Ejection Fraction vs Inducible VT for Predicting Morbid Events

Abstract & Commentary

Synopsis: Both low ejection fraction and inducible ventricular tachycardia can be used to identify patients with coronary disease at increased mortality risk.

Source: Buxton AE, et al, for the MUSTT Investigators. Circulation. 2002;106:2466-2472.

The multicenter, unsustained tachycardia trial (MUSTT) was a randomized clinical trial designed to test whether antiarrhythmic therapy guided by electrophysiologic testing would reduce arrhythmic death and total mortality in high-risk patients. The patients selected for the trial had documented coronary artery disease (CAD) with a left ventricular ejection fraction of 40% or less and asymptomatic nonsustained tachycardia. As part of the protocol, patients underwent a baseline electrophysiologic (EP) study with ventricular stimulation. Patients who did not have an inducible ventricular tachycardia that met study criteria were followed in a registry. Patients who did manifest a ventricular arrhythmia at EP study were randomized to either no drug therapy or antiarrhythmic therapy guided by serial electrophysiologic studies. A total of 2202 patients were enrolled in the study. Antiarrhythmic therapy was not used in 1791 of the patients either because they had no inducible arrhythmia (1362 patients), were randomized to no antiarrhythmic therapy (332 patients), or were not treated with antiarrhythmic drugs for various reasons (97 patients). Groups of patients with ejection fractions below 30% and above or equal to 30% were then identified for a cohort with and without inducible arrhythmias. Events during the trial were reviewed by an events committee. The description of events were redacted by the data coordinating center and reviewed by the events committee without knowledge of therapy. Deaths were classified as either arrhythmic or nonarrhythmic cardiac deaths or noncardiac deaths.

The 5-year mortality rate of all patients with ejection fractions less than 30% was, as expected, significantly higher than that of patients having an ejection fraction greater than or equal to 30% (54% vs 36%, P = 0.0001). The 5-year risk of arrhythmic death or cardiac arrest for all patients with ejection fractions less than 30% (33%) was also significantly higher than that of patients having ejection fractions greater than or equal to 30% (33% vs 20%, P = 0.0001). Multivariate analysis confirmed that both ejection fraction and the presence of inducible ventricular tachycardia were independent predictors of total mortality, arrhythmic death, and cardiac arrest. The ability to induce ventricular tachycardia was only a modest predictor of total mortality (hazard ratio 1.22), but it had a stronger relationship with arrhythmic death or cardiac arrest (hazard ratio 1.62). When the mode of death was analyzed, it was observed that patients who had inducible ventricular tachycardia were more likely to have an arrhythmic cause of death as opposed to a nonarrhythmic cause of death. This was a trend among patients with an ejection fraction less than 30% but did reach statistical significance among those with ejection fractions greater than or equal to 30%. Increasing age was a predictor of nonarrhythmic death. Fatal events in older patients were significantly less likely to be arrhythmic than fatal events in younger patients.

Buxton and associates conclude that both low ejection fraction and inducible ventricular tachycardia can be used to identify patients with coronary disease at increased mortality risk. The ability of an electrophysiologic study to determine high risk seems to be significant merely in those whose ejection fractions are greater than 30%.

Comment by John P. DiMarco, MD, PhD

The data presented in this paper confirm that ejection fraction is a powerful predictor of death. Buxton et al argue that EP studies will be helpful in selecting patients who have a greater propensity for arrhythmic death, particularly in patients with ejection fractions between 30% and 40%. However, although these differences are statistically significant, it is not certain whether these differences will be clinically helpful. For example, among patients with ejection fractions of ³ 30%, the 5-year mortality was 43% if they manifested an inducible arrhythmia and 34% if they did not. Does this 9% difference between the groups justify classifying them into high- and low-risk groups? For arrhythmic death or cardiac arrest, the differences are a bit more significant. For example, at 2 and 5 years, the event rates among those with an ejection fraction greater than 30% were 16% and 30% among those who had inducible arrhythmias and 8% and 17% among those who did not. Thus, it is clear that induction of an arrhythmia is a statistically significant risk factor, but unfortunately, it is not enough to really stratify individual patients into low- and high-risk groups.

Clinicians often deal with patients in whom they have to weigh their risks for future events. Since the intervention that has been shown to favorably affect mortality in this population has been the implantable cardioverter defibrillator (ICD), doctors have looked for markers that would stratify patients into very high-risk vs very low-risk groups. Unfortunately, none of the markers really do that. Each serves as a marker for a small increment in risk. If one requires that patients have many "high risk" markers, only a few will be identified, but there will be many events in the large pool of "lower risk" patients. In addition, if numerous tests are used to screen patients, not only does the sensitivity for detection of sudden death victims decrease, but also the cost of screening becomes substantial.

Other studies have shown that there was no benefit of the ICD over amiodarone therapy in secondary prevention of sudden death. It may well be that amiodarone or other drug therapy can be effective in patients with better ventricular function, but that the benefits of the drug are overwhelmed in patients with markedly decreased ejection fractions. It will be interesting to see if data from the still ongoing Sudden Cardiac Death Heart Failure Trial, which compares amiodarone, ICD therapy, and placebo in patients with ejection fractions below 35%, will settle this issue.

Dr. DiMarco is Professor of Medicine, Division of Cardiology University of Virginia, Charlottesville.