SEARCH

SEARCH BY CITATION

Keywords:

  • Accelerometry;
  • doubly labelled water;
  • intervention study;
  • obesity;
  • observational study;
  • weight gain

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conflict of Interest Statement
  8. Acknowledgement
  9. References

This study aimed at synthesizing the prospective associations between measured physical activity (PA) and change in adiposity in children, adolescents and adults following from two previous reviews. Search terms were adapted and a systematic literature search was conducted (January 2000–September 2008) and later updated (up to October 2009), considering observational and intervention studies of weight gain that measured both PA and body composition. Sixteen observational studies (six comprising adults) and five trials (one comprising adults) were eligible. For consistency, whenever possible either baseline PA energy expenditure or accelerometer output (counts min−1) and change in per cent body fat were the extracted exposure and outcome measures. Results of observational studies suggest that PA is not strongly prospectively related with adiposity: five studies on children and three on adults reported no association between baseline PA and change in adiposity, one study found a weak positive association and the other studies observed a weak negative association. Negative associations were more frequently observed in studies that analysed the association between change in the exposure and outcome. Intervention studies show generally no effect on either PA or adiposity. In conclusion, despite the well-established health benefits of PA, it may not be a key determinant of excessive gain in adiposity.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conflict of Interest Statement
  8. Acknowledgement
  9. References

Overweight and obesity are a global health problem (1,2). This is a major public health concern, as excess body weight is associated with many immediate and long-term health consequences in both children and adults (3–6).

Although gain in body weight and subsequent development of obesity is multifactorial, the fundamental physiological cause is a positive energy imbalance over time, caused by a higher energy intake than is expended. Hence, in theory, higher energy expenditure (EE) should help to maintain energy balance and therefore a stable body weight. Accordingly, it has been suggested that an increase in EE due to increased physical activity (PA) may protect against excess weight gain (5). In line with this, previous cross-sectional studies reported that lower levels of PA are related to a higher risk of obesity in children, adolescents and adults (7,8). However, the cross-sectional design of these studies hampers the interpretation of the results as it is not clear whether low levels of PA cause excess weight gain or whether overweight people are less likely to engage in PA. In contrast, the evidence for a prospective association between PA and the risk of excessive weight gain and developing obesity from intervention and observational studies is less clear (9,10). Most previous intervention studies and many of the observational studies assessed PA by questionnaires or recall interviews that are susceptible to recall bias and misclassification (11). Objective methods such as doubly labelled water, heart rate methods and accelerometry estimate total PA with both higher precision and higher accuracy than self-reports (12–16) and may be preferable for aetiological studies.

The increasing number of published studies using objective measurement methods for assessing PA in relation to change in adiposity makes it timely to scrutinize the results from these studies. Therefore, the aim of this review was to examine whether higher levels of baseline PA prevent excess gain in fat mass by synthesizing the evidence for the prospective association between objectively measured PA and subsequent change in adiposity in children, adolescents and adults taking into account both prospective observational and intervention studies.

Methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conflict of Interest Statement
  8. Acknowledgement
  9. References

Systematic literature search

This review is a systematic update of two previous reviews (10,17). The keywords were adopted from one of these reviews (10), although slightly adapted to fit current PubMed search terms. The search covers the associations between PA and body composition measures in children, adolescents and adults and considers both prospective observational studies and intervention studies that have been published since December 1999. In contrast to the two previous reviews that included only few studies with objective measurements of PA (10,17), the exposure (PA) had to be measured objectively, for instance by accelerometers or methods that allow assessing total daily EE such as doubly labelled water, indirect calorimetry or heart rate monitoring. Subtracting the basal metabolic rate and an estimate of the thermal effect of food from the total EE yields the physical activity energy expenditure (PAEE) (18). Accelerometers measure the acceleration of the body or the body segment to which it is attached. It provides an objective assessment of the total amount of time spent in PA, the PA intensity and the total body movement (accelerometer counts min−1) (19), the latter correlating well with PAEE measured by doubly labelled water (20). The outcome of interest was a body composition measure, assessed objectively, e.g. by anthropometrical methods, under water weighing or dual X-ray absorptiometry. For this review, the preferred exposure measure was either baseline PAEE or baseline total body movement (accelerometer counts min−1) and the preferred outcome measure was the change in per cent body fat. Therefore, whenever possible these were extracted. If they were not reported, surrogate measures for the PA level and adiposity were used. Studies were excluded if their samples were either limited to clinically ill participants, if they originated from trials involving intentional weight loss or if a clear separation between the effects of PA and nutrition in interventions was not possible.

PubMed was used for the literature search (1 January 2000 to 1 September 2008), applying the following text keywords: weight AND (gain OR change) OR maintenance OR management AND (exercise OR physical activity OR walking OR training) AND ‘humans’[MeSH Terms] AND English[lang] AND (‘longitudinal studies’[MeSH Terms] OR prospective OR randomized controlled trial[ptyp]). Two of the authors carried out the literature search and extracted data. When a title or an abstract could not be rejected with certainty, the full text of the article was obtained for further evaluation. A third person was consulted if there was uncertainty about the eligibility of studies. Where duplicate publications were located, either the first published paper or the one with the most comprehensive information was used. Based on this criterion, two studies were neglected (21,22).

The PubMed search yielded 3002 different studies and nine additional ones were found by hand searching reference lists. In total, 79 studies were considered potentially eligible. Of these, 10 observational studies (23–32) and four intervention studies (33–36) on children and adolescents met our inclusion criteria. Moreover, five eligible observational studies (37–41) and one intervention study (42) were identified reporting data in adults. An update of our initial search was carried out using the above-mentioned keywords (up to 29 October 2009). One additional observational study in adults was retrieved (43).

The eligible studies differed considerably in their design, quality and statistical analyses and it was not possible to quantitatively summarize the results by a conventional meta-analysis. Therefore, studies are reviewed in a narrative way.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conflict of Interest Statement
  8. Acknowledgement
  9. References

Children and adolescents – observational studies

Table 1 gives an overview of the ten observational studies reporting on the prospective associations between PA and estimates of adiposity in children (23–32). All studies were carried out in the USA, six of them involved children of mixed ethnicity (23,25,26,28,31,32) and the remaining four studies involved Caucasians (27,29), Hispanics (24) or Pima Indians (30). Four studies examined girls only (23,26,31,32), whereas all other studies analysed both boys and girls together. Children were aged between 3 and 12 years at baseline in all studies except for one study, in which the ages ranged between 4 and 19 years (24). The number of included children varied between 47 and 1576 at baseline, with four studies including <130 children (26,28,29,32) and two studies including >1000 children (24,31). Six studies reported dropouts of 15% or less (24,25,28–30,32), three studies reported a dropout rate of approximately 20% (23,26,27) and one study had a 38% dropout rate between baseline and follow-up (31). The duration of follow-up ranged between 1 and 8 years.

Table 1.  Prospective studies in children and adolescents
StudyPopulationFollow-upExposure*OutcomeModel (confounding)Size of effect
  • *

    EE was assessed by doubly labelled water at BL, except for (26), where whole room indirect calorimetry was used.

  • PAEE was calculated as 0.9 total EE – resting metabolic rate in all studies, except for two studies that calculated PAEE as total EE – resting metabolic rate – thermal effect of food (25) or 0.9 total EE – sleeping EE (26).

  • The outcome measure was assessed by dual X-ray absorptiometry in all studies except for four: these used bioimpedance or total body water (23), skin-fold callipers (29), dual X-ray absorptiometry along with 18O (30), or anthropometric measures along with published equations (31).

  • §

    The analysis is not truly longitudinal.

  • inline image no significant association; inline image significant association; Δ change; BF, body fat; BL, baseline; BMI, body mass index; BW, body weight; CI, confidence interval; EE, energy expenditure; FFM, fat free mass; FU, follow-up; LTM, lean total mass; MVPA, baseline moderate-to-vigorous PA; OR, odds ratio; PA, physical activity; PAEE, physical activity energy expenditure; SD, standard deviation; SSF, sum of skin-folds.

Bandini et al. (23)BL: 196 (FU: 156) 8- to 12-year-old pre-menarcheal healthy, non-obese girls of different ethnicity from Massachusetts, USA4 years after menarche; mean 7 years (SD 2.6 years)PAEE (adjusted for fat, fat free mass and age)Mean of %BF (repeated annually)Linear mixed effect modelling (age relative to menarche). Covariates: Model 1: none; Model 2: parental overweightBL PAEE was not associated with mean %BF. Model 1: β = 0.003; P = 0.16; Model 2: β = 0.0021; P = 0.35 inline image
Butte et al. (24)BL: 1030 (FU: 879) 4- to 19-year-old Hispanic children (partly siblings) from families with ≥1 overweight child from Houston, USA1 yearTotal accelerometer counts d−1ΔBWGeneralized estimating equations (Model 1: baseline age, age2, Tanner stage, sex; Model 2: Model 1 plus baseline BMI)BL PA was not associated with ΔBW. Model 1: β = −0.001, P = 0.72; Model 2: β = 0.001, P = 0.53 inline image
DeLany et al. (25)BL: 131 (FU: 114) 9- to 11-year-old healthy African–American and Caucasian, lean and obese children from Baton Rouge, USA2 yearsTotal EEΔ%BFLinear regression (no confounders used in the principal results extracted)BL total EE was inversely associated with Δ%BF. β = −0.73, R2 = 0.039, P < 0.04 inline image
Figueroa-Colon et al. (26)BL: 47 (FU1: 39; FU2: 36) 5- to 9-year-old healthy, normal weight Caucasian, African–American or Asian–American girls from Birmingham, USA1.6 years, 2.7 yearsPAEEΔBFLinear regression (BL FFM and BF)BL PAEE was a predictor of Δ%BF at FU1 but not at FU2. Model FU1: −17.35 + 0.004 girls' BL age + 0.29 fathers' %BF −0.01 PAEE inline imageinline image
Janz et al.§(27)BL: 470 (FU: 378) 4- to 6-year-old predominantly Caucasian (95%) children with from Iowa, USA3 yearsMean of BL and FU accelerometer countsLower vs. higher quartile of FU %BF (vice versa)Wilcoxon rank-sum test (no adjustment)Compared with children in the higher/lower three quartiles, those in the lowest/highest quartile of FU %BF had greater/lower overall PA (P < 0.05/P < 0.005) inline image
Johnson et al. (28)BL and FU: 115, 4- to 11-year-old African–American and Caucasian healthy children from Birmingham, USAAnnual FU for 3–5 yearsPAEETotal BF LTM−1Linear regression (ethnicity, BL BF, LTM, Tanner stage and age. Sex according to the abstract)BL PAEE was not a predictor of FU total BF LTM−1. β = 0, P = 0.74 inline image
Moore et al.§(29)BL: 106 (FU: 94) 3- to 5-year-old Caucasian healthy children, whose parents are third or fourth generation of the Framingham Heart Study, USAAnnual FU for 8 yearsAverage accelerometer counts over 8 yearsSSFMultivariable modelling (Sex, FU age and BL BMI)Children in the highest activity tertile had lower average SSF at age 11 years than did those at the lowest tertiles (trend: P = 0.045) inline image
Salbe et al. (30)BL and FU: 138, 5-year-old healthy Pima Indian children from Arizona, USA5 yearsPAEE%BFLinear regression (BL %BF and sex)BL PAEE was a predictor for FUΔ%BF (positive direction; P of the whole model = 0.003) inline image
Stevens et al. (31)BL: 1576 (FU: 984) 12-year-old girls of Caucasian, African–American, Hispanic or Asian and other ethnicity from different parts of the USA2 yearsBL MVPA by accelerometerOR for having (i) >32% FU BF or (ii) increase in %BF by ≥3.5% pointsMixed model logistic regression (no adjustment)(i) OR = 1.001 (95% CI 0.985, 1.017) (ii) OR = 1.009 (95% CI 0.997, 1.020) inline image
Treuth et al. (32)BL: 101 (FU: 88) 8- to 9-year-old healthy lean African–American and Caucasian girls in Tanner stage 1 living in Houston, USA1 year; 2 yearsPAEEΔ%BFRepeated measures anova (PAEE; group [according to parental obesity]; ethnicity; interaction of group and BL time; and BL Tanner stage, %BF or BF and time)BL PAEE was not associated with Δ%BF; β = 0.002, P = 0.14 inline image

Total daily EE was assessed by either doubly labelled water (23,25,28,30,32) or by 24 h whole room indirect calorimetry (26). Resting metabolic rate was assessed by either indirect calorimetry (23,26,30) or estimated from published equations (25,28,32). Five studies reported data on PAEE (23,26,28,30,32), whereas one study only reported data on total EE and therefore this variable was used (25). Four additional studies used accelerometers to assess PA and presented categorical data of either moderate-to-vigorous PA at baseline (31), total counts per day at baseline (24) or mean counts per minute or hour of several measurements including the baseline and follow-up measurements (27,29). The latter analyses are not truly prospective as prospective analyses would report on the association between baseline PA (as the exposure) and change in adiposity from baseline to follow-up (as the outcome). A cross-sectional element is introduced particularly in one of these studies (27) as it uses the average of two measurements (baseline and follow-up) as the exposure.

The outcome measure varied amongst studies: four of the nine studies reported change in per cent body fat as the outcome (25,26,30,32), one study presented change in body weight (24) and one study used the average of annually repeated measures of per cent body fat (23). Johnson et al. (28) used the follow-up ratio of total body fat and lean total mass, whereas Moore et al. (29) presented the sum of skin-folds at follow-up, both analyses adjusting for a baseline body composition measure. Two studies used categorical outcome measures: Stephens and co-workers reported the odds ratio for having more than 32% body fat at follow-up (31) and Janz et al. compared quartiles of follow-up per cent body fat (the lowest vs. the three highest quartiles, vice versa) (27). With the exception of three studies (25,27,31), the estimated association between PA and change in adiposity was adjusted for confounding factors; these included a baseline measure of body composition in all studies except for (23) and sex, age or ethnicity depending on the study population.

Six of the 10 studies did not find a prospective association between PA and adiposity (23,24,26,28,31,32). In the study by Figueroa-Colon et al., PAEE predicted per cent body fat at the first follow-up (β = −0.01) but not at the second follow-up (26). One study found a negative association between total EE and adiposity, whereas prospective data on the association between PAEE and adiposity were not presented (25). A negative relationship between PA and adiposity was observed in two studies (27,29), although the limitations in the analyses of these two studies need to be considered (see above). In contrast to what would be expected, Salbe et al. (30) reported higher levels of PAEE to be associated with greater subsequent gain in per cent body fat. In general, the magnitude of effect was low. For instance, beta coefficients were close to zero (23,24,26,28,32) and adding PAEE to a model that included baseline per cent body fat and sex, increased the explained variance of the whole model by only 3% (30). Similarly, in one study PAEE was the last and therefore least significant variable entering the stepwise multiple regression model and the magnitude of the association depended on the covariates included in the model (26).

Children and adolescents – intervention studies

Table 2 provides an overview of the four studies that investigated the effect of a PA intervention on surrogates for adiposity. Two studies evaluated the effectiveness of a community-based intervention (34,35), one study was a nursery-based (33) and one was a school-based (36) exercise intervention. These programmes included a combination of additional exercise classes and sport opportunities along with health and exercise education programmes aimed at either the children (36) or the parents and the family (33–35). The mode of delivery was always with personal contact, partly in combination with an impersonalized component such as leaflets. Only one study reported that the intervention was based on a theoretical framework for behaviour change, the social cognitive theory (34). The interventions were carried out in New Zealand (35), Europe (33,36) and the USA (34). They involved children aged between 4 and 12 years at baseline, who were of mainly Caucasian ethnicity, except for one study that focused on African–American children (34). Two studies presented data from the pilot phase of the intervention (34,35). The pilot intervention by Robinson et al. (34) lasted for 12 weeks and involved 61 girls at baseline. All other interventions had a duration of 6 months to 2 years; two studies included approximately 500 children at baseline (33,35) and one study included 810 children, although data on objectively measured PA were restricted to 123 children (36). Dropout was less than 12%, except for one study that had lost 25% of the participants at the end of the intervention (35).

Table 2.  Intervention studies in children and adolescents
StudyPopulationFollow-upInterventionExposure*OutcomeModel (confounding)Size of effect
  • *

    PA was assessed by accelerometers in all studies (Computer Science Application accelerometer (33,34), Mini Mitter Unidirectional Actical (35) and Actigraph (36)).

  • GD is the difference of the mean follow-up values of the treatment group and the control group, adjusted for the baseline value of the dependent variable.

  • FU PA monitoring occurred after completing the intervention.

  • Δ change; BL, baseline; BMI, body mass index; CI, confidence interval; FU, follow-up; GD, adjusted group difference; PA, physical activity; SD, standard deviation; SSF, sum of skin-folds.

Reilly et al. (33)BL: 36 nurseries, 545 children aged 4 years from Glasgow, UK; FU1: 36 nurseries, 481 children (250 controls). Randomization6 monthsNursery-based: three 30-min sessions of PA per week; families received PA guidance and leaflets promoting less TV timeFU log counts min−1FU BMI SD-scoreMulti-level modelling; levels: child, nursery; fixed effects: BL age, sex, group (BL PA)No intervention effect on PA and BMI
Robinson et al.(34)BL: 61 (33 controls) African–American healthy girls aged 8–10 years from low-income neighbourhoods of Oakland and East Palo Alto, USA. FU: 59 PA data4 monthsCommunity-based: five dance classes per week (2.5 h); five home-based lessons aiming to reduce TV time; health education by monthly lectures and newsletters to the familyFU average counts min−1FU BMIancova (BL PA, mean BMI; The interaction between treatment assignment and the centred BL BMI-value was also included into the model)No intervention effect on PA and BMI. PA: GD = 7.3 (95% CI −25.8, 40.4), P = 0.67; BMI: GD = −0.32 (95% CI −0.77, 0.12), P = 0.16
Taylor et al. (35)BL: seven primary schools; 513 (235 controls) 5- to 12-year-old mainly Caucasian children from Otago, New Zealand. FU: 384 children (177 controls); 303 children (141 controls) had valid accelerometer data. Control and intervention communities were not randomly selected1 yearCommunity-based, involving schools but focusing on non-curricular activities. Activity coordinators encouraged children to participate in lifestyle-based PA and non-traditional sports and aimed to increase PA levels in the communityFU average counts min−1FU BMI z-scoreGeneralized estimating equations (PA: age, sex, BL PA group-differences; BMI: BL age and BMI z-score, sex, school)Intervention effect on PA (P < 0.05) but not BMI z-score. GD = −0.10 (95% CI −0.19, 0.00)
Verstraete et al.(36)BL: 16 schools, 810 children aged 9.7 years (SD 0.7) from East Flanders, Belgium. FU: 16 schools, 764 children; PA data: BL 123, FU 111 children. Random allocation of schools to the groups2 yearsSchool-based: extra-curricular PA, classroom-based PA lessons; classroom-based health-related PA education programmeFU min d−1FU SSFLinear mixed models: factors condition and school, sex (BL SSF, PA values)No intervention effect on PA (positive; P = 0.06) but on SSF (negative; P < 0.05)

PA was assessed by accelerometers in all studies; three studies reported results for average counts per minute (33–35) and one used PA intensity categories and expressed accelerometer data in minutes per day (36). For this review results of the low-to-vigorous activity intensity category were extracted, as they reflect the total PA engagement. Three studies used body mass index (BMI; z-score, standard deviation score) as the main outcome, whereas Verstraete et al. (36) used the sum of five skin-folds. All studies reported on intervention vs. control group differences of follow-up PA and adiposity, adjusted for the respective baseline measures.

One study found a significant difference in change of PA between the intervention and the control group (35) and a positive trend in favour of the intervention group was reported in two studies (34,36). Only one of these interventions found a significant effect on adiposity, i.e. the intervention group had a smaller increase in skin-folds compared with the controls (36).

Adults – observational studies

Six studies examining the prospective association between PA and a measure of adiposity in adults were identified (37–41,43) (Table 3). One study involved middle-aged Europeans (38) and three studies were carried out in the USA, of which two studies included healthy pre-menopausal women aged 20–46 years (37,41) and one additional study was restricted to male and female Pima Indians aged between 19 and 70 years (40). Luke and co-workers examined the associations between PAEE and body fat in 18- to 55-year-old male and female Nigerians in their primary publication (39). In a follow-up study, the authors examined the influence of culture on the association between PAEE and body fat in 18- to 59-year-old Nigerian and African–American women (43). The studies included in this review involved between 58 and 1120 participants at baseline, with only three studies reporting data of >100 individuals (37,38,43). Five studies had dropout rates of approximately 20–30% (37–40,43) whereas one study did not report any losses to follow-up (41). The length of follow-up varied between 1.5 and 5.6 years.

Table 3.  Prospective studies in adults
StudyPopulationFollow-upInclusionExposure*OutcomeModel (confounding)Size of effect
  • *

    Total EE was assessed by doubly labelled water in all studies but (38), where individually calibrated heart rate monitoring was used and (37) where Caltrac accelerometers were used.

  • PAEE was calculated as 0.9 total EE – resting metabolic rate in all studies but (41), where it was calculated as 0.9 total EE – sleeping EE.

  • BF was either assessed by BodPod (37) or by bioimpedance (38); BW was assessed by scale in all studies but (40), where under water weighing was used.

  • §

    The analysis is not truly longitudinal.

  • It is not stated in this and reference papers, which locality participants are from.

  • inline image significant association; inline image no significant association; Δ change; BF, body fat; BL, baseline; BMI, body mass index; BW, body weight; CI, confidence interval; EE, energy expenditure; FFM, fat free mass; FU, follow-up; PA, physical activity; PAEE, physical activity energy expenditure; SD, standard deviation.

Bailey et al.§¶(37)BL: 275, 35- to 45-year-old mainly Caucasian women from the USA. FU: 228 women20 monthsHealthy, BMI ≤ 30 kg m−2, non-smoking, pre-menopausalAccelerometer categories: low, moderate or high intensity. Analysis: change in category (BL to FU): decreased, maintained or increased PA intensity%BF categorized: increase from BL to FU (by >1%pt); maintenance; decrease (by >1%pt)General Linear Model (age, total PA)Compared with other PA groups, higher percentage of ‘PA decrease group’ participants were in the ‘%BF increase group’; lower percentage were in the ‘%BF decrease group’ (both P < 0.05) inline image
Ekelund et al. (38)BL: 1120, 54-year-old (median) Caucasian adults from the UK. FU: 739 adults (311 men)5.6 years (median)Healthy and free of known diabetesPAEEFU BFGeneral Linear Model (Sex, age, duration FU, BL BF and FFM)BL PAEE was inversely associated with ΔBF. β = −0.0024 (95% CI −0.0048, −0.0001); P < 0.05. A significant PAEE by age interaction suggested that the effect of PAEE on ΔBF was restricted to younger individuals inline image
Luke et al. (39)BL: 58 (30 men) 18- to 55-year-old adults from Nigeria. FU: 47 adults (24 men)1.5 years (mean)Healthy, not engaged in weight-loss practicesPAEERate of ΔBW per yearRegression (BL BW)BL PAEE was not correlated with ΔBW. Women: R = −0.37 (95% CI −0.68, 0.05); Men: R = −0.05 (95% CI −0.44, 0.36). When PAEE was normalized by BW (kJ kg−1 d−1) results were also non-significant inline image
Tataranni et al. (40)BL: 64 male and 26 female 19- to 70-year-old (mean: 37 years and 32 years) Pima Indians from the USA. FU: 74 adults4 years (mean)Healthy, non-diabetic, no medication affecting energy metabolismPAEEΔBWRegression (Sex, BL age, BW and/ or body composition and duration of FU)BL PAEE was not associated with ΔBW. R = −0.03; P = 0.772 inline image
Weinsier et al.§(41)BL: 61, 20- to 46-year-old premenopausal, Caucasian or African–American women recruited from a previous weight loss study, USA. Comparison between two extreme groups (weight maintainers vs. gainers)1 yearHealthy, normal weight sedentary non-smokers, no medications affecting EE or thyroid statusPAEEBWProcMixed (Age, race, BL BF and lean body mass)Maintainers lost a mean (±SD) of 0.5 ± 2.2 kg year−1 and gainers gained 9.5 ± 2.1 kg year−1. Significantly higher mean PAEE in maintainers compared with gainers; P < 0.02 inline image
Luke et al. (43)BL: 172 African–American women and 149 Nigerian women, 18- to 59-year-old. Three FU: 127 African–American and 107 Nigerian women3 yearsHealthy, neither pregnant nor lactating, not engaged in weight-loss practices and planning to stay in the areaPAEERate of ΔBW per yearRegression (BL FFM, FM)BL PAEE was not correlated with ΔBW in either group inline image

Five studies used PAEE as the exposure measure (38–41,43). Total EE was assessed either by doubly labelled water (39–41,43), or by individually calibrated minute-by-minute heart rate monitoring (38). Resting or sleeping EE was assessed by indirect calorimetry in all of these studies (38–41,43). One study assessed PA by accelerometers and categorized their participants into low-, moderate- and high-intensity groups based on activity counts of the seven highest 10-min epochs recorded in 1 week (37). In this study, the change in PA group categorization from baseline to follow-up was used as the exposure measure and the change in per cent body fat categories was used as the outcome measure (decreased, maintained or increased PA intensity or per cent body fat), thus the statistical analysis is not truly prospective. This study is furthermore limited by the categorical comparison of groups rather than individuals. Weinsier et al. (41) compared the mean of baseline and follow-up PAEE in two groups of body-weight maintainers and gainers. This analysis also has a cross-sectional element, which prohibits interpreting whether PA predicts gain in body weight. The outcome in the remaining studies was either change in body weight from baseline to follow-up (39,40,43) or follow-up total body fat (38). All studies except for one (37) adjusted for baseline body composition. Additional confounding variables included in these studies were age, sex and duration of follow-up.

Three of the six studies reported no association between baseline PAEE and subsequent change in body weight (39,40,43), whereas three studies observed associations in the expected direction (37,38,41). Bailey et al. (37) found that the ‘increased per cent body fat group’ consisted of a significantly higher percentage of participants from the ‘PA decrease group’ compared with the other PA groups (P < 0.05). Similarly, Weinsier et al. (41) observed that body-weight maintainers had a higher mean PAEE across time than gainers (P < 0.02). However, the results from both these studies should be interpreted bearing in mind that they do not properly examine the prospective associations between exposure and outcome measures (see above). Ekelund et al. (38) observed that baseline PAEE was inversely correlated with the subsequent change in total body fat (β = −0.0024; P < 0.05). Subanalyses suggested that in adults younger than the mean of 54 years, who on average gained weight during the follow-up period, baseline PAEE was weakly inversely related with change in body fat (β = −0.001; P < 0.05). In the generally weight stable older adults, baseline PAEE was positively associated with body fat (β = 0.0005; P < 0.001).

Adults – intervention study

One intervention study on adults was retrieved (Table 4): Cooper et al. (42) investigated the effect of a home-based unsupervised 30-min moderate-intensity exercise programme carried out for at least 5 days per week for 6 weeks in 18- to 64-year-old adults. The primary outcome of this study was change in blood pressure and change in adiposity was reported as a secondary outcome. Participants wore a Caltrac accelerometer on each exercising day and recorded EE variables and the type and duration of exercise sessions in diaries. The mean PAEE increased by 34% on exercise days (162 kcal; P < 0.05). No change in body weight was observed in either the intervention or the control group between baseline and follow-up.

Table 4.  Intervention study in adults
StudyPopulationFollow-upInterventionInclusion criteriaExposureOutcomeModel (confounding)Size of effect
  1. PA was assessed by a Caltrac accelerometer.

  2. Δ change; BL, baseline; BMI, body mass index; BW, body weight; EE, energy expenditure; FU, follow-up; GP, general practice; PA, physical activity; PAEE, physical activity energy expenditure.

Cooper et al.(42)BL: 90, 18- to 64 year-old adults (72 men) recruited from GPs, workplaces and companies; controls: 33 men, 9 women. FU: 86 adults (39 controls)6 weeks30-min additional PA for ≥5 d week−1 for 6 weeksHealthy blood pressure and no blood pressure or lipid lowering treatment, sedentaryPAEE on the exercise day, assessed by accelerometersΔBW from BL to FUancova (BL BMI, EE)Increase in PAEE by 34% in the intervention group from BL to FU. No intervention effect on BW in either group.

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conflict of Interest Statement
  8. Acknowledgement
  9. References

The five intervention studies and the 16 observational studies included in this review suggest that baseline PA is not an important predictor of change in adiposity in children, adolescents and adults. This observation confirms previous reviews (9,10) and strengthens the current evidence base, as this report only includes prospective observational and intervention studies that measured PA objectively.

Energy balance is achieved when energy intake equals EE. However, so far the relative contribution of either component to energy imbalance and subsequent unhealthy gain in body weight is unclear. The observed weak or non-existing association between baseline PA and subsequent change in adiposity suggests that PA may not be the main determinant of energy imbalance. It was recently proposed that increased energy intake is the major driver of the increasing prevalence of overweight and obesity observed in the USA, whereas PA is one of the moderators that influence the steepness of the trajectory of weight gain (44). This is in line with findings by Taylor and co-workers (21,35), who carried out a 1-year exercise intervention followed by a 1-year combined exercise and nutritional guidance intervention in children. After the first year of the study, a trend for an intervention effect on PA was observed and adjusted accelerometer counts were 28% higher in the intervention compared with the control children. During that time BMI z-score decreased but not significantly in the intervention relative to the control children (35). After 2 years, adding a nutritional component to the intervention, the effect on BMI z-score was significant, and independent of PA (21). In agreement with the observations by Taylor et al. (35), all except one of the intervention studies included in this review found no effect of PA on adiposity. The exception was the school-based intervention by Verstraete et al. (36), in which the relative increase in fat mass was slightly smaller in intervention children compared with the control children. The authors also observed a trend for a positive intervention effect on PA, although PA was only assessed in a subgroup of children, which may not be entirely representative for the whole study population (36). Hence, it is not clear whether the smaller relative body-weight change in the intervention compared with the control group can be explained by differing PA levels. It is worth noting that regardless of the comprehensiveness of the study design and the applied intervention strategy, most studies experienced difficulties in increasing the PA level of their participants. Thus, PA may still influence body-weight gain in the expected direction, although the current evidence does not support this conjecture. The heterogeneity in the designs of the interventions, the statistical analyses and the theoretical underpinning makes it difficult to synthesize the results. Moreover, only a small number of trials that aimed at primary prevention of weight gain have measured PA. Consequently, the evidence to base conclusions on is insufficient. These issues have also been emphasized in previous reviews (9,45,46).

This review also includes a large number of prospective observational studies, of which more than half report no prospective association between baseline PA and subsequent change in adiposity (23,24,26,28,31,32,39,40,43). Eight studies did suggest that PA had a protective effect on subsequent adiposity; however, associations were weak and the data presented in five of these studies were not ideal to answer our research question: four of the analyses were not truly prospective as they included a cross-sectional element (see results) (27,29,37,41); and their results are in agreement with findings from cross-sectional studies (7,8). One additional study analysed PAEE, but presented the prospective association between total EE and adiposity only (25). Two studies that did find a prospective inverse relationship between PA and adiposity indicate that the results may be affected by factors such as the follow-up time and characteristics of the study population. For instance, Figueroa-Colon et al. found a prospective inverse association between baseline PAEE and body fat after 1.6-year but not after 2.7-year follow-up in children (26). This indicates a possible time effect such that there may be a stronger link at shorter compared with longer follow-up times. Ekelund et al. (38) found a prospective inverse association between PAEE and subsequent change in total body fat in adults; interestingly, this association was modified by age. In younger adults, who on average gained weight during the follow-up period, baseline PAEE was weakly inversely related with change in body fat. In contrast, in the generally weight stable older adults, baseline PAEE was weakly positively associated with body fat, which was explained by an increase in both fat mass and fat free mass. The authors suggested that PA may be particularly important for older people to counteract sarcopenia and weight-loss related ill-health. There is some evidence that the association between PA and adiposity may also be modified by sex in both adults (47,48) and children (49) although not necessarily by culture (43): Luke et al. (43) aimed to determine whether PAEE differentially predicted weight gain in two cohorts of women of the same ethnicity but with substantial differences in socioeconomic and environmental living conditions. No significant association was found between baseline PAEE and 3-year weight change, adjusted for fat free mass and fat mass, in either Nigerian or African–American women. In contrast to these studies, Salbe et al. (30) reported that PAEE predicted greater adiposity in Pima Indian children, suggesting that a higher absolute PAEE was associated with greater subsequent gain in per cent body fat. As PAEE is dependent on body size and for any given level of PA the expended energy is higher in overweight compared with normal weight children, this association is probably explained by the lack of adjustment for body size. Several studies on children who likely progressed through puberty during the observational period did not adjust for a measure of sexual maturity at baseline, which is likely an important confounder for the association between PA and body fat. Furthermore, although a smaller positive change in per cent body fat is likely to decrease the risk of excess adiposity, change in per cent body fat is not a direct measure of excess adiposity. Comparable with the intervention studies, the retrieved observational studies were heterogeneous in their study designs, population characteristics and the statistical analyses. Moreover, the sample sizes were relatively small, with only five studies including more than 200 participants at baseline. In spite of these limitations, and bearing in mind that the direction of associations between PA and PAEE with gain in body weight or body fat may not necessarily be straightforward (50), the results indicate that PA is not strongly prospectively associated with change in adiposity. The results therefore suggest that PA-related EE may not be the main determinant of a positive energy balance in both children and adults.

Most of the observational studies included in the present review used PAEE as the primary exposure variable. It is, however, possible that the extent to which PA contributes to energy balance is determined by subcomponents of PA, such as the intensity of PA. For example, Stephens et al. (31) observed that the association with body fat was stronger for high-intensity PA than for total PA. However, the magnitude of association appears small as indicated by the results from the Avon Longitudinal Study of Parents and Children study including 4150 children (published after this review was finalized). An increase of 15 min (75%) of vigorous PA per day at age 12 was associated with approximately 1 kg (11%) lower fat mass at age 14 (51). Likewise, the authors reported a significant yet weak association between the total PA and adiposity: a 20% higher total PA level at 12 years was associated with approximately 5% lower fat mass at age 14, corresponding to approximately 500 g. It should be noted that the applied statistical model did not seem to control for a measure of baseline body composition, which may increase the likelihood of observing a negative association. This is because children with more body fat at baseline are likely to be less physically active at that time. These children are expected to have larger body fat at follow-up, not only because of their lower PA level, but also because their greater body fat at baseline. To obtain more information on whether different subcomponents of PA predict subsequent change in body fat and how much PA is needed to prevent gain in body weight and fat in people consuming a usual diet, more observational studies employing the most accurate assessment methods available, starting from early age and with repeated measures of exposures and outcomes and controlling for important confounders such as dietary intake are warranted.

Conclusion

In conclusion, the results from this narrative review suggest that objectively measured total PA or PAEE may not be a key determinant of body fat gain in children, adolescents and adults. However, as higher levels of PA are consistently and strongly associated with a range of health outcomes (52,53), public health efforts should still continue to promote PA across the life span, whereas the direct impact of PA on weight control should not be overstated. To further elucidate the complex relationships between PA, diet and adiposity, more high-quality intervention and observational studies with precisely measured subcomponents of PA, dietary intake and body composition are needed.

Conflict of Interest Statement

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conflict of Interest Statement
  8. Acknowledgement
  9. References

Désirée C Wilks, Hervé Besson, Anna-Karin Lindroos and Ulf Ekelund are funded by the Medical Research Council. The funding bodies had no role in the decision to publish this article.

Acknowledgement

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conflict of Interest Statement
  8. Acknowledgement
  9. References

We wish to thank Stephen Sharp for his statistical advice.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conflict of Interest Statement
  8. Acknowledgement
  9. References
  • 1
    World Health Organisation. Global strategy on diet, physical activity and health. 2010. [WWW document]. URL http://www.who.int/dietphysicalactivity/en/ (accessed 20 January 2010).
  • 2
    Lobstein T, Baur L, Uauy R. Obesity in children and young people: a crisis in public health. Obes Rev 2004; 5(Suppl. 1): 4104.
  • 3
    Deckelbaum RJ, Williams CL. Childhood obesity: the health issue. Obes Res Suppl 2001; 9/4: 239243.
  • 4
    Guo SS, Roche AF, Chumlea WC, Gardner JD, Siervogel RM. The predictive value of childhood body mass index values for overweight at age 35 y. Am J Clin Nutr 1994; 59: 810819.
  • 5
    AICR. Determinants of weight gain, overweight, and obesity. In: AICR (ed.). Food, Nutrition, Physical Activity, and the Prevention of Cancer: A Global Perspective. World Cancer Research Fund/American Institute for Cancer Research: Washington, DC, 2007, pp. 322341.
  • 6
    Dietz WH. Health consequences of obesity in youth: childhood predictors of adult disease. Pediatrics 1998; 101: 518525.
  • 7
    Jimenez-Pavon D, Kelly J, Reilly JJ. Associations between objectively measured habitual physical activity and adiposity in children and adolescents: systematic review. Int J Pediatr Obes 2009; 26: 116.
  • 8
    Besson H, Ekelund U, Luan J, May AM, Sharp S, Travier N, Agudo A, Slimani N, Rinaldi S, Jenab M, Norat T, Mouw T, Rohrmann S, Kaaks R, Bergmann M, Boeing H, Clavel-Chapelon F, Boutron-Ruault MC, Overvad K, Andreasen EL, Johnsen NF, Halkjaer J, Gonzalez C, Rodriguez L, Sanchez MJ, Arriola L, Barricarte A, Navarro C, Key TJ, Spencer EA, Orfanos P, Naska A, Trichopoulou A, Manjer J, Wirfalt E, Lund E, Palli D, Agnoli C, Vineis P, Panico S, Tumino R, Bueno-de-Mesquita HB, van den Berg SW, Odysseos AD, Riboli E, Wareham NJ, Peeters PH. A cross-sectional analysis of physical activity and obesity indicators in European participants of the EPIC-PANACEA study. Int J Obes (2005) 2009; 33: 497506.
  • 9
    Wareham NJ, van Sluijs EM, Ekelund U. Physical activity and obesity prevention: a review of the current evidence. Proc Nutr Soc 2005; 64: 229247.
  • 10
    Fogelholm M, Kukkonen-Harjula K. Does physical activity prevent weight gain-a systematic review. Obes Rev 2000; 1: 95111.
  • 11
    Shephard RJ. Limits to the measurement of habitual physical activity by questionnaires. Br J Sports Med 2003; 37: 197206.
  • 12
    Melanson EL, Freedson PS, Hendelman D, Debold E. Reliability and validity of a portable metabolic measurement system. Can J Appl Physiol 1996; 21: 109119.
  • 13
    Westerterp KR. Assessment of physical activity level in relation to obesity: current evidence and research issues. Med Sci Sports Exerc 1999; 31(Suppl.): 522525.
  • 14
    Murgatroyd PR, Shetty PS, Prentice AM. Techniques for the measurement of human energy expenditure: a practical guide. Int J Obes Relat Metab Disord 1993; 17: 549568.
  • 15
    Schoeller DA. Measurement of energy expenditure in free-living humans by using doubly labeled water. J Nutr 1988; 118: 12781289.
  • 16
    Ekelund U, Sjostrom M, Yngve A, Poortvliet E, Nilsson A, Froberg K, Wedderkopp N, Westerterp K. Physical activity assessed by activity monitor and doubly labeled water in children. Med Sci Sports Exerc 2001; 33: 275281.
  • 17
    Molnar D, Livingstone B. Physical activity in relation to overweight and obesity in children and adolescents. Eur J Pediatr 2000; 159(Suppl.): 4555.
  • 18
    Schoeller DA, van Santen E, Peterson DW, Dietz W, Jaspan J, Klein PD. Total body water measurement in humans with 18O and 2H labeled water. Am J Clin Nutr 1980; 33: 26862693.
  • 19
    Trost SG, McIver KL, Pate RR. Conducting accelerometer-based activity assessments in field-based research. Med Sci Sports Exerc 2005; 37(Suppl.): 531543.
  • 20
    Warren JM, Ekelund U, Besson H, Mezzani A, Geladas N, Vanhees L, for the Experts Panel. Assessment of physical activity – a review of methodologies with reference to epidemiological research: a report of the exercise physiology section of the European Association of Cardiovascular Prevention and Rehabilitation. Eur J Cardiovasc Prev Rehabil 2010; 17: 127139.
  • 21
    Taylor RW, McAuley KA, Barbezat W, Strong A, Williams SM, Mann JI. APPLE Project: 2-y findings of a community-based obesity prevention program in primary school age children. Am J Clin Nutr 2007; 86: 735742.
  • 22
    Proctor MH, Moore LL, Gao D, Cupples LA, Bradlee ML, Hood MY, Ellison RC. Television viewing and change in body fat from preschool to early adolescence: the Framingham Children's Study. Int J Obes Relat Metab Disord 2003; 27: 827833.
  • 23
    Bandini LG, Must A, Phillips SM, Naumova EN, Dietz WH. Relation of body mass index and body fatness to energy expenditure: longitudinal changes from preadolescence through adolescence. Am J Clin Nutr 2004; 80: 12621269.
  • 24
    Butte NF, Cai G, Cole SA, Wilson TA, Fisher JO, Zakeri IF, Ellis KJ, Comuzzie AG. Metabolic and behavioral predictors of weight gain in Hispanic children: the Viva la Familia Study. Am J Clin Nutr 2007; 85: 14781485.
  • 25
    DeLany JP, Bray GA, Harsha DW, Volaufova J. Energy expenditure and substrate oxidation predict changes in body fat in children. Am J Clin Nutr 2006; 84: 862870.
  • 26
    Figueroa-Colon R, Arani RB, Goran MI, Weinsier RL. Paternal body fat is a longitudinal predictor of changes in body fat in premenarcheal girls. Am J Clin Nutr 2000; 71: 829834.
  • 27
    Janz KF, Burns TL, Levy SM. Tracking of activity and sedentary behaviors in childhood: the Iowa Bone Development Study. Am J Prev Med 2005; 29: 171178.
  • 28
    Johnson MS, Figueroa-Colon R, Herd SL, Fields DA, Sun M, Hunter GR, Goran MI. Aerobic fitness, not energy expenditure, influences subsequent increase in adiposity in black and white children. Pediatrics 2000; 106: e50.
  • 29
    Moore LL, Gao D, Bradlee ML, Cupples LA, Sundarajan-Ramamurti A, Proctor MH, Hood MY, Singer MR, Ellison RC. Does early physical activity predict body fat change throughout childhood? Prev Med 2003; 37: 1017.
  • 30
    Salbe AD, Weyer C, Harper I, Lindsay RS, Ravussin E, Tataranni PA. Assessing risk factors for obesity between childhood and adolescence: II. Energy metabolism and physical activity. Pediatrics 2002; 110: 307314.
  • 31
    Stevens J, Murray DM, Baggett CD, Elder JP, Lohman TG, Lytle LA, Pate RR, Pratt CA, Treuth MS, Webber LS, Young DR. Objectively assessed associations between physical activity and body composition in middle-school girls: the Trial of Activity for Adolescent Girls. Am J Epid 2007; 166: 12981305.
  • 32
    Treuth MS, Butte NF, Sorkin JD. Predictors of body fat gain in nonobese girls with a familial predisposition to obesity. Am J Clin Nutr 2003; 78: 12121218.
  • 33
    Reilly JJ, Kelly L, Montgomery C, Williamson A, Fisher A, McColl JH, Lo Conte R, Paton JY, Grant S. Physical activity to prevent obesity in young children: cluster randomised controlled trial. BMJ 2006; 333: 10411045.
  • 34
    Robinson TN, Killen JD, Kraemer HC, Wilson DM, Matheson DM, Haskell WL, Pruitt LA, Powell TM, Owens AS, Thompson NS, Flint-Moore NM, Davis GJ, Emig KA, Brown RT, Rochon J, Green S, Varady A. Dance and reducing television viewing to prevent weight gain in African–American girls: the Stanford GEMS pilot study. Ethn Dis 2003; 13(Suppl.): 6577.
  • 35
    Taylor RW, McAuley KA, Williams SM, Barbezat W, Nielsen G, Mann JI. Reducing weight gain in children through enhancing physical activity and nutrition: the APPLE project. Int J Pediatr Obes 2006; 1: 146152.
  • 36
    Verstraete SJ, Cardon GM, De Clercq DL, De Bourdeaudhuij IM. A comprehensive physical activity promotion programme at elementary school: the effects on physical activity, physical fitness and psychosocial correlates of physical activity. Public Health Nutr 2007; 10: 477484.
  • 37
    Bailey BW, Tucker LA, Peterson TR, LeCheminant JD. A prospective study of physical activity intensity and change in adiposity in middle-aged women. Am J Health Promot 2007; 21: 492497.
  • 38
    Ekelund U, Brage S, Franks PW, Hennings S, Emms S, Wong MY, Wareham NJ. Physical activity energy expenditure predicts changes in body composition in middle-aged healthy whites: effect modification by age. Am J Clin Nutr 2005; 8: 964969.
  • 39
    Luke A, Durazo-Arvizu RA, Rotimi CN, Iams H, Schoeller DA, Adeyemo AA, Forrester TE, Wilks R, Cooper RS. Activity energy expenditure and adiposity among black adults in Nigeria and the United States. Am J Clin Nutr 2002; 75: 10451050.
  • 40
    Tataranni PA, Harper IT, Snitker S, Del Parigi A, Vozarova B, Bunt J, Bogardus C, Ravussin E. Body weight gain in free-living Pima Indians: effect of energy intake vs expenditure. Int J Obes Relat Metab Disord 2003; 27: 15781583.
  • 41
    Weinsier RL, Hunter GR, Desmond RA, Byrne NM, Zuckerman PA, Darnell BE. Free-living activity energy expenditure in women successful and unsuccessful at maintaining a normal body weight. Am J Clin Nutr 2002; 75: 499504.
  • 42
    Cooper AR, Moore LA, McKenna J, Riddoch CJ. What is the magnitude of blood pressure response to a programme of moderate intensity exercise? Randomised controlled trial among sedentary adults with unmedicated hypertension. Br J Gen Pract 2000; 50: 958962.
  • 43
    Luke A, Dugas LR, Ebersole K, Durazo-Arvizu RA, Cao G, Schoeller DA, Adeyemo A, Brieger WR, Cooper RS. Energy expenditure does not predict weight change in either Nigerian or African American women. Am J Clin Nutr 2009; 89: 169176.
  • 44
    Swinburn B, Sacks G, Ravussin E. Increased food energy supply is more than sufficient to explain the US epidemic of obesity. Am J Clin Nutr 2009; 90: 14531456.
  • 45
    Lemmens VE, Oenema A, Klepp KI, Henriksen HB, Brug J. A systematic review of the evidence regarding efficacy of obesity prevention interventions among adults. Obes Rev 2008; 9: 446455.
  • 46
    Hardeman W, Griffin S, Johnston M, Kinmonth AL, Wareham NJ. Interventions to prevent weight gain: a systematic review of psychological models and behaviour change methods. Int J Obes Relat Metab Disord 2000; 24: 131143.
  • 47
    Schulz LO, Schoeller DA. A compilation of total daily energy expenditures and body weights in healthy adults. Am J Clin Nutr 1994; 60: 676681.
  • 48
    Westerterp KR, Goran MI. Relationship between physical activity related energy expenditure and body composition: a gender difference. Int J Obes Relat Metab Disord 1997; 21: 184188.
  • 49
    Ness AR, Leary SD, Mattocks C, Blair SN, Reilly JJ, Wells J, Ingle S, Tilling K, Smith GD, Riddoch C. Objectively measured physical activity and fat mass in a large cohort of children. PLoS Med 2007; 4: 04760484.
  • 50
    Ekelund U, Brage S, Besson H, Sharp S, Wareham NJ. Time spent being sedentary and weight gain in healthy adults: reverse or bidirectional causality? AJCN 2008; 88: 612617.
  • 51
    Riddoch CJ, Leary SD, Ness AR, Blair SN, Deere K, Mattocks C, Griffiths A, Davey Smith G, Tilling K. Prospective associations between objective measures of physical activity and fat mass in 12–14 year old children: the Avon Longitudinal Study of Parents and Children (ALSPAC). BMJ 2009; 339: b4544.
  • 52
    Khan K, McKay HA, Haapasalo H, Bennell KL, Forwood MR, Kannus P, Wark JD. Does childhood and adolescence provide a unique opportunity for exercise to strengthen the skeleton? J Sci Med Sport/Sports Med Aust 2000; 3: 150164.
  • 53
    Steele RM, Brage S, Corder K, Wareham NJ, Ekelund U. Physical activity, cardiorespiratory fitness, and the metabolic syndrome in youth. J Appl Physiol 2008; 105: 342351.