Prospective physical fitness status and development of cardiometabolic risk in children according to body fat and lifestyle behaviours: The IDEFICS study

ORIGINAL RESEARCH Prospective physical fitness status and development of cardiometabolic risk in children according to body fat and lifestyle behaviours: The IDEFICS study Alba M. Santaliestra-Pasías 1,2,3,4 | Luis A. Moreno 1,2,3,4 | Luis Gracia-Marco 1,5 | Christoph Buck 6 | Wolfgang Ahrens 6,7 | Stefaan De Henauw 8 | Antje Hebestreit 6 | Yiannis Kourides 9 | Fabio Lauria 10 | Lauren Lissner 11 | Denes Molnar 12 | Toomas Veidebaum 13 | Esther M. Gonz alez-Gil 1,2,14 | on behalf the IDEFICS consortium 1 GENUD (Growth, Exercise, Nutrition and Development) Research Group, University of Zaragoza, Zaragoza, Spain 2 Centro de Investigaci on Biomédica en Red de Fisiopatología de la Obesidad y Nutrici on (CIBERObn), Instituto de Salud Carlos III, Madrid, Spain 3 Instituto Agroalimentario de Arag on (IA2), Zaragoza, Spain 4 Instituto de Investigaci on Sanitaria Arag on (IIS Arag on), Zaragoza, Spain 5 PROFITH (PROmoting FITness and Health through physical activity) Research Group, Department of Physical Education and Sports, Faculty of Sport Sciences, Sport and Health University Research Institute (iMUDS), University of Granada, Granada, Spain 6 Leibniz Institute for Prevention Research and Epidemiology –BIPS, Bremen, Germany 7 Institute of Statistics, Bremen University, Bremen, Germany 8 Department of Public Health and Primary Care, Faculty of Medicine and Health Sciences, Ghent University, Ghent, Belgium 9 Research and Education Institute of Child Health, Cyprus 10 Institute of Food Sciences, National Research Council, Avellino, Italy 11 Department of Public Health and Community Medicine, University of Gothenburg, Gothenburg, Sweden 12 Department of Pediatrics, Medical School, University of Pécs, Pécs, Hungary 13 National Institute for Health Development, Center of Health and Behavioral Science, Tallinn, Estonia 14 Department of Biochemistry and Molecular Biology II, Instituto de Nutrici on y Tecnología de los Alimentos, Center of Biomedical Research (CIBM), Universidad de Granada, Granada, Spain Correspondence Luis A. Moreno. GENUD (Growth, Exercise, Nutrition and Development) Research Group, University of Zaragoza, Zaragoza, Spain. Email: [email protected] Funding information Ministerio de Ciencia e Innovaci on, Grant/ Award Number: FJCI-2017-34967; Sixth Framework Programme, Grant/Award Number: 016181 FOOD Summary Background: Elevated cardiometabolic risk (CMR) is an important factor for cardiovascular diseases later in life while physical fitness seems to decrease CMR. Objective: Thus, the aim of the present study is to assess the association between muscular fitness (MF) and cardiorespiratory fitness (CRF) on CMR in European children, both cross-sectional and longitudinally. Methods: A total of 289 children (49.5% males) from eight European countries, aged 6 to 9, with longitudinal information on blood pressure, triglycerides, total cholesterol, HDL-cholesterol, homoeostasis model assessment, body mass index, data on fitness level, objectively measured physical activity (PA), diet quality, and total screen time Received: 28 November 2020 Revised: 24 March 2021 Accepted: 26 April 2021 DOI: 10.1111/ijpo.12819 This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made. © 2021 The Authors. Pediatric Obesity published by John Wiley & Sons Ltd on behalf of World Obesity Federation. Pediatric Obesity. 2021;e12819. wileyonlinelibrary.com/journal/ijpo 1of13 https://doi.org/10.1111/ijpo.12819 were included. A CMR score was calculated and dichotomized. MF and CRF were also dichotomized. Cross-sectional and longitudinal multilevel logistic regressions adjusting for lifestyle behaviours were performed. Results: Reaching a high level of MF during childhood as well as remaining in that level over-time were associated with an 82% and 62% lower probability of high CMR at follow-up, respectively. Also, children who became top CRF over time, showed a 77% lower probability (P< 0.05) of being in the highest CMR quartile at follow-up, independently of sociodemographic and lifestyle indicators. Conclusions: A high MF at early childhood and during childhood reduces the odds of having CMR. Same occurs with the improvement of CRF during childhood. These findings highlight the importance of enhancing fitness to avoid CMR already in children. KEYWORDS cardiometabolic, childhood, European, fitness, longitudinal 1|INTRODUCTION Cardiovascular disease (CVD) is currently the leading cause of death and health loss in adults 1 and has its origin in early life. 2 In addition, it has been suggested that childhood cardiovascular risk tracks into adulthood 3 and has been associated with several diseases even among individuals with normal-weight. 4 It has been suggested that clustering of CVD risk factors seem to be a better measure of cardiovascular health in children than single risk factors. 5 Some of these scores include relevant cardiometabolic risk (CMR) markers in the composite risk and have shown associations between physical activity and clustered cardiovascular risk. 6 However, fitness, instead of physical activity, has been considered a powerful marker of health and seems to play an important role in cardiometabolic health, even in children and adolescents. 7 In this regard, muscular fitness (MF) and cardiorespiratory fitness (CRF) have been considered key fitness components associated with cardiovascular risk factors. 7 In youth, MF has been associated with CMR 8 even after controlling for weight and height, body mass index (BMI), and body fat. 9 In adolescents, CRF and MF have been independently associated with metabolic risk in European adolescents. 10 It has been suggested that physical fitness in childhood and adolescence is a useful early predictor of CVD risk factors. 7 In childhood, a systematic review assessing the association between CRF and future cardiovascular risk factors found that CRF reported inverse associations with BMI, body fatness and metabolic syndrome. 11 In European children, it has also been observed an inverse association between fitness and CMR. 12 However, many of the previous studies did not account for body composition or lifestyle behaviours when investigating associations between physical fitness and CMR, such as dietary intake, 13 objectively measured physical activity, 14 or sedentary behaviours. 15 Nowadays, there is a lack of longitudinal studies using standardized and objective measures that help to understand the association between physical fitness and cardio metabolic health during childhood. Taking all this into consideration, the main aims of the present study are (1) to assess cross-sectional associations between MF and CRS levels and individual and grouped markers of CMR and (2) to analyse longitudinally the effect of a transition fitness change over a twoyear period on the individual and grouped markers of CMR in a sample of European children, taking into consideration body composition and lifestyle behaviours. 2|MATERIAL AND METHODS 2.1 |Study design The Identification and prevention of Dietaryand lifestyleinduced health EFects In Children and Infants (IDEFICS) study is a multi-centre population-based study performed in children from eight European countries: Belgium, Cyprus, Estonia, Germany, Hungary, Italy, Spain and Sweden, which included an intervention component. A community intervention was developed, including a control and intervention region per country, geographically apart; it included diet, physical activity and stress modules. Design and main procedures have been described in detail elsewhere. 16 Baseline (T0) measurements were performed between September 2007 and May 2008, and the followup (T1) measurements between September 2009 and May 2010, after 2 years. Ethics committees in each centre provided an authorization, and parents also provided written informed consent and children their oral assent. The study was performed according to the ethical guidelines of the Edinburgh revision of the 1964 Declaration of Helsinki (2000). 2.2 |Study sample At baseline (T0), the study included 16 229 children from 2 to 9 years and, at follow-up (T1), 11 038 children aged 4 to 11 years. The 2of13 SANTALIESTRA-PASÍAS ET AL. minimum age to perform the physical fitness tests was 6 years old so only those aged 6 and above were included in the analysis. A crosssectional sample of 6086 children meeting the age criteria, were measured either at T0 or T1. Children with a complete data set consisting of socio-demographic, cardiometabolic risk markers, body composition indicators, physical fitness and dietary and sedentary behaviours information in both time points were included in the current analysis (n =289, 49.7% males). Figure 1 summarizes the flow chart of the study population. 2.3 |Measurements Standardized procedures were used for the anthropometric measurements. 17 Height was measured with a stadiometer (SECA 225, Birmingham, UK), while weight and percentage of body fat were measured with a child-adapted Tanita BC 420 SMA. Sexand age-standardized body mass index z-scores (zBMI) according to Cole et al. 18 were calculated. Skinfolds thicknesses were measured twice with a Holtain caliper (Holtain Ltd., Croswell, UK) at the triceps, biceps, subscapular and suprailiac sites, and the mean was used for the analysis. Then, sum of the four skinfold thicknesses was calculated. Blood pressure was measured with an electronic sphygmomanometer (Welch Allyn 4200B-E2, Welch Allyn Inc., Skaneateles Falls, New York) in the right arm, with the child in sitting position. Two measurements were taken at 2 minutes intervals. Differences higher than 5% of magnitude lead to a third measurement. Means of replicate measurements were used in all analyses. The systolic blood pressure (SBP) will be included in the CMR score. Also, the highest parental education level according to the International Standard Classification of Education (ISCED) 19 was categorized. 2.4 |Physical fitness Physical fitness was measured following the ALPHA Health-Related Fitness Test Battery for Children and Adolescents. 20 The upper-body MF was assessed using the handgrip strength test through a dynamometer with an adjustable grip (TKK 5401 Grip D, Takey, Tokyo, Japan). Participants were instructed to squeeze continuously for ≥2 seconds with the elbow in full extension. The best score of the two attempts for each hand was chosen and averaged. Relative upper-body MF was expressed per kg of body mass (handgrip strength [kg/kg]). 21 The lower-body MF was assessed by the standing long jump test. Participants had to jump as far as forward possible. The distance reached FIGURE 1 Flow chart of the population involved in the current study from the IDEFICS study. Abbreviations: T0, at baseline; T1, at follow-up; HDL, high-density lipoprotein; HOMA, Homeostasis Model Assessment Index; FFQ, food frequency questionnaire; 24HDR, 24-hour dietary-recall SANTALIESTRA-PASÍAS ET AL.3of13 was taken from the take-off line and the heel of the nearest foot at landing. The longest attempt out of two was chosen. Based on these two fitness tests, a MF score (MF z-score) was computed by combining upper-body and lower-body results. Each of these variables was standardized as follows: z-score =(ith value  mean)/SD. The MF z-score was based on previous studies 21 and calculated as the mean of the two standardized scores (handgrip strength and standing long jump). The CRF level was assessed using the 20 m' shuttle run test, which estimates the aerobic capacity. The results of all the centres were unified according to the Leger test protocol. 22,23 The number of shuttles was used as an indicator of the cardiorespiratory level with a greater number of shuttles indicating better performance. Both indicators, MF and CRF, were dichotomized. Thus, the first group included those children in the first, second or third quartile (Q1-Q3); and the group II included those children at the top quartile (Q4). Additionally, combinations of grouping between surveys were created and the cumulative fitness score (MF +CRF), including the MF and CRF as continuous variables, was calculated at T0, T1, and the delta values of this score, that is, differences between T1 and T0. 2.5 |Biological samples Children were asked to participate in fasting blood collection, on a voluntary basis in the study. 24 Blood sampling was performed after an overnight-fast. Blood glucose, total cholesterol (TC), high-density lipoprotein cholesterol (HDL-c) and triglycerides (TG) were assessed at each study centre by point-of-care analysis (Cholestech LDX analyzer, Cholestech Corp., Hayward, California). Serum insulin concentrations were determined by luminescence immunoassay Immulite 2000 (Siemens, Eschborn, Germany) in a central laboratory. Insulin resistance was defined by the homoeostasis model assessment (HOMA) 25 and calculated from fasting glucose and plasma insulin via a standard equation: HOMA =(insulin (μIU/mL)  glucose (mg/dL))/405. The TC/HDL-c ratio was computed. 2.6 |Cardio metabolic risk 2.6.1 |CMR scores A continuous score of clustered CMR factors was computed according to Andersen et al. 6 : SBP, TG, ratio TC/HDL-c, HOMA and sum of four skinfolds. Ageand sex-specific z-scores were calculated for each risk factor. All individual z-scores were added up to create the clustered CMR score. The lower the score the lower the overall cardiovascular risk. Also, two additional CMR scores were developed including two body composition indices, WC or FMI, instead of sum of four skinfolds, resulting in three CMR scores: (1) CMR score with sum of skinfolds, (2) CMR score with WC, and (3) CMR score with FMI. The WC and FMI were standardized with and ageand sexspecific z-scores for subsequent inclusion in the corresponding clustered CMR score. 2.6.2 |CMR categories For the individual risk factors, the population for each crude indicator was allocated into two groups. The first group included those children in the first, second or third quartile (Q1-Q3) of each individual crude indicator; and the group II included those children at the top quartile (Q4). For the CMR scores that included the sum of the individual risk z-scores, participants were allocated in two groups, taking into consideration the cut-off proposed by Andersen of 1 SD. 6 2.6.3 |Physical activity Physical activity (PA) was objectively measured using a uniaxial accelerometer (ActiTrainer or GT1M Actigraph; ActiGraph, LLC, Pensacola, Florida). Children wore the accelerometer for up to seven consecutive days. 26 Only those with at least 3 days' worth of valid accelerometer data, with at least 8 hours of valid data were included. The average time spent in moderate-to-vigorous PA (MVPA) was calculated using the Evenson cut-offs 27 and used as marker of PA. 2.7 |Dietary intake The food frequency questionnaire (FFQ), 28 which was part of the Children's Eating Habits Questionnaire in the IDEFICS study was used to derive the diet quality index (DQI). A computer-assisted 24-HDR, called SACINA ('Self-Administered Children and Infant Nutrition Assessment') was used in T0 and T1. 29 As we used a qualitative FFQ, sex-, age-, and country-specific medians of the portion sizes of the corresponding food groups were derived based on the 24-hours dietary recall (24-HDR) and the obtained information was used to derive the DQI from the FFQ. The DQI, adapted for children, 30 was used as a proxy indicator for the overall children's diet. 2.8 |Total screen time Total screen time was derived from the parent-reported questionnaire. The questions included the time spent watching TV, videos and DVDs, and the time using a computer and/ or playing videogames on a weekday and a weekend day separately. 31 The average total screen time in hours per week was calculated. 2.9 |Statistical analysis Descriptive study characteristics are shown as mean and SD for continuous variables and number of cases and percentages for categorical variables. For the cross-sectional analysis, a multilevel logistic regression analysis (levels: country and intervention vs control area) was 4of13 SANTALIESTRA-PASÍAS ET AL. performed using the individual risk factors and CMR score at both time points as dependent variable. This analysis was performed to assess the odds for having a higher individual and CMR status when participants were in the top level of fitness (Group II, top fit, Q4), compared with those who were in the low-medium level of fitness (Group I, low-medium fit, Q1-Q3), of fitness which were considered as the reference group. The cross-sectional analysis between fitness and individual risks and CMR score was performed using three models adjusting for potential covariates and levels (country and intervention vs control region). Model 0 was not adjusted. Model 1 included sex, age, zBMI and parental education level as covariates. Model 2 included covariates of Model 1 and also MVPA, DQI, total screen time as covariates. The models for the individual body composition indicators (sum of skinfolds, WC or FMI) and also the CMR did not include the zBMI as a covariate in order to avoid over-adjustment. To analyse the longitudinal effect association between fitness and individual risk and CMR scores, several analyses were performed. A multilevel logistic regression analysis (levels: country and intervention vs control area) was performed using the individual risk and CMR scores at T1 as dependent variable to assess the odds for having a higher cardiometabolic risk status when participants presented a specific fitness level (MF or CRF) at T0 and T1. Four transition groups of fitness were created (Figure 2): Group I, unchanged fitness at lowmedium level over time (remain low-medium fit), which included children being in the low-medium fitness level (Q1-Q3) of MF or CRF at T0 and T1; Group II, decreased fitness over time (became low-medium fit), which included those children being in the highest quartile of the MF or CRF (Q4) at T0, and being in the low-medium fitness level (Q1-Q3) at T1; Group III, improved fitness over time (became top fit), which included those children being in the low-medium fitness level (Q1-Q3) of MF or CRF at T0 and being in the highest quartile at T1 (Q4); and Group IV, unchanged fitness at top level over time (remained top fit), which included children being in the highest fitness MF or cardiorespiratory fitness (Q4) at both time points (T0&T1). The longitudinal multilevel logistic regression analysis was applied using three models, adjusting for potential covariates and levels (country and intervention vs control region). Model 0 was the nonadjusted model, which included the corresponding baseline individual or CMR score category. Model 1 included also sex, age, zBMI at T1 and education level. Model 2 included, MVPA, DQI, total screen, also at follow-up (T1). The models for the sum of skinfolds and also the CMR did not include the Z BMI as a covariate in order to avoid an over-adjustment. Sensitivity analysis was applied between included and excluded participants in order to check differences in some of the common measurements. Included participants were older and having high SES than the excluded ones (P< 0.05). However, no differences were observed in terms of BMI categories according to Cole et al. 18 The analysis was performed using the Statistical Package for the Social Sciences (version 21.0, SPSS) and Stata (version 13.0) for the multilevel logistic regression. 3|RESULTS Table 1 shows the main characteristic of the study participants at baseline (T0) and follow-up (T1) by sex. Odds ratio (OR) and 95% confidence interval (CI) for the crosssectional associations between fitness MF and individual and CMR score categories are shown in Table 2. In T0, children at top MF levels had 85% and 50% lower probabilities of being allocated in the upper category of sum of skinfolds and CMR, respectively, compared with children at low-medium MF levels, after controlling for education and lifestyle behaviours (MVPA, DQI and total screen time). In T1, children at top MF levels had 47%, 82% and 75% lower probability of being allocated in the upper category of the ratio total cholesterol/HDL, sum of skinfolds and CMR, respectively, compared with the children with low-medium MF levels, after controlling for education and lifestyle behaviours. Table 3 shows OR and 95% CI for the cross-sectional associations between CRF fitness level groups, and individual and CMR score categories. In both T0 and T1, top fit children had 60% or 64% lower FIGURE 2 Muscular fitness and Cardiovascular fitness grouping design between baseline and follow-up. MF or CRF transition groups over time: Group I, unchanged fitness at low-medium level over time (remain low-medium fit) (N =188 for MF and N =184 for CRF); Group II, decreased fitness over time (became lowmedium fit) (N =30 for MF, and N =38 for CRF); Group III, improved fitness over time (became top fit) (N =30 for MF, and N =37 for CRF); Group IV, unchanged fitness at top level over time (remain top fit) (N =41 for MF, and N=30for CRF) SANTALIESTRA-PASÍAS ET AL.5of13 TABLE 1 Description of included study population and the continuous cardiometabolic factors by gender at baseline (T0) and follow-up (T1) T0 T1 p 2 Male n =143 Female n =146 p 1 Male n =143 Female n =146 p 1 Age (x¯±SD) 7.6 (0.7) 7.5 (0.6) 0.263 9.6 (0.7) 9.5 (0.6) 0.176 <0.001 Parental education n(%) Low 3 (2.1) 5 (3.4) 0.778 4 (2.8) 2 (1.4) 0.696 <0.001 Medium 65 (45.8) 65 (44.5) 66 (46.2) 68 (46.6) High 74 (52.1) 76 (52.1) 73 (51.0) 76 (52.1) Region n(%) Intervention 73 (51.0) 71 (48.6) 0.681 73 (51.0) 71 (48.6) 0.681 <0.001 Control 70 (49.0) 75 (51.4) 70 (49.0) 75 (51.4) Systolic Blood Pressure, mm Hg (x¯±SD) 105.3 (8.8) 104.4 (8.6) 0.366 108.6 (9.9) 107.7 (9.7) 0.460 <0.001 Triglycerides, mg/dL (x¯±SD) 50.4 (13.3) 53.9 (23.7) 0.130 54.4 (21.7) 57.6 (24.9) 0.234 <0.001 Ratio total cholesterol/HDL, (x¯±SD) 2.88 (1.0) 3.23 (1.0) 0.003 3.0 (1.2) 3.3 (1.3) 0.051 0.159 Homeostatic Model Assessment index, (x¯±SD) 0.93 (0.6) 1.01 (0.7) 0.233 1.49 (0.8) 1.61 (1.2) 0.287 <0.001 Sum of skinfolds, mm (x¯±SD) 31.1 (16.1) 36.0 (15.2) 0.009 36.9 (20.5) 42.2 (18.6) 0.023 <0.001 ZBMI (x¯±SD) 0.353 (1.05) 0.500 (1.01) 0.224 0.402 (1.10) 0.468 (0.99) 0.595 <0.001 Cardiometabolic risk score (x¯±SD) 0.03(2.9) 0.00 (3.4) 0.941 0.02 (3.3) 0.00 (3.5) 0.953 0.990 Cardiometabolic risk score categories a (n,%) Category I 100 (69.9) 104 (71.2) 0.808 101 (70.6) 105 (71.9) 0.809 <0.001 Category II 43 (30.1) 42 (28.8) 42 (29.4) 41 (28.1) MF z-score (x¯±SD) 0.01 (1.69) 0.00 (1.61) 0.957 0.007 (0.99) 0.00 (0.99) 0.955 0.978 CRF, number of shuttles (x¯±SD) 21.0 (10.9) 16.6 (8.7) <0.001 29.2 (14.0) 22.9 (11.3) <0.001 <0.001 MVPA, minutes (x¯±SD) 51.2 (22.9) 40.3 (20.2) <0.001 49.8 (22.9) 37.8 (19.0) <0.001 0.149 DQI (x¯±SD) 90.2 (18.3) 90.4 (13.7) 0.924 89.5 (15.2) 90.6 (12.9) 0.501 0.793 Total screen time, hours per week(x¯±SD) 13.8 (6.8) 11.6 (6.0) 0.004 15.0 (7.3) 12.8 (7.1) 0.008 0.001 Abbreviations: SD, SD; zBMI, body mass index z score by Cole et al.; MF, muscular fitness score; CRF, Cardiorespiratory fitness; MVPA, Moderate to vigorous physical activity; DQI, diet quality index. Notes: p 1 gender differences; p 2 time-point differences. Bold letters indicate p < 0.05 between gender. a Cardiometabolic Risk Score categories: Category I, < 1SD; Category II, ≥1 SD. Based on Andersen et al. 5 6of13 SANTALIESTRA-PASÍAS ET AL. probability of being allocated in the upper category of the sum of skinfolds, respectively, compared with low CRF, after controlling for all the potential confounders. Also, in T0, the top fit children had 69% lower probability of being allocated in the upper category of CMR score after controlling for all the confounders, and in T1, the significant effect disappeared after included the lifestyle behaviours as covariates in the model. Taking into consideration the transition fitness groups over time, the highest proportion of children was those who remained in the low-medium MF or CRF level (Group I), being respectively the 65% and 63.7% of total sample. A low proportion of children become top fit(10.4%forMFand12.8%forCRS)orbecomelowfitoveratwoyear period (10.4% for MF, and 13.1% for CRF). Finally, a 14.2% of the sample for MF, and 10.4% for CRF remained in the top fit level. Table 4 shows separately the OR and 95%CI for the longitudinal associations between the individual and CMR categories, and the transition MF and CRF group over time. The strongest associations were observed for those become top MF over time (group III), they showed a 71% and 82% lower probability of being in the highest sum of skinfolds and CMR score categories, respectively, than those who remained low-medium fit. Also, those children who remained in TABLE 2 Cross-sectional multilevel logistic regression between grouping of muscular fitness (MF) and individual indicators and cardiometabolic risk score at baseline (T0) and follow up (T1) Predictor and outcomes MF groups a At baseline (T0) At follow-up (T1) OR 95% CI OR 95% CI Systolic blood pressure b Model 0 Group I, lowmedium fit (Ref) 11 Group II, top fit 0.77 0.44;1.35 0.58 0.33;0.99 Model 1 Group I, lowmedium fit (Ref) 11 Group II, top fit 1.19 0.64;2.22 0.99 0.54;1.79 Model 2 Group I, lowmedium fit (Ref) 11 Group II, top fit 1.18 0.63;2.24 0.96 0.52;1.76 Triglycerides b Model 0 Group I, lowmedium fit (Ref) 11 Group II, top fit 0.78 0.43;1.42 1.13 0.66;1.94 Model 1 Group I, lowmedium fit (Ref) 11 Group II, top fit 0.99 0.53;1.86 1.40 0.80;2.49 Model 2 Group I, lowmedium fit (Ref) 11 Group II, top fit 1.06 0.55;2.06 1.38 0.77;2.50 Ratio total cholesterol/ high density lipoprotein b Model 0 Group I, lowmedium fit (Ref) 11 Group II, top fit 0.52 0.30;0.90 0.41 0.23;0.72 Model 1 Group I, lowmedium fit (Ref) 11 Group II, top fit 0.61 0.34;1.09 0.49 0.27;0.89 Model 2 Group I, lowmedium fit (Ref) 11 Group II, top fit 0.71 0.39;1.28 0.53 0.29;0.96 Homeostatic model assessment b Model 0 Group I, lowmedium fit (Ref) 11 Group II, top fit 0.67 0.38;1.20 0.73 0.43;1.25 Model 1 Group I, lowmedium fit (Ref) 11 Group II, top fit 1.17 0.61;2.23 1.07 0.60;1.89 Model 2 Group I, lowmedium fit (Ref) 11 Group II, top fit 1.36 0.70;2.67 1.06 0.59;1.90 Sum of skinfolds b Model 0 Group I, lowmedium fit (Ref) 11 Group II, top fit 0.14 0.07;0.30 0.18 0.10;0.33 Model 1 Group I, lowmedium fit (Ref) 11 Group II, top fit 0.14 0.06;0.29 0.17 0.09;0.31 Model 2 Group I, lowmedium fit (Ref) 11 (Continues) TABLE 2 (Continued) Predictor and outcomes MF groups a At baseline (T0) At follow-up (T1) OR 95% CI OR 95% CI Group II, top fit 0.15 0.07;0.32 0.18 0.10;0.33 Cardiometabolic risk score c Model 0 Group I, lowmedium fit (Ref ) 11 Group II, top fit 0.45 0.23;0.87 0.22 0.11;0.52 Model 1 Group I, lowmedium fit (Ref) 11 Group II, top fit 0.44 0.23;0.86 0.24 0.11;0.52 Model 2 Group I, lowmedium fit (Ref) 11 Group II, top fit 0.50 0.25;0.98 0.25 0.11;0.56 Abbreviations: MF, muscular fitness; OR, odds ratio; CI, confident interval. Notes: All models of the multilevel logistic regression included random effects (country, study region [intervention or control]). Multilevel logistic regression analysis between each MF groups and systolic blood pressure (SBP), triglycerides (TRG), ratio total cholesterol/high-density lipoprotein (TC/HDL), homeostatic model assessment (HOMA), sum of skinfolds (Skinfolds) or cardiometabolic risk score (CMR score). Model 0, non-adjusted; Model 1, adjusted by sex, age, parental education and zBMI (except for the Sum of skinfolds and CRM Score) at baseline (T0) or follow-up (T1); Model 2, Model 1 adjusted +Moderate to vigorous physical activity, diet quality index and total screen time at baseline (T0) or follow-up (T1), respectively. * Odds of being allocated to the highest SBP, TGR, TC/HDL, HOMA, Skinfolds, CMR score category. Bold letters indicates P< 0.05. a MF groups: Group I, low-medium MF quartiles (Q1-Q3) at T0 or T1; Group II, top MR quartile (Q4) at T0 or T1. b SBP, TGR, TC/HDL, HOMA, Skinfolds categories: Category I, first and second tertile; Category II, third tertile. c CMR score categories: Category I, <1SD; Category II, ≥1SD. SANTALIESTRA-PASÍAS ET AL.7of13 the top MF group over time (group IV) had 69% lower probabilities of being in the highest sum of skinfolds category after controlling for the potential confounders. In the same vein, those in the top fit group over-time (group IV) showed also lower probabilities of a high CMR score after controlling for all covariates (OR =0.38; CI: 0.13-1.08). Those in the remain top CRF group (group IV) had lower odds of being in the high group of sum of skinfolds in comparison with those who remain in the low-medium fit CRF group over-time in the fully adjusted model (OR =0.33; CI:0.12;0.91). Those children who became top fit (group III) had a 74% less odds of being in the highest sum of skinfolds quartile. Finally, those became low-medium CRF over-time (group II), had a 77% less odds of being in the highest CMR quartile in comparison to those who maintained a low-medium fit CRF group over-time in the fully adjusted level. 4|DISCUSSION In the present study, the impact of physical fitness as predictor of CMR was investigated, both cross-sectional and longitudinally, finding TABLE 3 Cross-sectional multilevel logistic regression between grouping of cardiorespiratory fitness (CRF) and individual indicators and cardiometabolic risk score at baseline (T0) and follow up (T1) Predictor and outcomes CRF groups a At baseline (T0) At follow-up (T1) OR 95% CI OR 95% CI Systolic blood pressure b Model 0 Group I, lowmedium fit (Ref) 11 Group II, top fit 1.61 0.93;2.79 0.55 0.32;0.96 Model 1 Group I, lowmedium fit (Ref) 11 Group II, top fit 2.05 1.11;3.80 0.80 0.44;1.46 Model 2 Group I, lowmedium fit (Ref) 11 Group II, top fit 2.14 1.12;4.08 0.71 0.38;1.33 Triglycerides b Model 0 Group I, lowmedium fit (Ref) 11 Group II, top fit 0.53 0.28;1.02 0.90 0.52;1.57 Model 1 Group I, lowmedium fit (Ref) 11 Group II, top fit 0.55 0.28;1.10 1.06 0.59;1.90 Model 2 Group I, lowmedium fit (Ref) 11 Group II, top fit 0.66 0.32;1.34 1.25 0.68;2.30 Ratio total cholesterol/ high-density lipoprotein b Model 0 Group I, lowmedium fit (Ref) 11 Group II, top fit 0.66 0.39;1.14 0.82 0.48;1.39 Model 1 Group I, lowmedium fit (Ref) 11 Group II, top fit 0.81 0.46;1.43 1.08 0.62;1.92 Model 2 Group I, lowmedium fit (Ref) 11 Group II, top fit 0.86 0.47;1.56 1.05 0.58;1.90 Homeostatic model assessment b Model 0 Group I, lowmedium fit (Ref) 11 Group II, top fit 0.55 0.30;1.02 0.64 0.37;1.09 Model 1 Group I, lowmedium fit (Ref) 11 Group II, top fit 0.60 0.30;1.20 0.93 0.52;1.66 Model 2 Group I, lowmedium fit (Ref) 11 Group II, top fit 0.65 0.32;1.34 0.98 0.53;1.81 Sum of skinfolds b Model 0 Group I, lowmedium fit (Ref) 11 Group II, top fit 0.44 0.24;0.79 0.35 0.20;0.61 Model 1 Group I, lowmedium fit (Ref) 11 Group II, top fit 0.34 0.18;0.64 0.33 0.19;0.57 Model 2 Group I, lowmedium fit (Ref) 11 TABLE 3 (Continued) Predictor and outcomes CRF groups a At baseline (T0) At follow-up (T1) OR 95% CI OR 95% CI Group II, top fit 0.40 0.21;0.76 0.36 0.20;0.64 Cardiometabolic risk score c Model 0 Group I, lowmedium fit (Ref) 11 Group II, top fit 0.30 0.14;0.64 0.44 0.22;0.88 Model 1 Group I, lowmedium fit (Ref) 11 Group II, top fit 0.26 0.12;0.57 0.44 0.22;0.88 Model 2 Group I, lowmedium fit (Ref) 11 Group II, top fit 0.31 0.14;0.68 0.55 0.27;1.12 Abbreviations: CRF, cardiorespiratory fitness; OR, odds ratio; CI, confident interval. Notes: All models of the multilevel logistic regression included random effects (country, study region [intervention or control]). Multilevel logistic regression analysis between each MF groups and systolic blood pressure (SBP), triglycerides (TRG), ratio total cholesterol/high-density lipoprotein (TC/HDL), homeostatic model assessment (HOMA), sum of skinfolds (Skinfolds) or cardiometabolic risk score (CMR Score). Model 0, non-adjusted; Model 1, adjusted by sex, age, parental education and zBMI (except for the Sum of skinfolds and CRM Score) at baseline (T0) or follow-up (T1); Model 2, Model 1 adjusted +Moderate to vigorous physical activity, diet quality index and total screen time at baseline (T0) or follow-up (T1), respectively. * Odds of being allocated to the highest SBP, TGR, TC/HDL, HOMA, Skinfolds, CMR score category. Bold letters indicate p < 0.05. a CRF groups: Group I, low-medium CRF quartiles (Q1-Q3) at T0 or T1; Group II, top CRF quartile (Q4) at T0 or T1. b SBP, TGR, TC/HDL, HOMA, Skinfolds categories: Category I, first and second tertile; Category II, third tertile. c CMR score categories: Category I, <1SD; Category II, ≥1SD. 8of13 SANTALIESTRA-PASÍAS ET AL. TABLE 4 Longitudinal multilevel logistic regression between grouping of muscular fitness (MF) or cardiorespiratory fitness (CRF) and individual indicators and cardiometabolic risk (CMR) score at follow up (T1) Models for the MF a Groups Models for the CRF a Groups Model 0 Model 1 Model 2 Model 0 Model 1 Model 2 OR 95% CI OR 95% CI OR 95% CI OR 95% CI OR 95% CI OR 95% CI Systolic blood pressure GroupI,remainlow-medium fit (Ref ) 111111 Group II, became low-medium fit 0.74 0.32;1.71 1.23 0.49;3.22 1.35 0.53;3.43 0.91 0.43;1.94 1.26 0.55;2.86 1.28 0.56;2.92 Group III, became top fit 0.33 0.13;0.84 0.52 0.20;1.40 0.52 0.19;1.38 0.40 0.18;0.91 0.56 0.24;1.34 0.51 0.21;1.24 Group IV, remain top fit 0.71 0.34;1.47 1.14 0.52;2.47 1.21 0.55;2.66 0.63 0.27;1.45 0.96 0.39;2.37 0.86 0.34;2.18 Triglycerides Group I, remain low-medium fit (Ref) 111111 Group II, became low-medium fit 0.54 0.23;1.28 0.67 0.27;1.65 0.65 0.26;1.61 1.08 0.51;2.27 1.11 0.50;2.42 1.08 0.49;2.42 Group III, became top fit 1.36 0.61;2.99 1.64 0.72;3.73 1.65 0.72;3.82 1.04 051;2.15 1.21 0.57;2.57 1.42 0.65;3.12 Group IV, remain top fit 1.11 0.55;2.21 1.26 0.61;2.62 1.22 0.58;2.58 0.90 0.40;2.02 1.00 0.43;2.33 1.18 0.50;2.79 Ratio total cholesterol/ high-density lipoprotein Group I, remain low-medium fit (Ref) 111111 Group II, became low-medium fit 1.02 0.44;2.43 1.27 0.53;3.06 1.42 0.58;3.43 0.60 0.26;1.36 0.87 0.37;2.09 0.88 0.37;2.13 Group III, became top fit 0.53 0.22;1.25 0.64 0.26;1.58 0.66 0.27;1.63 0.94 0.42;2.10 1.23 0.53;2.83 1.19 0.50;2.82 Group IV, remain top fit 0.37 0.16;0.85 0.38 0.16;0.94 0.42 0.17;1.05 1.74 0.78;3.87 2.44 1.05;5.70 1.17 0.90;5.27 Homeostatic model assessment Group I, remain low-medium fit (Ref) 111111 Group II, became low-medium fit 0.32 0.14;0.70 0.44 0.19;1.01 0.46 0.20;1.08 0.80 0.39;1.68 1.20 0.54;2.66 1.17 0.52;2.61 Group III, became top fit 0.60 0.27;1.32 0.80 0.35;1.83 0.78 0.33;1.80 0.71 0.35;1.44 0.99 0.47;2.09 0.98 0.45;2.13 Group IV, remain top fit 0.81 0.40;1.62 1.09 0.53;2.26 1.09 0.52;2.30 0.73 0.33;1.60 1.11 0.48;2.59 1.24 0.51;3.00 Sum of Skinfolds b Group I, remain low-medium fit (Ref) 111111 Group II, became low-medium fit 0.47 0.18;1.21 0.48 0.18;1.25 0.48 0.18;1.27 0.25 0.09;0.70 0.29 0.10;0.84 0.27 0.09;0.80 Group III, became top fit 0.34 0.13;0.89 0.31 0.11;0.85 0.29 0.10;0.81 0.31 0.12;0.80 0.31 0.12;0.83 0.26 0.09;0.71 Group IV, remain top fit 0.32 0.13;0.80 0.32 0.12;0.83 0.31 0.12;0.82 0.41 0.16;1.08 0.38 0.14;1.03 0.33 0.12;0.91 (Continues) SANTALIESTRA-PASÍAS ET AL.9of13