Correlates of sitting time in adults with type 2 diabetes

Greater use of modern technology and electronic entertainment has contributed to reductions in physical movement [1]. Adults with established type 2 diabetes are characterized by low levels of physical activity with step counts falling in the low active category according to the cut-offs proposed by Tudor-Locke and Bassett [2, 3]. We have previously determined pedometer-assessed daily step counts to average 5365 steps/day (Standard deviation, SD = 2655) in this clinical population, with higher step counts during the spring/summer seasons compared to the fall/winter seasons [4]. We have also determined higher step counts to be associated with lower blood pressure [5], A1C, adiposity measures [6], consistent with the impact of pedometer-based interventions [7] and a recent longitudinal analysis demonstrating the effects of higher step counts on reductions in cardiovascular mortality [8].

While increasing step counts is clearly an imperative in adults with type 2 diabetes, an area that is emerging as an additional issue is sitting time. Sedentary and sitting time have been linked not only with the development of type 2 diabetes and cardiovascular disease [9, 10], but also with both cardiovascular-specific and all-cause mortality (greatest compared to lowest sedentary time pooled hazard ratio 1.9, 95 % confidence interval (CI) 1.4, 2.7 [9] and >7 h/week compared to <1 h/day of television watching hazard ratio 1.5, 95 % CI 1.1, 2.0, [11]). Although there is between-country variation, average sitting time in adult populations, measured by self-report, is approximately 300 to 360 min per day (i.e., 5 to 6 h) [12, 13].

Both high sitting time and low step counts are thus undesirable behaviours for optimal health. It is not clear whether those with low step counts are the same group of individuals with high sitting time. Clarification of this issue would help to strengthen the developing of health behaviour-enhancing interventions aiming to reduce vascular disease risk in type 2 diabetes. To effectively address this newly-emerging vascular risk factor, it would be helpful to identify the sitting time predictors and correlates, particularly in groups at increased risk for mortality, such as people with type 2 diabetes [14].

Therefore, the aims of this study were, in adults with type 2 diabetes, 1) to quantify overall daily sitting time and variation across seasons, 2) to compare sitting and step counts in this clinical population, and 3) to identify the socio-demographic, contextual and health-related factors of self-reported sitting time in this population. We sought to determine if there was an inverse relationship between sitting and pedometer-assessed step counts and if there was seasonal variation in sitting time similar to that previously observed for step counts.

This study is a secondary analysis of a prospective cohort study of adults (n = 201) with physician-diagnosed type 2 diabetes recruited through McGill University-affiliated outpatient clinics conducted in Montréal (Canada). Following informed consent, participants underwent consecutive seasonal assessments (i.e., 4 assessments) over a one-year follow-up period between June 2006 and June 2009. Complete description of the protocol can be found elsewhere but important elements are described herein [4, 15]. Procedures were approved by McGill University’s Faculty of Medicine Institutional Review Board and all participating institutions (McGill University Health Centre Research Ethics Board, Research Ethics Committee at the Jewish General Hospital and the Comité d'éthique de la recherche du Centre de santé et de services sociaux de la Montagne). The present analysis included data on 198 participants with at least one self-reported value for sitting (i.e., at one of the 4 assessments).

In our study sample of adults with type 2 diabetes, an average of over 5 h of sitting time/day was reported. While university-educated individuals reported sitting 55 min/day more than other individuals, immigrants reported 56 min/day less than non-immigrants. There was no seasonal variation in sitting time observed. Further, there was no important association between self-reported sitting time and pedometer-assessed step counts. Therefore, strategies aimed at reducing sitting time appear to require targeting a subgroup that differs somewhat from those with lower step counts. The absence of association differed from a previous study among young adult workers among whom there was a negative association between self-reported sitting and walking [22].

Studies in other groups (e.g. Canadian Community Health Survey, Epic Norfolk) [23, 24] have largely focused on television and screen viewing as metrics of sedentary time. Older age, being unemployed or retired, higher body mass index and depressive symptomology have been shown to be associated with more television viewing [25]. However, increased sitting time may be related to work and transport-related factors, not just television viewing; less is known about overall sitting time and its correlates [25]. The amount of sitting time reported in our type 2 diabetes cohort is similar to that reported in population-based multi-country studies using self-reported measures [12, 13]. We further assessed for differences across seasons. In contrast to step counts (758 fewer steps/day in fall/winter compared to spring/summer) [4], we demonstrated that sitting time did not vary importantly across seasons. Thus, the seasonal reduction in steps does not appear to directly lead to an increase in overall sitting time. One possibility that we considered was that the absence of seasonal variation in sitting time was the result of using self-report methods. However, with respect to physical activity, our analyses demonstrated seasonal variation not only for objectively-assessed step counts but also for self-reported walking, as captured through the IPAQ-SV. Both measures of walking showed important increases during spring/summer compared to fall/winter.

Accordingly, correlates of reported sitting time differ from those associated with steps [26, 27]. Indeed, it has been demonstrated that interventions aimed at increasing physical activity do not necessarily reduce sitting time [28]. A recent systematic review also reported a weak to moderate inverse relationship between sedentary behaviors in general and physical activity [29]. Further, even among those achieving the recommended 150 min/week of physical activity, sitting more than 8 h/day still confers an elevated risk of all-cause mortality [30]. Thus, it appears that both low levels of physical inactivity and high amounts of sitting time need to be separately targeted to achieve maximum health benefit [31].

Given that the IPAQ-SV queries overall sitting time on weekdays, a large part of self-reported sitting time may be work-related. High educational level is often associated with professional office setting environments (i.e., so-called “white-collar” jobs). These are in contrast to “blue-collar” jobs that are generally more physically demanding. The higher time spent sitting observed among people with a university degree may thus reflect the type of job they have. Consistent with our findings, low education status was negatively related to self-reported weekday sitting time among young adult Australian women (unstandardized B −1.23, 95 % CI −1.38, −1.09) [32]. Similarly, in this Australian study, people with a blue-collar occupation and those working less than 35 h per week reported sitting less during the week [32]. However, given that 43 % of our cohort was retired or unemployed, having a university degree in some individuals was likely related to non work-related activities that increased sitting time.

Immigrants demonstrated lower sitting time, even after adjustment for work status, educational level and age. Recent immigrants are twice as likely as Canadian-born Montréal residents to use public transit in their daily commute [33]. Compared to Canadian born individuals, immigrants are more likely to work in processing and manufacturing industries and sales and service occupations [34]. These factors may in part explain the lower daily sitting time observed in this group of individuals.

Although having a car was associated with increased sitting time (92.0 min/day, 95 % CrI 6.7, 179.2) in univariate analysis, we did not include this variable in the multivariate analysis because data were available for only 44 % of our sample. This one and half hour increase in self-reported sitting time in this subgroup is nonetheless substantial. In addition to university education and not being an immigrant, in univariate analyses, younger age, currently working, high annual income, depressed mood, and higher waist circumference were all associated with higher sitting time; further, male sex, not being in a married or common-law relationship, and shorter diabetes duration showed trends towards associations with more sitting time. In multivariate analyses, associations of sitting time with university education and not being an immigrant were confirmed. There were trends suggestive of associations with younger age and currently working, but these were not conclusive.

The IPAQ is the most widely used instrument to assess sitting at the population level and thus its use facilitates comparison with other studies. We acknowledge that the IPAQ-SV does not distinguish between the different domains of sitting, such as work-related, transportation, home-related, and leisure-time nor does it provide information on extended sitting time versus sitting occurring in short bouts. Moreover it queries only the information about week days in contrast to the IPAQ long version. Both the short and long versions are subject to recall bias in contrast to objective accelerometry measures. Healy and colleagues, in a review of validation studies, reported low-to-moderate correlations between self-reported sitting and accelerometer-assessed sedentary time. The IPAQ-SV tends to underestimate sitting time when compared to accelerometer-assessed sedentary time [35]. While this may partly be the result of recall bias, it may also be due to the fact that IPAQ specifically queries sitting, a subset of sedentary behaviour, while accelerometry captures overall sedentary behaviors and not specifically sitting. The French version of the IPAQ long form has been validated in a cohort of diabetes patients (n = 143; 60.9 SD 10.5 years of age) and showed similar results to non-clinical populations [36]. Despite the limitations of using the IPAQ-SV in our study, it does allow us to specifically focus on the issue of sitting and its distinction from low step counts in type 2 diabetes.

Our analyses did not demonstrate a relationship between sitting time and clinical variables; however, there is evidence from other larger studies of such a relationship [9, 10]. In our study, adults with type 2 diabetes reported an average sitting time similar to the general population that was without variation across seasons. Socio-demographic factors, such as having completed a university degree and being a non-immigrant, were linked to self-reported sitting time. There was no correlation between sitting time and step counts. Therefore, high sitting time and low step counts are behaviours that may need to be independently targeted. Interventions to reduce sitting time could include home and work environments designed to allow tasks to be completing from a standing position (e.g., standing desks, standing-only office meetings/gatherings or standing up every hour to break prolonged sitting) [37].