Journal Logo


Aerobic Fitness and Mode of Travel to School in English Schoolchildren


Author Information
Medicine & Science in Sports & Exercise: February 2010 - Volume 42 - Issue 2 - p 281-287
doi: 10.1249/MSS.0b013e3181b11bdc
  • Free


Cycling to school is positively associated with aerobic fitness (1,10). Switching to cycling as the dominant mode of travel to school also improves aerobic fitness (9). Such associations have, however, only been demonstrated in Danish schoolchildren. In Denmark, two-thirds of adolescents cycle to school (1,9,10), which differs greatly from most western countries, such as the United Kingdom, the United States, and Australia, where cycling to school is uncommon (13,19,31). Active school commuting is associated with higher physical activity (PA) levels (7,8,27,30), which has led to many initiatives promoting active travel to school in an attempt to increase PA and encourage a healthy and active lifestyle (3,12). It remains to be established, however, if active travel to school is positively associated with aerobic fitness and body mass index (BMI) in countries where cycling is uncommon. The aim of the present study was to evaluate the association between school commuting habits and aerobic fitness in English schoolchildren.



The sample was representative of the east of England in ethnic mix, rural/urban dwelling, and socioeconomic status. In total, 23 schools participated in the study that was approved by the university's ethics committee. After obtaining parental consent, measurements were made during scheduled physical education classes in the summer of 2007 and 2008. In total, 7914 schoolchildren have participated in the East of England Healthy Hearts study, but only data from participants providing information on sex, age (10.0-15.9 yr), stature, body mass, 20-m shuttle run test performance (20-mSRT), travel mode, and home postcode are presented here (n = 6085, 47% girls).


Travel mode.

Schoolchildren were asked to self-report and to answer the question, "How do you usually get to school?" and given the following options: "walk," "cycle," "public transport," "car" or "other." Participants also provided their home postcode; school postcode and home postcodes were entered into Google maps to estimate the distance they traveled to school.


Stature was measured to the nearest 1 mm, and mass was measured to the nearest 0.1 kg with participants dressed in standard physical education clothing without shoes. BMI was calculated (kg·m−2) and converted to z-scores on the basis of UK reference data (6). BMI was categorized as underweight, normal weight, overweight, or obese according to the criteria of the International Obesity Task Force (IOTF), which use age- and sex-specific cutoffs that correspond to BMI values of 18.5, 25, and 30 kg·m−2 at age 18 yr, respectively (4,5).


PA was assessed by a 7-day recall using the Physical Activity Questionnaire (PAQ) for children or adolescents, depending on the participants' age (15). The questionnaire measures general levels of PA through a summary score from questions about sports, activities during and after school, and activities at weekends. The PAQ is scored on a 5-point scale, with an overall score of 1 indicating low PA and a score of 5 indicating high PA.

Aerobic fitness.

Participants with known medical conditions that could endanger them during maximal exertion, along with those with current acute infections or injuries at the time of testing, were prohibited from participating in the aerobic fitness test. Aerobic fitness was assessed using the 20-mSRT. The 20-mSRT was administered in the form of the FITNESSGRAM PACER (20), a modified version of the original protocol (16) as described previously (26). All participants had previously taken part in the 20-mSRT as part of their physical education. Participants were reminded by both the instructions on the PACER CD and a researcher to "run for as long as possible." The test requires volunteers to run back and forth over a marked distance of 20 m in time with an audible signal. The test starts at an initial running speed of 8.5 km·h−1 and increases by 0.5 km·h−1·min−1. Researchers acted as "spotters" and recorded the final shuttle count at either the point of volitional exhaustion or when the participant failed to maintain the required running speed twice. Final shuttle count was converted first to final running speed and then into z-scores on the basis of global performance indices (22). V˙O2max was predicted on the basis of final running speed and age (16). Lower FITNESSGRAM PACER Healthy Fitness Zone cutoffs (21) were used to categorize participants: if the participants' total completed shuttle count was above their age- and sex-specific cutoff, they were classed as "fit"; otherwise, participants were classed as "unfit."

Statistical Analyses

Descriptive (mean, SD) and frequency (%) statistics were calculated for relevant variables. Participants who selected "other" or more than one of the transport modality options (n = 158; 2.8%) were excluded from further analysis, leaving the following four transport categories: cycle, walk, public transport, or car. The sample was split by sex for statistical analyses. Pearson's χ2 analysis was applied to assess differences in travel mode by sex. One-way ANOVA with post hoc Tukey tests were used to analyze between-group differences in mean BMI z-scores and 20-mSRT z-scores by travel mode (cycle, walk, public transport, and car). The analysis was repeated using PAQ score as a covariate (ANCOVA) on a smaller sample (n = 5577, 94%) representing participants with complete PACER and PAQ scores. To enable comparison with previous studies, the public transport and car categories were subsequently combined to create a "passive transport" group. Binary logistic regression analysis was used to calculate odds ratios (OR) for the likelihood of being classed as "fit" by travel mode (cycle, walk, passive transport) using passive transport as the reference category for each sex. BMI z-scores and age in whole years were added to the regression model as covariates to adjust for their respective contributions to the shared variance. The same analyses were repeated after further adjusting for PA by adding the PAQ score to the model. Statistical significance was set at P < 0.05. SPSS 16.0 for Windows (Chicago, IL) was used for all statistical analyses.



Descriptive statistics are reported in Tables 1 and 2. Five percent of the schoolchildren were underweight, 69% were normal weight, 21% were overweight, and 5% were obese.

Anthropometric characteristics (mean ± SD).
Aerobic fitness test performance (mean ± SD).

Travel mode

Table 3 shows that nearly half (49.9%) of the schoolchildren surveyed walked to school. Approximately one-quarter (25.7%) used public transport, 16.4% traveled by car, and 8.0% cycled. Pearson's χ2 analysis revealed that there were significant between-sex differences in travel mode (χ2 = 218.21, P < 0.001). Bonferroni-corrected χ2 analysis showed that significantly more boys (12.9%) cycled than girls (2.5%).

Travel mode frequency (% (n)).

Anthropometric measurements, PA questionnaire, and aerobic fitness test performance.

Detailed results are reported in Table 4 for girls in and Table 5 for boys. One-way ANOVA showed no significant between-group differences in BMI z-scores in girls (F = 2.157, P = 0.091) or boys (F = 0.653, P = 0.581). PAQ scores were significantly different between groups in girls (F = 8.376, P < 0.001) and boys (F = 18.587, P < 0.001), and post hoc (Tukey) tests identified that cyclists in both sexes scored significantly higher than any other group (both sexes: P < 0.001). ANOVA also revealed significant differences in mean 20-mSRT z-score between travel modes in boys (F = 4.296, P = 0.005) and girls (F = 5.344, P = 0.001). Post hoc (Tukey) tests showed that mean 20-mSRT z-score was significantly higher in female cyclists compared with public transport (P = 0.049) and car users (P = 0.006) and in walkers when compared with car users (P = 0.012). In boys, mean 20-mSRT z-score was significantly higher in cyclists (P = 0.020) and walkers (P = 0.013) when compared with car users. These between-group differences remained significant when the analyses were controlled for PA (ANCOVA) in girls (F = 3.228, P = 0.022) and boys (F = 2.925, P = 0.003).

Girls' characteristics (mean ± SD) by travel mode.
Boys' characteristics (mean ± SD) by travel mode.

Logistic regression.

Table 6 provides OR for fitness categorization by travel mode calculated by logistic regression. Girls who walked to school were more likely to be classed as fit (OR = 1.31, 95% confidence interval (CI) = 1.03-1.66, P = 0.027) compared with those who used passive transport, and girls who cycled had nearly 10 times the odds for being fit (OR = 9.99, 95% CI = 1.30-76.59, P = 0.027). Compared with those who used passive transport, boys who walked were more likely to be classified as fit (OR = 1.20, 95% CI = 1.00-1.43, P = 0.045), as were cyclists (OR = 1.31, 95% CI = 1.00-1.72, P = 0.049). OR for PA-adjusted fitness categorization by travel mode are reported in Table 6. The OR for being fit in girls who walked to school changed very little (OR = 1.34, 95% CI = 1.05-1.71, P = 0.018). The OR were reduced slightly in girls who cycled, but these girls were still eight times morel likely to be classified as fit compared with passive transport users (OR = 7.94, 95% CI = 1.05-60.23, P = 0.045). In boys, however, there was no significant association between travel mode and fitness category after adjusting for PA.

OR (95% CI) for likelihood of being classed as "fit" by travel mode.


Travel mode and aerobic fitness.

Cycling to school is associated with better aerobic fitness in Danish schoolchildren (1,9,10). The aim of the present study was to determine whether school travel mode is associated with aerobic fitness in English schoolchildren. In common with previous studies (1,9,10), girls who regularly cycled to school had significantly higher mean aerobic fitness test scores when compared with public transport and car users. All girls, except for one, who cycled to school were categorized as fit, a significantly greater proportion than in any other travel mode group. Girls who cycled were 10 times more likely than passive transport users to be classified as fit (OR = 9.99, 95% CI = 1.30-76.59). The large CI in the female cyclists' group were mainly due to the low number of individuals in this group (n = 71). Similarly, boys who cycled had a higher mean aerobic fitness test score when compared with boys that were driven to school. In agreement with previous studies (1,9,10), male cyclists were significantly more likely to be classed as fit compared with passive transport users.

No previous data have demonstrated an association between walking to school and aerobic fitness. In the present study, however, girls and boys who regularly walked to school were fitter than those who were driven. Boys and girls who walked to school were also significantly more likely to be categorized as fit compared with passive transport users.

The present data agree with existing literature in showing that cycling to school is associated with higher levels of aerobic fitness but are the first to show such an association with walking to school. The novelty of this finding is most probably the result of differing national travel habits, but the possibility remains that methodological differences in the assessment of aerobic fitness may have also had an impact. The predominant travel mode to school in Denmark is cycling, where cycling is positively associated with aerobic fitness (1,9,10). In the present English sample, however, the predominant mode of transport is walking. The current data do show that walkers were fitter than passive transport users. Cyclists were, however, fitter still. Commuter cycling tends to be a more intense form of exercise than walking. These data, therefore, support the previous suggestion (10) that exercise intensity of the active commuting mode may play a role in the strength of the association between travel mode and aerobic fitness.

Sample characteristics.

Reported school travel habits were broadly similar to UK national averages (13), except that walking and cycling were slightly overrepresented. All existing data showing associations between cycling to school and aerobic fitness (1,9,10) come from schoolchildren in Denmark (Danish Youth Heart Study, European Youth Heart Study), where two-thirds of adolescents regularly cycle to school. Denmark has good infrastructural provision and a strong culture of cycling (10). Cycling to school is rare in the United Kingdom (13) and even less common in the United States (19) and Australia (31). The relative infrequency of cycling necessitated the large overall sample size of this study (n = 5927). Previous studies may have had larger relative and absolute samples of cyclists, but the present data allow more powerful analysis of aerobic fitness of walkers and public transport users.

Participants of the present study were broadly similar to English reference data (34) in the prevalence of overweight and obesity: roughly a quarter of children and adolescents were overweight and obese. These rates are much greater than those seen in Denmark, where only a sixth of children and adolescents are overweight and obese (18). This may allow some generalization of the present findings to countries such as the United States and Australia, where active school travel by bicycle is low, and the obesity epidemic and associated health burdens are equally prevalent.

Aerobic fitness versus PA as a marker of health.

Several studies have shown a positive association between active travel to school and PA levels in children (7,8,27) and adolescents (30). Some studies (9,10) have reported conflicting data with regard to the association between PA and school travel mode. Cyclists have been reported as being less physically active than walkers yet still having higher aerobic fitness. Such a paradox illustrates the challenges associated with measuring PA in schoolchildren. In particular, it demonstrates the difficulty in estimating the PA of cyclists via accelerometry.

Aerobic fitness in children and adolescents is arguably a better marker of health than PA because of its strong associations with adiposity, current and future cardiovascular health, and psychological well-being (23). Poor aerobic fitness in children and adolescents is strongly associated with the clustering of cardiovascular disease risk factors (2) and metabolic risk (25). In addition, the objective assessment of aerobic fitness in children and adolescents is less challenging than objectively assessing PA. In this respect, it is perhaps surprising that the present data add to only a few studies (1,9,10) that have investigated associations between travel mode and aerobic fitness. As in other studies (14,28), the present study did not identify any associations between travel mode and BMI.

Measurement of aerobic fitness.

All associations between school travel and aerobic fitness (1,9,10) have been shown using maximal cycle ergometry tests, which have some obvious advantages. Maximum power output corrected for body mass (9,10) is a valid and reliable way to assess maximum aerobic fitness in adolescents (24). It has been suggested that most Danish children and adolescents are accustomed to regular cycling, even if they do not commute to school by bike (10). Because of this, it is unlikely that noncycle commuters were disadvantaged by cycle ergometry in previous studies. In England, however, cycling to school is much less common, and recreational cycling habits in this sample are presently unknown. Cycle ergometry would, therefore, be inappropriate as a method to assess aerobic fitness in the present sample. The present study assessed aerobic fitness using the 20-mSRT. The test uses running, which is an exercise modality that all children and adolescents are accustomed to. The test is a valid (16,17,32) and easily administered field test that provokes a maximal effort in schoolchildren (33). It is extensively used to assess aerobic fitness of children and adolescents (22,29). Global normative data exist for the 20-mSRT performance allowing representation of test scores as z-scores relative to a global mean (22). Danish adolescents perform better in the 20-mSRT than the global mean (22), whereas the performance of the present sample was much closer to the global mean (Table 2), thus adding to the generalizability of the present sample.

Active travel to school and health.

Previous data (10) show that schoolchildren who cycle to school are more likely to be fit than those that use passive transport. The present study adds to the existing data in demonstrating that not only cycling but also walking is associated with an increased likelihood of being fit (Table 6). More importantly, perhaps, the present study does not use arbitrary cut points to define fitness categories. Existing 20-mSRT cut points (21) were used, which are directly associated with adult health (11). Schoolchildren who were categorized as fit in the present study may enjoy a reduced risk for developing chronic disease in adulthood. Using such criteria, however, also has drawbacks. The observed associations between travel mode and fitness were notably weaker than those reported previously (10), with some relatively small OR and 95% CI approaching 1.0 (Table 6). These are likely the result of the fitness classification criteria applied. The unfit category in the present study roughly corresponds to performance predictive of the bottom quintile of adult fitness, whereas previous data (10) classified low fitness on the basis of belonging to the lowest quartile within the study population. The criteria used presently may have reduced the strength of some associations. The meaningfulness of such cutoffs in generalizability of results, however, justifies their use.

PA as a potential mechanism.

In agreement with previous studies (7,8,27,30), cyclists were more physically active than pupils using other transport modes (Tables 4 and 5). In boys, the OR for active commuters being classed as fit was nonsignificant after adjusting for this higher PA (Table 6). It may be, therefore, that active school commuting is linked to higher fitness in boys via higher levels of overall PA. In girls, however, the association between travel modality and fitness remained largely unchanged when controlling for PA (Table 6), which suggests that travel was associated with fitness independently of PA. This may be because girls reported lower overall PA than boys, and as a result, the PA associated with active school commuting contributes directly to increased fitness in girls. Investigations on potential between-sex differences in the mechanism by which active school travel is associated with fitness, either independently of or via PA, are warranted. In the present study, PA was derived from 7-day recall, which has numerous associated limitations. However, a noteworthy strength of the current measure is that the PAQ score is largely independent of any contributions from active commuting: school travel is not assessed, and daily cycling and/or walking each account for less than a percent of the total PAQ score.

Limitations and recommendations.

The present sample was not randomly selected or nationally representative. The sample was designed to represent the East of England in ethnicity, deprivation, and rural/urban mix. This region is, in turn, representative of the United Kingdom in ethnicity and rural/urban mix but is more affluent than the UK average. The sample also lacks large numbers of children from within large-city areas (defined as a population of >250,000), which reduces the generalizability of the present findings to such populations.

Total and relative frequencies of cycling were low in the present study compared with previous studies, despite a much larger sample size. This was particularly evident in girls, where this created large CI in the logistic regression analysis (Table 6). Only 2.5% of girls regularly cycled to school, and although this number seems low, it should be borne in mind that the rate is greater than the UK national average for girls (1%) (13). The phenomenon of low numbers of schoolchildren cycling to school (particularly girls) in the United Kingdom warrants further investigation.

As noted previously, PA was assessed by a 7-day recall, which has obvious limitations. The questionnaire only provides a general estimate on the basis of self-report and does not provide any objective measurements. The PAQ was, however, specifically designed to enhance participants' recall ability and has been deemed a valid and reliable tool to estimate PA in large populations (15).

Naturally, because of the cross-sectional nature of the present study design, the direction of causality of the presently identified association between active travel to school and better aerobic fitness cannot be commented on. The present study does, however, provide some insight into possible mechanisms underlying this association and some potential differences between boys and girls. Further studies aimed at elucidating the mechanism by which active travel is associated with schoolchildren's fitness are required. Mean aerobic fitness and likelihood of being fit were higher in cyclists than in walkers; this supports previous speculations (10) that exercise intensity may be a mediating factor in this association. In light of the present findings, the efficacy of this theory clearly needs to be investigated further in boys and girls.


The present data report higher aerobic fitness in cyclists compared with passive transport users in schoolchildren outside Denmark. The present study population is more representative of other western countries, such as the United States and Australia, in school commuting habits and progression of the obesity epidemic than Denmark. Present findings thus allow better generalization to such countries. The present study also shows higher aerobic fitness in schoolchildren who walk to school compared with passive transport users. Regular cycling and walking to school were significantly associated with higher mean aerobic fitness test scores. More importantly still, the present data provide a link between active travel to school and future health outcomes. Cycling and walking to school significantly increased the odds for being fit, and the cutoff used to define fit was directly linked to a reduced risk to develop chronic diseases in adulthood. The association between active travel to school and fitness, especially cycling, was particularly evident in girls. These data suggest potential mechanisms through which this association is achieved. There were between-sex differences in the role of PA. Boys who actively commuted seemed to be fitter as a result of higher overall PA, whereas in girls, the association between active travel and fitness was independent of PA. Regardless of the underlying mechanisms, the positive associations between active travel and fitness are so strong that cycling should be encouraged, especially in girls.

This study was funded by the University of Essex Research Promotion Fund and was not externally reviewed. The results of the present study do not constitute endorsement by American College of Sports Medicine.


1. Andersen LB, Lawlor DA, Cooper AR, Froberg K, Anderssen SA. Physical fitness in relation to transport to school in adolescents: the Danish youth and sports study. Scand J Med Sci Sports. 2009;19(3):406-11.
2. Anderssen SA, Cooper AR, Riddoch C, et al. Low cardiorespiratory fitness is a strong predictor for clustering of cardiovascular disease risk factors in children independent of country, age and sex. Eur J Cardiovasc Prev Rehabil. 2007;14(4):526-31.
3. Centers for Disease Control and Prevention. Barriers to children walking and biking to school-United States, 1999. JAMA. 2002;288(11):1343-4.
4. Cole TJ, Bellizzi MC, Flegal KM, Dietz WH. Establishing a standard definition for child overweight and obesity worldwide: international survey. BMJ. 2000;320(7244):1240-3.
5. Cole TJ, Flegal KM, Nicholls D, Jackson AA. Body mass index cut offs to define thinness in children and adolescents: international survey. BMJ. 2007;335(7612):194-201.
6. Cole TJ, Freeman JV, Preece MA. Body-mass index reference curves for the UK, 1990. Arch Dis Child. 1995;73(1):25-9.
7. Cooper AR, Andersen LB, Wedderkopp N, Page AS, Froberg K. Physical activity levels of children who walk, cycle, or are driven to school. Am J Prev Med. 2005;29(3):179-84.
8. Cooper AR, Page AS, Foster LJ, Qahwaji D. Commuting to school: are children who walk more physically active? Am J Prev Med. 2003;25(4):273-6.
9. Cooper AR, Wedderkopp N, Jago R, et al. Longitudinal associations of cycling to school with adolescent fitness. Prev Med. 2008;47(3):324-8.
10. Cooper AR, Wedderkopp N, Wang H, Andersen LB, Froberg K, Page AS. Active travel to school and cardiovascular fitness in Danish children and adolescents. Med Sci Sports Exerc. 2006;38(10):1724-31.
11. Cureton KJ, Plowman SA. Aerobic capacity assessments. In: Welk GJ, Meredith MD, editors. Fitnessgram/Activitygram Reference Guide. Dallas (TX): The Cooper Institute; 2008. p. 9-1-25.
12. Department for Transport. School Travel Strategies and Plans: A Best Practice Guide for Local Authorities. London (UK): Department for Transport; 1999. p. 1-72.
13. Department for Transport. Travel to School. Personal Travel Factsheet-March 2008. London (UK): Department for Transport; 2008. p. 1-4.
14. Faulkner GE, Buliung RN, Flora PK, Fusco C. Active school transport, physical activity levels and body weight of children and youth: a systematic review. Prev Med. 2009;48(1):3-8.
15. Kowalski KC, Crocker RE, Donen RM. The Physical Activity Questionnaire for Older Children (PAQ-C) and Adolescents (PAQ-A) Manual. Saskatoon (Canada): University of Saskatchewan; 2004. p. 1-37.
16. Leger LA, Mercier D, Gadoury C, Lambert J. The multistage 20 metre shuttle run test for aerobic fitness. J Sports Sci. 1988;6(2):93-101.
17. Liu NYS, Plowman SA, Looney MA. The reliability and validity of the 20-meter shuttle test in American students 12 to 15 years old. Res Q Exerc Sport. 1992;63(4):360-5.
18. Lobstein T, Frelut ML. Prevalence of overweight among children in Europe. Obes Rev. 2003;4(4):195-200.
19. McDonald NC. Active transportation to school-trends among US schoolchildren, 1969-2001. Am J Prev Med. 2007;32(6):509-16.
20. Meredith MD, Welk GJ. Chapter 5: Aerobic Capacity. FITNESSGRAM/ACTIVITYGRAM Test Administration Manual. 3rd ed. Champaign (IL): Human Kinetics; 2007. p. 27-34.
21. Meredith MD, Welk GJ. Chapter 9: Interpreting FITNESSGRAM Results. FITNESSGRAM/ACTIVITYGRAM Test Administration Manual. 3rd ed. Champaign (IL): Human Kinetics; 2007. p. 59-70.
22. Olds T, Tomkinson G, Leger L, Cazorla G. Worldwide variation in the performance of children and adolescents: an analysis of 109 studies of the 20-m shuttle run test in 37 countries. J Sports Sci. 2006;24(10):1025-38.
23. Ortega FB, Ruiz JR, Castillo MJ, Sjostrom M. Physical fitness in childhood and adolescence: a powerful marker of health. Int J Obes (Lond). 2008;32(1):1-11.
24. Riddoch C, Edwards D, Page A, et al. The European Youth Heart Study-cardiovascular disease risk factors in children: rationale, aims, study design, and validation of methods. J Phys Act Health. 2005;2:115-29.
25. Ruiz JR, Ortega FB, Rizzo NS, et al. High cardiovascular fitness is associated with low metabolic risk score in children: the European Youth Heart Study. Pediatr Res. 2007;61(3):350-5.
26. Sandercock G, Voss C, Gladwell V. Twenty-metre shuttle run test performance of English children aged 11-15 years in 2007: comparisons with international standards. J Sports Sci. 2008;26(9):953-7.
27. Sirard JR, Riner WF Jr, McIver KL, Pate RR. Physical activity and active commuting to elementary school. Med Sci Sports Exerc. 2005;37(12):2062-9.
28. Sirard JR, Slater ME. Walking and bicycling to school: a review. Am J Lifestyle Med. 2008;2:372-96.
29. Tomkinson GR, Olds TS. Secular changes in pediatric aerobic fitness test performance: the global picture. Med Sport Sci. 2007;50:46-66.
30. Tudor-Locke C, Ainsworth BE, Adair LS, Popkin BM. Objective physical activity of Filipino youth stratified for commuting mode to school. Med Sci Sports Exerc. 2003;35(3):465-71.
31. van der Ploeg HP, Merom D, Corpuz G, Bauman AE. Trends in Australian children traveling to school 1971-2003: burning petrol or carbohydrates? Prev Med. 2008;46(1):60-2.
32. Van Mechelen W, Hlobil H, Kemper HCG. Validation of 2 running tests as estimates of maximal aerobic power in children. Eur J Appl Physiol Occup Physiol. 1986;55(5):503-6.
33. Voss C, Sandercock G. Does the twenty meter shuttle-run test elicit maximal effort in 11-to 16-year-olds? Pediatr Exerc Sci. 2009;21(1):55-62.
34. Wardle J, Brodersen NH, Cole TJ, Jarvis MJ, Boniface DR. Development of adiposity in adolescence: five year longitudinal study of an ethnically and socioeconomically diverse sample of young people in Britain. BMJ. 2006;332(7550):1130-5.


©2010The American College of Sports Medicine