Effect of interventions on the body mass index of school-age students

OBJECTIVE: To evaluate the effect of intervention programs using nutritional education, physical activity or both on the reduction of body mass index in school-age students. METHODS: The systematic review with meta-analysis included randomized controlled studies available from the following electronic databases for the years 1998 to 2010: PubMed, Lilacs, Embase, Scopus, Web of Science and Cochrane Library. The descriptors were: randomized controlled trial, overweight, obesity, body mass index, child, adolescent, physical activity, nutrition education and Schools. A weighted average was based on the standardized means difference and used a 95% confi dence interval. The inconsistency test was utilized to evaluate the heterogeneity of studies. RESULTS: Initially, 995 studies were identifi ed, of which 23 were included, and 3 meta-analyses were performed. Isolated physical activity interventions did not present a signifi cant reduction in body mass index, with a standardized mean difference of -0.02 (95%CI: -0.08; 0.04). A similar result (n= 3,524) was observed in the isolated interventions of nutritional education, with a standardized mean difference of -0.03 (95%CI: -0.10; 0.04).When the interventions with physical activity and nutritional education were combined, the result of the meta-analysis (n= 9,997) presented a statistically signifi cant effect in the reduction of body mass index in school-age students, with a standardized mean difference: -0.37 (95%CI: -0.63; -0.12). CONCLUSIONS: The interventions that combined physical activity and nutritional education had more positive effects in the reduction of body mass index among school-age students than when they were applied individually. DESCRIPTORS: Child. Adolescent. Obesity, prevention & control. Body Mass Index. Food and Nutrition Education. Physical Education and Training. Motor Activity. Weight Reduction Programs. Meta-Analysis.


INTRODUCTION
Obesity is a condition where the accumulation of body fat increases and results in harm to health.a Despite being an ideal, this defi nition is diffi cult to apply due to the limited availability of easy to administer methods to evaluate body composition. 8Body mass index (BMI) is used to defi ne obesity.The World Health Organization published the new child growth standards in 2006 and made a World Health Organization.Obesity: Preventing and Managing the Global Epidemic.Report of a WHO Consultation on Obesity.Geneva; 2000.
information about BMI available in graphic and tables with percentile values and Z scores, simplifying their use at a population level. 47The increasing prevalence of obesity threatens the health of an increasing portion of the population.It is a challenge to health services and necessitates practical methods of diagnosis and monitoring that are low cost.
The most current national level data on the nutritional status of Brazilian children age 5 years and less is the Pesquisa de Orçamentos Familiares (POF -Household Budget Survey) from 2008-2009 and the Pesquisa Nacional sobre Demografi a e Saúde (PNDS -National Demographic and Health Survey) of 2006.The results of the POF show that the prevalence of overweight varied from 32% to 40% in children age fi ve to nine years in the Southeast, South and Central-West regions and from 25% to 30% in the North and Northeast regions, which was the age group for which the increased prevalence of obesity was most dramatic.b There was an increased prevalence of overweight in the population age 10 to 19 years, from 3.7% to 21.7% among boys and from 7.5% to 19.4% among girls between the periods of 1974-1975 and 2009-2009. b The PNDS recorded a national prevalence of overweight of 6.6%, with the highest proportion in the South (8.8%) and the lowest in the North (5.2%).[43][44] Although there is no consensus on which interventions are more adequate to decrease obesity, the approaches tend to center on changes in lifestyle, nutritional education and promotion of physical activity. 10Schools are a strategic place to promote through education the formation of daily habits in physical activity and proper nutrition.
The objective of this study was to evaluate the effects of interventions using physical activity, nutritional education or both on the reduction of BMI in schoolage children.August of 2010, in the following electronic databases: Lilacs, PubMed, Web Of Science, Scopus, EMBASE and Cochrane Library.The following key words were used: randomized controlled trial, overweight, obesity, BMI, child, adolescent, physical activity, nutrition education and schools.Campbell et al 5 performed a systematic review in 2001 with studies published until 1998, which explains our decision to include studies beginning in that year.A search was performed from the bibliographic references of the relevant studies and of the systematic reviews concerning the topic of interest.The inclusion criteria were: randomized controlled studies with students age four to 19 years and with baseline and endline measurement of BMI, in addition to interventions using nutritional education, physical activity or both for a minimal duration of three months.The internal quality of the study was evaluated by criteria of blinded randomization proposed by Cochrane 19 and complemented by the Jadad Scale. 22he studies were classifi ed into four categories by the criteria of blinded randomization: Category A or Adequate (adequate randomization process); Category B or Undetermined (undescribed randomization process but the text mentions the study is randomized); Category C or Inadequate (randomization process inadequately described); Category D or Not Utilized (non-random study).Studies classifi ed as A and B, by the analysis of blinded randomization, were included.Studies classifi ed as C and D were excluded from review because they were considered inadequate experiments.

METHODS
The criteria described by Jadad et al 22 for evaluating internal quality were randomization, double-blinding and loss to follow-up and exclusions.The results were presented by scoring (maximum of fi ve points).A study is considered of poor quality if it receives a score ≤ three.
The information was independently extracted by two reviewers.The results were matched to verify concordance and the discordant results were resolved by consensus.The evaluation by the reviewers was not blinded from the authors and the study results.
A summary measure based on the standardized mean difference (SMD) was used.The calculation of this summary measure and its 95% confi dence interval was according to the fi xed effects model or the random effects model, depending on the heterogeneity between the studies.The test of inconsistency (I 2 ) was used to evaluate heterogeneity between the studies and random effects model was used for I 2 > 50%. 20,21The effect size estimated from the scale of magnitudes for effect statistics was evaluated by the SMD. 7he statistical analyses was performed using the software program Review Manager version 5.1, produced by Cochrane Collaboration, and results were presented through use of Forest Plot graphics.

RESULTS
There were 995 studies identified, including 231 duplicates; 642 were excluded after analysis of titles and abstracts because they did not meet the inclusion criteria.The full text was analyzed for 122 articles of which 37 were excluded because they did not meet the inclusion criteria.Eighty-fi ve studies were analyzed and classifi ed by the criteria of blinded randomization; 19 and 40 studies were classifi ed as A or B and selected.Of these, 17 were excluded for not presenting suffi cient data, for a fi nal total of 23 included studies (Figure 1; Table ).
Of the 23 studies, 16 evaluated the effect of physical activity and nutritional education as an intervention to reduce BMI, fi ve evaluated physical activity and two evaluated nutritional education.The majority of prevention programs promoted physical acti vity 4,9,[11][12]14,16,24,[26][27][28][29]38,40,46,50,51,53 and two programs recommended physical activity. 15,55 All the studies included pograms that encouraged healthy dietary habits through presentations and didactic materials.26][27][28][52][53] Identifi cation of studies Seven studies were considered adequate in their randomization procedure, and it was not described in 16 studies, although the text mentioned that the study was randomized.Twenty-two studies were considered poor quality and one as good quality (Table ).
Three meta-analyses were performed to evaluate the effect of interventions that isolated or combined physical activity and nutritional education about BMI among students: Five studies were grouped to evaluate the effect of the interventions with physical activity among students.The results with 4,172 participants did not show a statistically signifi cant effect of the intervenitons with physical activity in the reduction of BMI, with SMD (fi xed effects): -0.02 (95%CI -0.08;0.04),p = 0.46, between the intervention group (IG) compared to the control group (CG), with a trivial effect magnitude.There was no heterogeneity between studies (I 2 = 0%) (Figure 2).Two studies involving interventions with nutritional education were included and the results grouped for a total 3,524 participants, and the analysis demonstrated no statistically significant effect in the reduction of BMI, with a SMD (fi xed effects): -0.03 (95%CI -0.10;004), p = 0.39 between the IG compared to the CG, with an effect magnitude considered trivial.There was heterogeneity between stuides, with low variability (I 2 = 36%) (Figure 3).
To evaluate the effect of interventions with physical activity and nutritional education among students, 16 studies were grouped.The results with 9,997 participants presented a statistically signifi cant effect of the interventions with physical activity and nutritional education together upon the BMI, with SMD (random effects): -0.37 (95%CI -0.63;-0.12),p < 0.01, between the IG compared to the CG with a small effect magnitude.There was heterogeneity between the studies, with high variability (I 2 = 97%) (Figure 4).

DISCUSSION
The results of the meta-analysis for the isolated interventions with physical activity or nutritional education did not show a statistically signifi cant effect on the reduction of BMI among students.A similar result was presented in a meta-analysis by Harris et al 18  Although it does not measure physical composition, the widespread use of BMI in epidemiologic studies as an indicator for nutritional status is justifi ed by its ease of measurement, large availability of data on body mass, in addition to its relationship with morbidity and mortality. 2 These fi ndings reinforce the importance of developing actions and programs to change lifestyles among this age group, since their lifestyles are in formation.The literature documents the benefits of physical activity and of healthy nutrition upon skeletal health (mineral consent and bone density) [34][35][36]55 increased fl exibility and aerobic capacity 6,30,55 and in the inverse relationship with cardiovascular risk factors. The most challenging aspect for health promotion strategies is adherence outside of schools, since health is negatively impacted by the food industry through advertisements and commercials for calorie-dense foods.Technological advances such as videogames and computers attract children to a lifestyle of little physical activity and increased calorie consumption.
Researchers foresee growth in high-speed internet and videos on screens (cable TV, VCR, DVD, videogames, computer games), according to a conference by the National Institutes of Health in the United States, and these changes will continue to incentivize lifestyle changes among children and adolescents. 25Strategies to prevent obesity should focus on factors that prevent its development.
Interventions with families, especially parental involvement in the promotion of healthy habits, should be contemplated and estimulated by interventional programs.Children are infl uenced by their parents' habits; therefore, recommendations introduced in schools should be followed at home through the positive example of parents for their children, through healthy nutrition and the regular practice of physical exercise.
Interventions have better results when the strategy includes a family component. 13,31e 23 studies included in the meta-analysis have limitations, since the majority were performed with a small sample and were considered of low quality by the Jadad scale because they did not describe the allocation randomization procedure, blinding, losses and exclusions.This implies a need for more well-designed randomized controlled clinical studies.No Brazilian study was included in this review since they did not meet the inclusion criteria.
This systematic review is subject to publication bias, since it is easier to publish studies that report benefi cial effects for a given intervention, in detriment to studies that do not describe benefi cial effects.
This study suggests a need for randomized controlled studies with well-designed methodologic criteria in order to evaluate the effect of interventions, especially

Figure
Figure 4. Forest Plot for studies comparying IG and CG for interventions with physical activity and nutritional education on BMI among students.

Table .
Characteristics of studies included in the systematic review.
Legend: Year -Publication year; Location -Location of intervention; n -Sample size; Blinding rating -per Cochrane; Jadad -Jadad scale for study quality; PA -Physical activity; NE -nutritional education; PA + NE: Physical activity and nutritional education.
Forest Plot for the studies comparying the IG with the CG for interventions with physical activity and BMI among students.: Standard deviation; n: Sample size; SMD [PA]: standardized mean difference for fi xed effects: CI: Confi dence interval; I²: Heterogeneity: Chi 2 = 1.57, df = 1 (P = 0.21); I 2 = 36% Test for effect: Z = 0.86 (P = 0.39) Figure 3. Forest Plot for the studies comparying the IG with CG for interventions with nutritional education and BMI among students.SD: Standard deviation; n: Sample size; SMD [PA]: standardized mean difference for fi xed effects: CI: Confi dence SD 4. Forest Plot for studies comparying IG and CG for interventions with physical activity and nutritional education on BMI among students.Brazilian population.The results can help in calculating sample sizes during the planning of future studies to evaluate the effect of structured interventions among an intervention and control group.Interventions that combine physical activity and nutritional education present better effects on the reduction of BMI among students, as a strategy to prevent and control obesity, than if applied in an isolated manner.