Value of Programmed Ventricular Stimulation in High-risk Postinfarction Patients

Abstract & Commentary

Schmitt and colleagues from the Technical University of Munich conducted a prospective survey of 1436 patients at their hospital with a confirmed acute myocardial infarction between late 1994 and 1999. After confirmation of the acute infarction, all patients were managed with acute coronary angiography and then treated with either thrombolysis or a percutaneous intervention as appropriate. During the subsequent in-hospital recovery period, all patients underwent a noninvasive evaluation protocol that involved measurements of left ventricular ejection fraction (LVEF), spontaneous ventricular ectopy, heart rate variability, and late potentials on a signal-averaged (SA) ECG. Additional percutaneous interventions were performed on diseased, noninfarct-related arteries 7 to 14 days after infarction whenever appropriate. Patients who developed spontaneous, sustained, ventricular arrhythmias after the first 24 hours of infarction were excluded. A noninvasive risk-scoring system was used that assigned 3 points for an LVEF < 0.40 and 1 point each for an LVEF between 0.40 and 0.49, frequent (> 10 ventricular premature beats per hour) or complex (3 ³ consecutive beats) ventricular ectopy, abnormal heart rate variability (standard deviation of normal RR intervals < 80 msec), and a late potential on the SAECG.

Out of the total group of 1436 patients, 248 (17.6%) had a score of 3 or greater and met the predetermined definition of high risk. Of these, 54 were older than age 75 and were not invited to participate in the remainder of the trial. Of the remaining 194 "high risk" patients, 98 agreed to undergo programmed ventricular stimulation using a standard protocol similar to those used in the evaluation of patients with documented, sustained, ventricular arrhythmias. If sustained, monomorphic ventricular tachycardia (VT) with acycle length > 230 msec was induced with stimulation at the electrophysiologic study, the patient was offered an implantable cardioverter-defibrillator (ICD). All patients were then followed prospectively for the clinical end points of total mortality, cardiac mortality, sudden death, sustained VT, and appropriate ICD therapies.

Sustained, monomorphic VT was induced at the electrophysiologist study in 21 (22%) of the 98 patients in whom stimulation was performed. All but one of these received an ICD. During 607 ± 424 days to follow-up, 7 of the 21 patients with induced VT received appropriate ICD therapy. In contrast, only 2 of the 77 patients with a negative EP study experienced either sudden death or sustained VT. This latter group included 26 patients in whom either VT with a cycle length < 230 msec, polymorphic VT, or ventricular fibrillation had been induced. None of these 26 experienced an arrhythmic event. With only a single exception, all arrhythmic events occurred in patients with an LVEF below 0.40. Ejection fraction was a strong predictor of outcome with a mean LVEF of 24.5 ± 6.9% for those with an arrhythmia compared to 33.2 ± 7.6% among those without.

Schmitt et al conclude that programmed ventricular stimulation after noninvasive risk stratification is an effective tool for selecting postmyocardial infarction patients for prophylactic ICD therapy (Schmitt C, et al. J Am Coll Cardiol. 2001;37:1901-1907).

Comment by John P. DiMarco, MD, PhD

This study attempts to address 3 highly relevant questions. First, is noninvasive risk stratification for sudden death a useful or necessary step in evaluating patients after myocardial infarction? Second, are electrophysiologic studies with programmed ventricular stimulation useful in defining a high-risk subgroup who would then be candidates for prophylactic ICD therapy? Third, will ICD therapy reduce mortality and morbidity in the patients so identified? Although Schmitt et al claim that their study favors a "yes" answer to all 3 questions, the data presented here are inadequate for any conclusions firm enough to change current patient management.

Although 1406 patients were included in this study, Schmitt et al’s conclusions are based on the data obtained after electrophysiologic study in only 98 patients. The high event rate reported in those with inducible VT was based on only 7 events in 21 patients. Although these 7 events were classified as "appropriate" ICD therapies, the number of ICD therapy end points is usually greater than the number of clinical, symptomatic arrhythmias in an otherwise comparable population to a degree that is difficult to estimate accuracy. Thus, the positive predictive value reported here should be interpreted as having a wide confidence interval. Other larger studies have described a lower positive predictive value for programmed ventricular stimulation. In the Multicenter Automatic Implantable Defibrillator Trial, the 2- year mortality rate was 35% among the patients with induced VT who did not receive an ICD, but more than half of the deaths were not caused by an arrhythmia.1 In the Multicenter Unsustained Tachycardia Trial, the 2- year cardiac arrest or arrhythmic event rate was only 18% among those with inducible VT vs. 12% among those without.2,3 Although these latter 2 trials directly or indirectly supported the value of prophylactic ICD therapy, their data suggest that the current paper may well overestimate the magnitude of benefit that is likely to be achieved.

The approach favored by Schmitt et al involved multiple noninvasive tests in an attempt to select a group of high-risk patients for the EP study. However, LVEF alone appears to be the most significant predictor—an observation made by earlier studies. Since each noninvasive test adds cost, time, and complexity to the post myocardial infarction protocol, approaches based solely on LVEF seem more attractive, and this approach is currently being evaluated in clinical trials with sufficient statistical power to answer this question. We should wait until these trials are concluded before we select a more complex format for evaluating potential candidates for a prophylactic ICD.


1. Moss AJ, et al. N Engl J Med. 1996;335:1933-1940.

2. Buxton AE, et al. N Engl J Med. 1999;341:1882-1890.

3. Buxton AE, et al. N Engl J Med. 2000;342:1937-1945.