Full text of DOI:10.1002/clc.4950250508

Clin. Cardiol. 25, 237–241 (2002)
Onset of Acute Myocardial Infarction during Sleep
ROBERT W. PETERS, M.D., ROBERT G. ZOBLE, M.D., PH.D.,* MARIA M. BROOKS, PH.D.†
Division of Cardiology, Department of Medicine, Department of Veterans Affairs Medical Center, and the University of Maryland
School of Medicine, Baltimore, Maryland; *Division of Cardiology, Department of Medicine, The James A. Haley Veterans
Administration Hospital and the University of South Florida College of Medicine, Tampa, Florida; †BARI Coordinating Center,
University of Pittsburgh Graduate School of Public Health, Department of Epidemiology, Pittsburgh, Pennsylvania, USA
Summary
Background: Analysis of the chronology of acute cardiovas-
Key words: acute myocardial infarction, Cardiac Arrhythmia
Suppression Trial (CAST), circadian variation, trigger factors,
sleep
cular events may provide important pathophysiologic informa-
tion. There is a circadian pattern in the onset of acute myocar-
dial infarction (AMI) with a mid-morning peak, ascribed to the
catecholamine surge that accompanies awakening and assum-
ing the upright posture. However, in up to 27% of patients the
onset of AMI occurs during sleep (without apparent precipitat-
ing factors). The reasons for this finding are unknown.
Hypothesis: The aim of the study was to determine why the
onset of symptoms of AMI occurs during sleep in some indi-
viduals rather than being precipitated by known trigger factors
such as physical exertion.
Methods: Using the database from a large multicenter clin-
ical trial, patients were grouped according to whether or not
they were awakened from sleep by the symptoms of AMI.
Results: In all, 870 of 3,309 patients (26%) were awakened
by AMI. In general, these patients were older and sicker, with
poorer left ventricular function, lower quality of life indices,
more frequent heart failure, lower ejection fractions, higher in-
cidence of angina, and a greater frequency of atrial arrhyth-
mias. On multivariate analysis, only low ejection fraction and
older age were independently associated with awakening by
the symptoms of AMI.
Introduction
Analysis of the timing of acute cardiovascular events has
the potential for providing important clues to their pathophys-
iology. Numerous studies have demonstrated a circadian pat-
tern in the onset of acute myocardial infarction (AMI) with a
peak in mid to late morning and/or late afternoon and early
evening.1–4 The morning peak has been ascribed to the effects
of the surge in catecholamines that accompanies awakening
and assuming the upright posture.5, 6 The pathogenesis of the
afternoon–evening peak is less clear. These peaks have been
shown actually to increase in prominence when the onset is ad-
justed for the time of awakening.7, 8 Other known trigger fac-
tors of myocardial infarction include heavy physical exertion,
sexual activity, and emotional stress.9–13 In contrast, 10–27%
of patients have the onset of symptoms of AMI during sleep
(with no apparent precipitating factors),7–9, 14, 15 but the char-
acteristics of these patients have not been well delineated. We
hypothesized that this patient group might somehow be differ-
ent from individuals in whom the onset of the acute event oc-
curs during activity. Accordingly, we utilized the large data-
base from the Cardiac Arrhythmia Suppression Trial (CAST)
to examine the characteristics of the patient population who
were awakened from sleep by the onset of AMI.
Conclusions: Patients who are older and sicker are more
likely to be awakened from sleep by the onset of symptoms of
AMI. Although the reasons are unknown, we speculate that
these individuals are less active and therefore less vulnerable to
established trigger factors such as vigorous physical exertion.
Methods
The institutional Committee on Human Research approved
the study protocol in all participating institutions. Written in-
formed consent was obtained from all subjects. The CAST
was a prospective, placebo-controlled, randomized, double-
blind study designed to test the hypothesis that suppression of
unsustained ventricular arrhythmias by antiarrhythmic drugs
in survivors of AMI would reduce the incidence of arrhythmic
death. A detailed description of the CAST methodology has
been provided elsewhere.16, 17 The following is a brief review
of the more important aspects of the trial, with more detailed
information in the area relevant to the present analysis.
Address for reprints:
Robert W. Peters, M.D.
Cardiology Section
Department of Veterans Affairs Medical Center
10 North Greene Street
Baltimore, MD 21201, USA
Received: March 27, 2001
Accepted with revision: October 8, 2001

238
Clin. Cardiol. Vol. 25, May 2002
Patient Population
quality of life indices, a decreased level of activity due to poor
health, more frequent heart failure, higher incidence of angi-
na pectoris, and a greater frequency of atrial fibrillation or
flutter. More of them had ejection fractions ≤20%; 24% more
were >75 years of age; and 29% more were in NYHA func-
tional class II and III. There was no gender difference be-
tween groups, and no significant difference in mean serum
cholesterol or in the incidence of hypertension, diabetes, or
cigarette smoking.
Patients were screened for inclusion in the trial between
June 13, 1987, and August 1, 1991. Patients < 80 years of age
were eligible if they demonstrated an average of ≥6 premature
ventricular depolarizations/h on baseline 24-h ambulatory
electrocardiographic (ECG) (Holter) monitoring and had an
ejection fraction ≤55% within 90 days or ≤40% within 2
years of a documented AMI. Specifically excluded were pa-
tients with ≥15 consecutive beats of ventricular tachycardia at
a rate of ≥120 beats/min > 6 days after infarction, symptomat-
ic bradycardia (without a permanent pacemaker), New York
Heart Association (NYHA) functional class IV congestive
heart failure, significant (creatinine ≥2.5 mg%) renal failure,
or any other conditions likely to limit life span. Also excluded
were patients already receiving class I or III antiarrhythmic
drugs or those with contraindications to any of the three study
drugs (encainide, flecainide, moricizine).
After the first 22 months of the study, eligibility criteria
were altered so that the upper limit of ejection fraction was
40% in all patients, the time from the qualifying myocard-
ial infarction was shortened to ≤90 days, and up to 30 s of
ventricular tachycardia was allowed if the individual was
asymptomatic.
Before initiating the study medication, all patients under-
went a detailed baseline medical history and physical exami-
nation. Patients were specifically questioned about the timing
of the onset of symptoms of their qualifying AMI relative to
the time of awakening on that day. Symptoms considered typ-
ical for AMI included severe discomfort anywhere in the ante-
rior chest, back, neck, or shoulder for ≥30 min, frequently as-
sociated with nausea, vomiting, diaphoresis, or dyspnea.
The ECG characteristics of the study groups are shown in
Table II. These were derived from the ECG and Holter monitor
recording obtained at the time of the the CAST qualifying ex-
amination. Those awakened from sleep had a higher resting
heart rate and a higher incidence of left bundle-branch block.
There was no difference between groups in PR interval or QRS
duration, the incidence of left ventricular hypertrophy or abnor-
mal Q waves, or the frequency of ventricular premature beats.
In the multivariate analysis, only low ejection fraction and
older age were independently associated with awakening by
the symptoms of AMI. The odds that a patient with an ejection
fraction < 20% (n = 266) was awakened by symptoms were
1.48 times larger than those in a patient with an ejection frac-
tion > 20% (odds ratio [OR] = 1.48, 95% confidence interval
[CI] 1.14, 1.94, p = 0.0039) (Fig. 1A). “Age squares” entered
the model (OR = 1.009 per 100 squared years, 95% CI 1.002,
1.016, p = 0.0079) such that the odds of awakening with a my-
ocardial infarction were 10.5% higher for a change from 50 to
60 years old (n = 878), 12.5% higher for a change from 60 to
70 years old (n = 1291), and 14.6% higher for a change from
70 to 80 years old (n = 655) (Fig. 1B).
Discussion
Study Design
Although there have been important advances in preventive
cardiology in recent years, AMI remains an all too frequent
and catastrophic event. Despite the identification of major risk
factors and improved diagnostic modalities, there are current-
ly no accurate means of predicting impending AMI in asymp-
tomatic individuals. If those at greatest risk could be identified,
then appropriate prophylactic measures might be undertaken.
Current research has focused on the atherosclerotic plaque and
the factors that cause it to become unstable, precipitating an
acute event. Prior studies in this area have addressed the role
of specific activities as potential trigger factors.9–13 However,
in up to 25% of cases, the onset of AMI occurs during sleep,
in the absence of any apparent precipitating factors.7–9, 14, 15
Since this represents a considerable number of acute events,
information about the pathophysiology would be of consider-
able importance.
Patients were grouped according to whether or not they
were awakened from sleep by the symptoms of AMI. Con-
tinuous variables between groups were compared with a one-
way analysis of variance (ANOVA); categorical variables by
chi-square test. P values < 0.05 were considered statistically
significant. Stepwise regression analysis was used to select the
predictors for a multivariate model. All demographic and clin-
ical variables, but not quality of life variables, that were uni-
variately related (p < 0.10) to awakening by AMI were consid-
ered as candidate variables for multivariate analysis. Standard
stepwise methods with a p < 0.05 entry criterion were utilized
to construct a logistic regression model where the outcome
was “awakening by myocardial infarction.”
Results
The major finding of the present study is that patients awak-
ened from sleep by the onset of symptoms of AMI tend to be
older and have a higher incidence of severe cardiac disease
and symptomatic angina pectoris, and that this relationship is
most pronounced in the subgroup that is oldest (≥76 years)
and the subgroup that has the lowest ejection fraction (≤20%).
Although the mechanisms involved in these findings are un-
known, we speculate that these individuals are likely to be less
Data regarding timing of onset of AMI were available for
3,309 patients. Of these, 870 (26%) indicated that they were
awakened by symptoms of AMI and 2,439 (74%) were not.
The demographic characteristics of the two study groups are
shown in Table I. The patients awakened from sleep were, in
general, older, with poorer left ventricular function, lower

R. W. Peters et al.: Acute myocardial infarction during sleep
239
TABLE I Clinical characteristics of the study population grouped according to whether or not they were awakened by the symptoms of acute my-
ocardial infarction
Characteristic
Awakened by Sx
Not awakened by Sx
p Value
No. of patients (3309) (%)
Age (years)
Male gender (%)
Caucasian (%)
870 (26.3)
62.4 ± 10
706 (81.1)
698(80.2)
0.35 ± 0.10
214.3 ± 49
160 (18.4)
436 (50.1)
302 (34.7)
197 (22.6)
396 (45.5)
296 (34)
182 (20.9)
707 (81.3)
29 (3.4)
1.9 ± 1.2
1.5 ± 1.1
2439 (87.7)
61.4 ± 10
0.0096
NS
NS
0.02
NS
0.03
0.04
NS
1994 (81.8)
1969 (80.7)
0.36 ± 0.10
213.2 ± 49
371 (15.2)
1126 (46.2)
770 (31.6)
520 (21.3)
1033 (42.4)
734 (30.1)
468 (19.2)
1990 (81.6)
48 (2.0)
Ejection fraction
Cholesterol (mg %)
H/o heart failure (%)
H/o of angina (%)
Hypertension (%)
Diabetes (%)
NS
Previous MI (%)
0.10
0.03
NS
Decreased activity due to health (%)
H/o CABG/PTCA (%)
H/o smoking (%)
H/o A fib/A flut (%)
QL symptoms (1-6)
QL mood (1-6)
NS
0.02
0.04
0.3
1.8 ± 1.2
1.4 ± 1.0
Sitting BP
126 ± 19
125 ± 19
NS
Abbreviations: Sx = symptom, H/o = history of, MI = myocardial infarction, CABG = coronary artery bypass graft, PTCA = percutaneous trans-
luminal coronary angioplasty, A fib = atrial fibrillation, A flut = atrial flutter, QL symptoms = quality of life related to symptoms (0 = no symptoms
and 6 = symptoms all of the time), QL mood = quality of life related to mood with (0 = the best mood and 6 = the worst mood), BP = blood pres-
sure, NS = not significant.
TABLE II Electrocardiographic characteristics of the study population grouped according whether or not they were awakened by the onset of
symptoms of acute myocardial infarction
Characteristic
Awakened by Sx
Not awakened by Sx
p Value
No. of patients
QRS width (ms)
QT interval (ms)
Abnormal Q waves (%)
PVC/h on Holter
LBBB (%)
870
2439
—
NS
NS
NS
NS
0.947 ± 0.02
0.38 ± 0.4
611 (73.7)
126 ± 249
31 (3.6)
0.940 ± 0.02
0.38 ± 0.4
1776 (75.7)
125 ± 243
53 (2.2)
0.025
0.03
Sitting heart rate (beats/min)
76.1 ± 13
75.0 ± 13
Abbreviations: PVC = premature ventricular contractions, LBBB = left bundle-branch block. Other abbreviations as in Table I.
Ejection fraction (% )
Age (years)
35
30
25
20
35
30
25
20
> 40
31–40
21–30
≤ 20
≤ 55
56–65
66–75
≥76
(A)
(B)
FIG. 1 (A) The relationship between ejection fraction and the probability of being awakened from sleep by the onset of symptoms of acute my-
ocardial infarction. The patients with ejection fractions ≤20% are 1.48 times more likely to be awakened by the onset of symptoms than patients
with ejection fractions >20%. (B) The relationship between age and the likelihood of being awakened from sleep by the onset of acute myocar-
dial infarction. Similar to the pattern observed with ejection fraction, there is an abrupt increase in the oldest age group.

240
Clin. Cardiol. Vol. 25, May 2002
active and therefore less vulnerable to established trigger fac-
tors such as vigorous physical exertion and sexual activity. In
fact, a higher percentage of these patients admitted to de-
creased physical activity because of poor health.
Third, for the purposes of the present analysis, we have as-
sumed that the onset of symptoms of AMI is synonymous
with the onset of the event. In certain individuals, such as
those with a stuttering onset or those with prolonged episodes
of pre- or postinfarction angina, the actual initiation of myo-
cardial necrosis may be difficult to determine without hourly
serum enzyme determinations. Similarly, symptoms may be
an unreliable indicator in patients with altered pain percep-
tion, such as diabetics and the very elderly. However, these
limitations would apply similarly to both groups so that any
systematic bias is unlikely. Finally, practice patterns change;
for example, it is likely that a higher percentage of patients are
prescribed beta blockers and aspirin following AMI now than
was true 10 years ago. Accordingly, the applicability of our
data to present day practice may be limited.
This hypothesis is supported by a study of Pell and D’Alon-
zo, who examined the clinical and epidemiologic aspects of
AMI in 1,356 employed patients between 25 and 64 years of
age.18 They found that younger men were almost three times
more likely to be engaged in heavy activity or severe exertion at
the time of symptom onset than their older co-workers. Sim-
ilarly, French and Dock, and also Yater et al. describe a much
lower incidence of AMI occurring during sleep (10 and 13.1%,
respectively) in populations that were considerably younger
than ours.14, 15 In contrast, Ridker et al. found that 26.3% of pa-
tients were awakened by the symptoms of AMI (compared
with 26.2% in our study) in the population of the Physicians
Health Study, which was 20–30 years older than in the two
studies described above.19 Stewart et al. recently reported the
results of a study based upon a standardized questionnaire in
2,468 consecutive patients admitted to a coronary care unit
with a first myocardial infarction between 1975 and 1993.9
They found that patients with exercise-related symptom onset
were more likely to be younger and male, while those who had
onset of symptoms in bed were more likely to be older and
have a history of stable or unstable angina. The study did not
distinguish between those individuals who were asleep at the
time of symptom onset and those who were merely in bed (per-
haps, in some cases, because they were feeling ill).
Conclusions
Patients who are awakened from sleep by the symptoms of
acute myocardial infarction appear to be older and sicker than
individuals who develop symptoms during activity. These data
suggest that these former individuals may be less likely to be
exposed to recognized trigger factors such as heavy exertion or
sexual activity. Further studies are warranted to attempt to
identify factors that are likely to precipitate acute myocardial
infarction in individuals who are largely sedentary.
References
Limitations
1. Muller JE, Stone PH, Turi ZG, Rutherford JD, Czeisler CA, Parker
C, Poole WK, Pasamani E, Roberts R, Robertson T, Sobel BE,
Willerson JT, Braunwald E, and the MILIS Study Group: Circadian
variation in the frequency of onset of acute myocardial infarction.
N Engl J Med 1985;313:1315–1322
2. Goldberg RJ, Brady P, Muller JE, Chen Z, deGroot M, Zonnevald
P, Dalen JE: Time of onset of symptoms of acute myocardial infarc-
tion. Am J Cardiol 1990;66:140–144
3. Hjalmarson A, Gilpin EA, Nicod P, Dittrich H, Henning H, Engler
R, Blacky R, Smith SC, Ricou F, Ross J Jr: Differing circadian pat-
terns of symptom onset in subgroups of patients with acute my-
ocardial infarction. Circulation 1989;80:267–275
4. Willich SN, Linderer T, Wegscheider K, Leizorovicz A, Alamerc-
ery I, Schroder R, and the ISAM Study Group: Increased morning
incidence of myocardial infarction in the ISAM study: Absence
with prior beta-adrenergic blockade. Circulation 1989;80:853–858
5. Tofler GH, Brezinski D, Schafer AI, Czeisler CA, Rutherford JD,
Willich SN, Gleason RE, Williams GH, Muller JE: Morning in-
crease in platelet response to ADP and epinephrine: Association
with the time of increased risk of acute myocardial infarction and
sudden cardiac death. N Engl J Med 1987;316:1514–1518
6. Brezinski DA, Tofler GH, Muller JE, Pohjola-Sintonen S, Willich
SN, Schafer AI, Czeisler CA, Williams GH: Morning increase in
platelet aggregability: Association with assumption of the upright
posture. Circulation 1988;78:35–40
7. Peters RW, Zoble RG, Liebson PR, Pawitan Y, Brooks MM, Pro-
schan M: Identification of a secondary peak in myocardial infarc-
tion onset 11 to 12 hours after awakening: The Cardiac Arrhythmia
Suppression Trial (CAST) experience. J Am Coll Cardiol 1993;
22:998–1003
8. Goldberg RJ, Brady P, Muller JE, Chen Z, de Groot M, Zonneveld
P, Dalen JE: Time of onset of symptoms of acute myocardial infarc-
tion. Am J Cardiol 1990;66:140–144
Our data should be interpreted in light of several method-
ologic limitations. Although our hypothesis was defined pro-
spectively, prior to data analysis, the information was obtained
at the time of the baseline examination, which may have oc-
curred as long as 2 years after the qualifying myocardial in-
farction. Thus, the data exclude patients who died from the
acute infarct (or from other causes) and may be applicable
only to a select population of survivors. In addition, this time
differential raises the possibility that our findings could be
partly due to biased recall. However, when we analyzed the
percentage of patients awakened by symptoms as a function
of the time interval between their index myocardial infarction
and their CAST baseline examination, we found no signifi-
cant differences between groups ( i.e., there was no significant
difference among those patients examined within 14 days of
their myocardial infarction, 15–28 days, 29–42 days, 43–90
days, etc.). Second, we have no information about the type of
medication that individuals were receiving at the time of the
acute event. Some of these medications, such as beta blockers
or salicylates, may have some effect upon trigger factors of
AMI and may have affected our findings. However, only 30%
of patients were receiving beta blockers (and a smaller per-
centage were receiving salicylates on a regular basis) at the
time of their baseline examination,16 and it is likely that an
even smaller percentage of individuals were receiving these
medications prior to their qualifying myocardial infarction.

R. W. Peters et al.: Acute myocardial infarction during sleep
241
9. Stewart RAH, Robertson MC, Wilkins GT, Low CJS, Restieaux
NJ: Association between activity at onset of symptoms and out-
come of acute myocardial infarction. J Am Coll Cardiol 1997;29:
250–253
10. Willich SN, Lewis M, Lowell H, Arntz H, Schubert F, Schroeder R:
Physical exertion as a trigger of acute myocardial infarction. N Engl
J Med 1993;329:1684–1690
11. Willich SN, Lewis M, Lowell H, Arntz H-R, Schubert F, Schroeder
R, for the Triggers and Mechanisms of Myocardial Infarction
Study Group: Physical exertion as a trigger of acute myocardial in-
farction. N Engl J Med 1993;329:1684–1690
12. Mittelman MA, Maclure M, Sherwood JB, Mulry RP, Tofler GH,
Jacobs SC, Friedman R, Benson H, Muller JE: Triggering of acute
myocardial infarction by episodes of anger. Circulation 1995;92:
1720–1725
14. French AJ, Dock W: Fatal coronary arteriosclerosis in young sol-
diers. J Am Med Assoc 1944;124:1233–1237
15. Yater WM, Traum AH, Brown WG, Fitzgerald RP, Geisler MA,
Wilcox BB: Coronary artery disease in men 18 to 39 years in age:
Report of 866 cases, 450 with necropsy examinations. Am Heart J
1948;36:334–372
16. The Cardiac Arrhythmia Suppression Trial (CAST) Investigators:
Preliminary report. Effect of encainide and flecainide on mortality
in a randomized trial of arrhythmia suppression after myocardial in-
farction. N Engl J Med 1989;321:406–412
17. The CAST II Investigators: Effect of the antiarrhythmic agent mori-
cizine on survival after myocardial infarction. N Engl J Med 1992;
327:227–233
18. Pell S, D’Alonzo A: Acute myocardial infarction in a large indus-
trial population: Report of a 6-year study of 1,356 cases. J Am Med
Assoc 1963;185:831–838
13. Muller JE, Mittelman MA, Maclure M, Sherwood JB, Tofler GH:
Triggering myocardial infarction by sexual activity: Low absolute
risk and prevention by regular physical exertion. J Am Med Assoc
1996;275:1405–1409
19. Ridker PM, Manson JE, Buring JE, Muller JE, Hennekens CH:
Circadian variation of acute myocardial infarction and the effect of
low-dose aspirin in a randomized trial of physicians. Circulation
1990;82:897–902
Clin. Cardiol. 25, 241 (2002)
:
Osborn Waves of Severe Hypothermia
THOMAS FEHR, M.D., EDWIN H. STRAUMANN, M.D., OSMUND BERTEL, M.D.
Department of Internal Medicine, Stadtspital Triemli, Zurich, Switzerland
A 50-year-old alcohol-addicted man was admitted uncon-
scious with a core body temperature of 26˚ C. Blood tests re-
vealed normal serum values for sodium, chloride, potassium,
and calcium. Alcohol was negative in serum. The patient was
rewarmed by conventional means (warmed air and infusions).
No serious disturbances of cardiac rhythm occurred, respira-
tion remained sufficient, and the patient survived without seri-
ous complications.
I
aVR
V
V
4
1
V
2
II
aVL
aVF
V
5
V
3
III
VI
V
6
(A)
The initial electrocardiogram (ECG) (Fig. 1A, 26˚ C)
showed a normal heart rate of approximately 80 beats/min
with intermittent atrial fibrillation. PQ and QTc intervals were
prolonged (280 and 600 ms, respectively) as was the QRS
complex, which displayed the typical Osborn waves of hy-
pothermia most pronounced in V3–V6, I–III, and aVF.
Similar ECG alterations have been described in normother-
mic patients with hypercalcemia, acute central nervous sys-
tem abnormalities, or acute inferior wall myocardial infarc-
tion, and they have been referred to as Osborn wave, J wave,
camel hump, or hypothermic hump.
V
1
I
II
V
2
V
3
III
V
4
aVR
aVL
aVF
V
5
V
6
(B)
During rewarming the patient’s ECG gradually normalized.
After 4 h (Fig. 1B, 30˚ C) a regular ectopic atrial rhythm was
observed and QTc interval had shortened to 503 ms. The
Osborn waves were markedly reduced. In the final ECG after
24 h (Fig. 1C, 36.7˚ C) Osborn waves had disappeared and
only a slightly prolonged QTc interval (453 ms) remained.
V
1
I
V
2
II
V
3
III
V
4
aVR
aVL
aVF
V
5
V
6
Reference
(C)
Gussak I, Bjerregaard P, Egan TM, Chaitman BR: ECG phenomenon
called the J wave. History, pathophysiology, and clinical significance.
J Electrocardiol 1995;28:49–58
FIG. 1 Twelve-lead electrocardiograms at (A) admission, (B) after
4 h, and (C) after 24 h. See text for explication.