Curriculum in Cardiology

Efficacy of exercise-based cardiac rehabilitation

postmyocardial infarction: A systematic review and
meta-analysis of randomized controlled trials
Patrick R. Lawler, MD, a,b Kristian B. Filion, PhD, c and Mark J. Eisenberg, MD, MPH a,d Montreal, Quebec, Canada;
and Minneapolis, MN

Background Exercise-based cardiac rehabilitation (CR) remains an underused tool for secondary prevention post
myocardial infarction (MI). In part, this arises from uncertainty regarding the efficacy of CR, particularly with respect to
reinfarction, where previous studies have failed to show consistent benefit. We therefore undertook a meta-analysis of
randomized controlled trials (RCTs) to (1) estimate the effect of CR on cardiovascular outcomes and (2) examine the effect of CR
program characteristics on the magnitude of CR benefits.
Methods We systematically searched MEDLINE as well as relevant bibliographies to identify all English-language RCTs
examining the effects of exercise-based CR among post-MI patients. Data were aggregated using random-effects models.
Stratified analyses were conducted to examine the impact of RCT-level characteristics on treatment benefits.
Results We identified 34 RCTs (N = 6,111). Overall, patients randomized to exercise-based CR had a lower risk of
reinfarction (odds ratio [OR] 0.53, 95% CI 0.38-0.76), cardiac mortality (OR 0.64, 95% CI 0.46-0.88), and all-cause
mortality (OR 0.74, 95% CI 0.58-0.95). In stratified analyses, treatment effects were consistent regardless of study periods,
duration of CR, or time beyond the active intervention. Exercise-based CR had favorable effects on cardiovascular risk factors,
including smoking, blood pressure, body weight, and lipid profile.
Conclusions Exercise-based CR is associated with reductions in mortality and reinfarction post-MI. Our secondary
analyses suggest that even shorter CR programs may translate into improved long-term outcomes, although these results need to
be confirmed in an RCT. (Am Heart J 2011;162:571-584.e2.)

Despite guidelines recommending the use of cardiac
rehabilitation (CR) programs for patients with ST-
segment elevation myocardial infarction (MI) 1 and non
ST-segment elevation MI/unstable angina, 2 participation
in these programs continues to be low; only 10% to 20%
of patients who survive an acute MI participate in an
From the aDivisions of Cardiology and Clinical Epidemiology, Jewish General Hospital/
McGill University, Montreal, Quebec, Canada, bDepartment of Medicine, McGill
University Health Center, Montreal, Quebec, Canada, cDivision of Epidemiology and
Community Health, School of Public Health, University of Minnesota, Minneapolis, MN, and
d
Department of Epidemiology, Biostatistics, and Occupational Health, McGill University,
Montreal, Quebec, Canada.
Dr Eisenberg is a National Researcher of the Quebec Foundation for Health Research.
Submitted April 10, 2011; accepted July 21, 2011.
Reprint requests: Mark J. Eisenberg, MD, MPH, FAHA, FACC, Professor of Medicine, Divisions
of Cardiology and Clinical Epidemiology, Jewish General Hospital/McGill University, 3755
Cte Ste-Catherine Road, Suite H-421, Montreal, Quebec, Canada H3T 1E2.
E-mail: mark.eisenberg@mcgill.ca
0002-8703/$ - see front matter
2011, Mosby, Inc. All rights reserved.
doi:10.1016/j.ahj.2011.07.017
exercise-based secondary prevention CR program. 3,4 The
reason for such low participation is likely multifactorial, 4
but 1 important obstacle is the infrastructure to support
prolonged participation in these programs. Another
barrier to usage is likely the absence of large randomized
controlled trials (RCTs) evaluating its efficacy and
insufficient data on what features of CR programs result
in the greatest efficacy.
Individual RCTs studying the efficacy of exercise-based
CR have provided conflicting results, and the most recent
meta-analysis (published in 2004) found that CR de-
creased all-cause and cardiac mortality but had no
beneficial effects on reinfarction. 5 In addition, consider-
able heterogeneity among programs exists, and the
factors that define effective CR on a program-level
including optimal duration 6are currently unclear. 7
Given these areas of uncertainty, we undertook a meta-
analysis of RCTs to (1) accurately estimate the effect of CR
on cardiovascular outcomes and (2) conduct stratified
analyses to examine the effect of RCT-level characteristics
on the benefits of CR.

572 Lawler, Filion, and Eisenberg
Methods
We carried out this systematic review and meta-analysis
following the Preferred Reporting Items for Systematic Reviews
and Meta-Analyses guidelines. 8
Search strategy
We systematically searched MEDLINE to identify RCTs
studying the use of exercise-based CR among survivors of MI.
This search, which is described in detail in Online Appendix 1,
involved search terms including and related to exercise,
exercise therapy, rehabilitation, myocardial infarction, scale. 9 Briefly, RCTs were scored as either high or low
and coronary artery disease. The search was limited to RCTs
published in English and indexed before June 13, 2010. We also
hand-searched bibliographies of retrieved publications and used
PubMed's related articles feature to identify studies not
captured by our primary search strategy.
Inclusion criteria
The inclusion criteria were (1) randomized design aimed at
evaluating the efficacy of exercise-based CR; (2) all patients who
recently survived an MI; (3) the intervention under examination
that involved any form of supervised or unsupervised exercise-
based CR program (which may or may not include other
interventions) in an outpatient, community, or inpatient setting;
(4) a minimum intervention duration of 2 weeks; (5) a minimum
follow-up of 12 weeks; (6) inclusion of a nonexercising control
group; and (7) published in an English-language peer-reviewed
journal. Trials were excluded if they did not meet all of the
inclusion criteria or if they did not present data on at least 1 of
the following outcomes: all-cause mortality, cardiac mortality,
reinfarction, revascularization (percutaneous coronary interven-
tion or coronary artery bypass graft surgery), or modifiable
cardiovascular risk factors (weight, lipids, blood pressure, and
smoking). When RCTs with multiple treatment arms were
included, each treatment arm was considered to be a separate
RCT, and count data from the control arm were subdivided to
ensure that patients in the control arm were not double counted
in the meta-analytic calculations.
Data extraction
Two reviewers independently extracted data using a pre-
specified protocol and standardized reporting form. Disagree-
ments were resolved by consensus or, when necessary, by a
third reviewer. Information about study design, publication
date, country of origin, sample size, and intervention character-
istics was recorded. Data were also recorded regarding patient
baseline demographic and clinical characteristics, including age,
sex, and exposure to catheter-based or surgical revascularization
therapies. Exercise intervention was classified by type of
exercise (aerobic, anerobic, mixed, or not directed), location
(home based, hospital or rehabilitation center based, or mixed),
and whether it was part of a comprehensive rehabilitation
program or an exercise-only intervention. We also extracted
data regarding the following outcomes: all-cause mortality,
cardiac mortality, cardiovascular mortality (which included
cardiac mortality as well as mortality due to cerebrovascular
or peripheral vascular disease), reinfarction, revascularization
(either surgical or catheter based), and change in modifiable
cardiovascular risk factors (body weight, lipids, blood pressure,
American Heart Journal
October 2011
and smoking). Unlike previous meta-analyses, 5 reinfarction
included both fatal and nonfatal events (ie, the risk of
reinfarction independent of outcome), given that a changing
case fatality rate would affect this risk over time. In all cases, if
data for the same study population were reported in multiple
publications, these data were extracted and are presented as a
single RCT.
Quality assessment
The quality of included RCTs was assessed using the Jadad
quality based on 3 primary questions: was the randomization
adequate and explained? Was the study double-blinded
appropriately? and was there a description of withdrawals
and dropouts?
Statistical analyses
Data were aggregated using DerSimonian-Laird random
effects meta-analysis models. Meta-analysis models were
created for the each of the following outcomes: all-cause
mortality, cardiac mortality, cardiovascular mortality, reinfarc-
tion (including both fatal and nonfatal recurrent MI), and
revascularization. All outcome data were analyzed according
to the intention-to-treat principle, and a 0.5 continuity
correction was used for zero-event RCTs. Treatment effects
are presented as odds ratios (ORs) with corresponding 95%
CIs. Most RCTs did not provide cardiovascular risk factor data
in poolable formats; these data were therefore systematically
reviewed but not statistically pooled.
We examined the influence of RCT-level characteristics on the
treatment effects of CR using stratified meta-analyses. First, we
stratified our meta-analyses by characteristics, including study
date (articles published 1990, 1991-2000, and 2001), CR
duration (b3 vs N3 months), sample size (above and below the
median sample size of RCTs reporting a given outcome), and
duration of follow-up (6, 6-12, and N12 months). The sample
size and follow-up time characteristics were used as proxies for
RCT quality, as the open-label nature of these RCTs limited the
ability of the Jadad scale 9 to discriminate between low- and
high-quality RCTs. For analyses stratified by sample size, the
following median values were used: 167 for all-cause mortality,
129 for cardiac mortality, 114 for cardiovascular mortality,
and 169 for reinfarction. We also stratified our analyses by
duration of follow-up beyond the exercise period (b1, 1-12, and
N12 months) to determine if benefits persist long term.
In sensitivity analyses, we first repeated our primary analyses
excluding RCTs with no events in both arms and then repeated
analyses excluding RCTs with no events in 1 arm to examine the
effect of zero-event RCTs. In addition, because 1 RCT 10 did not
specify what criteria were used for MI diagnosis, we excluded
this RCT in additional sensitivity analyses. Because one of the
largest RCTs was published in 1979, 11 we excluded this RCT in a
sensitivity analysis. We also repeated our publication date
analyses with publication date dichotomized (b1995 vs N1995).
This date was used as a marker for the burgeoning acceptance
of several important modern post-MI therapies including
-blockers, 12 coronary stenting, and angiotensin-converting
enzyme (ACE) inhibitorsthe latter primarily among patients
with left ventricular systolic dysfunction. 13,14 Funnel plots were
American Heart Journal
Volume 162, Number 4
Lawler, Filion, and Eisenberg 573

Figure 1

2,169 potentially relevant titles identified and reviewed

Excluded (n = 1,800)
- Not relevant (n = 1,795)
- Review article (n = 3)
- Editorial (n = 1)
- Study design (n = 1)

369 abstracts retrieved

Excluded (n = 222)
- Not relevant (n = 116)
- Not randomized/no non-exercising
control group (n = 75)
- Review article (n = 13)
- Study design article (n = 11)
- Editorial (n = 7)
147 full text articles retrieved

40 articles from 34 RCTs included in meta-analysis
Excluded (n = 107)
- CAD but not exclusively post-MI (n = 53)
- Follow-up inadequate (n = 23)
- No non-exercising controls group (n = 12)
- Not randomized (n = 6)
- Not reporting outcomes of interest (n = 5)
- Exercise intervention inadequate (n = 5)
- Inadequate reporting of methods (n = 1)
- Review (n = 1)
- Study design article (n = 1)

Preferred Reporting Items for Systematic Reviews and Meta-Analyses flow diagram of RCTs included in the meta-analysis. CAD indicates coronary
artery disease.

constructed and visually assessed for publication bias. All
analyses were conducted using Stata version 11.1 (StataCorp
LP, College Station, TX).
No extramural funding was used to support this work. The
authors are solely responsible for the design and conduct of this
study, all study analyses, and drafting and editing of the paper.
Results
Search results
Our search strategy identified 2,169 potentially rele-
vant articles of which 369 were reviewed as abstracts
and 147 as full articles (Figure 1). Thirty-four RCTs were
included in the final analysis. 10,11 , 15-52 One study
reported 15-year follow-up data 27 ; these protracted
follow-up data were not included, as they represented
an outlier, and instead, data from the original report
were included in the analyses. 11
Study and patient characteristics
The characteristics of the 34 identified RCTs are shown
in Table I. All 34 RCTs were open-label RCTs. The exercise
intervention varied considerably and, in some cases, was
not clearly defined. Furthermore, the duration of time
between the index MI and commencement of the CR
program varied between and within studies considerably.
Overall, 19 RCTs used exercise-only CR, 14 used exercise-
based CR as part of a comprehensive secondary preven-
tion program, and 1 used both in 2 independent
intervention arms. 46 Twenty RCTs used aerobic-only
exercise programs, none used anaerobic-only (ie, resis-
tance training) programs, 7 used a mixed aerobic and
anaerobic or undirected program, and 7 RCTs did not
specify the type of exercise used. Duration of exercise-
based CR programs ranged from 2 weeks (the prespeci-
fied minimum for inclusion in the present study) to
3 years. Of note, only 1 RCT (Oya et al 44, N = 28 patients)
 American Heart Journal
574 Lawler, Filion, and Eisenberg October 2011

Table I. Study and patient characteristics of RCTs examining the effect of exercise-based CR on cardiovascular outcomes and modifiable
cardiovascular risk factors
Study Patients Exercise intervention

Key features Home- vs

Total, Age Male of exercise Duration Comprehensive hospital/ Follow-up
Author Year Country N (mean SD) (%) intervention (m) CR group-based CR duration (m)

Ballantyne 1982 Scotland 42 52.8 5.8 100 Canadian Air 6.0 No Home 6.0
et al 15 Force 5BX plan
Bengtsson 16 1983 Sweden 171 Interval training 3.0 Yes Hospital/group 14.0
(jogging,
calisthenics,
cycling) for 30 min
at 90% baseline
peak HR, 2/wk
Bertie 1992 England 110 52.4 1.3 Pulse-monitored 1.0 No Hospital/group 24.0
et al 17 group exercise
sessions, 2/wk
Bethell and 1990 England 229 53.7 7.5 100 Circuit training, 3.0 No Hospital/group 3.0
Mullee 18 3/week
Carlsson 1997 Sweden 168 62.1 75 Interval training 3.0 Yes Hospital/group 12.0
et al 19 with cycling and
jogg ing f or 40
min, 2-3/wk
Carson 1982 England 303 51.5 0.7 100 Circuit training 3.0 No Hospital/group 25.2
et al 20 2/wk
DeBusk 1994 United 585 57.0 8.0 79 Aerobic exercise 12.0 Yes Home 12.0
et al 21 States for 30 min per
day, 5/wk
supervised by a
RN case manager
Dugmore 1999 England 124 55.3 98 Aerobic training 12.0 No Hospital/group 60.0
et al 22 3/wk
Fridlund 1991 Sweden 178 Multimodal CR, 6.0 Yes Hospital/group 12.0
et al 23 including contact
with physiotherapist
for 2 h, 1/wk
Giallauria 2006 Italy 40 68.4 2.5 83 Bicycle ergometer 3.0 No Hospital/group 3.0
et al 24 for 30 min at 60%
of the VO2 peak
achieved at
baseline, 3/wk
Giannuzzi 1993 Italy 103 50.5 8.5 100 Bicycle ergometer 6.0 No Mixed 6.0
et al 26 for 30 min at 80%
baseline peak HR,
3/wk for 2 m
supervised, then
same at home
unsupervised
Giannuzzi 1997 Italy 80 53.5 8.5 95 Bicycle ergometer 6.0 No Mixed 6.0
et al 25 for 30 min at 80%
baseline peak HR,
3/wk for 2 m
supervised and
then same at home
unsupervised
Heldal 2000 Norway 37 53.0 7.8 100 Cycling or running 1.0 No Hospital/group 6.0
et al 28 for 2 h per session
at 85% baseline
peak HR, 5/wk
Heller 1993 Australia 450 58.5 8.0 71 Patients given 6.0 Yes Home 6.0
et al 10 written materials
directing a
walking plan
American Heart Journal
Volume 162, Number 4
Lawler, Filion, and Eisenberg 575

Table I (continued)

Study Patients Exercise intervention

Key features Home- vs

Total, Age Male of exercise Duration Comprehensive hospital/ Follow-up
Author Year Country N (mean SD) (%) intervention (m) CR group-based CR duration (m)

Holmback 1994 Sweden 69 97 Interval training 3.0 No Hospital/group 12.0

et al 29 (jogging,
calisthenics, cycling)
for 45 min at
70%-85% baseline
peak HR, 2/wk
Kallio 1995 Finland 375 54.3 80 Individually 36.0 Yes Mixed (most 36.0
et al 11 1979 tailored program hospital-based)
based on baseline
exercise testing
Karvetti and 1983 Finland 133 53.3 7.1 100 Gymnastics for 12.0 Yes Hospital/group 12.0
Knuts 30 40 min, 3/wk for
3 m, then 1/wk
for 9 m
La Rovere 2002 Italy 95 51.5 7.5 100 Graded exercise 1.0 No Hospital/group 120.0
et al 31 (cycling,
calisthenics) for
30 min at 75%-
95% baseline peak
HR, 5/wk
Lee 2009 Taiwan 39 52 8.0 100 Bicycle ergometer 3.0 No Hospital/group 3.0
et al 32,33 2008 20 min, 55%-70%
peak VO2 on
baseline CPX and
subjective 12-13
on Borg scale,
3/wk
Lewin 1992 Scotland 176 55.8 10.6 72 Patients given a 1.5 Yes Home 12.0
et al 34 manual counseling
them on a home
exercise program
Marchionni 2003 Italy 270 71 Cycling 3/wk, 30 2.0 Yes Hospital/group
et al 35 min at 70%-85%
(Hospital- max HR, 3/wk,
Based and 1 h stretching
Group) and flexibility
sessions, 2/wk
Marchionni Instructed on activity Home
et al 35 as per hospital-
(Home group above, given
Group) an HR monitor/
cycle ergometer for
home exercise
Marra 1985 Italy 167 50.0 7.7 Calisthenics, 2.3 No Hospital/group 55.0
et al 36 cycling 55-70 total
min, 4/wk
Miller 1984 United 198 100 Walking/jogging 5.75 No Hospital/group 6.5
et al 37 States in gymnasium for

(2A) 1 h; patients
Miller regulated their rate 2.0
et al37 (2B) by self-pulse
checks, 3/wk
Miller Stationary cycling 5.75 Home
et al37 (1A) or walking for 30
Miller min. at 70%-85% 2.0
et al37 (1B) baseline peak HR
(with home HR
monitors), 5/wk
(continued on next page)
 American Heart Journal
576 Lawler, Filion, and Eisenberg October 2011

Table I (continued)

Study Patients Exercise intervention

Key features Home- vs

Total, Age Male of exercise Duration Comprehensive hospital/ Follow-up
Author Year Country N (mean SD) (%) intervention (m) CR group-based CR duration (m)

Myers 2000 CHE and 25 55.5 6.1 100 Walking for 1 h 2.0 Yes Hospital/group 12.0
et al 38,39 2001 United 2/d, and cycling
States 45 min 1/wk
Naughton 2000 United 641 100 Aerobic exercises 2.0 No Hospital/group 36.0
et al 40 States at HR 75%-85%
baseline peak HR,
3/wk
Oldridge 1998 Canada 201 52.8 9.5 88 Treadmill, 2.0 Yes Hospital/group 12.0
et al 41-43 1995 stationary cycling,
1991 arm bike for
50 min at 65%
baseline peak HR,
2/wk
Oya 1999 Japan 28 58.6 7.0 93 Cycle ergometer 0.5 No Hospital/group 3.0
et al 44 for 30 min, 2/d
PRECOR 1991 France 121 50.0 100 Cycle ergometer 1.5 Yes Hospital/group 24.0
group plus walking for
25 min, 3/wk
Sivarajan 1982 United 258 56.3 8.4 Progressive 3.0 No Mixed 6.0
et al46 States walking and
(Group B1)# calisthenics
Sivarajan program modified 3.0 Yes Mixed 6.0
et al46 weekly depending
(Group B2)# on performance
Specchia 1996 Italy 256 52.9 7.5 91 Bicycle ergometer 1.0 No Hospital/group 34.5
et al 47 for 30 min
progressing to 75%
baseline peak
maximal work
capacity, with
calisthenics 5/wk
Stern 1983 United 71 85 Mixed aerobic 3.0 No Hospital/group 12.0
et al 48 States exercise for 1 h at
85% baseline peak
HR, 3/wk
Taylor 1988 United 210 100 Patient- 5.75 Yes Mixed 6.5
et al 49,50 1986 States individualized
exercise program
by RN
Vermeulen 1983 Holland 98 49.2 4.1 100 Multimodal CR 1.5 Yes Hospital/group 60.0
et al 51 program including
physical
intervention
Zheng 2008 China 60 Bicycle ergometer 6.0 No Hospital/group 6.0
et al 52 for 30 min at
baseline peak HR,
3/wk

CHE, Switzerland, CPX, cardiopulmonary exercise test; HR, heart rate; RN, registered nurse; /wk, number of times per week; VO2 is a measure of oxygen consumption.
Most programs included brief warm-up and cool-down periods before and after the exercise intervention. Determination of baseline peak heart rate was, in all cases, done with
cardiopulmonary testing after MI.
The Canadian Air Force 5BX (Five Basic Exercises) plan is an incremental calisthenics and aerobic exercise program.
Corresponds to fairly light to somewhat hard.
www.theheartmanual.com.
SD not provided.
Miller et al37 included 4 intervention groups: home group with 23 weeks of exercise (1A), home group with 8 weeks of exercise (1B), center-based group with 23 weeks, and
center-based group with 8 weeks (2B).
# Sivarajan et al46 included 2 groups: exercise-only CR (group B1) and comprehensive including exercise CR (group B2).
American Heart Journal
Volume 162, Number 4
Lawler, Filion, and Eisenberg 577

Figure 2

Forest plot of effect of exercise-based CR on reinfarction. Data were pooled using random effects models. Exercise-based CR significantly reduces
reinfarction among MI survivors.

Table II. Effect of exercise-based CR on overall cardiovascular outcomes and outcomes stratified by duration of exercise intervention, length of
follow-up beyond the intervention period, sample size, and date of publication

Cardiovascular
All-Cause mortality mortality Cardiac mortality Reinfarction

N OR (95% CI) N OR (95% CI) N OR (95% CI) N OR (95% CI)

All Studies 31 0.74 (0.58-0.95) 18 0.61 (0.40-0.91) 22 0.64 (0.46-0.88) 18 0.54 (0.38-0.76)

Stratified Analyses
Exercise duration of 0.5 to 3 m 20 0.71 (0.51-1.01) 9 0.83 (0.31-2.22) 12 0.59 (0.34-1.03) 13 0.69 (0.43-1.11)
Exercise duration of N3 m 11 0.77 (0.54-1.09) 9 0.57 (0.37-0.88) 10 0.66 (0.44-0.98) 5 0.40 (0.24-0.66)
Pre-1995 19 0.77 (0.59-1.01) 9 0.70 (0.48-1.01) 7 0.58 (0.38-0.89) 12 0.58 (0.40-0.85)
1995 or later 8 0.59 (0.32-1.11) 9 0.49 (0.26-0.91) 7 0.83 (0.26-2.65) 3 0.33 (0.14-0.82)
Small studies 15 0.54 (0.24-1.21) 9 0.81 (0.29-2.25) 11 0.59 (0.29-1.17) 9 0.48 (0.28-0.81)
Large studies 16 0.77 (0.59-0.99) 9 0.57 (0.37-0.89) 11 0.65 (0.45-0.94) 9 0.58 (0.37-0.93)
b1 m between end of CR 13 0.81 (0.57-1.15) 10 0.57 (0.36-0.90) 11 0.67 (0.45-1.00) 4 0.55 (0.27-1.10)
and end of follow-up
1-12 m between end of CR 12 0.95 (0.56-1.61) 6 1.25 (0.32-4.85) 6 1.25 (0.32-4.85) 7 0.45 (0.21-0.98)
and end of follow-up
N12 m between end of 6 0.54 (0.35-0.85) 2 0.51 (0.15-1.76) 5 0.51 (0.29-0.92) 7 0.56 (0.35-0.90)
CR and end of follow-up
Comprehensive CR 20 0.62 (0.43-0.91) 13 0.70 (0.29-1.70) 16 0.57 (0.34-0.97) 13 0.63 (0.39-1.02)
Exercise-only CR 11 0.85 (0.61-1.17) 5 0.58 (0.37-0.92) 6 0.68 (0.45-1.01) 5 0.45 (0.27-0.74)

Distinct trial arms were analyzed as separate studies referenced to a duplicate control group and were not double counted).
Cardiovascular mortality was defined as mortality due to cardiac, cerebrovascular, or peripheral vascular diseases.

578 Lawler, Filion, and Eisenberg
Figure 3
Forest plot of effect of exercise-based CR on cardiac mortality. Data were pooled using random effects models. Exercise-based CR significantly
reduces cardiac mortality among MI survivors.
used a CR program shorter than 4 weeks, whereas all
other RCTs used a minimum program duration of
1 month. Follow-up duration ranged from 3 months (the
prespecified minimum for inclusion) to 5 years. Twenty-
seven RCTs were single-center RCTs, and 8 were multi-
center RCTs. Twelve studies were published in 1990 or
earlier, 17 were published during 1991 to 2000, and
5 were published between 2001 and June 2010. Based on
the Jadad scale, 5 RCTs were classified as high quality,
whereas the remainder was classified as low quality.
The total number of patients randomized was 6,111.
Most patients randomized were men (88.9%). The mean
age of participants at the time of enrollment was
54.7 years. Only 5 RCTs reported that patients were
eligible for exercise-based CR after percutaneous coro-
nary intervention as treatment for patients' index MI, and
3 studies reported including subjects after coronary
artery bypass graft.
Overall cardiovascular outcomes
Overall, patients randomized to exercise-based CR
were at significantly lower risk of reinfarction than
those randomized to control (OR 0.53, 95% CI 0.38, 0.76)
(Figure 2; Table II). Patients randomized to exercise-
American Heart Journal
October 2011
based CR were also at significantly lower risk of cardiac
death (OR 0.64, 95% CI 0.46-0.88) (Figure 3), cardiovas-
cular death (OR 0.60, 95% CI 0.40-0.76), and all-cause
mortality (OR 0.74, 95% CI 0.58-0.95) (Figure 4). No
significant difference in revascularization was observed
(OR 0.92, 95% CI 0.68-1.25), although these data were
inconsistently reported, and our pooled estimates are
accompanied by a wide 95% CI.
Stratified analyses
The impact of several study-level covariates on treat-
ment benefits was examined using stratified meta-
analyses. To investigate whether there was evidence of
benefit among shorter duration exercise-based CR pro-
grams, we performed stratified analyses of outcomes
using a program duration of 3 months as a cutoff. When
this was done, the effect sizes were similar among those
studies with a program duration 3 months as compared
with those b3 months for the outcome examined
(Table II), although the effects in the shorter duration
programs did not reach statistical significance.
To examine if benefits on reinfarction and mortality
persisted beyond the period of active treatment, we
analyzed data based on duration of follow-up beyond the
American Heart Journal
Volume 162, Number 4
Figure 4
Forest plot of effect of exercise-based CR on all-cause mortality. Data were pooled using random effects models. Exercise-based CR significantly
reduces all-cause mortality among MI survivors.
period of exercise intervention. There were consistent
beneficial effects seen for reinfarction, cardiac mortality,
and all-cause mortality across these follow-up categories
(Table II), suggesting that the benefits of a CR program
post-MI persist beyond the period of active intervention.
There was no suggestion of difference between the
overall findings among small versus large studies using
the median sample size as the cutoff (Table II). The
impact of date of publication was investigated by
stratifying RCTs into time-related strata; the beneficial
effects of CR on cardiac outcomes persisted across study
periods (Table II). Finally, studies were stratified by
comprehensive versus exercise-only CR, which showed
persistent benefit among outcomes in both strata
(Table II).
Cardiovascular risk factors
Several RCTs examined the effect of CR on modifiable
cardiovascular risk factors, including blood pressure,
smoking cessation, lipids, and weight loss. Overall, RCTs
found a more favorable reduction in the prevalence of
Lawler, Filion, and Eisenberg 579
smoking among those randomized to exercise-based CR
(Online Appendix 2). Similarly, effects on systolic and
diastolic blood pressure were generally more favorable in
the intervention arms compared with the control arms
(Online Appendix 2). Total cholesterol showed a trend
toward reduction in the intervention arm of several RCTs,
whereas there was no evidence of a decrease in the
corresponding control arms (Online Appendix 3).
Changes in body weight were minimal in both groups
(Online Appendix 3).
Sensitivity analyses
Results of sensitivity analyses that excluded zero-event
RCTs (in either one or both arms) were consistent with
those of our primary analyses (data not shown).
Furthermore, exclusion of the RCT by Heller et al, 10
which did not confirm the diagnosis of MI, did not
appreciably alter the results. Initial stratification by
publication date was done using decades as cutoffs.
When we reanalyzed data using 1995 as a cutoff, similar
results were obtained (Table II). Finally, given that the

580 Lawler, Filion, and Eisenberg
Figure 5
Funnel plot for meta-analysis examining the effect of exercise-based
CR on the risk of reinfarction. Dashed lines represent pseudo 95% CIs.
largest all-cause and cardiac mortality RCT was published
in 1979, 11 we conducted sensitivity analyses excluding
this RCT and observed similar overall results (all-cause
mortality: OR 0.78, 95% CI 0.58-1.04; cardiac mortality:
OR 0.72, 95% CI 0.47-1.10).
Publication bias
Visual inspection of funnel plots for all-cause mortality,
cardiac mortality, and reinfarction (Figure 5) did not
indicate the presence of publication bias.
Discussion
Our study was designed to examine the therapeutic
effects of exercise-based CR as secondary prevention
post-MI. Our primary analysis focused on establishing the
overall benefit of exercise-based CR. Our results demon-
strate a statistically significant reduction in reinfarction,
cardiac mortality, cardiovascular mortality, and all-cause
mortality with exercise-based CR. To our knowledge, this
is the first meta-analysis to show a statistically significant
reduction in reinfarction with exercise-based CR post-MI.
Our secondary analyses investigating the effect of RCT-
level characteristics suggested that even short-duration
CR programs (1-3 months) could be of benefit and that
the overall observed benefits persisted beyond the period
of active intervention.
Exercise-based CR is currently a class I indication in
most clinical practice guidelines, including those for ST-
segment elevation MI 1 and unstable angina/nonST-
segment elevation MI. 2 However, despite decades of
recommended use, the available evidence comes from
several small mostly single-center RCTs. Results of these
RCTs have been individually inconclusive. We conducted
American Heart Journal
October 2011
our analyses through 30 years of evidence to synthesize
the overall treatment effect, and our results support
these guidelines.
To our knowledge, this study is the first meta-analysis to
show a significant reduction in reinfarction. Likely, this is
related to differences in the definition of reinfarction (we
included both nonfatal and fatal events) as well as
differences in inclusion criteria. Recent evidence from
the OASIS-5 Trial supported the finding that exercise (and
diet modification) reduces early reinfarction rates after
acute coronary syndromes, 53 as did a recent observational
study. 54 Establishing that CR reduces reinfarction is of
great importance in effecting policy changes in a dynamic
health care environment that is emphasizing prudent and
parsimonious resource allocation. Indeed, despite sub-
stantial advances in the treatment and secondary preven-
tion of MI, subsequent mortality in the first year after MI
remains as high as 18% among men and 23% among
women N40 years. 55 Similarly, 20% of patients are
rehospitalized in the year after a first MI, 55 and the cost
of this rehospitalization averages N30% that of the index
hospitalization. 56,57 Among therapies available after MI,
usage of exercise-based CR lags considerably behind that
of pharmacological secondary prevention strategies,
including -blockers, antiplatelet agents, statins, and
ACE inhibitors. Our results suggest that exercise-based
CR reduces reinfarction and death. Given that high event
rates continue to be observed in these patients, better
implementation of exercise-based CR resources could
prove a powerful tool for reducing morbidity, mortality,
and potentially health care costs after MI.
The secondary objective of our study was to examine
the effects of study-level characteristics, including treat-
ment duration, on the efficacy of CR. Referral rates for CR
lag behind prescribing rates of evidence-based pharma-
cotherapy after MI, and participation in CR after referral is
even lower. 3,4 Although there clearly are multiple
obstacles along the road from index MI to successful
completion of a CR program, one of the largest is
availability and accessibility of resources. Our results,
although only hypothesis generating, suggest that even
short-term CR programs (1-3 months) may be effective as
secondary prevention programs post-MI. With an aging
population and health care costs that continue to rise,
optimal duration of CR has become a recent area of
importance, and it is speculated that if shorter programs
prove to have similar efficacy as longer programs, 6 use of
existing infrastructure could be expanded to additional
survivors of MI at marginal added cost. 7 Unfortunately,
there is few data currently available on what defines the
optimal duration of exercise-based CR. 58 Our results
suggest that the benefits of exercise-based CR might be
incurred earlier than previously thought and encourage
further research into this question as available evidence is
clearly inadequate to make clinical and policy decisions.
They should, however, motivate additional studies with
American Heart Journal
Volume 162, Number 4
the specific objective to investigate at what point post-MI
optimal benefits occur.
There is biologic basis for the hypothesis that exercise
improves outcomes post-MI independent of its effects of
risk factors. Specifically, experimental evidence impli-
cates exercise post-MI as an important determinant of
contractility and myofilament Ca 2+ sensitivity 59,60 as well
as a potentially important regulator of the renin-angio-
tensin-aldosterone system post-MI 61 and myocardial
fibrosis and remodeling. 62 Overall, these studies suggest
that exercise favorably modulates several physiologic and
pathologic processes at play post-MI when important
remodeling changes are underway.
Several included RCTs were conducted before the era
of current evidence-based treatment and secondary
prevention strategies, including coronary angioplasty
with and without intracoronary stenting, statins,
-blockers, dual antiplatelet therapy, and ACE inhibitors.
When we investigated the potential effect of publication
date, a proxy for concomitant therapies, on the benefits
of CR using stratified analyses, these analyses revealed no
evidence that the beneficial effects of CR were changing
over time. However, with the advent and acceptance of
these other evidence-based secondary preventative
therapies, it is clear that the effect sizes of the benefits
seen in our study are unlikely to represent the effect sizes
that would be expected from exercise-based CR in
modern practice. These concerns are not limited to CR,
as evidence for several commonly used secondary
prevention therapies is derived from data collected
some years ago in the absence of the full modern
compliment of therapies. Nonetheless, although also true
for other therapies, this may be particularly true for
exercise-based CR. In addition, it is important to
emphasize that these results are drawn from a population
of patients who were predominantly male and who were
generally younger than the average age of presentation
for patients with MI currently. These demographic
characteristics highlight a shortcoming of the CR
literature, and as it is becoming clear in many other
fields of cardiology as well, women and older patients
represent very important demographics in whom
dedicated study is very much needed.
Two early meta-analyses from the 1980s suggested
benefit of exercise-based CR among patients after MI. 63,64
In 2003, one of these groups updated their analysis and
showed statistically significant reductions in cardiac and
all-cause mortality. 5 Our findings are consistent with
these previous results but include a significant reduction
in reinfarction. This would be expected to be the
underlying cause of the observed reductions in cardiac
and all-cause mortalities but was not found in this
previous meta-analysis. Likely, this arose because in
previous meta-analyses, the risk of reinfarction was
determined only for nonfatal reinfarction, whereas we
used a composite of fatal and nonfatal reinfarction. We
Lawler, Filion, and Eisenberg 581
chose this approach because mortality rates have
changed considerably over time, and limiting analysis to
nonfatal events, therefore, may bias the measured effect
of exercise-based CR on reinfarction risk. Another
possible explanation for these divergent findings is our
more rigorous selection criteria, limiting our analysis to
peer-reviewed journal articles, where the previous meta-
analysis also included data from abstracts and nonpeer-
reviewed sources (eg, thesis dissertations). Our results
also include expanded secondary analyses and several
newer RCTs. With the inclusion of these more recent
RCTs, the consistency of results over time suggests that
there is still a role for exercise-based CR in the presence
of newer evidence-based MI treatment. Our secondary
analyses offer a platform for hypothesis generation and
will hopefully motivate and direct further study. Ulti-
mately, a modern well-powered RCT could answer many
of the questions that remain herein, but regrettably, it is
unlikely that such a study will be undertaken given that it
would be of questionable ethics to randomize patients
into a nonexercising group with the breath of evidence in
support of exercise-based CR.
Our study has several potential limitations. First,
because of the nature of the intervention, all included
RCTs were open label. We attempted to control for this
by focusing on objective end points, including mortality
and reinfarction, and used sample size and duration of
follow-up time as proxies for study quality in our
secondary analyses. Furthermore, although not always
noted, some RCTs reported that outcomes were adjudi-
cated by investigators blinded to the subjects' random-
ization. Second, because of the limited number of RCTs
examining each outcome and their relatively small
number of events, some of our subgroup analyses had
wide 95% CIs. Third, approximately 89% of included
patients were men. As discussed above, additional
studies are required among women and older patients
post-MI. Fourth, our inclusion criteria were relatively
broad to permit our search strategy to capture all relevant
articles, as this is a field dominated by numerous small
trials. As such, there was heterogeneity in study design,
populations investigated, and intervention examined,
including variability in the type and quality of CR
program. We therefore used random effects models to
account for both within- and between-study variability.
Furthermore, through our stratified analyses, we inves-
tigated potential sources of heterogeneity and their
influence on the effects of CR programs on clinical
outcomes. With specific respect to the inclusion criteria
allowing RCTs using b4 weeks of CR, we would point out
that only 1 RCT used an exercise duration b4 weeks, 44
and this study contributed only N = 28 patients to the
total pool of 6,111 patients studied, and hence, it is
unlikely that changing the minimum program duration to
4 weeks would have significantly affected our results.
Finally, data for cardiovascular risk factors were often not

582 Lawler, Filion, and Eisenberg
presented in a poolable manner. Consequently, these
data were systematically reviewed but not formally
pooled. These data were also generally restricted to
subjects who attended follow-up visits and thus are
presented using a modified intention-to-treat approach,
which may lead to bias. Nonetheless, we have provided
these data for completeness.
Conclusion
Exercise-based CR reduces the risk of reinfarction and
cardiac, cardiovascular, and all-cause mortality when
used as secondary prevention post-MI. Even if more
modest benefits are obtained when CR is used in
everyday current practice, our results suggest that
exercise-based CR is efficacious for secondary preven-
tion. Our examination of RCT-level characteristics
suggests that even short-term CR programs post-MI
could be of benefit and that the overall observed benefits
persisted beyond the period of active intervention. These
results encourage future RCTs to compare the efficacy of
short- and long-term CR programs for the secondary
prevention of cardiovascular events.
Acknowledgements
We would like to thank Tara Dourian for her help with
data abstraction.
Disclosures
The authors declare no conflict of interest.
References
1. Antman EM, Anbe DT, Armstrong PW, et al. ACC/AHA guidelines
for the management of patients with ST-elevation myocardial
infarctionexecutive summary: a report of the American College of
Cardiology/American Heart Association Task Force on Practice
Guidelines (Writing Committee to Revise the 1999 Guidelines for the
Management of Patients With Acute Myocardial Infarction). Circu-
lation 2004;110:588-636.
2. Anderson JL, Adams CD, Antman EM, et al. ACC/AHA 2007
guidelines for the management of patients with unstable angina/
non ST-elevation myocardial infarction: a report of the American
College of Cardiology/American Heart Association Task Force on
Practice Guidelines (Writing Committee to Revise the 2002
Guidelines for the Management of Patients With Unstable Angina/
Non ST-Elevation Myocardial Infarction): developed in collabora-
tion with the American College of Emergency Physicians, the
Society for Cardiovascular Angiography and Interventions, and the
Society of Thoracic Surgeons: endorsed by the American Associ-
ation of Cardiovascular and Pulmonary Rehabilitation and the
Society for Academic Emergency Medicine. Circulation 2007;116:
e148-304.
3. Suaya JA, Shepard DS, Normand SL, et al. Use of cardiac
rehabilitation by Medicare beneficiaries after myocardial infarction
or coronary bypass surgery. Circulation 2007;116:1653-62.
American Heart Journal
October 2011
4. Leon AS, Franklin BA, Costa F, et al. Cardiac rehabilitation and
secondary prevention of coronary heart disease: an American Heart
Association scientific statement from the Council on Clinical
Cardiology (Subcommittee on Exercise, Cardiac Rehabilitation, and
Prevention) and the Council on Nutrition, Physical Activity, and
Metabolism (Subcommittee on Physical Activity), in collaboration with
the American association of Cardiovascular and Pulmonary Reha-
bilitation. Circulation 2005;111:369-76.
5. Taylor RS, Brown A, Ebrahim S, et al. Exercise-based rehabilitation
for patients with coronary heart disease: systematic review and meta-
analysis of randomized controlled trials. Am J Med 2004;116:
682-92.
6. Reid RD, Dafoe WA, Morrin L, et al. Impact of program duration and
contact frequency on efficacy and cost of cardiac rehabilitation:
results of a randomized trial. Am Heart J 2005;149:862-8.
7. Thompson DR, Clark AM. Cardiac rehabilitation: into the future.
Heart 2009;95:1897-900.
8. Moher D, Liberati A, Tetzlaff J, et al. Preferred reporting items for
systematic reviews and meta-analyses: the PRISMA statement. BMJ
2009;b2535:339.
9. Jadad AR, Moore RA, Carroll D, et al. Assessing the quality of reports
of randomized clinical trials: is blinding necessary? Control Clin Trials
1996;17:1-12.
10. Heller RF, Knapp JC, Valenti LA, et al. Secondary prevention after
acute myocardial infarction. Am J Cardiol 1993;72:759-62.
11. Kallio V, Hamalainen H, Hakkila J, et al. Reduction in sudden deaths
by a multifactorial intervention programme after acute myocardial
infarction. Lancet 1979;2:1091-4.
12. Teo KK, Yusuf S, Furberg CD. Effects of prophylactic antiarrhythmic
drug therapy in acute myocardial infarction. An overview of
results from randomized controlled trials. JAMA 1993;270:
1589-95.
13. Pfeffer MA, Braunwald E, Moye LA, et al. Effect of captopril on
mortality and morbidity in patients with left ventricular dysfunction
after myocardial infarction. Results of the survival and ventricular
enlargement trial. The SAVE Investigators. N Engl J Med 1992;327:
669-77.
14. Ambrosioni E, Borghi C, Magnani B. The effect of the angiotensin-
converting-enzyme inhibitor zofenopril on mortality and morbidity
after anterior myocardial infarction. The Survival of Myocardial
Infarction Long-Term Evaluation (SMILE) Study Investigators. N Engl J
Med 1995;332:80-5.
15. Ballantyne FC, Clark RS, Simpson HS, et al. The effect of
moderate physical exercise on the plasma lipoprotein subfractions
of male survivors of myocardial infarction. Circulation 1982;65:
913-8.
16. Bengtsson K. Rehabilitation after myocardial infarction. A controlled
study. Scand J Rehabil Med 1983;15:1-9.
17. Bertie J, King A, Reed N, et al. Benefits and weaknesses of a cardiac
rehabilitation programme. J R Coll Physicians Lond 1992;26:
147-51.
18. Bethell HJ, Mullee MA. A controlled trial of community based
coronary rehabilitation. Br Heart J 1990;64:370-5.
19. Carlsson R, Lindberg G, Westin L, et al. Influence of coronary nursing
management follow up on lifestyle after acute myocardial infarction.
Heart 1997;77:256-9.
20. Carson P, Phillips R, Lloyd M, et al. Exercise after myocardial
infarction: a controlled trial. J R Coll Physicians Lond 1982;16:
147-51.
21. DeBusk RF, Miller NH, Superko HR, et al. A case-management system
for coronary risk factor modification after acute myocardial
infarction. Ann Intern Med 1994;120:721-9.
American Heart Journal
Volume 162, Number 4
22. Dugmore LD, Tipson RJ, Phillips MH, et al. Changes in cardiorespi-
ratory fitness, psychological wellbeing, quality of life, and vocational
status following a 12 month cardiac exercise rehabilitation pro-
gramme. Heart 1999;81:359-66.
23. Fridlund B, Hogstedt B, Lidell E, et al. Recovery after myocardial
infarction. Effects of a caring rehabilitation programme. Scand J
Caring Sci 1991;5:23-32.
24. Giallauria F, Lucci R, De LA, et al. Favourable effects of exercise
training on N-terminal probrain natriuretic peptide plasma levels in
elderly patients after acute myocardial infarction. Age Ageing 2006;
35:601-7.
25. Giannuzzi P, Temporelli PL, Corra U, et al. Attenuation of unfavorable
remodeling by exercise training in postinfarction patients with left
ventricular dysfunction: results of the Exercise in Left Ventricular
Dysfunction (ELVD) trial. Circulation 1997;96:1790-7.
26. Giannuzzi P, Tavazzi L, Temporelli PL, et al. Long-term physical
training and left ventricular remodeling after anterior myocardial
infarction: results of the Exercise in Anterior Myocardial Infarction
(EAMI) trial. EAMI Study Group. J Am Coll Cardiol 1993;22:
1821-9.
27. Hamalainen H, Luurila OJ, Kallio V, et al. Reduction in sudden deaths
and coronary mortality in myocardial infarction patients after
rehabilitation. 15 year follow-up study. Eur Heart J 1995;16:
1839-44.
28. Heldal M, Sire S, Dale J. Randomised training after myocardial
infarction: short and long-term effects of exercise training after
myocardial infarction in patients on beta-blocker treatment. A
randomized, controlled study. Scand Cardiovasc J 2000;34:59-64.
29. Holmback AM, Sawe U, Fagher B. Training after myocardial
infarction: lack of long-term effects on physical capacity and
psychological variables. Arch Phys Med Rehabil 1994;75:551-4.
30. Karvetti RL, Knuts LR. Effects of comprehensive rehabilitation on
weight reduction in myocardial infarction patients. Scand J Rehabil
Med 1983;15:11-6.
31. La Rovere MT, Bersano C, Gnemmi M, et al. Exercise-induced
increase in baroreflex sensitivity predicts improved prognosis after
myocardial infarction. Circulation 2002;106:945-9.
32. Lee BC, Chen SY, Hsu HC, Su MY, et al. Effect of cardiac
rehabilitation on myocardial perfusion reserve in postinfarction
patients. Am J Cardiol 2008;101:1395-402.
33. Lee BC, Hsu HC, Tseng WY, et al. Effect of cardiac rehabilitation on
angiogenic cytokines in postinfarction patients. Heart 2009;95:
1012-8.
34. Lewin B, Robertson IH, Cay EL, et al. Effects of self-help post
myocardial-infarction rehabilitation on psychological adjustment and
use of health services. Lancet 1992;339:1036-40.
35. Marchionni N, Fattirolli F, Fumagalli S, et al. Improved exercise
tolerance and quality of life with cardiac rehabilitation of older
patients after myocardial infarction: results of a randomized,
controlled trial. Circulation 2003;107:2201-6.
36. Marra S, Paolillo V, Spadaccini F, et al. Long-term follow-up after a
controlled randomized post-myocardial infarction rehabilitation
programme: effects on morbidity and mortality. Eur Heart J 1985;6:
656-63.
37. Miller NH, Haskell WL, Berra K, et al. Home versus group exercise
training for increasing functional capacity after myocardial infarc-
tion. Circulation 1984;70:645-9.
38. Myers J, Gianrossi R, Schwitter J, et al. Effect of exercise training on
postexercise oxygen uptake kinetics in patients with reduced
ventricular function. Chest 2001;120:1206-11.
39. Myers J, Goebbels U, Dzeikan G, et al. Exercise training and
myocardial remodeling in patients with reduced ventricular function:
Lawler, Filion, and Eisenberg 583
one-year follow-up with magnetic resonance imaging. Am Heart J
2000;139:252-61.
40. Naughton J, Dorn J, Oberman A, et al. Maximal exercise systolic
pressure, exercise training, and mortality in myocardial infarction
patients. Am J Cardiol 2000;85:416-20.
41. Oldridge N, Guyatt G, Jones N, et al. Effects on quality of life with
comprehensive rehabilitation after acute myocardial infarction. Am J
Cardiol 1991;67:1084-9.
42. Oldridge N, Gottlieb M, Guyatt G, et al. Predictors of health-related
quality of life with cardiac rehabilitation after acute myocardial
infarction. J Cardiopulm Rehabil 1998;18:95-103.
43. Oldridge N, Streiner D, Hoffmann R, et al. Profile of mood states and
cardiac rehabilitation after acute myocardial infarction. Med Sci
Sports Exerc 1995;27:900-5.
44. Oya M, Itoh H, Kato K, et al. Effects of exercise training on the
recovery of the autonomic nervous system and exercise capacity after
acute myocardial infarction. Jpn Circ J 1999;63:843-8.
45. P.RE.COR Group. Comparison of a rehabilitation programme, a
counselling programme and usual care after an acute myocardial
infarction: results of a long-term randomized trial. P.RE.COR. Group.
Eur Heart J 1991;12:612-6.
46. Sivarajan ES, Bruce RA, Lindskog BD, et al. Treadmill test responses to
an early exercise program after myocardial infarction: a randomized
study. Circulation 1982;65:1420-8.
47. Specchia G, De SS, Scire A, et al. Interaction between exercise
training and ejection fraction in predicting prognosis after a first
myocardial infarction. Circulation 1996;94:978-82.
48. Stern MJ, Gorman PA, Kaslow L. The group counseling v exercise
therapy study. A controlled intervention with subjects following
myocardial infarction. Arch Intern Med 1983;143:1719-25.
49. Taylor CB, Houston-Miller N, Haskell WL, et al. Smoking cessation
after acute myocardial infarction: the effects of exercise training.
Addict Behav 1988;13:331-5.
50. Taylor CB, Houston-Miller N, Ahn DK, et al. The effects of exercise
training programs on psychosocial improvement in uncomplicated
postmyocardial infarction patients. J Psychosom Res 1986;30:581-7.
51. Vermeulen A, Lie KI, Durrer D. Effects of cardiac rehabilitation after
myocardial infarction: changes in coronary risk factors and long-term
prognosis. Am Heart J 1983;105:798-801.
52. Zheng H, Luo M, Shen Y, et al. Effects of 6 months exercise training on
ventricular remodelling and autonomic tone in patients with acute
myocardial infarction and percutaneous coronary intervention.
J Rehabil Med 2008;40:776-9.
53. Chow CK, Jolly S, Rao-Melacini P, et al. Association of diet, exercise,
and smoking modification with risk of early cardiovascular events
after acute coronary syndromes. Circulation 2010;121:750-8.
54. Witt BJ, Jacobsen SJ, Weston SA, et al. Cardiac rehabilitation after
myocardial infarction in the community. J Am Coll Cardiol 2004;44:
988-96.
55. Lloyd-Jones D, Adams R, Carnethon M, et al. Heart disease and
stroke statistics2009 update: a report from the American Heart
Association Statistics Committee and Stroke Statistics Subcommittee.
Circulation 2009;119:e21-181.
56. Patel NB, Balady GJ. The rewards of good behavior. Circulation
2010;121:733-5.
57. Menzin J, Wygant G, Hauch O, et al. One-year costs of ischemic
heart disease among patients with acute coronary syndromes:
findings from a multi-employer claims database. Curr Med Res Opin
2008;24:461-8.
58. Thompson DR, Lewin RJ. Coronary disease. Management of the post
myocardial infarction patient: rehabilitation and cardiac neurosis.
Heart 2000;84:101-5.

584 Lawler, Filion, and Eisenberg
59. de Waard MC, van d V, Bito V, et al. Early exercise training normalizes
myofilament function and attenuates left ventricular pump dysfunction in
mice with a large myocardial infarction. Circ Res 2007;100:1079-88.
60. Schober T, Knollmann BC. Exercise after myocardial infarction
improves contractility and decreases myofilament Ca2+ sensitivity.
Circ Res 2007;100:937-9.
61. Wan W, Powers AS, Li J, et al. Effect of post-myocardial infarction
exercise training on the renin-angiotensin-aldosterone system and
cardiac function. Am J Med Sci 2007;334:265-73.
American Heart Journal
October 2011
62. Xu X, Wan W, Powers AS, et al. Effects of exercise training on cardiac
function and myocardial remodeling in post myocardial infarction
rats. J Mol Cell Cardiol 2008;44:114-22.
63. O'Connor GT, Buring JE, Yusuf S, et al. An overview of randomized
trials of rehabilitation with exercise after myocardial infarction.
Circulation 1989;80:234-44.
64. Oldridge NB, Guyatt GH, Fischer ME, et al. Cardiac rehabilitation
after myocardial infarction. Combined experience of randomized
clinical trials. JAMA 1988;260:945-50.
American Heart Journal
Volume 162, Number 4
Lawler, Filion, and Eisenberg 584.e1

Appendix. Online Supplements

Description
Online of MEDLINE
Appendix literature search
1. Description conducted
of MEDLINE on June
literature 13,conducted
search 2010 on June 13, 2010
Search no. Search description No. of results

1 exercise[MeSH] OR exercise therapy[MeSH] OR rehabilitation[MeSH] OR rehabilitation[tiab] 314 120

OR exercise[tiab] OR rehab[tiab]
2 coronary artery disease[MeSH] OR coronary artery disease[tiab] OR myocardial infarction 203 201
[MeSH] OR myocardial infarction[tiab] OR acute coronary syndrome[MeSH] OR acute
coronary syndrome[tiab] OR percutaneous coronary intervention[tiab] OR PCI[tiab] OR PTCA[tiab]
OR stent[tiab] OR Angioplasty, Transluminal, Percutaneous Coronary[MeSH] OR
coronary artery bypass[tiab] OR CABG[tiab] OR cardiac surgery[tiab]
OR coronary artery bypass[MeSH]
3 Searches 2 and 3 17 899
4 Limit search 3 to humans 17 117
5 Limit search 4 to English language 13 057
6 Limit search 5 to Clinical Trial, Randomized Controlled Trial 2169

Online
Effect ofAppendix 2. Effect
exercise-based CR onofsmoking
exercise-based CR on
and blood smoking
pressure and blood
among pressure
patients among patients post-MI
post-MI
Smoking (n/N) Blood pressure (systolic/diastolic) (Mean SD)

Exercise-based CR Control Exercise-based CR Control

Study Year Baseline Follow-up Baseline Follow-up Baseline Follow-up Baseline Follow-up

Bertie et al 17 1992 20/57 4/43 21/53 9/38

18
Bethell and Mullee 1990 55/99 12/99 60/101 8/101
Carlsson et al 19 1997 32/87 21/87 35/81 25/81
DeBusk et al 21 1994 131/293 39/248 121/292 57/259
Fridlund et al 23 1991 30/53 30/52 27/63 27/58
Giallauria et al 24 2006 12/20 13/20 129 15/ 128 7/ 130 13/ 132 9/
Giannuzzi et al 26 1993 34/49 34/46 129 18/ 125 16/ 123 13/ 127 20/
Giannuzzi et al 25 1997 30/39 27/38 124 12/ 117 12/ 120 10/ 120 12/
Heller et al 10 1993 66/213 18/168 74/237 26/207
Kallio et al 11 1979 107/183 113/187 129 19/ 136 24/ 128 18/ 145 22/
84 12 85 11 83 11 88 11
Lee et al 32,33 2008/2009 0/20 0/19 126 16/ 122 16/ 130 20/ 124 19/
Marra et al 36 1985 74/81 18/81 72/80 25/80
Myers et al 39 2000 11/12 11/13 132 18/ 141 20/ 137 18/ 140 24/
84 11 90 10 83 10 88 13
PRECOR group 1991 43/60 4/60 42/61 7/61
Sivarajan et al 46 1982 107 13/ 126 14/ 104 13/ 124 13/
(Group B1) 83 8 79 11 82 11 78 9
Sivarajan et al 46 106 12/ 127 16/
(Group B2) 83 9 83 10
Taylor et al 49,50 1986/1988 42/97 14/32 26/45 10/26

Data are only presented for studies that reported data for smoking or blood pressure at follow-up. Duration of follow-up for each trial is described in Table I.
Cardiovascular risk factor follow-up data are generally presented for those who attended the follow-up clinic visit; consequently, these data are presented using a modified intention-
to-treat approach.
This study included 2 intervention groups (group B1, exercise-only CR; group B2, comprehensive CR including exercise) and 1 duplicate control group.
 American Heart Journal
584.e2 Lawler, Filion, and Eisenberg October 2011

Effect ofAppendix
Online exercise-based CR onofbody
3. Effect weight and
exercise-based CRtotal cholesterol
on body weightamong patients
and total post-MI
cholesterol among patients post-MI
Body weight (Mean SD)(kg) Total cholesterol (Mean SD) [mg/dL]

Exercise-based CR Control Exercise-based CR Control

Study Year Baseline Follow-up Baseline Follow-up Baseline Follow-up Baseline Follow-up

Ballantyne et al 15 1982 72 8 72 9 74 7 74 7 253 32 246 31 259 46 261 43

Bertie et al 17 1992 75 11 76 11 77 10 77 9
Bethell and Mullee 18 1990 239 50 243 39 236 43 236 39
Kallio et al 11 1979 74 11 72 11 74 12 75 12 232 46 251 46 232 50 286 54
Marra et al 36 1985 71 10 69 9 72 9 71 9 241 49 227 42 232 51 235 54
Vermeulen et al 51 1983 74 7 76 8 74 8 75 9 271 38 254 35 278 48 276 35

Data are only presented for studies that reported follow-up data for body weight or total cholesterol. Duration of follow-up for each trial is described in Table I.