Clinical Prediction of Inducible Sustained Ventricular Tachycardia
abstract & commentary
Synopsis: Clinical factors have limited ability to accurately predict induction of sustained VT in patients with coronary artery disease and low ejection fractions.
Source: Buxton AE, et al, for the Multicenter Unsustained Tachycardia Trial (MUSTT) Investigators. Circulation 1999;99:1843-1850.
In this study, buxton and colleagues report on clinical features that were associated with inducible sustained ventricular tachycardia (VT) in the Multicenter Unsustained Tachycardia Trial (MUSST). MUSST recruited patients with known coronary artery disease (CAD), a left ventricular ejection fraction (LVEF) less than 0.40, and asymptomatic nonsustained VT. The design of the study included a baseline electrophysiologic (EP) study that used a standard protocol for programmed ventricular stimulation. The end point for the stimulation protocol was initiation of either sustained monomorphic VT at any point or sustained polymorphic VT or ventricular fibrillation with one or two extra stimuli. After their EP study, patients without an inducible sustained arrhythmia were followed on routine therapy while those with an inducible arrhythmia were randomized between therapy guided by serial EP studies and no specific antiarrhythmic therapy.
Data from 1721 patients who underwent the baseline EP study were analyzed in this report. Sustained monomorphic VT was inducible in 549 patients (32%) and an additional 63 patients had inducible polymorphic tachycardias. Variables most strongly associated with induction of monomorphic VT by univariate analysis included the following: male gender, more fixed thallium defects, clinical history of myocardial infarction (MI), absence of other cardiac disease, beta-blocker therapy at enrollment, areas of dyskinesis, congestive heart failure or sustained VT or VF at the time of MI, and longer duration from last MI (all P < 0.01). Patients with prior coronary artery bypass were less likely to have an inducible arrhythmia. These univariate predictors were then included in several models that combined multiple variables. Receiver operating curves (ROC) were used to quantify the ability of these models to predict who would have inducible VT. An ROC area near 1.0 would signify perfect prediction, while an area of 0.5 correlates with no predictive value. The three models tested yielded ROC areas between 0.647 and 0.692. Buxton et al conclude that clinical factors have limited ability to accurately predict induction of sustained VT in patients with CAD and low EFs.
Comment by John P. DiMarco, MD, PhD
Several years ago, the Multicenter Automatic Defibrillator Implantation Trial (MADIT—N Engl J Med 1996;335:1933-1940) reported that implantable cardioverter defibrillator (ICD) implantation reduced mortality by 75% in patients with prior MI, low LVEF, nonsustained clinical VT, and inducible sustained VT. On the basis of this trial, induction of VT at EP study became an acceptable class I indication for ICD implantation. MADIT, however, did not systematically collect data on patients who had negative EP studies. Therefore, it has been difficult to characterize the population from which the MADIT subjects were drawn. In MUSST, data from patients both with and without inducible VT were collected. This study shows that clinical factors only add minor additional information to the basic findings of a low EF and documented nonsustained VT. Previously, the MUSST investigators had also reported that electrocardiographic characteristics of the nonsustained VT did not predict inducibility (Buxton AE, et al. Ann Int Med 1996;125:35-39).
Should these data be taken to imply that all patients with CAD, an LVEF less than 40, and nonsustained VT should undergo an EP study? The data from MADIT and the poor ability to use clinical features to select patients with inducible VT shown in this study would support that approach. However, in further data reported from MUSST at the recent North American Society for Pacing and Electrophysiology meeting, patients without inducible VT were compared to those patients with inducible VT who were not treated. Although a statistically significant higher mortality was shown among those with inducible VT, the increment of risk was not great. Both groups had relatively high mortality rates.
The conclusion to be reached from these data is that we have great difficulty further separating out ultra-high risk groups if we start with patients who already have severely depressed EFs. Rather, we should probably consider all these patients to be high risk and proceed accordingly.
In patients with CAD, systolic dysfunction, and nonsustained VT on ECG monitoring, inducible VT is reliably predicted by:
a. prior MI.
b. heart failure.
c. VT during prior acute MI.
d. None of the above