| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
(Circulation. 1995;91:1908-1909.)
© 1995 American Heart Association, Inc.
Articles |
ESVEM and the Hazards of Clinical Trials
From the University of North Carolina at Chapel Hill.
Correspondence to Dr Leonard S. Gettes, University of North Carolina at Chapel Hill, 349 Burnett-Womack Bldg, CB 7075, Chapel Hill, NC 27599-7075.
 |
Introduction |
|---|
 Top Introduction References |
|---|
During the past 15 years, our approach to patients with cardiovascular disease has undergone significant and often dramatic changes. These changes have primarily occurred as a result of large-scale multicenter clinical trials involving many patients and extending over several years. One such trial is the ESVEM study,1 which was initiated in 1985, involved 14 centers, screened more than 2000 patients, and ultimately enrolled 486 patients with life-threatening ventricular arrhythmias. The trial had two major purposes: (1) to determine whether serial electrophysiological testing or serial Holter monitoring provided the better method of predicting the long-term therapeutic efficacy of antiarrhythmic drugs and (2) to determine which of the six antiarrhythmic drugs tested was most effective. In this issue of Circulation, Reiter and coworkers2 report on results obtained in a subgroup of 146 patients who were among the 242 patients comprising the electrophysiological arm of the ESVEM trial. In this subgroup of 146, Holter monitoring was also performed. The authors report that the therapeutic efficacy of the various drugs was no greater when predicted by both tests (concordant results) than when predicted by only one of the two tests (discordant results). Although not specifically stated, the hypothesis being tested was that the ability of a drug to suppress the arrhythmia trigger (as reflected by the suppression of the ventricular premature beats recorded by Holter monitoring) and to alter the arrhythmia substrate (as reflected by the inability to induce ventricular tachycardia by electrophysiological testing) should predict better long-term efficacy than the ability to accomplish either alone. This hypothesis makes intuitive sense if we assume that both the trigger and the substrate remain unchanged for the duration of the study and that the drugs themselves are not capable of causing spontaneous beats and/or of producing an arrhythmia substrate, ie, are not proarrhythmic. We have learned from a variety of clinical trials performed in the past decade that neither of these assumptions is true. Thus, there is no a priori reason to suspect that the hypothesis would have been proved. One might equally well have postulated that the ability to accomplish both would have identified a drug with more pronounced electrophysiological effects and that these effects might actually have resulted in a greater frequency of proarrhythmic events.
After reading the report by Reiter and coworkers and other reports and editorials spawned by the ESVEM trial3 4 5 6 7 and observing practice patterns with which I am familiar, I find myself doubting that the main results of the ESVEM trial, ie, that serial Holter monitoring predicts long-term therapeutic efficacy more often than serial electrophysiological testing and that sotalol may be the treatment of choice in patients with life-threatening ventricular arrhythmias, have significantly altered the approach to such patients. In my estimation, there are several factors responsible for this lack of impact. These factors fall into two general categories: those that relate to the mechanism of the trial itself and those that relate to the therapeutic arm of the trial.
The entry requirements of the trial restrict the conclusions to a specific group of patients with documented or suspected ventricular arrhythmias, at least 480 ventricular premature beats on the 48-hour Holter monitor, and inducible ventricular tachycardia by electrophysiological testing. Although one can argue that the profile of patients in the ESVEM trial was similar to that of patients with life-threatening ventricular arrhythmias described for most other trials, these specific requirements are rarely met. In the ESVEM trial, the electrophysiological protocol used during drug assessment was restricted to the same protocol that induced the arrhythmia in the baseline study. This restriction might have resulted in a greater number of subsequent drug failures than would have occurred had the drug-testing protocol included three extra stimuli at two sites in all patients, regardless of the baseline protocol. In the ESVEM trial, the exercise test was used for different purposes in the two limbs of the study. The exercise test was included in the drug-assessment portion of the Holter arm, whereas it was used to gauge long-term efficacy in the electrophysiology arm. This difference favored a greater prediction of drug efficacy in the Holter limb. For these several reasons, those with a bias for sequential electrophysiological testing can rationalize a decision to ignore the results of the ESVEM trial.
There are several factors that apply to the therapeutic arm of the trial. There was neither a placebo group nor a group in whom a ß-adrenergic–blocking agent without class III effects was administered. However, the use of either in patients with documented life-threatening ventricular arrhythmias and without an implanted defibrillator would have been indefensible. Yet, their absence, at least in the short-term drug-testing portion of the study, raises interesting questions, since it is likely that a certain proportion of patients would have demonstrated arrhythmia suppression with one or both of these agents. In addition, it has been speculated since 19838 and, most recently, in the wake of CAST9 that patients whose arrhythmia is suppressed during short-term drug testing may represent a group who, for unknown reasons, does just as well and perhaps even better over time if untreated than if treated with the effective drug. It is therefore possible that in both the Holter and electrophysiological arms of the study, respondents might have been as well treated with placebo as with the active drug. On the other hand, it has also been shown that in patients with life-threatening ventricular arrhythmias in whom drug therapy guided by serial electrophysiological testing is used, the discontinuation of the effective drug is associated with an increase in mortality.8
It is not known whether the outcomes observed in the sotalol-treated patients were the result of ß-adrenergic blockade or of the class III effects of the drug. The CASH trial10 suggests no significant differences between patients randomized to treatment with ß-blocker and those treated with either amiodarone or implantable defibrillators regardless of the results of electrophysiological testing or Holter monitoring. This result suggests that the beneficial effects of sotalol may have been due to ß-adrenergic blockade rather than to its class III effects. However, as with placebo, the use of a ß-adrenergic–blocking agent alone in this group of patients, without the protection of an implantable defibrillator, would have been difficult to justify.
The perceived lack of behavior modification by physicians after the ESVEM trial may also relate to the exclusion of amiodarone, to the failure to use implantable defibrillators, or both. In the ESVEM trial, the amiodarone option was specifically rejected, most likely because of the long half-life and unusual pharmacokinetics,11 which would have virtually prevented other drugs from being tested in a timely fashion, and because of the frequent inability to demonstrate arrhythmia suppression during electrophysiological testing in patients whose clinical arrhythmia is suppressed by long-term administration of the drug.12 Moreover, the results from several studies indicating a beneficial effect of empirically administered amiodarone to patients after myocardial infarction13 or to patients surviving an episode of sudden cardiac death10 14 were not available at the time the trial was designed. Also, the implantable defibrillator was not routinely used at that time. Its use, initially with a transthoracic approach, would not necessarily have prevented the ESVEM trial or invalidated its results. However, the availability of implantable defibrillators, particularly those that can be inserted without thoracotomy, has certainly altered the therapeutic approach to patients with life-threatening ventricular arrhythmias. It has permitted the use of placebo and ß-adrenergic–blocking agents in these patients and has made the questions under investigation in the ESVEM trial appear less important.
For all of these reasons, the ESVEM trial may be viewed as a case study of the hazards of large-scale clinical trials. At initiation, the question under investigation was valid and important. Moreover, the results were provocative and could not have been predicted by the available information. However, in 1995, 9 years after initiation of the trial, a variety of choices and events have conspired to make the study less relevant than anticipated, despite new information such as that contained in the article by Reiter and coworkers.2 The results of concurrent studies, the development of new technologies, and the need to address important issues in the course of the trial exemplify the hazards facing clinical investigators and the choices that must be made when considering long-term clinical trials.
Received February 7, 1995; accepted February 7, 1995.
|
References |
|---|
|
TopIntroductionReferences |
|---|
1. The ESVEM Investigators. The ESVEM trial: electrophysiological study versus electrocardiographic monitoring for selection of antiarrhythmic therapy of ventricular tachyarrhythmias. Circulation. 1989;79:1345-1360.
2.
Reiter MJ, Mann DE, Reiffel JE, Hahn E, Hartz V, ESVEM
Investigators. Significance and incidence of concordance of drug
efficacy predictions by Holter monitoring and electrophysiological
study in the ESVEM trial. Circulation. 1995;91:1988-1995.
3.
Mason JW for the Electrophysiologic Study Versus
Electrocardiographic Monitoring Investigators. A comparison of
electrophysiologic testing with Holter monitoring to predict
antiarrhythmic-drug efficacy for ventricular tachyarrhythmias.
N Engl J Med. 1993;329:445-451.
4.
Mason JW for the ESVEM Investigators. A comparison of seven
antiarrhythmic drugs in patients with ventricular tachyarrhythmias.
N Engl J Med. 1993;329:452-458.
5.
Ward DE, Camm AJ. Dangerous ventricular arrhythmias—can we
predict drug efficacy? N Engl J Med. 1993;329:498-499.
6. Murgatroyd F. Malignant arrhythmias: tribulations post ESVEM. Lancet. 1993;342:569.
7.
Greenspan AM. Determinants of antiarrhythmic drug efficacy
for ventricular tachyarrhythmias using ambulatory monitoring and
electrophysiological techniques. Circulation. 1993;87:643-645.
8. Swerdlow CD, Winkle RA, Mason JW. Determinants of survival in patients with ventricular tachyarrhythmias. N Engl J Med. 1983;308:1436-1442. [Abstract]
9. Cardiac Arrhythmia Suppression Trial (CAST) Investigators. Preliminary report: effect of encainide and flecainide on mortality in a randomized trial of arrhythmia suppression after myocardial infarction. N Engl J Med. 1989;321:406-412. [Abstract]
10. Siebels J, Cappato R, Rüppel R, Schneider MAE, Kuck KH, and the CASH investigators. Preliminary results of the Cardiac Arrest Study Hamburg (CASH). Am J Cardiol. 1993;72:109F-113F. [Medline] [Order article via Infotrieve]
11. Roden DM. Pharmacokinetics of amiodarone: implications for drug therapy. Am J Cardiol. 1993;72:45F-50F. [Medline] [Order article via Infotrieve]
12. Horowitz LN, Greenspan AM, Speilman SR, Webb CR, Morganroth J, Rotmensch H, Sokoloff NM, Rae AP, Segal BL, Kay HR. Usefulness of electrophysiologic testing in evaluation of amiodarone therapy for sustained ventricular tachyarrhythmias associated with coronary heart disease. Am J Cardiol. 1985;55:367-371. [Medline] [Order article via Infotrieve]
13.
Nademanee K, Singh BN, Stevenson WG, Weiss JN. Amiodarone and
post–myocardial infarction patients. Circulation. 1993;88:764-773.
14. The CASCADE Investigators. Randomized antiarrhythmic drug therapy in survivors of cardiac arrest (the CASCADE study). Am J Cardiol. 1993;72:280-287.[Medline] [Order article via Infotrieve]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||