Relationship between changes in heart rate recovery afterca ascwJo wa HSt

Fr een MHo ongWa

BAstrinfwit

OBchange in HRR after exercise training on clinical outcomes in MIpat

MEcenAlllinlow

RESHR12.(10siocon0.3

6HRvenP 3.6ofbpm to 12 bpm after exercise training had significantly higher

COreccar

KErat

ABmilHRtio

(Heon

InThregtheass

rate recovery (HRR), defined as the fall in heart rate throughthe first minute after exercise, is a measure of the capacityofthebeposur

mo

(M

preno

improvement in HRR and clinical outcomes after MI re-mains unclear. The aim of this study was to evaluate the

AdProUnE-mDec

154the cardiovascular system to reactivate the parasympa-tic nervous system after exercise3,4 and has been found toan independent predictor of all-cause mortality in variouspulations.57 Previous studies have shown that HRR mea-ed at predischarge exercise stress test (EST) predicts

effect of exercise training on HRR in patients with recentMI and the impact of change in HRR on their clinicaloutcomes.

MethodsStudy populationThis prospective cohort study consisted of consecutive pa-tients who were enrolled into our cardiac rehabilitationprogram from 1996 to 2007. All patients had confirmeddiagnoses of acute MI based on World Health Organizationcriteria.14 A total of 446 patients who were able to perform

dress reprint requests and correspondence: Dr. Chung-Wah Siu orf. Hung-Fat Tse, Cardiology Division, Department of Medicine, Theiversity of Hong Kong, Queen Mary Hospital, Hong Kong, China.ail addresses: dcwsiu@hkucc.hku.hk; hftse@hkucc.hku.hk (Receivedember 2, 2009; accepted March 18, 2010.)THODS The study included 386 consecutive patients with re-t MI who were enrolled into our cardiac rehabilitation program.patients underwent symptom-limited treadmill testing at base-e and after exercise training, and were prospectively fol-ed-up in the outpatient clinic.

ULTS Treadmill testing revealed significant improvement inR after 8 weeks of exercise training (17.5 10.0 bpm to 19.03 bpm, P .011). After follow-up of 79 41 months, 40.4%) patients died of cardiac events. Multivariate Cox regres-n analysis revealed that diabetes (hazard ratio [HR] 2.28, 95%fidence interval [CI] 1.015.19, P .049), statin use (HR6, 95% CI 0.160.80, P .012), baseline resting heart rate

troductione autonomic nervous system plays an important role inulating the cardiovascular system. Both high sympa-tic and low parasympathetic tones have been shown to beociated with increased cardiovascular mortality.1,2 Heart7-5271/$ -see front matter 2010 Published by Elsevier Inc. on behalf of HearNCLUSION Exercise training improved HRR in patients withent MI, and patients with HRR increased to12 bpm had betterdiac survival.

YWORDS Cardiac death; Cardiac rehabilitation; Exercise; Hearte recovery; Myocardial infarction

BREVIATIONS CI confidence interval; DTS Duke tread-l score; EST exercise stress test; HR hazard ratio;R heart rate recovery; LVEF left ventricular ejection frac-n; METs metabolic equivalents; MI myocardial infarction

art Rhythm 2010;7:929936) 2010 Published by Elsevier Inc.behalf of Heart Rhythm Society.

rtality among survivors of acute myocardial infarctionI).8HRR appears to be modifiable by regular exercise,911sumably due to the favorable effect of exercise on auto-

mic regulation.12,13 However, the relationship betweenients. mortality (HR 6.2, 95% CI 1.329.2, P .022).rdiac rehabilitation on cardiovith myocardial infarction-Jo Hai, MBBS,* Chung-Wah Siu, MBBS,* Hee-Hephen Lee, MBBS,* Hung-Fat Tse, MD, PhD,*

om the *Cardiology Division, Department of Medicine, Qurmone and Healthy Aging, the University of Hong Kong, Hh Hospital, Hong Kong, China.

CKGROUND Heart rate recovery (HRR) at predischarge exerciseess test predicts all-cause mortality in patients with myocardialarction (MI), but the relationship between improvement in HRRh exercise training and clinical outcomes remains unclear.

JECTIVE The purpose of this study was to evaluate the effect ofular mortality in patients

o, MBBS,* Sheung-Wai Li, MBBS,

ary Hospital, and Research Center of Heart, Brain,Kong SAR, China, and Department of Medicine, Tung

5 bpm (HR 5.37, 95% CI 1.3321.61, P .018), post-trainingR 12 bpm (HR 2.49, 95% CI 1.105.63, P .028), lefttricular ejection fraction 30% (HR 4.70, 95% CI 1.3416.46,

.016), and exercise capacity 4 metabolic equivalents (HR3, 95% CI 1.1711.28, P .026) were independent predictorscardiac death. Patients who failed to improve HRR from 12t Rhythm Society. doi:10.1016/j.hrthm.2010.03.023

symphpatw60miin

StAmex

co

supses

ex

tro19sam

ou

tiatio

an

tricrac

lowco

folicawe

lowwe

Ofactdearr

defaide

ExSyex

tratertieateESachas

ratdehetreforSTde

an

cla4whdiglefindter

StCocattisor

wa

wiTrwe

an

(Cmiforan

16

ReTastuex

grocu

ou

LVrataftcisgroov

sim14mo

(1.theinbobleris

pasudpaTafrodiatiogio

930 Heart Rhythm, Vol 7, No 7, July 2010ptom-limited EST on treadmill, completed phase 1 andase 2 of the program with good attendance over 70%, nocemaker implantation, and no change in medications be-een the two ESTs were recruited. Among these patients,were excluded because of loss to follow-up (n 55) orssing data (n 5). As a result, 386 patients were includedthe analysis.

udy designong the 386 patients recruited, 334 underwent an 8-week

ercise training program in phase 2 of our program, whichnsisted of twice-weekly 45-minute scheduled sessions ofervised exercise at an intensity based on individual as-sment. The remaining 52 patients did not receive any

ercise training because they were included into the con-l arm of our prior randomized study conducted between97 and 2002.15 Nevertheless, all 386 patients received the

e magnitude of education, counseling, and advice fromr nurses, physiotherapists, occupational therapists, dieti-ns, and clinical psychologists during cardiac rehabilita-n.Patients demographic data, cardiovascular risk factors,

d medications were recorded at program entry. Left ven-ular ejection fraction (LVEF) was measured by transtho-ic echocardiography. All patients were prospectively fol-ed-up in our outpatient clinics or until they left the

hort through death. Information on deaths during thelow-up period after recruitment were retrieved from med-l records and discharge summaries from our hospital asll as other institutions. Patients who were lost to fol--up were contacted by phone. In addition, survival data

re obtained from the Births and Deaths General Registerfice. The survival rate, cause of death, and clinical char-eristics of these patients were examined. Sudden cardiacath was defined as death from an unexpected circulatoryest occurring within 1 hour of symptom onset.16 Cardiacath was defined as death due to MI, intractable heartlure, or malignant arrhythmia and included cases of sud-n cardiac death.

ercise testingmptom-limited ESTs were conducted in phase 1 (beforeercise training) and at the end of phase 2 (after exerciseining) using the modified Bruce protocol. Treadmill wasminated within 10 seconds after peak exercise, and pa-nts were allowed to assume a sitting position immedi-ly. Resting heart rate was defined as heart rate beforeT during sitting. Peak heart rate was defined as heart rateieved at peak exercise. Heart rate increment was defined

heart rate increase from baseline to peak (i.e., peak hearte resting heart rate). HRR was defined as heart ratecrease through the first minute of recovery (i.e., peakart rate heart rate at 1 minute of recovery phase). Dukeadmill score (DTS) was calculated using the followingmula: (Duration of exercise [min]) (5 Maximal-segment deviation [mm]) (4 treadmill angina in-

x), where index 0 no angina, 1 nonlimiting angina, thed 2 exercise-limiting angina, and each patient wasssified as high (DTS 11), intermediate (DTS 10 to), or low (DTS 5) risk, accordingly.1720 Patientso had ST-segment depression on resting ECG due tooxin therapy, left ventricular hypertrophy, or completet or right bundle branch block were labeled as DTSeterminate.21,22 Exercise capacity was expressed inms of calculated metabolic equivalents (METs).

atistical analysisntinuous variables are expressed as mean SD, andegorical variables are presented in frequency tables. Sta-tical comparisons were performed using Students t-testPearson Chi-square test, as appropriate. LVEF 30%s chosen as a predictive variable of cardiac death, in lineth the Multicenter Automatic Defibrillator Implantationial II (MADIT II).23 For all other variables, cutoff pointsre obtained from receiver operating characteristic curvealyses. Hazard ratio (HR) and 95% confidence intervalI) of each variable to predict cardiac death were deter-ned by multivariate Cox regression model using P .1inclusion. P .05 was considered significant. Statistic

alyses were performed using SPSS software (version.0, SPSS, Inc., Chicago, IL, USA).

sultsble 1 lists baseline clinical demographic features of thedy population. A total of 334 (86.5%) patients receivedercise training. More patients in the exercise trainingup received antiplatelets or warfarin, statins, and revas-

larizations before enrollment compared to patients with-t exercise training. Their changes in EST parameters andEF during follow-up are listed in Table 2. Peak heart

e, heart rate increment, and HRR increased significantlyer completion of phase 2 in patients who received exer-e training but not in those without training. In bothups, exercise capacity and LVEF improved significantly

er time compared to baseline. Although LVEF increasedilarly in the two groups (5.1% 11.4% vs 6.2%

.0%, P 0.53), the increase in exercise capacity wasre profound in patients who received exercise training9 0.1 METs vs 1.4 0.2 METs, P .001). Althoughre were significant changes in the prevalence of patientsdifferent DTS risk groups between phase 1 and phase 2 inth patients with and those without exercise training (Ta-

2), exercise training did not modify the change in DTSk between the two phases (Figure 1).After 78.5 40.5 months of follow-up, 40 (10.4%)

tients died of cardiac events, of which 26 (6.7%) wereden cardiac deaths. Baseline clinical characteristics of

tients with and those without cardiac death are listed inble 3. Compared with survivors, patients who sufferedm cardiac death were older, were more likely to bebetic, developed acute pulmonary edema on presenta-n, received angiotensin-converting enzyme inhibitors/an-tensin receptor blockers, and had lower initial LVEF, but

y were less likely to receive coronary revascularizations,

ex

ev

grama

ofgro

acu

LVinhass

cu

proon

an

Tab

DeAgSexRacSmDiaHyHyRenResPhaEveAcuVenTreRevAnBetAn erSta

*PCrRe tive pufibr

Tab

ResPeaHeHeExeLefDu

LIHI

*P

931Hai et al Improved HR Recovery After Exercise Training and Clinical Outcome Post-MIercise training, and statins (Table 3, all P .05). How-er, there were no significant differences in other demo-phic data and cardiovascular risk factors, incidence oflignant arrhythmias during index hospitalizations, or useantiplatelets/warfarin or beta-blockers between the twoups.Univariate analysis demonstrated that diabetes mellitus,te pulmonary edema on presentation, phase 1 or phase 2EF 30%, and use of angiotensin-converting enzymeibitors/angiotensin receptor blockers were significantlyociated with cardiac death, whereas use of statins, revas-

le 1 Baseline clinical characteristics of patients with and with

mographics and Cardiovascular Risk Factorse (years)

(M/F)e (Chinese/non-Chinese)okers (active or quit 3 years)betes mellituspertensionpercholesterolemiaal impairmentpiratory diseasese 1 left ventricular ejection fraction (%)nts Occurred During Index Admissionste pulmonary edematricular tachycardia or ventricular fibrillationatment Receivedascularizations

tiplatelet/warfarina-blocker

giotensin-converting enzyme inhibitor/angiotensin receptor blocktin

Values are given as mean SD, number, or number (percent)..05.eatinine 200 mol/L.spiratory disease consisted of clearly documented asthma, chronic obstrucosis.

le 2 Exercise stress test parameters and mean left ventricular

With exercise training (n

Phase 1 Phase 2

ting heart rate (bpm) 71.2 13.1 71.9 k heart rate (bpm) 121.7 22.8 125.4

art rate increment (bpm) 50.5 20.8 53.4 art rate recovery (bpm) 17.5 10.0 19.0 rcise capacity (METs) 5.4 3.2 7.2 t ventricular ejection fraction (%) 46.5 9.9 52.1 ke treadmill score (risk group)ow riskntermediate riskigh riskndeterminate risk

187 (56.0)91 (27.2)3 (0.9)

53 (15.9)

167 (5094 (287 (2.

66 (19

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

.05.larizations before enrollment, and exercise training weretective against cardiac death (Table 4, all P.05). Basedthe results of the receiver operating characteristic curve

alyses in defining cutoff values (Table 5), several ESTrameters, including phase 1 resting heart rate65 bpm,art rate increment 45 bpm, and exercise capacity 4

Ts, and phase 2 heart rate increment 45 bpm, HRR2 bpm, and exercise capacity4 METs were found to bedictors of cardiac death (Table 4, all P.05). In contrast,

k assessment by DTS was not predictive of cardiac deathable 4, P .05).

ercise training

Exercise training(n 334)

No exercise training(n 52) P value

63.8 11.2 65.0 10.2 .40256/78 41/11 .86321/13 52/0 .23124 (37.1) 24 (46.2) 1122 (36.5) 25 (48.1) .13178 (53.3) 30 (57.7) .65226 (67.7) 28 (53.8) .05929 (8.7) 4 (7.7) 136 (10.8) 6 (11.5) .81

46.5 9.9 46.3 10.1 .90

85 (25.4) 14 (26.9) .8721 (6.3) 4 (7.7) .76

255 (76.3) 23 (44.2) .001*333 (99.7) 48 (92.3) .001*266 (79.6) 38 (73.1) .28266 (79.6) 41 (78.8) .86276 (82.6) 30 (57.7) .001*

lmonary disease, bronchiectasis, interstitial lung disease, or pulmonary

n fraction

) Without exercise training (n 52)

P value Phase 1 Phase 2 P value

.26 71.3 16.3 72.3 15.6 .61

.001* 118.9 25.6 124.4 21.7 .068

.006* 47.6 23.2 52.2 21.4 .060

.011* 19.1 14.4 20.9 15.1 .37.001* 4.3 2.7 5.1 3.1 .001*.001* 46.3 10.1 53.2 12.1 .003*

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

20 (38.5)22 (42.3)2 (3.8)8 (15.4)

22 (42.3)18 (34.6)3 (5.8)9 (17.3)

.001*.001*.001*

.10paheME1preris(T

ejectio

334

12.823.322.012.33.413.4

.0)

.1)1).8)out ex

2.20.3preva

1.3ph9595

deass

aftefftie2ma

sur

1.3

Figphafromwo

Tab out ca

DeAgSexRacSmDiaHyHyRenResPhaEveAcuVenTreRevAnBetAn erStaExe

*P

932 Heart Rhythm, Vol 7, No 7, July 2010Multivariate analysis revealed that diabetes mellitus (HR8, 95% CI 1.015.19, P .049) and use of statins (HR6, 95% CI 0.160.80, P .012) were independentdictors of cardiac death. Whereas none of the phase 1

riables except resting HR65 bpm (HR 5.37, 95% CI321.61, P .018) significantly predicted mortality,

ase 2 parameters including LVEF 30% (HR 4.70,% CI 1.34 16.46, P .016), HRR 12 bpm (HR 2.49,% CI 1.10 5.63, P .028), and exercise capacity 4

ure 1 Change in Duke treadmill risk score between phase 1 andse 2 with reference to exercise training. Improved risk improved

higher-risk to lower-risk group; same risk no change in risk group;rsened risk worsened from lower-risk to higher-risk group.

le 3 Baseline clinical characteristics of patients with and with

mographics and Cardiovascular Risk Factorse (years)

(M/F)e (Chinese/non-Chinese)okers (active or quit 3 years)betes mellituspertensionpercholesterolemiaal impairmentpiratory diseasese 1 left ventricular ejection fractionnts Occurred During Index Admissionste pulmonary edematricular tachycardia or ventricular fibrillationatment Receivedascularizations*

tiplatelet/warfarina-blockergiotensin-converting enzyme inhibitor/angiotensin receptor blocktinrcise training

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

.05.Ts (HR 3.63, 95% CI 1.1711.28, P .026) remainedbe significant independent predictors of cardiac deathable 5).Comparison of patients characteristics in relation toase 2 HRR is given in Table 6. Patients with phase 2R 12 bpm were older and more often had diabetesllitus, hypertension, respiratory disease, and acute pul-nary edema on presentation. However, fewer of themeived beta-blockers or had undergone revasculariza-n (Table 6, all P .05). They had a lower HRR atase 1, achieved a lower heart rate increment and exer-e capacity at both phases, and showed a significantlyer LVEF and higher resting heart rate at the end of

ase 2 (Table 6, all P .05). Compared with patients whophase 2 HRR12 bpm, fewer patients with phase 2 HRR

2 bpm belonged to the low-risk group as assessed by DTS,ereas more of these patients belonged to the intermediate-k group (Table 6, both P .05). Nevertheless, the propor-ns of high-risk patients between the two groups were notnificantly different.In this study, a cutoff value 12 bpm obtained fromeiver operating characteristic curve analysis was used to

fine impaired HRR. This value has also been shown to beociated with increased mortality at predischarge ESTer MI.8 Further analysis was performed to investigate theect of improved HRR on survival in this group of pa-nts. Of the 109 patients identified, 56 (51.4%) had phaseHRR improved to 12 bpm. Patients whose HRR re-ined 12 bpm at the end of phase 2 had much worsevival than did those who improved (HR 6.15, 95% CI029.15, P .022; Figure 2).

rdiac death

Cardiac death(n 40)

No cardiac death(n 346) P value

66.8 12.0 63.6 11.0 .019*29/11 268/78 .5539/1 334/12 115 (37.5) 133 (38.4) .2623 (55.0) 124 (35.8) .01*22 (55.0) 186 (53.8) 123 (57.5) 231 (66.8) .294 (10.0) 29 (8.4) .766 (15.0) 36 (10.4) .42

43.0 10.3 46.9 9.8 .028*

19 (47.5) 80 (23.1) .002*4 (10.0) 21 (6.1) .31

16 (40.0) 262 (75.7) 0.001*38 (95.0) 343 (99.1) .08627 (67.5) 277 (80.1) .1037 (92.5) 270 (78.0) .037*21 (52.5) 285 (82.4) .001*26 (65.0) 308 (89.0) .001*MEto(T

phHRme

mo

rec

tiophcislowphhad1whristiosig

rec

DiOuLVthiESingfun

Tab

AgSexSmDiaHyHy 31.88Ren 22.51Res 02.84Pha 65.97Pha 710.0Acu 36.34Ven 85.37Rev 40.88An 71.28Bet 11.18An 112.7

Sta 10.75Exe 20.83PhaPhaPhaPhaPha

LIH

PhaPhaPhaPha

LIH

*P

Tab

Par

PhaPhaPhaPhaPhaPha

*P

933Hai et al Improved HR Recovery After Exercise Training and Clinical Outcome Post-MIscussionr results demonstrated exercise training improved HRR,EF, and exercise capacity in patients with recent MI. In

s study, none of the parameters during baseline phase 1T, except resting heart rate, predicted cardiac death dur-

long-term follow-up. In contrast, left ventricular dys-ction with LVEF 30%, impaired HRR 12 bpm, and

le 4 Predictors of cardiac death by Cox analysis for all patien

Univariate pred

HR 95%

e 1.03 1.00.86 0.4

oker 0.59 0.2betes mellitus 2.64 1.4pertension 0.94 0.5percholesterolemia 1.00 0.5al impairment 0.83 0.3piratory disease 1.19 0.5se 1 LVEF 30% 2.75 1.2se 2 LVEF 30% 4.10 1.6te pulmonary edema 3.40 1.8tricular tachycardia/ventricular fibrillation 1.91 0.6ascularizations 0.45 0.2

tiplatelet/warfarin 0.31 0.0a-blocker 0.61 0.3giotensin-converting enzyme inhibitor/

angiotensin receptor blocker3.92 1.2

tin 0.40 0.2rcise training 0.43 0.2se 1 resting heart rate 65 bpm 3.04 1.2se 1 heart rate increment 45 bpm 2.25 1.1se 1 heart rate recovery 12 bpm 1.56 0.8se 1 exercise capacity 4 METs 2.90 1.3se 1 Duke treadmill score (risk group)ow riskntermediate riskigh risk

0.771.420.47

0.40.70.0

se 2 heart rate increment 45 bpm 2.76 1.4se 2 heart rate recovery 12 bpm 4.00 2.1se 2 exercise capacity 4 METs 6.51 3.1se 2 Duke treadmill score (risk group)ow riskntermediate riskigh risk

0.532.100.50

0.21.00.0

CI confidence interval; HR hazard ratio; LVEF left ventricular eje.05.

le 5 Receiver operating characteristic curve analysis

ameter AUC 95% CI P

se 1 resting heart rate 65 bpm 0.62 0.570.67 .0se 1 heart rate increment 45 bpm 0.66 0.610.70 .0se 1 exercise capacity 4 METs 0.66 0.610.71 .0se 2 heart rate increment 45 bpm 0.66 0.610.71 .0se 2 heart rate recovery 12 bpm 0.66 0.610.71 .0se 2 exercise capacity 4 METs 0.75 0.700.79 .0

AUC area under the curve; CI confidence interval; METs metabolic

.05.ercise capacity 4 METs at phase 2 after training werewn to be independent predictors of cardiac death. Fur-rmore, persistently impaired HRR 12 bpm at phase 2s found to have incremental prognostic value over tradi-nal markers, including diabetes mellitus, resting hearte, LVEF, and exercise capacity.2428 Of interest, patientsth phase 1 HRR 12 bpm had a much better prognosis if

Multivariate predictors

P value HR 95% CI P value

.10

.66

.19

.002* 2.28 1.015.19 .049*

.84

.99

.89

.70

.011* 1.99 0.745.32 .176 .002* 4.70 1.3416.46 .016*

.001* 1.76 0.803.89 .160.22.018* 1.08 0.402.92 .877.10.14

1 .023* 2.80 0.6312.45 .177

.004* 0.36 0.160.80 .012*

.011* 0.59 0.221.11 .088

.012* 5.37 1.3321.61 .018*

.017* 1.79 0.565.70 .325

.18

.005* 0.49 0.141.71 .261

.33

.44

.30

.46

.002* 0.80 0.302.08 .641.001* 2.49 1.105.63 .028*

3 .001* 3.63 1.1711.28 .026*

.11

.10

.49

action; METs metabolic equivalents.

Sensitivity (%) Specificity (%) PPV (%) NPV (%)

80.0 36.71 12.7 94.167.5 54.9 14.8 93.677.5 50.0 15.2 95.162.5 62.1 16.0 93.552.5 77.8 21.4 93.475.0 72.0 23.6 96.1

nts; NPV negative predictive value; PPV positive predictive value.ex

shothewa

tioratwi

87.2664.3622.9686.10

01.5032.77016155.2547.48813.3

61.0914.3673.67

ction fr

value

144*002*001*001*001*001*

equivalets

ictors

CI

01.0631.7271.2914.9501.75

HRcar

prototio

floinlevan

tairev

co

sum

Tab se 2 h

DeAgSexRacSmDiaHyHyRenResPhaPhaEveAcuVenTreRevAnBetAnStaExeExePhaPhaPhaPhaPha

LIH

PhaPhaPhaPha

LIH

*PCrRefibr

934 Heart Rhythm, Vol 7, No 7, July 2010R improved to 12 bpm upon completion of phase 2 ofdiac rehabilitation. These findings suggest that im-vement of HRR after exercise training may contributethe long-term beneficial effects of cardiac rehabilita-n after MI.Following MI, there is an increase in sympathetic out-w to the heart and peripheral vasculature and a decreasecardiac vagal outflow. In addition, plasma norepinephrineel rises due to increased spillover and reduced clear-

ce.4 Although these alterations are important for main-ning blood pressure in the acute state, patients who fail toerse these changes may suffer from long-term adverse

nsequences, such as elevated myocardial oxygen con-

le 6 Comparison of patients characteristics in relation to pha

mographics and Cardiovascular Risk Factorse (years)

(M/F)e (Chinese/non-Chinese)okers (active or quit 3 years)betes mellituspertensionpercholesterolemiaal impairmentpiratory diseasese 1 LVEF (%)se 2 LVEF (%)nts Occurred During Index Admissionste pulmonary edematricular tachycardia/ventricular fibrillationatment Receivedascularizations

tiplatelet/warfarina-blocker

giotensin-converting enzyme inhibitor/angiotensin receptor blocktinrcise trainingrcise Stress Test Parametersse 1 resting heart rate (bpm)se 1 heart rate increment (bpm)se 1 exercise capacity (METs)se 1 HRR (bpm)se 1 Duke treadmill score (risk group)ow riskntermediate riskigh riskse 2 resting heart rate (bpm)se 2 heart rate increment (bpm)se 2 exercise capacity (METs)se 2 Duke treadmill score (risk group)ow riskntermediate riskigh risk

Values are given as mean SD, number, or number (percent).HRR heart rate recovery; LVEF left ventricular ejection fraction; ME .05.eatinine 200 mol/L.spiratory disease consisted of clearly documented asthma, chronic obstrucosis.ption, unfavorable cardiac remodeling, and increased mopensity to malignant arrhythmia and therefore would behigher risk for cardiac death.4Randomized trials have shown that cardiac rehabilitationer MI decreased long-term all-cause and cardiovascularrtality.29 Possible mechanisms include better cardiovas-

lar risk factors control, physical strength, psychologicalll-being, cardiac remodeling, and sympathovagal bal-ce. As a reflection of autonomic balance, improved HRRer rehabilitation can be regarded as a sign of restorationcardiovascular autonomic control. To our knowledge, noor study has addressed the potential prognostic value ofst-rehabilitation HRR on cardiac death among patientsth prior MI. Our results provide evidence supporting

eart rate recovery

HRR 12 bpm(n 99)

HRR 12 bpm(n 287) P value

68.4 9.1 62.4 11.3 .001*73/26 224/63 .3896/3 277/10 .8339 (39.4) 109 (38.0) .0848 (48.5) 99 (34.5) .013*64 (64.6) 144 (50.2) .013*61 (61.6) 193 (67.2) .3113 (13.1) 20 (7.0) .05918 (18.2) 24 (8.4) .007*

46.1 9.8 46.6 9.9 .6749.9 12.5 53.1 13.4 .047*

37 (37.4) 62 (21.6) .002*7 (7.1) 18 (6.3) .78

65 (65.7) 213 (74.2) .014*96 (97.0) 285 (99.3) .0870 (70.7) 234 (81.5) .023*80 (80.8) 227 (79.1) .7273 (73.7) 233 (81.2) .12

249 (86.8) 85 (85.9) .82

73.5 15.0 70.4 12.9 .07640.3 18.1 53.5 21.0 .001*3.6 2.1 5.8 3.2 .001*

12.2 9.6 19.6 10.3 .001*

43 (43.4)39 (39.4)0 (0)

164 (57.1)74 (25.8)5 (1.7)

.014*

.005*

.1975.2 14.9 70.9 12.4 .005*38.2 15.4 58.5 21.4 .001*5.1 2.6 7.5 3.5 .001*

33 (33.3)36 (36.4)2 (2.0)

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

.005*

.003*

.82

etabolic equivalents.

lmonary disease, bronchiectasis, interstitial lung disease, or pulmonaryproat

aftmo

cu

we

an

aftofpripowi

er

Ts m

tive pudulation of sympathovagal tone as underlying an impor-

tantalaft1

sysfroindtioraticativforpatheme

icacar

faiincprethibybeex

inMIisindredthi

DTsen

atas

sigstrscr

eart rat

935Hai et al Improved HR Recovery After Exercise Training and Clinical Outcome Post-MIt mechanism by which exercise training decreases mor-ity, as patients who had restoration of autonomic controler exercise training, as evidenced by HRR improved to2 bpm, had better survival during long-term follow-up.Assessment of autonomic control of the cardiovasculartem after MI provides completely distinct informationm conventional prognostic markers.4 Several differentexes have been used to assess cardiac autonomic func-n.4 Although widely studied, parameters such as hearte variability and baroreflex sensitivity are of limited clin-l value due to their intrinsic methodologic and interpre-e difficulties.30,31 In contrast, ESTs are routinely per-med for risk stratification after MI, and simple autonomicrameters obtained from ESTs are easy to interpret. Fur-rmore, unlike chronotropic index and heart rate incre-nt, which also depend on myocardial perfusion and phys-l strength, HRR is a more precise measurement of thediovascular autonomic balance.32 In particular, our studyled to demonstrate any correlation between heart raterement and cardiac death, in contrast to findings fromvious studies on healthy volunteers.32,33 The reason fors discrepancy remains unclear but can be explained partlythe fact that 80% of our patients were treated with

ta-blockers, which blunted heart rate response duringercise but either enhanced or had no effect on HRR.34,35DTS failed to predict survival for our patients, which iscontrast with findings of previous validation studies ofpopulations.17,19 Although the reason remains unclear, it

possible that a significant proportion of our patients hadeterminate DTS due to baseline abnormal ECG, whichuces the power of DTS to predict clinical outcomes in

Figure 2 Kaplan-Meier survival curves for hs study. On the other hand, HRR appears to be superior to ingS as a predictor of clinical outcome after MI, being moresitive and unaffected by preexisting ECG abnormalities.As shown in our study, patients who had impaired HRRbaseline seem to benefit more from cardiac rehabilitation,those with improved HRR after exercise training hadnificantly lower cardiac mortality. Whether therapeuticategies such as titration of medications and exercise pre-iption that target to HRR can further optimize autonomiclance and improve clinical outcomes in MI patients needsbe addressed in future clinical studies.

udy limitationsst, our study recruited patients who completed phase 1d phase 2 of the cardiac rehabilitation program with goodendance and were able to perform symptom-limited ESTtreadmill. As a result, high-risk populations of patientso could not complete the program due to repeated hos-alizations or could not perform treadmill tests due to poordiovascular function were excluded. Indeed, the cardiac

ath rate was 21% and sudden cardiac death rate was.3% among patients excluded (n 211), which werech higher than that in the population studied. Therefore,

neralization of our results to this group of patients shouldmade with caution. Second, this study was not designedevaluate the benefit of exercise training. Although thisdy included patients who were randomized to the control

in our previous trial, the nonrandomized nature of thesent study and the severe imbalance between treatment

d control groups preclude analysis of the effect of exer-e training. Nevertheless, previous randomized studiesve already proved that HRR improves with exercise train-

e recovery. P1 phase 1; P2 phase 2.bato

StFiran

atton

whpitcar

de12mu

gebetostuarm

prean

cisha,9,11 and our study only addressed the effect of changes

in HRR on survival. Third, although this was a prospectiveobservational study, the endpoints were not prespecified butwere derived from post hoc analysis. Finally, only data onHRR at 1 minute were prospectively collected for all pa-tielatun

CoOuHRbplonva

tribbiltarne

Re1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

recent myocardial infarction or percutaneous coronary intervention. Arch PhysMed Rehabil 2004;85:19151922.

16. Zipes DP, Camm AJ, Borggrefe M, et al. ACC/AHA/ESC 2006 guidelines formanagement of patients with ventricular arrhythmias and the prevention ofsudden cardiac death: a report of the American College of Cardiology/AmericanHeart Association Task Force and the European Society of Cardiology Com-

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

936 Heart Rhythm, Vol 7, No 7, July 2010nts in this study; whether HRR measured at 2 minutes orer would provide better prognostic information remainsclear.

nclusionr study demonstrated that exercise training improvedR and that patients with persistently impaired HRR 12

m after rehabilitation had higher cardiac mortality duringg-term follow-up, suggesting that modulation of cardio-

scular autonomic control with exercise training may con-ute to the long-term beneficial effects of a cardiac reha-

itation program. Its clinical application as a therapeuticget for exercise prescription and medication titrationeds to be further addressed in future studies.

ferencesCurtis BM, OKeefe JH Jr. Autonomic tone as a cardiovascular risk factor: thedangers of chronic fight or flight. Mayo Clin Proc 2002;77:4554.Schwartz PJ, La Rovere MT, Vanoli E. Autonomic nervous system and suddencardiac death. Experimental basis and clinical observations for post-myocardialinfarction risk stratification. Circulation 1992;85(1 Suppl):I77I91.Freeman JV, Dewey FE, Hadley DM, Myers J, Froelicher VF. Autonomicnervous system interaction with the cardiovascular system during exercise. ProgCardiovasc Dis 2006;48:342362.Frenneaux MP. Autonomic changes in patients with heart failure and in post-myocardial infarction patients. Heart 2004;90:12481255.Nishime EO, Cole CR, Blackstone EH, Pashkow FJ, Lauer MS. Heart raterecovery and treadmill exercise score as predictors of mortality in patientsreferred for exercise ECG. JAMA 2000;284:13921398.Cole CR, Foody JM, Blackstone EH, Lauer MS. Heart rate recovery aftersubmaximal exercise testing as a predictor of mortality in a cardiovascularlyhealthy cohort. Ann Intern Med 2000;132:552555.Cole CR, Blackstone EH, Pashkow FJ, Snader CE, Lauer MS. Heart-raterecovery immediately after exercise as a predictor of mortality. N Engl J Med1999;341:13511357.Nissinen SI, Makikallio TH, Seppanen T, et al. Heart rate recovery after exerciseas a predictor of mortality among survivors of acute myocardial infarction. Am JCardiol 2003;91:711714.Tiukinhoy S, Beohar N, Hsie M. Improvement in heart rate recovery aftercardiac rehabilitation. J Cardiopulm Rehabil 2003;23:8487.Adams BJ, Carr JG, Ozonoff A, Lauer MS, Balady GJ. Effect of exercisetraining in supervised cardiac rehabilitation programs on prognostic variablesfrom the exercise tolerance test. Am J Cardiol 2008;101:14031407.Streuber SD, Amsterdam EA, Stebbins CL. Heart rate recovery in heart failurepatients after a 12-week cardiac rehabilitation program. Am J Cardiol 2006;97:694698.Malfatto G, Facchini M, Sala L, Branzi G, Bragato R, Leonetti G. Relationshipbetween baseline sympatho-vagal balance and the autonomic response to cardiacrehabilitation after a first uncomplicated myocardial infarction. Ital Heart J2000;1:226232.Lucini D, Milani RV, Costantino G, Lavie CJ, Porta A, Pagani M. Effects ofcardiac rehabilitation and exercise training on autonomic regulation in patientswith coronary artery disease. Am Heart J 2002;143:977983.Nomenclature and criteria for diagnosis of ischemic heart disease. Report of theJoint International Society and Federation of Cardiology/World Health Organi-zation task force on standardization of clinical nomenclature. Circulation 1979;59:607609.Yu CM, Lau CP, Chau J, et al. A short course of cardiac rehabilitation programis highly cost effective in improving long-term quality of life in patients withmittee for Practice Guidelines. J Am Coll Cardiol 2006;48:e247e346.Villella M, Villella A, Santoro L, Santoro E, Franzosi MG, Maggioni AP; DukeTreadmill Score, Veterans Administration Medical Center Score, GISSI-2 In-vestigators. Ergometric score systems after myocardial infarction: prognosticperformance of the Duke Treadmill Score, Veterans Administration MedicalCenter Score, and of a novel score system, GISSI-2 Index, in a cohort ofsurvivors of acute myocardial infarction. Am Heart J 2003;145:475483.Mark DB, Hlatky MA, Harrell FE, Jr., Lee KL, Califf RM, Pryor DB. Exercisetreadmill score for predicting prognosis in coronary artery disease. Ann InternMed 1987;106:793800.Valeur N, Clemmensen P, Grande P, Saunamaki K; DANAMI-2 Investigators.Prognostic evaluation by clinical exercise test scores in patients treated withprimary percutaneous coronary intervention or fibrinolysis for acute myocardialinfarction (a Danish Trial in Acute Myocardial Infarction-2 Sub-Study). Am JCardiol 2007;100:10741080.Valeur N, Clemmensen P, Saunamaki K, Grande P; DANAMI-2 Investigators.The prognostic value of pre-discharge exercise testing after myocardial infarc-tion treated with either primary PCI or fibrinolysis: a DANAMI-2 sub-study. EurHeart J 2005;26:119127.Kwok JM, Miller TD, Christian TF, Hodge DO, Gibbons RJ. Prognostic valueof a treadmill exercise score in symptomatic patients with nonspecific ST-Tabnormalities on resting ECG. JAMA 1999;282:10471053.Villella A. The prognostic value of an ergometric scoring system in symptomaticpatients with nonspecific ST-T anomalies in the resting ECG. Ital Heart J Suppl2000;1:573574.Moss AJ, Zareba W, Hall WJ, et al. Prophylactic implantation of a defibrillatorin patients with myocardial infarction and reduced ejection fraction. N EnglJ Med 2002;346:877883.Vanhees L, Fagard R, Thijs L, Amery A. Prognostic value of training-inducedchange in peak exercise capacity in patients with myocardial infarcts andpatients with coronary bypass surgery. Am J Cardiol 1995;76:10141019.Feuerstadt P, Chai A, Kligfield P. Submaximal effort tolerance as a predictor ofall-cause mortality in patients undergoing cardiac rehabilitation. Clin Cardiol2007;30:234238.Specchia G, De Servi S, Scire A, et al. Interaction between exercise training andejection fraction in predicting prognosis after a first myocardial infarction.Circulation 1996;94:978982.Fox K, Borer JS, Camm AJ, et al. Resting heart rate in cardiovascular disease.J Am Coll Cardiol 2007;50:823830.Mukamal KJ, Nesto RW, Cohen MC, et al. Impact of diabetes on long-termsurvival after acute myocardial infarction: comparability of risk with priormyocardial infarction. Diabetes Care 2001;24:14221427.OConnor GT, Buring JE, Yusuf S, et al. An overview of randomized trials ofrehabilitation with exercise after myocardial infarction. Circulation 1989;80:234244.Bigger JT Jr, La Rovere MT, Steinman RC, et al. Comparison of baroreflexsensitivity and heart period variability after myocardial infarction. J Am CollCardiol 1989;14:15111518.Kleiger RE, Miller JP, Bigger JT Jr, Moss AJ. Decreased heart rate variabilityand its association with increased mortality after acute myocardial infarction.Am J Cardiol 1987;59:256262.Myers J, Tan SY, Abella J, Aleti V, Froelicher VF. Comparison of the chrono-tropic response to exercise and heart rate recovery in predicting cardiovascularmortality. 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 Med2005;352:19511958.Pavia L, Myers J, Rusconi C. Effect of beta-blockade on heart rate and VO2kinetics during recovery in patients with coronary artery disease. Heart Drug2002;2:6974.Racine N, Blanchet M, Ducharme A, et al. Decreased heart rate recovery afterexercise in patients with congestive heart failure: effect of beta-blocker therapy.J Card Fail 2003;9:296302.

Relationship between changes in heart rate recovery after cardiac rehabilitation on cardiovascular mortality in patients with myocardial infarctionIntroductionMethodsStudy population

Study designExercise testingStatistical analysisResultsDiscussionStudy limitationsConclusionReferences