Sustained ventricular tachycardia induced by programmed electrical stimulation has been shown to predict death—specifically sudden death—in myocardial infarction survivors who have reduced ventricular function and present with asymptomatic ventricular ectopy or nonsustained ventricular tachycardia [9-12]. We evaluated whether electrocardiographic characteristics of spontaneous nonsustained ventricular tachycardia could predict the inducibility of sustained ventricular tachycardia by programmed electrical stimulation in a large sample of patients with documented coronary artery disease, reduced ventricular function, and spontaneous asymptomatic nonsustained ventricular tachycardia.

Methods

Top Methods Results Discussion Author & Article Info References

Patients

We did an analysis of the first 1480 patients enrolled in the ongoing Multicenter Unsustained Tachycardia Trial (MUSTT), which is testing whether electrophysiologically guided anti-arrhythmic therapy can reduce the risk for sudden cardiac death in high-risk patients with coronary artery disease. The major criteria for trial entry are documented coronary artery disease, left ventricular ejection fraction of 0.40 or less, and asymptomatic spontaneous nonsustained ventricular tachycardia. The protocol is described elsewhere [13].

Nonsustained ventricular tachycardia is defined as three or more consecutive premature ventricular complexes that last no more than 30 seconds and terminate spontaneously. At least two consecutive cycles must last 550 milliseconds or less (equivalent to more than equals 110 beats/min), and the average rate of the entire episode must be more than 100 beats/min. The qualifying nonsustained ventricular tachycardia must have been documented within 6 months before entry into the trial. In addition, the qualifying tachycardia must have occurred at least 4 days after the most recent myocardial infarction, coronary bypass grafting, or percutaneous transluminal coronary angioplasty. The qualifying tachycardia must not have been attributable to transient metabolic abnormalities such as myocardial ischemia and must not have produced symptoms severe enough to warrant treatment. All patients have at least 18 hours of analyzable continuous electrocardiographic monitoring.

Electrophysiologic Studies

The protocol for programmed electrical stimulation, standardized at all participating centers, includes delivery of 1, 2, and 3 ventricular extrastimuli at two right ventricular sites at pacing cycle lengths of 600 and 400 milliseconds. In addition, bursts of 5 to 15 ventricular stimuli at cycle lengths of 350 to 250 milliseconds are delivered. The end point for the stimulation protocol is the reproducible induction of sustained ventricular tachycardia or fibrillation or completion of the protocol.

Data Abstraction

Patient data, including characteristics of the spontaneous nonsustained ventricular tachycardia, are reported to the data coordinating center. The electrocardiographic characteristics of nonsustained ventricular tachycardia episodes are assessed and reported by individual investigators. To assure that the qualifying nonsustained ventricular tachycardia meets the entry criteria, the recordings of every episode of qualifying nonsustained ventricular tachycardia are reviewed by executive committee members who are blinded to other patient characteristics.

For our study, accuracy of the nonsustained ventricular tachycardia characteristics recorded by the participating investigators was checked in several additional ways. Electrocardiograms in a subset of data from 222 patients were reviewed. We also reviewed 1) all electrocardiographic tracings of patients with inducible ventricular tachycardia submitted by the eight study sites enrolling the most patients with inducible sustained ventricular tachycardia and 2) data from a random sample of 3 patients without inducible sustained ventricular tachycardia from each study site. The rate and duration of ventricular tachycardia reported on the data forms were compared with the rate of ventricular tachycardia on the submitted electrocardiograms, as measured by one author.

Statistical Analysis

Characteristics of spontaneous nonsustained ventricular tachycardia were summarized in patients with and those without inducible monomorphic sustained ventricular tachycardia using the median and the 5th, 25th, 75th, and 95th percentiles. The distributions of each characteristic in these two groups (inducible and noninducible) were compared using the nonparametric Wilcoxon rank-sum test [14].

Further statistical assessment of the relation of characteristics of the spontaneous nonsustained ventricular tachycardia with inducibility was done using logistic regression analysis. Inducibility during programmed electrical stimulation was the outcome variable, and characteristics of the nonsustained ventricular tachycardia were predictor variables. The relation between each predictor variable and inducible ventricular tachycardia was considered individually and jointly. To accommodate possible nonlinear relations between these predictors and inducibility, a flexible model-fitting approach involving cubic splines (piecewise cubic polynomials) was used [15, 16]. Characteristics of the nonsustained ventricular tachycardia examined in these regression analyses included mean duration of the episodes documented, duration of the longest episode, mean rate of the episodes, and mean number of episodes per day. The ability of these characteristics (individually or jointly) to accurately predict inducible sustained ventricular tachycardia or discriminate between patients who did and did not have inducible sustained ventricular tachycardia with programmed stimulation was quantitated using the area under the receiver-operating characteristic (ROC) curve [17]. An area under the ROC curve near 0.5 indicates that predictions of inducibility made using characteristics of the nonsustained ventricular tachycardia are essentially random with respect to actual inducibility. An area under the ROC curve near 1.0 indicates that the predictions correctly discriminate between patients who did and did not have inducible ventricular tachycardia.

Results

Top Methods Results Discussion Author & Article Info References

Data were analyzed from the first 1480 consecutive patients enrolled in the trial. Of these patients, 438 (30%) had inducible sustained monomorphic ventricular tachycardia, and 1042 (70%) did not.

When the electrocardiographic tracings in the subset of 222 patients were reviewed, the cycle length and duration of the recordings of qualifying nonsustained ventricular tachycardia compared with the values reported on the case report form showed a mean difference in cycle length of only 25 milliseconds. The median duration of episodes in the sample was five beats, the same as the entire population. Thus, the submitted sample data were very close to the data reported by individual investigators on the case report forms.

Spontaneous nonsustained ventricular tachycardia qualifying patients for study entry was discovered during in-hospital telemetric electrocardiography in 793 (53%) of the 1480 patients, by continuous ambulatory electrocardiographic (Holter) recording in 609 (41%), during exercise electrocardiography in 39 (3%), and in the course of recording a standard 12-lead electrocardiogram in 25 (2%). One episode was documented by a transtelephonic event recorder, and the source of the documentation could not be obtained in 13 patients.

The median duration of all recorded episodes of spontaneous nonsustained ventricular tachycardia was five complexes. The median value for the longest episode of documented nonsustained ventricular tachycardia was seven complexes. The median value for the mean cycle length of recorded episodes of spontaneous nonsustained ventricular tachycardia was 420 milliseconds (equivalent to a rate of 143 beats/min). The median frequency of documented nonsustained ventricular tachycardia was two episodes per day (mode, 1 episode/d).

Morphologic data on the spontaneous nonsustained ventricular tachycardia are limited. Nonsustained monomorphic ventricular tachycardia was recorded in 86% of patients, and polymorphic nonsustained ventricular tachycardia was recorded in 34% of patients (some patients had both events). However, the accuracy of the morphologic classification is compromised by the fact that spontaneous nonsustained ventricular tachycardia was recorded on two electrocardiographic leads in 75% of patients (12-lead electrocardiograms of nonsustained ventricular tachycardia are rarely available).

No significant differences were seen between patients with and those without inducible sustained monomorphic ventricular tachycardia in the distributions of the electrocardiographic characteristics, except for a slightly higher rate of nonsustained ventricular tachycardia in patients without inducible sustained ventricular tachycardia (Table 1). The magnitude of difference in rates is small, and there is such considerable overlap in the heart rate distributions in both patient groups that the difference is not clinically useful.

No relations between characteristics of spontaneous nonsustained ventricular tachycardia with inducibility were significant by logistic regression analysis, except for the mean rate variable (Table 2). However, even for this characteristic, the relation does not provide the ability to accurately discriminate between patients who have inducible ventricular tachycardia and those who do not, as quantified by the area under the ROC curve (Table 2). Even when all four characteristics of spontaneous nonsustained ventricular tachycardia were considered together in a regression model to predict inducibility, the area under the ROC curve was only 0.552, reflecting almost no predictive ability. When any inducible sustained ventricular tachycardia (monomorphic or polymorphic) was considered as the measure of inducibility, the results were similar.

Discussion

Top Methods Results Discussion Author & Article Info References

Counter to intuition, electrocardiographic characteristics of spontaneous nonsustained ventricular tachycardia in patients with coronary artery disease cannot differentiate between patients with and those without sustained ventricular tachycardia inducible by programmed stimulation in the electrophysiology laboratory. The importance of this finding lies in the fact that programmed stimulation is one of the best techniques for identifying patients with coronary artery disease who are at high risk for sudden cardiac death. Other tests, such as measurement of left ventricular ejection fraction and the ambulatory electrocardiogram, can predict overall cardiac-related death but not whether death will or will not occur suddenly. Thus, these other tests cannot identify patients likely to benefit from antiarrhythmic therapy. In contrast, previous studies [9-12] have associated the presence of sustained ventricular tachycardia with an excess of sudden, as opposed to nonsudden, deaths. The signal-averaged electrocardiogram has yet to be tested prospectively in a large sample of patients with the characteristics of our study participants.

The noted electrocardiographic characteristics of spontaneous nonsustained ventricular tachycardia are remarkably similar to those reported in the Cardiac Arrhythmia Suppression Trial (CAST) [8]. We have confirmed that most episodes of nonsustained ventricular tachycardia in patients with coronary artery disease who are at high risk for sudden cardiac death are brief and occur infrequently. An analysis of cardiac arrest in CAST showed no electrocardiographic characteristic of nonsustained ventricular tachycardia that distinguished patients who had had cardiac arrest from those who had not. This result was similar to that of a previous analysis that did not control anti-arrhythmic therapy in survivors with acute infarction [3].

Limitations

We did not use a core laboratory to analyze the Holter tapes. Investigators at each study site reviewed the electrocardiographic data and reported it on forms for entry into the computer database. Although analysis of Holter tapes by a core laboratory might have increased data accuracy, our study funds were insufficient to pay for Holter recording or analysis. Various recording systems were used: The systems in use depended on the study site and its referral sources. Despite this limitation, our findings are remarkably similar to those of previous studies that did use core laboratories. One advantage of our methods is that our data reflect the results that are likely to occur in actual clinical practice. Thus, these results should be generalizable to the entire patient sample meeting the study's entry criteria.

A second limitation, noted previously, is that with current technology, most episodes of spontaneous nonsustained ventricular tachycardia can only be recorded on one or two electrocardiographic leads. This limits our ability to accurately describe the morphologic characteristics of such arrhythmias.

Conclusions

No electrocardiographic characteristic of spontaneous nonsustained ventricular tachycardia can accurately predict whether patients will have inducible sustained ventricular tachycardia. Thus, we conclude that the rate, duration, and frequency of spontaneous nonsustained ventricular tachycardia in patients with abnormal left ventricular function caused by coronary artery disease cannot be used to make clinical decisions about the use of electrophysiologic studies for risk stratification.

Appendix: MUSTT Trial Investigators

Dr. Lee: Multicenter Unsustained Tachycardia Trial (MUSTT) Data Coordinating Center, Duke University Medical Center, Box 3363, Durham, NC 27710.

Dr. DiCarlo: Michigan Heart P.C., 5325 Elliott Drive, Suite 203, Ypsilanti, MI 48197.

Dr. Echt: Vanderbilt University Medical Center, Cardiology Division, Room 315 MRB II, Nashville, TN 37232.

Dr. Fisher: Montefiore Medical Center, Department of Cardiology, Henry & Lucy Moss Hospital Division, 11 East 210th Street, Bronx, NY 10467.

Dr. Greer: Arkansas Cardiology Clinic, Suite 600, Medical Towers II, 9501 Lile Drive, Little Rock, AR 72205.

Dr. Josephson: Division of Cardiology, RW 453, The Beth Israel Hospital, 330 Brookline Avenue, Boston, MA 02215.

Dr. Packer: Mayo Clinic, 1216 2nd Street, Southwest, 4-416 Alfred Building, Rochester, MN 55905.

Dr. Prystowsky: Northside Cardiology, St. Vincent Professional Building, 8402 Harcourt Road, Suite 228, Indianapolis, IN 46260.

Dr. Talajic: Montreal Heart Institute, 5000 East Belanger Street, Montreal, Quebec H1T1C8, Canada.