pi is 0002870311008994

7/31/2019 Pi is 0002870311008994

1/14

Coronary Artery Disease

Exercise intervention and inflammatory markers in

coronary artery disease: A meta-analysisWalter Swardfager, PhD, a,b Nathan Herrmann, MD, FRCP(C), a,b Stephen Cornish, BScH, a

Graham Mazereeuw, BSc, a,b Susan Marzolini, MSc, a,b,c Lauren Sham, BSc, a and Krista L. Lanctt, PhD a,b,c

Ontario, Canada

Background Inflammatory activity plays a role in the development and progression of coronary artery disease (CAD),and exercise confers survival benefit. We performed a meta-analysis of changes in inflammatory biomarkers over the course ofexercise interventions in patients with CAD.

Methods We searched MEDLINE, Embase, the Cochrane Collaboration, AMED, and CINAHL for studies reportingperipheral inflammatory biomarker concentrations before and after exercise interventions of2 weeks in patients with CAD.Data were summarized using standard mean differences (SMD) and 95% CIs.

Results Twenty-three studies were included. Concentrations of C-reactive protein (CRP; SMD 0.345, 95% CI 0.444to 0.246, n = 1,466, Pb .001), interleukin 6 (SMD 0.546, 95% CI 0.739 to 0.353, n = 280, Pb .001), fibrinogen(SMD 0.638, 95% CI 0.953 to 0.323, n = 247, Pb .001), and vascular cell adhesion molecule 1 (SMD 0.413, 95%CI 0.778 to 0.048, n = 187, P= .027) were lower postintervention. Higher total cholesterol (B = 0.328, 95% CI 0.612to 0.043, P= .026) and higher total/high-density lipoprotein cholesterol ratios (B = 0.250, 95% CI 0.425 to 0.076,P= .008) at baseline were associated with greater reductions in CRP. In controlled studies, follow-up concentrations of CRP(SMD 0.500, 95% CI 0.844 to 0.157, nexercise/control = 485/284, P = .004), and fibrinogen (SMD 0.544, 95% CI1.058 to 0.030, nexercise/control = 148/100, P= .038) were lower in subjects who exercised compared with controls.

Conclusion Exercise training is associated with reduced inflammatory activity in patients with CAD. C-reactive proteinand fibrinogen have provided the strongest evidence. Higher baseline CRP and adverse baseline lipid profiles predictedgreater reductions in CRP. (Am Heart J 2012;163:666-676.e3.)

BackgroundIn patients with coronary artery disease (CAD),

exercise interventions confer long-term survival benefit

and reduce the risk of recurrent events.1 Roles of immune

activity in the development and progression of athero-

sclerosis are now appreciated.2,3 Monocytes recruited

by vascular endothelial cell signals (eg, vascular cell

adhesion molecule 1 [VCAM-1]) are activated during

plaque formation to produce metalloproteinases, nitric

oxide, and the proinflammatory cytokines tumor necrosis

factor (TNF-) and interferon-.4 T-lymphocytes also

infiltrate the intima where they are activated by oxidized

low-density lipoprotein (LDL) cholesterol to further

produce proinflammatory cytokines.5,6 Evidence sug-

gests that physical activity can decrease inflammation,

thereby attenuating the progression of atherosclerosis.7-9

The clinical utility of circulating proteins as biomarkers

depends on their plasma half-life, stability in the

collected sample, concentration relative to assay sensi-

tivity, and interlaboratory assay reproducibility. Among

inflammatory biomarkers, the acute-phase reactant C-

reactive protein (CRP) is best established as a predictorof cardiovascular disease and mortality,10,11 and a

second, fibrinogen, has shown strong associations with

CAD and with mortality in CAD patients.12,13 Evidence

suggests that VCAM-114 and interleukin (IL) 6 15 can

refine risk stratification, and TNF- has been associated

with CAD in the elderly.16 Conversely, higher concen-

trations of the anti-inflammatory cytokine IL-10 have

been associated with reduced cardiovascular risk.17

Despite evidence from epidemiologic studies,18 exer-

cise interventions may not consistently lower CRP

concentrations across adult populations19; however, in

patients with CAD, studies have reported changes in CRP

From the aSunnybrook Research Institute, Toronto, Ontario, Canada, bUniversity of

Toronto, Toronto, Ontario, Canada, cToronto Rehabilitation Institute, Toronto, Ontario,

Canada.

Submitted November 17, 2011; accepted December 21, 2011.

Reprint requests: Krista L. Lanctt, PhD, Sunnybrook Health Sciences Centre, FG05, 2075

Bayview Avenue, Toronto, Ontario, Canada M4N1J7.

E-mail: [email protected]

0002-8703/$ - see front matter

2012, Mosby, Inc. All rights reserved.

doi:10.1016/j.ahj.2011.12.017
mailto:[email protected]://dx.doi.org/10.1016/j.ahj.2011.12.017http://dx.doi.org/10.1016/j.ahj.2011.12.017mailto:[email protected]

7/31/2019 Pi is 0002870311008994

2/14

and other inflammatory biomarkers.20 This meta-analysis

summarizes these clinical data. Characteristics of the

exercise interventions and of included populations were

examined as possible sources of heterogeneity. Clarifying

these issues may better establish the effectiveness of

exercise intervention as an anti-inflammatory therapy for

patients with CAD.

MethodsData sources

Methodology followed PRISMA guidelines.21 English-language

literature was searched using MEDLINE, Embase, the Cochrane

Collaboration, AMED, and CINAHL up to June 2011 for original

reports of inflammatory biomarker changes after exercise in

patients with CAD (see online Appendix A).

Study selectionInclusion criteria were (1) inflammatory biomarkers mea-

sured before and after aerobic exercise intervention; (2)

intervention 2 weeks; (3) diagnosis of CAD by history of

myocardial infarction, percutaneous coronary intervention,

coronary artery bypass graft, stable angina, or angiographic

confirmation of 50% blockage in at least 1 major coronary

artery; (4) excluded nonischemic heart failure; and (5)

English publications.

Figure 1

Search and selection of articles.

Swardfager et al 667American Heart Journal

Volume 163, Number 4

7/31/2019 Pi is 0002870311008994

3/14

Table I. Baseline population characteristics

Author, year Marker(s) nGender

(% male)Age

(y)BMI

(kg/m2)Vo2 (mL/

kgmin)HDL

(mmol/L)LDL

(mmol/L)

Totalcholesterol

(mmol/L)TG

(mmo

Ades et al,28 2009* CRP 38 79 64 9 32.2 3.7 22 6 1.0 0.2 2.4 0.7 4.2 0.9 1.7 36 83 63 9 32.0 4.5 22 5 1.1 0.3 2.2 0.5 3.8 0.7 1.3

Balen et al,29 2008 TNF-, CRP,fibrinogen

30 70 59 9 28.8 3.8 20.0 4.4 0.7 0.2 3.0 1.0 4.7 1.6 1.8

Beckie et al,30 2010 IL-6, TNF-,CRP, ICAM-1

87 0 62 10 32.0 7.0 21.0 7.0 1.1 0.3 2.4 0.9 4.2 1.0 1.5

Caulin-Glaser et al,31 2005 CRP 172 78 64 11 29.9 5.8 21.7 8.8 1.2 0.4 2.3 0.7 4.4 1.0 1.9 Conraads et al,32 2002 IL-6, TNF- 12 NS 57 NS 16.4 NS NS NS NSDod et al,33 2010 IL-6, TNF-, CRP,

VCAM-1, ICAM-127 52 56 33.3 NS 1.0 0.1 2.5 0.2 4.2 0.2 1.4

Fernandes et al,34 2011 CRP, VCAM-1 15 27 61 7 28.6 5.9 25.0 15.0 1.4 0.3 3.2 0.8 5.2 0.9 1.4 Goldhammer et al,35 2005 IL-6, CRP 28 65 65 7 26.9 4 24.7 1.0 0.1 3.3 0.7 5.0 0.9 1.7 Hansen et al,36 2008 CRP 134 82 63 10 26.7 3.6 18.2 6.4 1.2 0.3 3.1 1.1 4.8 1.3 NSKim et al,372008 IL-6, TNF-, CRP,

fibrinogen29 69 59 2 25.7 0.6 27.6 1.3 1.2 0.1 2.6 0.1 4.4 0.2 1.5

Lavie et al,38 2009 CRP 393 77 65 10 30.4 4.3 16.8 5.2 1.0 0.3 2.6 1.0 4.3 1.0 1.8 Luk et al,39 2011 CRP 32 75 68 9 24.7 2.4 NS 1.2 0.3 2.1 0.5 3.9 0.6 1.6 Milani et al,40 2004 CRP 235 71 67 11 27.9 4.9 16.6 5.1 1.1 0.4 2.7 1.1 4.5 1.0 1.7 Onishi etal,41 2009 CRP 32 91 66 10 25.0 2.7 14.2 3.7 1.1 0.3 3.2 0.7 5.1 0.9 2.0 Pluss et al,42 2008 CRP, fibrinogen 111 77 63 7 26.4 3.7 NS 1.2 0.3 3.0 1.0 4.9 1.1 1.6 Rankovic et al,43 2009 CRP, VCAM-1, ICAM-1, 22 55 63 7 29.3 3.2 NS NS NS NS NSSchumacher et al,44 2006 IL-6, TNF-, CRP,

VCAM-1, ICAM-195 83 54 8 27.0 4.0 NS 1.0 0.3 3.0 0.8 4.8 0.8 1.7

Shin et al,45 2006 IL-6, CRP 14 58 61 3 25.8 0.7 27.6 1.8 1.2 0.1 2.7 0.2 4.5 0.2 1.5 Sixt et al,46 2008 CRP, fibrinogen 13 77 64 6 29.2 4.3 21.5 6.0 1.4 0.3 3.2 1.5 NS 1.5 Sixt et al,472010 CRP, VCAM-1, ICAM-1 11 91 NS 32.1 21.5 1.2 0.4 2.5 0.8 4.5 0.9 2.0 Suzuki et al,48 1992 Fibrinogen 56 88 59 11 NS NS NS NS NS NSWalther et al, 49 2008 IL-6, TNF-, CRP,

fibrinogen34 100 61 1 27.2 0.4 23.3 0.6 1.2 0.1 2.7 0.2 5.2 0.2 2.2

Wosornu et al,50 1992 Fibrinogen 35 100 59 7 NS NS NS NS NS NS

NS, Not stated; TG, triglyceride.Prospectively randomized into high (top) and low (bottom) intensity groups. Estimated from treadmill time. Estimated from Watts.

7/31/2019 Pi is 0002870311008994

4/14

Table II. Exercise intervention characteristics of included studies

Author, yearDuration

(wk) Frequency Intervention Intensity Session time

(min) M

Ades et al,28 2009 20 Group 1: 5-7 Group 2: 3

ExerciseDiet

Behavioral counseling

Group 1: 50%-60% VO2peakGroup 2: 65%-70% VO2peak

Group 1: 45-60Group 2: 25-40

GrGro

Balen et al,29 2008 3 5 Exercise 50%-60% V O2peak 45 (C) and 30 (W)

Beckie et al,

30

2010 12 3 ExercisePsychosocial 60%-80% HRmax 35-45 C

Caulin-Glaser et al,31 2005 12 3 ExerciseDiet

Behavioral counseling

NS NS

Conraads et al,32 2002 16 3 Exercise 90% VAT 20

Dod et al,33 2010 12 NS ExerciseDiet

Stress managementSmoking cessation,

Psychosocial counseling

NS Goal: 180/wk

Fernandes et al,34 2011 16 3 Exercise HR between VAT and respiratory

compensation point

40

Goldhammer et al,35 2005 12 3 Exercise 70%-80% HRmax 45 C

Hansen et al,36 2008 7 3 ExerciseDiet counseling

65% VO2peak Group 1: 40Group 2: 60

C

Kim et al,372008 14 7 Exercise 50%-85% V O2peak 30-40 Lavie et al,38 2009 12 3 Supervised

+ 1-3 HomeExercise

Hypertension counselingDiabetes counselingSmoking cessation

Diet counseling

VAT 30-40 C

Luk et al,39 2011 8 3 Exercise 80% HRmax 50

7/31/2019 Pi is 0002870311008994

5/14

Table II (continued)

Author, yearDuration

(wk) Frequency Intervention Intensity Session time

(min) M

Milani et al,40 2004 12 3 Supervised+ 1-3 Home

ExerciseDietary management

Weight loss diet

(overweight pts)Health education

VAT 30-40 C

Onishi et al,41 2009 24 1-2 ExerciseDiet

Health education

VAT 60

Pluss et al,42 2008 12 2 ExerciseDiet

Smoking cessationPsychosocial management

Health education

NS 45

Rankovic et al,43 2009 6 3 Exercise 70%-80% HRmax 45 Schumacher et al,44 2006 6 2 Exercise

Hypocaloric dietSubjective ratingsomewhat hard

20

Shin et al,45 2006 14 3 Exercise 50%-85% V O2peak 30-40 Sixt et al,46 2008 4 7 Exercise 70% HRmax Week 1: 15 6

Week 2: 30

Sixt et al,472010 4 5 Supervised+ 2 Home

Exercise 80% HRmax 1530

Suzuki et al,48 1992 4 6 Exercise HR b 120 beat/min75% HRmax

80

Walther et al,49 2008 2 7 Cycle+ 1 Walk

Exercise 70% HRmax 20/60

Wosornu et al,50 1992 24 3 Exercise NS 12-60

NS, Not stated; C, cycling; W, walking; J, jogging; O, other; AT, aerobic; RT, resistance; VAT, ventilatory anaerobic threshold; IT, interval training; CT, continuous training; ID, insufficient d

7/31/2019 Pi is 0002870311008994

6/14

Data extractionStudy eligibility was determined, and data (ie, pre- and

postexercise biomarker concentrations mean SD, population

characteristics, cardiac risk factors, and exercise intervention

details, reporting quality and risk of bias items) were extracted

into a preformatted spreadsheet by 2 raters, and discrepancies

Figure 2

Biomarker changes pre- to postintervention. Gray square areas are proportional to weighting. Diamonds indicate SMD and 95% CI. Significanceof overall effects: CRP (Z = 6.85, Pb .001), IL-6 (Z = 5.54, Pb .001), fibrinogen (Z = 3.97, Pb .001), TNF- (Z = 1.59, P= .112), ICAM-1 (Z =1.91, P= .056), VCAM-1 (Z = 2.22, P= .027).


http://image%20of/

7/31/2019 Pi is 0002870311008994

7/14

were resolved by consensus. Missing biomarker data were

requested from corresponding authors. Mean weekly energy

expenditure was estimated by multiplying the prescribed

exercise intensity (oxygen uptake; 3.5 mLkg1min1 =

0.0733 kJ) by body mass and exercise duration22. Meta-

analyses were conducted for biomarkers reported in at least

3 studies.

Statistical analysesStandardized mean differences (SMD) and 95% CIs were

calculated using random-effects models.23 An SMD was chosen

because variability in absolute biomarker concentrations is

expected between assays from different laboratories.24

Q statistics and I indices were calculated to investigate

heterogeneity and inconsistency, respectively.25 Heterogeneity

in CRP changes was explored by dividing studies into subgroups

by session duration and sessions per week and by investigating

associations between SMD and population characteristics,

baseline CRP concentrations, cardiac risk factors, and interven-tion details by inverse variance-weighted metaregression

analyses (coefficient, B, with 95% CI reported).

Risk of publication bias was assessed using funnel plots and

Egger test.26,27 Analyses were conducted using Stata (Release

10.1; StataCorp, College Station, TX).

No extramural funding supported this work. The authors

are solely responsible for study design, conduct and analyses,

and drafting and editing the final manuscript.

Results

Population characteristicsTwenty-three studies met inclusion criteria (Figure 1)

ranging in size from 12 to 393 subjects (Table I).

Subjects (73.0% male, mean age 63.1 7.6 years) had a

20.1% prevalence of diabetes (antidiabetic use inconsis-

tently reported), and 73.1% were prescribed a choles-

terol-lowering medication. Mean exercise prescriptions

were 41.5 15.8 (20-80) minutes, 3.9 1.7 (2-7)

sessions per week, for 11.3 5.3 (2-24) weeks with

estimated energy expenditures ranging from 1,918 to

14,654 kJ per week (Table II). Improvements in VO2Peak,

lipid profiles, and body mass index (BMI) were

observed (Table III).

Baseline to postintervention comparisonsBaseline and follow-up data were combined from 20 CRP

(n= 1,466), 8 IL-6(n = 280), 6 TNF- (n = 249), 6 fibrinogen

(n = 247), 5 intracellular adhesion molecule 1 (ICAM-1; n =

193), and 5 VCAM-1 (n = 187) studies. Concentrations ofCRP, fibrinogen, IL-6, and VCAM-1 were significantly lower

postintervention than at baseline (Figure 2).

Qualitatively, in 2 studies each, concentrations of IL-10

were higher,29,35 and concentrations of the soluble TNF-

receptor 1 were lower29,32 postintervention.

Controlled comparisonsPostintervention biomarker concentrations were com-

bined from 9 CRP (n exercising [ne]/n controls [nc] =

485/284), 4 VCAM-1 (ne/nc = 143/155), 4 TNF- (ne/nc =

166/152), 3 IL-6 (ne/nc = 136/122), 5 fibrinogen (ne/nc =

148/100), and 3 ICAM-1 (ne/nc = 128/136) studies. Mean

follow-up concentrations of CRP and fibrinogen were

lower in exercising subjects compared with controls, but

differences in TNF-, VCAM-1, IL-6, and ICAM-1 were not

significant (Figure 3).

Risk of biasFunnel plots and Egger tests did not reveal significant

risk of publication bias. The scope of unpublished

observations could not be ascertained because exercise

studies are infrequently registered; however, most studies

reported at least 1 nonsignificant comparison, suggesting

reduced risk of selective reporting. Attrition rates (0%-

40%) and adverse events (0%-7.9%) were reported in 16and 6 studies, respectively. Among controlled trials, 7

specified prospective randomized designs. Control and

intervention groups were similar in most important

characteristics in most studies; however, other indicators

of potential bias including randomization procedures and

blinding of biomarker assessment to group allocation

were not frequently reported (see online Appendix B).

Investigations of heterogeneity in CRP changesExercise intervention characteristics. A subgroup

of 13 studies that prescribed b4 sessions per week

showed a decrease in CRP concentrations (SMD 0.316,

Table III. Pooled cardiac outcome measures for included studies

Baseline Final Change

Studies (n) Mean Studies (n) Mean SMD (95% CI) df Z P

BMI (kg/m2) 18 (1588) 28.4 17 (1449) 27.9 0.111 (0.185 to 0.037) 16 2.96 .003HDL cholesterol (mmol/L) 19 (1566) 1.08 17 (1453) 1.13 0.249 (0.003 to 0.503) 16 1.93 .053LDL cholesterol (mmol/L) 19 (1566) 2.60 17 (1453) 2.40 0.411 (0.648 to 0.175) 16 3.41 .001Total cholesterol (mmol/L) 18 (1553) 4.38 16 (1440) 4.17 0.417 (0.608 to 0.226) 15 4.28 b.001Total/HDL cholesterol ratio 18 (1553) 4.03 16 (1440) 3.63 0.478 (0.695 to 0.262) 15 4.33 b.001Triglycerides (mmol/L) 18 (1432) 1.67 16 (1319) 1.50 0.217 (0.357 to 0.077) 15 3.03 .002VO2Peak (mL/kgmin) 14 (1274) 18.5 14 (1271) 21.5 0.865 (0.583 to 1.146) 13 6.03 b.001

In random-effects meta-analysis.

672 Swardfager et alAmerican Heart Journal

April 2012

7/31/2019 Pi is 0002870311008994

8/14

95% CI 0.422 to 0.210, P b .001) as did 7 studies

prescribing 4 sessions per week (SMD 0.453, 95% CI

0.698 to 0.207, Pb .001). A subgroup of 7 studies that

prescribed b40 minutes per session showed a decrease in

CRP (SMD 0.321, 95% CI 0.444 to 0.249, Pb .001) as

did 12 studies that prescribed 40 minutes per session

(SMD 0.392, 95% CI 0.575 to 0.066, Pb .001). Length

of study was not associated with decrease in CRP (B =

0.004, P = .672, df = 20). Among studies from which

mean weekly energy expenditure could be estimated, no

association was observed with CRP changes (B = 0.013,

P= .894, df= 15).

Population characteristics. Percentage male (B =

0.058, P = .839, df= 20), percentage with diabetes (B =

0.002, P = .545, df = 15), and proportion using a

cholesterol-lowering medication (B = 0.002,P= .521,df=

19) were not associated with CRP changes by metaregres-

sion, although there was a trend toward older populations

showing smaller reductions in CRP (B = 0.026, 95% CI

0.004 to 0.055, P= .082, df= 19). Higher baseline CRP

concentrations predicted greater reductions in CRP (B =

0.035, 95% CI 0.067 to 0.003, P= .032, df= 19).Cardiac risk factors. Higher baseline total cholesterol

concentrations were associated with greater decreases in

Figure 3

Follow-up biomarker concentrations in exercising and control subjects. Gray square areas are proportional to weighting. Diamonds indicate SMDand 95% CI. Significance of overall effects: CRP (Z = 2.85, P= .004), TNF- (Z = 0.18, P= .858), VCAM-1 (Z = 1.41, P= .158), IL-6 (Z = 1.20, P=.231), ICAM-1 (Z = 1.24, P= .216), fibrinogen (Z = 2.07, P= .038).


http://image%20of/

7/31/2019 Pi is 0002870311008994

9/14

CRP (B = 0.328, 95% CI 0.613 to 0.043, P= .026, df=

18; see online Appendix C). This remained significant

when controlling for baseline CRP (P= .044, df= 18) and

dietary cointervention (P = .042, df = 18), and a trend

remained when controlling for the proportion of patients

using a cholesterol-lowering medication (P = .066, df =17). Greater reductions in CRP were associated with

greater decreases in total cholesterol concentrations (B =

0.618, 95% CI 0.111-1.125, P= .020, df= 15).

Greater decreases in CRP were associated with higher

baseline LDL concentrations (B = 0.306, 95% CI 0.611

to 0.002, P= .049, df= 19; see online Appendix C). This

was not attenuated by dietary cointervention (P = .043,

df = 19). A trend remained when controlling for the

proportion of patients using a cholesterol-lowering

medication (P= .075, df= 18) but not when controlling

for baseline CRP (P= .148, df= 19).

Higher baseline mean total/high-density lipoprotein

(HDL) cholesterol ratios were associated with greater

reductions in CRP (B = 0.250, 95% CI 0.425 to 0.076,

P= .008, df= 18; see online Appendix C). This was not

attenuated by the proportion of subjects using a

cholesterol-lowering medication (P = .006, df = 17) or

dietary cointervention (P = .017, df = 18), and there

remained a trend when controlling for baseline CRP (P=

.066, df= 18).

Regression analyses did not reveal significant associa-

tions between CRP changes and baseline VO2Peak (B =

0.054, P= .931, df= 12) or triglycerides (B = 0.179, P=

.542, df = 18), although populations with higher mean

initial BMI trended toward smaller reductions in CRP (B =0.039, 95% CI 0.003 to 0.081, P= .069, df= 20).

DiscussionCollectively, the evidence supports a reduction in

inflammatory activity associated with exercise training in

patients with CAD as indicated by lower CRP, fibrinogen,

IL-6, and VCAM-1 after intervention. Associations be-

tween these biomarkers and risk of mortality10-12,14,15

emphasize the potential significance of these findings.

Controlled studies strengthened this evidence, showing

lower final concentrations of CRP and fibrinogen in those

who undertook exercise compared with controls.The reduction in CRP postintervention in patients with

CAD may contrast findings from other populations19

because of more pronounced effects in those with CAD,

significant vascular risk factors, and/or the presence of

higher inflammation at baseline; exploration of heteroge-

neity suggested that elevated CRP and adverse lipid

profiles were associated with greater CRP reductions.

Although LDL cholesterol, particularly when oxidized, can

promote inflammation,51 HDL may have anti-inflammatory

effects.52 In the pooled studies, there were favorable

changes in lipid profiles with exercise, consistent with

modest increases in HDL cholesterol concentrations22 or

reduced total/HDL cholesterol ratios53 observed with

exercise in medically healthy populations.

Although specific inflammatory processes are not

targeted clinically in CAD,54 antithrombotic doses of

acetylsalicylic acid are recommended,55 and the benefits

of -hydroxy--methylglutaryl-CoA reductase inhibitorsand peroxisome proliferator-activated receptor- agonists

may be caused partly by anti-inflammatory effects.56,57

For instance, improvement in vascular endothelial

function in patients without diabetes treated with

rosiglitazone has been associated with lowering of

CRP.58 In the present study, the proportion of patients

with diabetes, or that using a cholesterol-lowering

medication, did not contribute significantly to heteroge-

neity in CRP outcomes.

Exercise intervention characteristics associated with

decreases in CRP could not be identified, and effects of

exercise intensity and modality were investigated infre-

quently in included studies.28,36,50 Guidelines suggest

broad ranges of intensities (40%-80% of exercise capacity),

durations (20-60 minutes per session), and frequencies (4-

7 days per week)59 for cardiac benefit, but there remains

little guidance for targeting inflammatory activity.

This report was limited to the inflammatory bio-

markers searched and to those that could be meta-

analyzed, which did not encompass all biomarkers

potentially responsive to exercise. The observed

biomarker changes may have been influenced by

lifestyle, diet, medical management, recovery from

coronary events, and heterogeneity in the included

populations at baseline. In controlled studies, lack ofreporting of randomization procedures and adherence

detracted from the quality of evidence. While speaking

to effectiveness, the use of prescription characteris-

tics rather than actual exercise performed in investiga-

tions of heterogeneity, in addition to large ranges of

durations and intensities within studies relative to those

between studies, may have obscured dose effects.

AcknowledgementsThe authors thank Drs Anita Shumacher, Sanja Balen,

and Kari Peersen for their valued correspondence and

Maureen Pakosh, BA, MISt, for information resourcessupport.

References1. Alter DA, Oh PI, Chong A. Relationship between cardiac rehabili-

tation and survival after acute cardiac hospitalization within auniversal health care system. Eur J Cardiovasc Prev Rehabil 2009;16:102-13.

2. Hansson GK, Libby P. The immune response in atherosclerosis: adouble-edged sword. Nat Rev Immunol 2006;6:508-19.

3. Libby P, Ridker PM, Hansson GK. Inflammation in atherosclerosis:from pathophysiology to practice. J Am Coll Cardiol 2009;54:2129-38.


April 2012

7/31/2019 Pi is 0002870311008994

10/14

4. Janeway Jr CA, Medzhitov R. Innate immune recognition. Annu RevImmunol 2002;20:197-216.

5. Stemme S, Faber B, Holm J, et al. T lymphocytes from humanatherosclerotic plaques recognize oxidized low density lipoprotein.Proc Natl Acad Sci U S A 1995;92:3893-7.

6. Frostegard J, Ulfgren AK, Nyberg P, et al. Cytokine expression inadvanced human atherosclerotic plaques: dominance of pro-inflammatory (Th1) and macrophage-stimulating cytokines. Athero-sclerosis 1999;145:33-43.

7. Pynn M, Schafer K, Konstantinides S, et al. Exercise training reducesneointimal growth and stabilizes vascular lesions developing afterinjury in apolipoprotein e-deficient mice. Circulation 2004;109:386-92.

8. Fukao K, Shimada K, Naito H, et al. Voluntary exercise amelioratesthe progression of atherosclerotic lesion formation via anti-inflam-matory effects in apolipoprotein E-deficient mice. J AtherosclerThromb 2010;17:1226-36.

9. Niessner A, Richter B, Penka M, et al. Endurance training reducescirculating inflammatory markers in persons at risk of coronaryevents: impact on plaque stabilization? Atherosclerosis 2006;186:160-5.

10. Kaptoge S, Di Angelantonio E, Lowe G, et al. C-reactive proteinconcentration and risk of coronary heart disease, stroke, andmortality: an individual participant meta-analysis. Lancet 2010;375:132-40.

11. Ridker PM, Paynter NP, Rifai N, et al. C-reactive protein and parentalhistory improve global cardiovascular risk prediction: the ReynoldsRisk Score for men. Circulation 2008;118:2243-51 [4p following2251].

12. Benderly M, Graff E, Reicher-Reiss H, et al. Fibrinogen is a predictorof mortality in coronary heart disease patients. The BezafibrateInfarction Prevention (BIP) study group. Arterioscler Thromb Vasc Biol1996;16:351-6.

13. Danesh J, Lewington S, Thompson SG, et al. Plasma fibrinogen leveland the risk of major cardiovascular diseases and nonvascularmortality: an individual participant meta-analysis. JAMA 2005;294:1799-809.

14. Blankenberg S, Rupprecht HJ,Bickel C, et al. Circulating cell adhesionmolecules and death in patients with coronary artery disease.Circulation 2001;104:1336-42.

15. Lindmark E, Diderholm E, Wallentin L, et al. Relationship betweeninterleukin 6 and mortality in patients with unstable coronary arterydisease: effects of an early invasive or noninvasive strategy. JAMA2001;286:2107-13.

16. Bruunsgaard H, Skinhoj P, Pedersen AN, et al. Ageing, tumournecrosis factor-alpha (TNF-alpha) and atherosclerosis. Clin ExpImmunol 2000;121:255-60.

17. Oemrawsingh RM, Lenderink T, Akkerhuis KM, et al. Multimarker risk

model containing troponin-T, interleukin 10, myeloperoxidase andplacental growth factor predicts long-term cardiovascular risk afternon-ST-segment elevation acute coronary syndrome. Heart 2011;97:1061-6.

18. Kasapis C, Thompson PD. The effects of physical activity on serum C-reactive protein and inflammatory markers: a systematic review. J AmColl Cardiol 2005;45:1563-9.

19. Kelley GA, Kelley KS. Effects of aerobic exercise on C-reactiveprotein, body composition, and maximum oxygen consumption inadults: a meta-analysis of randomized controlled trials. Metabolism2006;55:1500-7.

20. Lavie CJ, Church TS, Milani RV, et al. Impact of physical activity,cardiorespiratory fitness, and exercise training on markers ofinflammation. J Cardiopulm Rehabil Prev 2011;31:137-45.

21. Liberati A, Altman DG, Tetzlaff J, et al. The PRISMA statement forreporting systematic reviews and meta-analyses of studies thatevaluate health care interventions: explanation and elaboration. PLoSMed 2009;6:e1000100.

22. Kodama S, Tanaka S, Saito K, et al. Effect of aerobic exercise training

on serum levels of high-density lipoprotein cholesterol: a meta-analysis. Arch Intern Med 2007;167:999-1008.23. Harris RJ, Bradburn MJ, Deeks JJ, et al. Metan: fixed- and random-

effects meta-analysis. Stata J 2008;8:3-28.24. Noble JE, Wang L, Cerasoli E, et al. An international comparability

study to determine the sources of uncertainty associated with a non-competitive sandwich fluorescent ELISA. Clin Chem Lab Med 2008;46:1033-45.

25. Higgins JP, Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat Med 2002;21:1539-58.

26. Egger M, Davey Smith G, Schneider M, et al. Bias in meta-analysisdetected by a simple, graphical test. Bmj 1997;315:629-34.

27. Begg CB, Mazumdar M. Operating characteristics of a rankcorrelation test for publication bias. Biometrics 1994;50:1088-101.

28. Ades PA, Savage PD, Toth MJ, et al. High-calorie-expenditureexercise: a new approach to cardiac rehabilitation for overweightcoronary patients. Circulation 2009;119:2671-8.

29. Balen S, Vukelic-Damijani N, Persic V, et al.Anti-inflammatoryeffectsof exercise training in the early period after myocardial infarction.Coll Antropol 2008;32:285-91.

30. Beckie TM, Beckstead JW, Groer MW. The influence of cardiacrehabilitation on inflammation and metabolic syndrome in womenwith coronary heart disease. J Cardiovasc Nurs 2010;25:52-60.

31. Caulin-Glaser T, Falko J, Hindman L, et al. Cardiac rehabilitation isassociated with an improvement in C-reactive protein levels in bothmen and women with cardiovascular disease. J Cardiopulm Rehabil2005;25:332-6 [quiz 337-8].

32. Conraads VM, Beckers P, Bosmans J, et al. Combined enduran-

ce/resistance training reduces plasma TNF-alpha receptor levels inpatients with chronic heart failure and coronary artery disease. EurHeart J 2002;23:1854-60.

33. Dod HS, Bhardwaj R, Sajja V, et al. Effect of intensive lifestylechanges on endothelial function and on inflammatory markers ofatherosclerosis. Am J Cardiol 2010;105:362-7.

34. Lara Fernandes J, Serrano Jr CV, Toledo F, et al. Acute and chroniceffects of exercise on inflammatory markers and B-type natriureticpeptide in patients with coronary artery disease. Clin Res Cardiol2011;100:77-84.

35. Goldhammer E, Tanchilevitch A, Maor I, et al. Exercise trainingmodulates cytokines activity in coronary heart disease patients. Int JCardiol 2005;100:93-9.

36. Hansen D, Dendale P, Berger J, et al. Importance of exercise training

session duration in the rehabilitation of coronary artery diseasepatients. Eur J Cardiovasc Prev Rehabil 2008;15:453-9.

37. Kim YJ, Shin YO, Bae JS, et al. Beneficial effects of cardiacrehabilitation and exercise after percutaneous coronary interventionon hsCRP and inflammatory cytokines in CAD patients. Pflugers Arch2008;455:1081-8.

38. Lavie CJ, Milani RV, Artham SM, et al. The obesity paradox, weightloss, and coronary disease. Am J Med 2009;122:1106-14.

39. Luk TH, Dai YL, Siu CW, et al. Effect of exercise training on vascularendothelial function in patients with stable coronary artery disease: arandomized controlled trial. Eur J Cardiovasc Prev Rehabil 2011.

40. Milani RV, Lavie CJ, Mehra MR. Reduction in C-reactive proteinthrough cardiac rehabilitation and exercise training. J Am CollCardiol 2004;43:1056-61.



7/31/2019 Pi is 0002870311008994

11/14

41. Onishi T, Shimada K, Sunayama S, et al. Effects of cardiacrehabilitation in patients with metabolic syndrome after coronaryartery bypass grafting. J Cardiol 2009;53:381-7.

42. Pluss CE, Karlsson MR, Wallen NH, et al. Effects of an expandedcardiac rehabilitation programme in patients treated for an acute

myocardial infarction or a coronary artery by-pass graft operation.Clin Rehabil 2008;22:306-18.43. Rankovic G, Milicic B, Savic T, et al. Effects of physical exercise on

inflammatory parameters and risk for repeated acute coronarysyndrome in patients with ischemic heart disease. Vojnosanit Pregl2009;66:44-8.

44. Schumacher A, Peersen K, Sommervoll L, et al. Physical performanceis associated with markers of vascular inflammation in patients withcoronary heart disease. Eur J Cardiovasc Prev Rehabil 2006;13:356-62.

45. Shin YO, Bae JS, Lee JB, et al. Effect of cardiac rehabilitation andstatin treatment on anti-HSP antibody titers in patients with coronaryartery disease after percutaneous coronary intervention. Int Heart J2006;47:671-82.

46. Sixt S, Rastan A, Desch S, et al. Exercise training but not rosiglitazoneimproves endothelial function in prediabetic patients with coronarydisease. Eur J Cardiovasc Prev Rehabil 2008;15:473-8.

47. Sixt S, Beer S, Bluher M, et al. Long- but not short-term multifactorialintervention with focus on exercise training improves coronaryendothelial dysfunction in diabetes mellitus type 2 and coronaryartery disease. Eur Heart J 2010;31:112-9.

48. Suzuki T, Yamauchi K, Yamada Y, et al. Blood coagulability andfibrinolytic activity before and after physical training during therecovery phase of acute myocardial infarction. Clin Cardiol 1992;15:358-64.

49. Walther C, Mobius-Winkler S, Linke A, et al. Regular exercisetraining compared with percutaneous intervention leads to areduction of inflammatory markers and cardiovascular events inpatients with coronary artery disease. Eur J Cardiovasc Prev Rehabil2008;15:107-12.

50. Wosornu D, Allardyce W, Ballantyne D, et al. Influence of power andaerobic exercise training on haemostatic factors after coronary arterysurgery. Br Heart J 1992;68:181-6.

51. Cybulsky MI, Gimbrone Jr MA. Endothelial expression of amononuclear leukocyte adhesion molecule during atherogenesis.

Science 1991;251:788-91.52. Wadham C, Albanese N, Roberts J, et al. High-density lipoproteinsneutralize C-reactive protein proinflammatory activity. Circulation2004;109:2116-22.

53. Kelley GA, Kelley KS, Roberts S, et al. Efficacy of aerobic exerciseand a prudent diet for improving selected lipids and lipoproteinsin adults: a Meta-Analysis of Randomized Controlled Trials. BMCMed 2011;9:74.

54. Gullestad L, Aukrust P. Review of trials in chronic heart failureshowing broad-spectrum anti-inflammatory approaches. Am JCardiol 2005;95:17C-23C [discussion 38C-40C].

55. Aspirin for the prevention of cardiovascular disease: U.S. PreventiveServices Task Force recommendation statement. Ann Intern Med2009;150:396-404.

56. Morrow DA, de Lemos JA, Sabatine MS, et al. Clinical relevance ofC-reactive protein during follow-up of patients with acute coronarysyndromes in the Aggrastat-to-Zocor Trial. Circulation 2006;114:281-8.

57. Zhao Y, He X, Huang C, et al. The impacts of thiazolidinediones oncirculating C-reactive protein levels in different diseases: a meta-analysis. Diabetes Res Clin Pract 2010;90:279-87.

58. Wang TD, Chen WJ, Cheng WC, et al. Relation ofimprovement in endothelium-dependent flow-mediated vasodila-tion after rosiglitazone to changes in asymmetric dimethylargi-nine, endothelin-1, and C-reactive protein in nondiabeticpatients with the metabolic syndrome. Am J Cardiol 2006;98:1057-62.

59. Medicine ACoS. American College of Sports Medicine's guidelinesfor exercise testing and prescription. Philadelphia: LippincottWilliams and Wilkins; 2009.


April 2012

7/31/2019 Pi is 0002870311008994

12/14

Supplementary Appendix A. Samplesearch strategy (OVID searchingMEDLINE)

1. [Concept 1: Exercise]2. exp Exercise/3. exercise.af.4. Physical Fitness/5. fitness.af.6. training.af.7. aerobic.af.8. cardiac rehab.af.9. or/2-8

10. [Concept 2: Biomarkers]11. exp Biological Markers/12. exp Inflammation/13. inflammatory biomarker.af.14. (inflammat adj3 biomarker).af.

15. exp Cytokines/16. cytokine.af.17. interferon.af.18. interleukin.af.19. chemokine.af.20. monokine.af.21. exp Adipokines/22. adipokine.af.23. exp Acute-Phase Proteins/24. acute phase protein.af.25. C-Reactive Protein/26. c?reactive protein.af.27. CRP.af.

28. IL-

.af.29. fibrinogen.af.30. tumor necrosis factor.af.31. necrosis.af.32. factor.af.33. TNF.af.

34. exp Cell Adhesion Molecules/35. intracellular adhesion molecule.af.36. ICAM.af.37. vascular cell adhesion molecule.af.38. VCAM.af.

39. Platelet Activating Factor/40. PAF.af.41. monocyte chemoattractant protein.af.42. MCP.af.43. myeloid?related protein.af.44. MRP.af.45. myeloperoxidase.af.46. Lysophosphatidylcholines/47. Peroxidase/48. exp Phospholipases/49. lipoprotein phospholipase.af.50. Lp?PLA.af.51. pentraxin.af.

52. PTX

.af.53. exp Matrix Metalloproteinases/54. matrix metalloproteinase.af.55. MMP.af.56. exp Transforming Growth Factors/57. transforming growth factor.af.58. TGF.af.59. Or/11-5860. [Concept 3: Patient Group]61. Coronary Artery Disease/62. coronary artery disease.af.63. exp Heart Diseases/64. heart disease.af.65. or/61-6466. 9 and 59 [Exercise & Biomarkers]67. 66 and 65 [Patient Group &

Biomarkers & Exercise]68. limit 67 to English language69. manual reference search of retrieved articles

Swardfager et al 676.e1American Heart Journal


7/31/2019 Pi is 0002870311008994

13/14

Supplementary Appendix B. Study reporting quality and risk of bias assessment

General risk of bias items Controlled t

Populationrepresentative

Prospectivedesign

Interventionadequately

described

Outcomesreported

objectively

Follow-uprate

reported/adequate

Cointerventionsdocumented Randomized

Adequatesequence

generationAssessments

blinded

Balen, 2008 + + + + + + + + ? Conraads,

2002 ? + + ? + ?

Fernandes,2011

+ + + + + + ? ?

Kim, 2008 + + + + + + ? ? ?Luk, 2011 + + + + + + + + + Milani, 2004 + + + ? + ?Rankovic,

2009+ + + + ? + ? ? ?

Schumacher,2006

+ + + + + + + ? +

Sixt, 2008 ? + + + + + ? + Sixt, 2010 ? + + + + + + ? ? Suzuki, 1992 + + + + + + ? Wosornu,

1992+ + + + + ? + ? ?

Ades, 2009 ? + + + + +Beckie, 2010 + + + + +CaulinGlaser,

2005+ + ? +

Dod, 2010 ? + + + + +Goldhammer,

2005+ + + ? + +

Hansen,2008

+ + + + + +

Lavie, 2009 ? + + ? +

Onishi, 2009 ? + + + ? +Pluss, 2008 + + ? + + +Shin, 2006 + + + + +Walther,

2008+ + + + + +

+ indicates yes; , no; ?, uncertain.Study quality was assessed using items from the Newcastle Ottawa Scale and the Cochrane Collaboration's risk of bias assessment tool, addressing key methodological criteria relevant to

7/31/2019 Pi is 0002870311008994

14/14

Supplementary Appendix C.Investigation of heterogeneity inmetaregression analyses

Metaregression analyses of CRP changes against(A) total cholesterol,(B) LDL cholesterol, and (C) ratio of total/HDL cholesterol.

Swardfager et al 676.e3American Heart Journal

http://image%20of/

pi is 0002870311008994

Documents