Heart Rate Variability During Specific Sleep Stages
A Comparison of Healthy Subjects With Patients After Myocardial Infarction
Background Heart rate variability (HRV) is typically higher during nighttime. This evidence supports the concept that overall, sleep is a condition during which vagal activity is dominant. Myocardial infarction (MI) results in a loss in the overall nocturnal HRV increase. However, the characteristics of HRV during specific sleep stages in normal subjects and, more importantly, after MI, are unknown. This study describes HRV during sleep stages in normal subjects and in patients with a recent MI.
Methods and Results HRV was measured from 5 minutes of continuous ECG recording in 8 subjects with no clinical evidence of coronary artery disease (age, 47±4 years) and in 8 patients with a recent MI (age, 51±2 years; NS versus control subjects) in the awake state, non–rapid eye movement (REM), and REM sleep. In normal subjects, the low- to high-frequency ratio (LF/HF) derived from power spectral analysis of HRV decreased significantly from the awake state to non-REM sleep (from 4±1.4 to 1.22±0.33, P<.01). During REM sleep, the LF/HF increased to 3±0.74 (P<.01 versus non-REM, NS versus awake). In post-MI patients, the LF/HF showed an opposite trend toward an increase from 2.4±0.7 to 5.11±1.4 (NS, P<.01 versus the control subjects). REM sleep produced a further increase in the LF/HF up to 8.9±1.6 (P<.01 versus awake and versus REM in control subjects).
Conclusions Myocardial infarction causes a loss in the capability of the vagus to physiologically activate during sleep. This results in a condition of relative sympathetic dominance even in a situation such as sleep, normally described as a condition of vagal dominance and, consequently, low risk for lethal events. The evidence that the sleep-related vagal activation is lost after MI may provide new insights to understanding the nocturnal occurrence of sudden death.
Sleep is known to be a dynamic state of consciousness that is characterized by rapid fluctuations in autonomic activity controlling coronary artery tone,1 2 systemic blood pressure,3 4 5 and heart rate.1 2 3 4 5 6 The understanding of autonomic activity during sleep stages is primarily based on data obtained from animal studies.7 In humans, autonomic control of cardiovascular function during sleep has been described by observations of heart rate and blood pressure.5 6 Analysis of 24-hour heart rate variability (HRV) typically shows a nocturnal increase in the standard deviation of mean RR intervals.8 Overall, sleep is considered a condition in which vagal activity is high and sympathetic activity is relatively quiescent.7 This is considered true both for non–rapid eye movement (REM) and REM sleep, although the phasic bursts of rapid eye movements characteristic of REM sleep may reflect sympathetic activation.7
There is growing interest in the analysis of HRV circadian variation to study abnormalities of autonomic activity.9 Several clinical reports indicate that patients with coronary artery disease9 10 lose this circadian pattern primarily because heart rate reductions and increases in HRV during nighttime are blunted or absent. These observations suggest that sleep may be the condition in which markers like HRV best identify autonomic derangements. However, data are scarce concerning HRV during the different sleep stages, specifically in patients with cardiovascular diseases. This knowledge is important not only for a better understanding of the variability of heart rate but primarily because sleep represents a unique condition in which autonomic activity can be studied in the absence of factors such as physical activity and higher cortical functions.
The present study was designed to extend the understanding of the mechanisms of cardiac control during sleep and to characterize variability measurements during identified sleep stages in normal humans and in patients with a recent myocardial infarction (MI). Preliminary data have been presented.11
The study group consisted of 8 normal individuals (1 woman, 7 men) and of 8 patients (2 women and 6 men) with a recent MI (10 to 15 days old). The control subjects and the post-MI patients had similar ages (mean: control subjects, 47±4.4 years versus post-MI patients, 51±1.8 years, NS). Each participant in the study gave informed consent, and the Institutional Review Boards at Baptist Medical Center of Oklahoma City and the Oklahoma Medical Center approved the protocol. The control group was characterized by normal sinus rhythm, absence of diabetes mellitus, and no previous diagnosis or symptomatology of coronary heart disease. None of the post-MI patients were, at the time of the study, under therapy with β-blockers or digitalis or other drugs known to influence autonomic activity or sinus node activity. Furthermore, all patients had no history of diabetes mellitus and were in normal sinus rhythm.
Standard polysomnographic recordings were digitized and stored on an optical disk system (CNS, Inc) throughout the night. Parameters recorded were 13 bipolar electroencephalograms, submental electromyogram, electrooculogram, and ECG (256 Hz). Sleep stages were identified during off-line analysis using standard scoring techniques.
Heart Rate Variability
Frequency domain analyses of HRV were performed using 5 consecutive minutes of digitized ECG recorded during the awake state, during stable non-REM, and during REM sleep stages. Care was taken to identify times in which no arousal occurred and the ECG signals appeared stable. The digitized ECG segments were analyzed using a commercially available software (Corazonix Corp) and a software developed in Montescano, Italy (GDP, Reference 12). Each QRS complex was visually inspected by an investigator, and files requiring rejection of >10% of QRS complexes were not used. After detrending, total power between 0.04 Hz and 0.5 Hz along with the power in the low-frequency band (LF, 0.04 to 0.15 Hz) and in the high-frequency band (HF, 0.15 to 0.5 Hz) were calculated and logarithm-transformed. The percent of total power in each frequency band is presented unless otherwise noted. LF to HF ratios (LF/HF) were computed using the power in each band before log transformation of the data.
ANOVA appropriate for repeated measures was used to determine differences in HRV during specific sleep stages. Differences between the two study groups were evaluated by unpaired Student’s t test. A value of P<.05 was considered significant. Values reported are mean±SEM.
No differences were observed in total power between the two groups in all the conditions studied. The logarithm-transformed total power was, in control subjects and in post-MI patients, respectively, 7.76±0.43 and 7.54±0.37 in wakefulness, 7.16±0.47 and 7.51±0.43 in non-REM sleep, and 7.92±0.42 and 8.15±0.36 in REM sleep.
Heart Rate Variability
In normal subjects, the transition from wakefulness to non-REM sleep was associated with a trend toward a decrease of percentage power in the LF band (from 53±9% to 41±5%) and with a marked increase in the HF band (from 19±4% to 40±6%, P=.006). This condition traditionally describes parasympathetic dominance, as expected in quiet sleep. During REM sleep, LF power did not change (47±9%), while HF power decreased in all subjects to values comparable with wakefulness (17±2% of the total power, P=.004 versus non-REM sleep, NS versus wakefulness).
The most striking difference between normal subjects and post-MI patients was evident during non-REM sleep (Fig 1⇓). The surge in HF power, typically evident in normal subjects, was completely absent in all post-MI patients. In this latter group, power in the HF band actually tended to decrease from 22±4.5% to 16±4.5% (NS). The difference in HF power between the two study groups in non-REM sleep was highly significant (P=.005). In post-MI patients, HF power further decreased to 8±1.6% of total power during REM sleep (P<.01 versus wakefulness and versus REM sleep in control subjects).
Low- to High-Frequency Ratio
Although parasympathetic mechanisms probably contribute to the power comprised in the LF band, LF/HF is a simple and accepted tool that allows a description of the balance between the two limbs of the autonomic nervous system.13 Fig 2⇓ summarizes the LF/HF data during each state of consciousness in the two study groups. During wakefulness, the LF/HF averaged 4±1.41 in the control subjects and 2.39±0.71 in post-MI patients (NS). The ratio decreased significantly (P<.05) to 1.22±0.33 during non-REM sleep in normal subjects, while it paradoxically increased in patients to 5.11±1.34 (NS). Thus, during non-REM sleep, LF/HF, which was similar in the two groups in wakefulness, became highly discriminatory between healthy and post-MI subjects (P<.01). In normal subjects, the LF/HF increased during REM sleep (3.04±0.74, P<.01 versus non-REM) and became similar to the awake state. In the same sleep stage, the LF/HF ratio was 8.9±1.63 in the post-MI patients, almost twofold greater than in non-REM sleep (NS) and threefold greater than during wakefulness (P<.005).
The present study demonstrates that spectral analysis of HRV is unique for the specific sleep stages examined and characterizes non-REM sleep in normal individuals as the state of consciousness with the highest vagal influence on heart rate. Transition to REM sleep causes a decrease in vagal contribution to heart rate control, resulting in a relative prevalence of low-frequency variability of heart rate; the latter can be interpreted as a condition of sympathetic dominance in the autonomic balance during REM sleep. Myocardial infarction resulted in a complete loss of vagal activation during sleep even in our post-MI patients, who all had a low-risk profile while awake. This produced a dramatic dominance of low-frequency HRV both during REM and non-REM sleep, which may have important implications in the understanding of mechanisms involved in sudden death during sleep.
Autonomic Activity During Sleep
Thus far, conflicting data exist about the neural mechanisms responsible for cardiovascular changes occurring during sleep. Experimental studies in cats suggested that REM sleep is a condition in which cardiac parasympathetic activity increases and tonic sympathetic activity decreases.7 In the study by Zemaityte and colleagues,14 heart rate decreased and the respiratory sinus arrhythmia increased during non-REM sleep stages, but the authors found no fluctuation in the sympathetic spectral frequency band throughout all sleep stages. Based on the pharmacological intervention used in their study, it was concluded that increased heart rate and decreased respiratory sinus arrhythmia found in REM sleep were due exclusively to parasympathetic withdrawal. These conclusions were based on changes in heart rate with β-adrenergic blockade using propranolol and with muscarinic receptor blockade using atropine sulfate. Both of these drugs have central15 and peripheral16 actions that may have influenced the findings in their report.
Raetz and colleagues17 described HRV in cats by using nonlinear statistics. The authors found that, when compared with wakefulness, non-REM sleep was associated with a lower overall HRV but with a higher beat-to-beat variability. During REM sleep, the opposite was observed: an increase in overall variability and a decrease in beat-to-beat variability. This latter finding would reflect an increased influence of sympathetic control of heart rate.
More recently, Kirby and Verrier1 demonstrated that sympathetic discharges during REM sleep are responsible for increases in coronary blood flow during this state of consciousness. Interestingly, sympathetic activity decreased coronary blood flow during REM sleep in dogs with coronary artery stenosis.18 Additionally, two studies examining peripheral sympathetic nerve activity in humans found the highest nocturnal activity during REM sleep19 20 despite no difference in heart rates between the sleep stages.
In the present study, heart rates were also not different during non-REM and REM sleep stages. However, HRV analysis indicated a doubling of relative power in the HF band going from quiet wakefulness to non-REM sleep. During REM sleep, HF power returned to values comparable with wakefulness. Overall, the observation in normal subjects confirms that non-REM sleep is a condition of very high vagal activity and indicates that REM sleep is associated with a significant withdrawal of vagal activity.
This novel description of power spectral analysis during selected sleep stages provides a definitive key that may be useful in interpreting data concerning HRV at night. On the other hand, these data suggest caution in the interpretation of data obtained during nighttime without a monitoring of the sleep stages. Evidence exists supporting the reproducibility of 24-hour HRV.8 Nonetheless, the present study strongly suggests a careful consideration of any condition that can alter the ratio between REM and non-REM sleep before embarking on a study aimed at evaluating the effect of any intervention on 24-hour HRV.
After Myocardial Infarction
The data derived from normal subjects constitute the basis for understanding of autonomic mechanisms in individuals with MI. A large body of evidence has linked sympathetic dominance in autonomic control of heart rate to increased risk for life-threatening arrhythmias,21 whereas increases in parasympathetic activity reduce the risk for ventricular tachyarrhythmias.22 23
A striking finding in this study is that in post-MI patients, the typical increase in respiratory sinus arrhythmia was completely absent during the transition from quiet wakefulness to non-REM sleep, revealing a complete loss of the ability to activate the cardiac vagus in a condition free of any emotional or physical activity. Importantly, HRV measurements in this study were made only during times free of arousal. These findings correlate with the observed reduction in baroreflex sensitivity, that is, a decreased capability of activating vagal reflexes after MI,21 and indicate that measurement of HRV and specifically the HF band during non-REM sleep may contain highly predictive information. This is at variance from 24-hour recording where, on average, LF and HF power represent a much smaller percentage of the total spectral power. It is worth recalling that in the present analysis, detrending was applied to remove the influence of very-low-frequency events in the short period of recording used.
During REM sleep, LF/HF was threefold greater than in wakefulness in post-MI patients. This may be viewed as consequent to the lack of vagal antagonism of sympathetic bursts during REM sleep. The type of HRV response to sleep was consistent in all patients, possibly suggesting that this analysis may have limited power in identifying high-risk subjects. However, the present study describes HRV changes during sleep in the early phase of MI. These findings, combined with recently presented evidence that the recovery pattern of HRV after MI clearly discriminates high-risk subjects,24 represent the background for follow-up studies in which nocturnal HRV analysis may be used to identify high-risk subjects.
The present data indicate that sleep contains information that is highly relevant to the identification of autonomic derangements associated with a higher risk for lethal events after MI. Specifically, the expected surge in cardiac vagal activity associated with non-REM sleep is completely lost after MI. The higher risk for ischemic events25 and the unopposed sympathetic activity evident during REM sleep creates a condition in which lethal arrhythmic events are more likely to occur and provide new information to the understanding of sudden death at night.
- Received December 7, 1994.
- Revision received January 30, 1995.
- Accepted February 8, 1995.
- Copyright © 1995 by American Heart Association
Kirby DA, Verrier RL. Differential effects of sleep stage on coronary hemodynamic function. Am J Physiol. 1989;256(Heart Circ Physiol 25):H1378-H1383.
Verrier RL, Dickerson LW. Autonomic nervous system and coronary blood flow changes related to emotional activation and sleep. Circulation. 1991;83(suppl II):II-81-II-89.
Guazzi M, Zanchetti A. Carotid sinus and aortic reflexes in the regulation of circulation during sleep. Science. 1965;148:397-399.
Mancia G, Baccelli G, Adams DB, Zanchetti A. Vasomotor regulation during sleep in the cat. Am J Physiol. 1971;220:1086-1093.
Snyder F, Hobson JA, Morrison DF, Goldfrank F. Changes in respiration, heart rate, and systolic blood pressure in human sleep. J Appl Physiol. 1964;19:417-422.
Parmeggiani PL, Morrison AR. Alterations in autonomic functions during sleep. In: Loewy AD, Spyer KM, eds. Central Regulation of Autonomic Functions. New York, NY: Oxford University Press; 1990.
Huikuri HV, Niemela MJ, Ojala S, Rantala A, Ikaheimo MJ, Airaksinen J. Circadian rhythm of frequency domain measures of heart rate variability in healthy subjects and patients with coronary artery disease: effects of arousal and upright posture. Circulation. 1994;90:121-126.
Adamson PB, Vanoli E, Lin B, Hull SS Jr, Orr WC. Cardiac autonomic expression during sleep stages in normal persons. Circulation. 1993;88(suppl II):I-115. Abstract.
Pinna GD, Maestri R, Di Cesare A. On the use of the Akaike information criterion in AR spectral analysis of cardiovascular variability signals: a case report study. Proc Comp Cardiol. Los Alamitos, Calif: IEEE Computer Society Press; 1993.
Malliani A, Pagani M, Lombardi F. Methods for assessment of sympatho-vagal balance: power spectral analysis. In: Levy MN, Schwartz PJ, eds. Vagal Control of the Heart: Experimental Basis and Clinical Implications. Armonk, NY: Futura Publishing Co; 1994:433-454.
Cruickshank JM, Neil-Dwyer G, Cameron MM, McAnish J. Beta-adrenoceptor-blocking agents and the blood-brain barrier. Clin Sci. 1980;59:453s-455s.
Lombardi F, Casalone C, Malfatto G, Ruscone TG, Casati R, Malliani A. Effects of propranolol on the impulse activity of cardiovascular sympathetic afferent fibers. Hypertension. 1986;8:50-55.
Hornyak M, Cejnar M, Elam M, Matousek M, Wallin BG. Sympathetic muscle nerve activity during sleep in man. Brain. 1991;114:1281-1295.
Schwartz PJ, LaRovere MT, Vanoli E. Autonomic nervous system and sudden cardiac death: experimental basis and clinical observation for post–myocardial infarction risk stratification. Circulation. 1992;85(suppl I):I-77-I-91.
Vanoli E, De Ferrari GM, Stramba-Badiale M, Hull SS Jr, Foreman RD, Schwartz PJ. Vagal stimulation and prevention of sudden death in conscious dogs with a healed myocardial infarction. Circ Res. 1991;68:1471-1481.
Hull SS Jr, Vanoli E, Adamson PB, Verrier RL, Foreman RD, Schwartz PJ. Exercise training confers anticipatory protection from sudden death during acute myocardial ischemia. Circulation. 1994;89:548-552.
Adamson PB, Huang MH, Vanoli E, Foreman RD, Schwartz PJ, Hull SS Jr. Unexpected interaction between β-adrenergic blockade and heart rate variability before and after myocardial infarction: a longitudinal study in dogs at high and low risk for sudden death. Circulation. 1994;90:976-982.
Nowlin JB, Troyer WG Jr, Collins WS, et al. The association of nocturnal angina pectoris with dreaming. Ann Intern Med. 1965;63:1040-1046.