==== Front ARYA AtherosclerARYA AtherosclerARYAARYA Atherosclerosis1735-39552251-6638Isfahan Cardiovascular Research Center, Isfahan University of Medical Sciences ARYA-14-08510.22122/arya.v14i2.1489Meta-AnalysisEffect of cardiac rehabilitation on inflammation: A systematic review and meta-analysis of controlled clinical trials Sadeghi Masoumeh 1Khosravi-Broujeni Hossein 2Salehi-Abarghouei Amin 3Heidari Ramin 4Masoumi Gholamreza 5Roohafza Hamidreza 6 1 Professor, Cardiac Rehabilitation Research Center, Cardiovascular Research Institute, Isfahan University of Medical Sciences, Isfahan, Iran 2 School of Medicine AND Menzies Health Institute, Griffith University, Southport, Queensland, Australia 3 Assistant Professor, Nutrition and Food Security Research Center AND Department of Nutrition, School of Health, Shahid Sadoughi University of Medical Sciences, Yazd, Iran 4 Assistant Professor, Isfahan Cardiovascular Research Center, Cardiovascular Research Institute, Isfahan University of Medical Sciences, Isfahan, Iran 5 Associate Professor, Cardiac Anesthesiology Research Center, Chamran Heart Center Hospital, Isfahan University of Medical Sciences, Isfahan, Iran 6 Assistant Professor, Psychosomatic Research Center, Isfahan University of Medical Sciences, Isfahan, IranCorrespondence to: Hamidreza Roohafza, Email: roohafza@crc.mui.ac.ir3 2018 14 2 85 94 17 9 2016 29 1 2018 © 2018 Isfahan Cardiovascular Research Center & Isfahan University of Medical Sciences2018This work is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported License which allows users to read, copy, distribute and make derivative works for non-commercial purposes from the material, as long as the author of the original work is cited properly.BACKGROUND This systematic review and meta-analysis aimed to assess the effect of cardiac rehabilitation (CR) on serum C-reactive protein (CRP) as an indicator of the inflammatory state and predictor of recurrent cardiovascular events. METHODS PubMed, SCOPUS, Cochrane library, and Google Scholar databases were searched up to January 2014 for original articles which investigated the effect of CR on CRP among adult patients with previous cardiovascular events. The random effects model was used to assess the overall effect of CR on the variation in serum CRP levels. RESULTS In the present systematic review and meta-analysis, 15 studies were included. The analysis showed that CR might significantly reduce high-sensitivity CRP (hs-CRP) levels [Difference in means (DM) = -1.81 mg/l, 95% confidence interval (CI): -2.65, -0.98; P = 0.004). However, the heterogeneity between studies was significant (Cochran's Q test, P < 0.001, I-squared = 84.9%). To find the source of variation, the studies were categorized based on study design (quality) and duration. The negative effect was higher among studies which followed their participants for 3 weeks or less (DM = -2.75 mg/l, 95% CI: -3.86, -1.64; P < 0.001) compared to studies which investigated the effect of CR for 3-8 weeks (DM = -0.89 mg/l, 95% CI: -1.35, -0.44; P < 0.001) and those which lasted more than 8 weeks (DM = -1.71 mg/l, 95% CI: -2.53, -0.89; P < 0.001). There was no evidence of heterogeneity when the categorization was based on the follow-up period. CONCLUSION Both short- and long-term CR have resulted in improvement in serum hs-CRP levels. CR can be perceived as a beneficial tool to reduce inflammatory markers among patients with previous cardiac events. Cardiac RehabilitationInflammationC-Reactive Protein ==== Body Introduction Cardiovascular diseases (CVDs) are a group of disorders which effect the heart and blood vessels and are the leading cause of death, worldwide.1 Based on the World Health Organization (WHO) reports, 17.3 million people (about 30% of global deaths) died due to CVDs in 2008, among which 7.3 and 6.2 million were reported to dye due to coronary heart disease (CHD) and stroke, the two major subclasses of CVDs, respectively. Inflammation is shown to be one of the important factors in the development and clinical course of most CVDs. Inflammation is actively involved in all levels of atherogenesis, from the initial lesions to the end-stage complications. Therefore, atherosclerosis is presently recognized as a low grade inflammatory vascular disease,2 which maximizes after the cardiac events. Moreover, reducing the inflammatory state after the cardiovascular event might be of great importance. Inflammatory markers, including interleukin-6 (IL-6), tumor necrosis factor-α (TNF-α) and particularly, C-reactive protein (CRP), are considered as risk markers of CVDs.3 The lipid-lowering therapy as a CVD risk-reducing strategy has also been reported to produce a parallel decrease in CRP. It has been shown that CRP concentrations decrease 15-50% with statin therapy.4-6 Cardiac rehabilitation (CR) has been defined as “comprehensive long-term services involving medical evaluation, prescribed exercise, cardiac risk-factor modification, health education, counseling, and behavioral interventions” by the United States Department of Health and Human Services (HHS) and the National Heart, Lung, and Blood Institute (NHLBI). CR has been shown to beneficially effect overall health and metabolic factors including inflammatory markers among patients who have experienced CVDs.7-11 For instance, it has been shown that exercise training, as an important part of CR programs, might be more effective in reducing the inflammatory markers than standard treatments provided early after acute myocardial infarction.12,13 Several studies have assessed the effect of CR on inflammatory markers including fibrinogen,14,15 TNF-α,16,17 and IL-6.17,18 However, the majority of studies have selected high-sensitivity CRP (hs-CRP) levels as a widely acceptable marker of inflammation to assess the beneficial effects of CR on inflammatory state.17,19-22 Furthermore, the use of serum CRP has been suggested as a predictor for CVDs and their recurrence.23 Studies have led to conflicting results regarding the effect of CR on CRP levels. Some researchers have shown the beneficial effect of CR on CRP levels,9,18,24,25 while others did not find the same results.20 This is while the majority of published data are from low quality before-after trials,8,20,24-26 and the parallel studies had not assessed the difference in the CRP level variation between their intervention group and controls.9,13,17-19,22,27-29 To the best of our knowledge, no study has attempted to summarize published data about the effect of CR on CRP levels. As the clinical trials provide the most qualified data regarding this association, the present study aimed to systematically review the controlled clinical trials investigating the effect of CR on inflammatory markers in patients with CVDs, and if possible quantify their results and search for their possible sources of heterogeneity through a meta-analysis. Materials and Methods The online databases of PubMed, ISI Web of Science, Scopus, Science Direct, and Embase were searched for relevant English and non-English publications up to January 2014. Moreover, experts in this field were contacted and reference lists of the published papers were searched. The keywords used in the present search strategy consisted of those selected from the Medical Subject Headings (MeSH) database and other related non-MeSH terms. The non-MeSH terms consisted of 3 groups. The first group were keywords related to cardiac rehabilitation ("cardiac rehabilitation", "cardiovascular rehabilitation", "cardiopulmonary rehabilitation", OR "cardiac exercise", OR "cardiovascular exercise", OR "cardiopulmonary exercise", "rehabilitation", "cardiac rehab*", "pulmonary rehabilitation"). The second group were keywords related to cardiometabolic markers ["intercellular cell adhesion molecule (ICAM)", "vascular cell adhesion molecule (VCAM)", "adhesion molecule", "E-selectin", "fibrinogen", "white blood cell", "serum amyloid A", "cytokines", "erythrocyte sedimentation rate (ESR)", "total cholesterol (TC)", "inflammatory marker", "interleukin", "smoking", "inflammation", "inflammation mediators", "IL" OR "C-reactive protein", "c reactive protein", "CRP", "inflammatory", "inflammation", "tumor necrosis factor", "TNF", "interleukins", "CRP", "TNF-α", "cholesterol", "lipoproteins, high-density lipoprotein (HDL)", "lipoproteins, low-density lipoprotein (LDL)", "triglycerides", "glucose tolerance test (GTT)", "insulin", "blood glucose", "insulin resistance", "cholesterol", "lipoproteins, HDL", "lipoproteins, LDL", "triglycerides", "GTT", "insulin", "blood glucose", "insulin resistance", "LDL", "HDL", "triglyceride", "triacylglycerol (TG)", "TC", "GTT", "fast blood sugar (FBS)", " fasting blood glucose (FBG)", "fasting insulin", "FBS", "FBG", "insulin sensitivity", "blood sugar", "lipid profile", "serum lipid", "plasma lipid", "blood pressure", “hypertension”, "blood pressure", and “hypertension”]. The third group consisted of keywords related to clinical trials ("intervention studies", "intervention", "controlled trial", "randomized", "randomized", "random", "randomly", "placebo", "assignment", "clinical trial", and "trial"). Databases were searched using keywords 1 in combination with 2 and 1 in combination with 3 and duplicate studies were removed. No filter or limitation was implemented while searching the mentioned databases. To be included in the meta-analysis, a published study had to meet the following criteria: be an original article, a controlled clinical trial, and conducted in an adult human population, and having assessed the effect of CR on serum CRP levels. As the majority of studies had reported CRP levels as their main outcome variable, in the present study, only these articles were included in the systematic review and meta-analysis. Data were extracted on publication (the first author's last name, year of publication, and country of the studied population), number of individuals in the intervention and control groups, duration of the intervention, age, gender, and mean and standard deviation (SD) of inflammatory factors at baseline and after CR. The mean and SD of CRP levels at baseline and after the intervention period was used to calculate the mean change in CRP levels and its corresponding SD which was included in the meta-analysis as an effect size (raw difference in means: DM) from each before and after study. The difference in CRP levels variation between intervention and control groups (DM) was also calculated and used as the effect size for controlled clinical trials. Summary mean estimates with their corresponding SDs were derived using the random effects model (method of DerSimonian and Laird),30 which incorporates between-study variability. Subgroup analyses were performed to check for specific sources of heterogeneity. Sensitivity analysis was used to explore the extent to which inferences might depend on a particular study or group of studies. Publication bias was assessed by visual inspection of funnel plots.31 In these funnel plots, the difference in mean CRP was displayed against the inverse of the square of the standard error (a measure of the precision of the studies). Funnel plots generally show a peak at the point where the studies with smaller standard errors are found which usually represents the point of the approximate true effect. Formal statistical assessment of funnel plot asymmetry was performed using Egger's regression asymmetry test and adjusted rank correlation test.32 Reported P-values are from the intercept of the regression analysis, which provides a measure of asymmetry. In addition, Begg's adjusted rank correlation test was used.32 Statistical analyses were carried out in Stata (version 11.2; StataCorp, College Station, TX, USA). P values of less than 0.05 were considered statistically significant. Results In the preliminary online search, 7418 studies were retrieved; about 7354 were excluded after reading titles or abstracts because they did not meet the inclusion criteria (Figure 1). Finally, 43 articles were found which had investigated the effect of CR on one of the inflammatory markers; among these, 15 articles had examined the association between CR and hs-CRP or CRP.8,9,13,17-22,24-29 One of these 15 eligible studies had examined the effect of a very short-term period (1 session) of CR on hs-CRP levels in patients with ischemic heart disease (IHD).21 As the objective of the present review was to assess the long-term effect of CR on CRP levels, this study was removed from the qualitative and quantitative synthesis. Moreover, a study performed by Hansen et al. was not included in the present review.9 This study was a long-term follow-up (about 72 weeks) of two groups of patients randomly undergoing a 40-minute and 60-minute CR program for 7 weeks, and there was no intervention in the follow-up period.9 Therefore, 13 articles were eligible to be included in the systematic review.8,13,17-20,22,24-29 The characteristics of studies are presented in table 1. Studies were conducted in USA,18,20,22 Italy,8,24,26 South Korea,17,27 Croatia,13 Serbia,28 Belgium,29 Japan,25 and Iran.19 Among these studies, 5 were of lower quality because they had a before-after,8,20,24-26 while others were parallel studies with control groups.9,13,17-19,22,27-29 However, only in 2 of the parallel studies the patients were randomly assigned to intervention and control groups.9,13 Although the study conducted by Hansen et al.29 was a randomized controlled clinical trial, the intervention and control groups underwent 60 and 40 minutes of CR, respectively; therefore, the study was considered a before-after study which assessed the effect of the CR program on all its participants. Results of meta-analysis: Of the 15 eligible article for the systematic review, 2 studies did not report data about baseline and post-intervention serum CRP levels;19,25 These data could not be obtained even after contacting the authors twice through email, 1 week apart. Therefore, 13 studies were included for the quantification of the effect of CR on serum CRP levels.8,9,13,17,18,20-22,24,26-29 In total, 992 adult participants with CVDs were included in the current analysis. The analysis showed that CR significantly reduces hs-CRP levels [DM = -1.81 mg/l, 95% confidence interval (CI): -2.65, -0.98; P = 0.004). However, the heterogeneity between studies was significant (Cochran's Q test, P < 0.001, I-squared = 84.9%) (Table 2). To find the source of variation, the studies were categorized based on their design (quality) and duration. When categorizing studies based on their design, the significant reducing effect of CR was observed in before-after studies (DM = -1.74 mg/l, 95% CI: -2.93, -0.55; P = 0.001), parallel studies (DM = -1.39 mg/l, 95% CI: -2.19, -0.59; P < 0.001), and randomized controlled (DM = -4.5 mg/l, 95% CI: -6.63, -2.37; P < 0.001) clinical trials. The heterogeneity was still significant among before-after studies (Cochran's Q test, P < 0.001, I-squared = 92%), while there was no evidence of heterogeneity between parallel studies (Cochran's Q test, P = 0.990, I-squared = 0.0%). Studies were also categorized based on their follow-up period (Table 3). Although, the reducing effect was statistically significant for all these subgroups, the negative effect was higher among studies which followed their participants for 3 weeks or lower (short follow-up period) (DM = -2.75 mg/l, 95% CI: -3.86, -1.64; P < 0.001) compared to studies which investigated the CR effect for 3-8 weeks (middle follow-up period) (DM = -0.89 mg/l, 95% CI: -1.35, -0.44; P < 0.001) and those which lasted more than 8 weeks (long follow-up period) (DM = -1.71 mg/l, 95% CI: -2.53, -0.89; P < 0.001). Although the I-squared showed a moderate heterogeneity among studies with short follow-up periods (48.2%), Cochran's Q test did not provide a significant evidence of heterogeneity (P = 0.122). The heterogeneity was not significant for studies with middle- (Cochran's Q test, P = 0.80, I-squared = 0.0%) and long-term follow-up periods (Cochran's Q test, P = 0.290, I-squared = 19.4%). In a sensitivity analysis, it was found that the effect of CR on CRP levels was not substantially modified by the result of a certain study. Although a slight asymmetry was seen in Begg’s funnel plot, no evidence of publication bias was found using asymmetry tests (Egger’s test, P = 0.192, Begg and Mazumdar test, P = 0.170) (Figure 2). Discussion The systematic review and meta-analysis on clinical trials revealed the significant reducing effect of CR on serum hs-CRP levels. The analysis also showed that CR beneficially affects inflammation both in a short and long period of time. Therefore, CR might reduce the risk of CVDs recurrence not only through its beneficial effects on the lipid profile, but also through its effect on the inflammatory state. To the best of our knowledge, this is the first systematic review and meta-analysis assessing the effects of CR on inflammatory markers. CRP, as a liver-derived molecule that is increased in inflammatory states, rapidly increases within hours after tissue injuries like CVDs. Since CVD is at least in part an inflammatory process, CRP has been investigated in the context of arteriosclerosis and subsequent vascular disorders. CRP is recommended as a predictive laboratory marker for CVD risk in patients with high susceptibility to CVDs or their recurrence.33 CRP has been shown to have some advantages as a marker: (1) It is a stable marker; (2) It does not have clinically significant circadian variation; and (3) It’s measurement is easy and reliable.34 Therefore, decreasing the inflammatory state markers including CRP levels might help patients to decrease the risk of recurrence of CVDs. Aerobic exercise, as an adjunct to pharmacological therapy in patients with chronic heart failure and CAD, is known to improve peripheral vascular function and skeletal muscle abnormalities. Although exercise causes stress-induced up-regulation of endothelial nitric oxide synthase (eNOS),35 it is supposed that the modification of inflammatory mediators through exercise reverses peripheral vascular endothelial dysfunction.16 The present study results are in line with the previous reports suggesting that CR improved the inflammatory markers of patients experiencing CVDs.8,13,19,22 Therefore, CR must be considered as an important strategy adjacent to conventional therapy used for patients experiencing cardiac events. Some limitations must be considered while interpreting the present results. The effect of CR on CRP levels was studied because it is an acceptable marker for inflammation and also might be used as a predictor of cardiovascular events. Furthermore, the majority of published studies had selected the same marker to assess the inflammatory state. A limited number of studies were found which had used other inflammatory markers like TNF-α and interleukins.16-18 Further studies on the effect of CR on other inflammatory markers are highly recommended in order to determine whether CR has the same beneficial effects. In addition, it has been proposed that the beneficial effect of CR might be because of its negative effect on body fat and obesity. Lavie et al. showed that the beneficial effect of CR on CRP levels is only significant in obese patients with CHD.20 it has also been proposed that circulating hs-CRP and inflammatory cytokines are in relation with obesity among healthy subjects.36 In the present review, there was only one study which had separately examined the effect of CR on hs-CRP levels among obese and lean participants.20 Furthermore, the majority of eligible studies did not report the effect of CR on weight or BMI variation; therefore, the association between obesity or body fat and CRP levels variation could not be explored. In conclusion, the present systematic review provides strong evidence of the beneficial effect of CR on the inflammatory state of patients with CVDs. It is highly recommended that further studies be conducted to explore whether CR affects other inflammatory markers and how dependent the changes in inflammatory markers as a result of CR are on the magnitude of the change in body composition. Conclusion Both short- and long-term CR program have resulted in improvement in serum hs-CRP levels. CR can be perceived as a beneficial tool to reduce inflammatory markers among patients with previous cardiac events. It is highly recommended that further studies be conducted to explore whether CR affects other ýinflammatory markers and how dependent the changes in inflammatory markers as a result of CR ýare on the magnitude of the change in body composition. Acknowledgments This study was financially approved and supported by Isfahan University of Medical Sciences (grant No: 191175) and Cardiovascular Research Institute. The authors wish to gratefully acknowledge the assistance of Mrs. Safoura Yazdekhasti. Conflicts of Interest Authors have no conflict of interests. Figure 1 Flow chart for the study selection process TNF-α: Tumor necrosis factor-α; IL-6: Interleukin-6; hs-CRP: High-sensitivity C-reactive protein; CR: cardiac rehabilitation Figure 2 Begg’s funnel plot with pseudo 95% confidence interval (CI) of the difference in means versus the standard errors of the difference in means for studies investigating the effect of cardiac rehabilitation on Creactive protein /high-sensitivity C-reactive protein (b) MD: Mean difference Table 1 Characteristics of studies evaluated the effect of cardiac rehabilitation (CR) on metabolic syndrome and/or its components Author/year Country Subjects and gender Mean Age (year) Design Intervention Control Duration Outcome variable Result Fukuda et al.25 Japan Patients with cardiovascular diseases (F: 6/M: 44) 61 Before-after Aerobic bicycle exercise 2 or 3 times per week for 3-6 months - 24 weeks hs-CRP hs-CRP levels decreased, but it was not statistically significant. Ferratini et al.24 Italy Patients after cardiac surgery (F: 68/M: 155) 67 Before-after Up to 30 minutes of cycling 5 times a week at 70% maximal heart rate - 3 weeks hs-CRP hs-CRP levels significantly decreased after CR program. Cesari et al.8 Italy Patients with acute coronary syndrome (F:20/M: 92) 58.2 Before-after 3 days/week of endurance training on a cycle-ergometer at 60-70% of VO2 level - 4 weeks hs-CRP hs-CRP levels significantly decreased after CR program. Aminlari et al.19 Iran Patients with myocardial infarction (56 M/F) 62.7 Parallel (randomization was not mentioned) Exercise training, education, and behavior modification therapy were performed 3 times per week. The exercise training included arm and leg ergometry and treadmills. Behavioral modifications were smoking cessation, healthy nutrition, hypertension control, and etc. No intervention 8 weeks CRP CRP levels significantly decreased in the intervention group compared to controls. Kim et al.27 South Korea Patients with acute myocardial infarction (F:32/M: 109) 63.24 Parallel without randomization Warm-up (10 minutes), exercise (30 minutes), and cool-down (10 minutes) The intensity of exercise was adjusted on a test result basis by calculation of heart rate reserve first, followed by the increased target heart rate from 40% to 85% of the value in phases General training on exercise or risk factors management. instructed to maintain their own exercise 16 weeks hs-CRP The exercise group showed a significantly lower value of hs-CRP than the control group. Rankovic et al.28 Serbia patients with ischemic heart disease (F: 23/M:29) 60.22 Parallel (randomization was not mentioned) Continual aerobic exercise for 45 minutes on a treadmill, room bicycle or walking The intensity of physical exercise was limited to the submaximal physical capacity at the level of 70-80% of maximal heart frequency at the stress test taken before cardiovascular rehabilitation. Physical exercise was applied 3 times a week. Did not have physical training in the last 6 months, except for usual household activities. 3 weeks hs-CRP CRP levels decreased significantly in the exercise group compared to controls. Cesari et al.26 Italy Patients after cardiac surgery (F: 35/M: 51) 72.5 Before-after Aerobic exercise at cycle ergometer and short lasting calisthenic exercises, with the resistance sequentially provided by the weight of single body segments and gentle, passive stretching involving all the main joints The training frequency was 6 times per week for a total of 12 training sessions. - 15 days hs-CRP hs-CRP levels significantly decreased after CR program. Lavie et al.20 USA Patients with coronary heart disease (F:72/M:73) 65.37 Before-after 3 times per week group exercise and educational sessions, and individual exercise (between 1 and 3 times per week) on non-rehabilitation days - 12 weeks hs-CRP hs-CRP levels significantly decreased only among obese participants but not among lean subjects. Kim et al.17 South Korea Patients with coronary artery disease (F:11/M: 28) 50.08 Parallel (randomization was not mentioned) Supervised exercise under prescription based on symptom-limited treadmill exercise test at hospital lasted 6 weeks + a home based and self-managed exercise lasting 8 weeks The exercise: warm-up, 30- to 40-min exercise on a treadmill or bicycle ergometer, and a cool-down followed up with standard care as outpatients 14 weeks hs-CRP hs-CRP significantly decreased only among patients undergoing cardiac rehabilitation compared to controls. Hansen et al.29 Belgium Patients with coronary artery disease (F: 25/M: 109) 63.15 before-after 60 or 40 minutes of exercise (42% on the treadmill, 33% on the cycle ergometer, and 25% on the arm cranking device) 7 weeks CRP hs-CRP levels decreased in both exercise groups. Balen et al.13 Croatia Patients with acute myocardial infarction (F: 16/M: 44) 60 Randomized controlled clinical trial 45-minute aerobic activity on a cycle-ergometer and 30-minute organized program of supervised walking on a standardized track standard care 3 weeks hs-CRP hs-CRP significantly decreased in both groups, but after the intervention, values were significantly lower among patients with cardiac rehabilitation. Shin et al.18 USA patients with coronary artery disease after percutaneous coronary intervention (F: 11/M: 28) 56.58 Parallel Cardiac rehabilitation following hospital discharge and 8 weeks of home stay exercise plus statin therapy All subjects were prescribed daily lipid lowering medication consisting of 100 mg aspirin and 75 mg clopidogrel throughout the experimental period. 80 mg daily of fluvastatin 14 weeks hs-CRP hs-CRP significantly decreased in both groups but after intervention values were significantly lower among patients with cardiac rehabilitation. Milani et al.22 USA Patients with coronary heart disease (F:75/M: 202) 66.27 Parallel Patients received formalized exercise instruction and met 3 times per week for group exercise classes, and were encouraged to exercise on their own (1 to 3 times per week) in between sessions. Did not have cardiopulmonary exercise tests. 12 weeks hs-CRP hs-CRP levels significantly decreased only among patients with cardiac rehabilitation. F: Female, M: Male; hs-CRP: High-sensitivity C-reactive protein Table 2 Forest plot demonstrating weighted mean differences with 95% confidence interval (CI) for all eligible studies investigating the effects of cardiac rehabilitation on C-reactive protein /high-sensitivity Creactive protein levels grouped by study designs using random effects model Study DM (95% CI) Before-after studies  Lavie et al.20 (obese) -3.40 (-4.94, -1.86)  Lavie et al.20 (lean) -1.00 (-2.60, 0.60)  Hansen et al.29 -0.80 (-1.66, 0.06)  Cesari et al.26 -0.20 (-6.36,5.96)  Ferratini et al.24 -2.90 (-3.15, -2.65)  Cesari et al.8 -0.93 (-1.47, -0.39) Subtotal (I-squared = 92.0%, P < 0.001) -1.74 (-2.93, 0.55) Parallel studies with no randomization  Milani et al.22 -2.40 (-15.04, 10.24)  Shin et al.18 -1.49 (-3.48, 0.50)  Kim et al.17 -1.00 (-2.68, 0.68)  Rankovic et al.28 -1.51 (-3.11, 0.09)  Kim et al.27 -1.50(-2.85, 0.15) Subtotal (I-squared = 0.0%, P = 0.991) -1.39 (-2.19, 0.59) Randomized clinical trial  Balen et al.13 -4.50 (-6.63, -2.37) Subtotal (I-squared = 0%, P < 0.001) -4.50 (-6.63, -2.37) Overall (I-squared = 84.9%, P < 0.001) -1.81 (-2.65, -0.98) DM: Difference in means; CI: Confidence interval Table 3 Forest plot demonstrating weighted mean differences with 95% confidence interval (CI) for all eligible studies investigating the effects of cardiac rehabilitation on C-reactive protein /high-sensitivity Creactive protein levels grouped by study follow-up period using random effects model Study DM (95% CI) Short follow-up period  Balen et al.13 -4.50 (-6.63, -2.37)  Rankovic et al.28 -1.51 (-3.11, 0.09)  Cesari et al.26 -0.20 (-6.36,5.96)  Ferratini et al.24 -2.90 (-3.15, -2.65) Subtotal (I-squared = 48.2%, P = 0.122) -2.75 (-3.86, -1.64) Middle follow-up period  Hansen et al.29 -0.80 (-1.66, 0.06)  Cesari et al.8 -0.93 (-1.47, -0.39) Subtotal (I-squared = 0.0%, P = 0.801) -0.89 (-1.35, -0.44) Long follow-up period  Milani et al.22 -2.40 (-15.04, 10.24)  Shin et al.18 -1.49 (-3.48, 0.50)  Lavie et al.20 (obese) -3.40 (-4.94, -1.86)  Lavie et al.20 (lean) -1.00 (-2.60, 0.60)  Kim et al.17 -1.00 (-2.68, 0.68)  Kim et al.27 -1.50(-2.85, 0.15) Subtotal (I-squared = 19.4%, P = 0.287) -1.71 (-2.53, 0.89) Overall (I-squared = 84.9%, P < 0.001) -1.81 (-2.65, -0.98) DM: Difference in means; CI: Confidence interval ==== Refs REFERENCES 1 Talaei M Sarrafzadegan N Sadeghi M Oveisgharan S Marshall T Thomas GN Incidence of cardiovascular diseases in an Iranian population: The Isfahan Cohort Study. Arch Iran Med 2013 16 3 138 44 23432164 2 Wilensky RL Hamamdzic D The molecular basis of vulnerable plaque: Potential therapeutic role for immunomodulation. Curr Opin Cardiol 2007 22 6 545 51 17921743 3 Libby P Inflammation and cardiovascular disease mechanisms. Am J Clin Nutr 2006 83 2 456S 60S 16470012 4 van de Ree ME Huisman MV Princen HM Meinders AE Kluft C Strong decrease of high sensitivity C-reactive protein with high-dose atorvastatin in patients with type 2 diabetes mellitus. Atherosclerosis 2003 166 1 129 35 12482559 5 Ridker PM Rifai N Clearfield M Downs JR Weis SE Miles JS Measurement of C-reactive protein for the targeting of statin therapy in the primary prevention of acute coronary events. N Engl J Med 2001 344 26 1959 65 11430324 6 Ridker PM Rifai N Pfeffer MA Sacks FM Moye LA Goldman S Inflammation, pravastatin, and the risk of coronary events after myocardial infarction in patients with average cholesterol levels. Cholesterol and Recurrent Events (CARE) Investigators. Circulation 1998 98 9 839 44 9738637 7 Wenger NK Froelicher ES Smith LK Ades PA Berra K Blumenthal JA Cardiac rehabilitation as secondary prevention. Agency for Health Care Policy and Research and National Heart, Lung, and Blood Institute. Clin Pract Guidel Quick Ref Guide Clin 1995 17 1 23 8595435 8 Cesari F Marcucci R Gori AM Burgisser C Francini S Sofi F Impact of a cardiac rehabilitation program and inflammatory state on endothelial progenitor cells in acute coronary syndrome patients. Int J Cardiol 2013 167 5 1854 9 22626841 9 Hansen D Dendale P Raskin A Schoonis A Berger J Vlassak I Long-term effect of rehabilitation in coronary artery disease patients: Randomized clinical trial of the impact of exercise volume. Clin Rehabil 2010 24 4 319 27 20176771 10 Sarrafzadegan N Rabiei K Kabir A Sadeghi M Khosravi A Asgari S Gender differences in risk factors and outcomes after cardiac rehabilitation. Acta Cardiol 2008 63 6 763 70 19157173 11 Ghashghaei FE Sadeghi M Marandi SM Ghashghaei SE Exercise-based cardiac rehabilitation improves hemodynamic responses after coronary artery bypass graft surgery. ARYA Atheroscler 2012 7 4 151 6 23205048 12 Sadeghi M Garakyaraghi M Khosravi M Taghavi M Sarrafzadegan N Roohafza H The impacts of cardiac rehabilitation program on echocardiographic parameters in coronary artery disease patients with left ventricular dysfunction. Cardiol Res Pract 2013 2013 201713 24459599 13 Balen S Vukelic-Damijani N Persic V Ruzic A Miletic B Samardiija M Anti-inflammatory effects of exercise training in the early period after myocardial infarction. Coll Antropol 2008 32 1 285 91 18494215 14 Solov'ev AV Ermolin GA Ignashenkova GV Dikov MM Fibrinogen and fibrin degradation products in the blood of acute myocardial infarct patients at the hospital rehabilitation stage. Ter Arkh 1987 59 10 21 3 15 Wosornu D Allardyce W Ballantyne D Tansey P Influence of power and aerobic exercise training on haemostatic factors after coronary artery surgery. Br Heart J 1992 68 2 181 6 1389734 16 Conraads VM Beckers P Bosmans J De Clerck LS Stevens WJ Vrints CJ Combined endurance/resistance training reduces plasma TNF-alpha receptor levels in patients with chronic heart failure and coronary artery disease. Eur Heart J 2002 23 23 1854 60 12445534 17 Kim YJ Shin YO Bae JS Lee JB Ham JH Son YJ Beneficial effects of cardiac rehabilitation and exercise after percutaneous coronary intervention on hsCRP and inflammatory cytokines in CAD patients. Pflugers Arch 2008 455 6 1081 8 17906875 18 Shin YO Bae JS Lee JB Kim JK Kim YJ Kim C Effect of cardiac rehabilitation and statin treatment on anti-HSP antibody titers in patients with coronary artery disease after percutaneous coronary intervention. Int Heart J 2006 47 5 671 82 17106138 19 Aminlari A Jazayeri SM Bakhshandeh AR Association of cardiac rehabilitation with improvement in high sensitive C-reactive protein post-myocardial infarction. Iran Red Crescent Med J 2012 14 1 49 50 22737555 20 Lavie CJ Morshedi-Meibodi A Milani RV Impact of cardiac rehabilitation on coronary risk factors, inflammation, and the metabolic syndrome in obese coronary patients. J Cardiometab Syndr 2008 3 3 136 40 18983328 21 Madarame H Kurano M Fukumura K Fukuda T Nakajima T Haemostatic and inflammatory responses to blood flow-restricted exercise in patients with ischaemic heart disease: A pilot study. Clin Physiol Funct Imaging 2013 33 1 11 7 23216760 22 Milani RV Lavie CJ Mehra MR Reduction in C-reactive protein through cardiac rehabilitation and exercise training. J Am Coll Cardiol 2004 43 6 1056 61 15028366 23 Ridker PM Clinical application of C-reactive protein for cardiovascular disease detection and prevention. Circulation 2003 107 3 363 9 12551853 24 Ferratini M Ripamonti V Masson S Grati P Racca V Cuccovillo I Pentraxin-3 predicts functional recovery and 1-year major adverse cardiovascular events after rehabilitation of cardiac surgery patients. J Cardiopulm Rehabil Prev 2012 32 1 17 24 22193930 25 Fukuda T Kurano M Iida H Takano H Tanaka T Yamamoto Y Cardiac rehabilitation decreases plasma pentraxin 3 in patients with cardiovascular diseases. Eur J Prev Cardiol 2012 19 6 1393 400 21914683 26 Cesari F Sofi F Caporale R Capalbo A Marcucci R Macchi C Relationship between exercise capacity, endothelial progenitor cells and cytochemokines in patients undergoing cardiac rehabilitation. Thromb Haemost 2009 101 3 521 6 19277414 27 Kim C Kim DY Moon CJ Prognostic influences of cardiac rehabilitation in Korean acute myocardial infarction patients. Ann Rehabil Med 2011 35 3 375 80 22506147 28 Rankovic G Milicic B Savic T Dindic B Mancev Z Pesic G Effects of physical exercise on inflammatory parameters and risk for repeated acute coronary syndrome in patients with ischemic heart disease. Vojnosanit Pregl 2009 66 1 44 8 19195263 29 Hansen D Dendale P Berger J Onkelinx S Reyckers I Hermans A Importance of exercise training session duration in the rehabilitation of coronary artery disease patients. Eur J Cardiovasc Prev Rehabil 2008 15 4 453 9 18677171 30 DerSimonian R Laird N Meta-analysis in clinical trials. Control Clin Trials 1986 7 3 177 88 3802833 31 Egger M Davey SG Schneider M Minder C Bias in meta-analysis detected by a simple, graphical test. BMJ 1997 315 7109 629 34 9310563 32 Egger M Davey-Smith G Altman D Systematic reviews in health care: Meta-analysis in context. PHoboken, NJ John Wiley & Sons 2001 33 Pfutzner A Forst T High-sensitivity C-reactive protein as cardiovascular risk marker in patients with diabetes mellitus. Diabetes Technol Ther 2006 8 1 28 36 16472048 34 Libby P Ridker PM Inflammation and atherosclerosis: Role of C-reactive protein in risk assessment. Am J Med 2004 116 Suppl 6A 9S 16S 15050187 35 Gielen S Adams V Mobius-Winkler S Linke A Erbs S Yu J Anti-inflammatory effects of exercise training in the skeletal muscle of patients with chronic heart failure. J Am Coll Cardiol 2003 42 5 861 8 12957433 36 Yudkin JS Stehouwer CD Emeis JJ Coppack SW C-reactive protein in healthy subjects: Associations with obesity, insulin resistance, and endothelial dysfunction: A potential role for cytokines originating from adipose tissue? Arterioscler Thromb Vasc Biol 1999 19 4 972 8 10195925