Randomized controlled trials on physical activity, diet, or both for public health through body weight management in 5- to 10-year-old pupils: a systematic review and meta-analysis

Abstract

Introduction

Since the 20th century, noncommunicable diseases (NCDs) such as cardiovascular diseases, cancers, and diabetes have been on the rise, accounting for approximately 90% of all deaths and disability in developed countries and becoming more common in youth [1–3]. Correspondingly, there exists a paradoxical contingency between the adversity of spiraling healthcare costs (totaling US$9.8 trillion) amidst poorer health stature [1–5]. The transtheoretical model of behavior change may explain the challenges of dependence in adulthood, including pre-contemplative facets, exhaustive planning, action, and relapse [6]. However, competent health behavior education during childhood may offer a loophole to surpass permanent reliance pathways [6, 7]. Hence, the integrity of the school setting to tapping the underlying directive of the UN Sustainable Development Goals (e.g. Quality Education and Good Health and Well-Being) given the aggregation of youth with various sociocultural ethnicities and financial backgrounds [8–10]. Elementary schools, in particular, offer a unique structural framework for delivering rigorous scientific proceedings due to the link between early personal behavior development and health outcomes [9, 10].

Children’s body weight (BW) and body mass index (BMI) have been routine health indicators reflecting dietary intake with the degree of physical activity (PA) principally through the energy balance mechanism [11, 12]. Previous research has circumstantiated the necessity of PA for health (Exercise is Medicine) [13], including NCD attenuation [14], treatment for musculoskeletal and cardiometabolic parameters [15], as well as the sharpening of mental and psychological acuity [16]. Nutrition, however, may be more important regarding the formation of atherosclerotic plaques, even with the confounding effect of moderate PA levels [17]. For addressing worldwide health problems, generalizing the depth of nutrition by analyzing common diet types (omnivorous, vegetarian, and vegan) may offer largescale differentiation [18, 19]. Several systematic reviews with meta-analyses have been performed to date to summarize the effects of PA interventions and/or dietary in primary schools regarding the outcome of BW and BMI in pupils [20–23]. No systematic review has, however, examined the sole impact of PA and/or dietary trials exclusive to the primary school setting. In line with the UNESCO cross-cutting competencies [24], the aim of the present study was to analyze the effect of elementary school health programs of PA and/or nutrition (including diet type) with at least 8 weeks’ duration on the BW and BMI of pupils concerning sustainable health. This investigation hypothesizes that PA intervention for elementary school pupils performed for at least 8 weeks during mandatory education time results in healthier body composition management.

Methods

Study design and selection criteria

Following the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines [25], a systematic review and meta-analysis was performed. A study protocol was published a priori [26] along with a trial registration on the International Prospective Register of Systematic Reviews (PROSPERO: CRD42023477701; https://www.crd.york.ac.uk/PROSPERO/). There was no funding source for this study.

Randomized controlled trials (RCTs) of any type (excluding quasi-randomized) that were published until March 2025 in either English or German were sought. The study population included pupils aged 5–10 years and enrolled at the primary school level or equivalent. It was required that the pupils had no major nutritional deficiency or unstable health condition. In addition, the exclusion criteria were used in refining the study sample: (i) lacking an outcome on BW or BMI or based on self-reports, (ii) only overweight and/or obese pupils were intervened, (iii) lack of a control group, (iv) if PA or diet was not the predominant focus in multi-component theoretical approaches on health, (v) interventions not entirely school-based or before/after regular school hours, (vi) shorter than 8 weeks of intervention duration, (vii) specialized branches of school mainly focused on physical exercise and/or sport, and (viii) school level not stated with pupils’ aged < 5 years or > 10 years.

The databases were originally searched in December 2021 by DT, including PubMed, EBSCOhost Education Source, and EMBASE, and the search was updated on the 28 of February 2025 excluding the use of non-randomized study design terminology (quasi randomized controlled trial, quasi randomized controlled trial, cohort, cohort study); the search matrices for each database were provided in the Supplementary Material S1 (Tables S1–S3; pp. 1–24). Reference lists of included articles as well as similar systematic reviews were scanned to aid in the search coverage. Likewise, to avoid duplication, PROSPERO was searched for similar reviews in progress or recently completed reviews.

Data extraction and quality assessment

D.T. and M.M. completed data extraction primarily with the standardized digital data extraction forms available in Covidence (https://www.covidence.org/). Results on BW and BMI outcomes were collected with Microsoft Excel (Version 2405). To avoid data duplication, multiple reports of the same study were linked. The following data from included studies were extracted: (i) study citation and contacts; (ii) eligibility for review or reason for exclusion; (iii) methods (overall design, duration of intervention, randomization sequence generation, allocation concealment, blinding, and other areas of possible bias); (iv) characteristics of participants (number enrolled, years of age, sex, setting/type of school, rural or urban location, country, any prespecified diagnostic criteria); (v) intervention and control groups (type of group, intervention specifics, regime: PA, dietary, or dual approach, details for reproduction: Overall duration, volume, physical intensity and type of training, diet type, and integrity); (vi) results (participant numbers for each intervention/control group and for each BW/BMI outcome, missing participants, overall outcome summary for each intervention group and with any subgroup analysis provided); and (vii) other (financial resources, study conclusions, special comments made by study authors, and citations of relevant articles). Risk of bias assessment was performed by D.T. and M.M. with the Cochrane Tool in Covidence and is available in Supplementary Material S2.

Statistical testing and tools

In order to complete a quantitative synthesis of the data and provide a precise measure of effect by meta-analysis, studies were required to administer data either within the article or by direct contact with the study authors on BMI (kg/m²: standard deviation [SD] or conversion with standard error [SE]; BMI z-score: SD or conversion with 95% confidence interval [95% CI]; or BMI mean change: difference estimates, including Cohen’s d and 95% CI) or BW (kg: SD or conversion with SE or 95% CI) in addition to the participants numbers for intervention and control groups at baseline and follow-up. Statistical analyses were conducted with the meta package available for R (version 4.3.2; 2023/10/31 ucrt) [27]. The effect size for the intervention (vs control) was calculated by quantitatively pooling baseline and follow-up data of BW (kg) or BMI (kg/m²) with a random-effects, mean difference (MD: 95% CI) meta-analysis [28]. Statistical heterogeneity was assessed with I². Funnel plots for each meta-analysis were provided in Supplementary Material S1 (pp. 45–46) to address possible publication bias. Data were combined in studies with multiple PA groups or control groups [29]. In order to assess immediate intervention effects, the follow-up measurement taken nearest to the termination of the intervention was used in studies with several follow-up points. In cases where SD, baseline or follow-up values, or change scores were not provided, calculations following the guidelines of the Cochrane Handbook were made using other available estimates (SE, 95% CI) [29].

Results

The study selection process is displayed in Fig. 1. A total of 9479 title/abstracts were identified from the comprehensive search of the databases, 2444 were removed by the systematic management software as duplicates, and 6477 were excluded by the reviewers. Of the 558 full-texts that were screened, 245 were screened in the present investigation at the primary school level. Additionally, 24 articles were identified from scanning reference lists; 39 studies (52 references provided in Supplementary Material S3) were finally included in the systematic review and 20 studies met the inclusion criteria for the meta-analysis (see Supplementary Material S4). Supplementary Material S1 (Table S4; pp. 25–44) displays the list of excluded articles that were screened by full-text (n = 217) provided with a brief reason for their exclusion. The included studies were published from 1998 to 2024 (meta-analysis studies: 2000–2024). The characteristics of the included studies are provided in Table 1.

Figure 1.

Study selection PRISMA approach [25].

Open in new tabDownload slide

A total of 20 462 elementary school pupils (49.9% female) with a mean age of 8.97 ± 1.25 years were included in this systematic review. The pupils were from various continents: Africa (South Africa), Asia (China: Beijing, Tianjin), Australia (Canberra, Gold Coast, Melbourne, New South Wales, Sydney), Europe (Denmark, England, France, Germany, Ireland, Italy, Portugal, Serbia, Spain, The Netherlands), North America (Canada: Manitoba, Richmond; United States: Kansas, Ohio, Oregon, Texas), and South America (Mexico: Mexico City). For the risk of bias assessment available in Supplementary Material S2, the study designs were found to have a high risk most frequently in the following areas: Blinding for the outcome assessment (n = 29) blinding of participants/staff (n = 27), and randomized allocation concealment from investigators (n = 26). One study (Santos et al., 2014) was found to have a low risk of bias for all areas investigated.

The interventions were primarily found to be for PA (n = 31), several were dietary (n = 6), and some were dual approach (n = 5). Of the PA interventions, two (8.33%) were found to meet the basic health recommendation of the World Health Organization (WHO) during compulsory education time (Larsen et al., 2018; Tan et al., 2016). Of the dietary interventions, one was vegetarian (Du et al., 2004), two were focused on a healthy omnivorous nutrition (e.g. increasing fruit and vegetable consumption) (Damsgaard et al., 2014), one was for increasing water consumption (Muckelbauer et al., 2009a; Muckelbauer et al., 2009b), one was a daily chewing micronutrient supplement aimed at reducing obesity (Long et al., 2022), and one was an energy-based nutrition educational board game (Viggiano et al., 2018). Of the dual approach interventions (Kipping et al., 2008; Long et al., 2022; Safdie et al., 2013; Santos et al., 2014), none were associated with the minimum dual approach to long-term, sustainable health [26].

As displayed in Fig. 2, it was found that PA intervention (n = 1193) in the primary school setting with at least 8 weeks’ duration has a mean effect size of 0.07 BW units (kg; 95% CI: −0.01 to 0.15) compared to controls (n = 1204). Statistical heterogeneity of the BW random effects meta-analysis model was found to be low (I² = 0%; P = 1.000). In Fig. 3, it was found that PA intervention (n = 2039) in the primary school setting with at least 8 weeks’ duration has a mean effect size of −0.12 BMI units (kg/m²; 95% CI: −0.23 to −0.01) compared to controls (n = 1913). Statistical heterogeneity of the BMI random effects meta-analysis model was found to be low (I² = 0%; P = .9688). In Supplementary Material S4, details of methodological heterogeneity across the meta-analysis included studies are provided. Across the meta-analysis studies, intervention durations ranged from 8 weeks to 3 years. The physical training types of the interventions were based on varieties of classroom exercises (n = 4), beach exercises (n = 1), exergaming (n = 1), jumping exercises (n = 3), cardiovascular activity (n = 6), strength training (n = 1), combined cardiovascular and strength activity (n = 3), high-intensity interval training (n = 1), and martial arts (n = 1). Some of the control group comparators performed usual physical education (PE) lessons in parallel with the PA intervention (n = 6), whereas most studies implemented extra PA time compared to the controls (n = 14). Intervention intensities were moderate (n = 8), moderate-to-vigorous (n = 9), or vigorous (n = 3).

Figure 2.

Meta-analysis: BW mean difference random effects in elementary school pupils. MD: mean difference; HK: Hartung-Knapp.

Open in new tabDownload slide

Figure 3.

Meta-analysis: BMI mean difference random effects in elementary school pupils. MD: mean difference; HK: Hartung-Knapp.

Open in new tabDownload slide

Discussion

The present study aimed to assess the effect of randomized controlled elementary school health interventions of PA and/or diet on BW and BMI [26]. The major findings include (i) almost all PA interventions (37 of 39; 94.9%) carried out as RCTs over at least 8 weeks in the primary school setting with an outcome on BW and/or BMI did not meet the minimum PA health recommendation of the WHO; (ii) the number of dietary and dual approach interventions were limited in the field; (iii) the BMI of pupils improved with additional PA intervention beyond the curriculum requirements; and (iv) the body composition profile (BW + BMI) of pupils improved with PA intervention. The present study confirms the hypothesis that PA intervention for elementary school pupils performed for a minimum of 8 weeks during mandatory school time has a positive effect on body composition.

For children and adolescents of school age (5–17 years), the WHO recommends a minimum of 60 min (on average) of moderate-to-vigorous PA daily [30]. Of all randomized controlled PA interventions (n = 31) carried out exclusively in primary schools over at least 8 weeks and investigating BW and BMI, only two studies (6.45%) met the minimum PA for health recommendation by the WHO (60 min/day) (Larsen et al., 2018; Tan et al., 2016). Likewise, several interventions (n = 8) included in the meta-analyses were not directly implemented to improve pupils’ BW and BMI (Fuchs et al., 2001; Ketelhut et al., 2016; Ketelhut et al., 2017b; Ketelhut et al., 2020; Ketelhut et al., 2021; Larsen et al., 2018; MacKelvie et al., 2002; McKay et al., 2000). In addition, studies (n = 6; Supplementary Material S4) included in the meta-analyses involved specialized PA interventions during PE but performed the same duration of PA as compared to the controls. Thus, for future scientific investigation in schools where overweight and obesity is a prevailing concern [3, 4, 9], it is a requirement that interventions achieve the minimum PA recommendation of the WHO (alongside PE classes, recess, and other planned PA opportunities) [30].

While increasing PA opportunities within mandatory school hours is a sound recommendation for improving future health [14], addressing the dietary component of pupil health is far more challenging [31–34], which may explain why only five studies with dietary interventions were identified in this study. Interestingly, in two of the dietary interventions identified, a “healthy” omnivorous approach was attempted with the underlying focus to increase fruit and vegetable consumption (Damsgaard et al., 2014; Rosário et al., 2012). One study met the criteria of a vegetarian approach considering the use of dairy products; however, a focus on healthy plant foods was not part of the intervention (Du et al., 2004). No studies were identified that attempted to integrate a healthy vegan approach, although the literature suggests this to be a safe, effective, and low-cost option for promoting long-term sustainable health through healthy body weight [18, 19, 35].

Of the five dual approach interventions identified by this study (Kipping et al., 2008; Long et al., 2022; Morgado et al., 2023; Safdie et al., 2013; Santos et al., 2014), the theoretical aspects that were attempted appear to outweigh the basic concept for improving the health situation in primary schools [36]. However, a minimum dual approach to health could serve as the most promising starting point for future interventions [36, 37], as the present results indicate that one-dimensional interventions are constrained in their potential to shape sustainable health outcomes. For future RCTs on the dual approach of permanently linking PA with a healthy diet, it is fundamental that activity is performed frequently throughout the day (outdoors at best) and meeting the minimum PA for health recommendation of the WHO [30]. Likewise, fueling the activity with healthy whole plant foods (fruits, vegetables, whole grains, and legumes) [26, 37] available in the classroom and cafeteria as well as eliminating the promotion of processed foods and advertising of specific brands from the school environment seems necessary [33].

The results of the BMI meta-analysis show an overall improvement of −0.12 kg/m² among elementary school pupils with PA intervention in the current epoch of an increasing BW crisis among youth (and adults) [3, 4, 9]. Although the BW benefits of PA intervention alone appear to be limited with short-term intervention [11, 15], and a healthy diet (high in dietary fiber from varieties of plant-based, whole foods) is likely the primary mechanism for a healthy, sustainable BW [32], the results are promising, showing that a simple, low-cost, and effective intervention of Exercise is Medicine [13, 38] may be immediately realized in primary schools. Exercise as Medicine [13] in primary schools could include moderate-to-vigorous PA via active lessons in any subject (which has no impact on learning outcomes) [39], active breaks during or between classes, or even daily exercise sessions, e.g. with additional PA opportunities (when practical) such as open spaces for active recess, voluntary bike desks, or voluntary standing desks (Barnes et al., 2021; Benden et al., 2011; Breheny et al., 2020; de Greeff et al., 2016a; de Greeff et al., 2016b; Donnelly et al., 2009; Drummy et al., 2016). Exercise as Medicine in primary schools should not be confined to the compulsory subject of PE but could include implementing daily PE classes (Ketelhut et al., 2016; Ketelhut et al., 2017a). Considering that the pupils in the intervention treatment gained additional BW (+0.07 kg) but had a reduced BMI overall, and despite small effects, it appears that the interventions not only reduced body fat among the pupils but increased fat-free mass such as muscle mass and bone density, favoring a healthier body composition profile [13–15]. Similar to the result of the present study, two previous meta-analyses reported a healthier BMI with intervention (−0.13 kg/m² and −0.39 kg/m²) [21, 22]; however, the latter study included a range of interventions on obesity-related behaviors, which may explain why there was a greater effect. One meta-analysis on the outcome of BW found an improvement from the intervention treatment (0.44 kg; Cochran’s Q statistic: 0.19) [23]. The previous studies reported considerable levels of statistical heterogeneity (I² = 66%; I² = 89.8%; and I² = 86%) [20–22].

One limitation of the dietary and dual approach results of the current study is the focus on interventions performed exclusively in the school setting [26]. This approach likely reduced the number of studies included from these domains but enhanced the overall precision on the school setting [26, 40]. The second limitation to be addressed is the quality of the PA interventions (duration, intensity, frequency, and type of activity) included in the meta-analyses to manage the pupil BW and BMI situation. Likewise, a particularly remarkable strength of the present study is the connection between BW and BMI funnel plots (Supplementary Material S1, pp. 45–46) with the low statistical heterogeneity (I² = 0%). The present results demonstrate that primary schools hold the potential to begin improving the BW crisis with immediate heightened PA implementation without necessarily requiring external funding, expert guidance, community, or parental involvement.

To conclude, the current behavioral restrictions of the primary school setting documented by the scientific investigations within the present review indicate poor PA levels among children. Improving the quantity and quality of PA behavior for pupils across the elementary school day contributes to short-term body composition management with the long-term implication of NCD regression. Furthermore, future randomized controlled trials are suggested to investigate more dietary (whole food plant-based, vegan) and dual approach interventions as a basic route to tapping vast health benefits from sustainably healthy BW to NCD prevention [18, 19]. To critically address global health crises, more progress towards the establishment of lifestyle medicine in the elementary school setting is required [38]; primary school boards are obligated to provide 60 min of moderate-to-vigorous intensity PA daily.

Acknowledgements

There are no professional relationships with companies or manufacturers benefiting from the article. D.T. was awarded a student scholarship from the Leopold-Franzens University of Innsbruck (LFUI award promoting young researchers, ID: 2020/2/PSY/SPORT-21; 02/2021-07/2022); however, there was and is no connection to the funding of this article and this had no role in study design, data collection or data analysis, preparation of the manuscript, or the decision to publish.

Author contributions

Derrick R. Tanous is the guarantor of this study. Katharina C. Wirnitzer provided conceptualization, supervision, and critical review. Gerhard Ruedl and Clemens Drenowatz provided supervision and critical review. Mohamad Motevalli performed investigation with the role of second reviewer and directly accessed and verified the underlying data in the article. Armando Cocca provided investigation and review. Werner Kirschner provided supervision regarding data management and review and Markus Schauer and Thomas Rosemann provided statistical methodology and review. All authors have had full access to all data in the study, and critically reviewed and agreed to the final version of the article to submit for publication.

Supplementary data

Supplementary data are available at EURPUB online.

Conflict of interest: The authors have no competing interests. The results of the study are presented clearly, honestly, and without fabrication, falsification, or inappropriate data manipulation. The results of the present study do not constitute endorsement by the American College of Sports Medicine.

Funding

There was no funding source for this study.

Data availability

Data will be provided upon request within reason. Further information can be found in the open access study protocol [26] as well as the PROSPERO trial registration (CRD42023477701).

References