buxton AE, Waxman HL, Marchlinski FE. Simson MB, Cassidy D, Josephson ME. Right ventricular tachycardia: clinical and electrophysiologic characteristics. Circulation 1983; 68,5: 917-27. Lemery R, Brugada P, Bella PD. Dugernier T, Van Der Do01 A, Wellens HJJ. Nonischaemic ventricular tachycardia: clinical course and long-term follow-up in patients without clinically overt heart disease. Circulation 1989; 79: 990-999. Aitkin M, Anderson D, Francis B, Hinde J. In: Statistical modelling in GLIM. Oxford: Clarendon Press, 1989; 287-289. Lunn D, McNeil D. In: SPIDA, Statistical package for interactive data analysis. Version 5. Sydney: Macquarie University, 1988; 101-104. Mitchell LB, Wyse DC. Gillis AM, Duff HJ. Electropharmacology of amiodarone therapy: time courses of onset of electrophysiologic and antiarrhythmic effects, Circulation 1989; 80: 34-42. Cooper M, Koo C, Skinner M, Mortensen P. Hunt L. Richards D et al. Comparison of immediate versus day to day variability of ventricular tachycardia induction by programmed stimulation. J Am Coil Cardiol 1989; 13: 1599- 1607. Rae AP, Spielman SR, Kutalek SP, Kay HR. Horowitz LN. Electrophysiologic assessment of anti-arrhythmic drug efftcacy for ventricular tachyarrhythmias associated

68

21

22

23

24

25

26

with dilated cardiomyopathy. Am J Cardiol 1987; 59: 291-295. Naccarelli GV. Prystowsky EN, Jackman WM. Heger JJ, Rahilly GT. Zipes DP. Role of ei~trophysiolo~ic testing in managing patients who have ventricular tachycardia unrelated to coronary artery disease. Am J Cardiol 1982; 50: 165-171. Liem LB, Swerdlow CD. Value of eIectropharma~[~log~~ testing in idiopathic dilated cardiomyopathy and sustained ventricular tachyarrhythmias. Am J Cardiol 1988: 62: 61 l-616. Deal BJ, Miller SM. Scagliotti D, Prechel D. Gallastegui JL, Hariman RJ. Ventricular tachycardia in a young population without overt heart disease. Circulation 1986: 73.6: 1111-1118. Ohe T, Shimomura K, Aihara N. Kamakura S, Matsuhisa M, Sato f et al. Idiopathic sustained left ventricular tachycardia: clinical and electrophysiologic characteristics. Circulation 1988; 77.3: 560-568. Cooper MJ, Hunt LJ. Palmer KJ, Denniss AR, Richards DA. Uther JB, et al. Quantiiication of day to day variability in mode of induction of ventricular tachycardia by programmed stimulation. J Am Coil Cardiol 1988: t 1: 101-108. Denes P. Wu D. Dhingra RC. Amat-Y-Leon F, Wyndham C, Mautner RK et al. Electrophys~oiogic studies in patients with chronic recurrent ventricular tachycardia. Circulation

1976: 54.2: 229-236.

27

28

29

30

31 32

33

34

Rouleau J, Shenasa M, De Champlain J. Nadeau R. Predictors of survival and sudden death in patients with stable severe congestive heart failure due to ischemic and nonischemic causes: A prospective long-term study of 200 patients. Can J Cardiol 1990; 6: 453-460. Hofmann T, Meinertz T. Kasper W, Geibel A. Zehender M, Hohnloser S et al. Mode of death in idiopathic dilated cardiomyopathy: A multivariate analysis of prognostic determinants. Am Heart J 1988; 116: 1455-1463. De Maria R, Gavazzi A, Caroli A, Ometto R. Biagini A, Camerini F et al. Ventricular arrhythmias in dilated cardiomyopathy as an independent prognostic hallmark. Am J Cardiol 1992; 69: 1451-1457. Swerdlow CD, Winkle RA, Mason JW. Determinants of survival in patients with ventricular tachyarrhythmias. N Engl J Med 1983; 308: 1436-1442. Adams CW. Functiona paroxysmal ventricular tachycardia. Am J Cardiol 1962: 9: 215-222. Lesch M, Lewis E, Humphries JO. Ross RS. Paroxysmal ventricular tachycardia in the absence of organic heart disease. Ann Intern Med 1967; 66: 950-960. Wei JY. Bulkley BH, Schaeffer AH, Greene HL, Reid PR. Mitral valve prolapse syndrome and recurrent ventricular tachyarrhythmias. Ann Intern Med 1978; 89: 6-9. Fulton DR. Chung KJ, Tabakin BS, Keane JF. Ventricular tachycardia in children without heart disease. Am J Cardiol 1985; 55: 1328-1331.