Background
Sedentary behavior [SB], defined as any waking behavior characterized by an energy expenditure ≤ 1.5 metabolic equivalents while in a sitting or reclining posture [
1], is highly prevalent [
2]. Prevalence has increased as utilization of media technologies in leisure time has increased. Physical activity [PA] has decreased, both at work and in leisure time [
3]. There is emerging evidence that changes in patterns of PA and SB are independently related to factors of cardiometabolic health including the metabolic syndrome [
2,
4]. Meta-analysis data revealed that spending high amounts of time engaging in SB increased the odds of having the metabolic syndrome by 73% compared to spending low amounts of time engaging in SB [
4].
The association between leisure-time SB and factors of cardiometabolic health depends on the type of SB and the age of individuals [
5]. Many studies have focused on television [TV] time or time spent using a computer and their associations with cardiovascular health [
2,
6]. Few studies have analyzed associations between other SBs such as reading or socializing in leisure time and cardiometabolic risk factors [
7‐
12]. Time spent watching TV was found to be positively associated with obesity [
10], type II diabetes [
7], overweight [
8,
9], or individual cardiometabolic biomarkers [
10‐
12]. The evidence for using a computer in leisure time is inconsistent [
7‐
12]. No associations have been found between other SBs such as reading or socializing in leisure time and cardiometabolic risk factors [
7,
8].
A growing body of literature recommends using a cluster of continuous cardiometabolic risk factors instead of a binary definition of the metabolic syndrome [
13,
14]. The reasons include (i) cardiovascular risk increases progressively with an increasing number of metabolic syndrome risk factors, and (ii) using a continuous risk score increases statistical power [
14].
Current evidence has suggested that a greater increase in overall sedentary time is associated with a greater increase in clustered cardiometabolic risk in adults at high risk of developing type II diabetes [
15] or in a population-based sample [
16]. Nevertheless, to the best of our knowledge, no study has examined a broader range of screen-based and other leisure-time SBs and their associations with clustered cardiometabolic risk. One study has analyzed associations between SBs in leisure time and two clustered cardiometabolic risk scores over a 2-year follow-up period among adults at increased cardiometabolic risk [
17]. A risk score of developing type II diabetes (Atherosclerosis Risk in Communities study) and a score of developing fatal cardiovascular disease (Systematic Coronary Risk Evaluation) was used. The findings suggested no associations between time spent watching TV, using a computer, or reading and individual cardiometabolic risk factors or clustered cardiometabolic risk scores [
17].
Given the ubiquitous nature of prolonged SB in leisure time, a deeper understanding is needed about whether behaviors are associated with clustered cardiometabolic risk in order to develop adequate prevention strategies. The present study aimed to examine associations between leisure-time SBs (watching TV, using the computer, and playing computer games, reading, doing household tasks, caring for others, pursuing hobbies, and socializing) and clustered cardiometabolic risk in apparently healthy adults.
Discussion
There were two main findings of our study. First, watching TV was positively associated with CMRS. In addition, depending on the quantiles of CMRS, QR analysis revealed a negative association between computer time and CMRS. However, this association disappeared after adjusting for PA in leisure time and time spent traveling in motor vehicles. Second, no associations were present between reading or socializing and CMRS.
Our results suggest that study participants who spend higher amounts of time watching TV are at higher cardiometabolic risk than individuals with low levels of TV time. This association remained significant after adjusting for leisure-time PA and time spent traveling in motor vehicles. Furthermore, QR analyses revealed an association between computer time and CMRS that otherwise is hidden if using the mean of CMRS in OLS regression analysis. Among study participants in the medium and in the highest cardiometabolic risk group, higher amounts of time using a computer were associated with a more favorable cardiometabolic profile. However, this association disappeared after adjusting for leisure-time PA and time spent traveling in motor vehicles.
Our findings on associations between TV time and CMRS are in line with current evidence on associations between TV time and individual cardiometabolic risk factors [
12,
10,
2]. Additionally, watching TV has been shown to be associated with lower energy expenditure [
31], an increased intake of food with high energy density and overall unhealthy dietary habits [
32,
33] compared with other sedentary activities such as using a computer. A combination of these factors may explain our findings [
12,
10]. According to the QR result, time spent sedentary while using a computer may be differentially associated with CMRS. Whereas a population-based study suggested no association between using a computer in leisure time and individual risk factors of cardiometabolic health [
11], Heinonen et al. [
10] reported a positive association between computer time and WC as well as body mass index among women but not among men in a middle-aged sample. In contrast to these studies, our results of QR revealed a negative association between using a computer and CMRS among individuals in the medium and in the highest cardiometabolic risk group. After adjusting for two PA variables, the statistically significant association between using a computer and CMRS disappeared whereas the negative direction of the association remained. Thus, different behaviors in which individuals spend their time in a sitting or reclining posture may not influence the magnitude and direction of the association with clustered cardiometabolic risk in the same manner.
Our second finding adds to the literature that there seems to be no association between reading and socializing and clustered cardiometabolic risk. This finding is in line with previous studies that examined associations between time spent sedentary while reading or socializing and individual cardiometabolic health factors [
7‐
12,
17]. The less time spent reading or socializing in leisure time might be an explanation for the findings in our sample. In addition, accuracy to recall across contexts may vary [
2] in the sense that it may be easier for people to remember how long they have watched TV than how much time they have spent reading or in company with others [
34].
Furthermore, we found that leisure-time SBs differed in their frequency of occurrence. Time spent sitting while doing household tasks, caring for others, or pursuing hobbies were less prevalent in our sample. Evidence suggests that different types of SB often co-occur compared to activities with higher energy expenditure (over 1.5 metabolic equivalents) [
22], e.g. using the computer or doing household tasks while watching TV. Thus, it is important to consider not only the frequency of occurrence but also that of co-occurrence of leisure-time SB and how different types of leisure-time SB are linked to one another. Future studies should examine those patterns of leisure-time SB in detail and might include separate analyses for weekdays and weekends, because leisure-time SB patterns have been shown to vary between weekends and weekdays [
35].
Some limitations of our study have to be discussed. First, subjects were assessed within a study aiming to test the feasibility of a tailored letter intervention regarding PA and leisure-time SB. The proportion of people who declined participation (53%) was high and a selection bias is likely. Thus, our findings may not be generalizable to the population as a whole. Second, there may be confounding variables such as diet, drinking habits, different activity patterns with certain energy expenditure during leisure-time SBs, sleep duration, or cardiorespiratory fitness that were not considered in our study. Third, we collected blood samples in the non-fasted state. Because levels of glucose or HDL-C are influenced by external factors like caloric intake or muscle activity [
36], this may have implications for the clustered cardiometabolic risk. Although using fasting blood samples is recommended, there is evidence that using non-fasting blood samples is appropriate for decision making in the context of primary preventions regarding cardiovascular or cardiometabolic diseases [
36,
37]. Fourth, we assessed leisure-time SB by self-report. Self-report assessments are sensitive to recall bias and social desirability [
2]. In comparison to accelerometer measures of SB, self-report appears to capture different aspects of behaviors [
38] because it provides information on the context. Keeping in mind that leisure-time SB is complex and includes multiple domains, dimensions, and correlates, there are still many methodological challenges of measuring leisure-time SB [
39]. Finally, the design of our study does not allow for causal inference. To address this issue, more longitudinal studies are needed to understand the directionality of potential associations between leisure-time SBs and cardiometabolic health.
Acknowledgements
The authors wish to thank funders, supporters, and participants of the study. Further, we thank Liane Müller and Ramona Mühlenbächer for their assistance in screening, recruitment, and with data collection.