Associations between accelerometry measured physical activity and sedentary time and the metabolic syndrome: A meta‐analysis of more than 6000 children and adolescents

Summary Background Metabolic syndrome is increasingly prevalent in the pediatric population. To prevent an early onset, knowledge about its association with modifiable lifestyle factors is needed. Objectives To estimate the prevalence of the metabolic syndrome and examine its cross‐sectional associations with physical activity and sedentary time. Methods Participants were 6009 children and adolescents from 8 studies of the International Children's Accelerometry Database. Physical activity and sedentary time were measured by accelerometer. Metabolic syndrome was defined based on International Diabetes Federation criteria. Logistic regression models adjusted for sex, age and monitor wear time were used to examine the associations between physical activity, sedentary time and the metabolic syndrome in each study and effect estimates were combined using random‐effects meta‐analysis. Results The overall prevalence of the metabolic syndrome was 2.9%. In crude models, a 10 min increase in moderate‐to‐vigorous intensity physical activity and vigorous‐intensity physical activity were inversely associated with the metabolic syndrome [OR 0.88, 95% CI 0.82‐0.94, OR 0.80, 95% CI 0.70‐0.92]. One hour increase in sedentary time was positively associated with the metabolic syndrome [OR 1.28, 95% CI 1.13‐1.45]. After adjustment for sedentary time, the association between moderate‐to‐vigorous‐intensity physical activity and the metabolic syndrome remained significant [OR 0.91, 95% CI 0.84‐0.99]. Sedentary time was not associated with the metabolic syndrome after adjustment for moderate‐to‐vigorous intensity physical activity [OR 1.14 95% CI 0.96‐1.36]. Conclusions Physical activity of at least moderate intensity but not sedentary time is independently associated with the metabolic syndrome.


| INTRODUCTION
Although the prevalence of the metabolic syndrome (MetS) in children and adolescents is low, it rises with age. 1,2 Furthermore MetS is more prevalent in youth with overweight and obesity 3 which is especially threatening with regard to the continuous rise of obesity levels in this population. 4 Cardiovascular disease (CVD) risk factors already begin to cluster during childhood and adolescence 5 and track into adulthood. 6 Evidence from longitudinal studies suggests that pediatric MetS may transfer into an increased risk for CVD, type 2 diabetes mellitus and premature death. [7][8][9] To develop effective public health strategies and prevent an early onset of MetS, knowledge about preventive lifestyle factors such as physical activity and time spent sedentary is required.
Existing evidence in children and adolescents indicates a beneficial association between objectively measured physical activity and several cardio-metabolic biomarkers as well as adiposity. 10 The association between sedentary time and cardio-metabolic health remains inconsistent. A recent systematic review suggested strong evidence for a prospective relationship between TV-viewing time and obesity and for an inverse relationship between different sedentary time measures and HDL-cholesterol but no or insufficient evidence for associations between sedentary time and other biomarkers and metabolic risk scores. 11 When total sedentary time measured by accelerometry is considered, some found cross-sectional associations between sedentary time and a metabolic risk score 12,13 or single risk factors 14 while others did not. 15,16 Further, associations between sedentary time and metabolic risk scores appear attenuated when additionally adjusted for physical activity. [17][18][19] Prospectively, no association between sedentary time and a metabolic risk score or single risk factors have been observed. 20,21 However, most studies with objective measurements of physical activity and sedentary time have used metabolic risk scores or single biomarkers as outcomes and only three studies have assessed the associations between objectively measured physical activity and MetS 1,22,23 and only one of these examined the associations between objectively measured sedentary time and MetS as a dichotomous outcome in children and adolescents. 23 The advantage of using MetS as a dichotomous variable rather than a continuous score is that MetS prevalence and effect estimates (e.g. OR, RR) can be compared across studies and countries.
The objectives of this study were therefore to estimate the prevalence of MetS and examine the associations between objectively measured physical activity, sedentary time and MetS in a large and diverse multicentre sample of 6009 children and adolescents using a metaanalytical approach.

| Participants
For the present analysis we used data from 8 studies from Europe and the United States, which provided data on objectively measured physical activity and comparable fasting blood samples. [25][26][27][28][29][30] The included studies' designs are longitudinal, 28

| Assessment of physical activity and sedentary time
Physical activity and sedentary time were measured using hip-worn, uniaxial ActiGraph accelerometers (models 7164, 71256 and GT1M), previously validated for energy expenditure assessed by the doubly labelled water method in free living children. 31 All available raw accelerometer files were centrally processed, cleaned and reanalyzed using Kinesoft software (Kinesoft, version 3.3.20), to provide physical activity variables that could be directly compared. Before analysis, all files with shorter epoch lengths than 60 seconds were reintegrated to 60 second epochs for comparability. Non-wear time was considered as 60 minutes of consecutive zeros, allowing for two minutes of non-zero interruptions. The valid day criterion for the present study was 480 minutes of measured wear time between 7 AM and midnight. Individuals with at least one valid day were included. To estimate the time spent in light (101-2295 cpm), moderate (2296-4011 cpm) and vigorous intensity (> 4011 cpm), as well as sedentary time (0-100 cpm), Evenson cut-points were used. 32 Total physical activity was expressed in counts per minute (CPM). For each day, total accelerometer counts were divided by monitor wear time in minutes and then averaged across all valid days.

| Assessment of the MetS
MetS was defined according to the IDF pediatric definition in children and adolescents aged 9 to 15 years and according to the IDF worldwide adult definition in adolescents aged 16 to 18 years. 33,34 The IDF pediatric definition defines MetS as having abdominal obesity (waist circumference ≥ 90 th age and sex specific percentile or adult cut-off if lower) and the presence of two or more of the following clinical features: elevated triglycerides (≥ 1.7 mmol/L), low HDL-cholesterol (< 1.03 mmol/L), high blood pressure (systolic ≥ 130/diastolic ≥ 85 mm Hg) and increased fasting plasma glucose (≥ 5.6 mmol/L) or known type 2 diabetes mellitus. 33 Reference values were taken from a populationbased sample of British children. 35 The IDF worldwide adult definition defines MetS as having abdominal obesity (≥ 94cm for Caucasian men and ≥ 80cm for Caucasian women) and the presence of two or more of the following clinical features: elevated triglycerides (≥ 1.7 mmol/L or specific treatment for this lipid abnormality), low HDL-cholesterol (< 1.03 mmol/L in males and < 1.29 mmol/L in females or specific treatment for this lipid abnormality), high blood pressure (systolic ≥ 130/diastolic ≥ 85 mm Hg or treatment of previously diagnosed hypertension) and increased fasting plasma glucose(≥ 5.6 mmol/L) or known type 2 diabetes mellitus. 34 Blood pressure was measured using standard procedures, described in detail elsewhere 19,[25][26][27] Height and weight were measured using standardized clinical procedures across studies. Body mass index (BMI) was calculated as weight in kilograms divided by the square of height in meters (kg/m 2 ). Except for the NHANES (National Health and Nutrition Examination Survey), waist circumference (WC) was measured at the end of gentle expiration midway between the lower rib margin and the iliac crest using a metal tape. [27][28][29][30] In NHANES WC was measured just above the iliac crest at the midaxillary line using similar equipment. 36 Triglycerides, HDL cholesterol Insulin and plasma glucose were measured in all studies. [25][26][27][28][29][30] All blood samples were taken according to standard clinical procedures as previously described. 25

| RESULTS
Cohort Characteristics are presented in Table 1 and differences between included and excluded individuals are presented in Table S1 in the Supporting Information. Excluded individuals were slightly younger (age 13.5 year (1.9) vs. 14.0 year (2.6), p <0.001) and more active (TPA 533 cpm (212) vs. 501 cpm (226), p < 0.001).
One hundred and seventy-six children and adolescents had MetS  Inter.

Boys (n)
Results of the random-effects meta-analysis are shown in Table 2.
Total physical activity was inversely associated with MetS (Fig. S1).
For every 100 cpm increase in total physical activity the odds of MetS After additional adjustment for sedentary time the association between MVPA and MetS was slightly attenuated but remained statistically significant (Fig. 1)    used. 42 Although the most optimal threshold used to define sedentary time from accelerometry is debated, it is likely higher cut-points than 100 CPM increase the risk for misclassification. 43 Further, nonwear criteria of at least 60 minutes of consecutive zero-counts are recommended. 44 Several studies have suggested associations between objectively measured sedentary time and cardiometabolic risk factors, but these associations seem to disappear when controlling for MVPA. 18,19,42 Similarly, our data suggests that the association between sedentary time and MetS is attenuated and non-significant following adjustment for time spent in MVPA. Thus, sedentary time appears not an independent risk factor for MetS in children and adolescents.
We were unable to include data from the KISS study in our meta-  The excluded children and adolescents were younger and more active (Table S1). This may have led to a slight overestimation of the prevalence of MetS overall, as MetS is associated with older age and physical inactivity. 1  cycling or swimming are recorded inadequately. Accelerometer measurements involve some subjective decisions that can influence the association between exposure and outcome. However, as previously discussed, the cut-off points and non-wear-time criteria we used seem realistic, 32 and our decision to include all individuals with at least one day of at least 8h monitor wear time seems robust. Sensitivity analyses in individuals providing at least 2, 3 or 4 valid days did not change our results notably (Table S4). Finally, although less likely, we cannot exclude the possibility that our results are explained by unmeasured confounding factors such as dietary intake, socioeconomic status and genotype.
The prevalence of MetS overall was low, but differed substantially between the single studies. MVPA was associated with MetS in children and adolescents independent of sedentary time. Sedentary time was not associated with MetS after adjustment for MVPA or VPA.
These findings suggest that the promotion of physical activity might be more useful as a public health focus than the reduction of sedentary time.