(Circulation. 1997;96:3321-3327.)
© 1997 American Heart Association, Inc.
Articles |
From the Institute for Prevention of Cardiovascular Disease, Cardiovascular Division, Department of Medicine, Beth Israel Deaconess Medical Center, Boston, Mass (C.E.L., M.AM, M.C.C., R.L.V.); Department of Epidemiology, Harvard School of Public Health, Boston, Mass (M.AM); and Division of Cardiovascular Medicine, Department of Internal Medicine, University of Kentucky Medical Center (Lexington) (J.E.M.).
| Abstract |
|---|
|
|
|---|
Methods and Results We conducted a review of the circadian pattern of the onset of myocardial infarction (n=19), sudden cardiac death (n=12), and AICD discharge (n=7). The nighttime period was chosen a priori as midnight to 5:59 AM. These reports documented 11 633 nocturnal myocardial infarctions (20% of the total myocardial infarctions), 1981 nocturnal sudden cardiac deaths (14.6% of the total sudden cardiac deaths), and 1200 nocturnal AICD discharges (15.0% of the total discharges). The distributions of myocardial infarction, sudden cardiac death, and AICD discharge were each significantly nonuniform (P<.001). The peak incidence of myocardial infarction and AICD discharge occurred between midnight and 0:59 AM, whereas the peak incidence of sudden cardiac death was between 1:00 and 1:59 AM. The trough in incidence occurred between 4:00 and 4:59 AM for sudden cardiac death and between 3:00 and 3:59 AM for myocardial infarction and AICD discharge.
Conclusions Nocturnal myocardial infarction, sudden cardiac death, and AICD discharge exhibit nonuniform distributions. This finding is consistent with the hypothesis that sleep-state dependent fluctuations in autonomic nervous system activity may trigger the onset of major cardiovascular events and provides further impetus for more directly testing this hypothesis at population, individual, and mechanistic levels. A better understanding of nocturnal triggers may make it possible to reduce the incidence of myocardial infarction, ventricular tachyarrhythmias, and sudden cardiac death during the nocturnal period.
Key Words: sleep nervous system, autonomic meta-analysis circadian rhythm
| Introduction |
|---|
|
|
|---|
Sleep is not a physiologically uniform state. Changes in brain state trigger surges in autonomic nervous system activity.3 Rapid eye movement (REM) sleep, which occurs at 90-minute intervals during sleep, is characterized by sharp surges of sympathetic nervous system activity reaching levels observed during wakefulness.4 This physiological process may be analogous to the surges in sympathetic nervous system activity thought to be responsible in part for triggering of myocardial infarction and sudden cardiac death associated with anger, heavy exertion, and sexual intercourse.5-9 We hypothesized that the physiological processes associated with normal sleep could precipitate acute cardiovascular events. If this hypothesis were correct, then the distribution of acute cardiovascular events would be nonuniform throughout the night. We therefore conducted a review of published reports of the hourly incidence of acute myocardial infarction, sudden cardiac death, and discharge of automatic implantable cardioverter-defibrillators (AICDs) to evaluate whether the nocturnal incidence of these cardiac events is inhomogeneous. If this proved to be the case, the finding would provide further impetus for more directly testing this hypothesis at population, individual, and mechanistic levels.
| Methods |
|---|
|
|
|---|
The studies that were included for the nonfatal myocardial infarction
analysis used the standard definition of myocardial infarction;
to be included, at least two of the three following criteria were
present: chest discomfort and/or symptoms suggestive of myocardial
infarction for
20 minutes, ECG changes suggestive of evolving
myocardial infarction according to the Minnesota coding system, and
typical elevation of at least one of three cardiac enzymes to at least
twice the upper limit of normal.
The definitions of sudden cardiac death were heterogeneous among the 12 studies but were clinically strict. Buff and coworkers28 included in-hospital general medical ward cardiopulmonary arrests, defined as physician-documented circulatory collapse that required resuscitation efforts. Aronow and coworkers33 studied a population at a long-term health care facility. The other 10 studies included only out-of-hospital deaths.29-32,34-39 The majority of the out-of-hospital studies and the Aronow study defined sudden cardiac death as unexpected nontraumatic death within 1 hour of symptom onset in adults where the death could not be attributed to a disease other than coronary artery disease.29-34,36 Moser and coworkers39 defined sudden death as unexpected cardiac arrest that occurred outside the hospital within 15 minutes of a change in symptoms or during sleep in a previously ambulatory patient in stable condition. The remaining studies reported subjects who were found in cardiac arrest, developed cardiac arrest in the presence of emergency personnel,35 or were witnessed collapsing or found in ventricular fibrillation, pulseless ventricular tachycardia, or asystole.37,38 Willich and coworkers29 were the only investigators to discriminate between definite and possible sudden cardiac deaths; both were included in our review.
Documentation of the timing of sudden cardiac death onset was based on the arrival time of the rescue squad36; receipt of the emergency call by the dispatch28,35,37,38; retrospective telephone interviews with witnesses, friends, or relatives29,32,39; or death certificate.31 In three studies, the method of timing of the sudden cardiac death was not disclosed.30,33,34
For the circadian pattern of AICD discharges, the time of each discharge was taken from the device at clinic visits. The devices used in each of the seven studies recorded time and date of the event. Some studies diagnosed ventricular tachyarrhythmias via RR intervals,41,42,44,45 whereas others used devices that had the capability to store intracardiac ECGs before each event.43,46 One study40 involved the use of some devices with cycle length determination of arrhythmias and others with intracardiac ECG capability. Each discharge was analyzed for appropriateness of therapy for sustained ventricular tachyarrhythmia based on information available from the device. Excluded from our review were the less rapid tachyarrhythmias referred to by Tofler and collegues.44 Only the rapid ventricular tachyarrhythmias (heart rate, >250 bpm; cycle length, <240 ms) were included. The other six studies did not distinguish between rapid and less rapid ventricular tachyarrhythmias.
We defined the time period of 12:00 to 5:59 AM, a
priori, as the nocturnal period, according to current
practice.29,38 Analyses of the circadian
pattern of onset of acute myocardial infarction, sudden cardiac death,
and AICD discharge were carried out using simple proportions. The data
were then combined, and the hourly intervals were tested by the
2 test for goodness-of-fit. A significant
nonuniform distribution was considered present if uniformity could
be rejected. In addition, a sensitivity analysis was conducted
by omitting the data from 5:00 to 5:59 AM. A two-sided
value of P<.05 was considered statistically
significant.
| Results |
|---|
|
|
|---|
2 goodness-of-fit, P<.001
(Fig 1
|
|
Sudden Cardiac Death
In the 24-hour period, 1981 of a total of 13 591 patients
(14.6%) experienced onset of sudden cardiac death between midnight and
5:59 AM. Results from the 12 studies of sudden cardiac
death are shown in Table 2
. The incidence
of sudden cardiac death during the nighttime hours reported in the 12
studies ranged from 6.4% to 19.6% of all sudden cardiac deaths in the
24-hour period, for an average of 14.6%. Fig 2
shows that the distribution of sudden
cardiac death onset was significantly nonuniform throughout the night
(P<.001). The peak incidence of sudden cardiac death during
the night hours occurred between 12:00 and 1:59 AM, with
the nadir between 4:00 and 4:59 AM. A statistically
significant nonuniform distribution of sudden cardiac death for the
period from midnight to 4:59 AM (P<.001)
remained after the period from 5:00 to 5:59 AM was
excluded.
|
|
AICD Discharge
A total of 7 studies were identified that included data on the
hourly incidence of AICD discharge in 1197 patients. These 1197
patients experienced 8006 defibrillator discharges, 1200 (15.0%) of
which occurred between midnight and 5:59 AM. Results from
the 7 studies of AICD discharge are shown in Table 3
. The incidence of AICD discharge during
the nighttime hours reported in the 7 studies ranged from 10.3% to
19.8% of all AICD discharges in the 24-hour period, for an average of
15.0%. Fig 3
shows that the distribution
of these AICD discharges was nonuniform over the 6-hour nocturnal
period (P<.001). The peak incidence during the nighttime
hours occurred between midnight and 0:59 AM and sloped down
to the nadir between 3:00 and 3:59 AM. After the period
from 5:00 to 5:59 AM was excluded, a statistically
significant nonuniform distribution of AICD discharge persisted
(P<.001).
|
|
| Discussion |
|---|
|
|
|---|
The mechanisms accounting for the nonuniform nighttime distribution of
acute cardiovascular events are unknown. Despite the
common belief that sleep is a time of relative protection from cardiac
events, presumably because of the absence of triggering activities that
occur during wakefulness, in fact, normal sleep is a dynamic process
that involves complex regulation of the autonomic nervous
system.3 Slow-wave sleep is associated with
increased baroreceptor sensitivity, increased parasympathetic tone, and
decreased heart rate. Conversely, REM sleep, which occurs four to six
times per night for a total of
90 minutes, or 20% to 25% of total
sleep time, is characterized by surges in sympathetic activity and
decreased baroreceptor regulation and
control.47-49
A possible mechanism underlying the nonuniform distribution of cardiovascular events during the nighttime hours is that nocturnal surges in sympathetic activity that occur during REM sleep could precipitate myocardial ischemia50 and infarction by stimulating thrombotic processes21 and increasing fibrinolytic aggregability51 or by increasing hemodynamic stress on vessel walls via increased heart rate, blood pressure, and myocardial demand. REM sleep is associated with sympathetic activity, which may cause plaque rupture, increased platelet aggregability, and coronary vasospasm.4 REM-induced surges in sympathetic and parasympathetic nervous system activity can lead to surges or pauses in heart rhythm, increased electrical instability, and ventricular tachyarrhythmias.47-49,52 Under pathological conditions such as myocardial infarction, impairment in the ability to activate the vagus nerve may contribute to a documented overshoot in unbridled sympathetic nervous system activity during both REM and non-REM sleep.53 REM bouts also increase in intensity and duration toward the morning, which may explain the increased incidence of cardiac events between 5:00 and 5:59 AM.
Deep sleep, or slow-wave sleep, may also play a role in triggering cardiovascular events. Decreased cardiac output and blood pressure due to increased vagal tone have been documented during normal deep sleep. In patients with stenotic coronary artery lesions, the decrease in blood pressure associated with deep sleep54 may reduce the volume and velocity of blood flow, leading to ischemia, thrombi, and possibly evoking emboli either before or after arousal from sleep.55 Electrical instability secondary to poor myocardial perfusion may lead to potentially fatal arrhythmias.
Alternatively, it is possible that the nonuniform distribution observed here reflects many unknown and uncontrolled variables in our pooled data. This ecological comparison is problematic due to limitations inherent in the design. Many alternate explanations may be responsible for the nonuniform distribution. Unfortunately, the pooled data did not include age, gender, individual variation in timing of the sleep/wake cycle, and cultural variation in sleep patterns for the individual subjects. The reviewed studies include a diverse international population. It is impossible to ascertain the effect of cultural patterns, weekday versus weekend variation, and individual variations. The studies also likely included shift workers, who may have a different circadian rhythm and certainly have altered sleep/wake cycles. Data on sleep state and nocturnal or morning arousal are not available for the individual patients. Specifically, it is not known whether the individuals awoke from sleep and arose before the onset of the cardiovascular symptoms, thus omitting sleep statedependent surges as a trigger. As Barry and coworkers41 have shown, for patients with extensive coronary artery disease, rising at night often is associated with episodes of myocardial ischemia. Finally, and most importantly, it is possible that some, if not all, of the nonuniformity observed in this study is attributable to events during wakefulness because some patients may have retired after midnight and others may have awakened before 6:00 AM. Despite this, in such a large population, it is likely that the majority of these subjects were in bed, sleeping, before the onset of their cardiac symptoms.
As with all reviews of the literature, there is a potential for publication and data extraction bias. Multiple observers were used to reduce mismeasurement. Initial measurement error may be present in the original studies of myocardial infarction because investigators relied on patient self-report. However, Muller and colleagues conducted a validation study using data from the Multicenter Investigation of Limitation of Infarct Size Study2 and found that the time of myocardial infarction onset as estimated by serial creatine kinase measurements correlated well with self-reported time of myocardial infarction symptom onset. There also is a potential problem of misclassification of the time of reported sudden cardiac death onset because not all sudden cardiac deaths are witnessed. The differing definitions of sudden cardiac death and various differences in data handling may also have an impact on the nighttime distribution of sudden cardiac death. Random misclassification tends to make the distribution of events appear more uniform throughout the night. On the other hand, if more subjects were likely to be found early in the night (ie, midnight to 12:59 AM) or early in the morning hours (ie, 5:00 to 5:59 AM) rather than during other nocturnal hours, nonuniformity of the sudden cardiac death distribution may be false due to differential misclassification. Martens and coworkers38 found in their registry of 4719 subjects that the number of unwitnessed cardiac arrests was not significantly higher during the nighttime hours compared with the daytime hours. Willich and coworkers,29 in their study of sudden cardiac deaths occurring in the Framingham cohort, distributed the hourly probability of death equally over the time interval in which the death was known to have occurred when the hour of onset of sudden cardiac death could not be determined accurately. In addition, Maron and coworkers excluded six subjects for whom it was impossible to determine the nocturnal time of death.32 The nonfatal myocardial infarction distribution may also been affected by differential misclassification. Subjects may not experience symptoms of acute myocardial infarction, even if the pathophysiological processes are present during sleep, until after waking. This fact may artificially inflate the number of myocardial infarctions during the 5:00-to-5:59 AM time period. It is impossible to estimate the effect of differential misclassification in the sudden cardiac death and nonfatal myocardial infarction data sets. The AICD discharge data are not plagued with this problem because each discharge was recorded directly from the device. However, the mechanisms underlying the arrhythmias in the patients with AICDs may not be representative of the pathophysiology involved in other clinical situations involving life-threatening ventricular arrhythmias. As mentioned previously, 6 of the 7 studies of AICD discharge did not differentiate between rapid and less rapid ventricular tachyarrhythmias.
In conclusion, nocturnal myocardial infarction, sudden cardiac death, and AICD discharge exhibit a nonuniform distribution. This finding is consistent with the hypothesis that sleep statedependent fluctuations in autonomic nervous system activity may trigger the onset of major cardiac events and provides further impetus for more directly testing this hypothesis at population, individual, and mechanistic levels. Definitive elucidation of this mechanism will be aided by simultaneous monitoring of sleep state and ECG, which is now practical with home-based technology,56 by determining time of myocardial infarction and sudden cardiac death in relation to time of onset of sleep, time of awakening, and nocturnal arousal from sleep. A better understanding of nocturnal triggers may make it possible to reduce the incidence of myocardial infarction, ventricular arrhythmias, and sudden cardiac death during the nocturnal period.
| Acknowledgments |
|---|
| Footnotes |
|---|
Received May 6, 1997; revision received July 29, 1997; accepted August 2, 1997.
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
I. Gogenur, G. Wildschiotz, and J. Rosenberg Circadian distribution of sleep phases after major abdominal surgery Br. J. Anaesth., January 1, 2008; 100(1): 45 - 49. [Abstract] [Full Text] [PDF] |
||||
![]() |
T. Laitio, J. Jalonen, T. Kuusela, and H. Scheinin The Role of Heart Rate Variability in Risk Stratification for Adverse Postoperative Cardiac Events Anesth. Analg., December 1, 2007; 105(6): 1548 - 1560. [Abstract] [Full Text] [PDF] |
||||
![]() |
J. M. Legramante and A. Galante Sleep and Hypertension: A Challenge for the Autonomic Regulation of the Cardiovascular System Circulation, August 9, 2005; 112(6): 786 - 788. [Full Text] [PDF] |
||||
![]() |
P. C. Strike and A. Steptoe Behavioral and Emotional Triggers of Acute Coronary Syndromes: A Systematic Review and Critique Psychosom Med, March 1, 2005; 67(2): 179 - 186. [Abstract] [Full Text] [PDF] |
||||
![]() |
F. Iellamo, F. Placidi, M. G. Marciani, A. Romigi, M. Tombini, S. Aquilani, M. Massaro, A. Galante, and J. M. Legramante Baroreflex Buffering of Sympathetic Activation During Sleep: Evidence From Autonomic Assessment of Sleep Macroarchitecture and Microarchitecture Hypertension, April 1, 2004; 43(4): 814 - 819. [Abstract] [Full Text] [PDF] |
||||
![]() |
L J Gula, A D Krahn, A C Skanes, R Yee, and G J Klein Clinical relevance of arrhythmias during sleep: guidance for clinicians Heart, March 1, 2004; 90(3): 347 - 352. [Full Text] [PDF] |
||||
![]() |
P. A. Lanfranchi, A. S.M. Shamsuzzaman, M. J. Ackerman, T. Kara, P. Jurak, R. Wolk, and V. K. Somers Sex-Selective QT Prolongation During Rapid Eye Movement Sleep Circulation, September 17, 2002; 106(12): 1488 - 1492. [Abstract] [Full Text] [PDF] |
||||
![]() |
V. Crasset, S. Mezzetti, M. Antoine, P. Linkowski, J. P. Degaute, and P. van de Borne Effects of Aging and Cardiac Denervation on Heart Rate Variability During Sleep Circulation, January 2, 2001; 103(1): 84 - 88. [Abstract] [Full Text] [PDF] |
||||
![]() |
N. Peled, E. G. Abinader, G. Pillar, D. Sharif, and P. Lavie Nocturnal ischemic events in patients with obstructive sleep apnea syndrome and ischemic heart disease: Effects of continuous positive air pressure treatment J. Am. Coll. Cardiol., November 15, 1999; 34(6): 1744 - 1749. [Abstract] [Full Text] [PDF] |
||||
![]() |
K. Rowe, R. Moreno, T. R. Lau, U. Wallooppillai, B. D. Nearing, B. Kocsis, J. Quattrochi, J. A. Hobson, and R. L. Verrier Heart rate surges during REM sleep are associated with theta rhythm and PGO activity in cats Am J Physiol Regulatory Integrative Comp Physiol, September 1, 1999; 277(3): R843 - R849. [Abstract] [Full Text] [PDF] |
||||
![]() |
R. L. Verrier, T. R. Lau, U. Wallooppillai, J. Quattrochi, B. D. Nearing, R. Moreno, and J. A. Hobson Primary vagally mediated decelerations in heart rate during tonic rapid eye movement sleep in cats Am J Physiol Regulatory Integrative Comp Physiol, April 1, 1998; 274(4): R1136 - R1141. [Abstract] [Full Text] [PDF] |
||||
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Circulation Home | Subscriptions | Archives | Feedback | Authors | Help | AHA Journals Home | Search Copyright © 1997 American Heart Association, Inc. All rights reserved. Unauthorized use prohibited. |