This study examined associations between the volume, patterns, and types of sedentary behavior with cardio-metabolic risk factors in 6-19 year olds. Although this representative sample spent 50.8% of their waking hours in sedentary behavior, the volume of sedentary behavior was not an independent predictor of high-risk cardio-metabolic factor values. Similarly, patterns of sedentary behavior, such as the amount of time in bouts of sedentary behavior ≥30 minutes, was not related to cardio-metabolic risk factors. However, the type of sedentary behavior was important. More specifically, the amount of time spent watching TV was related to cardio-metabolic risk factors, while computer use was not.
Among adults, the volume of sedentary behavior, as measured objectively by accelerometers, is associated with a clustering of cardio-metabolic risk factors , waist circumference , and glucose intolerance  that are independent of MVPA and other confounders. The relationship between the volume of sedentary behavior and cardio-metabolic risk factors appears to be stronger and more consistent in adults than young people. In our study, the volume of objectively measured sedentary behavior was not associated with high CRS or its individual components. Two previous cross-sectional studies within children and/or adolescents have examined the association between the volume of sedentary behavior, measured by an accelerometer, with a summary cardio-metabolic risk score [10, 11]. Positive associations were observed in both studies. However, neither study examined whether these associations were independent of MVPA, which is an important limitation, given that MVPA is related to sedentary behavior (see Table 2). In addition, five previous studies within children and adolescents, all cross-sectional in design, have examined the association between the volume of sedentary behavior, measured by an accelerometer, with individual risk factors such as insulin resistance , blood pressure , and various measures of obesity [42–44]. One of these studies found moderate positive associations (r = 0.21) between the volume of sedentary behavior and insulin resistance among 9-10 year old children independent of obesity . Similarly, this study did not determine whether the associations were independent of MVPA. In the three studies that examined obesity measures, associations with volume of sedentary behavior did not exist or was severely attenuated after adjustment for MVPA [42–44].
Emerging evidence in adults suggests that patterns in which sedentary behavior is accumulated may independently impact cardio-metabolic risk. More specifically, a cross-sectional study of 168 Australian adults found that the frequency of breaks in sedentary behavior was negatively related to waist circumference, triglycerides, and glucose levels, independent of MVPA and the overall volume of sedentary behavior . We are unaware of previous studies that have examined these associations in children or adolescents. Thus, our observation that patterns of sedentary behavior, including sedentary behavior bouts and breaks in bouts of sedentary behavior, were not related to cardio-metabolic risk factors in 6-19 year olds makes a novel contribution to the literature. It is possible that the differences in results between the present study and the previously mentioned adult study  is explained by a physiological difference in the way sedentary behavior impacts health in adults and young people. It is also possible that the different results are due to differences in the way "breaks" were measured in the two studies. While the present study looked at the frequency of breaks within prolonged (≥30 minutes) bouts, the Australian study counted a break any time the participant moved from a sedentary minute to a minute above the 100 count per minute accelerometry threshold . Due to the dearth of information, more research is needed to better understand the relationship between patterns of sedentary behavior and cardio-metabolic health in all ages.
Our third objective was to determine if different types of sedentary behavior impact cardio-metabolic health in a similar manner. Several studies among young people have found associations between TV and total screen time with individual cardio-metabolic risk factors and the metabolic syndrome [12, 14, 45, 46]. To our knowledge only one of these studies reported on the impact of different screen time measures on cardio-metabolic risk factors other than obesity . This particular study found that blood pressure was associated with TV but not with computer use . Also, a recent literature review found that TV use is more strongly associated with obesity in children and adolescents than is computer use . Likewise, we found that the odds of a high CRS increased in a dose-response manner within increasing TV volume, independent of MVPA, but that computer use did not predict CRS. Similar associations were seen with the individual components of CRS. There are two possible explanations for these findings. First, amongst the sedentary behaviors, TV may be at the lowest end of the energy expenditure spectrum. In fact, one study reported that energy expenditure in children and adolescents was lower while watching TV then while sleeping . Second, TV encourages between meal snacking  and is associated with a greater exposure to junk food advertisements than other screen time measures . Even though various dietary measures (total fat, saturated fat, cholesterol, and sodium) were adjusted for in this study, residual confounding may have been present. Future research needs to consider the impact of other types of sedentary behavior (reading, homework, etc.) on the health of young people.
Interestingly, in the present study the CRS variable was predicted by a self-report measure of TV use but not an objective measure of overall sedentary behavior volume or the overall volume of sedentary behavior accumulated in prolonged bouts. Also, TV use was poorly correlated (r ≤ 0.08) to these two objective measures. There are three possible explanations for these observations. First, the uniaxial accelerometer used in NHANES may not be sensitive enough to differentiate between sitting and standing like an inclinometer . Also, participants may have been more likely to keep their accelerometer on during their daily activities and MVPA, and take it off later in the evening while watching TV . Therefore, the objectively measured sedentary behavior may not have captured 100% of the sedentary behavior for some participants. Second, the specific sedentary behavior of TV may have a unique impact on cardio-metabolic risk factors due to its impact on energy expenditure and intake, as previously discussed. Third, the catchment period of sedentary behavior differed between the self- report measure (past 30 days) and the accelerometer measure (7 days). Perhaps, the longer catchment period better reflects typical behavior compared to the shorter period.
We also examined the association between objectively measured MVPA with CRS and non-obesity CRS. The finding that MVPA was strongly and independently associated with cardio-metabolic risk factors in a dose-response manner is consistent with previous literature . For example, a dose-response relationship between MVPA and cardio-metabolic risk factors was observed in approximately 2000 participants of the European Youth Heart Study . As with the present study, the associations between MVPA and clustered cardio-metabolic risk factors within children and adolescents have been reported to be independent of TV use and obesity . However, similar to the present study, the association between TV use and clustered cardio-metabolic risk factors do not appear to be independent of obesity . This suggests that obesity mediates or confounds the relationship between TV use and clustered cardio-metabolic risk factors .
Strengths of this study include the objective measures of MVPA and most of the sedentary behavior variables as well as our novel approach used to examine patterns of sedentary behavior. Limitations of the study include the cross-sectional design, which limits the ability to make causal inferences about the relationships. Also, our final sample was not representative of the population in terms of ethnicity. In addition, the accelerometers used may not be sensitive enough to differentiate between sitting and standing . Furthermore, we only considered two types of sedentary behavior, both of which were measured via self-report. The biases with these self-reported measures may have results in an underestimation of the strength of associations between the TV, computer, and CRS variables. Finally, although a variety of confounders were considered, we were not able adjust for pubertal development, a factor which influences physiological processes .