Use of non-invasive cardiac investigations to

predict clinical endpoints after coronary bypass

graft surgery in coronary artery disease patients:
results from the prognosis and evaluation of risk
in the coronary operated patient (PERISCOP)
study
Philippe Sellier
a*
, Gilles Chatellier
b
, Marie C. D'Agrosa-Boiteux
c
,
Herve Douard
d
, Claude Dubois
e
, Pierre C. Goepfert
f
,
Catherine Monpe `re
g
, Alain Saint Pierre
h
, on behalf of the investigators
of the PERISCOP study
a
Service de readaptation cardiaque, Hopital Broussais, 96 rue Didot-75014, Paris, France
b
Departement d'informatique hospitalie`re, HEGP, 20-40 Rue Leblanc, Paris, France
c
Clinique Medicale cardio-pneumologique, 63830 Durtol, France
d
Service de cardiologie, Hopital cardiologique du Haut Leve`que, Avenue de Magellan 33604 Pessac,
France
e
Service de chirurgie cardiovasculaire, Hopital Foch, 40 Rue Worth, 92151 Suresnes, France
f
Departement de Reeducation, Hopital Brabois, Avenue du Morvan, 54500 Vandoeuvre le`s Nancy, France
g
Centre Bois-Gibert, BP 37510 Ballan-Mire, France
h
Cabinet Medical, 1 Rue des Belges, 69006 Lyon, France
Received 20 November 2002; revised 14 January 2003; accepted 14 January 2003
Aims Little is known about which patients who have undergone coronary bypass
surgery are at risk of future clinical cardiovascular events and may benefit from
further medical treatment. We sought to determine if routine non-invasive
cardiac investigations performed early after surgery were able to stratify the risk of
cardiovascular events in this population.
Methods Two thousand and sixty-five consecutive patients were enrolled in a prospec-
tive multicenter study (PERISCOP). Exercise testing, echocardiography, and 24-h
ambulatory ECG monitoring were performed at day 2010 after coronary bypass
surgery. Follow-up was performed 1 year after coronary bypass surgery. Causes of all
hospitalisation and death occurring within 1 year were documented and classified by
an End-point Committee. The principal endpoint was the combination of all-cause
deaths and cardiovascular events requiring hospitalisation (myocardial infarction,
unstable or severe angina, stroke, congestive heart failure).
Results The 1-year frequency of first events was 155 (8%). In multivariate analysis,
exercise duration <420 s (RR1.68; 95% CI: 1.132.49), exercise induced ST segment
KEYWORDS
Prognostic evaluation;
Non-invasive cardiac
investigations;
Exercise test;
Echocardiography;
Coronary artery bypass
surgery
Funding: the PERISCOP study was supported by a grant from NOVARTIS PHARMA.
* Corresponding author. Tel.: +33-1-43-95-94-50; fax: +33-1-43-95-94-58
E-mail address: philippe.sellier@brs.ap-hop-paris.fr (P. Sellier).
European Heart Journal (2003) 24, 916926
0195-668X/03/$ - see front matter 2003 The European Society of Cardiology. Published by Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0195-668X(03)00084-8

b
y

g
u
e
s
t

o
n

A
u
g
u
s
t

1
8
,

2
0
1
3
h
t
t
p
:
/
/
e
u
r
h
e
a
r
t
j
.
o
x
f
o
r
d
j
o
u
r
n
a
l
s
.
o
r
g
/
D
o
w
n
l
o
a
d
e
d

f
r
o
m

depression >1 mm (RR1.90; 95% CI: 1.183.05), and left ventricular (LV) dysfunction
(wall motion index <1.15) (RR1.97; 95% CI: 1.103.51) were independent predictors
of cardiovascular events and deaths. Ambulatory ECG monitoring had no predictive
value.
Conclusion Exercise testing and echocardiography performed early after coronary
bypass surgery are able to identify high-risk patients who may benefit from intensive
secondary prevention.
2003 The European Society of Cardiology. Published by Elsevier Science Ltd. All
rights reserved.
Introduction
Coronary artery bypass graft surgery is commonly
used to treat patients with coronary heart disease,
and has been shown to improve prognosis in
patients with three-vessel coronary artery disease
and those with left ventricular (LV) dysfunction.
13
However, it is well-known that myocardial revascu-
larisation does not completely protect the patient
from future cardiac events. Some patients remain
at high risk of fatal or non-fatal cardiovascular
events during follow-up.
4,5
Since several classes of
drugs have a confirmed efficacy to improve prog-
nosis in subgroups of coronary artery disease
patients,
6
it is therefore necessary to identify high-
risk patients as soon as possible after coronary
artery bypass graft to optimise treatment.
Non-invasive cardiac investigations including
exercise testing,
79
24-h ambulatory ECG record-
ing,
10
LV function analysis by radionuclide angiog-
raphy
11
, and ventricular late potential detection
12
have identified patients at high risk of future car-
diac events after recent myocardial infarction,
thereby allowing physicians to adopt appropriate
management strategies in these patients. However,
there are limited data available to help physicians
to distinguish between patients at high and low
risk of future events following coronary artery
bypass graft surgery. Previous studies conducted in
patients before coronary artery bypass surgery
have shown that long-term prognosis in post-CABG
patients is related to age, number of coronary
arteries involved, and LV function.
4,5
However,
prognostic evaluation before coronary artery
bypass surgery may be inaccurate, especially given
that the intervention itself would be expected to
modify the prognosis,
13
but not necessarily to the
same extent in all patients. The few studies which
have examined prognostic factors in post-CABG
patients were conducted in small study populations
lacking statistical power and designed to evaluate
the prognostic value of only one non-invasive
method at each time. In one study, the rate of
cardiac events was 50% among 36 patients with
persistent myocardial ischaemia detected by
ambulatory ECG monitoring compared to 8% in
those without ischaemia at 9-month follow-up,
14
whereas ambulatory ECG monitoring findings had
no predictive value in 79 patients with up to 10-
year follow-up.
15
In another study in 70 patients,
moderate-intensity treadmill exercise test was also
found to be ineffective to identify known prog-
nostic factors such as current cardiac function and
later graft status.
16
But this study did not directly
deal with prognosis. More recently, the ability of
exercise thallium-201 scintigraphy to stratify risk
after coronary artery bypass surgery was shown in
two studies: in a prospective study of 873 symptom-
free patients,
17
as in a retrospective study of
411 patients within 2 years of coronary artery
bypass surgery,
18
both reversible thallium-
perfusion defects and exercise capacity were
strong and independent predictors of subsequent
death or non-fatal myocardial infarction. However,
the use of exercise thallium-201 scintigraphy to
determine prognosis in a clinical setting is limited
because it is not widely available and it is relatively
expensive. Thus, the question of whether simple,
non-invasive investigations can be used to identify
post-CABG patients who would benefit from further
medical treatment and more intensive manage-
ment, and consequently reduce treatment costs in
this population, remains open. The objective of our
study was to determine prospectively in a large
population of patients the prognostic value of
widely available, relatively low-cost, non-invasive
cardiac investigations which are routinely per-
formed early after coronary artery bypass surgery.
Methods
Patients
The prognosis and evaluation of risk in the coronary
artery operated patient (PERISCOP) study was a
nationwide prospective multicentre cohort study.
All patients who had undergone coronary artery
bypass surgery within 20 days, had been referred to
a cardiac rehabilitation centre, had no concomit-
ant surgical interventions (valve replacement,
Use of non-invasive cardiac investigations 917

b
y

g
u
e
s
t

o
n

A
u
g
u
s
t

1
8
,

2
0
1
3
h
t
t
p
:
/
/
e
u
r
h
e
a
r
t
j
.
o
x
f
o
r
d
j
o
u
r
n
a
l
s
.
o
r
g
/
D
o
w
n
l
o
a
d
e
d

f
r
o
m

aneurysm resection, carotid intervention), did not
have an artificial pacemaker, or contraindication or
inability to perform an exercise test, and were
willing and able to participate in the follow-up
1 year after discharge were included in the study.
Written informed consent was obtained from all
study participants.
Sample size
According to the available literature, cardiovascu-
lar mortality in this population was estimated to be
1% per year.
13
After a 1-year follow-up, 10 deaths
for 1000 included patients were expected. Based on
the results of the 4S study,
19
we could expect four
to five events requiring hospitalisation (myocardial
infarction, acute coronary events, myocardial
revascularisation) for one death. As a whole, 4050
major events for 1000 patients (a 45% event rate
per year) were expected. According to Peduzzi
et al.,
20
the description of a valid multivariate
statistical model requires at least 10 events per
predictive variable. We therefore decided to
include 2000 patients, giving between 100 and 120
major events requiring hospitalisation, and allow-
ing us to describe a 10-variable predictive model.
Non-invasive cardiac evaluation
The PERISCOP study comprised three mandatory
cardiac explorations, one symptom-limited exer-
cise test, one 24-h ambulatory ECG recording, and
one echocardiography. These three cardiac tests
had to be performed between 10 and 30 days after
coronary artery bypass surgery.
To validate LV function evaluation by echo-
cardiography in comparison to reference radio-
nuclide LV angiography, a sub-study was set up. In
centres where appropriate equipment was avail-
able, LV function (ejection fraction) was also evalu-
ated by radionuclide LV angiography under the
same treatment and at a few days interval.
Exercise test
The exercise test was performed on an ergometric
bicycle using the same standardised protocol.
Briefly, the test began with a starting level of 30 W,
with a 10-W increase per minute. This symptom-
limited test was stopped if the patient complained
of angina, dyspnoea, or fatigue. In the absence of
symptoms, the test was stopped according to the
Recommendations of the European Society of
Cardiology.
21
The test was considered abnormal if
at least one of the following characteristics was
present: ST segment depression >1 mm compared
to baseline, horizontal or downsloping, measured
0.06 s after J point, in leads V4, V5, or V6; frequent
or complex premature ventricular contractions or
runs of ventricular tachycardia; or drop in systolic
blood pressure >10 mmHg or absence of blood
pressure increase during two successive stages.
Exercise duration was recorded for each patient.
Echocardiography
LV systolic function was determined using the wall
motion index (WMI), which has been previously
validated
22
as a method to estimate risk after myo-
cardial infarction. The LV wall kinetics were evalu-
ated in nine segments and a score between 1 and
+3 was attributed according to the segmental
ventricular wall motion and thickening.
Continuous 24-h ambulatory ECG monitoring
The following were considered abnormal:
ischaemic episodes with ST segment depression
>1 mm in comparison to baseline, horizontal or
downsloping, measured 0.06 s after J point, for
more than 60 s, separated from the precedent one
by more than 1 min; ventricular arrhythmias with
the number of premature ventricular contractions
and episodes of ventricular tachycardia.
Data collection
At enrolment, participants were interviewed and
underwent a complete clinical examination
performed by the cardiologists participating in the
study. Clinical data and results of the cardiac tests
were recorded in a standardised case report form.
One year after coronary artery bypass surgery, a
self-administered questionnaire was sent to all
patients by the rehabilitation centres. Patients had
to record detailed information about their clinical
condition and the occurrence of any medical event
leading to hospitalisation, and to send back this
questionnaire to the rehabilitation centre. When a
hospitalisation had occurred, the coordinating
centre obtained all necessary information from the
patient's cardiologist, general practitioner, or the
admitting hospital (discharge summary, investiga-
tion reports, etc.). When a patient was lost to
follow-up (no answer to the questionnaire), his
family and the city hall of the town in which he was
born were contacted to find out whether he had
died. An End-point Committee comprising three
trained senior physicians then validated and classi-
fied all events using a set of pre-defined rules.
Finally, data were entered (double entry) in a
computerised database for statistical analyses.
918 P. Sellier et al.

b
y

g
u
e
s
t

o
n

A
u
g
u
s
t

1
8
,

2
0
1
3
h
t
t
p
:
/
/
e
u
r
h
e
a
r
t
j
.
o
x
f
o
r
d
j
o
u
r
n
a
l
s
.
o
r
g
/
D
o
w
n
l
o
a
d
e
d

f
r
o
m

Study endpoints
The study end-points were all-cause deaths and
non-fatal cardiovascular events (new or recurrent
myocardial infarction, unstable or severe angina,
stroke, congestive heart failure) requiring hos-
pitalisation during the 1-year follow-up period. To
avoid multiple statistical tests, the primary out-
come measure was defined as the combination of
all-cause deaths and the defined cardiovascular
events. In patients with multiple events, only the
first event was considered. All statistical analyses
were performed using this composite end-point.
Statistical analysis
Statistical associations between occurrence of the
outcome and prognostic variables were tested by
Pearson's chi-square test for categorical variables
and either t-test (Gaussian variables) or non-
parametric tests (non-Gaussian variables) for con-
tinuous variables. The risk of all-cause deaths and
cardiovascular events was estimated by the
KaplanMeier method, and the significance of
the difference between two survival curves was
assessed by the logrank test. Multivariate analysis
of associations between outcome and prognostic
factors were done by Cox's proportional hazards
model. We included all variables that were signifi-
cant in univariate analysis, plus age and sex since
these are usually strongly associated with the inci-
dence of cardiovascular disease. We checked the
proportional hazards assumption by plotting
KaplanMeier curves. To comply with the hypoth-
esis of proportional hazards, continuous variables
were dichotomised using either standard thresholds
(ST depression >1 mm, LV ejection fraction <0.40)
or the median value of the considered variable.
Statistical analysis was performed using the
Statview statistical software (Version 5.01, SAS
Institute, Cary, NC, USA). A p value <0.05 was
considered significant.
Results
General characteristics of the cohort
Between May, 1998 and February, 1999, 41 centres
participated in the study. During this period, 2375
patients were screened and 2065 consecutive
patients were enrolled. The most frequent reason
for non-inclusion was the impossibility to partici-
pate in the 1 year follow-up. Each centre included
a median number of 25 patients (range 4136).
The study design and general characteristics of
the patients at inclusion have already been
published.
23
These are summarised in Table 1. Indi-
cations for surgery were stable angina in 864
patients (42%), unstable angina in 736 patients
(36%), recent myocardial infarction in 281 patients
(14%), silent myocardial ischaemia in 187 patients
(9%), and complication of coronary angioplasty in 54
patients (3%). Surgery was performed in emergency
in 79 cases (4%). A history of previous myocardial
revascularisation was found in 444 (21%) patients,
either by coronary angioplasty in 381 patients (18%)
or by coronary bypass surgery in 63 patients (3%).
All the 2065 patients underwent exercise test-
ing, 2064 underwent ambulatory ECG monitoring,
and 2024 had interpretable echocardiography. The
mean interval between surgery and evaluation was
18.95.5 days for exercise testing, 21.112.0 days
for echocardiography, and 19.95.9 days for ambu-
latory ECG monitoring. A total of 130 patients in
seven centres underwent the optional evaluation of
LV function (ejection fraction) by radionuclide LV
angiography. Treatment at discharge is summarised
in Table 2.
Incidence of events
The mean duration of follow-up was 41264 days.
The study flow chart is shown in Fig. 1. Among the
2065 patients, 42 (2%) did not respond to the
follow-up questionnaire. Among these, vital status
could not be traced in 22 (1%), and the remaining
20 were still alive. Among the patients followed
according to the protocol, there were 35 deaths
Table 1 Patients' baseline characteristics (n2065)
Variables Value
a
Age (years) 63.19.9
Male (%) 1779 (86)
Current smoker (%) 598 (29)
Hypertension (%) 943 (46)
Diabetes mellitus (%) 406 (20)
Lipid abnormalities (%) 1362 (66)
Body mass index (kg/m
2
) 26.13.6
History of myocardial infarction (%) 753 (37)
Number of diseased coronary vessels:
One (%) 148 (7)
Two (%) 601 (29)
Three (%) 1316 (64)
LVEF before intervention (1315 patients) 0.600.13
Number of grafts (mean) 2.70.86
Revascularisation of all vessels (%) 1499 (73)
Arterial graft (exclusive) (%) 461 (22)
Early recurrent myocardial ischaemia (%) 50 (2.4)
Early congestive heart failure (%) 51 (2.4)
Early ventricular arrhythmias (%) 181 (9)
a
Values are expressed as either meanstandard deviation or
number (%).
Use of non-invasive cardiac investigations 919

b
y

g
u
e
s
t

o
n

A
u
g
u
s
t

1
8
,

2
0
1
3
h
t
t
p
:
/
/
e
u
r
h
e
a
r
t
j
.
o
x
f
o
r
d
j
o
u
r
n
a
l
s
.
o
r
g
/
D
o
w
n
l
o
a
d
e
d

f
r
o
m

(1.7%). Eighteen were of cardiac origin, mainly due
to sudden death (n8) or heart failure (n5).
Among the 17 non-cardiovascular deaths, 13 were
due to cancer. Of the total of 159 events or deaths,
155 (8% of the total included population) were first
events (Table 3).
Predictors of poor outcome
Univariate analysis
Among the clinical variables including sex, risk
factors, history of myocardial infarction, and
number of coronary vessels involved, only gender
(p0.041), smoking habits (proportion of events 10
vs. 6.7%, p0.041), and occurrence of a cardiovas-
cular event before discharge from the rehabilita-
tion centre (proportion of events 12.6 vs. 7.2%,
p0.013) were predictive of the 1-year incidence
of death or non-fatal cardiovascular events. The
results of initial non-invasive investigations in
patients experiencing at least one end-point (group
A) and in patients without end-point (group B) are
compared in Table 4.
Exercise test
A large majority of the patients (1778: 86%) were
receiving one or several antianginal therapy at the
time of exercise testing (betablockers: 66%; cal-
cium channel blockers: 32% nitrates: 16%; amio-
darone: 17%). There was no significant relationship
between the presence of exercise induced ST
depression >1 mm and a history of myocardial
infarction. A positive exercise test was found in 119
out of 1312 patients (9%) without a history of
myocardial infarction and in 56 out of 753 (7.4%)
patients with a history of myocardial infarction
(p0.53).
The rate of cardiovascular events and deaths at
1 year was 12.4% in the group of patients who had a
positive exercise test (exercise induced ST depres-
sion >1 mm) and 7.2% in those who had a negative
test (p0.016). The prognostic value of exercise
testing is summarised in.Table 4 Heart rate, systolic
blood pressure at rest, and systolic blood press-
ure at peak exercise did not differ significantly
between the two groups. The percentage of
patients with exercise induced ST segment depres-
sion >1 mm, abnormal systolic blood pressure
response, and exercise duration differed signifi-
cantly between group A and group B. When combin-
ing these exercise test abnormalities, the risk of
cardiovascular events and deaths in patients with 0,
1, 2, and 3 abnormalities was 4.7, 8.3, and 17.5%,
respectively (p<0.001).
Echocardiography
There was a good correlation (Fig. 2) between
the WMI determined by echocardiography and the
reference radionuclide LV angiography (n130,
r0.73, p<0.0001). LV ejection fraction (LVEF) may
be roughly estimated by the following equation:
LVEF = 6.51 + (29WMI). Using this equation, a WMI
value of 1.15 corresponds to the usual 0.40 LVEF
threshold defining a poor LV function. LV systolic
function was significantly different between
groups A and B (p0.009), but there was only a
slight difference in LVEF between the two groups
(Table 4).
Twenty-four hour ambulatory ECG monitoring
Heart rate (24-h average), number of premature
ventricular contractions and the frequency of runs
of ventricular tachycardia did not differ signifi-
cantly between group A and B (Table 4). Although a
trend towards the prognostic value of ischaemic
episodes was found, the percentage of patients
with ischaemic episodes did not differ significantly
between the two groups.
The presence of ischaemic episodes on ambula-
tory ECG monitoring was significantly related to the
presence of a positive exercise test (ST segment
depression >1 mm at peak exercise). The incidence
of ischaemic episodes was 6.1% in case of negative
exercise test, whereas it was 20.1% in case of
positive exercise test (p<0.0001).
Survival curves
The median value of exercise duration (420 s) was
used as the cut-off point between long and short
duration of exercise. The 1-year survival prognosis
was significantly worse in patients with short
exercise duration (p0.00008) (Fig. 3), those with
exercise induced ST segment depression >1 mm
(p0.022) (Fig. 4), and those who had an abnor-
mal systolic blood pressure response during the
Table 2 Type of treatments prescribed at discharge
(n2065)
Drugs Number (%)
Aspirin 1722 (84)
Anticoagulants (oral) 365 (18)
Beta-blockers 1469 (71)
Calcium antagonists 539 (26)
Nitrates 240 (12)
Diuretics 287 (14)
ACE inhibitors 459 (22)
Digitalis 27 (1)
Amiodarone 209 (10)
Lipid lowering drugs (statins) 1165 (56) (1097)
920 P. Sellier et al.

b
y

g
u
e
s
t

o
n

A
u
g
u
s
t

1
8
,

2
0
1
3
h
t
t
p
:
/
/
e
u
r
h
e
a
r
t
j
.
o
x
f
o
r
d
j
o
u
r
n
a
l
s
.
o
r
g
/
D
o
w
n
l
o
a
d
e
d

f
r
o
m

exercise test (p0.0028) (Fig. 5). Patients with
a WMI <1.15 were also at higher risk (p0.003)
(Fig. 6).
Multivariate analysis
The independent predictive value of the variables
significantly associated with the risk of death or
events was assessed using a Cox proportional
hazard model. Gender and number of coronary
vessels involved were forced into the model
although they were not associated significantly to
the prognosis in univariate analysis. Only smoking
habits, presence of a cardiovascular event before
discharge from the rehabilitation centre, short
exercise duration (<420 s), exercise induced ST
segment depression >1 mm, LV dysfunction (corre-
sponding to an ejection fraction <0.40) were signifi-
cantly predictive of the risk of cardiovascular
events and deaths (Table 5).
In addition, since many of these factors have also
been identified as prognostic factors in patients
with myocardial infarction, we performed an
unplanned analysis to evaluate the influence of
a history of myocardial infarction on the model.
Similar results were obtained in patients with and
without such a history.
Discussion
Our large prospective study of patients who under-
went coronary bypass surgery in France showed
that early exercise testing (exercise duration, ST
segment depression) is a useful method to identify
those patients who were at high subsequent risk of
death or non-fatal cardiovascular events. These
results were obtained despite concerns raised
about the validity of early exercise testing after
coronary bypass surgery due to anaemia, fatigue,
and pericardial effusion which are often present at
this stage.
24
This study also confirms the prognostic value of
alteration in LV function, even if the impact on
prognosis was limited in this study. This may be
explained by the fact that our patient population
had almost normal LV function (mean LVEF: 0.60)
before coronary bypass surgery. This is consistent
with the fact that the relationship between LV
alteration and risk of death is incremental.
11
On the contrary, we found that heart rate and
frequency of arrhythmias during 24-h ambulatory
ECG monitoring had no prognostic value in our study
population. This may be accounted for by the fact
that a majority of our patients (71%) were treated
by betablockers. However, the presence of
ischaemic episodes had an almost significant prog-
nostic value of death or events. This finding seems
to be in agreement with some studies performed in
patients with recent myocardial infarction.
10
Interestingly, variables traditionally related to
an increased risk of future events in coronary artery
disease patients, such as age, gender, and number
Fig. 1 Flow chart of the study.
Table 3 Number and type of first events or deaths during
1-year follow-up
Type of events
a
n (%)
All-cause deaths 35 (2)
Including cardiac deaths 18 (1)
Events with hospitalisation 120 (6)
Severe or unstable angina 71 (4)
Myocardial infarction 9 (1)
Stroke 17 (1)
Heart failure 23 (1)
a
Among the 155 patients with death or events, four had
more than one event. Only the first one was considered for
analysis.
Use of non-invasive cardiac investigations 921

b
y

g
u
e
s
t

o
n

A
u
g
u
s
t

1
8
,

2
0
1
3
h
t
t
p
:
/
/
e
u
r
h
e
a
r
t
j
.
o
x
f
o
r
d
j
o
u
r
n
a
l
s
.
o
r
g
/
D
o
w
n
l
o
a
d
e
d

f
r
o
m

of coronary arteries involved, had no independent
prognostic value in our study. The absence of pre-
dictive value of age may be explained by the fact
that the follow-up was limited to 1 year in this
study and that most of the events occurred after
1 year in other studies.
5
In contrast, the presence of
events during the rehabilitation programme had a
high predictive value, probably because this com-
posite clinical variable reflects the expertise of the
clinician which has already been emphasised in
other studies.
7
Studies of risk stratification in patients who have
undergone coronary bypass surgery are rare. Most
have been based on preoperative data.
4,5,2527
In
these studies, the significant prognostic factors
were age, number of coronary arteries involved,
and LV dysfunction. It is important to note that
the event rate during the follow-up period of the
PERISCOP study was similar to that observed in
other published studies.
5,28
Also, the few studies of
the prognostic value of the presence of ischaemic
episodes during 24-h ambulatory ECG monitoring
after coronary bypass surgery concluded that this
factor was unable to identify patients at high risk of
cardiac events during follow-up.
14,15
On the other
hand, although exercise thallium-201 scintigraphy
may identify patients at high risk of future cardio-
vascular events, this investigation is not widely
available and was shown to increase the cost from
US$553 to US$1285 per patient identified to be at
Table 4 Prognostic value of non-invasive investigations (univariate analysis)
Variables Whole population
(n=2023)
With events
(n=155)
Without events
(n=1868)
p
value
Exercise test
Mean resting heart rate (b/min) 80.315 80.216 80.415 0.91
Mean peak heart rate (b/min) 115.321 112.222 115.621 0.06
Mean resting systolic blood pressure (mmHg) 124.217 123.719 124.317 0.66
Mean peak systolic blood pressure (mmHg) 164.827 160.929 165.227 0.06
Peak ST segment depression >1 mm (%) 8.4 13.5 8.0 0.016
Abnormal systolic blood pressure response (%) 5.5 11 5.1 0.002
Exercise duration (s) 430169 372167 435168 <0.001
Echocardiography
Wall motion index 1.750.31 1.690.37 1.760.31 0.009
24-h AECG
Mean heart rate (b/min) 7511 7511 7512 0.65
PVCs/24 h 44 50 42 0.77
Presence of VT (%) 3.4 3.9 3.3 0.71
Presence of ischaemic episodes (%) 7.2 11.0 6.9 0.06
Values are expressed as meanone standard deviation or number (%), except for PVC where median values are shown.
AECG=ambulatory ECG monitoring; PVC=premature ventricular contraction; VT=ventricular tachycardia.
Fig. 2 Correlation between echocardiographic WMI and the
reference radionuclide LVEF (validation study performed in
130 patients).
Fig. 3 KaplanMeier survival curve of the proportion of
patients remaining free of cardiovascular events according to
exercise duration (cut-off value: 420 s).
922 P. Sellier et al.

b
y

g
u
e
s
t

o
n

A
u
g
u
s
t

1
8
,

2
0
1
3
h
t
t
p
:
/
/
e
u
r
h
e
a
r
t
j
.
o
x
f
o
r
d
j
o
u
r
n
a
l
s
.
o
r
g
/
D
o
w
n
l
o
a
d
e
d

f
r
o
m

high risk.
17
Furthermore, poor exercise capacity
was considered a much stronger predictor of
death and major events than thallium-perfusion
defects.
17
Any comparison with the study per-
formed by Miller et al.
18
is hardly possible: first,
this study was retrospective; second, the evalua-
tion investigation (thallium-201 imaging) was per-
formed later than in PERISCOP study (within 2 years
after coronary bypass surgery). Interestingly, the
author found that exercise variables alone were
predictive of mortality and myocardial infarction.
In patients with recent myocardial infarction,
risk stratification by exercise testing has shown the
prognostic value of similar variables: short duration
of exercise,
9
occurrence of ST segment depression
>1 mm,
7,8
exercise induced premature ventricular
contractions, and abnormal systolic blood pressure
response.
9
The prognostic value of alteration in
LV function has already been assessed in
post-myocardial infarction patients, with an
inverse relationship between LVEF and 1-year mor-
tality.
11
The LVEF threshold determined as a nega-
tive factor in the latter study was close to 0.40, a
figure which was comparable in our study.
The prognostic value of ambulatory ECG moni-
toring after myocardial infarction is more contro-
versial. The presence of ischaemic episodes is
related
10
or is not related
29
to an increased risk of
future cardiovascular events. The results of the
present study show only a trend in favour of the
prognostic value of the presence of ischaemic epi-
sodes after coronary bypass surgery. The presence
of frequent premature ventricular beats has been
shown to be an independent risk factor of total and
sudden death in the large GISSI-2 trial involving
8676 patients.
30
This was not found in the PERISCOP
study, perhaps because only one-third (753) of the
patients only had a history of myocardial infarction.
Our results are in keeping with those of Myers
et al.,
31
who showed that exercise testing, and
particularly peak exercise capacity, was a stronger
predictor of an increased risk of death than clinical
variables or established risk factors in a population
of patients with cardiovascular risk factors or
established coronary artery disease.
Limitations of the study
Firstly, the inclusion of patients only at the
beginning of a cardiac rehabilitation programme
1520 days after the intervention may have induced
a bias because this study did not take into account
Fig. 4 KaplanMeier survival curve of the proportion of
patients remaining free of cardiovascular events according to
exercise induced ST segment depression (<1 mm vs. >1 mm in
leads V4V6).
Fig. 5 KaplanMeier survival curve of the proportion of
patients remaining free of cardiovascular events according to
exercise systolic blood pressure response abnormalities.
Fig. 6 KaplanMeier survival curve of the proportion of
patients remaining free of cardiovascular events according to
WMI (cut-off value 1.15). LVEF, left ventricular ejection
fraction.
Use of non-invasive cardiac investigations 923

b
y

g
u
e
s
t

o
n

A
u
g
u
s
t

1
8
,

2
0
1
3
h
t
t
p
:
/
/
e
u
r
h
e
a
r
t
j
.
o
x
f
o
r
d
j
o
u
r
n
a
l
s
.
o
r
g
/
D
o
w
n
l
o
a
d
e
d

f
r
o
m

the events which occurred during the first 2 or
3 weeks after coronary bypass surgery. However,
the objective of the PERISCOP study was not to
assess the short-term post-operative prognosis
after coronary bypass surgery. And the rate of
events occurring during the early post-operative
period is limited, around 3% in the first month after
coronary bypass surgery.
5
Secondly, as the ability to perform an exercise
test was an inclusion criterion, patients with the
poorest prognosis may have been excluded from our
study population. Patients unable to perform an
exercise test are usually a high-risk group of
patients.
7
However, the percentage of coronary
bypass surgery patients excluded for this reason
was very low in the PERISCOP study (less than 1%)
and may not bias the study.
Thirdly, the size of the population prevented a
centralised analysis of cardiac investigations for
logistic reasons and this may have induced a lack of
precision in the results. However, all the cardiac
investigations were performed in rehabilitation
centres where these techniques are routinely
performed by trained physicians. Moreover, any
heterogeneity among investigators would only have
decreased the value of the hazard ratio, and
thus would not have masked any significant
relationships.
Fourthly, the follow-up period of the study was
voluntarily limited to 1 year. So our study was
unable to detect prognostic factors related to later
events.
Furthermore, by using a self-administered
questionnaire, we may have missed some events.
However, the number of patients who did not
answer to the questionnaire is very small (2%). And
if we may have been missed some events, there is
no reason to think that there is a relation-
ship between missing events and the results of
non-invasive evaluation.
Finally, the extent to which cardiac rehabilita-
tion influenced the prognosis of our coronary artery
disease patients after coronary bypass surgery
could not be assessed because, to our knowledge,
there is no evidence available in the literature to
answer this question. However, all our patients
underwent a similar cardiac rehabilitation pro-
gramme at the same time after coronary bypass
surgery. Furthermore, as coronary bypass surgery
is widely accepted as an indication to cardiac
rehabilitation, the conclusions of this study are
fully applicable to the whole population of
post-CABG patients.
Clinical implications
This study demonstrates that in post-coronary
bypass surgery patients, low-cost, non-invasive
evaluation, especially by exercise testing, can
identify a subgroup of patients at high risk of future
events. The risk of an event at short or medium
term after coronary bypass surgery is still present,
confirming that myocardial revascularisation does
not completely protect the patient. Consequently,
it is important to ensure that appropriate evidence-
based medical therapy is used in all patients, and
should be reinforced in those patients who are at
highest risk of future events. The occurrence of
cardiac events in coronary bypass surgery patients
who are often already receiving medical treatment
suggests that further improvement, especially in
secondary prevention, is needed to achieve optimal
management of these patients.
Acknowledgements
This study was undertaken by the Working Group on
Functional Evaluation and Cardiac Rehabilitation of
the French Society of Cardiology, with grant sup-
port from NOVARTIS PHARMA. We thank all the
investigators who made this study possible. We
Table 5 Cox proportional analysis of the prognostic factors of death or cardiovascular events in 1998 patients
Variable Hazard ratio 95% CI p value
Age (years) 1.01 0.991.03 0.217
Gender (male=1) 1.54 1.002.38 0.051
Current smoker 1.70 1.192.43 0.004
No. of coronary vessels involved
2 vs. 1 2.10 0.894.95 0.088
3 vs. 1 1.82 0.794.17 0.159
Presence of post-operative events (yes=1) 1.70 1.062.72 0.027
Exercise duration (<420 s=1) 1.68 1.132.49 0.011
Exercise induced ST depression >1 mm 1.90 1.183.05 0.008
Abnormal SBP response (yes=1) 1.59 0.922.73 0.094
LV dysfunction (WMI <1.15) 1.97 1.103.51 0.022
SBP=systolic blood pressure; WMI=wall motion index.
924 P. Sellier et al.

b
y

g
u
e
s
t

o
n

A
u
g
u
s
t

1
8
,

2
0
1
3
h
t
t
p
:
/
/
e
u
r
h
e
a
r
t
j
.
o
x
f
o
r
d
j
o
u
r
n
a
l
s
.
o
r
g
/
D
o
w
n
l
o
a
d
e
d

f
r
o
m

also thank Mrs Pamela Johnson for reviewing the
manuscript.
Appendix A
List of investigators: B. Avias (83406 Giens-
Hyeres), J.M. Feige (84000 Avignon), J. Garaix
(83407 Hyeres), M. Ferriere (34295 Montpellier), C.
Rocca (38660 St Hilaire du Touvet), S. Frolichman
(69565 Saint Genis Laval), K. Sourdais (23000 Sainte
Feyre), M.C. D'Agrosa-Boiteux (63830 Durtol), H.
Ducrot (42400 Saint Chamond), M. Bousquet (31650
Saint Orens de Gameville), J. Breton (82500
Beaumont de Lomagne), J.L. Breda (66240 Saint
Esteve), M. Fischbach (33150 Cenon), Pic (33950
Lege Cap Ferret), Borgat (44474 Carquefou), Pavy
(44270 Machecoul), J.P Marty (17138 Puilboreau),
J. Le Henaff (86021 Poitiers), Monnet de Lorbeau
(41600 Lamotte-Beuvron), Monpere (37510 Ballan
Mire), P. Guillo (35044 Rennes), L. Michel (29684
Roscoff), P. Sellier (75014 Paris), Fattell (91000
Evry), D.M. Marcadet (75116 Paris), O. Stechepin-
sky (50190 Saint Martin d'Aubigny), Belin (14360
Trouville sur mer), S. Corone-Alden (91640 Briis
sous Forges), F. Larrazet (77174 Villeneuve Saint
Denis), C. Cabanis (78605 Maisons Laffitte), C.
Puechavy (78740 Evecquemont), P. Aeberhard
(93200 Saint Denis), Dominique (80800 Corbie), P.
Delelis (59160 Lomme), D. Souris (57038 Metz), J.J.
Maureira (54201 Dommartin-les-touls), M. Ross
(57560 Abreschviller), Kessler (57565 Niderviller),
R. Grudet Bald (68460 Lutterbach), Verges (21000
Dijon), P. Poncelet (62590 Oignies).
Executive committee: P. Sellier, G. Chatellier,
E. Dufour (Novartis Pharma), C. Dubois, A. Costa.
Scientific committee: P. Sellier (Chairman),
C. Dubois, G. Chatellier, P.C. Goepfert, M.C.
D'Agrosa-Boiteux, C. Monpe`re, H. Douard, A,
Saint-Pierre.
End-point committee: A. Costa, C. Dubois,
P. Sellier.
References
1. Alderman EL, Bourassa MG, Cohen LS et al. Ten-year
follow-up of survival and myocardial infarction in the
randomized Coronary Artery Surgery Study. Circulation
1990;82:162946.
2. The Veterans Administration coronary artery bypass
surgery cooperative study group. Eleven-year survival in the
Veterans Administration randomized trial of coronary
bypass surgery for stable angina. N Engl J Med 1984;
311:13339.
3. European coronary surgery study group. Prospective
randomized study of coronary artery bypass surgery in
stable angina pectoris: a progress report on survival.
Circulation 1982;65(Suppl II):II-6771.
4. Van Brussel BL, Plokker HWM, Voors AA et al. Multivariate
risk factor analysis of clinical outcome 15 years after venous
coronary artery bypass graft surgery. Eur Heart J 1995;
16:12006.
5. Stahle EZ, Bergstrom R, Holmberg L et al. Survival after
coronary artery bypass grafting. Experience from 4661
patients. Eur Heart J 1994;15:120411.
6. Management of stable angina pectoris. Recommendations of
the Task Force of the European Society of Cardiology. Eur
Heart J 1997; 18:394413.
7. DeBusk RF, Kraemer HC, Nash E. Stepwise risk stratification
soon after acute myocardial infarction. Am J Cardiol 1983;
52:11616.
8. Theroux P, Waters D, Halphen C et al. Prognostic value of
exercise testing soon after myocardial infarction. N Engl J
Med 1979;301:3415.
9. Krone RJ, Gillespie JA, Weld FM et al. The Multicenter
Postinfarction Research Group. Low-level exercise testing
after myocardial infarction: usefulness in enhancing clinical
risk stratification. Circulation 1985;71:809.
10. Tzivoni D, Gavish A, Zin D et al. Prognostic significance of
ischemic episodes in patients with previous myocardial
infarction. Am J Cardiol 1988;62:6614.
11. Gadsboll N, Hoilund-Carlsen PF, Madsen EB et al. Right and
left ventricular ejection fractions: relation to one-year
prognosis in acute myocardial infarction. Eur Heart J 1987;
8:12019.
12. Breithardt G, Schwarzmaier J, Borggrefe M et al. Prognostic
significance of late ventricular potentials after acute
myocardial infarction. Eur Heart J 1983;4:48795.
13. Yusuf S, Zucker D, Peduzzi P et al. Effect of coronary artery
bypass graft surgery on survival: overview of 10-year results
from randomised trials by the Coronary Artery Bypass Graft
Surgery Trialists Collaboration. Lancet 1994;344:56370.
14. Egstrup K. Asymptomatic myocardial ischemia as a predictor
of cardiac events after coronary artery bypass grafting for
stable angina pectoris. Am J Cardiol 1988;61:24852.
15. Patel D, Mulcahy D, Curzen N et al. Prognostic significance
of transient ST segment changes after coronary artery
bypass surgery: a long-term (410 year) follow up study. Br
Heart J 1993;70:33741.
16. Pollock ML, Foster C, James D et al. Diagnostic capabilities
of exercise testing soon after myocardial revascularization
surgery. Cardiology 1982;69:35865.
17. Lauer MS, Lytle B, Pashkow F et al. Prediction of death and
myocardial infarction by screening with exercise-thallium
testing after coronary-artery-bypass grafting. Lancet 1998;
351:61522.
18. Miller TD, Christian TF, Hodge DO et al. Prognostic value of
exercise thallium-201 imaging performed within 2 years of
coronary artery bypass graft surgery. J Am Coll Cardiol
1998;32:14834.
19. Scandinavian Simvastatin Survival Study Group. Randomised
trial of cholesterol lowering in 4.444 patients with coronary
heart disease: the Scandinavian Simvastatin Survival Study
(4S). Lancet 1994;344:13839.
20. Peduzzi P, Concato J, Feinstein AR et al. Importance of
events per independent variables in proportional hazards
regression analysis. II. Accuracy and precision of regression
estimates. J Clin Epidemiol 1995;48:150310.
21. ESC Working Group on Exercise Physiology, Physiopathology
and Electrocardiography. Guidelines for cardiac exercise
testing. Eur Heart J 1993;14:96988.
22. Berning J, Steensgaard-Hansen F. Early estimation of risk by
echocardiographic determination of wall motion index in an
Use of non-invasive cardiac investigations 925

b
y

g
u
e
s
t

o
n

A
u
g
u
s
t

1
8
,

2
0
1
3
h
t
t
p
:
/
/
e
u
r
h
e
a
r
t
j
.
o
x
f
o
r
d
j
o
u
r
n
a
l
s
.
o
r
g
/
D
o
w
n
l
o
a
d
e
d

f
r
o
m

unselected population with acute myocardial infarction. Am
J Cardiol 1990;65:56776.
23. Sellier P, Chatellier G, Dubois C et al. Prognosis and
risk evaluation in postoperative coronary patients: the
PERISCOP Study: methodology and characteristics of the
included population. Arch Mal Coeur 2001;94:915.
24. Ikaheimo MJ, Huikuri HV, Airaksinen KE et al. Pericardial
effusion after cardiac surgery: incidence, relation to the
type of surgery, antithrombotic therapy, and early coronary
bypass graft patency. Am Heart J 1988;116:97102.
25. Risum , Nitter-Hauge S, Abdelnoor M et al. Mortality and
morbidity after coronary artery bypass surgery related to
pre-operative left ventricular ejection fraction: a follow-up
study. Eur Heart J 1996;17:8749.
26. Huikuri HV, Yli-Mayry Y, Juhani Airaksinen KE et al. Clinical
and angiographic prediction of cardiac death after coronary
artery bypass graft surgery. Br Heart J 1992;67:21620.
27. Vogt A, Grube E, Glunz HG et al. Determinants of mortality
after cardiac surgery: results of the registry of the Arbeits-
gemeinschaft Leitender Kardiologischer Krankenhausarzte
(ALKK) on 10 525 patients. Eur Heart J 2000;21:2832.
28. Myers WO, Blackstone EH, Davis KD et al. CASS Registry.
Long term surgical survival. J Am Coll Cardiol 1999;
33:48898.
29. Moss JA, Goldstein RE, Hall WJ et al. Detection and signifi-
cance of myocardial ischemia in stable patients after recov-
ery from an acute coronary event. JAMA 1993;18:237988.
30. Maggioni AP, Zuanetti G, Franzosi MG et al. Prevalence and
prognostic significance of ventricular arrhythmias after
acute myocardial infarction in the fibrinolytic era.
Circulation 1993;87:31222.
31. Myers J, Prakash M, Froelicher V et al. Exercise capacity and
mortality among men referred for exercise testing. N Engl J
Med 2002;346:793801.
926 P. Sellier et al.

b
y

g
u
e
s
t

o
n

A
u
g
u
s
t

1
8
,

2
0
1
3
h
t
t
p
:
/
/
e
u
r
h
e
a
r
t
j
.
o
x
f
o
r
d
j
o
u
r
n
a
l
s
.
o
r
g
/
D
o
w
n
l
o
a
d
e
d

f
r
o
m