- Email: [email protected]
Cardiorespiratory Fitness and Muscular Strength as Mediators of the Influence of Fatness on Academic Achievement
ARTICLE IN PRESS THE JOURNAL OF PEDIATRICS • www.jpeds.com
ORIGINAL ARTICLES
Cardiorespiratory Fitness and Muscular Strength as Mediators of the Influence of Fatness on Academic Achievement Antonio García-Hermoso, PhD1, Irene Esteban-Cornejo, PhD2, Jordi Olloquequi, PhD3, and Robinson Ramírez-Vélez, PhD4 Objective To examine the combined association of fatness and physical fitness components (cardiorespiratory fitness [CRF] and muscular strength) with academic achievement, and to determine whether CRF and muscular strength are mediators of the association between fatness and academic achievement in a nationally representative sample of adolescents from Chile. Study design Data were obtained for a sample of 36 870 adolescents (mean age, 13.8 years; 55.2% boys) from the Chilean System for the Assessment of Educational Quality test for eighth grade in 2011, 2013, and 2014. Physical fitness tests included CRF (20-m shuttle run) and muscular strength (standing long jump). Weight, height, and waist circumference were assessed, and body mass index and waist circumference-to-height ratio were calculated. Academic achievement in language and mathematics was assessed using standardized tests. The PROCESS script developed by Hayes was used for mediation analysis. Results Compared with unfit and high-fatness adolescents, fit and low-fatness adolescents had significantly higher odds for attaining high academic achievement in language and mathematics. However, in language, unfit and lowfatness adolescents did not have significantly higher odds for obtaining high academic achievement. Those with high fatness had higher academic achievement (both language and mathematics) if they were fit. Linear regression models suggest a partial or full mediation of physical fitness in the association of fatness variables with academic achievement. Conclusions CRF and muscular strength may attenuate or even counteract the adverse influence of fatness on academic achievement in adolescents. (J Pediatr 2017;■■:■■-■■). See related article, p •••
uccessful academic development during adolescence underpins potential occupational success later in life.1 Obesity might have a detrimental effect on academic achievement in adolescents2-6; however, this influence seems weak,7 and other factors, such as socioeconomic status,8 screen time,9 or physical fitness,10 might be interrelated. Physical fitness is a marker of health in children and adolescents.11 The main components of physical fitness are cardiorespiratory fitness (CRF), muscular strength, and motor ability; however, CRF and muscular strength are the components with the greatest potential to improve health.12 Previous studies have shown that CRF and muscular strength may prevent the development of cardiovascular and metabolic diseases, obesity, and mental illness. In addition, a growing body of evidence suggests that physical fitness is associated with better academic achievement,13 although this association may differ among specific components of physical fitness. More specifically, evidence for a relationship between muscular strength and academic achievement is scarce, but there is evidence of a positive association with CRF.14,15 Most previous studies did not take into consideration the combined association of physical fitness and fatness. Sardinha et al3 found that aerobically fit and normal-weight students were more likely to have better school performance. When examining the interdependence among fitness, fatness, and academic achievement, previous studies usually performed multiple linear regression, logistic regression, or analysis of covariance to adjust for confounding or mediator variables; however, these multivariate methods did not account for the percentage of the total effect explained by the potential From the 1Physical Activity, Sport and Health Sciences covariates.
S
BMI CRF SIMCE TEM WC WHtR
Body mass index Cardiorespiratory fitness System for the Assessment of Educational Quality Technical error of measurement Waist circumference Waist-to-height ratio
Laboratory, Faculty of Medical Sciences, Universidad de Santiago de Chile, USACH, Santiago, Chile; 2Promoting Fitness and Health Through Physical Activity Research Group, Department of Physical Education and Sport, Faculty of Sport Sciences, University of Granada, Granada, Spain; 3Faculty of Health Sciences, Universidad Autónoma de Chile, Talca, Chile; and 4Center of Studies in Physical Activity Measurements, School of Medicine and Health Sciences, Universidad del Rosario, Bogotá, Colombia The authors declare no conflicts of interest. 0022-3476/$ - see front matter. © 2017 Elsevier Inc. All rights reserved. http://dx.doi.org10.1016/j.jpeds.2017.04.037
1 FLA 5.4.0 DTD ■ YMPD9146_proof ■ May 16, 2017
THE JOURNAL OF PEDIATRICS • www.jpeds.com The purpose of this study was twofold: to examine the combined association of fatness and physical fitness components with academic achievement and to determine whether CRF and muscular strength are mediators of the association between fatness and academic achievement.
Methods The methodology of this study is described in detail elsewhere.16 In brief, this cross-sectional analysis is based on data drawn from a national sample of students (the Chilean Agency of Quality for Education [Agencia de Calidad de la Educación, Gobierno de Chile]), including all students enrolled in eighth grade (mean age, 13.83 years) who completed a standardized test, the System for Assessment of Educational Quality (SIMCE) and the SIMCE in physical education, administered by the Chilean Ministry of Education in November 2011, 2013, and 2014.17 The SIMCE test was not carried out in 2012. The sample was stratified by 15 regions in Chile (with the exception of Easter Island, the Juan Fernández archipelago, and the Antarctic) and by 3 school types (public, private subsidized, and private nonsubsidized). Within each stratum, schools were the primary sampling unit, and all students in the selected schools were sampled. A total of 50 549 students participated in the SIMCE in physical education. After exclusion of students because of erroneous data entry, disability, temporary illness or injury, chronic illness, absenteeism, age ≥18 years, missing body mass index (BMI) data, missing waist circumference (WC) data, or missing academic data, the present series comprised a total of 36 870 adolescents (72.94 % of the original sample) from 3 different cohorts and 669 schools. The SIMCE test for physical education consists of a fitness and anthropometric assessment using a standardized battery of fitness tests administered in November of each test year. The full protocol and battery of tests have been described previously.16 For the anthropometric measurements, the students wore light clothing and were barefoot. Data were recorded on paper by Ministry of Education evaluators.17 Weight was measured to the nearest 0.1 kg, and height was measured to the nearest 0.1 cm. BMI was calculated as body weight in kilograms divided by the square of height in meters. Weight status and increased adiposity were defined as a BMI above the age- and sex-specific thresholds of the International Obesity Task Force.18 WC was measured by placing a nonelastic tape measure midway between the lowest rib margin and the iliac crest and recorded to the nearest 0.1 cm. WC was classified using criterion-referenced, health-related cutpoints derived from de Ferranti et al19 because of the large size, age-specificity, and relatively generalizable ethnicity (ie, the Chilean population is roughly 59% non-Hispanic white) of the sample. WC and height were used to calculate waist-to-height ratio (WHtR); central obesity was defined as a WHtR ratio ≥0.5.20 Testing procedures were consistent with guidelines for schoolbased physical fitness assessment.21 At each school, a team of trained Ministry of Education evaluators (n = 5 each year) ad-
Volume ■■ ministered the tests with the same instruments in partnership with the physical education instructor. Tests were administered in the school gymnasium or on another available hard surface.17 Consistent with recommendations,22,23 our analysis was limited to health-related, valid, and reliable fieldbased tests, such as the 20-m shuttle run and standing long jump test. The 20-m shuttle run was used to assess CRF. The test was performed as described by Léger et al.24 Each participant ran in a straight line between 2 lines 20 m apart while keeping pace with prerecorded audio signals. The initial speed of 8.5 km/ hour was increased by 0.5 km/hour each minute. The test was completed when the participant failed to reach the end lines keeping pace with the audio signals on 2 consecutive occasions or when he or she stopped because of fatigue. Results were recorded to the nearest stage completed. The Léger equation was used to determine peak oxygen uptake (mL/kg/ minute) in each adolescent.24 Fit adolescents were defined using the age- and sex-specific cutoffs listed for the healthy fitness zone in the 2011 FITNESSGRAM.25 The standing long jump was used to assess muscular strength. In this test, the participant stood behind the starting line and was instructed to push off vigorously and jump as far as possible. The participant had to land with the feet together and remain upright. The test was repeated twice, and the longer distance was recorded to the nearest 0.1 cm. To classify muscular strength, we considered the risk group to be the age- and sex-specific 20th percentile of standing long jump for European adolescents,26 because the lower quintile in this sample had the strongest association with a poor cardiometabolic risk profile.27 All measurements in a subsample of 50 adolescents (28 girls, 22 boys; mean [SD] age, 13.7 [1.0] years; mean [SD] weight, 56.2 [10.4] kg; mean [SD] height, 1.60 [7.0] m; mean [SD] BMI, 20.9 [3.5]) were repeated by administering the tests again 1 week later with several physical education instructors from the last SIMCE (2015). Technical errors of measurement (TEMs) for weight, height, and ANOVA-based intraclass correlations (R) and corresponding 95% CIs were used to estimate test-retest reliability. For anthropometric measures, TEMs were 0.1 cm for height, 0.1 kg for weight, and 0.2 cm for WC. For physical fitness tests, R ranged from 0.89 (95% CI, 0.840.92) for the 20-m shuttle run to 0.96 (95% CI, 0.94-0.98) for the standing broad jump. The SIMCE measures national curricular objectives for language and mathematics as established by the Chilean Ministry of Education. The language test evaluates reading and writing ability. The mathematics test evaluates the ability to understand concepts and numerical operations, ability to use simple nonfractional algebraic expressions and apply them to solve problems using the correct approach, and ability to solve first-degree equations with a single unknown quantity. The test, administered by the Ministry of Education evaluators, comprises both multiple-choice items and open-ended questions, and is scored on a scale of 0-400. In the present study, scores were categorized according to the achievement levels established by the Ministry of Educa-
2
Garcia-Hermoso et al FLA 5.4.0 DTD ■ YMPD9146_proof ■ May 16, 2017
■■ 2017
ORIGINAL ARTICLES
tion.17 The minimum scores for demonstrating proficiency were 292 for language and 297 for mathematics. The National Physical Education Survey was authorized under the Chilean Law of Sport, number 19.712, article 5.17 The Ministry of Education solicited consent from the schools before testing and instructed each school to inform parents and students with a standardized letter about the nature and importance of the fitness tests, the assessment date and time, and recommended test preparation.17 The nature and purpose of the study were communicated to potential participants and their parents or guardians, with the explanation that data would be available to the Chilean education authorities in accordance with the Law of Data Protection (Law 21120/2016). The Ministry of Education requested that all adolescents and parents consent to be part of the sample; therefore, written informed consent was obtained from parents and subjects before participation in the study. Because this was a secondary analysis of deidentified data, the University Manuela Beltran (Colombia) Institutional Review Board considered the study exempt from review. The first author applied to the Ministry of Education and obtained permission to use the publicly available data for research and teaching learning purposes. Further details are available at http://www.simce.cl/. Socioeconomic status was defined by the type of school. The Chilean educational system is decentralized and consists of 3 types of schools: public, private subsidized, and private nonsubsidized. As might be expected, a family’s socioeconomic status is highly predictive of the type of school their children attend; that is, with very few exceptions, low socioeconomic status families tend to send their children to public
schools, middle socioeconomic status families tend to send their children to private subsidized schools, and high socioeconomic status families tend to send their children to private nonsubsidized schools.28 Statistical Analyses Results are presented as mean (SD) or relative frequency (%). Differences between the sexes were tested using the t test or the c2 test for unadjusted means or frequencies, respectively. The normality of the distribution of the variables was tested using both graphical (normal probability plot) and statistical (Kolmogorov-Smirnov test) procedures. Because of their skewed distribution, the following variables were log-transformed before analysis: BMI, WC, WHtR, CRF, and muscular strength. ANCOVA models were used to assess mean differences in academic achievement (language and mathematics) among the categories of BMI, WC, WHtR, and physical fitness components (CRF and muscular strength), controlling for age, sex, assessment year, and socioeconomic status (model 1), and also adjusting for CRF and muscular strength or BMI depending on the fixed factor (model 2). Pairwise post hoc hypotheses were tested using the Bonferroni correction for multiple comparisons. The ORs and 95% CIs of having a score of proficiency in language and mathematics by combined factors of fatness and physical fitness variables (ie, CRF-BMI, CRF-WC, CRFWHtR, muscular strength-BMI, muscular strength-WC, and muscular strength-WHtR) were calculated through multinomial logistic regression adjusted by age, sex, assessment year, and socioeconomic status. For the combined analysis, CRF and
Table I. Characteristics of the sample by sex Characteristics Physical characteristics Age, y, mean (SD) Weight, kg, mean (SD) Height, cm, mean (SD) Fatness parameters BMI, kg/m2, mean (SD) High fat, n (%) WC, cm, mean (SD) High fat, n (%) WHtR, mean (SD) High fat, n (%) Physical fitness CRF, stage, mean (SD) CRF, mL/kg/min, mean (SD)* Unfit, n (%) Muscular strength, cm, mean (SD)† Unfit, n (%) Academic achievement (SIMCE) Language, raw score, mean (SD) Score for proficiency, n (%) Mathematics, raw score, mean (SD) Score for proficiency, n (%) Socioeconomic status, n (%) Low Medium High
Total (n = 36 870)
Boys (n = 19 827)
Girls (n = 17 043)
P value for sex
13.8 (0.7) 57.0 (10.6) 161.0 (7.9)
13.9 (0.8) 58.3 (11.0) 164.3 (7.7)
13.8 (0.7) 55.4 (10.0) 157.0 (6.0)
<.001 <.001 <.001
21.9 4940 72.3 3503 0.44 4056
(3.6) (13.4) (8.9) (9.5) (0.05) (11.0)
21.5 2102 73.3 1983 0.44 2022
(3.5) (10.6) (8.8) (10.0) (0.05) (10.2)
22.5 2838 71.0 1520 0.45 2033
(3.7) (16.6) (8.8) (9.0) (0.06) (12.0)
<.001 <.001 <.001 .004 <.001 <.001
5.5 43.9 4498 151.4 6673
(2.5) (6.6) (12.2) (32.4) (18.1)
6.8 (2.3) 47.2 (6.0) 1864 (9.4) 168.0 (28.5) 3053(15.4)
4.0 39.9 2634 130.8 3620
(1.9) (5.1) (15.6) (24.0) (21.4)
<.001 <.001 <.001 <.001 <.001
252.0 8554 261.5 8812
(50.9) (23.2) (48.3) (23.9)
246.3 3965 263.3 4897
259.0 4589 259.4 3915
(49.1) (26.4) (49.0) (23.0)
<.001 <.001 <.001 .001
7636 (44.9) 8098 (47.6) 1269 (7.4)
<.001
16 518 (44.8) 17 218 (46.7) 3093 (8.39)
(51.7) (20.0) (47.8) (24.7)
8882 (44.8) 9120 (46.0) 1824 (9.2)
*Measured by the 20-minute shuttle run test. †Measured by the standing long jump test.
Cardiorespiratory Fitness and Muscular Strength as Mediators of the Influence of Fatness on Academic Achievement FLA 5.4.0 DTD ■ YMPD9146_proof ■ May 16, 2017
3
THE JOURNAL OF PEDIATRICS • www.jpeds.com Values are mean ± SD. Categories are L, M, and H, representing the first, second-third, and fourth quartiles, respectively. Categories of BMI are normal weight, overweight, and obese, according to sex- and age-specific cutoffs as defined by the International Obesity Task Force guidelines. Model 1: adjusted for age, sex, assessment year, and socioeconomic status. Model 2: model 1 covariates + CRF and muscular strength or BMI, depending on the fixed factor. *Bonferroni-adjusted pairwise comparisons.
L>H L>H L>H L>H L>H .009 <.001 <.001 <.001 <.001 L>M>H L > H; M > H L>M>H L>M>H L>M>H <.001 <.001 <.001 <.001 <.001 254.4 ± 46.1 255.1 ± 46.9 253.8 ± 46.6 266.7 ± 49.4 269.9 ± 50.4 260.6 ± 48.2 262.3 ± 48.7 261.0 ± 48.4 260.7 ± 47.5 259.5 ± 47.3 263.3 ± 48.9 265.1 ± 48.7 268.5 ± 48.9 250.7 ± 46.6 253.7 ± 46.5 — L>H L>H L>H L>H .071 <.001 <.001 <.001 <.001 L>M>H L > H; M > H L>M>H L>M>H L>M>H <.001 <.001 <.001 <.001 <.001 251.8 ± 50.5 252.8 ± 51.5 251.4 ± 51.4 251.2 ± 50.8 250.3 ± 50.4 BMI WC WHtR CRF Muscular strength
252.4 ± 51.7 254.9 ± 50.8 258.0 ± 50.8 242.4 ± 50.7 254.7 ± 49.9
247.7 ± 49.0 245.8 ± 49.6 245.6 ± 49.2 254.4 ± 51.0 258.3 ± 51.9
Post hoc* P value Medium (M) Categories
Low (L)
High (H)
P
Post hoc*
P value
Post hoc*
Low (L)
Medium (M)
High (H)
P value
Post hoc*
Model 2 Model 1
Mathematics (0-400)
Crude data Model 2
4
Model 1
Language (0-400)
Table I presents the demographic descriptive statistics of the sample. The final sample had a mean age of 13.8 ± 0.7 years and contained more boys than girls (55.2% vs 44.8%). Compared with the boys, the girls had lower WC, CRF, muscular strength, and mathematics scores, as well as a higher prevalence of obesity according to the International Obesity Task Force criteria and unfit profile (P < .001). Differences in academic achievement across fatness and physical fitness categories are shown in Table II. In model 1, overweight and obese adolescents had significantly lower language and mathematics scores compared with their normalweight peers (P < .001), but in model 2, after adjusting for CRF and muscular strength, these significant differences disappeared or were attenuated for language (P = .071) and mathematics (P = .009), respectively. Regarding WC and WHtR, adolescents in the medium and high quartiles had significantly lower scores in language and mathematics compared with their normal-weight peers in both models. Adolescents with high levels of CRF and muscular strength had significantly higher academic achievement scores compared with their peers with low and medium fitness levels in both models (P < .001). Table III shows the ORs of the relationship between high academic achievement in language and mathematics and the combined association of fatness (BMI, WC, and WHtR) and physical fitness (CRF and muscular strength). Adolescents classified in the fit and high fatness group and in the fit and low fatness group had higher odds of having high academic
Crude data
Results
Table II. Mean differences in academic achievement parameters by fatness and physical fitness categories controlling for potential confounders
muscular strength were categorized as “fit” or “unfit.”25 Each measure of adiposity (BMI, WC, and WHtR) was categorized as “high fatness” or “low fatness.” Four groups were created: unfit and high fatness, unfit and low fatness, fit and high fatness, and fit and low fatness. To examine whether the associations between measures of fatness or physical fitness and academic achievement were mediated by physical fitness or fatness, linear regression models were fitted using bootstrapped mediation procedures included in the PROCESS SPSS script.29 This script used bootstrapping methods30 for testing mediation hypotheses, using a resampling procedure of 10 000 bootstrap samples. The aim of the model was to investigate the total effects (equation 3) and direct effects (equations 1, 2, and 3’), reflected by the unstandardized regression coefficient and significance between the independent and dependent variables in each model. To establish mediation, equations 1-3 must be significant, and the unstandardized coefficients between the independent and dependent variable must be attenuated (partial mediation) or become (close to) null (full mediation) when the mediator is included in the regression model (equation 3’).31 The Sobel test was used to test the hypothesis that the indirect effect was equal to zero.32 This analysis was adjusted by age, sex, assessment year, and socioeconomic status. All the analyses were performed using SPSS 21 (IBM, Armonk, New York). The level of statistical significance was established as P < .05.
Volume ■■
Garcia-Hermoso et al FLA 5.4.0 DTD ■ YMPD9146_proof ■ May 16, 2017
■■ 2017
ORIGINAL ARTICLES
Table III. Combined effects of physical fitness and fatness parameters on score for proficiency in language and mathematics Parameters
CRF and BMI Unfit and low fat Fit and high fat Fit and low fat CRF and WC Unfit and low fat Fit and high fat Fit and low fat CRF and WHtR Unfit and low fat Fit and high fat Fit and low fat Muscular fitness and BMI Unfit and low fat Fit and high fat Fit and low fat Muscular fitness and WC Unfit and low fat Fit and high fat Fit and low fat Muscular fitness and WHtR Unfit and low fat Fit and high fat Fit and low fat
Language
Mathematics
OR
95% CI
P
OR
95% CI
P
1.055 1.187 1.349
0.905-1.231 1.024-1.377 1.184-1.538
.493 .023 <.001
1.287 1.444 1.728
1.090-1.521 1.229-1.695 1.497-1.995
.003 <.001 <.001
1.162 1.298 1.538
0.965-1.497 1.087-1.551 1.322-1.790
.067 .004 <.001
1.360 1.515 1.805
1.135-1.629 1.258-1.825 1.536-2.120
.001 <.001 <.001
1.102 1.214 1.498
0.920-1.416 1.035-1.425 1.303-1.721
.058 .017 <.001
1.366 1.419 1.812
1.150-1.623 1.191-1.691 1.559-2.107
<.001 <.001 <.001
1.159 1.261 1.396
0.975-1.323 1.096-1.451 1.238-1.573
.059 .001 <.001
1.289 1.437 1.717
1.119-1.486 1.273-1.669 1.509-1.953
<.001 <.001 <.001
1.149 1.169 1.425
0.973-1.358 1.012-1.351 1.247-1.628
.102 .034 <.001
1.288 1.422 1.719
1.104-1.502 1.195-1.691 1.491-1.981
.001 <.001 <.001
1.140 1.196 1.450
0.976-1.332 1.044-1.370 1.281-1.641
.099 .010 <.001
1.382 1.443 1.809
1.193-1.599 1.224-1.702 1.583-2.068
<.001 <.001 <.001
Reference group (OR, 1.0): adolescents classified as unfit and with high fatness. Analysis adjusted by age, sex, assessment year, and socioeconomic status.
achievement in both mathematics (all P < .001) and language (all P < .034) compared with unfit and high fatness adolescents. In contrast, compared with adolescents in the unfit and high fatness group, those in the unfit and low fatness group had higher odds of having high academic achievement in mathematics (all P < .003) but not in language (P = .058-.493). We performed a mediation analysis to test whether CRF and muscular strength acted as mediator variables between academic achievement (dependent variable) and fatness (independent variable) (Figures 1, 2, and 3; available at www.jpeds.com). As shown in Figure 1, the effect of BMI on language was fully mediated, but the effect of BMI on mathematics was only partially mediated, by CRF and muscular strength. In the first regression step (equation 1), BMI was negatively correlated with the physical fitness component (CRF and muscular strength) (P < .001). In the second step (equation 3), the regression coefficient of BMI on the dependent variable (language and mathematics) was negatively correlated (P < .001). In the last regression model, the mediator variable was positively correlated with the dependent variable (equation 2) (P < .001), but when both CRF and muscular strength were included in the model (equation 3’), the regression coefficient did not maintain its statistical significance (full mediation) in language, but did so in mathematics (partial mediation). Figures 2 and 3 show that the effect of WC and WHtR on academic achievement (both language and mathematics) was partially mediated by CRF and muscular strength.
Discussion Experimental evidence in neuroscience indicates that unhealthy body composition may hamper learning and aca-
demic performance.33 For example, gray matter reductions in the prefrontal cortex, a brain region involved in cognitive control, appear to occur in a dose-dependent manner with increasing BMI.33 Regarding physical fitness, CRF seems to be related to angiogenesis in the motor cortex and increased blood flow, resulting in improved brain vascularization, which could affect cognitive performance.34 In addition, CRF has been shown to play roles in spatial memory, hippocampal volume, and white matter integrity, whereas muscular strength has been related to synaptogenesis.35 Numerous studies have examined the associations of fitness and fatness individually with academic achievement; however, the findings are inconsistent.12 To date, only 1 study, of 1531 youths from 3 different cohorts has examined the combined association of CRF and weight status.3 That study showed that in students of good cardiorespiratory fitness and normal weight, the odds of high academic achievement were increased by 449% compared with unfit and overweight peers, suggesting that the combination of these components is a predictor of high academic achievement in language-related areas. Our findings confirm these results regardless of the physical fitness or fatness variables used. In mathematics, the odds of high academic achievement were significantly higher in adolescents classified as unfit and low fatness compared with those classified as unfit and high fatness. In addition, fit adolescents with either low or high fatness were more likely to have high academic achievement in both mathematics and language. Our results highlight the higher odds of WC or WHtR compared with BMI for predicting academic achievement. In this study, we also aimed to test the phenotype known as “fat but fit.”36 Our results suggest that the effect of fatness on academic achievement is fully mediated (BMI) or
Cardiorespiratory Fitness and Muscular Strength as Mediators of the Influence of Fatness on Academic Achievement FLA 5.4.0 DTD ■ YMPD9146_proof ■ May 16, 2017
5
THE JOURNAL OF PEDIATRICS • www.jpeds.com partially mediated (WC and WHtR) by CRF and muscular strength. BMI appears to be less independent of context to assess the relationship between fatness and academic achievement (ie, only for language); this measure is commonly used to define being overweight/obese, but is not a specific measure of fatness. The study reported by Van Dusen et al37 suggested a nonlinear association between BMI and academic achievement after covariate adjustment, indicating that physical fitness may be more important than weight control per se for academic development. According to Biddle and Asare,38 several explanations for the mediation role of physical fitness based on psychosocial aspects have been suggested: (1) the physical fitness of students is associated with better health, which can positively contribute to academic achievement; (2) increased physical fitness can improve the attention and behavior of adolescents in the classroom; and (3) greater physical fitness can improve mental health and self-esteem and can relieve stress, anxiety and depression, which in turn can positively affect academic achievement. The main limitation of the present study is its crosssectional design, which does not allow any inference of temporality or possible causal relationships. In addition, it is possible that the correlation among academic achievement, fatness, and physical fitness was mediated by variables not included in this study, such as cognitive outcomes, maturational stage, or other socioeconomic indicators (eg, maternal education). Finally, other tests to determine muscular strength are required, such as a handgrip test or an assessment of motor ability. EstebanCornejo et al39 showed that motor ability is more strongly related to academic achievement in youth compared with CRF. It may be true that muscular strength, as measured by the standing long jump, is a proxy for motor ability, which may explain the strength of its relationship with academic achievement. This study has several analytical strengths, including the mediation analysis and the large sample collected at a national level, including 3 different cohorts, which allowed us to extend previous findings to adolescents born in different years. Every effort should be made to sustain an effective physical education curriculum during the key period of adolescence, given that improving physical fitness can be a possible strategy for further improving the school performance of this population. The inclusion of a physical education assessment in the national education survey reflects an understanding of the importance of physical fitness and fatness for the current and future academic and health status of Chilean youth.40 ■ Submitted for publication Dec 9, 2016; last revision received Mar 31, 2017; accepted Apr 17, 2017 Reprint requests: Antonio García-Hermoso, PhD, Laboratorio de Ciencias de la Actividad Física, el Deporte y la Salud, Facultad de Ciencias Médicas, Universidad de Santiago de Chile, Avenida Libertador Bernardo, O’Higgins n° 3363, Estación Central, Santiago, Chile. E-mail: [email protected].
Volume ■■
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17. 18.
19.
20.
References
21.
1. Arora T, Hosseini-Araghi M, Bishop J, Yao G, Thomas GN, Taheri S. The complexity of obesity in UK adolescents: relationships with quantity and
22.
type of technology, sleep duration and quality, academic performance and aspiration. Pediatr Obes 2013;8:358-66. García-Hermoso A, Marina R. Relationship of weight status, physical activity and screen time with academic achievement in adolescents. Obes Res Clin Pract 2017;11:44-50. Sardinha LB, Marques A, Martins S, Palmeira A, Minderico C. Fitness, fatness, and academic performance in seventh-grade elementary school students. BMC Pediatr 2014;14:176. Esteban-Cornejo I, Tejero-González CM, Castro-Piñero J, CondeCaveda J, Cabanas-Sanchez V, Sallis JF, et al. Independent and combined influence of neonatal and current body composition on academic performance in youth: the UP & DOWN study. Pediatr Obes 2015;10:15764. Olivares PR, García-Rubio J. Associations between different components of fitness and fatness with academic performance in Chilean youths. PeerJ 2016;4:e2560. Morita N, Nakajima T, Okita K, Ishihara T, Sagawa M, Yamatsu K. Relationships among fitness, obesity, screen time and academic achievement in Japanese adolescents. Physiol Behav 2016;163:161-6. Suchert V, Hanewinkel R, Isensee B. Longitudinal relationships of fitness, physical activity, and weight status with academic achievement in adolescents. J Sch Health 2016;86:734-41. Datar A, Sturm R, Magnabosco JL. Childhood overweight and academic performance: national study of kindergartners and first-graders. Obes Res 2004;12:58-68. Aguilar MM, Vergara FA, Velásquez EJ, Marina R, García-Hermoso A. Screen time impairs the relationship between physical fitness and academic attainment in children. J Pediatr (Rio J) 2015;91:339-45. Torrijos-Niño C, Martínez-Vizcaíno V, Pardo-Guijarro MJ, GarcíaPrieto JC, Arias-Palencia NM, Sánchez-López M. Physical fitness, obesity, and academic achievement in schoolchildren. J Pediatr 2014;165:104-9. Ortega FB, Ruiz JR, Castillo MJ, Sjöström M. Physical fitness in childhood and adolescence: a powerful marker of health. Int J Obes (Lond) 2008;32:1-11. Donnelly JE, Hillman CH, Castelli D, Etnier JL, Lee S, Tomporowski P, et al. Physical activity, fitness, cognitive function, and academic achievement in children: a systematic review. Med Sci Sports Exerc 2016;48:12234. Chaddock L, Hillman CH, Pontifex MB, Johnson CR, Raine LB, Kramer AF. Childhood aerobic fitness predicts cognitive performance one year later. J Sports Sci 2012;30:421-30. Santana CC, Azevedo LB, Cattuzzo MT, Hill JO, Andrade LP, Prado WL. Physical fitness and academic performance in youth: a systematic review. Scand J Med Sci Sports 2016; doi:10.1111/sms.12773. [Epub ahead of print]. Sardinha LB, Marques A, Minderico C, Palmeira A, Martins S, Santos DA, et al. Longitudinal relationship between cardiorespiratory fitness and academic achievement. Med Sci Sports Exerc 2016;48:839-44. Garber MD, Sajuria M, Lobelo F. Geographical variation in healthrelated physical fitness and body composition among Chilean 8th graders: a nationally representative cross-sectional study. PLoS ONE 2014;9:e108053. Agencia de Calidad de la Educación, Gobierno de Chile. SIMCE. Available at: http://www.simce.cl/. Accessed October 1, 2016. Cole TJ, Bellizzi MC, Flegal KM, Dietz WH. Establishing a standard definition for child overweight and obesity worldwide: international survey. BMJ 2000;320:1240. de Ferranti SD, Gauvreau K, Ludwig DS, Neufeld EJ, Newburger JW, Rifai N. Prevalence of the metabolic syndrome in American adolescents: findings from the Third National Health and Nutrition Examination Survey. Circulation 2004;110:2494-7. Bauer KW, Marcus MD, El ghormli L, Ogden CL, Foster GD. Cardiometabolic risk screening among adolescents: understanding the utility of body mass index, waist circumference and waist to height ratio. Pediatr Obes 2015;10:329-37. Santos R, Mota J. The ALPHA health-related physical fitness test battery for children and adolescents. Nutr Hosp 2011;26:1199-200. Pate R, Oria M, Pillsbury L. Fitness measures and health outcomes in youth. Washington (DC): The National Academies Press; 2012.
6
Garcia-Hermoso et al FLA 5.4.0 DTD ■ YMPD9146_proof ■ May 16, 2017
■■ 2017
ORIGINAL ARTICLES
23. Ruiz JR, Castro-Piñero J, España-Romero V, Artero EG, Ortega FB, Cuenca MM, et al. Field-based fitness assessment in young people: the ALPHA health-related fitness test battery for children and adolescents. Br J Sports Med 2011;45:518-24. 24. Léger LA, Mercier D, Gadoury C, Lambert J. The multistage 20 metre shuttle run test for aerobic fitness. J Sports Sci 1988;6:93-101. 25. Welk GJ, Laurson KR, Eisenmann JC, Cureton KJ. Development of youth aerobic-capacity standards using receiver operating characteristic curves. Am J Prev Med 2011;41(4 Suppl 2):S111-6. 26. Ortega FB, Artero EG, Ruiz JR, España-Romero V, Jiménez-Pavón D, Vicente-Rodriguez G, et al. Physical fitness levels among European adolescents: the HELENA study. Br J Sports Med 2011;45:209. 27. Artero EG, Ruiz JR, Ortega FB, España-Romero V, Vicente-Rodríguez G, Molnar D, et al. Muscular and cardiorespiratory fitness are independently associated with metabolic risk in adolescents: the HELENA study. Pediatr Diabetes 2011;12:704-12. 28. Taut S, Cortéés F, Sebastian C, Preiss D. Evaluating school and parent reports of the national student achievement testing system (SIMCE) in Chile: access, comprehension, and use. Eval Program Plann 2009;32:12937. 29. Hayes AF. Introduction to mediation, moderation, and conditional process analysis: a regression-based approach. New York (NY): Guilford Press; 2013. 30. Preacher KJ, Hayes AF. Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behav Res Methods 2008;40:879-91.
31. Baron RM, Kenny DA. The moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations. J Pers Soc Psychol 1986;51:1173-82. 32. Sobel ME. Asymptotic confidence intervals for indirect effects in structural equation models. Sociol Methodol 1982;13:290-312. 33. Taki Y, Kinomura S, Sato K, Inoue K, Goto R, Okada K, et al. Relationship between body mass index and gray matter volume in 1,428 healthy individuals. Obesity (Silver Spring) 2008;16:119-24. 34. Hillman CH, Erickson KI, Kramer AF. Be smart, exercise your heart: exercise effects on brain and cognition. Nat Rev Neurosci 2008;9:58-65. 35. Adkins DL, Boychuk J, Remple MS, Kleim JA. Motor training induces experience-specific patterns of plasticity across motor cortex and spinal cord. J Appl Physiol (1985) 2006;101:1776-82. 36. McAuley PA, Blair SN. Obesity paradoxes. J Sports Sci 2011;29:773-82. 37. Van Dusen DP, Kelder SH, Kohl HW 3rd, Ranjit N, Perry CL. Associations of physical fitness and academic performance among schoolchildren. J Sch Health 2011;81:733-40. 38. Biddle SJ, Asare M. Physical activity and mental health in children and adolescents: a review of reviews. Br J Sports Med 2011;45:886-95. 39. Esteban-Cornejo I, Tejero-González CM, Martinez-Gomez D, del Campo J, González-Galo A, Padilla-Moledo C, et al. Independent and combined influence of the components of physical fitness on academic performance in youth. J Pediatr 2014;165:306-12.e2. 40. Weintraub WS, Daniels SR, Burke LE, Franklin BA, Goff DC Jr, Hayman LL, et al. Value of primordial and primary prevention for cardiovascular disease a policy statement from the American Heart Association. Circulation 2011;124:967-90.
Cardiorespiratory Fitness and Muscular Strength as Mediators of the Influence of Fatness on Academic Achievement FLA 5.4.0 DTD ■ YMPD9146_proof ■ May 16, 2017
7
THE JOURNAL OF PEDIATRICS • www.jpeds.com
Volume ■■
Figure 1. Physical fitness mediation models of the relationship between BMI and academic achievement score in adolescents, adjusted by age, sex, assessment year, socioeconomic status, and CRF or muscular fitness according to mediation variable included in the model. MS, muscular strength. *P < .05; **P < .001.
7.e1
Garcia-Hermoso et al FLA 5.4.0 DTD ■ YMPD9146_proof ■ May 16, 2017
■■ 2017
ORIGINAL ARTICLES
Figure 2. Physical fitness mediation models of the relationship between WC and academic achievement score in adolescents, adjusted by age, sex, assessment year, socioeconomic status, and CRF or muscular fitness according to mediation variable included in the model. *P < .05; **P < .001.
Cardiorespiratory Fitness and Muscular Strength as Mediators of the Influence of Fatness on Academic Achievement FLA 5.4.0 DTD ■ YMPD9146_proof ■ May 16, 2017
7.e2
THE JOURNAL OF PEDIATRICS • www.jpeds.com
Volume ■■
Figure 3. Physical fitness mediation models of the relationship between WHtR and academic achievement score in adolescents, adjusted by age, sex, assessment year, socioeconomic status, and CRF or muscular fitness according to mediation variable included in the model. *P < .05; **P < .001.
7.e3
Garcia-Hermoso et al FLA 5.4.0 DTD ■ YMPD9146_proof ■ May 16, 2017
ORIGINAL ARTICLES
Cardiorespiratory Fitness and Muscular Strength as Mediators of the Influence of Fatness on Academic Achievement Antonio García-Hermoso, PhD1, Irene Esteban-Cornejo, PhD2, Jordi Olloquequi, PhD3, and Robinson Ramírez-Vélez, PhD4 Objective To examine the combined association of fatness and physical fitness components (cardiorespiratory fitness [CRF] and muscular strength) with academic achievement, and to determine whether CRF and muscular strength are mediators of the association between fatness and academic achievement in a nationally representative sample of adolescents from Chile. Study design Data were obtained for a sample of 36 870 adolescents (mean age, 13.8 years; 55.2% boys) from the Chilean System for the Assessment of Educational Quality test for eighth grade in 2011, 2013, and 2014. Physical fitness tests included CRF (20-m shuttle run) and muscular strength (standing long jump). Weight, height, and waist circumference were assessed, and body mass index and waist circumference-to-height ratio were calculated. Academic achievement in language and mathematics was assessed using standardized tests. The PROCESS script developed by Hayes was used for mediation analysis. Results Compared with unfit and high-fatness adolescents, fit and low-fatness adolescents had significantly higher odds for attaining high academic achievement in language and mathematics. However, in language, unfit and lowfatness adolescents did not have significantly higher odds for obtaining high academic achievement. Those with high fatness had higher academic achievement (both language and mathematics) if they were fit. Linear regression models suggest a partial or full mediation of physical fitness in the association of fatness variables with academic achievement. Conclusions CRF and muscular strength may attenuate or even counteract the adverse influence of fatness on academic achievement in adolescents. (J Pediatr 2017;■■:■■-■■). See related article, p •••
uccessful academic development during adolescence underpins potential occupational success later in life.1 Obesity might have a detrimental effect on academic achievement in adolescents2-6; however, this influence seems weak,7 and other factors, such as socioeconomic status,8 screen time,9 or physical fitness,10 might be interrelated. Physical fitness is a marker of health in children and adolescents.11 The main components of physical fitness are cardiorespiratory fitness (CRF), muscular strength, and motor ability; however, CRF and muscular strength are the components with the greatest potential to improve health.12 Previous studies have shown that CRF and muscular strength may prevent the development of cardiovascular and metabolic diseases, obesity, and mental illness. In addition, a growing body of evidence suggests that physical fitness is associated with better academic achievement,13 although this association may differ among specific components of physical fitness. More specifically, evidence for a relationship between muscular strength and academic achievement is scarce, but there is evidence of a positive association with CRF.14,15 Most previous studies did not take into consideration the combined association of physical fitness and fatness. Sardinha et al3 found that aerobically fit and normal-weight students were more likely to have better school performance. When examining the interdependence among fitness, fatness, and academic achievement, previous studies usually performed multiple linear regression, logistic regression, or analysis of covariance to adjust for confounding or mediator variables; however, these multivariate methods did not account for the percentage of the total effect explained by the potential From the 1Physical Activity, Sport and Health Sciences covariates.
S
BMI CRF SIMCE TEM WC WHtR
Body mass index Cardiorespiratory fitness System for the Assessment of Educational Quality Technical error of measurement Waist circumference Waist-to-height ratio
Laboratory, Faculty of Medical Sciences, Universidad de Santiago de Chile, USACH, Santiago, Chile; 2Promoting Fitness and Health Through Physical Activity Research Group, Department of Physical Education and Sport, Faculty of Sport Sciences, University of Granada, Granada, Spain; 3Faculty of Health Sciences, Universidad Autónoma de Chile, Talca, Chile; and 4Center of Studies in Physical Activity Measurements, School of Medicine and Health Sciences, Universidad del Rosario, Bogotá, Colombia The authors declare no conflicts of interest. 0022-3476/$ - see front matter. © 2017 Elsevier Inc. All rights reserved. http://dx.doi.org10.1016/j.jpeds.2017.04.037
1 FLA 5.4.0 DTD ■ YMPD9146_proof ■ May 16, 2017
THE JOURNAL OF PEDIATRICS • www.jpeds.com The purpose of this study was twofold: to examine the combined association of fatness and physical fitness components with academic achievement and to determine whether CRF and muscular strength are mediators of the association between fatness and academic achievement.
Methods The methodology of this study is described in detail elsewhere.16 In brief, this cross-sectional analysis is based on data drawn from a national sample of students (the Chilean Agency of Quality for Education [Agencia de Calidad de la Educación, Gobierno de Chile]), including all students enrolled in eighth grade (mean age, 13.83 years) who completed a standardized test, the System for Assessment of Educational Quality (SIMCE) and the SIMCE in physical education, administered by the Chilean Ministry of Education in November 2011, 2013, and 2014.17 The SIMCE test was not carried out in 2012. The sample was stratified by 15 regions in Chile (with the exception of Easter Island, the Juan Fernández archipelago, and the Antarctic) and by 3 school types (public, private subsidized, and private nonsubsidized). Within each stratum, schools were the primary sampling unit, and all students in the selected schools were sampled. A total of 50 549 students participated in the SIMCE in physical education. After exclusion of students because of erroneous data entry, disability, temporary illness or injury, chronic illness, absenteeism, age ≥18 years, missing body mass index (BMI) data, missing waist circumference (WC) data, or missing academic data, the present series comprised a total of 36 870 adolescents (72.94 % of the original sample) from 3 different cohorts and 669 schools. The SIMCE test for physical education consists of a fitness and anthropometric assessment using a standardized battery of fitness tests administered in November of each test year. The full protocol and battery of tests have been described previously.16 For the anthropometric measurements, the students wore light clothing and were barefoot. Data were recorded on paper by Ministry of Education evaluators.17 Weight was measured to the nearest 0.1 kg, and height was measured to the nearest 0.1 cm. BMI was calculated as body weight in kilograms divided by the square of height in meters. Weight status and increased adiposity were defined as a BMI above the age- and sex-specific thresholds of the International Obesity Task Force.18 WC was measured by placing a nonelastic tape measure midway between the lowest rib margin and the iliac crest and recorded to the nearest 0.1 cm. WC was classified using criterion-referenced, health-related cutpoints derived from de Ferranti et al19 because of the large size, age-specificity, and relatively generalizable ethnicity (ie, the Chilean population is roughly 59% non-Hispanic white) of the sample. WC and height were used to calculate waist-to-height ratio (WHtR); central obesity was defined as a WHtR ratio ≥0.5.20 Testing procedures were consistent with guidelines for schoolbased physical fitness assessment.21 At each school, a team of trained Ministry of Education evaluators (n = 5 each year) ad-
Volume ■■ ministered the tests with the same instruments in partnership with the physical education instructor. Tests were administered in the school gymnasium or on another available hard surface.17 Consistent with recommendations,22,23 our analysis was limited to health-related, valid, and reliable fieldbased tests, such as the 20-m shuttle run and standing long jump test. The 20-m shuttle run was used to assess CRF. The test was performed as described by Léger et al.24 Each participant ran in a straight line between 2 lines 20 m apart while keeping pace with prerecorded audio signals. The initial speed of 8.5 km/ hour was increased by 0.5 km/hour each minute. The test was completed when the participant failed to reach the end lines keeping pace with the audio signals on 2 consecutive occasions or when he or she stopped because of fatigue. Results were recorded to the nearest stage completed. The Léger equation was used to determine peak oxygen uptake (mL/kg/ minute) in each adolescent.24 Fit adolescents were defined using the age- and sex-specific cutoffs listed for the healthy fitness zone in the 2011 FITNESSGRAM.25 The standing long jump was used to assess muscular strength. In this test, the participant stood behind the starting line and was instructed to push off vigorously and jump as far as possible. The participant had to land with the feet together and remain upright. The test was repeated twice, and the longer distance was recorded to the nearest 0.1 cm. To classify muscular strength, we considered the risk group to be the age- and sex-specific 20th percentile of standing long jump for European adolescents,26 because the lower quintile in this sample had the strongest association with a poor cardiometabolic risk profile.27 All measurements in a subsample of 50 adolescents (28 girls, 22 boys; mean [SD] age, 13.7 [1.0] years; mean [SD] weight, 56.2 [10.4] kg; mean [SD] height, 1.60 [7.0] m; mean [SD] BMI, 20.9 [3.5]) were repeated by administering the tests again 1 week later with several physical education instructors from the last SIMCE (2015). Technical errors of measurement (TEMs) for weight, height, and ANOVA-based intraclass correlations (R) and corresponding 95% CIs were used to estimate test-retest reliability. For anthropometric measures, TEMs were 0.1 cm for height, 0.1 kg for weight, and 0.2 cm for WC. For physical fitness tests, R ranged from 0.89 (95% CI, 0.840.92) for the 20-m shuttle run to 0.96 (95% CI, 0.94-0.98) for the standing broad jump. The SIMCE measures national curricular objectives for language and mathematics as established by the Chilean Ministry of Education. The language test evaluates reading and writing ability. The mathematics test evaluates the ability to understand concepts and numerical operations, ability to use simple nonfractional algebraic expressions and apply them to solve problems using the correct approach, and ability to solve first-degree equations with a single unknown quantity. The test, administered by the Ministry of Education evaluators, comprises both multiple-choice items and open-ended questions, and is scored on a scale of 0-400. In the present study, scores were categorized according to the achievement levels established by the Ministry of Educa-
2
Garcia-Hermoso et al FLA 5.4.0 DTD ■ YMPD9146_proof ■ May 16, 2017
■■ 2017
ORIGINAL ARTICLES
tion.17 The minimum scores for demonstrating proficiency were 292 for language and 297 for mathematics. The National Physical Education Survey was authorized under the Chilean Law of Sport, number 19.712, article 5.17 The Ministry of Education solicited consent from the schools before testing and instructed each school to inform parents and students with a standardized letter about the nature and importance of the fitness tests, the assessment date and time, and recommended test preparation.17 The nature and purpose of the study were communicated to potential participants and their parents or guardians, with the explanation that data would be available to the Chilean education authorities in accordance with the Law of Data Protection (Law 21120/2016). The Ministry of Education requested that all adolescents and parents consent to be part of the sample; therefore, written informed consent was obtained from parents and subjects before participation in the study. Because this was a secondary analysis of deidentified data, the University Manuela Beltran (Colombia) Institutional Review Board considered the study exempt from review. The first author applied to the Ministry of Education and obtained permission to use the publicly available data for research and teaching learning purposes. Further details are available at http://www.simce.cl/. Socioeconomic status was defined by the type of school. The Chilean educational system is decentralized and consists of 3 types of schools: public, private subsidized, and private nonsubsidized. As might be expected, a family’s socioeconomic status is highly predictive of the type of school their children attend; that is, with very few exceptions, low socioeconomic status families tend to send their children to public
schools, middle socioeconomic status families tend to send their children to private subsidized schools, and high socioeconomic status families tend to send their children to private nonsubsidized schools.28 Statistical Analyses Results are presented as mean (SD) or relative frequency (%). Differences between the sexes were tested using the t test or the c2 test for unadjusted means or frequencies, respectively. The normality of the distribution of the variables was tested using both graphical (normal probability plot) and statistical (Kolmogorov-Smirnov test) procedures. Because of their skewed distribution, the following variables were log-transformed before analysis: BMI, WC, WHtR, CRF, and muscular strength. ANCOVA models were used to assess mean differences in academic achievement (language and mathematics) among the categories of BMI, WC, WHtR, and physical fitness components (CRF and muscular strength), controlling for age, sex, assessment year, and socioeconomic status (model 1), and also adjusting for CRF and muscular strength or BMI depending on the fixed factor (model 2). Pairwise post hoc hypotheses were tested using the Bonferroni correction for multiple comparisons. The ORs and 95% CIs of having a score of proficiency in language and mathematics by combined factors of fatness and physical fitness variables (ie, CRF-BMI, CRF-WC, CRFWHtR, muscular strength-BMI, muscular strength-WC, and muscular strength-WHtR) were calculated through multinomial logistic regression adjusted by age, sex, assessment year, and socioeconomic status. For the combined analysis, CRF and
Table I. Characteristics of the sample by sex Characteristics Physical characteristics Age, y, mean (SD) Weight, kg, mean (SD) Height, cm, mean (SD) Fatness parameters BMI, kg/m2, mean (SD) High fat, n (%) WC, cm, mean (SD) High fat, n (%) WHtR, mean (SD) High fat, n (%) Physical fitness CRF, stage, mean (SD) CRF, mL/kg/min, mean (SD)* Unfit, n (%) Muscular strength, cm, mean (SD)† Unfit, n (%) Academic achievement (SIMCE) Language, raw score, mean (SD) Score for proficiency, n (%) Mathematics, raw score, mean (SD) Score for proficiency, n (%) Socioeconomic status, n (%) Low Medium High
Total (n = 36 870)
Boys (n = 19 827)
Girls (n = 17 043)
P value for sex
13.8 (0.7) 57.0 (10.6) 161.0 (7.9)
13.9 (0.8) 58.3 (11.0) 164.3 (7.7)
13.8 (0.7) 55.4 (10.0) 157.0 (6.0)
<.001 <.001 <.001
21.9 4940 72.3 3503 0.44 4056
(3.6) (13.4) (8.9) (9.5) (0.05) (11.0)
21.5 2102 73.3 1983 0.44 2022
(3.5) (10.6) (8.8) (10.0) (0.05) (10.2)
22.5 2838 71.0 1520 0.45 2033
(3.7) (16.6) (8.8) (9.0) (0.06) (12.0)
<.001 <.001 <.001 .004 <.001 <.001
5.5 43.9 4498 151.4 6673
(2.5) (6.6) (12.2) (32.4) (18.1)
6.8 (2.3) 47.2 (6.0) 1864 (9.4) 168.0 (28.5) 3053(15.4)
4.0 39.9 2634 130.8 3620
(1.9) (5.1) (15.6) (24.0) (21.4)
<.001 <.001 <.001 <.001 <.001
252.0 8554 261.5 8812
(50.9) (23.2) (48.3) (23.9)
246.3 3965 263.3 4897
259.0 4589 259.4 3915
(49.1) (26.4) (49.0) (23.0)
<.001 <.001 <.001 .001
7636 (44.9) 8098 (47.6) 1269 (7.4)
<.001
16 518 (44.8) 17 218 (46.7) 3093 (8.39)
(51.7) (20.0) (47.8) (24.7)
8882 (44.8) 9120 (46.0) 1824 (9.2)
*Measured by the 20-minute shuttle run test. †Measured by the standing long jump test.
Cardiorespiratory Fitness and Muscular Strength as Mediators of the Influence of Fatness on Academic Achievement FLA 5.4.0 DTD ■ YMPD9146_proof ■ May 16, 2017
3
THE JOURNAL OF PEDIATRICS • www.jpeds.com Values are mean ± SD. Categories are L, M, and H, representing the first, second-third, and fourth quartiles, respectively. Categories of BMI are normal weight, overweight, and obese, according to sex- and age-specific cutoffs as defined by the International Obesity Task Force guidelines. Model 1: adjusted for age, sex, assessment year, and socioeconomic status. Model 2: model 1 covariates + CRF and muscular strength or BMI, depending on the fixed factor. *Bonferroni-adjusted pairwise comparisons.
L>H L>H L>H L>H L>H .009 <.001 <.001 <.001 <.001 L>M>H L > H; M > H L>M>H L>M>H L>M>H <.001 <.001 <.001 <.001 <.001 254.4 ± 46.1 255.1 ± 46.9 253.8 ± 46.6 266.7 ± 49.4 269.9 ± 50.4 260.6 ± 48.2 262.3 ± 48.7 261.0 ± 48.4 260.7 ± 47.5 259.5 ± 47.3 263.3 ± 48.9 265.1 ± 48.7 268.5 ± 48.9 250.7 ± 46.6 253.7 ± 46.5 — L>H L>H L>H L>H .071 <.001 <.001 <.001 <.001 L>M>H L > H; M > H L>M>H L>M>H L>M>H <.001 <.001 <.001 <.001 <.001 251.8 ± 50.5 252.8 ± 51.5 251.4 ± 51.4 251.2 ± 50.8 250.3 ± 50.4 BMI WC WHtR CRF Muscular strength
252.4 ± 51.7 254.9 ± 50.8 258.0 ± 50.8 242.4 ± 50.7 254.7 ± 49.9
247.7 ± 49.0 245.8 ± 49.6 245.6 ± 49.2 254.4 ± 51.0 258.3 ± 51.9
Post hoc* P value Medium (M) Categories
Low (L)
High (H)
P
Post hoc*
P value
Post hoc*
Low (L)
Medium (M)
High (H)
P value
Post hoc*
Model 2 Model 1
Mathematics (0-400)
Crude data Model 2
4
Model 1
Language (0-400)
Table I presents the demographic descriptive statistics of the sample. The final sample had a mean age of 13.8 ± 0.7 years and contained more boys than girls (55.2% vs 44.8%). Compared with the boys, the girls had lower WC, CRF, muscular strength, and mathematics scores, as well as a higher prevalence of obesity according to the International Obesity Task Force criteria and unfit profile (P < .001). Differences in academic achievement across fatness and physical fitness categories are shown in Table II. In model 1, overweight and obese adolescents had significantly lower language and mathematics scores compared with their normalweight peers (P < .001), but in model 2, after adjusting for CRF and muscular strength, these significant differences disappeared or were attenuated for language (P = .071) and mathematics (P = .009), respectively. Regarding WC and WHtR, adolescents in the medium and high quartiles had significantly lower scores in language and mathematics compared with their normal-weight peers in both models. Adolescents with high levels of CRF and muscular strength had significantly higher academic achievement scores compared with their peers with low and medium fitness levels in both models (P < .001). Table III shows the ORs of the relationship between high academic achievement in language and mathematics and the combined association of fatness (BMI, WC, and WHtR) and physical fitness (CRF and muscular strength). Adolescents classified in the fit and high fatness group and in the fit and low fatness group had higher odds of having high academic
Crude data
Results
Table II. Mean differences in academic achievement parameters by fatness and physical fitness categories controlling for potential confounders
muscular strength were categorized as “fit” or “unfit.”25 Each measure of adiposity (BMI, WC, and WHtR) was categorized as “high fatness” or “low fatness.” Four groups were created: unfit and high fatness, unfit and low fatness, fit and high fatness, and fit and low fatness. To examine whether the associations between measures of fatness or physical fitness and academic achievement were mediated by physical fitness or fatness, linear regression models were fitted using bootstrapped mediation procedures included in the PROCESS SPSS script.29 This script used bootstrapping methods30 for testing mediation hypotheses, using a resampling procedure of 10 000 bootstrap samples. The aim of the model was to investigate the total effects (equation 3) and direct effects (equations 1, 2, and 3’), reflected by the unstandardized regression coefficient and significance between the independent and dependent variables in each model. To establish mediation, equations 1-3 must be significant, and the unstandardized coefficients between the independent and dependent variable must be attenuated (partial mediation) or become (close to) null (full mediation) when the mediator is included in the regression model (equation 3’).31 The Sobel test was used to test the hypothesis that the indirect effect was equal to zero.32 This analysis was adjusted by age, sex, assessment year, and socioeconomic status. All the analyses were performed using SPSS 21 (IBM, Armonk, New York). The level of statistical significance was established as P < .05.
Volume ■■
Garcia-Hermoso et al FLA 5.4.0 DTD ■ YMPD9146_proof ■ May 16, 2017
■■ 2017
ORIGINAL ARTICLES
Table III. Combined effects of physical fitness and fatness parameters on score for proficiency in language and mathematics Parameters
CRF and BMI Unfit and low fat Fit and high fat Fit and low fat CRF and WC Unfit and low fat Fit and high fat Fit and low fat CRF and WHtR Unfit and low fat Fit and high fat Fit and low fat Muscular fitness and BMI Unfit and low fat Fit and high fat Fit and low fat Muscular fitness and WC Unfit and low fat Fit and high fat Fit and low fat Muscular fitness and WHtR Unfit and low fat Fit and high fat Fit and low fat
Language
Mathematics
OR
95% CI
P
OR
95% CI
P
1.055 1.187 1.349
0.905-1.231 1.024-1.377 1.184-1.538
.493 .023 <.001
1.287 1.444 1.728
1.090-1.521 1.229-1.695 1.497-1.995
.003 <.001 <.001
1.162 1.298 1.538
0.965-1.497 1.087-1.551 1.322-1.790
.067 .004 <.001
1.360 1.515 1.805
1.135-1.629 1.258-1.825 1.536-2.120
.001 <.001 <.001
1.102 1.214 1.498
0.920-1.416 1.035-1.425 1.303-1.721
.058 .017 <.001
1.366 1.419 1.812
1.150-1.623 1.191-1.691 1.559-2.107
<.001 <.001 <.001
1.159 1.261 1.396
0.975-1.323 1.096-1.451 1.238-1.573
.059 .001 <.001
1.289 1.437 1.717
1.119-1.486 1.273-1.669 1.509-1.953
<.001 <.001 <.001
1.149 1.169 1.425
0.973-1.358 1.012-1.351 1.247-1.628
.102 .034 <.001
1.288 1.422 1.719
1.104-1.502 1.195-1.691 1.491-1.981
.001 <.001 <.001
1.140 1.196 1.450
0.976-1.332 1.044-1.370 1.281-1.641
.099 .010 <.001
1.382 1.443 1.809
1.193-1.599 1.224-1.702 1.583-2.068
<.001 <.001 <.001
Reference group (OR, 1.0): adolescents classified as unfit and with high fatness. Analysis adjusted by age, sex, assessment year, and socioeconomic status.
achievement in both mathematics (all P < .001) and language (all P < .034) compared with unfit and high fatness adolescents. In contrast, compared with adolescents in the unfit and high fatness group, those in the unfit and low fatness group had higher odds of having high academic achievement in mathematics (all P < .003) but not in language (P = .058-.493). We performed a mediation analysis to test whether CRF and muscular strength acted as mediator variables between academic achievement (dependent variable) and fatness (independent variable) (Figures 1, 2, and 3; available at www.jpeds.com). As shown in Figure 1, the effect of BMI on language was fully mediated, but the effect of BMI on mathematics was only partially mediated, by CRF and muscular strength. In the first regression step (equation 1), BMI was negatively correlated with the physical fitness component (CRF and muscular strength) (P < .001). In the second step (equation 3), the regression coefficient of BMI on the dependent variable (language and mathematics) was negatively correlated (P < .001). In the last regression model, the mediator variable was positively correlated with the dependent variable (equation 2) (P < .001), but when both CRF and muscular strength were included in the model (equation 3’), the regression coefficient did not maintain its statistical significance (full mediation) in language, but did so in mathematics (partial mediation). Figures 2 and 3 show that the effect of WC and WHtR on academic achievement (both language and mathematics) was partially mediated by CRF and muscular strength.
Discussion Experimental evidence in neuroscience indicates that unhealthy body composition may hamper learning and aca-
demic performance.33 For example, gray matter reductions in the prefrontal cortex, a brain region involved in cognitive control, appear to occur in a dose-dependent manner with increasing BMI.33 Regarding physical fitness, CRF seems to be related to angiogenesis in the motor cortex and increased blood flow, resulting in improved brain vascularization, which could affect cognitive performance.34 In addition, CRF has been shown to play roles in spatial memory, hippocampal volume, and white matter integrity, whereas muscular strength has been related to synaptogenesis.35 Numerous studies have examined the associations of fitness and fatness individually with academic achievement; however, the findings are inconsistent.12 To date, only 1 study, of 1531 youths from 3 different cohorts has examined the combined association of CRF and weight status.3 That study showed that in students of good cardiorespiratory fitness and normal weight, the odds of high academic achievement were increased by 449% compared with unfit and overweight peers, suggesting that the combination of these components is a predictor of high academic achievement in language-related areas. Our findings confirm these results regardless of the physical fitness or fatness variables used. In mathematics, the odds of high academic achievement were significantly higher in adolescents classified as unfit and low fatness compared with those classified as unfit and high fatness. In addition, fit adolescents with either low or high fatness were more likely to have high academic achievement in both mathematics and language. Our results highlight the higher odds of WC or WHtR compared with BMI for predicting academic achievement. In this study, we also aimed to test the phenotype known as “fat but fit.”36 Our results suggest that the effect of fatness on academic achievement is fully mediated (BMI) or
Cardiorespiratory Fitness and Muscular Strength as Mediators of the Influence of Fatness on Academic Achievement FLA 5.4.0 DTD ■ YMPD9146_proof ■ May 16, 2017
5
THE JOURNAL OF PEDIATRICS • www.jpeds.com partially mediated (WC and WHtR) by CRF and muscular strength. BMI appears to be less independent of context to assess the relationship between fatness and academic achievement (ie, only for language); this measure is commonly used to define being overweight/obese, but is not a specific measure of fatness. The study reported by Van Dusen et al37 suggested a nonlinear association between BMI and academic achievement after covariate adjustment, indicating that physical fitness may be more important than weight control per se for academic development. According to Biddle and Asare,38 several explanations for the mediation role of physical fitness based on psychosocial aspects have been suggested: (1) the physical fitness of students is associated with better health, which can positively contribute to academic achievement; (2) increased physical fitness can improve the attention and behavior of adolescents in the classroom; and (3) greater physical fitness can improve mental health and self-esteem and can relieve stress, anxiety and depression, which in turn can positively affect academic achievement. The main limitation of the present study is its crosssectional design, which does not allow any inference of temporality or possible causal relationships. In addition, it is possible that the correlation among academic achievement, fatness, and physical fitness was mediated by variables not included in this study, such as cognitive outcomes, maturational stage, or other socioeconomic indicators (eg, maternal education). Finally, other tests to determine muscular strength are required, such as a handgrip test or an assessment of motor ability. EstebanCornejo et al39 showed that motor ability is more strongly related to academic achievement in youth compared with CRF. It may be true that muscular strength, as measured by the standing long jump, is a proxy for motor ability, which may explain the strength of its relationship with academic achievement. This study has several analytical strengths, including the mediation analysis and the large sample collected at a national level, including 3 different cohorts, which allowed us to extend previous findings to adolescents born in different years. Every effort should be made to sustain an effective physical education curriculum during the key period of adolescence, given that improving physical fitness can be a possible strategy for further improving the school performance of this population. The inclusion of a physical education assessment in the national education survey reflects an understanding of the importance of physical fitness and fatness for the current and future academic and health status of Chilean youth.40 ■ Submitted for publication Dec 9, 2016; last revision received Mar 31, 2017; accepted Apr 17, 2017 Reprint requests: Antonio García-Hermoso, PhD, Laboratorio de Ciencias de la Actividad Física, el Deporte y la Salud, Facultad de Ciencias Médicas, Universidad de Santiago de Chile, Avenida Libertador Bernardo, O’Higgins n° 3363, Estación Central, Santiago, Chile. E-mail: [email protected].
Volume ■■
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17. 18.
19.
20.
References
21.
1. Arora T, Hosseini-Araghi M, Bishop J, Yao G, Thomas GN, Taheri S. The complexity of obesity in UK adolescents: relationships with quantity and
22.
type of technology, sleep duration and quality, academic performance and aspiration. Pediatr Obes 2013;8:358-66. García-Hermoso A, Marina R. Relationship of weight status, physical activity and screen time with academic achievement in adolescents. Obes Res Clin Pract 2017;11:44-50. Sardinha LB, Marques A, Martins S, Palmeira A, Minderico C. Fitness, fatness, and academic performance in seventh-grade elementary school students. BMC Pediatr 2014;14:176. Esteban-Cornejo I, Tejero-González CM, Castro-Piñero J, CondeCaveda J, Cabanas-Sanchez V, Sallis JF, et al. Independent and combined influence of neonatal and current body composition on academic performance in youth: the UP & DOWN study. Pediatr Obes 2015;10:15764. Olivares PR, García-Rubio J. Associations between different components of fitness and fatness with academic performance in Chilean youths. PeerJ 2016;4:e2560. Morita N, Nakajima T, Okita K, Ishihara T, Sagawa M, Yamatsu K. Relationships among fitness, obesity, screen time and academic achievement in Japanese adolescents. Physiol Behav 2016;163:161-6. Suchert V, Hanewinkel R, Isensee B. Longitudinal relationships of fitness, physical activity, and weight status with academic achievement in adolescents. J Sch Health 2016;86:734-41. Datar A, Sturm R, Magnabosco JL. Childhood overweight and academic performance: national study of kindergartners and first-graders. Obes Res 2004;12:58-68. Aguilar MM, Vergara FA, Velásquez EJ, Marina R, García-Hermoso A. Screen time impairs the relationship between physical fitness and academic attainment in children. J Pediatr (Rio J) 2015;91:339-45. Torrijos-Niño C, Martínez-Vizcaíno V, Pardo-Guijarro MJ, GarcíaPrieto JC, Arias-Palencia NM, Sánchez-López M. Physical fitness, obesity, and academic achievement in schoolchildren. J Pediatr 2014;165:104-9. Ortega FB, Ruiz JR, Castillo MJ, Sjöström M. Physical fitness in childhood and adolescence: a powerful marker of health. Int J Obes (Lond) 2008;32:1-11. Donnelly JE, Hillman CH, Castelli D, Etnier JL, Lee S, Tomporowski P, et al. Physical activity, fitness, cognitive function, and academic achievement in children: a systematic review. Med Sci Sports Exerc 2016;48:12234. Chaddock L, Hillman CH, Pontifex MB, Johnson CR, Raine LB, Kramer AF. Childhood aerobic fitness predicts cognitive performance one year later. J Sports Sci 2012;30:421-30. Santana CC, Azevedo LB, Cattuzzo MT, Hill JO, Andrade LP, Prado WL. Physical fitness and academic performance in youth: a systematic review. Scand J Med Sci Sports 2016; doi:10.1111/sms.12773. [Epub ahead of print]. Sardinha LB, Marques A, Minderico C, Palmeira A, Martins S, Santos DA, et al. Longitudinal relationship between cardiorespiratory fitness and academic achievement. Med Sci Sports Exerc 2016;48:839-44. Garber MD, Sajuria M, Lobelo F. Geographical variation in healthrelated physical fitness and body composition among Chilean 8th graders: a nationally representative cross-sectional study. PLoS ONE 2014;9:e108053. Agencia de Calidad de la Educación, Gobierno de Chile. SIMCE. Available at: http://www.simce.cl/. Accessed October 1, 2016. Cole TJ, Bellizzi MC, Flegal KM, Dietz WH. Establishing a standard definition for child overweight and obesity worldwide: international survey. BMJ 2000;320:1240. de Ferranti SD, Gauvreau K, Ludwig DS, Neufeld EJ, Newburger JW, Rifai N. Prevalence of the metabolic syndrome in American adolescents: findings from the Third National Health and Nutrition Examination Survey. Circulation 2004;110:2494-7. Bauer KW, Marcus MD, El ghormli L, Ogden CL, Foster GD. Cardiometabolic risk screening among adolescents: understanding the utility of body mass index, waist circumference and waist to height ratio. Pediatr Obes 2015;10:329-37. Santos R, Mota J. The ALPHA health-related physical fitness test battery for children and adolescents. Nutr Hosp 2011;26:1199-200. Pate R, Oria M, Pillsbury L. Fitness measures and health outcomes in youth. Washington (DC): The National Academies Press; 2012.
6
Garcia-Hermoso et al FLA 5.4.0 DTD ■ YMPD9146_proof ■ May 16, 2017
■■ 2017
ORIGINAL ARTICLES
23. Ruiz JR, Castro-Piñero J, España-Romero V, Artero EG, Ortega FB, Cuenca MM, et al. Field-based fitness assessment in young people: the ALPHA health-related fitness test battery for children and adolescents. Br J Sports Med 2011;45:518-24. 24. Léger LA, Mercier D, Gadoury C, Lambert J. The multistage 20 metre shuttle run test for aerobic fitness. J Sports Sci 1988;6:93-101. 25. Welk GJ, Laurson KR, Eisenmann JC, Cureton KJ. Development of youth aerobic-capacity standards using receiver operating characteristic curves. Am J Prev Med 2011;41(4 Suppl 2):S111-6. 26. Ortega FB, Artero EG, Ruiz JR, España-Romero V, Jiménez-Pavón D, Vicente-Rodriguez G, et al. Physical fitness levels among European adolescents: the HELENA study. Br J Sports Med 2011;45:209. 27. Artero EG, Ruiz JR, Ortega FB, España-Romero V, Vicente-Rodríguez G, Molnar D, et al. Muscular and cardiorespiratory fitness are independently associated with metabolic risk in adolescents: the HELENA study. Pediatr Diabetes 2011;12:704-12. 28. Taut S, Cortéés F, Sebastian C, Preiss D. Evaluating school and parent reports of the national student achievement testing system (SIMCE) in Chile: access, comprehension, and use. Eval Program Plann 2009;32:12937. 29. Hayes AF. Introduction to mediation, moderation, and conditional process analysis: a regression-based approach. New York (NY): Guilford Press; 2013. 30. Preacher KJ, Hayes AF. Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behav Res Methods 2008;40:879-91.
31. Baron RM, Kenny DA. The moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations. J Pers Soc Psychol 1986;51:1173-82. 32. Sobel ME. Asymptotic confidence intervals for indirect effects in structural equation models. Sociol Methodol 1982;13:290-312. 33. Taki Y, Kinomura S, Sato K, Inoue K, Goto R, Okada K, et al. Relationship between body mass index and gray matter volume in 1,428 healthy individuals. Obesity (Silver Spring) 2008;16:119-24. 34. Hillman CH, Erickson KI, Kramer AF. Be smart, exercise your heart: exercise effects on brain and cognition. Nat Rev Neurosci 2008;9:58-65. 35. Adkins DL, Boychuk J, Remple MS, Kleim JA. Motor training induces experience-specific patterns of plasticity across motor cortex and spinal cord. J Appl Physiol (1985) 2006;101:1776-82. 36. McAuley PA, Blair SN. Obesity paradoxes. J Sports Sci 2011;29:773-82. 37. Van Dusen DP, Kelder SH, Kohl HW 3rd, Ranjit N, Perry CL. Associations of physical fitness and academic performance among schoolchildren. J Sch Health 2011;81:733-40. 38. Biddle SJ, Asare M. Physical activity and mental health in children and adolescents: a review of reviews. Br J Sports Med 2011;45:886-95. 39. Esteban-Cornejo I, Tejero-González CM, Martinez-Gomez D, del Campo J, González-Galo A, Padilla-Moledo C, et al. Independent and combined influence of the components of physical fitness on academic performance in youth. J Pediatr 2014;165:306-12.e2. 40. Weintraub WS, Daniels SR, Burke LE, Franklin BA, Goff DC Jr, Hayman LL, et al. Value of primordial and primary prevention for cardiovascular disease a policy statement from the American Heart Association. Circulation 2011;124:967-90.
Cardiorespiratory Fitness and Muscular Strength as Mediators of the Influence of Fatness on Academic Achievement FLA 5.4.0 DTD ■ YMPD9146_proof ■ May 16, 2017
7
THE JOURNAL OF PEDIATRICS • www.jpeds.com
Volume ■■
Figure 1. Physical fitness mediation models of the relationship between BMI and academic achievement score in adolescents, adjusted by age, sex, assessment year, socioeconomic status, and CRF or muscular fitness according to mediation variable included in the model. MS, muscular strength. *P < .05; **P < .001.
7.e1
Garcia-Hermoso et al FLA 5.4.0 DTD ■ YMPD9146_proof ■ May 16, 2017
■■ 2017
ORIGINAL ARTICLES
Figure 2. Physical fitness mediation models of the relationship between WC and academic achievement score in adolescents, adjusted by age, sex, assessment year, socioeconomic status, and CRF or muscular fitness according to mediation variable included in the model. *P < .05; **P < .001.
Cardiorespiratory Fitness and Muscular Strength as Mediators of the Influence of Fatness on Academic Achievement FLA 5.4.0 DTD ■ YMPD9146_proof ■ May 16, 2017
7.e2
THE JOURNAL OF PEDIATRICS • www.jpeds.com
Volume ■■
Figure 3. Physical fitness mediation models of the relationship between WHtR and academic achievement score in adolescents, adjusted by age, sex, assessment year, socioeconomic status, and CRF or muscular fitness according to mediation variable included in the model. *P < .05; **P < .001.
7.e3
Garcia-Hermoso et al FLA 5.4.0 DTD ■ YMPD9146_proof ■ May 16, 2017