Medicine & Science in Sports & Exercise:
APPLIED SCIENCES: Physical Fitness and Performance
Factors that alter body fat, body mass, and fat-free mass in pediatric obesity
LeMURA, LINDA M.; MAZIEKAS, MICHAEL T.
Exercise Physiology Laboratory, Graduate Program in Exercise Science, Bloomsburg University of Pennsylvania, Bloomsburg, PA
Submitted for publication March 2001.
Accepted for publication June 2001.
LeMURA, L. M., and M. T. MAZIEKAS. Factors that alter body fat, body mass, and fat-free mass in pediatric obesity. Med. Sci. Sports Exerc., Vol. 34, No. 3, pp. 487–496, 2002.
Purpose: The purpose of this study was to quantify the effects of exercise treatment programs on changes in body mass, fat-free mass, and body fat in obese children and adolescents.
Methods: By using the meta-analytic approach, studies that met the following criteria were included in our analyses: 1) at least six subjects per group; 2) subject groups consisting of children in the 5- to 17-yr age range; 3) pretest and posttest values for either body mass, percent body fat, or fat-free mass (FFM); 4) used exercise as a mode of treatment (e.g., walking, jogging, cycle ergometry, high-repetition resistance exercise, and combinations); 6) training programs ≥ 3 wk; 7) full-length publications (not conference proceedings); 8) apparently “healthy” children (i.e., free from endocrine diseases and disorders); and 9) published studies in English language journals only.
Results: A total of 120 investigations were located that addressed the issue of exercise as a method of treatment in pediatric obesity. Of those, 30 met our criteria for inclusion. Across all designs and categories, fixed-effects modeling yielded significant decreases in the following dependent variables: 1) percent body fat (mean = 0.70 ± 0.35; 95% CI = 0.21 to 1.1); 2) FFM (mean = 0.50 ± 0.38; 95% CI = 0.03 to 0.57); 3) body mass (mean = 0.34 ± 0.18; 95% CI = 0.01 to 0.46); 4) BMI (mean = 0.76 ± 0.55; 95% CI = 4.24 to 1.7), and 5) V̇O2max (mean = 0.52 ± 0.16; 95% CI = 0.18 to 0.89), respectively. Significant differences were found as a function of the type intervention groups (exercise vs exercise + behavioral modification;P < 0.04); body composition assessment methods (skinfold vs hydrostatic weighing, DEXA, and total body water;P < 0.006); exercise intensity (60–65%, vs ≥ 71% V̇O2max;P < 0.01); duration (≤ 30 min vs > 30 min;P < 0.03); and mode (aerobic vs aerobic + resistance training;P < 0.02). Stepwise linear regression suggested that initial body fat levels (or body mass), type of treatment intervention, exercise intensity, and exercise mode accounted for most of the variance associated with changes in body composition after training.
Conclusions: Exercise is efficacious for reducing selected body composition variables in children and adolescents. The most favorable alterations in body composition occurred with 1) low-intensity, long-duration exercise; 2) aerobic exercise combined with high-repetition resistance training; and 3) exercise programs combined with a behavioral-modification component.
Obesity has emerged as the most common pediatric chronic illness in Western countries. The prevalence of childhood obesity in North America ranges from 10 to 25% and has skyrocketed during the last two decades (16). Prospective research has shown that obesity can cause complications in many organ systems. For example, obese adults have an added risk of morbidity and mortality from coronary heart disease, lipid abnormalities, hypertension, diabetes mellitus, sleep apnea, infertility, gall bladder disease, and some cancers (28). In children, obesity substantially increases the risk of orthopedic, respiratory, and psychosocial disorders (46). Because studies have demonstrated that obesity in childhood tracks into adulthood, public health officials commonly refer to pediatric obesity as an “alarming trend” and as an “epidemic” (16).
The serious consequences of pediatric obesity have propelled health care professionals to study the impact of various treatment programs. Numerous studies have been published on the role of physical activity in obesity, but until recently, the majority of research has focused on treating adults (16,25,45). Although there is agreement within the scientific community that exercise is an empirically validated method of treating pediatric obesity, numerous questions remain regarding the type of exercise program that is most effective to favorably alter body composition (2,12). In light of the new direction in obesity research that involves molecular and behavioral genetics (i.e., the next generation of treatments), it seems prudent to synthesize what we now know about exercise as a palliative treatment in pediatric obesity.
Meta-analysis is a quantitative approach in which studies are converted to individual data points and then subjected to parametric statistical analyses (17). This allows one to make comparisons on a common research problem that would be difficult or impractical to perform using conventional research paradigms. Thus, the purpose of this study was to use the meta-analytic approach to examine the effects of exercise programs on changes in body mass, fat-free mass, and body fat in obese children. A second purpose was to identify areas of potential strengths and weaknesses in the literature and to provide directions for future research on this topic.
The search for literature was limited to human exercise-training studies published between 1960 and 2001 in which body mass, fat-free mass (FFM), and percent body fat were measured in children aged 5–17. Studies were located via computer generated citation searches of the following databases: Current Contents, MEDLINE, Dissertation Abstracts International, Psychological Abstracts, and Sport Discus. Extensive hand searching and cross-referencing was done from the bibliographies of previously retrieved studies and from review articles. In addition, an expert in exercise and pediatric obesity (Dr. Melinda S. Sothern) reviewed our reference list for thoroughness and completeness. A sample of the descriptive terms that were used to locate relevant studies in English research journals included: children, exercise, obesity, physical activity, and weight loss.
Studies were included if they met the following criteria: 1) at least six subjects per group; 2) subject groups consisting of children in the 5–17 yr age range; 3) pretest and posttest body mass and body mass index (BMI; kg·m−2) (21), percent body fat, and FFM values; 4) used exercise as a mode of training (e.g., walking, jogging, cycle ergometry, high-repetition resistance exercise, and combinations); 6) training programs ≥ 3 wk; 7) full-length publications (not conference proceedings); 8) apparently “healthy” children (i.e., free from endocrine diseases and disorders); and 9) published studies in English language journals, only.
The studies in this review were coded according to the following characteristics: 1) exercise program characteristics (length, frequency, intensity, duration, mode, and compliance defined as the number of exercise sessions attended); 2) study characteristics (author(s), year, definition used to establish obesity, study design, comparison groups [e.g., exercise plus diet versus exercise plus behavior intervention] and number of subjects); 3) subject characteristics (age, weight, sex, health status, race, and fitness status [e.g., V̇O2max]); 4) body composition measurement characteristics (e.g., DEXA versus BIA versus skinfold); and 5) primary outcomes (changes in body mass, percent body fat, and FFM). To avoid interobserver bias, the authors of this study independently extracted all data. The authors then met and reviewed each item for accuracy and consistency. Disagreements were resolved by consensus. Several studies met our inclusion criteria but were missing data vital to our analyses (i.e., means and standard deviations); therefore, we contacted the authors and requested raw data to calculate the effect sizes (ES) for the primary and secondary outcomes.
The primary outcome in this study was ES changes in body mass, percent body fat, and FFM. Subgroup analysis was performed to examine between group differences. The subgroups were analyzed according to: 1) intensity (the percent of V̇O2max and HRmax at which exercise training was prescribed); 2) exercise duration (the amount of time spent during each exercise session); 3) length of training (the weeks or months the training protocols lasted); 4) frequency (the number of exercise sessions per week); 5) mode (the type of activity used to train the subjects); 6) the method of assessing body composition; 7) subject characteristics such as sex and V̇O2max); 8) research design (randomized control trials, controlled trials, or no control group); and 9) the type of intervention program (exercise versus diet, or exercise versus behavior modification). Analyses of the dependent variables partitioned by subgroup were performed when the data for each coded characteristic were available.
The meta-analytic approach was popularized by Glass (17) as a mode of quantitatively integrating findings from various studies. Each study serves as a unit of analysis, and the findings between studies are compared through the calculation of a common metric, the ES. The ES is defined as the difference between the means of the experimental group (ME) and the means of the control group (MC), divided by the control group standard deviation (SC). The ES formula is applied as follows:MATH
Effect size computation and analysis.
The research findings from each study were transformed into the common ES metric that was submitted for further statistical analysis. The purpose of the ES analysis was to allow us to draw conclusions regarding the efficacy of exercise on obese children in the existing literature. In the present study, the primary dependent variables of interest used for analysis were body mass, FFM, percent body fat, and V̇O2max. For those studies whose research design included a control group (i.e., nonexercise, dietary, or behavioral intervention), the standard ES formula was applied. When a research design did not include a control group, the pre- to post-training changes in body mass, FFM, and percent body fat were used to determine the ESs for each study (i.e., the pretraining value minus the posttraining value). To calculate the ESs for V̇O2max, we reversed the pretraining and posttraining values to keep the algebraic sign of the ES positive (i.e., the posttraining value minus the pretraining value). Thus, positive ESs for any of the dependent variables indicated improvements after the training period. The differences between the pre- and post-training measures for the primary outcomes were then divided by a pooled variance (17). A pooled estimate of the variance provided a more precise estimate of the population variance. The pooled variance weighted for sample size was obtained by using the following formula:MATH for which Sp stands for the pooled standard deviation, S12 is the variance for the control group or group 1, S22 is the variance for the experimental group or group 2, n 1 is the number of subjects in group 1, and n 2 stands for the number of subjects in group 2.
The variance for each ES and a correction for small sample bias were calculated using the procedures developed by Efron and Tibshirani (7). The meta-analytic approach is based upon the principle of normally distributed data; however, because this is not always the case, we corrected for bias by bootstrap resampling to generate 95% confidence intervals (CI) around the mean ES. The estimate generated from this approach is based upon the sample itself, rather than from a theoretical distribution. This nonparametric method (which is not restricted by large sample assumptions) of estimating the reliability of the original sample estimate is calculated by randomly drawing from the available sample, with replacement (22). Each time an observation is selected for a new sample, each element of the original sample has an equal chance of being selected. If the CI included zero (0.00), we concluded that there was no effect of exercise on obesity in children. For studies that generated more than one ES because of the presence of more than one treatment group, the ESs were treated as independent data points but were also combined to determine the impact of ESs on overall results. To obtain a measure of the variability of the data, we examined the heterogeneity of the ESs by identifying the outliers beyond the 10th and 90th percentiles. The authors examined each study represented by an outlier further. If a methodological flaw or physiologic reason existed to explain the variability, the study was precluded from additional analysis. Differences of opinion were resolved through discussion and consensus. In addition, because there is a propensity for studies to be published that generate statistically significant results, we also addressed the issue of publication bias (i.e., the tendency for authors to submit and journal editors to publish studies that yield statistically significant results). This was a pertinent concern because the unpublished studies that were located were ultimately published; therefore, the studies included in this review were derived exclusively from research journals. We examined publication bias with the Kendal τ rank (3) correlation test (rt). This consisted of correlating observed outcomes, i.e., changes in the ESs of body mass, FFM, percent body fat, and V̇O2max with sample size.
We used a random-effects model to pool the ES data on changes in body composition if heterogeneity was present. ESs were then averaged across studies to determine treatment effects and were further stratified according to coded characteristics of interest. ANOVA-like procedures for meta-analysis were used to determine whether significant differences in body mass, FFM, and percent body fat existed for exercise prescription components. Linear regression was used to indicate the magnitude and direction of relations among variables, and stepwise regression was used to find which combination of factors best accounted for changes in body weight and composition after training. Finally, a subgroup analysis of the primary outcomes partitioned by research design (randomized controlled trials [RCTs], controlled trials [CTs], and studies with no control group [NC]) was performed to determine the impact of design on components of the exercise prescription. A Tukey post hoc test was used to determine which means were statistically different in the presence of significant F-ratios. An alpha level of P ≤ 0.05 was established a priori to establish the presence of significant differences. All ESs were calculated with Meta-Stat (version 1.3) (40). Mixed effects analysis, resampling, and randomization analyses were performed with Meta-Win (version 1.0) (33).
We coded for age, sex, ethnicity, and health status. Because the subjects’ ages were often reported a range of years, an analysis of older and younger children was not possible. Additionally, many investigators combined boys and girls in their studies; thus, we were precluded from conducting an analysis of gender. Most of the children in the studies were Caucasian. Although Hispanic and African-American children were also studied, there were too few children in these ethnic groups to perform a meaningful analysis. Finally, all of the subjects were identified as “healthy,” i.e., free from endocrine diseases or other significant comorbidities that might affect their ability to respond to the intervention programs.
A total of 120 investigations were located that addressed the issue of exercise as a method of treatment in pediatric obesity. From this group of studies, 30 met our criteria for inclusion. The 30 studies generated 92 ESs based upon a total of 945 subjects ranging from 5 to 17 yr. Reasons for the rejection of studies in this analysis included a lack of or an inability to obtain pre- to post-data for body composition, the use of the same subject-pool in more than one study, cross-sectional rather than longitudinal designs, and acute response study designs. The time to code each study ranged from 1.0 to 3.5 h (mean = 1.8 ± 0.45 h). An examination of the outliers beyond the 10th and 90th percentiles revealed two outliers. These outliers remained in the analysis because neither physiological nor methodological reasons were found to exclude them. Additionally, no publication bias was found for ES changes in body mass (P = 0.71), FFM (P = 0.24), percent body fat (P = 0.09), or V̇O2max (P = 0.89), respectively. Several studies generated more than one ES as a result of subject assignment into exercise-only groups, diet-only groups, behavioral-modification-only groups, any combination of these groups, or along gender lines. The average length of training ranged from 3 to 30 wk (mean = 12.75 ± 5.9), exercise duration per session from 20 to 60 min (mean = 38 ± 17.5), and frequency ranged 1–7 d (mean = 3.9 ± 1.5). The rate of compliance, defined as the percentage of exercise sessions completed, ranged from 90 to 100%. Tables 1 and 2 detail the coding characteristics of studies and their associated ESs.
The methods used to assess body composition included skinfold (SF) calipers (N = 11; 1 Lange and 10 Harpenden calipers), hydrostatic weighing (HSW;N = 3), dual-photon x-ray absorptiometry (DEXA;N = 3), and both total body water (TBW) and bioelectric impedance analysis (BIA;N = 1). There were not enough studies to conduct multiple comparisons for each assessment method; therefore, we compared SF measurement to all other methods (HSW, DEXA, and TBW) combined. Investigators that used SF calipers (mean ES = 1.0 ± 0.24; CI = 0.33 to 1.6) to assess body composition reported significantly greater reductions in percent body fat than investigators that used alternative methods of assessment (mean ES = 0.54 ± 0.25; CI = 0.19 to 0.92;P < 0.006).
The reported data permitted an analysis of changes in percent body fat and body mass partitioned by research design. Changes in percent body fat in the RCTs (mean = 0.68 ± 0.35; CI = 0.28 to 1.1) when compared with the NCs (mean = 0.70 ± 0.38; CI = 0.39 to 1.2) were not different. Similarly, changes in body mass as a function of research design did not result in a significant difference. The NCs and CTs resulted in ES means of 0.34 ± 0.14 (CI = −.04 to 0.56) and 0.38 ± 0.21 (CI = −.02 to 0.58), respectively.
We also analyzed the effects of intervention method or treatment groups on body mass. Specifically, we found that the exercise-only, exercise plus diet, and exercise plus behavior-modification groups yielded mean ESs of 0.36 ± 0.18, 0.33 ± 0.16, and 0.63 ± 0.20, respectively (P < 0.04). The post hoc analysis revealed that a significant difference was found between the diet and exercise and the exercise plus behavior modification groups (P < 0.05), only.
Across all designs, intervention strategies, and categories, the pre- to post-ESs after exercise training in children were: 1) percent body fat (mean = 70 ± 0.35; 95% CI = 0.21 to 1.1); 2) FFM (mean = 0.50 ± 0.38; 95% CI = 0.03 to 0.57); 3) body mass (mean = 0.34 ± 0.18; 95% CI =. 01 to 0.46); 4) BMI (mean = 0.76 ± 0.55; 95% CI = 0.24 to 1.7); and 5) V̇O2max (mean = 0.52 ± 0.16; 95% CI = 0.18 to 0.89), respectively. According to Cohen’s categories to classify ESs (< 0.41 = small; 0.41–0.70 = moderate; > 0.70 = large) (5), the degree to which exercise treatment programs in pediatric obesity produced favorable changes in body composition and V̇O2max ranged from small to large, depending upon the variable measured.
Changes in percent body fat were significantly different when the data were partitioned by the exercise program characteristics. These included exercise intensity, duration, mode, and program length. 1) Intensity: (60–65%, ≥ 66–70%, and ≥ 71% of V̇O2max). The mean ESs and associated confidence intervals for each intensity range were 1) 0.97 ± 0.33; 95% CI = 0.31 to 1.9, 2) 0.65 ± 0.19; 95% CI = 0.27 to 1.1), and 3) 0.29 ± 0.10; 95% CI = −0.01 to 0.35), respectively. The Tukey post hoc test revealed that the highest and lowest pairs of ES means (60–65 and ≥ 71% of V̇O2max) were significantly (P < 0.01) different from each other. A linear regression was used to determine the level of association between increases in V̇O2max and percent body fat. This analysis produced a significant r-value of 0.70 (P < 0.05); 2) Duration: (≤ 30 min > 30 min) This analysis yielded significantly different ESs of 0.57 ± 0.21; 95% CI = 0.14 to 0.71 and 0.95 ± 0.27; 95% CI = 0.29 to 1.7, respectively (P < 0.03); 3) Mode: (aerobic and aerobic plus resistance training). All of the studies appeared to be aerobic in nature, although nine studies combined aerobic with high-repetition resistance training. A comparison of studies using aerobic exercise (mean = 0.58 ± 0.31; 95% CI = 0.18 to 0.90) and aerobic plus resistance exercise (mean = 1.2 ± 0.35; 95% CI = 0.34 to 2.1) on percent body fat resulted in a significant difference (P < 0.02); and 4) Program length: (≤10 wk and >10 wk). An analysis of these data resulted in ESs and associated CIs of mean = 0.70 ± 0.40; 95% CI = 0.36 to 1.1 and mean = 0.76 ± 0.30; 95% CI = 0.38 to 1.4, respectively; however, the difference was not significant. An analysis of exercise frequency was not possible because of the large range of exercise sessions reported by the investigators (i.e., 1–7 d·wk−1). The ES values for the dependent variables of interest and for the coded characteristics of the studies are summarized in Table 3.
A stepwise linear regression was used to find which combination of factors best accounted for changes in percent body fat, body mass, and FFM after training. Independent variables were considered to be significant predictors if its P-value was < 0.05. Table 4 includes the results of this analysis for all of the children combined. The major predictors of change in body composition were preintervention percent body fat (body mass and body fat are highly correlated), exercise intensity (≤60–65% of V̇O2max), exercise mode (aerobic plus resistance exercise), and type of intervention (i.e., exercise plus behavior modification). These variables accounted for 30–59% of the variance associated with changes in percent body fat, body mass, and FFM.
Obesity is a disorder of energy metabolism (1). Small, chronic imbalances between energy intake and energy expenditure lead to the storage of excess energy in the form of triglycerides in adipose tissue (16,19). The meteoric rise in childhood obesity indicates that shifts in energy intake and expenditure have adversely affected energy metabolism. As a result, exercise has emerged as a cornerstone of pediatric obesity treatment, along with dietary and behavioral-intervention programs.
To have confidence in the results of a meta-analysis, two conditions must be satisfied. First, the data must not be biased by the absence of unpublished studies where changes do not achieve statistical significance. Second, the results of studies meeting the minimum criteria for analysis must be included. We addressed the former issue by searching vigorously for unpublished and published studies and the latter issue by retrieving and including all relevant studies.
We had difficulty locating research studies designed to identify the optimal exercise program characteristics to maximize weight loss in obese children. This underscores the necessity to design controlled, prospective studies to determine the impact of exercise prescription components on weight loss in obese children. The overall results of this research synthesis suggest that exercise has a modest to strong impact on selected body composition variables. Our analyses showed that after exercise training, percent body fat, body mass, and BMI decreased, whereas FFM increased in obese children and adolescents.
Most analyses were conducted with percent body fat as the dependent variable. When the data were partitioned by exercise intensity, duration, mode, and intervention program, we found that the lowest exercise intensity, the longest exercise duration, the combination of aerobic plus high-repetition (8–12 repetitions) resistance exercise, and an intervention of exercise plus behavior modification resulted in the greatest decreases in percent body fat in children. These data were supported further by the identification of predictors for changes in the dependent variables (i.e., exercise intensity, mode, and intervention program). It seems reasonable to speculate that the changes in body fat were the result of exercise-induced lipolysis and substrate utilization of fat through beta-oxidation. This is supported by the low-intensity exercise in concert with a relatively long-exercise duration that is known to accompany this energy system (25,45). It is also important to note that in studies that included changes in FFM (or when we could calculate FFM from the available data), it was found that FFM increased approximately 0.5 SD higher than untrained control subjects. We attribute this finding to the combination of aerobic-type exercise with high-repetition resistance exercise. It would appear that whereas both aerobic and combination training reduce percent body fat in children, high-repetition resistance exercise potentially serves to simultaneously increase FFM. This is an important finding because investigators have speculated that the use of resistance training could be a safe, effective means to promote weight loss and to increase FFM in obese children (38). From a practical standpoint, low-intensity exercise is more sustainable and therefore may facilitate a longer duration of exercise in obese children with limited aerobic and anaerobic capacities (2,10,34).
We coded for body composition assessment techniques to determine the magnitude of ESs across all reported methods. Because most investigators used SF assessments, we were limited to only one meaningful analysis, i.e., SF versus all of methods combined (DEXA, HSW, TBW, and BIA). When SF measurements were used, the reduction in percent body fat was nearly 0.5 SD over all other methods combined. This points to the need to incorporate a unified approach in the interpretation and use of existing techniques so that various equations and methods can be compared between laboratories and across studies. Another critical need is to reach consensus on reference data and standards to assess percent body fat and patterns of distribution throughout childhood (19,20).
Our analysis revealed a significant, positive relationship for exercise intensity and V̇O2max. Although correlation analyses do not establish cause and effect, perhaps a pertinent question that might emerge from this finding is: what is the nature of the relationship between aerobic fitness and percent body fat in obese children (24,27)? Alternatively, does an increased percent body fat preclude the attainment of aerobic fitness? In obese adults, it has been demonstrated recently that body fat does not influence maximal aerobic capacity, rather the major influence of body composition on V̇O2max is FFM (18). This also appears to be the case in pediatric obesity. In support of this finding, research indicates that when V̇O2max is expressed per kilogram of body weight or per kilogram of FFM, it is comparable in obese and normal-weight children (21,28,34). Furthermore, researchers have reported that FFM is either unaffected or increased along with weight loss after training obese children (44,50). The results of this study demonstrated that V̇O2max improved after training. It is reasonable to speculate that the obese children in our analysis increased V̇O2max because the effects of carrying excess fat were removed and FFM increased (19). It is also important to note that obese children have a significantly larger pulmonary ventilatory response to exercise before training (4,21,39). After training, the ventilatory response improves while body fat is reduced. These favorable physiologic alterations are reflected by increases in exercise performance and V̇O2max (50).
We also found that the inclusion of a behavioral-modification program in conjunction with exercise training significantly affected changes in body composition. The studies that used behavioral strategies included the family to support exercise (at home and in the clinical setting) (10,41,43,48), offered classes on variations of caloric and nutrient intake (2,9,11,15), and encouraged and reinforced participation in spontaneous over sedentary activities (i.e., walking vs television viewing). Research with adults has shown excellent results for home-based over on-site exercise programs (30). It would be interesting to study this exercise program issue with children and adolescents but the paucity of “at-home” designs eliminated this possibility. We were surprised that there were no ES differences when the data were partitioned by research design, as we expected the RCTs (N = 7) and CTs (N = 6) to generate significantly lower treatment effects when compared with the NCs (N = 17). Nevertheless, it seems prudent to approach these research questions with carefully selected controls for the purpose of generalizability. Other issues that emerged from this research synthesis with important clinical implications include the absence of studies stratified by ethnicity and the paucity of studies related to gender differences. There are ample data that indicate that children and adolescents from minority backgrounds tend to be less physically active (45). Prospective research and epidemiological study is needed to determine the factors that either support or discourage an active lifestyle in ethnic or minority groups. Furthermore, the study of gender differences remains ripe for investigation. Despite a limited, gender-focused database, we know that gender significantly influences fat deposition and removal. But in addition to biological differences, girls are socialized differently than boys with reference to participation in physical activity. Research has shown significant differences in physical activity and energy expenditure exist in children as young as age 10 (26). Additional study is needed to develop strategies that will encourage and reinforce activity in girls from late childhood through adolescence.
It is important to discuss the limitations of the present investigation in order to evaluate the results in their proper context. First, a potential limitation of these analyses relates to the definition of obesity provided in the studies (Table 5). Researchers used different measures of obesity, some definitions were unclear, and others referred to their subjects as obese but did not include any definitions. This may be partly explained by the widespread, current use of the BMI definition that was not available in previous years (16,41). The use of a standard measure of obesity would make comparisons of treatment effectiveness easier to accomplish and interpret. Second, many authors neglected to include pertinent subject characteristics, such as body weight, percent body fat, and BMI. Only five studies (14,21,28,39,40) included data on ethnicity and descriptions of matching procedures. Without these data, meta-analysts are unable to conduct analyses that might explain treatment effects and associated variation. Concerning this meta-analysis, we would be remiss if we did not ask investigators to include and editors to publish all relevant data, even in the absence of significant findings.
In summary, this research synthesis demonstrated the effectiveness of exercise as an adjunctive treatment for childhood and adolescent obesity. With regard to the development of an appropriate exercise prescription, the most favorable alterations in body composition were associated with low-intensity, long-duration exercise and aerobic exercise combined with high-repetition resistance training. Our analysis also showed the added benefit of exercise combined with a behavioral-modification component. In studies that reported V̇O2max values, the data revealed that obesity did not prohibit an improvement in aerobic fitness. Finally, the method of percent body fat measurement has a significant impact on the outcomes of assessments.
These data are reported in the wake of an exciting, new direction in pediatric obesity research, i.e., molecular and behavioral genetics. These new paradigms will most likely be used to focus on factors that influence craving and satiety, the role of neurotransmitters on mood and caloric regulation, the potential for muscle fiber type as an etiologic factor in obesity, and many other genetic influences on obesity. The research tools of organ-systems physiology and exercise biochemistry should be used in synergistic ways with the predictive tools of molecular biology to reduce an increasingly complex problem and serious public health concern.
The authors gratefully acknowledge Dr. Bernard Gutin for his contribution to this manuscript. We also thank the Office of Graduate Studies and the Department of Exercise Science of Bloomsburg University for funding this study.
Address for correspondence: Linda M. LeMura, Ph.D., FACSM, Professor and Graduate Program Director, Exercise Physiology Laboratory, Centennial Hall, Bloomsburg University of Pennsylvania 17815; E-mail: firstname.lastname@example.org.
1. Baldwin, K. M. Research in the exercise sciences: where do we go from here? J. Appl. Physiol. 88: 332–336, 2000.
2. Becque, D. M., V. L. Katch, A. P. Rocchini, C. R. Marks, and C. Moorehead. Coronary risk incidence obese adolescents: reduction by exercise plus diet intervention. Pediatrics 81: 605–612, 1988.
3. Begg, C. B. Publication bias. In: The Handbook of Research Synthesis, H. Cooper and L. V. Hedges (Eds.). New York: Russell Sage, 1994, pp. 399–400.
4. Blaak, E. E., K. R. Westerterp, O. Bar-Or, L. J. Wouters, and W. H. Saris. Total energy expenditure and spontaneous activity in relation to training in obese boys. Am. J. Clin. Nutr. 55: 777–82, 1992.
5. Cohen, J. Statistical Power Analysis for the Behavioral Sciences. Hillsdale, NJ: Erlbaum, 1988, pp. 22–45.
6. Ebbeling, C. B., and N. R. Rodriguez. Effects of exercise combined with diet therapy on protein utilization in obese children. Med. Sci. Sports Exerc. 31: 378–385, 1999.
7. Efron, B., and R. Tibshirani. An Introduction to the Bootsratp. London: Chapman and Hall, 1993, pp. 2–56.
8. Epstein, L. H., L. H. Kuller, R. R. Wing, A. Valoski, and J. McCurley. The effect of weight control on lipid changes in obese children. Am. J. Dis. Child. 143: 454–457, 1989.
9. Epstein, L. H., A. Valoski, R. R. Wing, et al. Perception of eating and exercise in children as a function of child and parent weight status. Appetite 12: 105–118, 1989.
10. Epstein, L. H., A. Valoski, R. Koeske, and R. R. Wing. Family-based behavioral weight control in obese young children. J. Am. Diet. Assoc. 86: 481–484, 1986.
11. Epstein, L. H., R. R. Wing, B. C. Penner, and M. J. Kress. Effect of diet and controlled exercise on weight loss in obese children. J. Pediatr. 107: 358–361, 1985.
12. Epstein, L. H., R. Koeske, J. Zidansek, and R. R. Wing. Effects of weight loss on fitness and obese children. Am. J. Dis. Child. 137: 654–657, 1983.
13. Escobar, O., H. Mizuma, M. S. Sothern, et al. Hepatic insulin clearance increases after weight loss in obese children and adolescents. Am. Med. Sci. 317: 282–286, 1999.
14. Figueroa-Colon, R., M. S. Mayo, R. A. Aldridge, T. Winder, and R. L. Weinsier. Body composition changes in Caucasian and African American children and adolescents with obesity using dual-energy x-ray absorptiometry measurements after a 10-week weight loss program. Obes. Res. 6: 326–331, 1998.
15. Figueroa-Colon, R., K. V. Almen, F. A. Franklin, C. Schuftan, and R. M. Suskind. Comparison of two hypocaloric diets in obese children. Am. J. Dis. Child. 147: 160–166, 1993.
16. Flegal, K. M. The obesity epidemic in children and adults: current evidence and research issues. Med. Sci. Sports Exerc. 31 (Suppl. 11): S509–S514, 1999.
17. Glass, G. V. Integrating findings: the meta-analysis of research. Rev. Res Educ. 5: 351–379.
18. Goran, M., D. A. Fields, G. R. Hunter, S. L. Herd, and R. L. Weinsier. Total body fat does not influence maximal aerobic capacity. Int. J. Obes. 24: 841–848, 2000.
19. Goran, M. I. Metabolic precursors and effects of obesity in children: a decade progress, 1990–1999. Am. J. Clin. Nutr. 73: 158–171, 2001.
20. Goran, M. I. Measurement issues related to studies of childhood obesity: assessment of body composition, body fat distribution, physical activity, and food intake. Pediatrics 101 (Suppl. 3): 505–518, 1998.
21. Gutin, B., N. Cucuzzo, S. Islam, C. Smith, R. Moffatt, and D. Pargman. Physical training improves body composition of black obese 7 to 11 year old girls. Obes. Res. 3: 305–312, 1995.
22. Hedges, L. V., and I. Olkin. Statistical Methods for Meta-Analysis.
San Diego, CA: Academic Press. 1985, pp. 3, 18.
23. Hills, A. P., and A. W. Parker. Obesity management via diet and exercise intervention. Child Care Health 14: 409–416, 1988.
24. Lemura, L. M., S. P. von Duvillard, S. L. Cohen, et al. Treadmill or cycle ergometry testing in 5–6 year old children: which mode is best? Eur. J. Appl. Physiol. 85: 472–478, 2001.
25. Lemura, L. M., S. P. von Duvillard, J. L. Andreacci, J. M. Klebez, S. A. Chelland, and J. Russo. Lipid and lipoprotein profiles, cardiovascular fitness, body composition, and diet during and after resistance, aerobic and combination training in young women. Eur. J. Appl. Physiol. 82: 451–458, 2000.
26. Lemura, L. M., J. L. Andreacci, R. Carlonas, J. M. Klebez, and S. Chelland. Evaluation of physical activity measured via accelerometry in rural fourth-grade children. Percept. Mot. Skills 90: 329–337, 2000.
27. Lemura, L. M., S. P. von Duvillard, and R. Carlonas. Can exercise training improve maximal aerobic power (VO2max) in children: a meta-analytic review. J. Exerc. Physiol. 2: 1–17, 1999.
28. Owens, S., B. Gutin, J. Allison, et al. Effect of physical training on total and visceral fat in obese children. Med. Sci. Sports Exerc. 31: 143–148, 1999.
29. Pena, M., J. Bacallao, L. Barta, M. Amador, and F. E. Johnston. Fiber and exercise in the treatment of obese adolescents. J. Adolesc. Health Care 10: 30–34, 1989.
30. Perri, M. G., A. D. Martin, E. A. Leermakers, S. F. Sears, and M. Notelovitz. Effects of group-versus home-based exercise in the treatment of obesity. J. Consult. Clin. Psychol. 65: 278–285, 1997.
31. Reybrouck, T., J. Vinckx, G. Berghe, and M. Vanderschueren-Lodeweyckx. Exercise therapy and hypocaloric diet in the treatment of obese children and adolescents. Acta Paediatr. Scand. 79: 84–89, 1990.
32. Rocchini, A. P., V. Katch, J. Anderson, et al. Blood pressure in obese adolescents: effect of weight loss. Pediatrics 82: 16–23, 1988.
33. Rosenberg, M. S., D. C. Adams, and J. Gurevitch. Meta-Win: Statistical Software for Meta-Analysis with Resampling Tests, version 1:0. Sunderland, MA: Sinauer, 1997, pp. 51–97.
34. Rowland, T. W., M. R. Varzeas, and C. A. Walsh. Aerobic responses to walking training in sedentary adolescents. J. Adolesc. Health 12: 30–34, 1991.
Rudner, L. Meta-stat: Software to Aid in the Meta-Analysis of Research Findings, version 1.3. Chevy Chase, MD: LMP, 1993, pp. 38–62.
36. Sasaki, J., M. Shindo, H. Tanaka, M. Ando, and K. Arakawa. A long-term aerobic exercise program decreases the obesity index, and increases the high density lipoprotein cholesterol concentration in obese children. Int. J. Obes. 11: 339–345, 1987.
37. Schwingshandl, J., and M. Borkenstein. Changes in lean body mass in obese children during a weight reduction program: effect on short term and long-term outcome. Int. J. Obes. 19: 752–755, 1995.
38. Sothern, M. S., J. M. Loftin, J. N. Udall, Jr., et al. Safety, feasibility, and efficacy of a resistance training program in preadolescent obese children. Am. J. Med. Sci. 319: 370–375, 2000.
39. Sothern, M. S., M. Loftin, U. Blecker, and J. N. Udall, Jr. Impact of significant weight loss on maximal oxygen uptake in obese children and adolescents. J. Investig. Med. 48: 411–416, 2000.
40. Sothern, M. S., J. N. Udall, Jr., R. M. Suskind, A. Vargas, and U. Blecker. Weight loss growth velocity in obese children after very low calorie diet, exercise and behavior modification. Acta Paediatr. 89: 1036–1043, 2000.
41. Sothern, M. S., B. Despinasse, R. Brown, R. M. Suskind, J. N. Udall, Jr., and U. Blecker. Lipid profiles of obese children and adolescents before and after significant weight loss: differences according to sex. South. Med. J. 93: 278–282, 2000.
42. Sothern, M. S., T. K. Almen, H. D. Schumacher, R. M. Suskind, and U. Blecker. A multidisciplinary approach to the treatment of childhood obesity. Del. Med. J. 71: 255–261, 1999.
43. Sothern, M. S., S. Hunter, R. M. Suskind, R. Brown, J. N. Udall, Jr., and U. Blecker. Motivating the obese child to move: the role of structured exercise in pediatric weight management. South. Med. J. 92: 577–584, 1999(b).
44. Sothern, M. S., M. Loftin, R. M. Suskind, J. N. Udall, Jr., and U. Blecker. Impact of significant weight loss on resting energy expenditure in obese youth. J. Investig. Med. 47: 222–226, 1999.
45. Szmedra, L., L. M. Lemura, and W. M. Shearn. Exercise tolerance, body composition and blood lipids in obese African-American women following short-term training. J. Sports Med. Phys. Fitness 38: 59–65, 1998.
46. Troiano, R. P., K. M. Flegal, R. J. Kuczmarski, S. M. Campbell, and C. L. Johnson. Overweight prevalence and trends for children and adolescents. Arch. Pediatr. Adolesc. Med. 149: 1085–1091, 1995.
47. Wabitsch, M., U. Braun, E. Heinze, et al. Body composition in 5–18-y-old obese children and adolescents before and after weight reduction as assessed by deuterium dilution and bioelectrical impedance analysis. Am. J. Clin. Nutr. 64: 1–6, 1996.
48. Wadden, W. A., A. J. Stunkard, L. Rich, C. J. Rubin, G. Sweidel, and S. Mckinney. Obesity in black adolescent girls: a controlled clinical trial of treatment by diet, behavioral modification, and parental support. Pediatrics 85: 345–352, 1990.
49. Williford, H. N., D. L. Blessing, W. J. Duey, et al. Exercise training in black adolescents: changes in blood lipids and V̇O2max
. Ethn. Dis. 6: 279–285, 1996.
50. Zanconato, S., E. Baraldi, P. Santuz, et al. Gas exchange during exercise in obese children. Eur. J. Pediatr. 148: 614–617, 1989.
ADOLESCENTS; BODY WEIGHT; CHILDREN; FAT MASS; FAT-FREE MASS; OBESITY; WEIGHT LOSS
©2002The American College of Sports Medicine
What does "Remember me" mean?
By checking this box, you'll stay logged in until you logout. You'll get easier access to your articles, collections,
media, and all your other content, even if you close your browser or shut down your
To protect your most sensitive data and activities (like changing your password),
we'll ask you to re-enter your password when you access these services.
What if I'm on a computer that I share with others?
If you're using a public computer or you share this computer with others, we recommend
that you uncheck the "Remember me" box.
Highlight selected keywords in the article text.
Data is temporarily unavailable. Please try again soon.