Physical Activity and C-reactive Protein Levels: The Confounding Role of Body Fat

Transcription

1 Journal of Physical Activity and Health, 2011, 8, Human Kinetics, Inc. Physical Activity and C-reactive Protein Levels: The Confounding Role of Body Fat James LeCheminant, Larry Tucker, and Kenric Russell Background: This study investigated the relationship between objectively-measured total physical activity (PA), and intensity of PA and high-sensitivity C-reactive protein (CRP) in 211 healthy, middle-age women (43.1 ± 3.0 y). In addition, this study examined the extent to which age, BMI, abdominal circumference, and body fat percentage operated as confounders in these associations. Methods: PA was objectively measured for 7 continuous days using accelerometry. Fasting blood samples were taken, from which CRP was measured using a solid phase ELISA. Body mass index (BMI) (kg/m 2 ), abdominal circumference measured at the umbilicus, and body fat percentage using air displacement plethysmography, were assessed. Results: Total PA (activity counts) was significantly and inversely related to CRP concentrations (F = 7.76, P =.006) as was vigorous-intensity PA. After adjusting for differences in body fat percentage, total PA and vigorous-intensity PA were no longer significant predictors of CRP. Abdominal circumference and BMI also tended to weaken the relationship between total or vigorous-intensity PA and CRP but not to the same extent as body fat percentage. Conclusions: These findings suggest that higher total and vigorous-intensity PA levels are significantly related to lower CRP levels in healthy, middle-age women; however, this relationship is largely a function of differences in body fat percentage. Keywords: women, cardiovascular risk It is well-documented that inflammation is associated with the process of atherosclerosis and the development of cardiovascular disease (CVD). 1 4 Mounting evidence suggests that high-sensitivity C-reactive protein (CRP), a marker of inflammation, is an independent predictor of CVD. 5,6 Further, some investigators have suggested that CRP may better predict future cardiovascular events than LDL cholesterol 7 and that CRP be used as one of the primary screening tools to assess cardiovascular risk. 8 Low levels of physical activity (PA) and physical fitness also increase the risk for CVD, with an attributable risk as high as 35% for coronary heart disease (CHD), the top CVD killer. 9 Investigations also consistently report that moderate to high levels of PA and/or changing from a sedentary lifestyle to a more active lifestyle decreases the risk of CHD This connection between increased PA and decreased CHD risk may be partially due to a decrease in endothelial inflammation as detected by CRP. In a review study of the relationship between PA and CRP, Plaisance and Grandjean suggested that CRP levels were 6 35% lower in adults with higher levels of PA and physical fitness. 6 Additionally, Kasapis and Thompson systematically reviewed 18 cross-sectional studies comparing CRP by activity level and 5 prospective The authors are with the Dept of Exercise Sciences, Brigham Young University, Provo, UT. exercise studies and reported an anti-inflammatory effect of habitual PA, indicated by lowered plasma CRP. 14 This relationship has been found among men and women of various racial groups. 15 Despite the mounting evidence linking PA and CRP, there are multiple shortcomings related to this literature. First, most investigations have used self-reported measures to determine PA. 6 The limitations of self-reported data are well-known. Second, while a higher body fat percentage has been shown to be associated with higher CRP levels, the extent that body fat influences the relationship between PA and CRP is undetermined. Crosssectional, longitudinal, and intervention studies suggest PA level is an independent predictor of CRP, 4,5,14,16 while others suggest body fat attenuates or eliminates this relationship Moreover, most studies have used BMI as a covariate rather than a more precise measure of body composition. 6 Further, other potential confounders are not always considered, such as abdominal circumference, injury, illness, infection, or acute exercise. The current study sought to improve upon common weaknesses found in previous studies that have examined the relationship between PA and CRP. Therefore, the purpose of the current study was to determine the extent to which objectively-measured total PA, using accelerometry, was associated with resting CRP levels in healthy, middle-age women. In addition, this study examined the relationship between duration of either moderate- or vigorous-intensity activity and CRP. A 481

2 482 LeCheminant, Tucker, and Russell secondary purpose was to measure the extent to which potentially confounding factors, such as age, total body fat percentage, abdominal circumference, and BMI affect the PA and CRP association. This study also controlled for injury, illness, infection, and acute exercise, which can directly influence CRP levels, and used a relatively large sample of women. Participants Methods This study received university Institutional Review Board (IRB) approval before initiation. In addition, all participants signed an IRB approved informed consent before participating in the study. Participants were recruited through newspaper advertisements, fliers, , and word of mouth. Participants had no apparent health problems as determined by a physical activity readiness questionnaire (PAR-Q), and were nonsmokers. Measurement of Physical Activity Accelerometers were used to objectively measure PA. Originally, the accelerometers were purchased from Computer Science and Application but are currently known as Actigraph (Health One Technology, Inc., Fort Walton Beach, FL). These devices have been shown to give valid and reliable activity counts, and can precisely monitor everyday activities The accelerometers were worn over the left hip, in line with the outer seam of the pants and the umbilicus, and were secured by a small pouch fastened to a nylon belt. They were worn at all times during 7 continuous days, except while bathing and during water activities; thus, there was very little nonwear time (<2 hours per day). The monitors summed the motion counts of each participant into 10-minute intervals. Therefore, the motion of each participant was recorded in 144 time segments (epochs) each day, each with a length of 10 minutes. The length of 10 minute epochs was chosen based upon recommendations from the American College of Sports Medicine (ACSM) that physical activity can be accumulated in 10 minute bouts. 24 Over the course of the 7 days, each participant had a total of 1008 epochs which were summed to index total PA. To assess and categorize PA intensity, the following cut-points were used based upon our previous pilot data to describe low-intensity PA, moderate-intensity PA, and vigorous-intensity PA, respectively: 0 29,999 counts in 1 10-min epoch (<3 mph), 30,000 49,999 counts in 1 10-min epoch (3 4 mph), and 50,000 counts or greater in 1 10-min epoch (>4 mph). 25 The amount of time subjects spent in moderateintensity PA and vigorous-intensity PA was assessed separately. For moderate-intensity PA, the ACSM cutpoint of 150 minutes per week was applied, 24 and subjects were divided into 3 categories, Low: 0 min per week, Moderate: minutes, and High: 150 minutes or more per week. For vigorous-intensity PA, the ACSM cut-point of 60 minutes per week was employed, 24 and subjects were divided into Low (Level 1): 0 min per week, Moderate (Level 2): min, and High (Level 3): 60 or more min per week. Measurement of C-reactive Protein High-sensitivity CRP was measured using the solid phase ELISA (cat no. 1000; Alpha Diagnostics International (ADI), Inc., San Antonio, TX). This kit was chosen because of its sensitivity, precision, availability, and cost. 26 The ADI assay is most sensitive within the normal range of 0 to 5 mg/l. 26 The intra-assay variation in the normal range is low (CV = 3.0%; SD = 0.3), and the interassay variation is acceptable (CV = 7.0%; SD = 0.4). 26 Blood serum was collected by professionals at a certified hospital laboratory. Participants were asked not to exercise for at least 48 hours before being tested. If a participant was sick, recovering from an illness or infection, bruised, or injured, she was asked to wait until all symptoms were gone before having her blood drawn. If she was experiencing hay-fever symptoms, she was asked to rate the severity of the symptoms. Blood samples were drawn, centrifuged, and stored in aliquots at 20 C. After participants had their blood drawn, the serum was transported to the biochemistry laboratory at the university where an ELISA for C-reactive protein was performed for each participant. Assays were performed in duplicate and the mean of the 2 analyses was used to index CRP. Measurement of Body Composition and Anthropometrics Body fat percentage was measured using the BOD POD (Life Measurement Instruments, Concord, CA), which uses air displacement plethysmography. Test-retest reliability of the BOD POD was evaluated using a subset of 100 women from our sample and resulted in a strong intraclass correlation (r =.999). 27 Body composition for the same 100 participants was also measured simultaneously using Duel Energy X-ray Absorptiometery (DXA) (Hologic 4500W, Bedford, MA). When the DXA results were compared with the BOD POD findings, the Pearson correlation was 0.94 (P <.001), and the intraclass correlation was 0.97 (P <.001). 28 Others have also shown that the BOD POD produces reliable and valid measurements of body fat. 29,30 Participants fasted for 4 hours before being tested, and wore a standardized swimsuit and a swim cap. Before each test, the BOD POD was calibrated with a knownvolume cylinder according to the manufacturer s instructions. At least 2 tests were performed, and the average taken. If the 2 tests were not within 1 percentage point of each other, subsequent tests were performed until 2 were within 1 percentage point. Approximately 25% of

3 Physical Activity and CRP 483 participants required a third measurement. The average of the 2 results that were within 1 percentage point was used. Abdominal circumference was assessed at the umbilicus using a standard measuring tape. At least 2 abdominal assessments, within 1 cm, were made for each participant and the average was reported for this study. Height was assessed using a standard wall-mounted stadiometer and weight was determined using an electronic scale (Tanita Corporation, Japan). Body mass index was calculated in kg/m 2. Procedures Participants reported to the Human Performance Research Center at the University for 2 appointments. These methodologies and procedures have been published elsewhere. 25 During the first appointment, participants were asked to eliminate any bodily waste and to change into a standard nylon swimsuit. Subsequently, height, weight, abdominal circumference, and body fat percentage were assessed. After changing back into street clothes, the proper technique for wearing the accelerometer was demonstrated and discussed with the participant. An accelerometer was properly fitted, located over the left hip. A second appointment was made for 8 days later. Participants wore the monitor at all times (except bathing) for the next 7 days. During the 7-day data collection period, participants were called to encourage them to follow proper procedure. At the second appointment, the activity monitor was returned and the data were collected. Each participant was then given a blood requisition form for a nearby regional hospital blood laboratory, and was directed to have her blood drawn within the next week. Participants were instructed to fast for 12 hours, to be free of any illness, injury, or infection, and to refrain from moderate and intense PA 48 hours before the draw. At the time of the blood draw, participants completed a short questionnaire to confirm that they had fasted, were free from illness, injury and infection, and that they had not exercised for at least 48 hours. Participants who did not meet these conditions were asked to return to the laboratory to have their blood drawn at another time. Statistical Analyses Alpha was set at the 0.05 level and all analyses were performed using PC-SAS software (version 9.1, SAS Institute, Inc., Cary, NC). Because the CRP data for the participants in this study were not normally distributed, they were log transformed and used in the analyses. However, to allow for easier interpretation, CRP data in the results section and tables are reported in common clinical units. The magnitude and direction of the bivariate relationships between objectively measured total PA or moderate- or vigorous-intensity PA, resting CRP, and the potential confounders were indexed using Pearson product-moment correlation coefficients. As there was no significant relationship for moderate-intensity PA and CRP, only total and vigorous PA and their association with CRP received additional analyses (see below). Participants were divided into quartiles based on their total PA levels, and then the middle 2 quartiles were collapsed, forming 3 groups. The mean CRP differences were compared across the PA categories using regression analysis and contrast coding. Likewise, to examine the influence of vigorous-intensity PA, participants were divided into 3 vigorous-intensity duration categories: no vigorous-intensity PA (Level 1), 10 to 59 min per week (Level 2), and 60 min or more per week (Level 3) and then the mean CRP differences were compared across the PA categories using regression analysis and contrast coding. The affect of potentially confounding factors (age, BMI, abdominal circumference, and body fat percentage) on the PA and CRP associations, were analyzed using partial correlation. Results In this investigation, 211 women were studied. Participants were primarily white (~90%), middle-aged (43.1 ± 3.0 y), ~50% attended at least some college, ~60% were employed part-time or full-time, and ~80% were married. Table 1 shows participants weight, BMI, abdominal circumference, and body fat percentage. Participants average CRP concentration was 1.28 ± 1.70 mg/l, and Table 1 Descriptive Information for All Participants (n = 211) Variables Mean SD 25th percentile 50th percentile 75th percentile Low High Ages (years) Weight (kg) BMI (kg/m 2 ) Abdominal circumference (cm) Body fat (%) Physical activity* CRP (mg/l) *Average weekly activity counts measured objectively using accelerometers, divided by 1000.

4 484 LeCheminant, Tucker, and Russell the average weekly activity count was 2.6 million ± 0.9 million counts. With both of the key variables treated as continuous measures, total PA was significantly and inversely related to CRP concentrations (r = 0.19, P =.006). In addition, Pearson product-moment coefficients indicated that CRP was significantly correlated with abdominal circumference (r =.37, P <.0001), BMI (r =.38, P <.0001) and body fat percentage (r =.44, P <.0001). Further, regression analysis showed that for each 100,000 count increase in total PA, there was a decrease of mg/l of CRP (F = 7.76, P =.006). As shown in Table 2, when the data were analyzed according to PA quartiles with the middle 2 quartiles combined, CRP differed significantly across PA categories (F = 5.06, P =.026). Particularly, the women in the highest total PA category (Level 3) had lower CRP levels than women in the moderate (Level 2) and low (Level 1) PA categories. Control of age did not change this relationship (F = 5.19, P =.024). However, controlling for body fat percentage weakened the relationship by 98% (F = 0.12, P =.725) and the association was no longer significant. Likewise, controlling for differences in abdominal circumference weakened the association by 61% (F = 2.23, P =.137), whereas controlling for BMI weakened the relationship by 45% (F = 3.19, P =.076). When PA was analyzed according to moderateintensity PA only, and both moderate-intensity PA and CRP were treated as continuous variables, there was no significant relationship (r = 0.015, P =.8327). Thus, further analysis of the relationship between moderateintensity activity and CRP was not conducted. However, when analyzed according to vigorous-intensity PA only, there was a statistically significant relationship with CRP (r = 0.20, P =.003). Age did not change the relationship between vigorous-intensity PA and CRP; however, BMI (P =.065) significantly weakened this relationship as did abdominal circumference (P =.106) and body fat percentage (P =.990). As shown in Table 3, with vigorous-intensity PA divided into 3 categories: no vigorous-intensity PA (Level 1), 10 to 59 min per week (Level 2), and 60 or more min per week (Level 3), there was a significant difference in CRP (F = 12.06, P < 0.001). Particularly, women in the highest vigorous-intensity duration category had lower CRP levels than women in the moderate and low duration vigorous-intensity PA categories. Control of age did not change this relationship (F = 12.17, P <.001). Controlling for body fat percentage weakened the relationship to the point of nonsignificance (F = 0.86, P =.354). However, controlling for differences in abdominal circumference weakened the relationship (F = 5.08, P =.025) as did BMI (F = 5.35, P =.022) but neither to the point of nonsignificance. Discussion Results of the current study indicate that there is a significant relationship between objectively measured PA and CRP in healthy, middle-age women before adjusting for body fat, abdominal circumference, or BMI. Utilization of an objective measure of PA represented a significant strength of this study as it eliminates some potential bias associated with self-reported PA data, such as recall bias and over- or under-reporting. It is noteworthy that the population used in the current study was healthy and tended toward low CRP levels. As a continuous relationship between CRP level and risk of CVD is likely, examining the influence of PA on CRP is warranted in both healthy and nonhealthy populations. Other investigators examining the relationship between PA and CRP reported similar results to the current study, with and without consideration of body composition in the analysis. Reviews by Kasapis and Thompson 14 and Plaisance et al, 6 of multiple crosssectional, longitudinal, as well as training studies indicated that lower CRP levels were associated with habitual PA; however, it is noteworthy that virtually all the cross-sectional studies reviewed used a self-reported Table 2 Differences in CRP Levels by Total Physical Activity Categories Physical activity category Level 1 (n = 53) Level 2 (n = 104) Level 3 (n = 54) Variable controlled Mean (SD) Mean (SD) Mean (SD) F P CRP level None 1.36 (1.83) a,b 1.43 (1.89) a 0.90 (1.02) b Age (years) 1.36 a,b 1.43 a 0.90 b BMI (kg/m 2 ) Abdominal circumference (cm) Body fat (%) Note. Physical activity categories represent quartiles, with the 2 middle quartiles combined. Level 1 = lowest total PA category; Level 2 = moderate total PA category; Level 3 = highest total PA category. Means on the same row with different superscript letters are significantly different (P <.05). On rows showing that a variable was controlled, means are adjusted means. a,b Indicates that the Level 1 category was not statistically different than the Level 2 or Level 3 categories.

5 Physical Activity and CRP 485 Table 3 Differences in CRP Levels by Vigorous-Intensity Physical Activity Duration Vigorous-intensity physical activity category Level 1 (n = 126) Level 2 (n = 39) Level 3 (n = 46) Variable controlled Mean (SD) Mean (SD) Mean (SD) F P CRP level None 1.42 (1.85) a 1.42 (1.76) a 0.76 (1.0) b <0.001 Age (years) 1.42 a 1.42 a 0.76 b <0.001 BMI (kg/m 2 ) 1.42 a 1.42 a 0.76 b Abdominal circumference (cm) 1.31 a,b 1.48 a 1.00 b Body fat (%) Note. Level 1 = no vigorous-intensity PA; Level 2 = min of vigorous-intensity PA; Level 3 = 60 min of vigorous-intensity PA. Means on the same row with different superscript letters are significantly different (P <.05). On rows showing that a variable was controlled, means are adjusted means. a,b Indicates that the Level 1 category was not statistically different than the Level 2 or Level 3 categories. assessment of PA. With regard to intervention studies, Plaisance and Grandjean suggested that the majority of training studies they reviewed resulted in CRP reductions, 6 but they acknowledged the difficulty in determining the independent effect of PA on CRP without weight loss because many studies used BMI or other less precise methods of assessing body composition. 6 Further, Mora et al examined the cross-sectional relationship between multiple cardiovascular biomarkers, including CRP, and self-reported PA in 27,158 women. 16 The lower quintiles of energy expended in PA per week showed significantly higher CRP levels with and without adjustment for several variables, including BMI. 16 However, Mora et al noted that when CRP levels were categorized by quintiles of BMI, individuals in the highest quintile of BMI (>29.3 kg/m 2 ) had a much higher CRP level than the lowest quintile of PA, with and without statistical adjustments. They concluded that BMI was more strongly associated with CRP and other CV biomarkers than PA. 16 This study had the additional strength of highlighting the relationship between intensity of PA and CRP. These data showed that vigorous-intensity PA of at least 60 min per week was an important predictor of lower CRP levels; however, moderate-intensity PA was not significantly associated with CRP levels. As noted in the methods section, vigorous-intensity PA was associated with activities equivalent to >4 mph. This is equivalent to a minimum of fast walking. Sixty minutes plus per week of vigorous-intensity PA is equivalent to the recommendation from the American College of Sports Medicine to obtain 20 minutes of vigorous-intensity PA 3 times per week or more 24 and roughly equal to recent recommendations by the US Department of Health and Human Services to obtain 75 minutes of vigorous-intensity PA per week for more substantial health benefits. 31 As with total PA, it is noteworthy that body fat percentage eliminated the relationship between vigorous-intensity PA and CRP and additional studies are needed to elucidate this relationship. However, this study highlights the potential benefit of accumulating higher intensity PA and is consistent with other findings from intervention studies. For example, Pitsavos et al reported that among subjects who exercised more than once per week, those engaging in high-intensity activities (>7 kcal/min) had 38% lower CRP levels than those engaging in light-intensity activities (<4 kcal/min). 32 In the current study, there was a significant linear relationship between CRP and body fat percentage (r =.44, P =.0001), suggesting that high levels of CRP are partly a function of high levels of body fat. There is a great deal of literature that supports these findings. Hence, the potential influence of body fat percentage on CRP, and outcomes related to CVD, are not surprising. For example, Rogowski et al reported a substantial decrease in the association between the 10-year Framingham Coronary Heart Disease Risk Score and CRP after controlling for BMI in men and women. 33 Visser et al showed higher CRP levels as BMI increased in young adults. 34 In addition, Lemieux et al suggested that obesity and abdominal fat are the most significant contributors to elevated plasma CRP in men. 35 Accordingly, as with the current study, it is reasonable to assume that body fat percentage is an important factor when considering the relationship between PA and CRP and seems to account for a large portion of the variance in this relationship. This is supported by other investigations, particularly longitudinal and intervention studies. Rawson et al, longitudinally examined the relationship among changes in BMI, self-reported PA, and CRP over 1 year. 18 These researchers reported that BMI and CRP were unchanged during the study and only BMI was significantly associated with CRP rather than previous PA or change in the participants current PA level. 18 Further, CRP was higher in the obese individuals compared with overweight or normal weight individuals. 18 Merrill et al reported the influence of a 6-month, education-based community PA intervention on CRP. 17 They found that at baseline, only BMI significantly predicted CRP, accounting for 28% of the variance. In addition, at 6 weeks and 6 months, the PA intervention was not associated with CRP. Of note, obese individuals were 3.1, 1.3, and 1.3 times more likely to have high

6 486 LeCheminant, Tucker, and Russell CRP (>3 mg/l) at baseline, 6 weeks, and 6 months than normal weight individuals. 17 Verdaet et al reported that the association between CRP and various levels of leisure time PA in men was explained by differences in BMI or smoking. 19 Hamer and Steptoe reported the association between physical fitness and CRP at baseline and a 3-year follow-up in 176 healthy middle-aged men and women. They found that BMI and weight gain, but not fitness, were related to CRP. 20 In the current study, it is noteworthy that body fat percentage weakened the relationship between total PA and CRP and vigorous-intensity PA and CRP to a much greater extent than abdominal circumference and BMI. These data indicate that abdominal circumference and BMI may not be as sensitive of measures as body fat percentage and may not play as important a role in the PA and CRP relationship compared with body fat percentage. This becomes an important methodological issue in identifying the independent relationship between PA and CRP and deserves further attention. The current study was limited by its cross-sectional design. Without the use of a randomized controlled design, cause-and-effect conclusions are not warranted. Although PA may have a direct effect on CRP levels, given the design of this study, other factors could influence the relationship between activity and CRP. However, because potentially confounding factors were controlled statistically in this study, the influence of body fat percentage on the PA and CRP association was identified. The current study suggests that high PA levels are significantly related to lower CRP levels in healthy, middle-age women, and 60 min or more per week of vigorous-intensity PA, but not duration of moderateintensity PA, was associated with lower levels of CRP; however, this relationship is almost entirely a result of differences in body fat percentage. In other words, physically active women tend to have lower levels of CRP, not necessarily because they are physically active, but may be because physically active women tend to be leaner. It is noteworthy that the participants in the current study were healthy and tended to have low CRP levels. It is possible that PA would have a greater impact on CRP (and CVD risk), in women with higher initial CRP levels, regardless of body composition. Nevertheless, as with other studies, the current study indicates that decreased body fat may be the critical factor for lowering levels of CRP and perhaps in reducing the risk of heart disease. Additional research to evaluate the intricacies of this relationship is warranted. Specifically, randomized, controlled trials would be of particular value. Moreover, experiments using only lean or only obese participants that use objective measures of PA and precise measures of body composition would help to ascertain the extent to which PA and exercise influence CRP levels directly. References 1. Libby P. Inflammation in atherosclerosis. Nature. 2002;420: Shishehbor MH, Bhatt DL. Inflammation and atherosclerosis. Curr Atheroscler Rep. 2004;6: Ridker PM, Silvertown JD. Inflammation, C-reactive protein, and atherothrombosis. J Periodontol. 2008;79: Ridker PM, Hennekens CH, Buring JE, Rifai N. C-reactive protein and other markers of inflammation in the prediction of cardiovascular disease in women. N Engl J Med. 2000;342: Abramson JL, Vaccarino V. Relationship between physical activity and inflammation among apparently healthy middle-aged and older US adults. Arch Intern Med. 2002;162: Plaisance EP, Grandjean PW. Physical activity and highsensitivity C-reactive protein. Sports Med. 2006;36: Ridker PM, Rifai N, Rose L, Buring JE, Cook NR. Comparison of C-reactive protein and low-density lipoprotein cholesterol levels in the prediction of first cardiovascular events. N Engl J Med. 2002;347: Patrick L, Uzick M. Cardiovascular disease: C-reactive protein and the inflammatory disease paradigm: HMG- CoA reductase inhibitors, alpha-tocopherol, red yeast rice, and olive oil polyphenols. A review of the literature. Altern Med Rev. 2001;6: Centers for Disease Control and Prevention. Physical activity and the prevention of coronary heart disease. Morbidity and Mortality Weekly Report. 1993;42: Wood PD. Physical activity, diet, and health: independent and interactive effects. Med Sci Sports Exerc. 1994;26: Sesso HD, Paffenbarger RS, Jr, Lee IM. Physical activity and coronary heart disease in men: The Harvard Alumni Health Study. Circulation. 2000;102: Blair SN, Kohl HW, Barlow CE. Physical activity, physical fitness, and all-cause mortality in women: do women need to be active? J Am Coll Nutr. 1993;12: Blair SN, Jackson AS. Physical fitness and activity as separate heart disease risk factors: a meta-analysis. Med Sci Sports Exerc. 2001;33: Kasapis C, Thompson PD. The effects of physical activity on serum C-reactive protein and inflammatory markers: a systematic review. J Am Coll Cardiol. 2005;45: Majka DS, Chang RW, Vu TH, et al. Physical activity and high-sensitivity C-reactive protein: the multi-ethnic study of atherosclerosis. Am J Prev Med. 2009;36: Mora S, Lee IM, Buring JE, Ridker PM. Association of physical activity and body mass index with novel and traditional cardiovascular biomarkers in women. JAMA. 2006;295: Merrill RM, Massey MT, Aldana SG, Greenlaw RL, Diehl HA, Salberg A. C-reactive protein levels according to physical activity and body weight for participants in the coronary health improvement project. Prev Med. 2008;46:

7 Physical Activity and CRP Rawson ES, Freedson PS, Osganian SK, Matthews CE, Reed G, Ockene IS. Body mass index, but not physical activity, is associated with C-reactive protein. Med Sci Sports Exerc. 2003;35: Verdaet D, Dendale P, De Bacquer D, Delanghe J, Block P, De Backer G. Association between leisure time physical activity and markers of chronic inflammation related to coronary heart disease. Atherosclerosis. 2004;176: Hamer M, Steptoe A. Prospective study of physical fitness, adiposity, and inflammatory markers in healthy middleaged men and women. Am J Clin Nutr. 2009;89: Bassett DR, Jr, Ainsworth BE, Swartz AM, Strath SJ, O Brien WL, King GA. Validity of four motion sensors in measuring moderate intensity physical activity. Med Sci Sports Exerc. 2000;32:S471 S Trost SG, Ward DS, Moorehead SM, Watson PD, Riner W, Burke JR. Validity of the computer science and applications (CSA) activity monitor in children. Med Sci Sports Exerc. 1998;30: Nelson T, Leenders N, Sherman WM. Comparison of activity monitors worn during treadmill walking. Med Sci Sports Exerc. 1998;30:11S. 24. Haskell WL, Lee IM, Pate RR, et al. Physical activity and public health: updated recommendation for adults from the American College of Sports Medicine and the American Heart Association. Med Sci Sports Exerc. 2007;39: LeCheminant JD, Tucker LA, Bailey BW, Peterson TR. The relationship between intensity of physical activity and HDL cholesterol in 272 women. J Phys Act Health. 2005;3: Massarrat A. Alpha Diagnostic International, Inc. Accessed October 12, Bailey BW, Tucker LA, Peterson TR, LeCheminant JD. Test-retest reliability of body fat percentage results using dual energy X-ray absorptiometry and the Bod Pod. Med Sci Sports Exerc. 2001;33:174S. 28. LeCheminant JD, Tucker LA, Peterson TR, Bailey BW. Differences in body fat percentage measured using dual energy X-ray absorptiometry and the Bod Pod in 100 women. Med Sci Sports Exerc. 2001;33:174S. 29. Maddalozzo GF, Cardinal BJ, Snow CA. Concurrent validity of the BOD POD and dual energy x-ray absorptiometry techniques for assessing body composition in young women. J Am Diet Assoc. 2002;102: Noreen EE, Lemon PW. Reliability of air displacement plethysmography in a large, heterogeneous sample. Med Sci Sports Exerc. 2006;38: US Department of Health and Human Services Physical Activity Guidelines for Americans. health.gov/paguidelines. Accessed September 24, Pitsavos C, Chrysohoou C, Panagiotakos DB, et al. Association of leisure-time physical activity on inflammation markers (C-reactive protein, white cell blood count, serum amyloid A, and fibrinogen) in healthy subjects (from the ATTICA study). Am J Cardiol. 2003;91: Rogowski O, Shapira I, Toker S, et al. Obesity-related correlation between C-reactive protein and the calculated 10-y Framingham Coronary Heart Disease Risk Score. Int J Obes. 2005;29: Visser M, Bouter LM, McQuillan GM, Wener MH, Harris TB. Elevated C-reactive protein levels in overweight and obese adults. JAMA. 1999;282: Lemieux I, Pascot A, Prud homme D, et al. Elevated C-reactive protein: another component of the atherothrombotic profile of abdominal obesity. Arterioscler Thromb Vasc Biol. 2001;21: