Changes in adiposity and movement behaviors were revealed during the transition from childhood to adolescence. All adiposity markers, total SB, and VPA increased, while LPA and MVPA decreased. A compositional prospective association between the changes in the proportion of time spent in middle sedentary bouts and FM% in adolescence was found. Moreover, a change in VPA was associated with VAT in adolescence. Favorable changes in adiposity status were identified when the amounts of time spent in total SB and sedentary bouts of different durations were substituted with VPA.
Our findings on the age-related changes in adiposity and movement behaviors are in accordance with available evidence [29–31]. We found that the increase in the proportion of time spent in total SB was predominantly caused by an increase in the time spent engaging in prolonged SB. Similar results were observed in 5 991 participants included in a pooled analysis of longitudinal data from the International Children’s Accelerometry Database . Moreover, the proportion of time spent in middle and long sedentary bouts increased at the expense of LPA and MPA in the current study. This implies that SB becomes less fragmented, as it is probably less interrupted by brief bouts of PA. This assumption might be supported by the longitudinal analysis performed by Janssen and colleagues , who found a decrease in the frequency of sedentary breaks and an increase in the sedentary bout duration across childhood and adolescence.
One possible explanation for the decrease in sedentary fragmentation could be a developmental shift in sleep duration. It is recommended that adolescents might sleep less than children to achieve similar health benefits . However, our previous investigation  suggests that school-aged children tend to spend extra waking hours in prolonged SB. Such time-use patterns could potentially modify the association between SB patterns and adiposity. Another explanation could be a change in movement behaviors associated with the transition to secondary school. As most school time corresponds to SB [33, 34] and considering the SB patterns during lessons (i.e., prolonged uninterrupted sitting), it can be assumed that the increase in prolonged SB is caused mainly by an increase in school-based SB. In addtion, the age-related decrease in non-organized leisure time activities may also be associated with unhealthy time use across childhood and adolescence . If further longitudinal studies confirm these assumptions, interventions targeting the interruption of prolonged sitting during school and leisure time should be implemented into public health strategies.
The change in SB patterns during the transition from childhood to adolescence could have several health implications. According to this study, an increase in the proportion of middle sedentary bouts is prospectively associated with adiposity in adolescence. Similar to our findings, Mann and colleagues  showed that a decrease of one bout per sedentary hour per year (i.e., SB becomes less fragmented) was associated with an increase in BMI and FMI between 7 and 12 years of age. However, the ability to compare our findings with those previously published is limited, as, to the best our knowledge, the current study is the first to investigate prospective associations between SB patterns and adiposity using the CoDA approach. Despite this limitation, it appears that changes in SB patterns play an important role in the accumulation of adipose tissue in adolescence. Future studies are needed to identify the potential determinants of SB pattern changes, as this was beyond the scope of the present study.
We emphasize that an increase in time spent in VPA is associated with lower central adiposity in adolescence. This is in accordance with the review by Gralla and colleagues , who suggested that VPA is a stronger predictor for central adiposity than other intensities of PA. In light of this finding, we suggest one possible strategy to reduce adiposity is interrupting prolonged SB by bouts of VPA. This hypothesis is supported by our results from the isotemporal substitution analysis, which indicate an improved adiposity status when 15 min of total SB or all sedentary bouts are reallocated in favor of VPA. However, it is questionable whether that change could be sustainable, as it would mean more than doubling the amount of time spent in VPA in our sample. An alternative strategy may be interrupting SB by very short bouts (i.e., < 1 min) of VPA. This specific point could not be addressed in the current study, as the movement behaviors were assessed using a 60 s sampling interval. However, the findings from an experimental study conducted in children with overweight and obesity  support this strategy by showing an acute improvement in metabolic markers in response to interrupting SB with very short bouts of walking.
The current study has several strengths. First, this is the only study that considers the compositional nature of movement behavior data in analyzing the prospective associations between changes in SB patterns and adiposity. The utilization of the CoDA approach allows the control of regression models for all movement behaviors (i.e., using the set of the first pivot coordinates in the models) and the avoidance of biased estimates due to multicollinearity. Second, the device-based assessment of both adiposity markers and movement behaviors provides reliable and valid data. Finally, the SB patterns were analyzed as well, which allowed the accurate differentiation of prolonged SB from total SB.
There are also limitations within this study. Our findings should be interpreted with caution, as the study is limited to movement behaviors. Omitting time spent asleep may lead to biased estimates of regression analysis. Another source of bias relates to analyzing SB without distinguishing body postures (i.e., the differentiation of sedentary postures from standing). Thus, longitudinal CoDA-based studies based on a 24-h wear time protocol and the combination of intensity- and posture-specific assessments of movement behaviors are warranted to analyze the associations between SB patterns and adiposity across childhood and adolescence. Although the regression models were adjusted for several confounding factors, there are still potential endo- and exogenous determinants of adiposity that have not been considered. Since youth movement behaviors are characterized by their intermittent nature, the long sampling interval used in this study could lead to the underestimation of high-intensity PA and may not capture all breaks in SB . Finally, our findings are not fully generalizable due to the small sample size and the higher proportion of girls.