| J Thomas Bigger, Jr, MD |
Feb 8, 2000 |
There is a large body of clinical and experimental evidence that indicates an important role for the autonomic nervous system in the triggering or sustaining of malignant ventricular arrhythmias [1,2,3,4,5]. Experimentally, sympathetic stimulation reduces the ventricular refractory period and the ventricular fibrillation threshold, promotes triggered activity afterpotentials, and enhances automaticity. All of these are factors that promote arrhythmia. Vagal stimulation opposes these changes and reduces the effects of sympathetic stimulation; it prolongs refractoriness, elevates the ventricular fibrillation threshold, and reduces automaticity.
There are two noninvasive or minimally invasive measures to evaluate the autonomic nervous system in intact humans:
• RR interval (heart rate) variability
• Baroreflex sensitivity (BRS)
RR variability measurements provide a low-cost, widely available method for assessing the status of the parasympathetic nervous system in humans and for predicting cardiovascular events, especially cardiac death and sustained ventricular arrhythmias in coronary heart disease patients. BRS also is useful for predicting cardiovascular events and adds prognostic information to RR variability.
This card will review the clinical applications of RR variability other than after myocardial infarction. The technical aspects of this modality, its predictive value in patients who have had a myocardial infarction, and the use of BRS are discussed separately. (See "Heart rate variability: Technical aspects" and see "Heart rate variability: Use after myocardial infarction").
Variation in RR intervals can be measured by many methods which can be categorized as time domain measures or frequency domain measures (show table 1) [6]. The measures most often used clinically include:
• Standard deviation of NN (normal to normal RR) intervals over a 24-hour period (SDNN)
• Standard deviation of the average NN intervals for the 288 five-minute intervals in a 24-hour continuous ECG recording (SDANN)
• Heart rate variability (HRV) triangular index
• Total power, ultra-low frequency (ULF) power, very low frequency (VLF) power, low-frequency (LF) power, high frequency (HF) power, the ratio of LF/HF, and power law regression parameters.
The data documenting the reproducibility of these measurements and the role of the autonomic nervous system in promoting the development of arrhythmias are discussed elsewhere. (See "Heart rate variability: Use after myocardial infarction").
CLINICAL USES OF HEART RATE VARIABILITY ！ As stated by the Task Force on Heart Rate Variability, there are two proven clinical uses of RR variability [7]:
• Prediction risk of cardiac death or arrhythmic events after acute myocardial infarction
• Detection and quantification of autonomic neuropathy in patients with diabetes mellitus
The most frequent use of RR variability is for risk stratification for cardiac and arrhythmic death after a myocardial infarction. However, this technique has been used in a number of other settings. The predictive value of RR variability for nonfatal myocardial infarction or death when measured in patients who have angina pectoris without previous myocardial infarction is not clear. Similarly, it has not yet been definitively demonstrated that RR variability predicts mortality, particularly arrhythmic death, in patients with nonischemic dilated cardiomyopathy, hypertrophic cardiomyopathy, valvular heart disease, or congenital heart disease. It has been suggested that, at the present time, the measurement of RR variability is a research technique, but not a routine clinical tool [8]. In 1999 a task force of the American College of Cardiology/American Heart Association has published guidelines for the routine use of ambulatory monitoring to determine heart rate variability (show table 2) [9].
Low RR variability in the elderly ！ Data from the Framingham Heart study suggest that reduced RR variability predicts mortality in a population-based sample of elderly subjects. In a study of 1825 surviving participants, 736 with an average age of 72 years who underwent two hour ambulatory ECG monitoring were selected for analysis [10]. The clinical characteristics of the participants were similar to those of the nonparticipants. Frequency domain analysis was performed on 100 sec blocks of data, sampled at 128 samples/sec, and power spectral density was calculated using a fast Fourier transform method.
After an average follow-up of 3.9 years, LF power was the best of eight time and frequency domain measures of RR variability for predicting death of all causes with a relative risk of 1.87. After statistical adjustment for age, sex, and clinical risk factors (history of myocardial infarction, heart failure, diuretic use, and frequent or complex ventricular premature complexes), the relative risk was 1.70.
Low RR variability in patients referred for 24-hour Holter recording ！ The hypothesis that a disturbance in autonomic nervous system activity may play a role in sudden cardiac death was evaluated in 6693 consecutive patients who underwent 24-hour ambulatory electrocardiographic recordings; the reasons for the test included evaluation of palpitations, dizziness, syncope, or angina pectoris; risk assessment after myocardial infarction; evaluation of antiarrhythmic therapy; or as a search for a cardiac cause of transient cerebral events [11]. Three time domain measures of RR variability were used: pNN50, short-term variation (mean over 24 hours of per-minute standard deviations of RR intervals), and long-term variation (standard deviation over 24 hours of per-minute means of RR intervals). Although the last two measures are not orthodox, the measure of short-term variation is similar to ASDNN or LF power and the measure of long-term variation is similar to SDANN or ULF power.
After a two-year follow-up period, patients with low values for these three measures of RR variability had a higher relative risk of experiencing sudden cardiac death (1.8 for pNN50, 3.0 for short-term variation, and 2.7 for long-term variation). These results indicate that RR variability has predictive value in a consecutive, heterogeneous group of patients referred to a Holter laboratory.
Effect of exercise on heart rate variability ！ Exercise training produces a reduction in resting heart rate, probably due to an increase in parasympathetic tone. In a study of 13 healthy older (
60
years) and 11 healthy younger patients (
32
years) who underwent six months of intensive aerobic training, heart rate
variability in the time domain at rest increased by 68 percent in older
patients and 17 percent in the younger patients [12].
This increase in parasympathetic tone may, in part, account for the benefit of
exercise in reducing all cause and cardiovascular mortality. (See
"Preventive cardiology: Role of exercise").RR variability in heart failure ！ There are limited data that RR variability may be of predictive value in dilated cardiomyopathy and congestive heart failure in general [13,14,15,16,17]. As an example, one study of 64 patients with dilated cardiomyopathy found that measures of RR variability were reduced compared to controls and that reduced RR variability was associated with disease severity measures such as NYHA functional class, left ventricular diastolic dimension, reduced left ventricular ejection fraction, and peak O2 consumption [13]. At 12 months, patients with an SDNN less than 50 ms had a lower survival rate free of progressive heart failure those those with a higher value (p = 0.0002).
A much larger group of 433 outpatients with congestive heart failure was evaluated in the UK-HEART study [17]. After a follow up of 482 days, SDNN was found to be an independent predictor of all-cause mortality and the most powerful predictor of death from progressive heart failure. The risk ratio for 41.2 ms decrease in SDNN was 1.62. The annual mortality for those with SDNN of >100 ms, 50 to 100 ms, and <50 ms was 5.5, 12.7, and 51.4 percent, respectively (show figure 1). In another study of 116, a reduced SDNN <100 ms was predictive of arrhythmic events and sudden cardiac death (show figure 2) [18].
The predictive value of depressed heart rate variability appears to be independent of other known risk factors, including NYHA functional class, ejection fraction, peak O2 consumption, and the presence of ventricular arrhythmia [16].
RR variability in spontaneous ventricular tachycardia ！ Changes in autonomic tone play an important role in arrhythmogenesis. Increased sympathetic activity, manifested as an increase in heart rate, precedes the onset of sustained ventricular tachycardia (VT) in the majority of patients in whom the initiation of the VT has been recorded [19,20]. In one study of 47 patients with spontaneous VT recorded on 24 hours of ambulatory monitoring, all patients had an increase in heart rate 15 minutes prior to VT onset [21]. However, 68 percent had a decrease in total power, LF power, and LF/HF ratio 15 minutes prior to VT, and these patients had a higher baseline level of these parameters during the two hours before VT. The remaining patients who had a lower baseline level of total power, LF power, and LF/HF ratio had an increase in these parameters prior to VT. These data suggest that change in the dynamics of RR intervals, rather than the direction of change, facilitates VT induction in most patients.
RR variability to study drug effects ！ A number of drugs can influence heart rate and may have an effect on RR variability. These include atropine, beta blockers, calcium channel blockers, digoxin, angiotensin converting enzyme (ACE) inhibitors, and antiarrhythmic drugs.
Anticholinergic drugs ！ A high dose of atropine can abolish HF power and markedly attenuate LF power [22]. These findings indicate that parasympathetic modulation of RR intervals is totally responsible for HF power and contributes importantly to LF power. Low doses of atropine or scopolamine enhance vagal modulation of RR intervals, ie, they increase HF power, via a central action. This has been demonstrated in both normal subjects [23,24] and those with congestive heart failure [25].
Beta blockers ！ The effect of the beta blocker atenolol on time and frequency domain measures of RR variability calculated from 24-hour continuous ECG recordings was evaluated in 18 normal volunteers given atenolol, diltiazem, and placebo in random order [26]. During treatment with atenolol, the 24-hour average normal RR interval increased 24 percent and measures of phasic vagal activity (HF power, r-MSSD, pNN50) were also significantly increased [26]. In another report, beta blockade with propranolol enhanced resting HF power and reduced LF power [27].
The effects may be different with certain other beta blockers that have additional activities. As an example, pindolol, a beta blocker that has intrinsic sympathomimetic activity, decreased total, LF, and HF power in normal subjects, while labetalol, which has combined alpha- and beta-adrenergic blocking activity, produced no significant change in these parameters [28].
Thus, beta blockade causes a substantial increase in vagal modulation of RR variability unless the beta blocker has other actions, such as intrinsic sympathomimetic activity or alpha-adrenergic blocking activity. Since there is little sympathetic nervous activity in resting normal persons, the effects of beta blockade on sympathetic modulation of RR intervals are variable and smaller. One study showed that chronic beta blockade blunted the marked increase in LF power that normally occurs during 90º head-up tilt [29].
Calcium channel blockers ！ Variable effects on RR variability have been reported with the different calcium channel blockers. In one study, diltiazem had no significant effect on 24-hour average RR interval or on any measure of heart period variability in normal subjects [26]. In another report, however, the administration of diltiazem, but not nifedipine, during head-up tilt reduced the increase in LF power to the same degree as metoprolol in nine postinfarction patients [30].
Digoxin ！ The effect of digoxin on RR variability was investigated in a study of 20 normal subjects who were randomized to digoxin or placebo [31]. Ambulatory monitoring for 24 hours was performed after five days of treatment, and 10 subjects also had tilt table studies. The average RR interval did not change significantly during digoxin treatment, but HF power increased by approximately 50 percent and LF power by approximately 30 percent, suggesting that digoxin increased the vagal modulation of RR intervals without changing the mean value for the RR intervals. Digoxin had no significant effect on the autonomic response to head-up tilt. It is likely that the effects of digitalis are different in patients with heart failure and neurohumoral activation.
ACE inhibitors ！ The effect of ACE inhibitors on RR variability appears to vary with the patient population being studied. There may be no significant effect on RR variability or the autonomic response to head-up tilt in normal subjects [31]. In comparison, in a series of 32 patients with class III heart failure, captopril increased the median NN50, a clear indication of increased cardiac parasympathetic modulation [32].
It has been suggested that ACE inhibitors increase parasympathetic activity by reducing angiotensin II levels, thereby removing the cardiac vagal inhibitory activity of angiotensin II [33]. (See "Actions of angiotensin II on the heart"). It is possible that the increase in vagal activity during ACE inhibitor therapy might contribute to the increase in survival produced by these agents in patients with congestive heart failure. (See "ACE inhibitors and receptor antagonists in congestive heart failure: Clinical use").
Antiarrhythmic drugs ！ The effect of three antiarrhythmic drugs on RR variability was evaluated in a study of patients with cardiac disease being treated for nonsustained ventricular arrhythmias [34]. NN50 was used to assess the impact on parasympathetic modulation of RR intervals. Flecainide and propafenone decreased NN50 by 56 and 64 percent, respectively, while amiodarone had no effect. It was suggested that the substantial decrease of NN50 during the administration of class IC antiarrhythmic agents reflected an antivagal effect. It is possible that this contributes to the occasionally deleterious effect of these drugs on mortality in postinfarction patients. (See "Ventricular arrhythmias after acute myocardial infarction: Treatment").
