Autonomic Tone Post MI

Abstract & Commentary

Synopsis: Depressed baroreceptor sensitivity and the presence of nonsustained VT are independent risk factors for both cardiac mortality and arrhythmic events after MI.

Source: LaRovere MT, et al. Circulation. 2001;103:2072-2077.

The Autonomic Tone and Reflexes After Myocardial Infarction (ATRAMI) project was a study that examined the predictive value of analysis of autonomic nervous system function in survivors of acute myocardial infarction. The study, whose primary end points were reported several years ago, enrolled 1284 patients younger than 80 years of age with a myocardial infarction within 1 month of enrollment.1 This substudy reports the prognostic significance of the following variables in ATRAMI: left ventricular ejection fraction (LVEF), nonsustained ventricular tachycardia (VT), abnormal heart rate variability (HRV), and baroreceptor sensitivity. These definitions were used: nonsustained VT was present if 3 or more consecutive ventricular beats at a rate of over 100 beats per minutes were seen on a 24-hour holter recording. Abnormal HRV was defined as a standard deviation of normal-to-normal RR intervals less than 70 msec. Depressed baroreceptor sensitivity was present if there was less than 3 msec heart rate slowing per mm Hg blood pressure increase during a phenylephrine infusion. Patients were dichotomized at a LVEF of 35%. The end points analyzed were total cardiac mortality and a combined end point of sudden death and/or sustained VT. Associations were assessed with univariate and multivariate Cox analysis and relative risks calculated.

Complete data were available on 1071 patients. The mean age was 59 ± 10 years and the mean LVEF was 49 ± 11%. The LVEF was less than 35% in 157 patients (14.6%). During 21 ± 8 months of follow-up there were 43 cardiac deaths or nonfatal cardiac arrests and 5 episodes of sustained ventricular tachycardia.

Life-table analysis of mortality showed that patients with both nonsustained VT and depressed baroreceptor sensitivity had a higher mortality (21%) than those without either of these findings (2.4%). Patients with either nonsustained VT or depressed baroreceptor sensitivity, but not both, had an intermediate mortality of 7.5%. Almost identical observations were made for patients with abnormal HRV with an overall mortality of 29% for patients with nonsustained VT and abnormal HRV, a 2.5% mortality for those with neither finding, and a 6-7% mortality for those with 1 of the findings. Multivariate Cox analysis showed that the combination of nonsustained VT plus abnormal HRV increased the relative risk for cardiac mortality to 17, and the combination of nonsustained VT plus depressed baroreceptor sensitivity resulted in a relative risk of 9.6. In patients with all 3 of these findings, the relative risk for cardiac death was 22.2. The prognostic significance of these variables was also examined in the subgroup of patients with ejection fractions less than 35%. Both nonsustained VT and depressed baroreceptor sensitivity maintained an independent prognostic association with cardiac mortality in this group but this was not the case for abnormal HRV. In patients with an ejection fraction of less than 35% who did not have nonsustained VT, depressed baroreceptor sensitivity still influenced mortality risk with a relative risk of 4.1.

LaRovere and colleagues conclude that depressed baroreceptor sensitivity and the presence of nonsustained VT are independent risk factors for both cardiac mortality and arrhythmic events after MI. A reasonable compromise between sensitivity and specificity is achieved by combining reduced LVEF with the presence of either nonsustained VT or depressed baroreceptor sensitivity.

Comment by John P. DiMarco, MD, PhD

Effective identification of patients at risk for death after MI remains challenging. Ejection fraction and the presence of nonsustained VT are commonly used markers for an adverse prognosis. Recently, the Multicenter Automatic Defibrillator Implantation Trial directly, and the Multicenter Unsustained Tachycardia Trial indirectly indicated that implantable cardioverter defibrillator (ICD) therapy reduced total mortality, largely by preventing sudden death. However, ICD therapy is quite expensive using current systems; and more precise methods to identify the patients in whom prophylactic ICD therapy would be the most cost-effective are needed.

The ATRAMI study tested whether baroreflex sensitivity and heart rate variability would be effective predictors of mortality after myocardial infarction. Nonsustained VT is the most commonly used predictor of arrhythmic risk but use of nonsustained VT in decisions regarding an individual patient is limited by day-to-day variability in the postinfarction period. Baroreflex sensitivity, a test infrequently performed in North America, and heart rate variability are measures of autonomic nervous system tone that reflect sympathetic predominance. They should remain stable from day-to-day unless the patient’s clinical status changes. This study shows that these 2 markers, particularly when combined with the presence of nonsustained VT and a low LVEF, can help identify a group of high-risk patients who would have a substantial probability of benefiting from ICD implant insertion.

Unfortunately, there was only a relatively small number of patients with low LVEF (157) and this resulted in large confidence intervals for any of the estimates given. However, it seems clear that patients with a low LVEF who have either nonsustained VT and abnormal baroreceptor sensitivity, or both, constitute a group in whom studies to assess the benefits of prophylactic ICD therapy would be appropriate.

Unfortunately, this paper does not contain data analyzing a number of other factors associated with outcome after MI. Patients who had recurrent ischemia and underwent bypass surgery or patients with signs or symptoms of heart failure were excluded. In addition, the use of other drugs, including statins, ACE inhibitors, and beta blockers, which have been shown to favorably influence survival, is not discussed in this manuscript. It is likely that a much larger study population would be required to include these variables in a survival analysis when planning any clinical trials. 


1. LaRovere M, et al. Lancet. 1998;351:478-484.