Sedentary Time in Relation to Cardio-metabolic Risk Factors

Differential Associations for Self-Report vs Accelerometry in Working Age Adults

Emmanuel Stamatakis; Mark Hamer; Kate Tilling; Debbie A Lawlor


Int J Epidemiol. 2012;41(5):1328-1337. 

In This Article


In this study of a representative sample of the UK population, we observed associations between self-reported ST and a number of cardio-metabolic risk factors, which confirms previous findings.[1–3] ST domain-specific analyses suggested that these associations were more consistent for TV time than for other domains. Importantly, these associations were not attenuated by adjustments for self-reported and objectively assessed MVPA. We did not, however, confirm most of these associations when using objective (accelerometry-based) measures of ST. These findings suggest that associations of TV viewing with adverse cardio-metabolic risk factors may not represent a causal association between ST and these outcomes, but might reflect the association of TV viewing with other behavioural risk factors, such as consuming high-energy snack foods and/or the influence of TV advertisements on unhealthy behaviours.[10] In the present analyses, we were only able to adjust for simple measurements of diet and hence residual confounding from these characteristics might remain. Besides behavioural residual confounding, we[25] and others[26] have reported TV viewing to be associated with adverse psychological outcomes. Given that cardio-metabolic risk is associated with psychological factors,[27,28] this may be another potential pathway through which TV viewing adversely affects cardio-metabolic risk profile.

The associations between TV viewing and cardio-metabolic risk we observed are consistent with several cross-sectional[1,2,29,30] and some longitudinal[8,31] studies that have found deleterious associations with adiposity markers,[1,2,8,29–31] blood lipids[2,8,29–31] or clustered cardio-metabolic risk profiles.[8,31] On the other hand, our results showing limited associations between accelerometry-assessed ST and the cardio-metabolic risk outcomes partly contrast with another recent cross-sectional analysis among 4757 US adults who participated in the 2003–06 National Health and Nutrition Examination Survey and who were also given an Actigraph accelerometer to wear for 7 days.[32] In this study by Healy et al., accelerometry-assessed ST showed adverse associations with waist circumference and HDLC but, as in our study, there was no evidence for associations with systolic or diastolic blood pressure. In addition, Healy et al. reported detrimental associations between accelerometry-assessed ST and a number of cardio-metabolic risk factors that were not available to us (C-reactive protein and fasting triglycerides). In terms of glycaemic status markers, Healy et al. found an association with fasting insulin but we did not find any association with HbA1c.[32] We examined whether the differential associations in the overlapping outcomes and glycaemic status markers were due to the different accelerometry cutoff (200 counts/min vs 100 counts/min used by Healy et al.) by repeating all our analyses using a ST cutoff of 100 counts/min. As with the main analysis (Table 3), we found evidence only for an association between ST and total cholesterol (data not shown). Another smaller study among obese individuals[33] failed to find evidence for an association between accelerometry-assessed ST and glucose metabolism.

The discrepancy observed between self-reported and objectively assessed ST might also be explained by other methodological issues. For example, our self-report measure enquired about ST in the past 28 days, whereas objective measures were collected over a 7-day period. Nevertheless, one might expect TV viewing and occupational ST (the likely major contributors to total ST) to be relatively stable over time, particularly within the same season.[34] It is possible that self-reported TV time is able to better capture prolonged periods of sitting compared with the currently available methods of objective assessment. Sitting, rather than standing, is thought to be potentially detrimental to cardio-metabolic health,[35] but this idea is far from verified as it is only based on animal models.[36,37] Like in previous studies,[3,32] the accelerometry device participants wore could not distinguish between sitting and standing and one could argue that this has diluted the associations between objectively assessed ST and risk factors. However, it is far more likely that low accelerometer counts will represent sitting rather than standing (especially in office-based occupations[38,39]) as few occupations involve standing still for more than a few seconds. According to the UK Time Use Survey 2005,[40] the activities taking up most adults' waking time are employment, housework, watching TV/DVD, travelling and eating.[40] Despite the fact that accelerometers and self-report differ in their ability to capture MVPA, both of these MVPA measures were associated with cardio-metabolic risk in our study (data not shown), providing further evidence that methodological differences are unlikely to explain all of the differential associations of self-reported and accelerometer-assessed ST with these risk factors.

The key strength of this study is that, to our knowledge, it is the first to examine the associations between both self-reported and objectively assessed ST and cardio-metabolic risk markers in a sample of working age adults. Our study is one of the largest to date looking at the associations of objectively assessed ST with cardio-metabolic risk. We were able to adjust for objectively measured MVPA in our analyses, and the sample is roughly representative of the population of England, which adds to the ecological validity of our results. The main limitation of this study is the cross-sectional design, which precludes us from making any inferences about causality or examining the meaning of the attenuation of associations after adjustment for BMI/waist.