Relationship between changes in heart rate recovery after

cardiac rehabilitation on cardiovascular mortality in patients

with myocardial infarction
Jo-Jo Hai, MBBS,* Chung-Wah Siu, MBBS,* Hee-Hwa Ho, MBBS,* Sheung-Wai Li, MBBS,
Stephen Lee, MBBS,* Hung-Fat Tse, MD, PhD,*
From the *Cardiology Division, Department of Medicine, Queen Mary Hospital, and Research Center of Heart, Brain,
Hormone and Healthy Aging, the University of Hong Kong, Hong Kong SAR, China, and Department of Medicine, Tung
Wah Hospital, Hong Kong, China.
BACKGROUND Heart rate recovery (HRR) at predischarge exercise
stress test predicts all-cause mortality in patients with myocardial
infarction (MI), but the relationship between improvement in HRR
with exercise training and clinical outcomes remains unclear.
OBJECTIVE The purpose of this study was to evaluate the effect of
change in HRR after exercise training on clinical outcomes in MI
patients.
METHODS The study included 386 consecutive patients with recent MI who were enrolled into our cardiac rehabilitation program.
All patients underwent symptom-limited treadmill testing at baseline and after exercise training, and were prospectively followed-up in the outpatient clinic.
RESULTS Treadmill testing revealed significant improvement in
HRR after 8 weeks of exercise training (17.5 10.0 bpm to 19.0
12.3 bpm, P .011). After follow-up of 79 41 months, 40
(10.4%) patients died of cardiac events. Multivariate Cox regression analysis revealed that diabetes (hazard ratio [HR] 2.28, 95%
confidence interval [CI] 1.015.19, P .049), statin use (HR
0.36, 95% CI 0.16 0.80, P .012), baseline resting heart rate

Introduction
The autonomic nervous system plays an important role in
regulating the cardiovascular system. Both high sympathetic and low parasympathetic tones have been shown to be
associated with increased cardiovascular mortality.1,2 Heart
rate recovery (HRR), defined as the fall in heart rate through
the first minute after exercise, is a measure of the capacity
of the cardiovascular system to reactivate the parasympathetic nervous system after exercise3,4 and has been found to
be an independent predictor of all-cause mortality in various
populations.57 Previous studies have shown that HRR measured at predischarge exercise stress test (EST) predicts

65 bpm (HR 5.37, 95% CI 1.3321.61, P .018), post-training

HRR 12 bpm (HR 2.49, 95% CI 1.10 5.63, P .028), left
ventricular ejection fraction 30% (HR 4.70, 95% CI 1.34 16.46,
P .016), and exercise capacity 4 metabolic equivalents (HR
3.63, 95% CI 1.1711.28, P .026) were independent predictors
of cardiac death. Patients who failed to improve HRR from 12
bpm to 12 bpm after exercise training had significantly higher
mortality (HR 6.2, 95% CI 1.329.2, P .022).
CONCLUSION Exercise training improved HRR in patients with
recent MI, and patients with HRR increased to 12 bpm had better
cardiac survival.
KEYWORDS Cardiac death; Cardiac rehabilitation; Exercise; Heart
rate recovery; Myocardial infarction
ABBREVIATIONS CI confidence interval; DTS Duke treadmill score; EST exercise stress test; HR hazard ratio;
HRR heart rate recovery; LVEF left ventricular ejection fraction; METs metabolic equivalents; MI myocardial infarction
(Heart Rhythm 2010;7:929 936) 2010 Published by Elsevier Inc.
on behalf of Heart Rhythm Society.

mortality among survivors of acute myocardial infarction

(MI).8
HRR appears to be modifiable by regular exercise,9 11
presumably due to the favorable effect of exercise on autonomic regulation.12,13 However, the relationship between
improvement in HRR and clinical outcomes after MI remains unclear. The aim of this study was to evaluate the
effect of exercise training on HRR in patients with recent
MI and the impact of change in HRR on their clinical
outcomes.

Methods
Study population

Address reprint requests and correspondence: Dr. Chung-Wah Siu or

Prof. Hung-Fat Tse, Cardiology Division, Department of Medicine, The
University of Hong Kong, Queen Mary Hospital, Hong Kong, China.
E-mail addresses: dcwsiu@hkucc.hku.hk; hftse@hkucc.hku.hk (Received
December 2, 2009; accepted March 18, 2010.)

This prospective cohort study consisted of consecutive patients who were enrolled into our cardiac rehabilitation
program from 1996 to 2007. All patients had confirmed
diagnoses of acute MI based on World Health Organization
criteria.14 A total of 446 patients who were able to perform

1547-5271/$ -see front matter 2010 Published by Elsevier Inc. on behalf of Heart Rhythm Society.

doi:10.1016/j.hrthm.2010.03.023

930
symptom-limited EST on treadmill, completed phase 1 and
phase 2 of the program with good attendance over 70%, no
pacemaker implantation, and no change in medications between the two ESTs were recruited. Among these patients,
60 were excluded because of loss to follow-up (n 55) or
missing data (n 5). As a result, 386 patients were included
in the analysis.

Study design
Among the 386 patients recruited, 334 underwent an 8-week
exercise training program in phase 2 of our program, which
consisted of twice-weekly 45-minute scheduled sessions of
supervised exercise at an intensity based on individual assessment. The remaining 52 patients did not receive any
exercise training because they were included into the control arm of our prior randomized study conducted between
1997 and 2002.15 Nevertheless, all 386 patients received the
same magnitude of education, counseling, and advice from
our nurses, physiotherapists, occupational therapists, dietitians, and clinical psychologists during cardiac rehabilitation.
Patients demographic data, cardiovascular risk factors,
and medications were recorded at program entry. Left ventricular ejection fraction (LVEF) was measured by transthoracic echocardiography. All patients were prospectively followed-up in our outpatient clinics or until they left the
cohort through death. Information on deaths during the
follow-up period after recruitment were retrieved from medical records and discharge summaries from our hospital as
well as other institutions. Patients who were lost to follow-up were contacted by phone. In addition, survival data
were obtained from the Births and Deaths General Register
Office. The survival rate, cause of death, and clinical characteristics of these patients were examined. Sudden cardiac
death was defined as death from an unexpected circulatory
arrest occurring within 1 hour of symptom onset.16 Cardiac
death was defined as death due to MI, intractable heart
failure, or malignant arrhythmia and included cases of sudden cardiac death.

Exercise testing
Symptom-limited ESTs were conducted in phase 1 (before
exercise training) and at the end of phase 2 (after exercise
training) using the modified Bruce protocol. Treadmill was
terminated within 10 seconds after peak exercise, and patients were allowed to assume a sitting position immediately. Resting heart rate was defined as heart rate before
EST during sitting. Peak heart rate was defined as heart rate
achieved at peak exercise. Heart rate increment was defined
as heart rate increase from baseline to peak (i.e., peak heart
rate resting heart rate). HRR was defined as heart rate
decrease through the first minute of recovery (i.e., peak
heart rate heart rate at 1 minute of recovery phase). Duke
treadmill score (DTS) was calculated using the following
formula: (Duration of exercise [min]) (5 Maximal
ST-segment deviation [mm]) (4 treadmill angina index), where index 0 no angina, 1 nonlimiting angina,

Heart Rhythm, Vol 7, No 7, July 2010

and 2 exercise-limiting angina, and each patient was
classified as high (DTS 11), intermediate (DTS 10 to
4), or low (DTS 5) risk, accordingly.1720 Patients
who had ST-segment depression on resting ECG due to
digoxin therapy, left ventricular hypertrophy, or complete
left or right bundle branch block were labeled as DTS
indeterminate.21,22 Exercise capacity was expressed in
terms of calculated metabolic equivalents (METs).

Statistical analysis
Continuous variables are expressed as mean SD, and
categorical variables are presented in frequency tables. Statistical comparisons were performed using Students t-test
or Pearson Chi-square test, as appropriate. LVEF 30%
was chosen as a predictive variable of cardiac death, in line
with the Multicenter Automatic Defibrillator Implantation
Trial II (MADIT II).23 For all other variables, cutoff points
were obtained from receiver operating characteristic curve
analyses. Hazard ratio (HR) and 95% confidence interval
(CI) of each variable to predict cardiac death were determined by multivariate Cox regression model using P .1
for inclusion. P .05 was considered significant. Statistic
analyses were performed using SPSS software (version
16.0, SPSS, Inc., Chicago, IL, USA).

Results
Table 1 lists baseline clinical demographic features of the
study population. A total of 334 (86.5%) patients received
exercise training. More patients in the exercise training
group received antiplatelets or warfarin, statins, and revascularizations before enrollment compared to patients without exercise training. Their changes in EST parameters and
LVEF during follow-up are listed in Table 2. Peak heart
rate, heart rate increment, and HRR increased significantly
after completion of phase 2 in patients who received exercise training but not in those without training. In both
groups, exercise capacity and LVEF improved significantly
over time compared to baseline. Although LVEF increased
similarly in the two groups (5.1% 11.4% vs 6.2%
14.0%, P 0.53), the increase in exercise capacity was
more profound in patients who received exercise training
(1.9 0.1 METs vs 1.4 0.2 METs, P .001). Although
there were significant changes in the prevalence of patients
in different DTS risk groups between phase 1 and phase 2 in
both patients with and those without exercise training (Table 2), exercise training did not modify the change in DTS
risk between the two phases (Figure 1).
After 78.5 40.5 months of follow-up, 40 (10.4%)
patients died of cardiac events, of which 26 (6.7%) were
sudden cardiac deaths. Baseline clinical characteristics of
patients with and those without cardiac death are listed in
Table 3. Compared with survivors, patients who suffered
from cardiac death were older, were more likely to be
diabetic, developed acute pulmonary edema on presentation, received angiotensin-converting enzyme inhibitors/angiotensin receptor blockers, and had lower initial LVEF, but
they were less likely to receive coronary revascularizations,

Hai et al
Table 1

Improved HR Recovery After Exercise Training and Clinical Outcome Post-MI

931

Baseline clinical characteristics of patients with and without exercise training

Demographics and Cardiovascular Risk Factors

Age (years)
Sex (M/F)
Race (Chinese/non-Chinese)
Smokers (active or quit 3 years)
Diabetes mellitus
Hypertension
Hypercholesterolemia
Renal impairment
Respiratory disease
Phase 1 left ventricular ejection fraction (%)
Events Occurred During Index Admissions
Acute pulmonary edema
Ventricular tachycardia or ventricular fibrillation
Treatment Received
Revascularizations
Antiplatelet/warfarin
Beta-blocker
Angiotensin-converting enzyme inhibitor/angiotensin receptor blocker
Statin

Exercise training
(n 334)

No exercise training
(n 52)

63.8 11.2
256/78
321/13
124 (37.1)
122 (36.5)
178 (53.3)
226 (67.7)
29 (8.7)
36 (10.8)
46.5 9.9

65.0 10.2
41/11
52/0
24 (46.2)
25 (48.1)
30 (57.7)
28 (53.8)
4 (7.7)
6 (11.5)
46.3 10.1

85 (25.4)
21 (6.3)
255
333
266
266
276

.40
.86
.23
1
.13
.65
.059
1
.81
.90

14 (26.9)
4 (7.7)

(76.3)
(99.7)
(79.6)
(79.6)
(82.6)

23
48
38
41
30

P value

.87
.76
.001*
.001*
.28
.86
.001*

(44.2)
(92.3)
(73.1)
(78.8)
(57.7)

Values are given as mean SD, number, or number (percent).

*P .05.
Creatinine 200 mol/L.
Respiratory disease consisted of clearly documented asthma, chronic obstructive pulmonary disease, bronchiectasis, interstitial lung disease, or pulmonary
fibrosis.

exercise training, and statins (Table 3, all P .05). However, there were no significant differences in other demographic data and cardiovascular risk factors, incidence of
malignant arrhythmias during index hospitalizations, or use
of antiplatelets/warfarin or beta-blockers between the two
groups.
Univariate analysis demonstrated that diabetes mellitus,
acute pulmonary edema on presentation, phase 1 or phase 2
LVEF 30%, and use of angiotensin-converting enzyme
inhibitors/angiotensin receptor blockers were significantly
associated with cardiac death, whereas use of statins, revas-

Table 2

cularizations before enrollment, and exercise training were

protective against cardiac death (Table 4, all P .05). Based
on the results of the receiver operating characteristic curve
analyses in defining cutoff values (Table 5), several EST
parameters, including phase 1 resting heart rate65 bpm,
heart rate increment 45 bpm, and exercise capacity 4
METs, and phase 2 heart rate increment 45 bpm, HRR
12 bpm, and exercise capacity 4 METs were found to be
predictors of cardiac death (Table 4, all P .05). In contrast,
risk assessment by DTS was not predictive of cardiac death
(Table 4, P .05).

Exercise stress test parameters and mean left ventricular ejection fraction
With exercise training (n 334)
Phase 1

Resting heart rate (bpm)

Peak heart rate (bpm)
Heart rate increment (bpm)
Heart rate recovery (bpm)
Exercise capacity (METs)
Left ventricular ejection fraction (%)
Duke treadmill score (risk group)
Low risk
Intermediate risk
High risk
Indeterminate risk

71.2
121.7
50.5
17.5
5.4
46.5
187
91
3
53

Phase 2
13.1
22.8
20.8
10.0
3.2
9.9

(56.0)
(27.2)
(0.9)
(15.9)

Values are given as mean SD or number (percent).

METs metabolic equivalents.
*P .05.

71.9
125.4
53.4
19.0
7.2
52.1
167
94
7
66

12.8
23.3
22.0
12.3
3.4
13.4

(50.0)
(28.1)
(2.1)
(19.8)

Without exercise training (n 52)

P value

Phase 1

.26
.001*
.006*
.011*
.001*
.001*

71.3
118.9
47.6
19.1
4.3
46.3

.001*
.001*
.001*
.001*

20
22
2
8

Phase 2
16.3
25.6
23.2
14.4
2.7
10.1

(38.5)
(42.3)
(3.8)
(15.4)

72.3
124.4
52.2
20.9
5.1
53.2
22
18
3
9

P value
15.6
21.7
21.4
15.1
3.1
12.1

(42.3)
(34.6)
(5.8)
(17.3)

.61
.068
.060
.37
.001*
.003*
.001*
.001*
.001*
.10

932

Heart Rhythm, Vol 7, No 7, July 2010

Figure 1
Change in Duke treadmill risk score between phase 1 and
phase 2 with reference to exercise training. Improved risk improved
from higher-risk to lower-risk group; same risk no change in risk group;
worsened risk worsened from lower-risk to higher-risk group.

Multivariate analysis revealed that diabetes mellitus (HR

2.28, 95% CI 1.015.19, P .049) and use of statins (HR
0.36, 95% CI 0.16 0.80, P .012) were independent
predictors of cardiac death. Whereas none of the phase 1
variables except resting HR65 bpm (HR 5.37, 95% CI
1.3321.61, P .018) significantly predicted mortality,
phase 2 parameters including LVEF 30% (HR 4.70,
95% CI 1.34 16.46, P .016), HRR 12 bpm (HR 2.49,
95% CI 1.10 5.63, P .028), and exercise capacity 4

Table 3

METs (HR 3.63, 95% CI 1.1711.28, P .026) remained

to be significant independent predictors of cardiac death
(Table 5).
Comparison of patients characteristics in relation to
phase 2 HRR is given in Table 6. Patients with phase 2
HRR 12 bpm were older and more often had diabetes
mellitus, hypertension, respiratory disease, and acute pulmonary edema on presentation. However, fewer of them
received beta-blockers or had undergone revascularization (Table 6, all P .05). They had a lower HRR at
phase 1, achieved a lower heart rate increment and exercise capacity at both phases, and showed a significantly
lower LVEF and higher resting heart rate at the end of
phase 2 (Table 6, all P .05). Compared with patients who
had phase 2 HRR 12 bpm, fewer patients with phase 2 HRR
12 bpm belonged to the low-risk group as assessed by DTS,
whereas more of these patients belonged to the intermediaterisk group (Table 6, both P .05). Nevertheless, the proportions of high-risk patients between the two groups were not
significantly different.
In this study, a cutoff value 12 bpm obtained from
receiver operating characteristic curve analysis was used to
define impaired HRR. This value has also been shown to be
associated with increased mortality at predischarge EST
after MI.8 Further analysis was performed to investigate the
effect of improved HRR on survival in this group of patients. Of the 109 patients identified, 56 (51.4%) had phase
2 HRR improved to 12 bpm. Patients whose HRR remained 12 bpm at the end of phase 2 had much worse
survival than did those who improved (HR 6.15, 95% CI
1.30 29.15, P .022; Figure 2).

Baseline clinical characteristics of patients with and without cardiac death

Demographics and Cardiovascular Risk Factors

Age (years)
Sex (M/F)
Race (Chinese/non-Chinese)
Smokers (active or quit 3 years)
Diabetes mellitus
Hypertension
Hypercholesterolemia
Renal impairment
Respiratory disease
Phase 1 left ventricular ejection fraction
Events Occurred During Index Admissions
Acute pulmonary edema
Ventricular tachycardia or ventricular fibrillation
Treatment Received
Revascularizations*
Antiplatelet/warfarin
Beta-blocker
Angiotensin-converting enzyme inhibitor/angiotensin receptor blocker
Statin
Exercise training
Values are given as mean SD, number, or number (percent).
*P .05.

Cardiac death
(n 40)

No cardiac death
(n 346)

66.8 12.0
29/11
39/1
15 (37.5)
23 (55.0)
22 (55.0)
23 (57.5)
4 (10.0)
6 (15.0)
43.0 10.3

63.6 11.0
268/78
334/12
133 (38.4)
124 (35.8)
186 (53.8)
231 (66.8)
29 (8.4)
36 (10.4)
46.9 9.8

19 (47.5)
4 (10.0)
16
38
27
37
21
26

(40.0)
(95.0)
(67.5)
(92.5)
(52.5)
(65.0)

80 (23.1)
21 (6.1)
262
343
277
270
285
308

(75.7)
(99.1)
(80.1)
(78.0)
(82.4)
(89.0)

P value
.019*
.55
1
.26
.01*
1
.29
.76
.42
.028*
.002*
.31
0.001*
.086
.10
.037*
.001*
.001*

Hai et al
Table 4

Improved HR Recovery After Exercise Training and Clinical Outcome Post-MI

933

Predictors of cardiac death by Cox analysis for all patients

Univariate predictors

Age
Sex
Smoker
Diabetes mellitus
Hypertension
Hypercholesterolemia
Renal impairment
Respiratory disease
Phase 1 LVEF 30%
Phase 2 LVEF 30%
Acute pulmonary edema
Ventricular tachycardia/ventricular fibrillation
Revascularizations
Antiplatelet/warfarin
Beta-blocker
Angiotensin-converting enzyme inhibitor/
angiotensin receptor blocker
Statin
Exercise training
Phase 1 resting heart rate 65 bpm
Phase 1 heart rate increment 45 bpm
Phase 1 heart rate recovery 12 bpm
Phase 1 exercise capacity 4 METs
Phase 1 Duke treadmill score (risk group)
Low risk
Intermediate risk
High risk
Phase 2 heart rate increment 45 bpm
Phase 2 heart rate recovery 12 bpm
Phase 2 exercise capacity 4 METs
Phase 2 Duke treadmill score (risk group)
Low risk
Intermediate risk
High risk

Multivariate predictors

HR

95% CI

P value

1.03
0.86
0.59
2.64
0.94
1.00
0.83
1.19
2.75
4.10
3.40
1.91
0.45
0.31
0.61
3.92

1.001.06
0.431.72
0.271.29
1.414.95
0.501.75
0.531.88
0.322.51
0.502.84
1.265.97
1.6710.06
1.836.34
0.685.37
0.240.88
0.071.28
0.311.18
1.2112.71

.10
.66
.19
.002*
.84
.99
.89
.70
.011*
.002*
.001*
.22
.018*
.10
.14
.023*

0.40
0.43
3.04
2.25
1.56
2.90

0.210.75
0.220.83
1.287.26
1.164.36
0.822.96
1.386.10

0.77
1.42
0.47
2.76
4.00
6.51

0.40 1.50
0.732.77
0.00 161.33
1.455.25
2.147.48
3.1813.33

0.53
2.10
0.50

0.26 1.09
1.01 4.36
0.073.67

.004*
.011*
.012*
.017*
.18
.005*
.44
.30
.46
.002*
.001*
.001*

HR

95% CI

P value

2.28

1.015.19

.049*

1.99
4.70
1.76

0.745.32
1.3416.46
0.803.89

.17
.016*
.160

1.08

0.402.92

.877

2.80

0.6312.45

.177

0.36
0.59
5.37
1.79

0.160.80
0.221.11
1.3321.61
0.565.70

.012*
.088
.018*
.325

0.49

0.141.71

.261

0.80
2.49
3.63

0.302.08
1.105.63
1.1711.28

.641
.028*
.026*

.11
.10
.49

CI confidence interval; HR hazard ratio; LVEF left ventricular ejection fraction; METs metabolic equivalents.
*P .05.

Discussion
Our results demonstrated exercise training improved HRR,
LVEF, and exercise capacity in patients with recent MI. In
this study, none of the parameters during baseline phase 1
EST, except resting heart rate, predicted cardiac death during long-term follow-up. In contrast, left ventricular dysfunction with LVEF 30%, impaired HRR 12 bpm, and

Table 5

Receiver operating characteristic curve analysis

Parameter
Phase
Phase
Phase
Phase
Phase
Phase

1
1
1
2
2
2

exercise capacity 4 METs at phase 2 after training were

shown to be independent predictors of cardiac death. Furthermore, persistently impaired HRR 12 bpm at phase 2
was found to have incremental prognostic value over traditional markers, including diabetes mellitus, resting heart
rate, LVEF, and exercise capacity.24 28 Of interest, patients
with phase 1 HRR 12 bpm had a much better prognosis if

resting heart rate 65 bpm

heart rate increment 45 bpm
exercise capacity 4 METs
heart rate increment 45 bpm
heart rate recovery 12 bpm
exercise capacity 4 METs

AUC

95% CI

P value

Sensitivity (%)

Specificity (%)

PPV (%)

NPV (%)

0.62
0.66
0.66
0.66
0.66
0.75

0.570.67
0.610.70
0.610.71
0.610.71
0.610.71
0.700.79

.0144*
.0002*
.0001*
.0001*
.0001*
.0001*

80.0
67.5
77.5
62.5
52.5
75.0

36.71
54.9
50.0
62.1
77.8
72.0

12.7
14.8
15.2
16.0
21.4
23.6

94.1
93.6
95.1
93.5
93.4
96.1

AUC area under the curve; CI confidence interval; METs metabolic equivalents; NPV negative predictive value; PPV positive predictive value.
*P .05.

934
Table 6

Heart Rhythm, Vol 7, No 7, July 2010

Comparison of patients characteristics in relation to phase 2 heart rate recovery

Demographics and Cardiovascular Risk Factors

Age (years)
Sex (M/F)
Race (Chinese/non-Chinese)
Smokers (active or quit 3 years)
Diabetes mellitus
Hypertension
Hypercholesterolemia
Renal impairment
Respiratory disease
Phase 1 LVEF (%)
Phase 2 LVEF (%)
Events Occurred During Index Admissions
Acute pulmonary edema
Ventricular tachycardia/ventricular fibrillation
Treatment Received
Revascularizations
Antiplatelet/warfarin
Beta-blocker
Angiotensin-converting enzyme inhibitor/angiotensin receptor blocker
Statin
Exercise training
Exercise Stress Test Parameters
Phase 1 resting heart rate (bpm)
Phase 1 heart rate increment (bpm)
Phase 1 exercise capacity (METs)
Phase 1 HRR (bpm)
Phase 1 Duke treadmill score (risk group)
Low risk
Intermediate risk
High risk
Phase 2 resting heart rate (bpm)
Phase 2 heart rate increment (bpm)
Phase 2 exercise capacity (METs)
Phase 2 Duke treadmill score (risk group)
Low risk
Intermediate risk
High risk

HRR 12 bpm
(n 99)

HRR 12 bpm
(n 287)

P value

68.4 9.1
73/26
96/3
39 (39.4)
48 (48.5)
64 (64.6)
61 (61.6)
13 (13.1)
18 (18.2)
46.1 9.8
49.9 12.5

62.4 11.3
224/63
277/10
109 (38.0)
99 (34.5)
144 (50.2)
193 (67.2)
20 (7.0)
24 (8.4)
46.6 9.9
53.1 13.4

.001*
.38
.83
.08
.013*
.013*
.31
.059
.007*
.67
.047*

37 (37.4)
7 (7.1)

62 (21.6)
18 (6.3)

65
96
70
80
73
249

(65.7)
(97.0)
(70.7)
(80.8)
(73.7)
(86.8)

213
285
234
227
233
85

(74.2)
(99.3)
(81.5)
(79.1)
(81.2)
(85.9)

73.5
40.3
3.6
12.2

70.4
53.5
5.8
19.6

15.0
18.1
2.1
9.6

43 (43.4)
39 (39.4)
0 (0)
75.2 14.9
38.2 15.4
5.1 2.6
33 (33.3)
36 (36.4)
2 (2.0)

.002*
.78
.014*
.08
.023*
.72
.12
.82

12.9
21.0
3.2
10.3

.076
.001*
.001*
.001*

164 (57.1)
74 (25.8)
5 (1.7)
70.9 12.4
58.5 21.4
7.5 3.5

.014*
.005*
.19
.005*
.001*
.001*

156 (54.4)
76 (26.5)
8 (2.8)

.005*
.003*
.82

Values are given as mean SD, number, or number (percent).

HRR heart rate recovery; LVEF left ventricular ejection fraction; METs metabolic equivalents.
*P .05.
Creatinine 200 mol/L.
Respiratory disease consisted of clearly documented asthma, chronic obstructive pulmonary disease, bronchiectasis, interstitial lung disease, or pulmonary
fibrosis.

HRR improved to 12 bpm upon completion of phase 2 of

cardiac rehabilitation. These findings suggest that improvement of HRR after exercise training may contribute
to the long-term beneficial effects of cardiac rehabilitation after MI.
Following MI, there is an increase in sympathetic outflow to the heart and peripheral vasculature and a decrease
in cardiac vagal outflow. In addition, plasma norepinephrine
level rises due to increased spillover and reduced clearance.4 Although these alterations are important for maintaining blood pressure in the acute state, patients who fail to
reverse these changes may suffer from long-term adverse
consequences, such as elevated myocardial oxygen consumption, unfavorable cardiac remodeling, and increased

propensity to malignant arrhythmia and therefore would be

at higher risk for cardiac death.4
Randomized trials have shown that cardiac rehabilitation
after MI decreased long-term all-cause and cardiovascular
mortality.29 Possible mechanisms include better cardiovascular risk factors control, physical strength, psychological
well-being, cardiac remodeling, and sympathovagal balance. As a reflection of autonomic balance, improved HRR
after rehabilitation can be regarded as a sign of restoration
of cardiovascular autonomic control. To our knowledge, no
prior study has addressed the potential prognostic value of
post-rehabilitation HRR on cardiac death among patients
with prior MI. Our results provide evidence supporting
modulation of sympathovagal tone as underlying an impor-

Hai et al

Improved HR Recovery After Exercise Training and Clinical Outcome Post-MI

Figure 2

935

Kaplan-Meier survival curves for heart rate recovery. P1 phase 1; P2 phase 2.

tant mechanism by which exercise training decreases mortality, as patients who had restoration of autonomic control
after exercise training, as evidenced by HRR improved to
12 bpm, had better survival during long-term follow-up.
Assessment of autonomic control of the cardiovascular
system after MI provides completely distinct information
from conventional prognostic markers.4 Several different
indexes have been used to assess cardiac autonomic function.4 Although widely studied, parameters such as heart
rate variability and baroreflex sensitivity are of limited clinical value due to their intrinsic methodologic and interpretive difficulties.30,31 In contrast, ESTs are routinely performed for risk stratification after MI, and simple autonomic
parameters obtained from ESTs are easy to interpret. Furthermore, unlike chronotropic index and heart rate increment, which also depend on myocardial perfusion and physical strength, HRR is a more precise measurement of the
cardiovascular autonomic balance.32 In particular, our study
failed to demonstrate any correlation between heart rate
increment and cardiac death, in contrast to findings from
previous studies on healthy volunteers.32,33 The reason for
this discrepancy remains unclear but can be explained partly
by the fact that 80% of our patients were treated with
beta-blockers, which blunted heart rate response during
exercise but either enhanced or had no effect on HRR.34,35
DTS failed to predict survival for our patients, which is
in contrast with findings of previous validation studies of
MI populations.17,19 Although the reason remains unclear, it
is possible that a significant proportion of our patients had
indeterminate DTS due to baseline abnormal ECG, which
reduces the power of DTS to predict clinical outcomes in
this study. On the other hand, HRR appears to be superior to

DTS as a predictor of clinical outcome after MI, being more

sensitive and unaffected by preexisting ECG abnormalities.
As shown in our study, patients who had impaired HRR
at baseline seem to benefit more from cardiac rehabilitation,
as those with improved HRR after exercise training had
significantly lower cardiac mortality. Whether therapeutic
strategies such as titration of medications and exercise prescription that target to HRR can further optimize autonomic
balance and improve clinical outcomes in MI patients needs
to be addressed in future clinical studies.

Study limitations
First, our study recruited patients who completed phase 1
and phase 2 of the cardiac rehabilitation program with good
attendance and were able to perform symptom-limited EST
on treadmill. As a result, high-risk populations of patients
who could not complete the program due to repeated hospitalizations or could not perform treadmill tests due to poor
cardiovascular function were excluded. Indeed, the cardiac
death rate was 21% and sudden cardiac death rate was
12.3% among patients excluded (n 211), which were
much higher than that in the population studied. Therefore,
generalization of our results to this group of patients should
be made with caution. Second, this study was not designed
to evaluate the benefit of exercise training. Although this
study included patients who were randomized to the control
arm in our previous trial, the nonrandomized nature of the
present study and the severe imbalance between treatment
and control groups preclude analysis of the effect of exercise training. Nevertheless, previous randomized studies
have already proved that HRR improves with exercise training,9,11 and our study only addressed the effect of changes

936
in HRR on survival. Third, although this was a prospective
observational study, the endpoints were not prespecified but
were derived from post hoc analysis. Finally, only data on
HRR at 1 minute were prospectively collected for all patients in this study; whether HRR measured at 2 minutes or
later would provide better prognostic information remains
unclear.

Heart Rhythm, Vol 7, No 7, July 2010

16.

17.

Conclusion
Our study demonstrated that exercise training improved
HRR and that patients with persistently impaired HRR 12
bpm after rehabilitation had higher cardiac mortality during
long-term follow-up, suggesting that modulation of cardiovascular autonomic control with exercise training may contribute to the long-term beneficial effects of a cardiac rehabilitation program. Its clinical application as a therapeutic
target for exercise prescription and medication titration
needs to be further addressed in future studies.

References
1. Curtis BM, OKeefe JH Jr. Autonomic tone as a cardiovascular risk factor: the
dangers of chronic fight or flight. Mayo Clin Proc 2002;77:4554.
2. Schwartz PJ, La Rovere MT, Vanoli E. Autonomic nervous system and sudden
cardiac death. Experimental basis and clinical observations for post-myocardial
infarction risk stratification. Circulation 1992;85(1 Suppl):I77I91.
3. Freeman JV, Dewey FE, Hadley DM, Myers J, Froelicher VF. Autonomic
nervous system interaction with the cardiovascular system during exercise. Prog
Cardiovasc Dis 2006;48:342362.
4. Frenneaux MP. Autonomic changes in patients with heart failure and in postmyocardial infarction patients. Heart 2004;90:1248 1255.
5. Nishime EO, Cole CR, Blackstone EH, Pashkow FJ, Lauer MS. Heart rate
recovery and treadmill exercise score as predictors of mortality in patients
referred for exercise ECG. JAMA 2000;284:13921398.
6. Cole CR, Foody JM, Blackstone EH, Lauer MS. Heart rate recovery after
submaximal exercise testing as a predictor of mortality in a cardiovascularly
healthy cohort. Ann Intern Med 2000;132:552555.
7. Cole CR, Blackstone EH, Pashkow FJ, Snader CE, Lauer MS. Heart-rate
recovery immediately after exercise as a predictor of mortality. N Engl J Med
1999;341:13511357.
8. Nissinen SI, Makikallio TH, Seppanen T, et al. Heart rate recovery after exercise
as a predictor of mortality among survivors of acute myocardial infarction. Am J
Cardiol 2003;91:711714.
9. Tiukinhoy S, Beohar N, Hsie M. Improvement in heart rate recovery after
cardiac rehabilitation. J Cardiopulm Rehabil 2003;23:84 87.
10. Adams BJ, Carr JG, Ozonoff A, Lauer MS, Balady GJ. Effect of exercise
training in supervised cardiac rehabilitation programs on prognostic variables
from the exercise tolerance test. Am J Cardiol 2008;101:14031407.
11. Streuber SD, Amsterdam EA, Stebbins CL. Heart rate recovery in heart failure
patients after a 12-week cardiac rehabilitation program. Am J Cardiol 2006;97:
694 698.
12. Malfatto G, Facchini M, Sala L, Branzi G, Bragato R, Leonetti G. Relationship
between baseline sympatho-vagal balance and the autonomic response to cardiac
rehabilitation after a first uncomplicated myocardial infarction. Ital Heart J
2000;1:226 232.
13. Lucini D, Milani RV, Costantino G, Lavie CJ, Porta A, Pagani M. Effects of
cardiac rehabilitation and exercise training on autonomic regulation in patients
with coronary artery disease. Am Heart J 2002;143:977983.
14. Nomenclature and criteria for diagnosis of ischemic heart disease. Report of the
Joint International Society and Federation of Cardiology/World Health Organization task force on standardization of clinical nomenclature. Circulation 1979;
59:607 609.
15. Yu CM, Lau CP, Chau J, et al. A short course of cardiac rehabilitation program
is highly cost effective in improving long-term quality of life in patients with

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.
28.

29.

30.

31.

32.

33.

34.

35.

recent myocardial infarction or percutaneous coronary intervention. Arch Phys

Med Rehabil 2004;85:19151922.
Zipes DP, Camm AJ, Borggrefe M, et al. ACC/AHA/ESC 2006 guidelines for
management of patients with ventricular arrhythmias and the prevention of
sudden cardiac death: a report of the American College of Cardiology/American
Heart Association Task Force and the European Society of Cardiology Committee for Practice Guidelines. J Am Coll Cardiol 2006;48:e247 e346.
Villella M, Villella A, Santoro L, Santoro E, Franzosi MG, Maggioni AP; Duke
Treadmill Score, Veterans Administration Medical Center Score, GISSI-2 Investigators. Ergometric score systems after myocardial infarction: prognostic
performance of the Duke Treadmill Score, Veterans Administration Medical
Center Score, and of a novel score system, GISSI-2 Index, in a cohort of
survivors of acute myocardial infarction. Am Heart J 2003;145:475 483.
Mark DB, Hlatky MA, Harrell FE, Jr., Lee KL, Califf RM, Pryor DB. Exercise
treadmill score for predicting prognosis in coronary artery disease. Ann Intern
Med 1987;106:793 800.
Valeur N, Clemmensen P, Grande P, Saunamaki K; DANAMI-2 Investigators.
Prognostic evaluation by clinical exercise test scores in patients treated with
primary percutaneous coronary intervention or fibrinolysis for acute myocardial
infarction (a Danish Trial in Acute Myocardial Infarction-2 Sub-Study). Am J
Cardiol 2007;100:1074 1080.
Valeur N, Clemmensen P, Saunamaki K, Grande P; DANAMI-2 Investigators.
The prognostic value of pre-discharge exercise testing after myocardial infarction treated with either primary PCI or fibrinolysis: a DANAMI-2 sub-study. Eur
Heart J 2005;26:119 127.
Kwok JM, Miller TD, Christian TF, Hodge DO, Gibbons RJ. Prognostic value
of a treadmill exercise score in symptomatic patients with nonspecific ST-T
abnormalities on resting ECG. JAMA 1999;282:10471053.
Villella A. The prognostic value of an ergometric scoring system in symptomatic
patients with nonspecific ST-T anomalies in the resting ECG. Ital Heart J Suppl
2000;1:573574.
Moss AJ, Zareba W, Hall WJ, et al. Prophylactic implantation of a defibrillator
in patients with myocardial infarction and reduced ejection fraction. N Engl
J Med 2002;346:877 883.
Vanhees L, Fagard R, Thijs L, Amery A. Prognostic value of training-induced
change in peak exercise capacity in patients with myocardial infarcts and
patients with coronary bypass surgery. Am J Cardiol 1995;76:1014 1019.
Feuerstadt P, Chai A, Kligfield P. Submaximal effort tolerance as a predictor of
all-cause mortality in patients undergoing cardiac rehabilitation. Clin Cardiol
2007;30:234 238.
Specchia G, De Servi S, Scire A, et al. Interaction between exercise training and
ejection fraction in predicting prognosis after a first myocardial infarction.
Circulation 1996;94:978 982.
Fox K, Borer JS, Camm AJ, et al. Resting heart rate in cardiovascular disease.
J Am Coll Cardiol 2007;50:823 830.
Mukamal KJ, Nesto RW, Cohen MC, et al. Impact of diabetes on long-term
survival after acute myocardial infarction: comparability of risk with prior
myocardial infarction. Diabetes Care 2001;24:14221427.
OConnor GT, Buring JE, Yusuf S, et al. An overview of randomized trials of
rehabilitation with exercise after myocardial infarction. Circulation 1989;80:
234 244.
Bigger JT Jr, La Rovere MT, Steinman RC, et al. Comparison of baroreflex
sensitivity and heart period variability after myocardial infarction. J Am Coll
Cardiol 1989;14:15111518.
Kleiger RE, Miller JP, Bigger JT Jr, Moss AJ. Decreased heart rate variability
and its association with increased mortality after acute myocardial infarction.
Am J Cardiol 1987;59:256 262.
Myers J, Tan SY, Abella J, Aleti V, Froelicher VF. Comparison of the chronotropic response to exercise and heart rate recovery in predicting cardiovascular
mortality. Eur J Cardiovasc Prev Rehabil 2007;14:215221.
Jouven X, Empana JP, Schwartz PJ, Desnos M, Courbon D, Ducimetiere P.
Heart-rate profile during exercise as a predictor of sudden death. N Engl J Med
2005;352:19511958.
Pavia L, Myers J, Rusconi C. Effect of beta-blockade on heart rate and VO2
kinetics during recovery in patients with coronary artery disease. Heart Drug
2002;2:69 74.
Racine N, Blanchet M, Ducharme A, et al. Decreased heart rate recovery after
exercise in patients with congestive heart failure: effect of beta-blocker therapy.
J Card Fail 2003;9:296 302.