|
|
| |||||||||||||||||||||||||||||||||||||||
Background Invasive electrophysiologic study and noninvasive Holter monitoring (in conjunction with exercise testing) have both been used to evaluate the efficacy of antiarrhythmic drugs in patients with sustained ventricular tachycardia and in survivors of cardiac arrest. We directly compared these two approaches to the prediction of drug efficacy.
Methods A total of 486 patients who had documented ventricular tachyarrhythmias that were inducible during electrophysiologic study and 10 or more premature ventricular complexes per hour during Holter monitoring were randomly assigned to undergo serial testing of antiarrhythmic-drug efficacy by electrophysiologic study or Holter monitoring. The patients received up to six drugs in random order until one was predicted to be effective either in suppressing inducible arrhythmia (in the electrophysiologic-study group) or in suppressing premature ventricular complexes (in the Holter-monitoring group). The patients were then followed for recurrences of arrhythmia or death.
Results In the electrophysiologic-study group, 108 of 242 patients (45 percent) received a prediction of efficacy, as compared with 188 of 244 patients (77 percent) in the Holter-monitoring group (P<0.001). Over a six-year follow-up period, there were 150 recurrences of arrhythmia and 46 deaths among the 296 patients receiving drugs predicted to be effective. Thirty-four of the deaths were from arrhythmic causes, and eight were from cardiac causes. There was no significant difference between the two study groups in the actuarial probabilities of these events. The risk of a recurrence of arrhythmia was significantly lower in patients who received sotalol than in those who received other antiarrhythmic drugs, and the risk was lower in those who had not previously failed to respond to antiarrhythmic drugs than in those who had.
Conclusions Although Holter monitoring led to predictions of antiarrhythmic-drug efficacy more often than did electrophysiologic study in patients with sustained ventricular tachyarrhythmias, there was no significant difference in the success of drug therapy as selected by the two methods.
It has been shown that a prediction of drug efficacy by electrophysiologic study3 or by Holter monitoring with exercise testing4 identifies patients who have a lower risk of death from arrhythmia and other cardiac causes than those without such a prediction. Several retrospective analyses have suggested that electrophysiologic study provides more reliable predictions of efficacy than Holter monitoring5,6,7. A small, randomized trial reached the same conclusion8. The ESVEM trial was a large, randomized trial directly comparing electrophysiologic study with Holter monitoring for the prediction of antiarrhythmic-drug efficacy.
Methods
The design of the trial has been described elsewhere1,9. Patients were screened for enrollment at 14 centers, listed in the Appendix. A total of 2103 patients were enrolled in the study log on the basis of one or more of the following three characteristics: electrocardiographically documented sustained ventricular tachycardia or fibrillation (lasting 15 seconds or more), resuscitation from cardiac arrest, and syncope without electrocardiographic documentation of the responsible rhythm. In the case of syncope, ventricular tachycardia inducible by electrophysiologic study was also required. Medical conditions requiring exclusion1 from the study were present in 600 patients, and 498 other patients declined to be enrolled. Of the remaining 1005 patients, 141 did not meet the entry criterion for Holter monitoring (an average of 10 or more premature ventricular complexes per hour during 48 hours of monitoring), 349 did not meet the criterion for electrophysiologic study (reproducibly inducible ventricular tachyarrhythmia), and 29 did not meet either set of criteria.
The remaining 486 patients (23 percent of those enrolled in the study) gave informed consent to the protocol, which was approved by the institutional review board. They were randomly assigned to undergo serial testing of drug efficacy by either electrophysiologic study (242 patients) or Holter monitoring (244 patients). Seven antiarrhythmic drugs (imipramine, mexiletine, pirmenol, procainamide, propafenone, quinidine, and sotalol) were used during the course of the trial, but a maximum of six were available at any one time1,10. In each study group the patients received the drugs in random order until one was predicted to be effective, until all drugs had been tried that the patients were eligible to receive, or until the patients were withdrawn from the study. In the Holter-monitoring group a prediction of efficacy was defined on the basis of substantial suppression of premature ventricular complexes,1 as well as the absence of ventricular tachycardia of five or more consecutive beats on exercise testing. In the electrophysiologic-study group a prediction of efficacy required the suppression of inducible ventricular tachyarrhythmias lasting more than 15 beats.
The follow-up of patients for whom predictions of efficacy were made included examination and interview by a study investigator. Surviving patients who did not receive such a prediction were evaluated annually for vital status and any recurrence of arrhythmia. Before one year of follow-up was completed, four patients (1 percent) who received predictions of efficacy and six patients (3 percent) who did not were lost to follow-up. An additional 9 and 11 patients (3 and 6 percent), respectively, were lost to follow-up after one year.
Statistical Analysis
The primary end point was the recurrence of arrhythmia in patients who were receiving a drug predicted to be effective by serial testing. The study was designed to have a power of 0.80 at a two-tailed significance level of 0.05 to detect a difference of 15 percentage points in the actuarial probability of a recurrence of arrhythmia in the two groups, assuming an annual probability of recurrence of 25 percent. A recurrence of arrhythmia was defined as any of the following: electrocardiographically documented ventricular tachycardia more than 15 beats in duration, ventricular fibrillation, death caused by arrhythmia, cardiac arrest, torsade de pointes, or unmonitored syncope with no explanation other than arrhythmia. Three end points related to mortality (death from any cause, death from a cardiac cause, and death from arrhythmia) were also examined, although the trial was not designed to have sufficient power to identify differences in mortality between the two groups. A Safety Monitoring Committee regularly reviewed the relative frequencies of events and complications in the two groups.
Deaths from cardiac causes were defined as those resulting from cardiogenic circulatory failure and those due to arrhythmia. An End Points Committee reviewed all deaths and classified them according to the criteria of Hinkle and Thaler11.
The base-line characteristics of the patients and the categories of recurrence were compared by standard statistical tests for categorical and continuous data. The proportion of occurrences of the four end points was summarized with Kaplan-Meier actuarial methods12. The total duration of follow-up analyzed was 6.2 years. The actuarial analyses are presented here truncated at four years, by which time approximately 90 percent of the events had occurred. The frequencies of events were analyzed separately in two groups of patients: the 296 patients who had predictions of drug efficacy, and all 486 randomized patients. End points for the latter group included events that occurred after a failure to attain drug efficacy or the discontinuation of a drug for any reason other than a recurrence. The P values shown here to compare actuarial outcomes in the two study groups are the values obtained after adjustment for covariates by a multivariate Cox proportional-hazards regression model13 that assessed the effects of the study group and other clinical factors on the recurrence of arrhythmia and mortality. Statistical interaction between the study group and each stratification factor, prespecified clinical variable, or drug was evaluated first. A basic analytic model was then constructed, with the study group and the five stratification factors entered as main effects. A hierarchical step-up approach was used next, in which the clinical variables were added one at a time. A two-tailed significance level of 0.05 was used to test the primary hypothesis and to develop an exploratory model. The Ingres relational data-base management system14 was used to enter, store, and verify data, and SAS software15 was used for all the statistical analyses.
Results
Predictions of Drug Efficacy
An antiarrhythmic drug was predicted to be effective for 108 of the 242 patients in the electrophysiologic-study group (45 percent) and 188 of the 244 patients in the Holter-monitoring group (77 percent) (P<0.001). Predictions of efficacy based on Holter monitoring were rescinded by the results of exercise testing in only 3 of the 143 patients in the Holter-monitoring group who had predictions of efficacy and subsequently underwent exercise testing. Drug efficacy was predicted in a total of 296 patients (61 percent of those randomized); the remaining 190 patients, in whom no drug was predicted to be effective, were treated outside the study protocol.
Characteristics of the Patients
Table 1 shows the clinical characteristics of all 486 patients randomly assigned to the two study groups and those of the 296 patients who received predictions of efficacy. Among the 486 randomized patients, more patients in the Holter-monitoring group than in the electrophysiologic-study group used beta-blockers at the time of enrollment (P = 0.02). Among the 296 patients with predictions of efficacy, a larger proportion had coronary disease (P = 0.04) and at least 30 premature ventricular complexes per hour (P = 0.03) in the Holter-monitoring group than in the electrophysiologic-study group. There were no other significant differences.
|
There were 150 recurrences of arrhythmia among the 296 patients with predictions of drug efficacy while they continued to receive the drugs that were predicted to be effective. The mean (±SE) actuarial recurrence rates of arrhythmia were 37 ±3, 49 ±3, 61 ±4, and 66 ±4 percent at one, two, three, and four years, respectively. In the electrophysiologic-study group and the Holter-monitoring group, respectively, these probabilities were 32 ±5 and 41 ±4 percent at one year, 47 ±5 and 51 ±4 percent at two years, 64 ±6 and 59 ±4 percent at three years, and 64 ±6 and 67 ±5 percent at four years. The primary analysis in this trial, a comparison of the actuarial probability of a recurrence of arrhythmia in the two groups among the patients receiving a drug predicted to be effective, is shown in Figure 1. There was no significant difference between the two groups (risk ratio, 1.07; 95 percent confidence interval, 0.76 to 1.51; P = 0.69).
|
Characteristics of Recurrences of Arrhythmia
Table 2 categorizes all recurrences of arrhythmia in the 296 patients who received drugs predicted to be effective and compares the distribution of categories in the two study groups. Most of the recurrences (85 percent) involved sustained arrhythmias. Of these, 79 percent resulted in cardiac arrest or required emergency intervention with electrical cardioversion or the administration of a drug. All the episodes of sustained ventricular tachycardia lasted more than 30 seconds, except eight that lasted 15 to 30 seconds.
|
Predictors of a Recurrence of Arrhythmia
In analyses of the 296 patients with predictions of efficacy and of all 486 randomized patients, Cox proportional-hazards models were used to identify variables that influenced the recurrence of arrhythmia. Six variables were specified in advance, as follows: method of testing (electrophysiologic study vs. Holter monitoring), type of presenting arrhythmia (ventricular tachycardia or syncope vs. ventricular fibrillation or cardiac arrest), frequency of premature ventricular complexes, classification on the symptom-activity scale of Goldman et al.16 (class 1 vs. classes 2 through 4), enrolling center (Utah vs. other), and underlying cardiac disease (coronary artery disease vs. other). The following eight variables were also studied: antiarrhythmic drug used (from among the seven study drugs), age, unsustained ventricular tachycardia (present vs. absent), sex, failure of previous antiarrhythmic therapy (yes vs. no), beta-blocker therapy at the time of randomization (yes vs. no), left ventricular ejection fraction, and previous myocardial infarction (yes vs. no). The method of testing did not correlate significantly with recurrences of arrhythmia and did not interact with any of the variables examined to predict recurrence in either model. Treatment with sotalol10 and the absence of failure of antiarrhythmic-drug therapy before enrollment (risk ratio, 0.55; 95 percent confidence interval, 0.38 to 0.79; P = 0.001) were the only significant independent predictors of a reduction in the recurrence of arrhythmia among the 296 patients with predictions of drug efficacy. Only left ventricular ejection fraction (approximate risk reduction, 1.3 percent per ejection-fraction unit [defined as a 1 percent increase in the left ventricular ejection fraction]; 95 percent confidence interval, 1.0 to 2.4 percent; P = 0.03) was an independent predictor when all 486 randomized subjects were studied. In the models for the entire group of randomized subjects, drug treatment could not be included, because therapy was not controlled in patients who did not receive an efficacy prediction.
Mortality
By design, the statistical power of our study to detect differences in mortality was low. Among the 296 patients receiving drugs predicted to be effective, there were 46 deaths. Forty-two of these deaths were from cardiac causes, 34 of which were from arrhythmia. Among all 486 randomized patients, there were 180 deaths, 134 of which were from cardiac causes. Of these, 87 were from arrhythmia. Table 3 compares the actuarial probabilities of death from any cause and from arrhythmia and other cardiac causes in the two study groups among the 296 patients with predictions of efficacy and the 486 randomized patients. There were no statistically significant differences among the patients receiving drugs predicted to be effective. Among the patients with coronary artery disease, however, survival was better in the patients randomly assigned to electrophysiologic study than in those assigned to Holter monitoring, whereas those with no coronary artery disease had better survival in the Holter-monitoring group (P = 0.02).
|
Treatment of Patients outside the Protocol
The patients who discontinued a drug that had been predicted to be effective and the 190 patients in whom no drug was predicted to be effective were treated outside the protocol. The intention-to-treat analysis of recurrence of arrhythmia included 379 years of patient time outside the protocol (47 percent of the total), and the intention-to-treat analyses of mortality included 755 such years (64 percent of the total).
The 190 randomized patients in whom no drug was predicted to be effective were treated outside the study protocol according to the preferences of their physicians. The proportion of patients receiving an invasive or surgical therapy (implantation of a cardioverter-defibrillator, catheter ablation, surgical myocardial ablation or resection, coronary-bypass graft surgery, or implantation of an antitachycardia pacemaker) was significantly higher in the electrophysiologic-study group than in the Holter-monitoring group (49 percent vs. 22 percent, P = 0.002). Patients treated outside the protocol in the Holter-monitoring group were more likely to cross over to testing of drug efficacy by electrophysiologic study than were the patients in the electrophysiologic-study group to do the reverse (patients without efficacy predictions, 40 percent vs. 23 percent, respectively; P = 0.059; all patients treated outside the protocol, 40 percent vs. 26 percent; P = 0.010).
Among the 190 patients in whom no drug was predicted to be effective, the risk of recurrence of arrhythmia was similar to that among the 296 patients with predictions of efficacy, but the risk of death was higher. In the former group, the percentages with recurrence of arrhythmia at one and four years were 38 ±4 and 62 ±4, respectively, and the percentages who died from any cause were 22 ±3 and 49 ±5. Among these 190 patients, the percentages who had a recurrence of arrhythmia were similar in the two study groups, but cumulative mortality from any cause differed: Holter monitoring, 37 ±7 percent at one year and 64 ±9 percent at four years; electrophysiologic study, 15 ±3 percent at one year and 41 ±5 percent at four years; P = 0.022. There was no difference between groups, however, in the percentage who died from cardiac causes or arrhythmia.
Discussion
We found no significant difference in the probability of recurrence of ventricular tachyarrhythmia between patients who received antiarrhythmic drugs predicted to be effective on the basis of electrophysiologic study and those who received drugs predicted to be effective on the basis of Holter monitoring combined with exercise testing. The characteristics of the recurrences of arrhythmia were similar in the two groups. The probability of death from any cause, from a cardiac cause, or from a cardiac arrhythmia was not significantly different between the patients in the two groups who received a prediction of drug efficacy and continued to be treated. In addition, there were no significant differences in the recurrence of arrhythmia among all randomized patients in the two groups, regardless of treatment status. The principal difference between the two methods was a larger yield of predictions of drug efficacy with Holter monitoring9. These data indicate that the accuracy of the predictions made by the two methods is equivalent but that predictions of efficacy are obtained more readily by Holter monitoring.
In a previous randomized comparison of electrophysiologic study with Holter monitoring, Mitchell and coworkers8 concluded that electrophysiologic study was superior. They used an intention-to-treat approach for the primary analysis. In our intention-to-treat analysis of recurrence of arrhythmia, we did not find a significant difference between electrophysiologic study and Holter monitoring. The disparity between our study and that of Mitchell et al.8 may be explained in part by the fact that their sample was considerably smaller than ours -- 57 patients as compared with 486. In addition, our follow-up was longer. We recorded recurrences of arrhythmia over a period of 64 months, whereas the last recurrence in the study of Mitchell et al.8 was at approximately 20 months.
The findings of our trial may not be applicable to patients who do not have both inducible ventricular tachyarrhythmia and frequent spontaneous ventricular ectopy. Also, the extent to which our findings apply to patients presenting with cardiac arrest can be questioned, because only 108 of our randomized patients (22 percent) were survivors of cardiac arrest, and some clinicians are hesitant to use Holter monitoring alone to evaluate therapy in such patients. Multivariate analyses did not identify the type of presenting arrhythmia as a determinant of the accuracy of predictions of efficacy, however, and we did not detect any interaction between the type of presenting arrhythmia and the method of testing. Our findings may also not apply to patients for whom a decision to implant a cardioverter-defibrillator or similar device has already been made. We excluded patients with implanted cardioverter-defibrillators from the trial.
A relatively high proportion of all patients screened for enrollment in the study (23 percent) were randomized in the study trial. Nevertheless, it is likely that various practices and opinions of the investigators and referring physicians influenced the decisions to offer patients enrollment. There was wide variation in enrollment among the centers, which may have resulted from differing levels of effort to recruit patients. However, we did not find by multivariate analysis that the outcomes in patients from centers with low enrollments differed from the outcomes in patients from centers with high rates. Nevertheless, we cannot measure directly the possible influence of bias on the part of the physicians and investigators in the trial.
The observation that the risk of recurrence of arrhythmia among patients taking antiarrhythmic drugs predicted to be effective was lower in patients receiving sotalol is considered in detail in an accompanying paper10. The absence of failure of antiarrhythmic-drug therapy before enrollment in the trial was also associated with a lower probability of recurrence of arrhythmia. Other investigators have found this variable to be predictive of the suppression of inducibility on electrophysiologic study,17,18 but it has not previously been examined as a predictor of recurrence.
The probability of recurrence of arrhythmia among patients receiving drugs predicted to be effective was higher in our study than in most previous reports. Few reports include patients with a distribution of presenting arrhythmias similar to that of the study population. Mason and colleagues19 reported a two-year actuarial probability of recurrence of 22 percent in 60 patients in whom a similar distribution of ventricular tachyarrhythmias was treated with drugs selected on the basis of electrophysiologic study. Waller et al.20 and Rae and colleagues21 reported actuarial probabilities of recurrence of approximately 14 percent at 3 years and 15 percent at 2 1/2 years, respectively. Both these studies included a large proportion of patients with undiagnosed syncope, unsustained ventricular tachycardia, or undiagnosed palpitations.
Our higher rate of recurrence of arrhythmia may be due in part to differences in study design. No previous publications have reported on recurrences of arrhythmia as defined here. We included as recurrences asymptomatic episodes, documented episodes as brief as 15 beats in duration, and unmonitored syncopal episodes. Among the 150 recurrences studied in patients receiving drugs predicted to be effective, 22 (15 percent) were either unsustained runs lasting 15 beats to 15 seconds or unmonitored syncope.
Our method of detecting recurrences differs from those of most other reported studies in that we required frequent Holter monitoring on a fixed schedule. In addition, recurrences of arrhythmia were recorded prospectively according to written criteria, whereas in most previously reported studies data were collected retrospectively.
The patients in our study were required to have both inducible ventricular tachyarrhythmias and frequent spontaneous ventricular premature beats. We excluded patients who had only unsustained ventricular tachycardia and those who had previously responded to one of the antiarrhythmic drugs we used. These differences would be expected to increase the measured probability of a recurrence of arrhythmia in our study as compared with others. Another possible explanation for the high probability of recurrence in this study is the low level of effectiveness of the drugs we used (primarily Vaughan-Williams class I agents). In two of the previously cited studies,19,21 however, approximately the same proportion of patients received class I agents, and few received amiodarone. In the study by Waller et al.20 approximately half the patients received amiodarone. This drug was not used in our study because its long elimination half-life did not allow its inclusion in a random order of drug administration.
No clinical variables were clearly predictive of death among patients receiving a drug predicted to be effective, although treatment with sotalol and classification in class 1 on the symptom-activity scale of Goldman et al.16 were associated with trends toward reduced mortality from any cause. Clinical variables that often influence outcome, such as left ventricular ejection fraction and previous myocardial infarction, were not independent predictors of recurrent ventricular tachyarrhythmia or death in patients in whom a drug was predicted to have efficacy. Their apparent unimportance in determining survival may be related to homogeneity in those characteristics among our patients.
The lower mortality among patients with coronary artery disease in the electrophysiologic-study group in the intention-to-treat analysis may have resulted from differences in the treatments received by the patients treated according to the protocol and outside the protocol. A known difference in their care was the greater use of invasive and surgical therapies among those who did not respond to treatment in the electrophysiologic-study group. This difference in management may be explained by physicians' bias in interpreting the clinical meaning of failure to receive a prediction of efficacy on the basis of electrophysiologic study as compared with failure to receive a prediction with Holter monitoring. The higher proportion of crossovers to the alternate method of testing among the patients treated outside the protocol in the Holter-monitoring group supports the possibility of such a bias. This dissimilar treatment may also have contributed to the observed difference in mortality from any cause between the two groups among the patients who did not have predictions of efficacy.
Although the probability of recurrence of arrhythmia was higher in our study than in previous ones, the mortality due to arrhythmia of 10 ±2 percent at one year and 20 ±4 percent at four years was similar to or lower than that of most other studies. In the three reports cited above, the probabilities of death due to arrhythmia were about 16 percent at 1 1/2 years,3,19 approximately 7 percent at 2 years,20 and 16 percent at 2 1/2 years21.
The accuracies of different methods and criteria of efficacy for Holter monitoring and electrophysiologic study have not been rigorously compared. The results of this trial do not necessarily apply to the numerous other methods and criteria that are in use or that could be devised.
The design of this trial does not permit conclusions about the absolute value of electrophysiologic study or Holter monitoring for guiding therapy, because no empirical-therapy group (that did not receive guidance by either method) was included. The results of objective testing to predict drug efficacy have not yet been compared with the results of empirical administration of the same therapy in a large, randomized trial. Also, this study evaluated electrophysiologic study and Holter monitoring only with regard to the prediction of drug efficacy in patients with ventricular tachyarrhythmias; it did not assess the diagnostic use of either method.
Given the similar predictive accuracies of electrophysiologic study and Holter monitoring, other factors must be considered in deciding which method to use in patients with ventricular tachyarrhythmias who qualify for either technique of testing. Among those factors are the yield of predictions of efficacy, the duration of hospitalization, the cost of testing, the safety of the procedure, the comfort and convenience of the patient, and the experience and familiarity of physicians and staff members with the two methods.
Supported by a grant (RO-1-HL34071) from the National Heart, Lung, and Blood Institute, a grant (M01-RR00064) from the National Center for Research Resources, and by Bristol-Myers Squibb Company, Knoll Pharmaceutical Company, Boehringer-Ingelheim, Parke-Davis, and Ciba-Geigy.
Source Information
A complete list of the study investigators appears in the Appendix.
From the Cardiology Division, University of Utah School of Medicine, 50 N. Medical Dr., Salt Lake City, UT 84132, where reprint requests should be addressed to Dr. Mason.
References
The enrollment centers and study investigators participating in the ESVEM trial are listed below, with the number of patients randomized at each center given in parentheses. For each center, the first person listed was the principal investigator.
University of Utah, Salt Lake City (154): J.W. Mason, J.L. Anderson, K.P. Anderson, R.A. Freedman, L. Karagounis, D.A. Rawling, M. Hutson, D. Mannis, and M. Roskelley; Columbia University, New York (44): J.T. Bigger, Jr., J. Coromilas, F.D. Livelli, Jr., J. Reiffel, J.S. Steinberg, J. Campion, and A.M. Squatrito; University of Arizona, Tucson (43): F.I. Marcus, A. Caruso, H.L. Faitelson, T.E. Raya, Z. Garcia, K. Gear, and M.K. Pierce; University of Oklahoma, Oklahoma City (41): R. Lazzara, K. Beckman, K. Friday, W.M. Jackman, T. Deaton, K. Drennan, J. Foster, and S. Harris; University of New Mexico, Albuquerque (40): R.C. Klein, C. Machell, L. Widman, C. Acosta-Miller, and G. Lomeli; Baylor College of Medicine, Houston (33): C.M. Pratt, A. Boahene, A. Pacifico, C. Wyndham, and M. Francis; University of California, San Francisco (32): J.C. Griffin, M. Lesh, M.M. Scheinman, M. Wong, and M. Wong; University of Colorado, Denver (28): M.J. Reiter, D. Mann, T. Heyborne, and C. Kenney; Oregon Health Sciences University, Portland (28): J.H. McAnulty, J. Kron, B.D. Halperin, K. Sinner, and K. Martin; University of Pennsylvania-Presbyterian Medical Center, Philadelphia (20); L.N. Horowitz, C.D. Gottlieb, and C. Vrabel; University of Massachusetts, Worcester (14): S.K. Huang, R. Mittleman, P. Collett-Willey, and K. Rofino; Newark Beth Israel Medical Center, Newark, N.J. (5): S. Saksena, R.B. Krol, and L. Duque; University of British Columbia, Vancouver (3): C.R. Kerr, J.A. Yeung, and S. Vorderbrugge; and Northwestern University, Chicago (1): R. Kehoe, T.A. Zheutlin, T. Mattioni, and C. Dunnington.
Data Coordinating Center (University of Arizona): T. Moon, E. Hahn, V. Hartz, A. Rico, and N. Jenrow.
Safety Monitoring Committee: R. Bressler (chair), H.L. Greene, M. Lebowitz, T. Moon, and E. Morkin.
End Points Committee: W.D. Weaver (chair), T. Bump, F. Morady, and B. Olshansky.
| |||||||||||||||||||||||||||||||||||||||
Related Letters:
The Treatment of Ventricular Tachyarrhythmias
Damle R. S., Ehlert F. A., Winters S. L., Rubinstein D., Gomes J. A., Mason J. W., The ESVEM Investigators
Extract |
Full Text
N Engl J Med 1994;
330:286-288, Jan 27, 1994.
Correspondence
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | TERMS OF USE | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2009 Massachusetts Medical Society. All rights reserved. |
