1958
Is 'Silent' Myocardial Ischemia Really as Severe
as Symptomatic Ischemia?
The Analytical Effect of Patient Selection Biases
Jacob Klein, MD; Susan Y. Chao, MS; Daniel S. Berman, MD, FACC; Alan Rozanski, MD, FACC
Downloaded from http://circ.ahajournals.org/ by guest on June 16, 2017
Bacgond The clinical significance of exercise-induced
chest pain remains controversial, as reflected by sharply discordant clinical results within the medical literature. Thus, we
developed a prospective study to compare the functional
significance of silent versus symptomatic ischemia and to
evaluate whether patient selection biases influence this
analysis.
Metds and Rss We evaluated 117 patients (mean age,
63±9 years) with ischemic ST-segment depression during
treadmill testing. Each patient underwent T1-201 myocardial
perfusion single-photon emission computed tomography
(SPECT) after exercise followed by 24-ambulatory ECG
monitoring. Patients were divided into silent versus symptomatic cohorts and were compared for the degree of hemodynamic, exercise and ambulatory ECG, and thallium abnormalities during stress testing. Analyses were repeated as the
patient population became increasingly restricted. Compared
with the silent patients, patients with chest pain during exercise had a shorter exercise duration (P<.009), lower peak
heart rate (P=.009) and double product (P=.005), lower heart
rate threshold for ST depression (P<.05), more episodes of
ambulatory ST-segment depression (P<.05), a higher fre-
quency of ischemia abnormalities during T1-201 SPECT
(P=.02), and higher summed 11 reversibility scores (P=.002).
As the population became increasingly restricted, the relative
magnitude of differences in silent versus symptomatic cohorts
diminished, whereas the absolute magnitude of ischemic abnormalities progressively increased in both cohorts. For example, within the restricted group having ischemia on both
exercise and ambulatory ECG, 50% of the silent cohort had
severe ischemia on TI SPECT (five or more reversible defects)
and more than one third demonstrated the ominous finding of
transient left ventricular dilation after exercise.
Conclusions The induction of chest pain is associated with
substantially more functional abnormalities when it is analyzed
in a relatively "broad-spectrum" coronary artery disease population; by contrast, chest pain tends to lose its apparent value
as a clinical test parameter when its analysis is restricted to
coronary artery disease populations with a greater a priori
likelihood of manifesting inducible ischemia. These findings
may help resolve some of the previous discordant literature
reports. (Circulation. 1994;89:.1958-1966.)
Key Words * ischemia * tomography * exercise .
chest pain
T he clinical significance of exercise-induced chest
pain remains controversial. A variety of previous studies,1-6 including a large follow-up study
by Cole and Ellestad,1 have demonstrated that the
induction of chest pain conveys an adverse prognosis in
coronary artery disease (CAD) patients. However, a
variety of other studies have indicated that silent and
symptomatic ischemia are associated with a similar
angiographic magnitude of coronary disease,7-10 similar
functional abnormalities during stress testing,9-12 and/or
similar prognostic findings.13-18 These discordant results
cannot be attributed to the evaluation of small patient
populations. Rather, among studies of 250 patients or
more, the number reporting a similar magnitude of
abnormalities among silent and symptomatic patients15'16.18 versus the number reporting more abnormality among symptomatic patientsl13'5'6 were nearly
equal. Thus, the cause(s) for this discordance among
studies remains unresolved.
We previously noted that different forms of selection
bias may affect the perceived significance of test parameters.19-26 For example, "pretest" and "posttest" referral biases are biases that powerfully influence the measurement of test parameters.'9-21 Along these lines, it is
notable that the significance of exercise-induced chest
pain has been studied in widely different populations,
ranging from asymptomatic patients without documented CAD at one extreme1 to the study of patients
having only documented CAD and a severe magnitude
of exercise-induced ischemia at the other extreme.14
Hence, we developed the following prospective study
to assess the functional significance of silent versus
symptomatic ischemia during exercise testing and to
evaluate the effect of common selection biases on this
assessment.
Received December 22,1993; revision accepted January 6,1994.
From the Division of Cardiology, Department of Medicine, and
Department of Nuclear Medicine, Cedars-Sinai Medical Center
and the UCLA School of Medicine, Los Angeles, Calif, and the
Division of Cardiology, Department of Medicine, St Luke's/
Roosevelt Hospital Center and the Department of Medicine,
Columbia University College of Physicians and Surgeons, New
York, NY.
Correspondence to Alan Rozanski, MD, St Luke's/Roosevelt
Hospital Center, Division of Cardiology, 114th St on Amsterdam
Ave, New York, NY 10025.
Methods
Patient Population
The patients for this study were recruited from those
routinely referred to our laboratories for stress-redistribution
thallium single-photon emission computed tomography
(SPECT) over a 14-month period. We prospectively recruited
a random sample of patients from among those who satisfied
the following two entry criteria: (1) they had documented
coronary disease based on angiography or the occurrence of
Klein et al Silent vs Symptomatic Ischemia
prior myocardial infarction or a high (>80%) pretest probability of coronary disease, based on Bayesian analysis of
clinical symptoms, risk factors, and results of exercise electrocardiography27 and (2) they manifested an ischemic ECG
response during exercise testing. Patients with an uninterpretable ECG for ST analysis were excluded, including those with
significant resting ST-segment abnormalities, conduction abnormalities, left ventricular hypertrophy, and those on digoxin.
Each recruited patient was required to undergo additional
24-hour ambulatory ECG monitoring. Recruited patients who
did not reliably record chest pain symptoms in their diary
during ambulatory monitoring (five patients) also were excluded from this study. There were 117 patients who satisfied
all criteria. These included 103 men and 14 women with a
mean age of 63±9 years (range, 39 to 83).
Exercise Testing
Downloaded from http://circ.ahajournals.org/ by guest on June 16, 2017
Maximal exercise treadmill testing was performed using the
standard Bruce protocol. The ECG was monitored continuously in three leads during exercise using leads aVF, V1, and
V5. A 12-lead ECG was recorded at rest, during each minute
of exercise, at peak exercise, and during each minute after
exercise for at least 5 minutes or until all exercise-induced
ECG changes disappeared. The exercise ECG response was
considered ischemic when planar or downsloping ST-segment
depression was .1 mm in magnitude from the baseline or
upsloping depression was .1.5 mm at measurements obtained
0.08 seconds after the J point. Blood pressure was recorded
using a cuff sphygmomanometer at rest, at the third minute of
each exercise stage, at peak exercise, and at 2-minute intervals
after exercise. Exercise was terminated by the physician if
severe angina, high-grade ventricular arrhythmia, or exertional
hypotension occurred during stress testing. Patients were
instructed to withhold P-blocking medication for 48 hours,
calcium channel blocking agents for 24 hours, and long-acting
nitrates on the day of testing (139 patients [92%] complied
with this request).
Ambulatory ECG
All patients underwent ambulatory ECG monitoring, usually immediately after the conclusion of the stress-redistribution thallium study in the laboratory (median, 1 day; range, 0
to 61 days, with 93% tested within the same week). Ambulatory monitoring was performed for either 24 hours (87 patients) or 48 hours (30 patients). An AM recorder was used
(Cardiodata MK4) with automatic calibration of the ECG
signal with a 1-mV square wave signal at the start of the
recording. This recorder has a frequency of 0.05 to 100 Hz/s.
After careful skin preparation, two exploring bipolar leads
were attached to the V5 and modified inferior electrode
positions. Patients were instructed to carry out their usual
daily activities and to fill out a comprehensive structured
diary,28 including the occurrence and time of each episode of
chest pain.
The calibrated 24-hour ambulatory monitor tapes were
visually analyzed by an experienced technician at 60 times real
time using an operator-interactive Cardiodata MK4 computer.
Episodes of transient ST-segment depression on ambulatory
ECG were considered ischemic when horizontal or downsloping ST-segment depression of >1 mm or >1.5 mm upsloping
ST-segment depression persisted for at least 60 seconds.29
Measurements were made at 0.08 seconds after the J point and
compared with the isoelectric baseline at the PR interval.
Separation of one episode from the next required that the
ECG return to baseline for at least 3 minutes after an earlier
episode. For data analysis, baseline strips and representative
strips for each ischemic episode were saved and printed at real
time. Blinded interpretation of all recordings was performed
by two independent physicians, with resolution of disagreements by consensus interpretation of the two observers. The
time and heart rate at onset, duration, and magnitude of
1959
ST-segment depression and the presence or absence of chest
pain were noted for each episode of transient myocardial
ischemia during ambulatory monitoring. In patients who were
monitored for 48 hours, only the first 24 hours were considered
for analysis.
Thallium-201 SPECT
Three to four millicuries of thallium-201 was injected intravenously at near-maximal exercise. Imaging was initiated approximately 6 minutes after the injection of thallium, beginning with a 5-minute planar image in the anterior view
followed by tomographic imaging.30 Thirty-two projections (40
seconds) were obtained over 1800. Redistribution imaging was
repeated at 4 hours and at 24 hours if a stress defect was
persistent at 4 hours. A large-field-of-view scintillation camera
containing 75 photomultiplier tubes and a 0.25-in. (0.64 cm)
sodium iodine crystal, equipped with a low-energy, all-purpose
hole collimator was used. The raw data were initially smoothed
with a 9-point weighted averaged algorithm, and filtered
back-projection was then performed using a low-resolution,
fifth-order Butterworth filter with a cutoff frequency of 0.20
cycles per pixel to reconstruct transverse axial tomograms,
each of 6.2-mm thickness, encompassing the entire heart.
Vertical and horizontal long-axes tomograms parallel to the
long and short axes of the left ventricle were extracted from
the filtered transaxial tomograms.
Blinded experienced physicians interpreted all short-axis
and long-axis tomograms displayed on transparency film. The
intensity of each image was normalized to the maximal pixel
value in that image. Thallium-201 myocardial tomograms were
divided into 20 segments: 6 evenly spaced regions in the apical,
midventricular, and basal cuts of the short-axis views and the
two apical segments of the midvertical long-axis cut.20 Each
segment was scored for intensity of thallium activity using a
four-point system: 0, no; 1, mild; 2, moderate; and 3, severe
decrease in regional thallium-201 uptake.30 Any poststress
segmental score of .2 was considered abnormal. Stress defects with a score of ' 1 at 4 or 24 hours of redistribution
imaging were deemed "reversible." A study was considered
positive for ischemia if .2 of the 20 myocardial segments
showed a reversible thallium defect. For each study, we
assessed the number of reversible segments and the thallium
"reversibility score," defined as the sum of thallium score in
the 20 myocardial segments at stress minus the sum of the
segmental scores at 4 hours (or at 24 hours if late imaging was
performed). Transient ischemic dilation of the left ventricle
after stress was evaluated by comparing left ventricular size
on the poststress and 4-hour anterior view planar thallium
images.31
Assessment of Anginal Presentation
Patients with chest pain syndromes were characterized
according to their frequency, precipitants, mode of relief, and
course before the performance of testing. Patients were considered to have an increase in symptoms if they reported a
recent increase in chest pain frequency or severity.
Coronary Angiography
Selective cine coronary arteriography was performed in
multiple oblique projections using the Judkins technique. The
results were interpreted by experienced angiographers. Significant stenosis was considered present when there was 250%
diameter narrowing of a major coronary vessel.
Statistical Methods
Based on the clinical response to exercise, all patients were
divided into those with ST depression and no chest pain during
exercise (silent exercise group) and those with ST depression
and accompanying chest pain during exercise (symptomatic
exercise group). Similarly, for the group of these patients who
also had ST-segment depression during ambulatory ECG
1960
Circulation Vol 89, No 5 May 1994
TABLE 1. Subdivisions of the Patient Population
Total Population
(n=1 17)
l
I
Stratification by chest pain occurrence
during ambulatory ischemic episodes
(n=only the 34 patients with ischemia
on ambulatory ECG monitoring)
Stratification by chest pain response to exercise
(n=all 117 patients)
No chest pain
(silent exercise group)
(n=88)
Chest pain during ischemia
(symptomatic ambulatory group)
(n=12)
No chest pain during ischemia
(silent ambulatory group)
(n=22)
Chest pain
(symptomatic exercise group)
(n=29)
Downloaded from http://circ.ahajournals.org/ by guest on June 16, 2017
monitoring, two additional subgroups were formed: those with
no chest pain during episodes of ST-segnent depression
(silent ambulatory subgroup) and those with at least one or
more ischemic episodes accompanied by chest pain during
episodes of ST-segment depression (symptomatic ambulatory
subgroup).
Data were expressed as frequency (proportion) for categorical variables and as mean±SD for normally distributed continuous variables; otherwise, they are reported as median and
range. Univariate analysis of categorical variables was performed using X2 analysis or Fisher's exact test. When only two
groups were compared, continuous variables were analyzed
using the Student's t test or Wilcoxon rank-sum test. When
more than two groups were compared, continuous variables
were analyzed using ANOVA or the Kruskal-Wallis test. All
significance testing was done at the .05 level.
Results
Table 1 illustrates the subdivision of our patient
population, based on the presence or absence of chest
pain during exercise testing and ambulatory ECG monitoring. During exercise testing, 88 (75%) of the 117
patients had silent ST-segment depression (silent exercise group), whereas only 29 patients (25%) developed
chest pain (symptomatic exercise group). During 24hour ambulatory monitoring, 34 (29%) of the 117
patients had a total of 76 episodes of ST-segment
depression, with a range of 1 to 7 episodes per patient
and a median duration of 7.9 minutes per episodes
(range, 1 to 51 minutes). Of these 34 patients, 22
patients (65%) had only silent ischemic episodes, and
these 22 form the silent ambulatory subgroup. Of the
other 12 patients (35%) who experienced chest pain
with ambulatory ST-segment depression, forming the
symptomatic ambulatory subgroup, 6 had only symptomatic episodes and 6 had both silent and symptomatic
episodes. Overall, 59 (78%) of the ischemic ECG episodes occurred silently, including 20 of the 37 episodes
(54%) occurring in the symptomatic group.
Clinical descriptors for the silent and symptomatic
patients during exercise testing and ambulatory monitoring are listed in Table 2. The majority of symptomatic patients had atypical or typical angina by clinical
history. Conversely, nearly half of the patients in the
silent groups were asymptomatic, and nearly one quarter had nonanginal chest pain. Diabetes mellitus was
not more common in the silent groups.
Comparison of Clinical, Exercise ECG, and
Ambulatory ECG Responses in the Two Exercise
ECG Subgroups
The results in silent and symptomatic exercise subgroups are listed in Table 3. Compared with the silent
subgroup of 88 patients, the symptomatic exercise subgroup of 29 patients had a significantly shorter duration
of exercise (P=.001), lower peak heart rate (P=.009)
TABLE 2. Clinical Descriptors of Silent Versus Symptomatic Patients Within the Ambulatory and
Exercise Groups
(+) Exercise Groups
(+) Ambulatory Groups
Silent
Symptomatic
Silent
Symptomatic
(n=88)
(n=29)
(n=22)
(n=12)
61+10
66+9
64±9
Age (mean+SD)
22
23
80
men
(100)
No.
(79)
(91)
Sex,
(%)
Symptom class, No. (%)
10 (45)
42 (48)
0 (O)*
Asymptomatic patients
5 (23)
6 (21)
18 (20)
NACP patients
3 (14)
6 (21)
15 (17)
Atypical angina patients
4 (18)
17 (59)
13 (15)
Typical angina patients
0 (0)
4 (15)
6 (8)
On medications, No. (%)
3 (15)
7 (27)*
5 (6)
History of diabetes mellitus, No. (%)
6 (27)
10 (34)
39 (44)
History of prior Ml No. (%)
(+) indicates positive; NACP, nonanginal chest pain; and Ml, myocardial infarction.
*P<.01, silent vs symptomatic subgroup.
63±9
8 (67)
0 (0)*
1 (8)
3 (25)
8 (67)
2 (18)
3 (27)
4 (33)
Klein et al Silent vs Symptomatic Ischemia
1961
TABLE 3. Comparison of Exercise ECG and Clinical Variables In the Positive Exercise
ECG Subgroups
Positive Exercise ECG Group
Silent Subgroup
(n=88)
Symptomatic Subgroup
Downloaded from http://circ.ahajournals.org/ by guest on June 16, 2017
P
(n=29)
Clinical variables
.001
5.4±1.7
7.7±2.6
Exercise duration, min
.009
135±20
146±18
Peak heart rate, bpm
.0004
85±10
93±10
% MHPR achieved
NS
173±29
179±24
Peak SBP, mm Hg
.005
23 309±5132
26 319±4819
Peak double product
Exercise ECG variables
NS
2.7±1.0
2.5±0.8
ST magnitude, mm
.04
123±18
131 ±18
HR at ST depression onset, bpm
1 (0-10)
2 (1-6)
.05
ST duration into recovery, min*
Ambulatory ECG variables
NS
38%
30%
% With test positive
3 (1-7)
.04
1 (1-4)
Number of ST episodest
.007
22 (3-66)
11 (2-59)
Duration of episodes (min)*
.009
2.3+1.0
1.9±0.5
ST depression magnitude, mm
NS
100±20
106±16
HR threshold at ischemia, bpm
% MPHR indicates % maximal predicted heart rate; SBP, systolic blood pressure; double product, heart rate times
SBP product; HR, heart rate; and bpm, beats per minute.
*Values are expressed as median (ranges are in parentheses).
tNumber of ST episodes equal number of transient ischemic episodes per 24-hour ambulatory ECG monitoring
period.
and double product (P=.005), and a lower heart rate
threshold (P=.037) for exercise-included ST-segment
depression as well as a greater frequency (P<.05) and
duration (P<.07) of ST-segment depression episodes
during ambulatory ECG monitoring. Thus, the symptomatic subgroup had significantly greater clinical, exercise ECG, and ambulatory ECG abnormalities compared with the silent exercise subgroup.
Comparison of Silent and Symptomatic Ambulatory
ECG Subgroups
The 34 patients with ambulatory ST-segment depression were also divided into a symptomatic subgroup of
12 patients and a silent subgroup of 22 patients (Table
4). The symptomatic subgroup manifested a significantly
greater frequency of inducible chest pain during treadmill testing (P<.05) as well as a shorter exercise duration (P<.05), and they tended toward having a lower
peak exercise double product (P=.08) as well as more
prolonged ST-segment depression after cessation of
exercise (P=.09).
Ischemia During Thallium-201
Stress-Redistribution Scintigraphy
As illustrated in Fig 1, the frequency of a positive
thallium scan was substantially greater in the symptomatic patients during exercise testing and during ambulatory ECG monitoring. The global magnitude of inducible ischemia, as measured by the summed thallium
reversibility score, was also greater in both symptomatic
subgroups. The score was 7.2+6.7 versus 12.2+8.6 in
the silent versus symptomatic exercise subgroups
(P=.002) and 9.2±9.1 versus 13.1±7.1 in the silent
versus symptomatic ambulatory subgroups (P=.2).
Other markers of ischemic magnitude are shown in Fig
2. Although there was a greater frequency of transient
ischemic left ventricular dilations after exercise in the
symptomatic versus silent exercise subgroup, this difference did not reach statistical significance. The frequency of transient ischemic dilation was quite similar
in the symptomatic versus silent ambulatory ECG cohorts, but the frequency of transient dilation was quite
high, occurring in over one third of patients in both
cohorts. Severe ischemia, defined as five or more reversible thallium defects, was significantly more common in
the symptomatic ischemic patients both during exercise
testing and during ambulatory ECG monitoring.
The results of exercise thallium scintigraphy were
concordant with the clinical and ECG data: Symptomatic patients manifested a substantially greater frequency and magnitude of myocardial hypoperfusion
during exercise testing.
Comparison of Exercise Results in Selected
Patient Subgroups
To evaluate the potential influence of patient selection bias on these comparisons, we reassessed the
comparative magnitude of ischemic abnormalities in the
symptomatic versus silent subgroups as the patient
population became progressively restricted. Thus, Table
5 lists differences in exercise variables for symptomatic
versus silent patient subgroups following the application
1962
Circulation Vol 89, No 5 May 1994
TABLE 4. Comparison of Exercise ECG and Clinical Variables in the Positive Ambulatory ECG Group
Positive Ambulatory ECG Group
Silent Subgroup
(n=22)
Clinical variables
Exercise duration, min
Exercise chest pain
Peak heart rate, bpm
% MHPR achieved
Peak SBP, mm Hg
Peak double product
Exercise ECG variables
ST depression magnitude, mm
HR at ST depression onset, bpm
ST duration into recovery, min*
7.0+2.4
3 (14%)
139±21
90±12%
175+26
24 496±5792
Symptomatic Subgroup
(n=12)
.04
.03
NS
5.2+2.5
6 (50%)
130+23
83+13%
159+35
20 611 +6094
NS
NS
.08
NS
NS
Downloaded from http://circ.ahajournals.org/ by guest on June 16, 2017
3.1 (1-6)
2.8±1.0
121 +23
4.2 (2-10)
Silent Group
Symptomatic Group
(n=22)
(n=12)
3.3+0.9
120+19
P
.09
Ambulatory groups
1 (1-7)
2.5 (1-6)
NS
Number of ST episodes*
NS
16 (3-54)
10 (2-61)
Duration of episodes, min*
2.1±1.0
NS
2.3+0.9
ST depression magnitude, mm
NS
102±24
104±14
HR threshold at ischemia, bpm
% MPHR indicates % maximal predicted heart rate; SBP, systolic blood pressure; double product, heart rate times
SBP product; HR, heart rate; and bpm, beats per minute.
*Values are expressed as median (ranges are in parentheses).
of increasingly selected inclusion criteria. First, we
restricted our analyses to patients with known coronary
disease, excluding the "diagnostic" patients who had
high Bayesian CAD likelihood but unconfirmed coronary disease. The same results were noted in this group
as in our population at-large: The symptomatic exercise
subgroup had significantly worse clinical, exercise ECG,
E
p=0.06
p=0.02
P<0.05
92%
%
100
63% _
80CL
3: 60
40200
(n=22) (n=1 2)
(n=88) (n=29)
AMBULATORY ECG()
EXERCISE ECG (+)
M SILENT EM SYMPTOMATIC
FIG 1. Bar graph: Comparison of the frequency of a positive
exercise thallium study (vertical axis) in the patients with a
positive exercise ECG response (left) and positive ambulatory
ECG response (right). Each group was subdivided into two
subgroups based on the presence or absence of chest pain with
exercise. Patients who had silent ST-segment depression are
represented with black bars and those with symptomatic STsegment depression during exercise are represented by striped
bars. In both groups, the symptomatic cohort had a higher
frequency of positive thallium studies.
(n=88) (n=29)
(n=22) (n=12)
EXERCISE ECG (+)
AMBULATORY ECG (+)
-
SIIENT
Z SYMPTOMATIC
FIG 2. Top graph: Frequency of transient ischemic dilation (TID)
of the left ventricle after stress (on the vertical axis) in the
exercise and ambulatory EGG subgroups. Note the high frequency of TID in the patients with a positive ambulatory ECGG
study regardless of the presence or absence of chest pain
during ischemic episodes. Bottom graph: Percent of patients
manifesting severe ischemia (five or more reversible thallium
defects) on the vertical axis. Patients with chest pain and
exercise-induced ST-segment depression had nearly twice the
frequency of severe ischemia on thallium scintigraphy compared
with patients with silent ST-segment depression.
Klein et al Silent vs Symptomatic Ischemia
1963
Comparison of Test Variates in Silent Versus Symptomatic Subgroups in Selected Samples of the
Study Population
Ti Variables
Hemodynamics
ST-Segment Depression
TABLE 5.
Exercise
Known CAD patients
Silent (n=64)
Symptomatic (n=25)
P
Recent cath (within 6 mo)
Silent (n=25)
Symptomatic (n=1 0)
Peak HR
Peak SBP
Duration
mm
144±19
135±20
.06
178±10
171±23
NS
7.5±2.5
3.5±1.9
.0005
2.5±10
2.5±0.7
147±18
133±17
.04
174±23
161±22
7.0±2.6
5.2±2.0
.06
2.7±0.8
3.1±0.7
NS
HR at
Onset, bpm
Duration
Median (Range)
130±18
123±17
.07
1 (0,10)
2 (1,6)
129±17
122±16
NS
1 (1-6)
2 (1,5)
NS
% With
%+
TID
72%
92%
.04
19%
25%
76%
90%
NS
28%
40%
NS
Downloaded from http://circ.ahajournals.org/ by guest on June 16, 2017
NS
NS
P
NS
NS
+Ambulatory ECG, recent cath
7.2±2.4 3.3±1.0
121 ±18
3 (1-6)
79%
42%
140±19
170±22
Silent (n=14)
4
128±24
165±26
5.0±2.4
2.5±0.7
120±23
90%
50%
(2-10)
Symptomatic (n=10)
P
.004
NS
.05
NS
NS
NS
NS
NS
Increase in anginal symptoms
4.8±1.5 2.0±0.8
130±20
5 (1-6)
162±10
100%
60%
136±15
Silent (n=5)
121±15
3 (1-5)
88%
50%
131±17
161±25
3.2±2.3 3.2±0.8
Symptoms (n=8)
P
NS
NS
NS
NS
NS
NS
NS
NS
HR indicates heart rate; SBP, systolic blood pressure; +, positive (evidence of ischemia, as defined in the text); TID, transient ischemic
dilation of the left ventricle after exercise.
and exercise thallium abnormalities. Second, we restricted our analyses to only patients who were recently
catheterized (within 6 months of thallium testing).
Differences in exercise ECG abnormalities persisted,
although they were not significant in this sample size,
and differences in thallium abnormalities (number of
reversible defects and summed thallium reversibility
score) remained significantly greater in the symptomatic
exercise subgroup. Third, we restricted our analysis to
the even more selected population of patients with
angiographically documented CAD who had a positive
ambulatory ECG response. In this subgroup, the differences between the silent and symptomatic patient subgroups narrowed. For example, both groups now had a
similar heart rate threshold for the onset of ST-segment
depression with exercise, and the difference in summed
thallium reversibility score between the two groups was
now mild. Fourth, we evaluated the subgroup of patients manifesting a recent increase in anginal symptoms. Thallium abnormalities of ischemic magnitude
remained mildly greater in the symptomatic subgroups,
not reaching statistical significance in this very small
sample size. Fig 3 compares the summed thallium
reversibility score in each of these selected populations.
As the patient population became increasingly selected,
there was a progressive increase in the magnitude of the
summed thallium reversibility score among both the
silent and symptomatic subgroups.
Discussion
In our study, the induction of chest pain during
exercise-induced ischemia signified ischemia of a
greater severity compared with the induction of silent
ischemia. Our results were based on the analyses of
hemodynamic responses and multiple markers of myocardial ischemia. Each measurement was concordant in
confirming the functional significance of chest pain
during the evocation of ischemia. Among hemodynamic
parameters, patients with symptomatic ischemia manifested a significantly shorter duration of exercise, a
lower peak heart rate, and lower double product. Concomitantly, they manifested a lower threshold for the
induction of exercise-induced ST-segment depression
_ silent* patients
symptomatic patients
20
W
Q0 15
U/)
C1 0
(88) (29)
(64) (25)
(25) (10)
>80% CAD
LIKELIHOOD
CONFIRMED
CAD
CATHETERIZED
PATIENTS
(14) (10)
POSITIVE
HOLTER
(5) (8)
INCREASE
IN SYMPTOMS
FIG 3. Bar graph: Summed thallium reversibility score (vertical
axis) is shown for silent and symptomatic patients in five groups
as increasing selection criteria were applied to define the population of interest. There was a substantial difference in the
summed thallium reversibility score between symptomatic and
silent cohorts for patients having a high likelihood of coronary
artery disease (CAD) (P=.002), confirmed CAD (P=.004), or
recent catheterization (P=.06) (first three groups on the left). By
contrast, in the patients defined on the basis of having a positive
Holter or increasing anginal symptoms, no significant differences were noted. Note that the absolute values for the summed
thallium reversibility scores increased progressively as the population became increasingly selected.
1964
Circulation Vol 89, No 5 May 1994
TABLE 6. Summary of Comparisons of Silent Vaes Symptomtc lachemla
Principle
% With
Investigator
Patients,
Selection Criteria
No.
Silent lschemia
(at al)
Primary end points greater in symptomatic patents
46%
+Cath, mild symptoms
131
1) Bonow
66%
Any +Ex ECG
1402
2) Cole
40%
+Ex ECG (>2 mm)
298
3) Dagenais
Primary End Points
ischemialanatomy*
cardiac events
cardiac eventst
ischemia
anatomy/cardiac events#
ischemia*
Downloaded from http://circ.ahajournals.org/ by guest on June 16, 2017
74
Ex RNV and cath
58%
4) Iskandrian
842
29%
+Cath, symptoms
5) Mark
540
47%
+Ex ECG and cath or Mi
6) Stern
in
Primary end points similar silent vs symptomatic patients
+Ex thallium, +cath in 6 mo
49%
cardiac eventst
55
1) Assey
EF fall >10%, or >5% fall and WMA
60%
cardiac events
140
2) Breitenbacher
65%
+Ex ECG
cardiac eventst
484
3) Callaghan
+Ex ECG, +cath (within 1 mo)
76%
cardiac eventst#
473
4) Falcone
57%
+Ex thallium
103
ischemia/anatomy
5) Gasperatti
112
75%
+Ex thallium, +cath in (1 mo)
6) Hecht
ischemialanatomyt
52%
+Ex ECG, +cath (within 3 mo)
95
ischemia
7) Hikita
72
50%
Selected cath patients
ischemia
8) Hirzel
+EX ECG s/p acute Ml
90
cardiac events
N/R
9) Koppes
+Ex ECG, +cath
122
39%
anatomy
10) Undsey
+Ex
63%
ECG
acute
Ml
60
anatomy
s/p
11) Ouyang
48%
+Cath (from the CASS registry)
cardiac events
880
12) Weiner
Ex indicates exercise; WMA, wall motion abnormalities with exercise.
*In subgroups who were more sick within the study population, the primary end points were similar for silent and symptomatic cohorts.
tThe frequency of revascularization was 45% in the symptomatic cohort vs only 21% in the silent cohort in this study.
tSelected hemodynamic or exercise ECG variables were significantly more abnormal in patients with symptomatic ischemic.
and more prolonged ST-segment depression after exercise. The symptomatic group manifested more ischemic
abnormalities during ambulatory ECG, including a
greater number of transient episodes of ST-segment
depressions during daily life activities. On a clinical
basis, patients in the symptomatic group also were more
likely to have a history of typical angina. Within our
study, there was also a higher frequency of diabetes
mellitus in the symptomatic group, but caution must be
exercised in analyzing this finding, since selection bias
could have influenced the distribution of conventional
risk factors in our patient population.
Our study was designed to use stress-redistribution
thallium-201 myocardial perfusion scintigraphy as an
independent marker of myocardial ischemia and as a
means of comparing functional disease severity among
clinically relevant subgroups of our patient population.
The results of thallium scintigraphy were concordant
with those noted above: The induction of chest pain
with exercise was associated with a greater frequency of
exercise-induced thallium abnormalities, a greater
summed thallium reversibility score, and nearly twice
the frequency of severe ischemia (five or more reversible thallium defects). Hence, as indicated by multiple
ischemic measurements and hemodynamic variables,
the induction of chest pain was associated with greater
ischemic and functional abnormalities during exercise
testing, and the differences between the symptomatic
and silent subgroups were substantial.
Our results were based on the combined analysis of
uncatheterized patients with a high (>80%) likelihood
of coronary disease patients as well as patients with
angiographically documented disease. By contrast,
many other studies limited their assessment to patients
with angiographically documented coronary disease
only. Accordingly, we also restricted our own analyses to
patients with angiographically documented coronary
disease only; this yielded the same observations noted
for our population at large.
Comparison With Other Studies
There have been markedly discordant results among
18 published studies that compared the significance of
silent versus symptomatic ischemia during exercise testingl-18 (Table 6). These differences were present regardless of primary end point (frequency of cardiac events,
magnitude of myocardial ischemia, and/or extent of
anatomic coronary disease). Among the seven largest
studies (more than 250 patients per study), four concluded that symptomatic ischemia was associated with a
higher frequency of abnormality among primary end
points,13-5'6 whereas silent and symptomatic ischemia
were comparable in the other three studies.15,16'18 There
was a wide variation in the criteria used for patient
selection among these studies. At one extreme, in their
large follow-up of 1402 patients, Cole and Ellestad1
used the more lenient inclusion of any patient with a
positive exercise ECG response, even if the test was
Klein et al Silent vs Symptomatic Ischemia
Downloaded from http://circ.ahajournals.org/ by guest on June 16, 2017
done for screening purposes and concomitant angiography was not a required inclusion criterion. At the other
extreme, Breitenbucher et al'4 used the more strict
inclusion criterion of a >10% fall in exercise ejection
fraction or a >5% fall in association with exerciseinduced wall motion abnormalities.14 Similarly, patients
varied in terms of clinical descriptors ranging from the
inclusion of patients who had only mild anginal symptoms or no symptoms, used by Bonow et al,2 to the study
of patients exercising after acute myocardial infarction,
as performed by Ouyang et al.8 Other potential descriptors of functional disease severity do not appear in
Table 6. For example, Gasperetti et al9 were probably
evaluating a functionally sick population, since 25% of
the patients in their study manifested exercise hypotension, and the mean duration of ST-segment postexercise
depression was >5 minutes. Generally, however, even
based on the variables listed in Table 6, there was a
tendency for the strictest selection criteria to be concentrated among the studies in which silent and symptomatic ischemia were found to be associated with
comparable degrees of clinical abnormality.
To mimic selection criteria found in the literature, we
also subdivided our patient population into a series of
clinical subgroups and compared hemodynamic and
ischemic variables in each subgroup. Our results revealed two trends. First, as the patient population
becomes more selected, the magnitude of thallium
abnormalities tended to narrow between silent and
symptomatic cohorts. Second, the more selected the
patient population, the greater the magnitude of thallium defect reversibility in both silent and symptomatic
patients (Fig 3). This also was true for other test
abnormalities. For example, there was a progressive
decline in peak heart rate and peak systolic blood
pressure and a progressive increase in ST-segment
depression duration after exercise as well as a progressive increase in the frequency of transient left ventricular dilation after exercise, as silent ischemic patients
were culled from progressively more selected subgroups. Thus, our results confirm the trend that is
suggested from our literature review: As patients become increasingly selected toward a higher a priori
likelihood of developing ischemia, there is a tendency
for chest pain to lose its significance as an additional
differentiating factor in patient assessment.
A further subanalysis of previously reported studies
provides further support for this conclusion. In three of
the studies cited in Table 6, the comparisons of silent
versus symptomatic ischemia also were assessed after
dividing patients according to the functional severity of
CAD.26,15 In each study, different conclusions could be
derived on the basis of this patient stratification. For
example, Stern et al6 found a greater degree of test
abnormalities among symptomatic versus silent ischemia when the analysis was restricted to patients manifesting ischemia that was between 1 and 2 mm in
magnitude, but they noted comparable degrees of abnormality for symptomatic versus silent ischemia among
patients with .2 mm ST-segment depression. Similar
findings were reported by Callaghan et al.15 Bonow et
a12 found more test abnormalities in symptomatic patients in their population at large but similar degrees of
abnormality among silent and symptomatic patients
when their analysis was restricted to only high-risk
1965
patients manifesting both ST-segment depression and
an abnormal ejection fraction during exercise. These
results further emphasize the importance of patient
selection biases in assessing the significance of chest
pain.
Comparison of Silent and Symptomatic Ischemia
During Ambulatory ECG
Symptomatic ischemia also was associated with more
test abnormalities compared with silent ischemia during
ambulatory ECG monitoring in our study. However, the
magnitude of differences was less than those noted for
silent versus symptomatic ischemia during exercise testing. These ambulatory test results parallel the majority
of 10 prior ambulatory ECG studies that compared
silent versus symptomatic ischemia during ambulatory
ECG monitoring,32-4' including two studies that showed
symptomatic episodes to be associated with greater test
abnormality3334 and five others that showed a strong
tendency toward support of this finding.35-37,3940 More
striking in our data was the scintigraphic findings linking
the presence of ambulatory ST-segment depression per
se-whether silent or symptomatic-to a severe magnitude of inducible ischemia. Thus, even among patients
demonstrating silent ischemia only during ambulatory
ECG monitoring, over one third of these patients
manifested transient ischemia dilation of the left ventricle after exercise and one half also demonstrated
reversible thallium defects in five or more myocardial
segments. These results help explain reports regarding
the adverse prognosis associated with ST-segment depression during ambulatory ECG monitoring.42-44
Study Limitations
Evaluation of the prognostic significance of chest pain
would have been of further interest in our study. Along
these lines, as emphasized elsewhere,2' increasing caution must be exercised in interpreting the results of
follow-up when "interventional referral bias," the preferential selection of sicker patients for revascularization, becomes prevalent. For example, Falcone et al'6
concluded that there was a similar frequency of cardiac
events in patients with silent and symptomatic ischemia
in their study, but nearly twice as many symptomatic
patients (45% versus 24%) were referred for revascularization. This may have resulted in a preferential
lowering of cardiac event rates in the symptomatic
group of patients. Second, Cole et al reported a doubling of cardiac event rates when chest pain was induced
at low versus high exercise workloads, but our study and
most other prior studies have provided no measurable
quantification of chest pain to confirm this possibility.
We did, however, identify a small cohort with increasing
anginal symptoms who appeared to form a functionally
sick cohort in our study (Fig 3). Similar findings have
been reported by Harris et al.45 Further analysis of these
and other chest pain features deserve further clinical
investigation.
Acknowledgments
This study was funded in part by the John D. and Catherine
T. MacArthur Foundation and the KROC Foundation. Dr
Klein was a Save-A-Heart Foundation Fellow in Preventive
Cardiology.
1966
Circulation Vol 89, No S May 1994
References
1. Cole JP, Ellestad MH. Significance of chest pain during treadmill
exercise: correlation with coronary events. Am J Cardiol. 1978;41:
227-232.
2. Bonow RO, Bacharach SL, Green MV, LaFreniere RL, Epstein
SE. Prognostic implications of symptomatic versus asymptomatic
(silent) myocardial ischemia induced by exercise in mildly symptomatic and in asymptomatic patients with angiographically documented coronary artery disease. Am J Cardiol. 1987;60:778-783.
3. Dagenais GR, Rouleau JR, Hochart P, Magrina J, Cantin B,
Dumesnil JG. Survival with painless strongly positive exercise electrocardiogram. Am J Cardiol. 1988;62:892-895.
4. Iskandrian AS, Hakki A. Left ventricular function in patients with
coronary heart disease in the presence or absence of angina
pectoris during exercise radionuclide ventriculography. Am J
Cardiol. 1984;53:1239-1243.
5. Mark DB, Hlatky MA, Calift RM, Morris JJ, Sisson SD, McCants
C, Lee KL, Harrell FE, Pryor DB. Painless exercise ST deviation
on the treadmill: long-term prognosis. JAm Coil CardioL 1989;14:
Downloaded from http://circ.ahajournals.org/ by guest on June 16, 2017
885-892.
6. Stern S, Weisz G, Gavish A, Keren A, Tzivoni D. Comparison
between silent and symptomatic ischemia during exercise testing in
patients with coronary artery disease. J Cardiopulmonary Rehabil.
1988;12:507-512.
7. Lindsey HE Jr, Cohn PF. 'Silent' myocardial ischemia during and
after exercise testing in patients with coronary artery disease. Am
Heart J. 1978;95:441-447.
8. Ouyang P, Shapiro EP, Chandra C, Gottlieb SH, Chew PH,
Gottlieb SO. An angiographic and functional comparison of
patients with silent and symptomatic treadmill ischemia early after
myocardial infarction. Am J Cardiol 1987;59:730-734.
9. Gasperetti CM, Burwell LR, Beller GA. Prevalence of and
variables associated with silent myocardial ischemia on exercise
thallium-201 stress testing. JAm CoU Cardiol. 1990;16:115-123.
10. Hecht HS, Shaw RE, Bruce T, Myler RK. Silent ischemia: evaluation by exercise and redistribution tomographic thallium-201
myocardial imaging. JAm CoU Cardiol. 1989;14:895-900.
11. Hikita H, Kurita A, Takase B, Nagayoshi H, Uehata A, Nishioka T,
Mitani H, Mizuno K, Nakamura H. Usefulness of plasma betaendorphin level, pain threshold and autonomic function in
assessing silent myocardial ischemia in patients with and without
diabetes mellitus. Am J CardioL 1993;72:140-143.
12. Hirzel HO, Leutwyler R, Krayenbuehl HP. Silent myocardial
ischemia: hemodynamic changes during dynamic exercise in
patients with proven coronary artery disease despite absence of
angina pectoris. JAm Coil Cardiol. 1985;6:275-284.
13. Assey ME, Walters GL, Hendrix GH, Carabello BA, Usher BW,
Spann JF. Incidence of acute myocardial infarction in patients with
exercise-induced silent myocardial ischemia. Am J Cardiol. 1987;
59:497-500.
14. Breitenbucher R, Pfisterer M, Hoffmann A, Burckhardt D.
Long-term follow-up of patients with silent ischemia during exercise
radionuclide angiography. JAm Coi Cardiol 1990;15:999-1001.
15. Callaghan PR, Froeicher VF, Klein J, Risch M, Dubach P, Friis R.
Exercise-induced silent ischemia: age, diabetes mellitus, previous
myocardial infarction and prognosis. J Am Coil Cardiol. 1989;14:
1175-1180.
16. Falcone C, Servi S, Poma E, Campana C, Scire A, Montemartini C,
Specchia G. Clinical significance of exercise-induced silent myocardial ischemia in patients with coronary artery disease. JAm Coll
Cardiol. 1987;9:295-299.
17. Koppes MG, Kruyer W, Beckman CH, Jones FG. Response to
exercise early after uncomplicated acute myocardial infarction in
patients receiving no medication: long-term follow-up. Am J
CardioL 1980;46:764. Abstract.
18. Weiner DA, Ryan TJ, McCabe CH, Luk S, Chaitman BR, Sheffield
LT, Tristani F, Fisher LD. Significance of silent myocardial
ischemia during exercise testing in patients with coronary artery
disease. Am J Cardiol. 1987;59:725-729.
19. Rozanski A, Diamond G, Berman DS, Forrester JS, Morris D,
Swan HJC. The declining specificity of exercise radionuclide ventriculography. N Engi J Med. 1983;309:518-522.
20. Rozanski A, Diamond GA, Forrester JS, Berman D, Morris D,
Swan HJC. Comparison of alternative referent standards for
cardiac normality: implications for diagnostic testing. Ann Intern
Med. 1984;70:192-201.
21. Rozanski A. Referral bias and the efficacy of radionuclide stress
tests: problems and solutions. J Nuci Med. 1992;33:2074-2079.
22. Rozanski A, Diamond GA, Forrester JS, Berman DS, Morris D,
Jones RH, Okada R, Freeman M, Swan HJC. Should the intent of
testing influence its interpretation? JAm Coll Carliol. 1986;7:17-24.
23. Rozanski A, Berman DS. Silent myocardial ischemia, I: pathophysiology, frequency of occurrence, and approaches toward detection.
Am Heart J. 1987;114:615-626.
24. Berman DS, Rozanski A, Knoebel S. The detection of silent
ischemia: cautions and precautions. Circulation. 1987;75:101-105.
25. Rozanski A, Berman DS. The efficacy of cardiovascular nuclear
medicine studies. Semin Nucl Med. 1987;27:104-120.
26. Bairey CN, Rozanski A, Berman DS. Exercise echocardiography:
ready or not? J Am Coil CardioL 1988;11:1355-1358.
27. Diamond GA, Forrester JS. Analysis of probability as an aid in the
clinical diagnosis of coronary artery disease. N Engl J Med. 1979;
300:1350-1358.
28. Hedges SM, Krantz DS, Contrada RJ, Rosen S, Rozanski A.
Development of a diary for use with ambulatory monitoring of
mood, activities, and physiologic function. J Psychopathol Behav
Assessment. 1990;12:203-217.
29. Klein J, Rodriguez EA, Berman DS, Prigent F, Chao SY, Maryon
T, Rozanski A. The prevalence and functional significance of
transient ST-segment depression during daily life activity with
comparisons of ambulatory ECG to stress-redistribution
thallium-201 SPECT. Am Heart J. 1993;125:1247-1257.
30. Kiat H, Maddahi J, Yang LD, Train KV, Freidman J, Rozanski A,
Berman DS. Late reversibility of tomographic myocardial
thallium-201 defects: an accurate marker of myocardial viability.
JAm Coll CardioL 1988;12:1456-1463.
31. Weiss AT, Berman DS, Lew AS, Jielsen J, Potkin B, Swan HJC,
Waxman A, Maddahi J. Transient ischemic dilation of the left
ventricle on stress thallium-201 scintigraphy: a marker of severe
and extensive coronary artery disease. JAm Coil Candiol 1987;9:
752-759.
32. Allen D, Gettes LS, Phalan C, Avington F. Painless ST-segment
depression in patients with angina pectoris: correlation with daily
activities and cigarette smoking. Chest. 1976;69:467-473.
33. Cecchi AC, Dovellini EV, Marchi F, Pucci P, Santoro GM, Fazzini
PF. Silent myocardial ischemia during ambulatory electrocardiographic monitoring in patients with effort angina. J Am Coil
Cardiol. 1983;103:934-939.
34. Chierchia S, Lazzari M, Freedman B, Brunelli C, Maseri A.
Impairment of myocardial perfusion and function during painless
myocardial ischemia. JAm Coll Cardiol. 1983;1:924-930.
35. Deanfield JE, Maseri A, Selwyn AP, Ribeiro P, Chierchia S, Kikler
S, Morgan M. Myocardial ischaemia during daily life in patients
with stable angina: its relation to symptoms and heart rate changes.
Lancet. 1983;2:753-758.
36. Quyyumi AA, Wright CM, Mockus U, Fox KM. How important is
a history of chest pain in determining the degree of ischaemia in
patients with angina pectoris? Br Heart J. 1985;54:22-26.
37. Currie P, Saltissi S. Transient myocardial ischaemia after acute
myocardial infarction. Br Heart J. 1990;64:299-303.
38. Stern S, Gavish A, Weisz G, Benhorin J, Keren A, Tzivoni D.
Characteristics of silent and symptomatic myocardial ischemia
during daily activities. Am J CardioL 1988;61:1223-1228.
39. Carboni GP, Lahiri A, Cashman PMM, Raftery EB. Ambulatory
heart rate and ST-segment depression during painful and silent
myocardial ischemia in chronic stable angina pectoris. Am J
CardioL 1987;59:1029-1034.
40. Deedwamia PC, Carbajal VE. Silent ischemia during daily life is an
independent predictor of mortality in stable angina. Circulation.
1990;81:748-756.
41. Levy RD, Shapiro LM, Wright C, Mockus U, Fox KM. The
haemodynamic significance of asymptomatic ST segment
depression assessed by ambulatory pulmonary artery pressure
monitoring. Br Heart J. 1986;56:526-530.
42. Rocco MB, Nabel ECG, Campbell S, Goldman L, Barry J, Mead
K, Selwyn AP. Prognostic importance of myocardial ischemia
detected by ambulatory monitoring in patients with stable
coronary artery disease. Circulation. 1988;78:877-884.
43. Tzivoni D, Gavish A, Zin D, Gottlieb S, Moriel M, Keren A, Banai
S, Sken S. Prognostic significance of ischemic episodes in patients
with previous myocardial infarction. Am J CardioL 1988;62:661-664.
44. Gottlieb SO, Gottlieb SH, Achuff SC, Baumgardner R, Mellits
ED, Weisfeldt ML, Gerstenblith G. Silent ischemia on Holter
monitoring predicts mortality in high-risk postinfarction patients.
JAMA. 1988;159:1030-1035.
45. Harris PJ, Lee KL, Harrell FE, Behar VS, Rosati RA. Outcome in
medically treated coronary artery disease. Circulation. 1980;62:
718-726.
Is 'silent' myocardial ischemia really as severe as symptomatic ischemia? The analytical effect
of patient selection biases.
J Klein, S Y Chao, D S Berman and A Rozanski
Downloaded from http://circ.ahajournals.org/ by guest on June 16, 2017
Circulation. 1994;89:1958-1966
doi: 10.1161/01.CIR.89.5.1958
Circulation is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 1994 American Heart Association, Inc. All rights reserved.
Print ISSN: 0009-7322. Online ISSN: 1524-4539
The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://circ.ahajournals.org/content/89/5/1958
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published in
Circulation can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office.
Once the online version of the published article for which permission is being requested is located, click Request
Permissions in the middle column of the Web page under Services. Further information about this process is
available in the Permissions and Rights Question and Answer document.
Reprints: Information about reprints can be found online at:
http://www.lww.com/reprints
Subscriptions: Information about subscribing to Circulation is online at:
http://circ.ahajournals.org//subscriptions/