Listen to this manuscript’s

audio summary by

JACC Editor-in-Chief

Dr. Valentin Fuster.

J O U R N A L O F T H E AM E R I C A N C O L L E G E O F C A R D I O L O G Y VO L . 7 0 , N O . 1 4 , 2 0 1 7

ª 2 0 1 7 B Y T H E AM E R I C A N C O L L E G E O F C A R D I O L O G Y F O U N D A T I O N

P U B L I S H E D B Y E L S E V I E R

I S S N 0 7 3 5 - 1 0 9 7 / $ 3 6 . 0 0

h t t p : / / d x . d o i . o r g / 1 0 . 1 0 1 6 / j . j a c c . 2 0 1 7 . 0 8 . 0 1 7

ORIGINAL INVESTIGATIONS

Physical Activity and Mortality in PatientsWith Stable Coronary Heart Disease

Ralph A.H. Stewart, MD,a Claes Held, MD, PHD,b,c Nermin Hadziosmanovic, MSC,c Paul W. Armstrong, MD,d

Christopher P. Cannon, MD,e Christopher B. Granger, MD,f Emil Hagström, MD, PHD,b,c Judith S. Hochman, MD,g

Wolfgang Koenig, MD,h,i,j Eva Lonn, MD,k José C. Nicolau, MD,l Philippe Gabriel Steg, MD,m,n,o,p Ola Vedin, MD,b,c

Lars Wallentin, MD, PHD,b,c Harvey D. White, MB CHB, DSC,a on behalf of the STABILITY Investigators

ABSTRACT

Fro

CadC

Bo

Yo

MejDe

an

On

UnnP

dio

BACKGROUND Recommendations for physical activity in patients with stable coronary heart disease (CHD) are

based on modest evidence.

OBJECTIVES The authors analyzed the association between self-reported exercise and mortality in patients with

stable CHD.

METHODS A total of 15,486 patients from 39 countries with stable CHD who participated in the STABILITY (Stabilization

of Atherosclerotic Plaque by Initiation of Darapladib Therapy) study completed questions at baseline on hours spent

each week takingmild, moderate, and vigorous exercise. Associations between the volume of habitual exercise inmetabolic

equivalents of task hours/week and adverse outcomes during a median follow-up of 3.7 years were evaluated.

RESULTS A graded decrease in mortality occurred with increased habitual exercise that was steeper at lower compared

with higher exercise levels. Doubling exercise volume was associated with lower all-cause mortality (unadjusted

hazard ratio [HR]: 0.82; 95% confidence interval [CI]: 0.79 to 0.85; adjusting for covariates, HR: 0.90; 95% CI: 0.87 to

0.93). These associations were similar for cardiovascular mortality (unadjusted HR: 0.83; 95% CI: 0.80 to 0.87;

adjusted HR: 0.92; 95% CI: 0.88 to 0.96), but myocardial infarction and stroke were not associated with exercise volume

after adjusting for covariates. The association between decrease in mortality and greater physical activity was

stronger in the subgroup of patients at higher risk estimated by the ABC-CHD (Age, Biomarkers, Clinical–Coronary

Heart Disease) risk score (p for interaction ¼ 0.0007).

CONCLUSIONS In patients with stable CHD, more physical activity was associated with lower mortality.

The largest benefits occurred between sedentary patient groups and between those with the highest mortality risk.

(J Am Coll Cardiol 2017;70:1689–700) © 2017 by the American College of Cardiology Foundation.

m the aGreen Lane Cardiovascular Service, Auckland City Hospital, Auckland, New Zealand; bDepartment of Medical Sciences,

rdiology, Uppsala University, Uppsala, Sweden; cUppsala Clinical Research Center, Uppsala University, Uppsala, Sweden;

anadian Vigour Centre, University of Alberta, Edmonton, Canada; eCardiovascular Division, Brigham and Women’s Hospital,

ston, Massachusetts; fDuke Clinical Research Institute, Duke Medicine, Durham, North Carolina; gDepartment of Medicine, New

rk University Langone Medical Center, New York, New York; hDepartment of Internal Medicine II–Cardiology, University of Ulm

dical Center, Ulm, Germany; iDeutsches Herzzentrum München, Technische Universität München, Munich, Germany;

utsches Zentrum fur Herz-Kreislauf-Forschung (German Centre for Cardiovascular Research), partner site Munich Heart Alli-

ce, Munich, Germany; kDepartment of Medicine and Population Health Research Institute, McMaster University, Hamilton,

tario, Canada; lHeart Institute (InCor), University of São Paulo Medical School, São Paulo, Brazil; mDépartement Hospitalo-

iversitaire FIRE (Fibrosis Inflammation REmodeling), Assistance Publique–Hôpitaux de Paris, Hôpital Bichat, Paris, France;

aris Diderot University, Sorbonne Paris Cité, Paris, France; oNational Heart and Lung Institute, Imperial College, Institute of Car-

vascular Medicine and Science, Royal Brompton Hospital, London, United Kingdom; and the pFrench Alliance for Cardiovascular

ABBR EV I A T I ON S

AND ACRONYMS

CABG = coronary artery bypass

graft

CHD = coronary heart disease

CI = confidence interval

CV = cardiovascular

eGFR = estimated glomerular

filtration rate

HDL = high-density lipoprotein

HR = hazard ratio

LDL = low-density lipoprotein

METs = metabolic equivalents

MI = myocardial infarction

Trials, Fren

France. Th

were funde

Hagström,

grants and

honoraria f

Roche. Mr.

Merck & Co

Cannon ha

GlaxoSmith

Ingelheim,

Myers Squi

pany, and N

Lilly, Gilea

tional resea

Sanofi and

research gr

& Dohme. D

Institutes o

lecture and

from Glaxo

Diagnostics

conduct of

Dr. Nicolau

Daiichi-San

board mem

received pe

Merck Shar

personal fe

fees, hono

GlaxoSmith

consultanc

Ingelheim;

GDF-15 pro

board mem

National In

peutics Inc.

Manuscript

Stewart et al. J A C C V O L . 7 0 , N O . 1 4 , 2 0 1 7

Exercise and Coronary Heart Disease Mortality O C T O B E R 3 , 2 0 1 7 : 1 6 8 9 – 7 0 0

1690

C linical practice guidelines forprevention of cardiovascular (CV)disease recommend $150 min of

moderate intensity or $60 to 75 min ofvigorous exercise each week (1–3). Guide-lines on secondary prevention of stable coro-nary heart disease (CHD) have recommendedsimilar levels of regular moderate orvigorous exercise (4,5). These recommenda-tions are based in part on studies that indi-cate cardiorespiratory fitness predictsmortality, and regular moderate or vigorousexercise improves physical fitness morethan mild intensity exercise does (6). Mostinformation on the relationship between

SEE PAGE 1701

physical activity and mortality comes from large gen-eral population studies (7–9). These studies suggestthat there is a graded association between a combina-tion of the intensity and duration of self-reported reg-ular exercise and mortality, even at levels belowthose recommended in current guidelines (1–3).

ch Clinical Research Infrastructure Network, Institut National de

e STABILITY (Stabilization of Atherosclerotic Plaque by Initiation o

d by GlaxoSmithKline. The charge for Open Access has been p

Lonn, Armstrong, Granger, Hochman,Wallentin, andWhite are STA

nonfinancial support from GlaxoSmithKline. Dr. Held has received

romAstraZeneca; and institutional research grants fromBristol-My

Hadziosmanovic has received grants from GlaxoSmithKline. Dr. A

. and Bayer; grants from Sanofi; and personal fees fromAstraZenec

s received research grants and consulting fees from Arisaph, Astra

Kline, Merck & Co., Takeda; research grants from Janssen; and

Eli Lilly, Kowa, Lipimedix, Pfizer, Regeneron, and Sanofi. Dr. Grang

bb, Pfizer, Bayer, Daiichi-Sankyo, Boehringer Ingelheim, Janssen,

ovartis; grants from Armetheon, Medtronic Foundation, U.S. Food

d, Hoffmann-La Roche, Medtronic Inc., National Institutes of Hea

rch grants fromGlaxoSmithKline,AstraZeneca,Amgen, Sanofi, and

Amgen; has received lecture fees and institutional research grant

ants from AstraZeneca and GlaxoSmithKline; and has served as an

r. Hochman has received travel reimbursement from GlaxoSmit

f Health; and support for drug distribution related to the ISCHE

consultancy fees fromNovartis, Amgen, andAstraZeneca; lecture f

SmithKline, TheMedicines Company, Pfizer, DalCor, Merck Sharp

, Abbott, Singulex, andBeckmann. Dr. Lonn has received institutio

the study; and institutional research grants fromAstraZeneca,Hoffm

has received grant support from Eli Lilly, AstraZeneca, Johnson &

kyo; has receivedgrant support andpersonal fees fromGlaxoSmith

ber for AstraZeneca, Bayer, and Sanofi; and has received lecture fe

rsonal fees fromGlaxoSmithKline, Amarin, Bayer, Boehringer Inge

p & Dohme, Novartis, Pfizer, The Medicines Company, CLS-Behrin

es, honoraria, and nonfinancial support from Sanofi and Servier;

raria, and nonfinancial support from AstraZeneca. Dr. Vedin

Kline; and lecture fees from Fresenius and Novartis. Dr. Wa

y fees, lecture fees, and travel support from Bristol-Myers Squibb, P

institutional research grants fromMerck&Co. andRoche; consulta

teins. Dr. White has received research grants and personal fees f

ber fees from AstraZeneca; advisory board member fees from Acet

stitute of Health, Merck Sharp & Dohme, George Institute, Omthera

, Elsai Inc., Dal-GenE, and Daiichi-Sankyo Pharma Development. Pau

received April 18, 2017; revised manuscript received August 3, 2

Milder intensity exercise, less sedentary time, andmore time spent standing are also associated withlower mortality in general population cohorts (10,11).

Few studies have evaluated the potential benefitsof lower intensity exercise in CHD populations,although several have evaluated more vigorous ex-ercise (12,13). Runners with a history of myocardialinfarction (MI) have lower mortality than nonrunners,but very high durations and intensities of runningmay increase CV risk (13). Randomized clinical trialsof exercise training after MI suggest that increasingexercise lowers CV risk (14). However, these trialsprovide limited evidence on the importance of theintensity and duration of exercise interventions forprognosis; most studies were small, and reporting ofexercise interventions was often poor (15).

We analyzed relationships between the amount ofmild, moderate, and vigorous physical activityassessed by self-reported questionnaire (16) and

la Santé et de la Recherche Mèdicale U1148, Paris,

f Darapladib Therapy) trial and the lifestyle substudy

aid by GlaxoSmithKline. Drs. Stewart, Held, Vedin,

BILITY Study Investigators. Dr. Stewart has received

an institutional research grant and Speakers Bureau

ers Squibb, Pfizer, Merck& Co., GlaxoSmithKline, and

rmstrong has received grants and personal fees from

a, Axio/Orexigen, Eli Lilly, andMast Therapeutics. Dr.

Zeneca, Bristol-Myers Squibb, Boehringer Ingelheim,

consulting fees from Alnylam, Amgen, Boehringer

er has received grants and personal fees fromBristol-

AstraZeneca, GlaxoSmithKline, The Medicines Com-

and Drug Administration; and personal fees from Eli

lth, and Verseon. Dr. Hagström has received institu-

Ariad; has served as an expert committeemember for

s from Sanofi and Amgen; has received institutional

expert committeemember for Ariad andMerck Sharp

hKline; a grant for the ISCHEMIA trial from National

MIA trial from AstraZeneca. Dr. Koenig has received

ees fromActavis andBerlin-Chemie; consultancy fees

& Dohme, and Kowa; and research grants from Roche

nal research grants fromGlaxoSmithKline, during the

ann-LaRoche,Novartis, Eli Lilly, Amgen, andBayer.

Johnson, Bristol-Myers Squibb, Astellas, Sanofi, and

Kline, during the conduct of the study; has served as a

es from AstraZeneca, Bayer, and Sanofi. Dr. Steg has

lheim, Bristol-Myers Squibb, Daiichi-Sankyo, Eli Lilly,

g, and Janssen; grants, institutional research grants,

honoraria from Amgen and Regeneron; and personal

has received institutional research grants from

llentin has received institutional research grants,

fizer, AstraZeneca, GlaxoSmithKline, and Boehringer

ncy fees fromAbbott; andholds twopatents involving

rom GlaxoSmithKline; research grants and advisory

elion and Sirtex; research grants from Sanofi, Eli Lilly,

Pharmaceuticals, Pfizer New Zealand, Intarcia Thera-

l Thompson,MD, served as Guest Editor for this paper.

017, accepted August 4, 2017.

J A C C V O L . 7 0 , N O . 1 4 , 2 0 1 7 Stewart et al.O C T O B E R 3 , 2 0 1 7 : 1 6 8 9 – 7 0 0 Exercise and Coronary Heart Disease Mortality

1691

subsequent all-cause mortality, CV mortality, non-CVmortality, MI, and stroke in a large cohort of patientswith stable CHD who participated in the globalSTABILITY (Stabilization of Atherosclerotic Plaque byInitiation of Darapladib Therapy) trial (17).

METHODS

STUDY POPULATION. STABILITY was a global out-comes trial designed to determine whether dar-apladib, a specific inhibitor of lipoprotein associatedphospholipase A2, would reduce risk of CV death, MI,and stroke in patients with chronic CHD (18). Subjectsfrom 39 countries (n ¼ 15,828) were randomized. Allpatients had chronic stable CHD, defined as prior MI(>1 month before randomization), prior percutaneouscoronary intervention, coronary artery bypass graft(CABG), or multivessel CHD confirmed by coronaryangiography. In addition, patients had to meet atleast 1 of the following CV risk criteria: age $60 years;diabetes mellitus requiring pharmacotherapy; high-density lipoprotein (HDL)-cholesterol <1.03 mmol/l;current or previous smoker defined as $5 cigarettesper day on average; significant renal dysfunction(estimated glomerular filtration rate [eGFR]: $30and <60 ml/min/1.73 m2 or urine albumin-creatinineratio $30 mg albumin/g creatinine); or polyvasculardisease (CHD and cerebrovascular disease or CHD andperipheral arterial disease). Detailed descriptions ofthe study design and population have been publishedpreviously (18,19).

PHYSICAL ACTIVITY. At baseline, patients under-went a detailed medical history, physical examina-tion, and fasting blood samples, and they wereinvited to complete a lifestyle questionnaire. Ques-tions related to physical activity (based on the Inter-national Physical Activity Questionnaire [20]), werecompleted by 15,486 subjects (97.8%). Each subjectwas asked “How many hours during a typical week doyou spend doing the following activities for 10 mi-nutes or more? Please estimate to the nearest 1 h foreach category: 1) Doing MILD physical activity such aseasy walking, yoga, Tai Chi, mild house work? 2)Doing MODERATE physical activity such as fastwalking, jogging, aerobics, gardening, bicycling,dancing, swimming or house cleaning? 3) DoingVIGOROUS exercise such as running, lifting heavyobjects, playing strenuous sports or strenuous work?”Each level of exercise includes a range of intensitiesestimated to be <3 metabolic equivalents (METs) fortask for mild, 3 to 6 METs for moderate, and >6 METsfor vigorous intensity physical activity (21). To esti-mate the METs h/week, 2 METs were assigned formild, 4 METs for moderate, and 8 METs for vigorous

intensity activity, as previously reported (16). Theaverage intensity of physical activity was calculatedby dividing the total METs h/week by the total hoursof exercise per week. The baseline characteristics ofthe study population have been described by tertile ofself-reported physical activity previously (16).

Additionally, participants were asked “How activeare you at work and during leisure time? At work(check 1) ‘mainly sedentary,’ ‘predominantly walkingon 1 level, no heavy lifting,’ ‘mainly walking,including climbing stairs, or walking uphill or liftingheavy objects,’ ‘heavy physical activity,’ or ‘I do notwork.’ During your leisure time (check 1) ‘mainlysedentary,’ ‘mild exercise,’ ‘moderate exercise,’ or‘strenuous physical exercise.’”

CLINICAL OUTCOMES. The primary endpoint was all-cause mortality. Secondary endpoints were CV death,MI, stroke, non-CV mortality, and major adverse cor-onary events defined as first occurrence of CV death,or stroke. For 90 deaths that occurred outside theperiod defined for evaluation of darapladib treatment,the cause of death was not defined. All other eventswere independently reviewed by an adjudicationcommittee who had no knowledge of physical activitydata. Event definitions and main results of the studyhave been presented elsewhere (17).

STATISTICAL ANALYSIS. The baseline characteris-tics of the study population were summarized byphysical activity tertiles, with categorical variablespresented as counts with proportions and continuousvariables as medians and interquartile ranges. Toinvestigate differences across the 3 groups of pa-tients, categorical variables were compared with thechi-square test. Continuous variables were comparedwith Mann-Whitney nonparametric tests.

The associations between self-reported exerciseand study outcomes were assessed using Cox propor-tional hazards regression models and expressed withhazard ratios (HRs) and 95% confidence intervals (CIs).In themultivariatemodel, we adjusted for randomizedtreatment, age, body mass index, sex, systolic bloodpressure, hypertension, geographic region for finalreporting, prior MI, prior percutaneous coronaryintervention or CABG, prior multivessel CHD, baselinehistory of diabetes, smoking, polyvascular disease,significant renal dysfunction, hemoglobin, whiteblood cell count, low-density lipoprotein (LDL)-cholesterol, HDL-cholesterol, triglycerides, eGFRaccording to CKD-EPI (Chronic Kidney Disease Epide-miology research group) calculator, and history ofcongestive heart failure at baseline.

HR for change in different measures of self-reportedphysical activity were also analyzed. Spline plots were

TABLE 1 Baseline Characteristics of Study Population by Physical Activity Tertile

All Subjects(N ¼ 15,487)

Least Active(n ¼ 5,281)

Intermediate Activity(n ¼ 5,055)

Most Active(n ¼ 5,151) p Value

Age, yrs 65.0 � 12.0 65.0 � 12.0 65.0 � 11.0 64.0 � 12.0 <0.0001

Female 2,885 (18.6) 921 (17.4) 989 (19.6) 975 (18.9) 0.0170

Time from qualifying event to randomization, yrs 3.42 � 6.38 3.22 � 6.32) 3.49 � 6.66 3.57 � 6.25 <0.0001

Follow-up time from randomization, yrs 3.79 � 0.31 3.75 � 0.32 3.80 � 0.31 3.81 � 0.31 <0.0001

Attended cardiac rehabilitation 5,397 (35.1) 1,608 (30.7) 1,865 (37.2) 1,924 (37.6) <0.0001

Geographic region

Asia/Pacific 3,052 (19.7) 1,275 (24.1) 1,054 (20.9) 723 (14.0) <0.0001

Eastern Europe 3,442 (22.2) 805 (15.2) 1,107 (21.9) 1,530 (29.7) <0.0001

North America 3,969 (25.6) 1,397 (26.5) 1,323 (26.2) 1,249 (24.2) <0.0001

South America 1,177 (7.6) 646 (12.2) 320 (6.3) 211 (4.1) <0.0001

Western Europe 3,847 (24.8) 1,158 (21.9) 1,251 (24.7) 1,438 (27.9) <0.0001

Risk factors

Current smoker 2,789 (18.0) 938 (17.8) 813 (16.1) 1,038 (20.2) <0.0001

Body mass index, kg/m2 28.3 � 6.2 28.4 � 6.9 28.2 � 6.0 28.3 � 5.8 0.0066

Systolic blood pressure, mm Hg 131.0 � 22.0 130.0 � 23.0 130.0 � 22.0 131.0 � 22.0 0.0100

Diastolic blood pressure, mm Hg 79.0 � 13.0 78.0 � 14.0 79.0 � 14.0 80.0 � 13.0 <0.0001

Hypertension 11,075 (71.5) 3,854 (73.0) 3,579 (70.8) 3,642 (70.7) 0.0144

LDL-cholesterol, mmol/l 2.07 � 0.99 2.01 � 1.01 2.06 � 0.96 2.14 � 1.00 <0.0001

HDL-cholesterol, mmol/l 1.15 � 0.38 1.13 � 0.40 1.17 � 0.40 1.19 � 0.40 <0.0001

Triglycerides, mmol/l 1.52 � 1.03 1.58 � 1.03 1.50 � 1.02 1.46 � 1.02 <0.0001

HB, g/l 144.0 � 18.0 142.0 � 19.0 144.0 � 18.0 145.0 � 17.0 <0.0001

WBC, gi/l 6.5 � 2.3 6.7 � 2.4 6.5 � 2.2 6.4 � 2.2 <0.0001

eGFR (according to CKD-EPI) 74.4 � 25.1 72.9 � 27.6 73.9 � 24.5 75.8 � 23.1 <0.0001

Diabetes mellitus 6,000 (38.7) 2,297 (43.5) 1,965 (38.9) 1,738 (33.7) <0.0001

Cardiovascular and renal disease markers

Significant renal dysfunction 4,682 (30.2) 1,828 (34.6) 1,557 (30.8) 1,297 (25.2) <0.0001

Prior myocardial infarction 9,114 (58.8) 3,156 (59.8) 2,907 (57.5) 3,051 (59.2) 0.0528

Prior PCI or CABG 11,610 (75.0) 3,881 (73.5) 3,855 (76.3) 3,874 (75.2) 0.0045

Multivessel coronary heart disease 2,337 (15.1) 918 (17.4) 711 (14.1) 708 (13.7) <0.0001

Polyvascular disease 2,325 (15.0) 829 (15.7) 759 (15.0) 737 (14.3) 0.1389

Congestive heart failure 3,326 (21.5) 1,139 (21.6) 1,087 (21.5) 1,100 (21.4) 0.9628

STABILITY CHD risk score 0.44 � 0.76 0.50 � 0.99 0.44 � 0.76 0.39 � 0.62 <0.0001

Measures of physical activity

All self-reported physical activity, METs h/week 40.0 � 52.0 14.0 � 12.0 40.0 � 14.0 90.0 � 52.0 <0.0001

Duration of physical activity, h/week 14.0 � 16.0 5.0 � 4.0 14.0 � 7.0 28.0 � 13.0 <0.0001

Average exercise intensity, METs 2.6 � 1.2 1.9 � 0.6 2.6 � 0.9 3.2 � 1.0 <0.0001

Mild exercise, h/week 7.0 � 10.0 4.0 � 5.0 9.0 � 8.0 14.0 � 14.0 <0.0001

Moderate exercise, h/week 4.0 � 10.0 0.0 � 2.0 4.0 � 5.0 11.0 � 12.0 <0.0001

Vigorous exercise, h/week 0.0 � 0.0 0.0 � 0.0 0.0 � 0.0 1.0 � 5.0 <0.0001

Moderate or vigorous intensity exercise, METs h/week 20.0 � 48.0 0.0 � 8.0 20.0 � 20.0 64.0 � 40.0 <0.0001

Continued on the next page

Stewart et al. J A C C V O L . 7 0 , N O . 1 4 , 2 0 1 7

Exercise and Coronary Heart Disease Mortality O C T O B E R 3 , 2 0 1 7 : 1 6 8 9 – 7 0 0

1692

constructed to show relations between self-reportedexercise and mortality—all-cause, CV, and non-CV(22,23). Because the association between METsh/week and mortality was nonlinear (Online Figure 1),splines were also evaluated for the natural logarithmof physical activity volume (Online Figure 2). Subjectswere grouped in increasing categories representing anapproximate doubling of exercise volumes: 0; 0 to <5;5 to <10; 10 to <20; 20 to <40; 40 to <80; 80 to <160;and >160 METs h/week.

Secondary analyses evaluated associations be-tween physical activity and mortality in subgroupsdefined by baseline clinical characteristics and

physical activity level. The ABC-CHD (Age, Bio-markers, Clinical–Coronary Heart Disease) riskscore, which estimates the risk of CV death basedon N-terminal pro-B-type natriuretic peptide, high-sensitivity troponin T, LDL-cholesterol, smoking,diabetes mellitus, and peripheral arterial disease, wasused to estimate overall CHD risk (24). The chi-squarevalue minus degrees of freedom was used to comparethe relative strength of association of all covariateswith mortality (Online Figure 3). All analyses wereperformed using SAS software version 9.4 (SASInstitute, Cary, North Carolina). A 2-sided p valueof <0.05 was considered statistically significant.

TABLE 1 Continued

All Subjects(N ¼ 15,487)

Least Active(n ¼ 5,281)

Intermediate Activity(n ¼ 5,055)

Most Active(n ¼ 5,151) p Value

Highest reported intensity of habitual physical activity

Mild 3,755 (24.2) 2,637 (50.6) 1,038 (20.5) 80 (1.6) <0.0001

Moderate 7,511 (48.4) 1,910 (36.2) 3,099 (61.3) 2,502 (48.6) <0.0001

Vigorous 3,651 (23.6) 164 (3.1) 918 (18.1) 2,569 (49.9) <0.0001

Physical activity during leisure time

Mainly sedentary 4,905 (32.5) 2,509 (48.9) 1,416 (28.8) 980 (19.5) <0.0001

Mild exercise 6,655 (44.2) 2,126 (41.4) 2,388 (48.5) 2,141 (42.7) <0.0001

Moderate exercise 3,377 (22.4) 484 (9.4) 1,089 (22.1) 1,804 (36.0) <0.0001

Strenuous physical exercise 136 (0.9) 15 (0.3) 32 (0.6) 89 (1.8) <0.0001

Physical activity at work

Mainly sedentary 2,499 (16.4) 1,157 (22.3) 819 (16.5) 523 (10.3) <0.0001

Predominantly walking on 1 level, no heavy lifting 2,558 (16.8) 856 (16.5) 869 (17.6) 833 (16.5) <0.0001

Mainly walking, including climbing stairs, or walkinguphill or lifting heavy objects

1,488 (9.8) 316 (6.1) 360 (7.3) 812 (16.0) <0.0001

Heavy physical activity 222 (1.5) 32 (0.6) 25 (0.5) 165 (3.3) <0.0001

I do not work 8,438 (55.5) 2,832 (54.5) 2,878 (58.1) 2,728 (53.9) <0.0001

Symptoms that limit exercise

Dyspnea 7,692 (50.1) 2,767 (52.9) 2,505 (50.0) 2,420 (47.3) <0.0001

Chest pain or discomfort 5,751 (37.5) 2,052 (39.3) 1,890 (37.9) 1,809 (35.4) 0.0003

Fatigue 9,347 (61.1) 3,226 (61.9) 3,022 (60.7) 3,099 (60.8) 0.3459

Values are mean � SD or n (%).

CABG ¼ coronary artery bypass graft; CHD ¼ coronary heart disease; CKD-EPI ¼ Chronic Kidney Disease Epidemiology research group calculator; eGFR ¼ estimatedglomerular filtration rate; HB ¼ hemoglobin; HDL ¼ high-density lipoprotein; LDL ¼ low-density lipoprotein; METs ¼ metabolic equivalents; PCI ¼ percutaneous coronaryintervention; STABILITY ¼ Stabilization of Atherosclerotic Plaque by Initiation of Darapladib Therapy (trial); WBC ¼ white blood cell count.

J A C C V O L . 7 0 , N O . 1 4 , 2 0 1 7 Stewart et al.O C T O B E R 3 , 2 0 1 7 : 1 6 8 9 – 7 0 0 Exercise and Coronary Heart Disease Mortality

1693

RESULTS

STUDY POPULATION. Baseline characteristics of thestudy population are presented in Table 1. Participantsin the most active tertile were less likely to havechronic renal disease, multivessel coronary arterydisease, or diabetes. Patients from South America andAsia/Pacific were less active, and those from Easternand Western Europe on average were more physicallyactive. White blood cell count was lower, and hemo-globin, HDL-cholesterol, and eGFR were higher in themost active tertile. LDL-cholesterol was slightly higherand current smoking slightly more common in themost active group. More physical activity was associ-ated with a lower ABC-CHD risk score.

For the least active tertile, the majority of exercisewas of mild intensity and few subjects in this groupreported more than 10 METs h/week of moderate orvigorous exercise each week (Table 1). Both the re-ported average exercise intensity and time spentexercising each week increased progressively fromthe least active to the intermediate and most activetertiles of physical activity. There were small differ-ences in the proportion of subjects who reported atleast some limitation of exercise by dyspnea andchest discomfort.

OUTCOMES BY TERTILE OF PHYSICAL ACTIVITY.

Outcomes by physical activity tertile before and afteradjusting for covariates are displayed in Figures 1Aand 1B, respectively. There was a graded lower total,CV, and non-CV mortality with higher physical ac-tivity. Total mortality in the most active tertile wasw50% that of the least active tertile in the unadjustedmodel, and w30% lower after adjusting for baselinecovariates. Physical activity was associated with CVdeath and with major adverse CV events. The risk ofMI was lower at higher physical activity before butnot after adjusting for covariates. There was no sig-nificant association between stroke and exercise ineither the unadjusted or fully adjusted models.

VOLUME, INTENSITY, AND DURATION OF HABITUAL

EXERCISE AND MORTALITY. Doubling of exercisevolume was associated with lower all-cause mortalityboth before (HR: 0.82; 95% CI: 0.79 to 0.85) and afteradjusting for covariates (HR: 0.90; 95% CI: 0.87 to0.93) (Central Illustration). The association betweendoubling of exercise volume was observed for bothCV (unadjusted HR: 0.83; 95% CI: 0.80 to 0.87;adjusted HR: 0.92; 95% CI: 0.88 to 0.96) and non-CVmortality (unadjusted HR: 0.83; 95% CI: 0.78 to0.88; adjusted HR: 0.88; 95% CI: 0.83 to 0.95)(Figure 2).

FIGURE 1 HR for Outcomes for the Middle and Most Active Physical Activity Tertiles of the Study Population Compared With the Most Sedentary

Tertile, Before and After Adjustment for Covariates

1.61.41.2Lower Risk Higher Risk

10.80.60.4

528150555151

528150555151

528150555151

528150555151

528150555151

528150555151

No ofPatients

HR (95% CI) LeastActive as Reference

Outcome PhysicalActivity

A

B

Least ActiveAll-cause mortality

CV death

Non-CV death

MACE

MI

Stroke

Intermediate activityMost Active

Least ActiveIntermediate activityMost Active

Least ActiveIntermediate activityMost Active

Least ActiveIntermediate activityMost Active

Least ActiveIntermediate activityMost Active

Least ActiveIntermediate activityMost Active

No of Events(%/3 yrs)

528 (8.72)326 (5.47)273 (4.45)

320 (5.29)229 (3.84)162 (2.64)

159 (2.63)76 (1.28)91 (1.49)

621 (10.61)515 (8.91)413 (6.93)

277 (4.70)251 (4.31)220 (3.67)

104 (1.73)109 (1.84)89 (1.46)

10.62 (0.54-0.72)0.50 (0.44-0.58)

10.72 (0.61-0.86)0.50 (0.41-0.60)

10.48 (0.37-0.63)0.56 (0.43-0. 72)

10.84 (0.75-0.94)0.65 (0.58-0.74)

10.92 (0.77-1.09)0.78 (0.66-0.94)

11.06 (0.81-1.39)0.84 (0.63-1.12)

p-value

<0.0001

<0.0001

<0.0001

<0.0001

0.0255

0.2498

1.61.41.2Lower Risk Higher Risk

10.80.60.4

508148574958

508148574958

508148574958

508148574958

508148574958

508148574958

No ofPatients

HR (95% CI) LeastActive as Reference

Outcome PhysicalActivity

Least ActiveAll-cause mortality

CV death

Non-CV death

MACE

MI

Stroke

Intermediate activityMost Active

Least ActiveIntermediate activityMost Active

Least ActiveIntermediate activityMost Active

Least ActiveIntermediate activityMost Active

Least ActiveIntermediate activityMost Active

Least ActiveIntermediate activityMost Active

No of Events(%/3 yrs)

504 (8.65)313 (5.46)264 (4.47)

307 (5.28)219 (3.82)157 (2.66)

154 (2.65)73 (1.27)88 (1.49)

597 (10.61)496 (8.93)397 (6.92)

267 (4.71)244 (4.36)211 (3.65)

98 (1.70)105 (1.85)86 (1.47)

10.75 (0.65-0.87)0.70 (0.60-0.82)

10.89 (0.74-1.06)0.71 (0.58-0.88)

10.54 (0.41-0.72)0.73 (0.55-0.96)

10.96 (0.85-1.08)0.81 (0. 71-0.92)

11.02 (0.86-1.22)0.90 (0.74-1.08)

11.15 (0.87-1.52)0.97 (0.71-1.31)

p-value

<0.0001

0.0052

<0.0001

0.0054

0.3514

0.4392

Plots are of hazard ratios (HR) and 95% confidence intervals (CI) (A) before and (B) after adjustment for covariates. Covariates included in the fully

adjusted model are randomized treatment, age, body mass index, sex, systolic blood pressure, hypertension, geographic region for final reporting, prior

myocardial infarction (MI), prior coronary revascularization percutaneous coronary intervention or coronary artery bypass graft, coronary heart disease,

baseline diabetes, smoking, polyvascular disease, significant renal dysfunction, hemoglobin, white blood cell count, low-density lipoprotein cholesterol,

high-density lipoprotein cholesterol, triglycerides, estimated glomerular filtration rate according to CKD-EPI (Chronic Kidney Disease-Epidemiology

research group) calculator, and congestive heart failure. CV ¼ cardiovascular; MACE ¼ major adverse cardiac event(s).

Stewart et al. J A C C V O L . 7 0 , N O . 1 4 , 2 0 1 7

Exercise and Coronary Heart Disease Mortality O C T O B E R 3 , 2 0 1 7 : 1 6 8 9 – 7 0 0

1694

CENTRAL ILLUSTRATION Habitual Physical Activity and Mortality in Patients With Stable CoronaryArtery Disease

All-cause mortality risk associated with each doubling of habitual physical activity volume, andby linear increase in physical activity

Characteristics associated with greatest potential to benefit from increase in physical activity

All-C

ause

Mor

talit

y Ri

sk

1

0.9

0.8

0.7

0.6

0.5

0.4

0.3

0.2

0.1

00 0

to 5>5

to 10>10

to 20>20

to 40>40

to 80Amount of Physical Activity (MET.hours/week)

↑ ABC-CHD risk score

↑ Age

↑ Troponin T

↑ NT-proBNP

↑ LDL cholesterol

Smoker

Diabetes

Peripheral arterydisease

Sedentary Limited by dyspnea

>80to 160

>160

1

0.8

0.6

0.4

0.20 50 100

MET h/w

All-C

ause

Mor

talit

y Ri

sk

150 200

Stewart, R.A.H. et al. J Am Coll Cardiol. 2017;70(14):1689–700.

All-cause mortality risk associated with each doubling of habitual exercise volume and by linear increase in physical activity (inset) in 15,486 patients with

stable coronary heart disease. Selected characteristics of patients who had the greatest reduction in mortality associated with increase in physical activity

are also illustrated. ABC-CHD ¼ Age, Biomarkers, Clinical Variables–Coronary Heart Disease; LDL ¼ low-density lipoprotein; METs ¼metabolic equivalents;

NT-proBNP ¼ N-terminal pro–B-type natriuretic peptide.

J A C C V O L . 7 0 , N O . 1 4 , 2 0 1 7 Stewart et al.O C T O B E R 3 , 2 0 1 7 : 1 6 8 9 – 7 0 0 Exercise and Coronary Heart Disease Mortality

1695

Results of analyses were similar when absoluteincrease in METs h/week of exercise (Online Figure 1),and natural logarithm–transformed (Online Figure 2)measures of exercise were evaluated (Online Table 1).

A 1-level increase in average exercise intensity(from mild to moderate, or from moderate tovigorous) between subjects was associated with adecrease in total mortality (unadjusted HR: 0.69; 95%

FIGURE 2 Associations Among the Volume of Mild, Moderate, and Vigorous Physical

Activity and CV and Non-CV mortality

36-M

onth

Eve

nt R

ate

(%)

11

10

9

8

7

6

5

4

3

2

1

00 0

to 5>5

to 10>20

to 40>10

to 20>40

to 80>80

to 160>160

n = 570 544 1121 33842382 4286 2592

CV Mortality Non CV Mortality

608

Physical Activity MET.hours/week

Self-reported habitual exercise in metabolic equivalents (METs) h/week. Ten METs

h/week of exercise is equivalent to 5 h of mild intensity exercise, 2.5 h of moderate

intensity exercise, and 1.25 h of vigorous exercise. CV ¼ cardiovascular.

Stewart et al. J A C C V O L . 7 0 , N O . 1 4 , 2 0 1 7

Exercise and Coronary Heart Disease Mortality O C T O B E R 3 , 2 0 1 7 : 1 6 8 9 – 7 0 0

1696

CI: 0.64 to 0.73; adjusted HR: 0.84; 95% CI: 0.78 to0.89), and CV mortality (unadjusted HR: 0.66; 95% CI:0.61 to 0.71; adjusted HR: 0.81; 95% CI: 0.75 to 0.88).

Doubling the duration of exercise was also associ-ated with lower all-cause mortality (unadjusted HR:0.83; 95% CI: 0.80 to 0.86; adjusted HR: 0.90; 95% CI:0.87 to 0.94), and with CV mortality (unadjusted HR:0.85; 95% CI: 0.81 to 0.89; adjusted HR: 0.93; 95% CI:0.89 to 0.98).

Associations between the volume of exercise atthe highest reported intensity and mortality are dis-played in Figure 3. Increase in the amount of mildintensity exercise between subjects was associatedwith mortality in the subgroup who reported takingno moderate or vigorous exercise. In participants whoreported taking no vigorous exercise, more moderateintensity exercise was also associated with lowermortality. Subjects who reported taking any vigorousexercise had lower mortality than those who took novigorous exercise, but there was no clear associationbetween increasing duration of vigorous exercise andeither higher or lower mortality.

SUBGROUP ANALYSIS. Adjusted HR for all-causemortality by increase in total physical activity aredisplayed for subgroups in Figure 4. The associationbetween exercise volume and mortality was similar

across all geographic regions. There was a largerdecrease in mortality with increase in METs h/week ofexercise between subjects in the least active tertile,and between subjects who reported <10 METs h/weekcompared with those reporting $10 METs h/week ofmoderate or vigorous exercise.

Diabetes, significant renal dysfunction, poly-vascular disease, and ABC-CHD risk score above thepopulation median were associated with a greaterreduction in mortality between participants withincrease in habitual physical activity (p forinteraction <0.05 for all). The decrease in mortalityassociated with increase in exercise was also greaterbetween patients who reported limitation of exerciseby dyspnea and similar between patients who did anddid not report limitation because of chest pain ordiscomfort.

RELATIVE IMPORTANCE OF EXERCISE AND OTHER

PROGNOSTIC VARIABLES. Of the covariates evalu-ated, habitual exercise was the second strongest in-dependent predictor of mortality, after history ofcongestive heart failure (Online Figure 3). However,its predictive value for mortality when consideredalone was modest (C-statistic ¼ 0.586; 95% CI: 0.569to 0.604), and it provided a modest incrementalpredictive improvement after considering all othercovariates, (C-statistic for all covariates, excludingphysical activity ¼ 0.746; 95% CI: 0.732 to 0.761;C-statistic including physical activity ¼ 0.750; 95% CI:0.735 to 0.764).

DISCUSSION

In this study of patients with stable CHD, there was agradually lower all-cause, CV, and non-CV mortalityat gradually higher self-reported habitual exercise.These associations were not linear: the differencein mortality was greater for different activities atlower levels of habitual exercise, and less pronouncedfor differences in activity at higher levels of exercise.This suggests important potential benefits ofincreasing habitual exercise in persons with stableCHD who are sedentary.

Analyses from large general population studiesand others also suggest the relationship between theamount of exercise and mortality is nonlinear;greater reductions are associated with modest in-creases in physical activity in sedentary persons, anddecrease progressively at higher levels of exercise(7–9). In a large Taiwanese population study (25),persons who exercised on average for only 15 mineach day had 14% lower mortality and an increase inlife expectancy of 3 years compared with personswho were completely sedentary. Mortality was lower

FIGURE 3 Associations Between CV Mortality and the Amount of Habitual Exercise at the Highest Reported Intensity

Reported by Each Participant

CV M

orta

lity

Risk

1

0.8

0.6

0.4

0.2

00 >10

to 20Mild Moderate Vigorous

0to 10

>20 >10to 20

0to 10

>20 >80>40to 80

>20to 40

>10to 20

0to 10

n = 570 12641281 1210 19821613 3916 466

0.23

0.29

0.25

0.34

0.24

0.33

0.35

0.52

0.69

0.770.73

1

664968738815

Volume of Physical Activity at Highest Reported Intensity (MET.hours/week)

The duration of mild intensity exercise each week is reported for subjects who undertake no moderate or vigorous exercise. Moderate intensity

physical activity is given for patients who had no vigorous physical activity. The amount of vigorous exercise is reported for patients who

reported any habitual vigorous exercise. Ten METs h/week ¼ 5 h of mild intensity exercise, 2.5 h of moderate intensity exercise, and 1.25 h of

vigorous exercise. To evaluate possible associations at high intensities of exercise, vigorous physical activity is also reported for >20 to 40

METs h ($2.5 to 5 h), >40 to 80 METs h (>5 to 10 h), and >80 METs h (>8 h per week). Abbreviations as in Figure 2.

J A C C V O L . 7 0 , N O . 1 4 , 2 0 1 7 Stewart et al.O C T O B E R 3 , 2 0 1 7 : 1 6 8 9 – 7 0 0 Exercise and Coronary Heart Disease Mortality

1697

also in persons who exercised more, but the incre-mental benefit was less. In studies of runners,increase in the speed and duration of running wasnot associated with progressive reductions inmortality (26,27).

Ours is the largest analysis evaluating the dose-response relationship between habitual exerciseand mortality in a global cohort of patients withstable CHD. Subgroup analyses from large cohortstudies suggest associations are likely to be similar forpatients with and without known CV disease, butinformation about the nature and severity of CVdisease from these studies was limited (7–9).

Vigorous exercise may increase CV risk by trig-gering MI (28) or sudden death (29), and this riskcould be greater in patients with stable CHD. Previousstudies in both general (30) and CHD populations(12,13) suggest that mortality may be increased at veryhigh levels of vigorous exercise. A reverse J-shapedassociation between exercise levels and mortality wasreported in patients with stable CHD who participatedin a cardiac rehabilitation program (12). In a largestudy of readers of a U.S. running magazine whoindicated they had had a heart attack, mortality was

lower for runners than for nonrunners, but mortalitywas higher for runners who reported # comparedwith >7.2 METs h/day of vigorous exercise (thehighest category of vigorous exercise, equivalent torunning more than w7.1 km/day) (13). However, inour study long durations of vigorous physical activitymay not have been accurately quantified and the 95%CI for the risk estimates were wide, and therefore amodest increase in mortality associated with highintensities or durations of vigorous exercise cannotbe excluded.

We included evaluation of mild intensity exercise,which contributed relatively more to overall physicalactivity in persons exercising less. For personswho undertook no moderate or vigorous exercise,mild-intensity exercise was associated with lowermortality. Reduced sedentary time and greater timespent standing are also associated with lower mor-tality in general population studies, especially inpersons who are less active (10).

Subgroup analyses suggest the association be-tween increased exercise and mortality is strongerin patients with the highest ABC-CHD risk score,which estimated risk from several clinical risk

FIGURE 4 Association Between Increasing Physical Activity and Mortality in Subgroups Defined by Baseline Characteristics

Lower Risk Higher Risk1.11.0510.950.90.850.80.750.7

*p-valueHR (95%CI)No of

PatientsOutcome: All-Cause MortalityNo of Events

(%/3 yrs)Moderate or vigorous physical activity

Tertile of physical activity

Age

Sex

Geographic region

Smoking

Body mass index

Diabetes Mellitus

Significant renal dysfunction

Prior myocardial infarction

Prior PCI or CABG

Multi-vessel coronary heart disease

Poly-vascular disease

Congestive heart failure

Stability risk score

Limited by Dyspnea

Limited by Fatigue

1.01 (0.99-1.03)9339≥10 MET h/week 484 (4.36)

0.96 (0.94-0.98)8263≥64 years 755 (7.84)

0.98 (0.96-0.99)12148Male 900 (6.32)

0.98 (0.95-1.01)3306Eastern Europe 286 (7.30)0.96 (0.93-0.99)3870North America 250 (5.62)0.96 (0.91-1.01)1133South America 131 (10.33)0.97 (0.94-1.01)3667Western Europe 237 (5.37)

0.97 (0.95-0.98)12214No Current smoker 864 (6.02)

0.97 (0.94-0.99)5421≥30 379 (5.96)

0.99 (0.97-1.01)9052No 545 (5.09)

0.98 (0.96-1.00)10373No 544 (4.42)

0.98 (0.95-1.00)6153No 379 (5.26)

0.96 (0.93-0.99)3706No 348 (8.10)

0.95 (0.91-0.98)2246Yes 210 (8.13)

0.94 (0.91-0.97)2233Yes 266 (10.39)

0.96 (0.93-0.99)3173Yes 391 (10.71)

0.95 (0.94-0.97)6437>Median 783 (10.56)

Limited by Chest0.95 (0.93-0.97)7374Yes 665 (7.75)

0.96 (0.94-0.99)5496Yes 507 (7.95)

0.97 (0.95-0.99)8989Yes 755 (7.21)

0.96 (0.85-1.09)4857Middle tertile 313 (5.46)1.00 (0.98-1.03)4958Most Active 264 (4.47)

0.94 (0.87-1.00)5444<10 MET h/week 585 (9.39) 0.0282

0.98 (0.96-1.01)6633<64 years 326 (4.16) 0.2426

0.94 (0.89-0.98)2748Female 181 (5.63) 0.0895

0.96 (0.91-1.01)2920Asia/Pacific 177 (5.19) 0.7782

0.98 (0.95-1.01)2682Current smoker 217 (6.97) 0.1926

0.97 (0.95-0.99)9475<30 702 (6.33) 0.9356

0.95 (0.92-0.97)5844Yes 536 (7.93) 0.0195

0.95 (0.93-0.97)4523Yes 537 (10.40) 0.0456

0.97 (0.95-0.99)8743Yes 702 (6.84) 0.4918

0.98 (0.96-0.99)11190Yes 733 (5.57) 0.4621

0.97 (0.96-0.99)12650No 871 (5.85)0.4972

0.98 (0.96-1.00)12663No 815 (5.47)0.0530

0.97 (0.96-0.99)11723No 690 (5.00)0.1865

1.02 (0.99-1.05)6447≤Median 151 (1.94) 0.0007

1.00 (0.98-1.03)7405No 405 (4.63) 0.0003

0.98 (0.96-1.00)9243No 563 (5.17) 0.3167

0.98 (0.95-1.00)5718No 308 (4.55) 0.4013

0.80 (0.71-0.91)5081Most Sedentary 504 (8.65) 0.0013

HR and 95% CI for all-cause mortality with increase in mild, moderate, and/or vigorous physical activity of 10 METs h/week, adjusted for randomized treatment, age,

body mass index, sex, systolic blood pressure, hypertension, geographic region for final reporting, prior MI, prior percutaneous coronary intervention (PCI) or coronary

artery bypass graft (CABG), prior multivessel coronary heart disease, baseline diabetes, smoking, polyvascular disease, significant renal dysfunction, hemoglobin,

white blood cell count, low-density lipoprotein cholesterol, high-density lipoprotein cholesterol, triglycerides, estimated glomerular filtration rate according to

CKD-EPI (Chronic Kidney Disease-Epidemiology research group) calculator, and congestive heart failure. *The p values are for test for interaction. Abbreviations as

in Figures 1 and 2.

Stewart et al. J A C C V O L . 7 0 , N O . 1 4 , 2 0 1 7

Exercise and Coronary Heart Disease Mortality O C T O B E R 3 , 2 0 1 7 : 1 6 8 9 – 7 0 0

1698

PERSPECTIVES

COMPETENCY IN PATIENT CARE AND PROCEDURAL

SKILLS: In patients with stable ischemic heart disease, a modest

increase in exercise is associated with reduced CV and all-cause

mortality, particularly in sedentary individuals and those at

highest CV risk. There is no evidence of either harm or benefit

from increasing the duration of vigorous activity or harm from

more exercise in patients limited by angina or dyspnea.

TRANSLATIONAL OUTLOOK: There is a need for simple

strategies that encourage regular exercise in sedentary patients

with stable coronary disease and for randomized trials to confirm

that increasing exercise and reduced mortality are causally

related.

J A C C V O L . 7 0 , N O . 1 4 , 2 0 1 7 Stewart et al.O C T O B E R 3 , 2 0 1 7 : 1 6 8 9 – 7 0 0 Exercise and Coronary Heart Disease Mortality

1699

factors and biomarkers (24). Diabetes, polyvasculardisease, and limitation by dyspnea were each inde-pendently associated with a greater decrease inmortality with increase in exercise, and the associ-ation was similar for persons who did and did notreport limitation of exercise by chest pain ordiscomfort. These observations suggest that personswith more advanced disease and/or limiting symp-toms, who are often also more sedentary, couldhave the most to benefit from increasing habitualexercise.

STUDY LIMITATIONS. Causality cannot be estab-lished from observational studies. We report dif-ferences in mortality between participants whoreported different levels of exercise, but random-ized trials are needed to reliably determine theindependent benefit from increasing habitual exer-cise. Socioeconomic, cultural, and societal factorswere important determinants of the amount ofusual physical activity (13). There were also baselinedifferences in risk factors, measures of disease,comorbidity, and ABC-CHD risk score by habitualphysical activity. Adjustment for all covariatesaccounted for about one-half of the associationbetween exercise and mortality. On the other hand,because exercise has favorable effects on multipleCV risks factors, adjustment for all covariates mayalso underestimate potential benefits fromincreasing exercise. General population studies,where pre-existing disease is less likely to confoundevaluations, have reported a similar dose-responserelationship with long-term mortality (8,25).

Assessment of usual exercise was based on asimple self-reported questionnaire, which is easierto administer than formal exercise testing duringroutine medical consultation and can also be usedto guide recommendations based on reportedlevels of exercise. However, questionnaires aresubjective; patients may overestimate habitualexercise, and estimates of physical activity volumeare likely to be imprecise. These limitations wouldbe expected to result in an underestimate of thetrue strength of association between exercise andmortality.

The study evaluated high-risk clinical trial partici-pants who may have different characteristics than doother patients with stable CHD. However, associa-tions between self-reported exercise and mortalitywere consistent across diverse geographic regionsand by various demographic variables, suggestingresults are probably broadly representative.

Different approaches were explored to describethe nonlinear association between habitual exercise

and outcomes. Doubling of exercise volume, whichcan be achieved by increasing average exerciseintensity by 1 level, doubling the duration of exer-cise, or a combination of both were chosen becauseit is relatively easy for patients and clinicians tounderstand. However, associations were stronglystatistically significant for all measures used todescribe increase in exercise volume, as reported inthe Online Appendix.

CONCLUSIONS

For sedentary persons, smaller amounts of habitualmild or moderate intensity exercise may have sub-stantial health benefits, whereas persons who alreadyundertake regular moderate or vigorous intensityphysical activity may benefit less from increasingexercise. At a population level, the greatest benefitsto health are likely to be achieved by modestincreases in exercise in sedentary persons, especiallyin persons who have a higher risk of adverse events,and those with exertional angina and dyspnea.However, because low physical activity may reflectpoorer health, randomized clinical trials are stillneeded to confirm that increasing physical activityalso reduces mortality.

ACKNOWLEDGMENTS The authors thank all patientswho participated in the STABILITY trial, as well asstudy nurses and investigators at 639 participatingsites.

ADDRESS FOR CORRESPONDENCE: Dr. Ralph A.H.Stewart, Auckland City Hospital, Green Lane Cardio-vascular Services, Park Road, Grafton, Auckland 1142,New Zealand. E-mail: rstewart@adhb.govt.nz.

Stewart et al. J A C C V O L . 7 0 , N O . 1 4 , 2 0 1 7

Exercise and Coronary Heart Disease Mortality O C T O B E R 3 , 2 0 1 7 : 1 6 8 9 – 7 0 0

1700

RE F E RENCE S

1. Perk J, De Backer G, Gohlke H, et al. Europeanguidelines on cardiovascular disease prevention inclinical practice (version 2012): the Fifth Joint TaskForce of the European Society of Cardiology andOther Societies on Cardiovascular Disease Pre-vention in Clinical Practice (constituted by repre-sentatives of nine societies and by invitedexperts). Eur Heart J 2012;33:1635–701.

2. Haskell WL, Lee IM, Pate RR, et al. Physicalactivity and public health: updated recommenda-tion for adults from the American College ofSports Medicine and the American Heart Associa-tion. Circulation 2007;116:1081–93.

3. Eckel RH, Jakicic JM, Ard JD, et al. 2013 AHA/ACC guideline on lifestyle management to reducecardiovascular risk: a report of the AmericanCollege of Cardiology/American Heart AssociationTask Force on Practice Guidelines. J Am CollCardiol 2014;63:2960–84.

4. Montalescot G, Sechtem U, Achenbach S, et al.2013 ESC guidelines on the management of stablecoronary artery disease: the Task Force on themanagement of stable coronary artery disease ofthe European Society of Cardiology. Eur Heart J2013;34:2949–3003.

5. Smith SC Jr., Allen J, Blair SN, et al. AHA/ACCguidelines for secondary prevention for patientswith coronary and other atherosclerotic vasculardisease: 2006 update endorsed by the NationalHeart, Lung, and Blood Institute. J Am Coll Cardiol2006;47:2130–9.

6. Ross R, Blair SN, Arena R, et al. Importanceof assessing cardiorespiratory fitness in clinicalpractice: a case for fitness as a clinical vitalsign: a scientific statement from the AmericanHeart Association. Circulation 2016;134:e653–99.

7. Sattelmair J, Pertman J, Ding EL, et al. Doseresponse between physical activity and risk ofcoronary heart disease: a meta-analysis. Circula-tion 2011;124:789–95.

8. Arem H, Moore SC, Patel A, et al. Leisure timephysical activity and mortality: a detailed pooledanalysis of the dose-response relationship. JAMAIntern Med 2015;175:959–67.

9. Moore SC, Patel AV, Matthews CE, et al. Leisuretime physical activity of moderate to vigorous in-tensity and mortality: a large pooled cohortanalysis. PLoS Med 2012;9:e1001335.

10. Biswas A, Oh PI, Faulkner GE, et al. Sedentarytime and its association with risk for disease

incidence, mortality, and hospitalization in adults:a systematic review and meta-analysis. Ann InternMed 2015;162:123–32.

11. Chau JY, Grunseit AC, Chey T, et al. Daily sittingtime and all-cause mortality: a meta-analysis.PLoS One 2013;8:e80000.

12. Mons U, Hahmann H, Brenner H. A reverseJ-shaped association of leisure time physical activitywith prognosis in patientswith stable coronaryheartdisease: evidence from a large cohort with repeatedmeasurements. Heart 2014;100:1043–9.

13. Williams PT, Thompson PD. Increased cardio-vascular disease mortality associated with exces-sive exercise in heart attack survivors. Mayo ClinProc 2014;89:1187–94.

14. Anderson L, Taylor RS. Cardiac rehabilitationfor people with heart disease: an overview ofCochrane systematic reviews. Cochrane DatabaseSyst Rev 2014;12:CD011273.

15. Abell B, Glasziou P, Hoffmann T. Reporting andreplicating trials of exercise-based cardiac reha-bilitation: do we know what the researchersactually did? Circ Cardiovasc Qual Outcomes 2015;8:187–94.

16. Stewart R, Held C, Brown R, et al. Physicalactivity in patients with stable coronary heartdisease: an international perspective. Eur Heart J2013;34:3286–93.

17. White HD, Held C, Stewart R, et al. Darapladibfor preventing ischemic events in stable coronaryheart disease. N Engl J Med 2014;370:1702–11.

18. White HD, Held C, Stewart RA, et al. Studydesign and rationale for the clinical outcomes ofthe STABILITY Trial (STabilisation of Atheroscle-rotic plaque By Initiation of darapLadIb TherapY)comparing darapladib versus placebo in patientswith clinical coronary heart disease. Am Heart J2010;160:655–61.

19. Vedin O, Hagström E, Stewart R, et al.Secondary prevention and risk factor targetachievement in a global, high-risk population withestablished coronary heart disease: baselineresults from the STABILITY study. Eur J PrevCardiol 2013;20:678–85.

20. Craig CL, Marshall AJ, Sjostrom M, et al.International physical activity questionnaire:12-country reliability and validity. Med Sci SportsExerc 2003;35:1381–95.

21. Ainsworth BE, Haskell WL, Herrmann SD, et al.2011 Compendium of physical activities: a second

update of codes and MET values. Med Sci SportsExerc 2011;43:1575–81.

22. Stone CJ, Koo C. Additive splines in statistics.In: Proceedings of the Statistical Computing Sec-tion. Alexandria, VA: American Statistical Associ-ation, 1985:45–8.

23. Devlin TF, Weeks BJ. Spline functions forlogistic regression modeling. In: Proceedings of the11th Annual SAS Users Group International Con-ference. Cary, NC: SAS Institute Inc., 1986:646–51.

24. Lindholm D, Lindbäck J, Johan MSC, et al.Biomarker-based risk model to predict cardio-vascular mortality in patients with stable coro-nary heart disease. J Am Coll Cardiol 2017;70:813–26.

25. Wen CP, Wai JP, Tsai MK, et al. Minimumamount of physical activity for reduced mortalityand extended life expectancy: a prospectivecohort study. Lancet 2011;378:1244–53.

26. Gebel K, Ding D, Chey T, Stamatakis E,Brown WJ, Bauman AE. Effect of moderate tovigorous physical activity on all-cause mortality inmiddle-aged and older Australians. JAMA InternMed 2015;175:970–7.

27. Lee DC, Pate RR, Lavie CJ, Sui X, Church TS,Blair SN. Leisure-time running reduces all-causeand cardiovascular mortality risk. J Am Coll Car-diol 2014;64:472–81.

28. Mittleman MA, Maclure M, Tofler GH, et al.,for the Determinants of Myocardial InfarctionOnset Study Investigators. Triggering of acutemyocardial infarction by heavy physical exertion:protection against triggering by regular exertion.N Engl J Med 1993;329:1677–83.

29. Albert CM, Mittleman MA, Chae CU, Lee IM,Hennekens CH, Mason JE. Triggering of suddendeath from cardiac causes by vigorous exertion.N Engl J Med 2000;343:1355–61.

30. Schnohr P, O’Keefe JH, Marott JL, Lange P,Jensen GB. Dose of jogging and long-term mor-tality: the Copenhagen City Heart Study. J Am CollCardiol 2015;65:411–9.

KEY WORDS cardiac rehabilitation,coronary artery disease, exercise, physicalactivity

APPENDIX For supplemental figures and atable, please see the online version of thisarticle.