The association between physical activity, physical fitness and ...

12 downloads 4523 Views 93KB Size Report
Jun 17, 2011 - 1Center for Research in Childhood Health, Institute of Sport Sciences ... association between composite risk factor score and physical fitness or ...
International Journal of Pediatric Obesity, 2011; 6(S1): 29–34

ORIGINAL ARTICLE

The association between physical activity, physical fitness and development of metabolic disorders

LARS B. ANDERSEN1,2, ANNA BUGGE1, MAGNUS DENCKER1,3, STIG EIBERG4 & BIANCA EL-NAAMAN1 1Center

for Research in Childhood Health, Institute of Sport Sciences and Clinical Biomechanics, University of Southern Denmark, 2Department of Sports Medicine, Norwegian School of Sport Science, 3Department of Clinical Sciences, Unit of Clinical Physiology and Nuclear Medicine, Malmö University Hospital, Lund University, Malmö, Sweden, 4National Institute of Public Health, University of Southern Denmark, Denmark Abstract Background. Cardiovascular (CVD) risk factors have been shown to cluster in some children. This has been shown in children from the age of nine years, but recently we found no clustering in six-year old children. It is uncertain when clustering develops and which parameters are related to the development of clustered CVD risk. Methods. A longitudinal study including 484 children aged six years. Three years later, 434 children participated in a follow-up. The main outcome was clustering of five CVD risk factors: homeostasis assessment insulin resistance (HOMA), total cholesterol:HDL ratio, triglyceride (TG), systolic blood pressure and sum of four skinfolds. Independent variables were physical activity and cardiorespiratory fitness. Results. CVD risk factors were independently distributed in the six-year-olds, and there was no association between composite risk factor score and physical fitness or activity even if there were obese and unfit children in the population. Clustering of CVD risk factors was found at the age of nine years, and the observed number with three or more CVD risk factors was 3.33 (95% CI: 1.41–7.87) times higher than expected if risk factors had been independently distributed. At the age of nine years, the lowest quartile of fitness had 34.9 (95% CI: 8.0–152.5) times higher risk of having clustered risk than the most fit quartile. Conclusion. Clustering of CVD risk factors developed between the age of six and nine years. At nine years of age clustered CVD risk was highly associated with low fitness level. Key words: Children, metabolic syndrome, physical fitness, physical activity, fatness

Introduction CVD risk factors have been shown to cluster in children nine years of age and older (1). Clustering of CVD risk factors means that the single risk factors are not independently distributed, and elevated levels are found in the same children for many CVD risk factors simultaneously. In an earlier study, we defined six CVD risk factors (1). Four or more risk factors were found in 3.03 (95% confidence interval [CI]: 2.24–4.10) times as many participants as expected from a random distribution and five risk factors were found in 8.70 (95% CI: 4.35–17.4) times as many participants as expected in this nine-year-old population. This clustering has been shown to occur in 10–15% of the population in some countries (2).

Also, this clustering is associated with physical activity level, physical fitness and fatness levels with an increased risk of clustering of around three times for the lower quartile of physical activity, 10 times for the upper quartile of fatness and 15 for the lower quartile of fitness compared to the most favorable quartiles (3). McMurray and Andersen recently reviewed the literature of the association between physical activity, fitness and metabolic syndrome (4). The risk of having metabolic syndrome was consistently lower in fitter individuals, but only two studies were found on longitudinal development of metabolic syndrome. Studies relating physical activity to metabolic syndrome were less consistent and results depended on assessment method of physical

Correspondence: Dr Lars B. Andersen, University of Southern Denmark, Institute of Sport Science and Clinical Biomechanics, Odense, 5230 Denmark. E-mail: [email protected] (Received 13 December 2010 ; final version received 17 June 2011) ISSN Print 1747-7166 ISSN Online 1747-7174 © 2011 Informa Healthcare DOI: 10.3109/17477166.2011.606816

30

L. B. Andersen et al.

activity. In studies using accelerometry, physical activity was consistently negatively associated with metabolic syndrome, but the association was weaker than found for fitness. In a recent study, we investigated six-year-old preschool children, and surprisingly we did not find clustering of CVD risk factors (5). Furthermore, we did not find an association between fitness or fatness and clustering of CVD risk factors. The population included obese children, children with low fitness level and insulin resistant children, and still there was no sign of clustered risk. The observation could suggest that clustering of CVD risk factors develops after the children start in school, and a possible explanation could be the behavioral change that happens at this time. We have not found other studies in children of this age where clustered CVD risk has been analyzed in the general population and it is therefore not possible to judge if the observation can be generalized to other populations. This population has since been followed until the age of nine years, and the purpose of the present study is to analyze whether clustering of CVD risk factors has developed in this population and if clustering is associated with physical activity, physical fitness and obesity.

progressive protocol and detailed criteria for an accepted test are provided elsewhere (7). Habitual physical activity was measured by the MTI 7164 activity monitor (Manufactory Technology Inc., Florida, USA). We chose an epoch of 10 sec. All MTI files were screened for sustained periods of zero activity. Periods of 10 min or more only with zero counts were interpreted as ‘MTI not worn’ and removed from the file. Data were included if the child had accumulated more than 8 h of activity per day for at least three days. Intravenous blood samples were taken between 08:00–09:30 h from the antecubital vein after an overnight fast (8 h). The samples were later analyzed at the Copenhagen Muscle Research Centre for insulin, TC, HDL and TG. Insulin was analyzed spectrophotometrically using an enzyme linked immunosorbent assay (DAKO Insulin, Code no. K6219). Blood lipids were analyzed on a COBAS FARA (Roche, Switzerland) using spectrophotometry (ABX diagnostics, Montpellier, France). Blood pressure was measured sitting with a Dinamap XL vital signs blood pressure monitor (Critikron, Inc., Tamapa, FL, USA) from the left arm five times in 10 min, with the mean of the last three measurements being recorded.

Methods All children in two communities in the Copenhagen area (from 18 schools with a total of 46 preschool classes) were invited to participate in the Copenhagen School Child Intervention Study in 2001. The communities are suburbs approximately 10 km from the central part of the city. The total population of children this age in the two suburbs was 1024, and 706 (69%) volunteered to take part in the study at the age of 6–7 years. Written informed consent was obtained from the parents/guardian. Only 484 children with complete data on all variables are included in the present analysis. A follow-up was conducted in 2004. Children with complete data were 434. The Ethical Committee of Copenhagen County (case no. KA00011gm) approved the study. Height and body weight were measured and body mass index (BMI) was calculated as body weight (kg)· height (m)2. Bicipital, tricipital, subscapular and suprailiac skinfolds were measured with a Harpenden skinfold caliper using standard procedures. Fat percentage was calculated according to the equation of Dezenberg et al. (6). Detailed information has been published earlier (7,8). VO2 was measured directly with an AMIS 2001 Cardiopulmonary Function Test System (Innovision, DK 5260 Odense). The children were instructed to run until exhaustion on a treadmill with a continuous

Risk factors used for clustering analysis Five risk factors for CVD were chosen in the analyses of associations between physical activity and fitness and clustered risk: HOMA score, the ratio of TC to HDL, TG, systolic blood pressure, and the sum of four skinfolds. These risk factors were chosen because they are known to be elements of the metabolic syndrome (9). The reason for choosing skinfold instead of waist circumference or BMI was that skinfold is less dependent on body height, and children at this age mature very differently. As can be seen from the selected risk factors only one measure was selected for fatness, insulin sensitivity, cholesterol, and BP, respectively. This was done to avoid any kind of cumulative effect of variables related with each other (e.g., systolic and diastolic BP). In a second analysis where both fitness and fatness were associated with clustered risk, skinfold was not included in the composite score. In the analysis of independence of the risk factors a child belonging to the top quartile of the distribution (the most unfavourable quartile) was defined as having that risk factor. Quartiles were calculated for each sex separately. Number of CVD risk factors was used to calculate if risk factors were independently distributed. The number of expected risk factors was calculated according to the binomial distribution (n!·(pr·(1-p)n-r)/(r!·(n-r)!)) (10), where n is the

Metabolic syndrome in children possible number of risk factors, p is the probability for being in the quartile defined as being at risk (0.25), and r is the number of risk factors which the probability is calculated for (0 through 5). The expected proportions from a binomial distribution for each number of risk factors were 0.237, 0.396, 0.264, 0.088, 0.015 and 0.001, respectively. Quartiles were only used for this analysis, and a continuous composite score was calculated to analyze associations with physical activity, fitness and fatness. Z-scores were calculated for the same risk factors and summed for the analysis of associations between exposure (physical activity and fitness) and outcome. Finally, a summed z-score which did not include skinfold was calculated and used for an analysis where fitness and fatness were independent variables.

Statistical analysis All analyses were performed using the Statistical Package for the Social Sciences (SPSS) version 18. The associations between individual risk factors and physical fitness were calculated by linear regression using the partial correlation between physical fitness and the risk factors. Logistic regression analysis was performed with cases being defined as 1 SD in sum of z-scores. Fitness was used as an independent variable. Fitness was ranked into quartiles by gender. A significance level of p  0.05 was chosen.

Results Anthropometry, physical activity, physical fitness and CVD risk factors are shown for baseline and followup in Table I. Clustering of CVD risk factors was defined from the distribution of five CVD risk factors.

31

The number of subjects being in the upper quartile of a specific number of CVD risk factors follows a binomial distribution if they are independent of each other. Independency was found in six-year-old preschool children. Three, four and five CVD risk factors were found in 57, eight and two subjects out of 484, respectively. This gave an odds ratio of 1.19 (not significant) for having three or more risk factors compared to the expected number of children. Four or five risk factors were found in 1.28 (not significant) times as many as expected. Among the preschool children, 3.2% had a fat percentage above 25% and 10.2% above 20% (6.8% above IOTF cut-off points for overweight). A HOMA score above 2 was found in 3.1% and fitness levels below 35 ml min–1 kg–1 was found in 1.6% of the boys and 3.2% of the girls. In 3rd grade children, risk factors were not independently distributed. Numbers of subjects with three, four and five CVD risk factors were found in 37, 14 and nine subjects, respectively, out of 434. This gave an odds ratio of 3.33 (95% CI: 1.41–7.87) for three or more risk factors and an odds ratio of 20.74 (95% CI: 1.0–437.4) for four or five risk factors. Three or more risk factors were found in 13.8% of the 3rd grade children. In 3rd grade children, 12% had a HOMA score above 2, and more extreme values were more common than in preschool children with 3.2% above 3. The number of obese and low fit children did not change substantially since preschool. In preschool children, there was only a weak association between cardiorespiratory fitness and the sum of z-scores (r   0.17, p  0.01) and fitness and physical activity (r  0.15, p  0.01), and there was no association between physical activity and sum of z-score. Table II shows the odds ratio of having 1 SD in the sum of z-scores in quartiles of cardiorespiratory fitness and physical activity with the lower quartile as reference. No significant increased risk

Table I. Mean and standard deviation (SD) of anthropometric variables, physical activity, cardiorespiratory fitness and CVD risk factors. Preschool Boys Mean (SD) Age (years) Body mass index (kg m–2) Physical activity (mean count min–1) VO2max (ml min–1 kg–1) Sum of four skinfolds (mm) Systolic Blood Pressure (mmHg) Diastolic Blood Pressure (mmHg) Insulin (mU/l) Insulin·glucose/22.5 Total cholesterol (mmol l–1) HDL (mmol l–1) Ratio total cholesterol:HDL Triglyceride (mmol–1)

6.8 16.0 789.3 48.5 24.0 97.8 57.7 4.08 0.82 4.49 1.50 3.05 0.57

(0.4) (1.6) (216.7) (6.1) (7.9) (8.0) (5.6) (3.70) (0.81) (0.76) (0.26) (0.58) (0.22)

3rd grade Girls Mean (SD) 6.7 (0.3) 15.9 (1.9) 705.9 (164.3) 44.9 (5.4) 27.9 (9.4) 97.4 (8.0) 57.5 (5.9) 4.37 (3.06) 0.83 (0.63) 4.58 (0.87) 1.48 (0.27) 3.15 (0.61) 0.61 (0.29)

Boys Mean (SD) 9.4 17.18 745.3 52.2 29.6 104.9 62.2 5.75 1.24 3.91 1.60 2.51 0.54

(1.5) (2.39) (211.8) (8.8) (14.9) (8.8) (5.9) (3.84) (0.81) (0.60) (0.31) (0.54) (0.27)

Girls Mean (SD) 9.5 (0.8) 17.10 (2.49) 646.5 (169.8) 46.8 (6.2) 36.0 (17.3) 102.3 (8.4) 61.0 (5.9) 6.41 (3.26) 1.39 (0.83) 4.06 (0.65) 1.58 (0.37) 2.64 (0.58) 0.57 (0.23)

32

L. B. Andersen et al.

Table II. Odds ratio (95% confidence interval) of having a sum of five z-scores  1 SD. Included in the sum of z-scores were total cholesterol:HDL ratio, triglyceride, sum of four skinfolds, systolic blood pressure and HOMA-IR score. Reference groups are the lowest quartiles of cardiorespiratory fitness and physical activity. Preschool children Quartiles 1st 2nd 3rd 4th

3rd grade children

Physical activity

Fitness

Physical activity

Fitness

0.7 (0.3–1.5) 0.6 (0.3–1.4) 0.8 (0.4–1.7) 1

1.9 (0.9–4.0) 1.0 (0.5–2.3) 1.0 (0.4–2.3) 1

1.1 (0.4–2.8) 0.9 (0.4–2.4) 1.3 (0.5–3.2) 1

34.9 (8.0–152.5) 8.2 (1.8–37.8) 1.0 (0.1–7.5) 1

was found for fitness in preschool children. When 3rd grade children were analyzed, the association between fitness and CVD risk became stronger. When sum of z-score was analyzed as a continuous variable, the association with fitness was r  0.49 ( p  0.001) and with physical activity r  0.11 ( p  0.05). The association between fitness and physical activity was r  0.19 ( p  0.001). The lower quartile of fitness had a 34.7 times increased risk of having a z-score of 1 SD compared to the fit quartile (Table II). A posthoc analysis was made, where a new clustered risk variable, which did not include fatness (sum of four skinfold), was analyzed against both fitness and fatness. The analysis was not significant in six-year-olds, and the analysis in nine-year-old children is presented in Table III and Figure 1. There was a significant increased risk of clustering of the four remaining risk factors both in relation to low fitness level and high fatness level (skinfold) with around 6–7 times increased risk in the least favorable quartile. Adjustment for the other variable attenuated the risk (Table III). Furthermore, the strength of association was compared for skinfold and waist circumference standardized for height (Figure 1). Waist circumference standardized for height showed stronger relationship with clustered CVD risk. Last, an analysis of physical activity and fitness in preschool as predictors of having a sum of z-scores 1 SD at follow-up was performed. Physical activity at baseline was not associated with clustered CVD risk at follow-up, but low cardiorespiratory fitness was a strong predictor of developing clustered CVD risk

with the risk in the first three quartiles being 6.8 (95% CI: 2.2–21.0), 2.9 (95% CI: 0.9–9.5) and 3.3 (95% CI: 1.0–10.5) compared to the upper quartile. Discussion Metabolic syndrome is characterized by an increase in many CVD risk factors simultaneously. In children, it can be difficult to define what an increased level in a specific risk factor is, because the absolute levels in most risk factors increase with increasing age. We analyzed whether five CVD risk factors were independently distributed, and found that clustering of CVD risk factors was not apparent in children at the age of 6–7 years, but developed in 13.8% of the children before the age of 9–10 years. Clustering of CVD risk factors was strongly related to low fitness level, and low fitness at baseline was also related to clustered risk at follow-up. We did not find an association between clustered CVD risk and accelerometer counts.When skinfold thickness was not included in the composite risk factor score, the score was associated with both fitness (negatively), skinfold and waist standardized for height (positively) with a stronger association for abdominal fat. Other studies have earlier shown an association between physical activity, cardiorespiratory fitness and clustered cardiovascular risk (2,11,12). For a comprehensive review of studies, see McMurray and Andersen (4). This study is the first longitudinal study showing that clustered CVD risk is not prevalent in preschool children, but develops after children start school. We are not able to conclude if it is a

Table III. Sum of skinfolds was removed from the clustered risk variable, which now included total cholesterol:HDL ratio, triglyceride, systolic blood pressure and HOMA-IR score. The association between fitness, fatness and risk of having  1 SD in the sum of four z-scores were calculated. Columns to the left shows results where fitness or fatness was entered into the model separately, and to the right estimates are adjusted for the other variable. Crude analysis

Analysis with adjustment for the other

Quartiles

Sum 4 skinfold

Fitness

Sum 4 skinfold

Fitness

1st 2nd 3rd 4th

1 1.5 (1.5–4.1) 1.9 (0.7–5.0) 7.6 (3.2–18.2)

6.4 (2.8–15.0) 2.3 (0.9–5.7) 1.3 (0.5–3.5) 1

1 1.4 (0.5–4.0) 1.5 (0.5–4.3) 4.3 (1.5–12.3)

2.5 (0.9–7.4) 1.5 (0.6–4.2) 1.1 (0.4–3.0) 1

Metabolic syndrome in children Clustring of BP+TG+HOMA+TC:HDL

Odds ratio of clustered risk

70 60

Skinfold Waist/h

50 40 30 20 10 0 1 Qt 2 Qt 3 QT 4 QT Quartiles of skinfold and waist/height

Figure 1. Odds ratio for quartiles of skinfold and waist circumference standardized for height for having high level (1 SD) in the sum of four risk factors including systolic blood pressure, triglyceride, HOMA score and ratio of total cholesterol and HDL in nine-year-olds. There was no association in six-year-olds.

biological phenomenon, which is founded in early childhood and just takes time to develop or if it is caused by the changes that happens in physical activity when children starts school. We did show that when clustered risk had developed, it was highly associated with low physical fitness, and also that low physical fitness level at the age of 6–7 years predicted later development of clustered CVD risk. If this observation is confirmed by other studies, it should be tested in an intervention study if it is possible to prevent the development of clustered CVD risk by making the children less sedentary or more physically active when they start school. We used gold standard assessment of cardiorespiratory fitness, and measured physical activity objectively. However, even if accelerometry is preferred compared to self-report in the assessment of physical activity it is far from perfect in Danish children. The problem is a substantial amount of cycling, and about 40% of children at the age of nine years cycle to school (13). Furthermore, children cycling to school have 8% higher fitness than children who travel by other means, so this activity which is not picked up by the accelerometer, is substantial (14). It is therefore not surprising that fitness related much stronger to CVD risk factors than physical activity. Fatness was part of the clustered risk score and we therefore did a separate analysis of fatness and fitness as exposures for clustered risk after exclusion of fatness from the composite risk score in 3rd grade children. Earlier studies have shown that fatness is highly related to CVD risk factors. We found that both low fitness and high fatness predicted clustering of CVD risk factors independently. Some studies claim that one is more important than the other, but

33

this is not possible to conclude from cross-sectional studies and not even from longitudinal observational studies (15,16). The problem is that we cannot be sure how the causal pathway is. The interpretation of a statistical analysis depends on how we believe the causal pathway is. If a variable is part of a causal chain, adjustment for this variable will remove the association between the real cause (first variable in the chain) and the outcome, but if it is a confounder, we must adjust for it. Whether fitness or fatness are most important in relation to clustered risk is therefore not possible to conclude from the existing evidence, which mainly include observational studies. It may not even be possible to conclude which is more important in a randomized trial, because it may be impossible to manipulate one of them without changing the other. It is possible to make a training intervention with no weight change, but probably not without a change in body composition. However, we may conclude from our analysis that abdominal fat is a better predictor of clustered risk than skinfold. Hormones such as insulin and adrenalin may play a role in development of clustered CVD risk. Many studies have supported this observation since Reaven first described syndrome X in the mid-1980s (17). Sensitivity of these hormones is associated with physical activity, fitness and abdominal fatness, and there is a strong relationship between glucose metabolism and clustering of CVD risk factors (18). We cannot explain why clustered CVD risk was not apparent in six-year olds, because we did find low fit, obese and insulin resistant children. They just had not yet developed a metabolic disorder. Future research should investigate whether diet and exercise interventions could prevent clustering of CVD risk factors in this age group, because interventions in children where it has already developed have shown modest results. In conclusion, in this study sample clustered CVD risk developed in 13.8% of the children between the ages of six and nine years or during the first three years in school. No clustering of CVD risk factors was seen in 6- to 7-year-olds even if some of the children were obese, low fit and partly insulin resistant. The observation raises the question if it is possible to prevent clustering of the risk factors by keeping the children physically active and fit. Acknowledgements The authors wish to thank Copenhagen Muscle Research Centre for analyzing the blood samples and the communities of Ballerup and Tårnby. This work was supported by grants from Danish Heart Foundation, The National Board of Health, The Danish Ministry of Health, The Danish Ministry of Culture, The Danish Sport Association.

34

L. B. Andersen et al.

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper. References 1. Andersen LB, Wedderkopp N, Hansen HS et al. Biological cardiovascular risk factors cluster in Danish children and adolescents. Danish part of the European Heart Study. Prev Med. 2003;37:363–7. 2. Andersen LB, Harro M, Sardinha LB et al. Physical activity and clustered cardiovascular risk in children: a cross-sectional study (The European Youth Heart Study). Lancet. 2006;368: 299–304. 3. Andersen LB, Sardinha LB, Froberg K et al. Fitness, fatness and clustering of cardiovascular risk factors in children from Denmark, Estonia and Portugal: the European Youth Heart Study. Int J Pediatr Obes. 2008;(3 Suppl. 1):58–66. 4. McMurray RG, Andersen LB. The influence of exercise on the metabolic syndrome in youth: a review. Am J Lifestyle Med. 2010;4:176–86. 5. Eiberg S, Hasselstrom H, Gronfeldt V et al. Physical fitness as a predictor of cardiovascular disease risk factors in 6- to 7-year-old Danish children: the Copenhagen School-Child Intervention study. Pediatr Exerc Sci. 2005;17:161–70. 6. Dezenberg CV, Nagy TR, Gower BA et al. Predicting body composition from anthropometry in pre-adolescent children. Int J Obes. 1999;23:253–259. 7. Eiberg S, Hasselstrom H, Gronfeldt V et al. Maximum oxygen uptake and objectively measured physical activity in Danish children 6–7 years of age: the Copenhagen school child intervention study. Br J Sports Med. 2005;39:725–30. 8. Hansen SE, Hasselstrom H, Gronfeldt V et al. Cardiovascular disease risk factors in 6- to 7-year-old Danish children: the Copenhagen School Child Intervention Study. Prev Med. 2005;40:740–6.

9. Grundy SM, Brewer HB Jr, Cleeman JI et al. Definition of metabolic syndrome: report of the National Heart, Lung, and Blood Institute/American Heart Association conference on scientific issues related to definition. Circulation. 2004; 109:433–8. 10. Altman DG. Theoretical distributions. Pract Stat Med Res. 1991;68–70. 11. Anderssen SA, Cooper AR, Riddoch C et al. Low cardiorespiratory fitness is a strong predictor for clustering of cardiovascular disease risk factors in children independent of country, age and sex. Eur J Cardiovasc Prev Rehabil. 2007; 14:526–31. 12. Rizzo NS, Ruiz JR, Hurtig-Wennlof A et al. Relationship of physical activity, fitness, and fatness with clustered metabolic risk in children and adolescents: the European youth heart study. J Pediatr. 2007;150:388–394. 13. Cooper AR, Page AS, Foster LJ et al. Commuting to school: are children who walk more physically active? Am J Prev Med. 2003;25:273–6. 14. Cooper AR, Andersen LB, Wedderkopp N et al. Physical activity levels of children who walk, cycle, or are driven to school. Am J Prev Med. 2005;29:179–84. 15. Klasson-Heggebo L, Andersen LB, Wennlof AH et al. Graded associations between cardiorespiratory fitness, fatness, and blood pressure in children and adolescents. Br J Sports Med. 2006;40:25–9. 16. Christou DD, Gentile CL, DeSouza CA et al. Fatness is a better predictor of cardiovascular disease risk factor profile than aerobic fitness in healthy men. Circulation. 2005;111: 1904–14. 17. Reaven GM, Chen YD. Role of insulin in regulation of lipoprotein metabolism in diabetes. Diabetes Metab Rev. 1988;4:639–52. 18. Andersen LB, Boreham CA, Young IS et al. Insulin sensitivity and clustering of coronary heart disease risk factors in young adults. The Northern Ireland Young Hearts Study. Prev Med. 2006;42:73–7.