- Email: [email protected]
Sedentary work—Associations between five-year changes in occupational sitting time and body mass index
Preventive Medicine 73 (2015) 1–5
Contents lists available at ScienceDirect
Preventive Medicine journal homepage: www.elsevier.com/locate/ypmed
Sedentary work—Associations between five-year changes in occupational sitting time and body mass index Dorte Eriksen a,⁎, Susanne Rosthøj b, Hermann Burr c, Andreas Holtermann a a b c
National Research Centre for the Working Environment, Copenhagen, Denmark Section of Biostatistics, Department of Public Health, University of Copenhagen, Copenhagen, Denmark Bundesanstalt für Arbeitsschutz und Arbeitsmedizin (BAuA), Germany
a r t i c l e
i n f o
Available online 6 January 2015 Keywords: Public health Work Sedentary behavior Occupational sitting Change in occupational sitting time Body mass index
a b s t r a c t Objective. The aim of this study is to investigate the association between five-year changes in occupational sitting and body mass index (BMI) in working adults. Methods. We analyzed data from The Danish Work Environment Cohort Study (2005 and 2010, n = 3.482). Data on occupational sitting, weight, height and several potential confounders were self-reported. The association between change in occupational sitting (hours) (categorized as large decrease b−7.5, moderate decrease −7.5 to b −2.5, no change −2.5 to 2.5, moderate increase N2.5 to 7.5 and large increase N7.5) and change in BMI was explored by multiple linear regression analyses. Results. 43.0% men and 36.1% women had high occupational sitting time (≥25 h per week) at baseline. 31.8% men and 27.2% women decreased while 30.0% men and 33.0% women increased occupational sitting. The proportion of obese (BMI ≥ 30) increased almost 3% for both genders. BMI changed 0.13 (CI: 0.06; 0.20, p = 0.0003), per category of change in occupational sitting in women, but no association was found in men. Conclusion. In women, there is a positive association between five-year changes in occupational sitting and BMI. © 2015 Elsevier Inc. All rights reserved.
Introduction The prevalence of overweight and obesity has been strongly increasing the last decades (Haslam and James, 2005; Kelly et al., 2008). Because overweight and obesity increase the risk for several diseases such as diabetes, cancer and cardiovascular disease, it is considered to be among the largest challenges for public health worldwide (Haslam and James, 2005; Kelly et al., 2008). A contributing factor for the obesity epidemic could be, that work has changed from mainly involving different forms of physical activity to becoming mainly sedentary for a high proportion of the working population (Stamatakis et al., 2007; Juneau and Potvin, 2010; Allman-Farinelli et al., 2010; Church et al., 2011). Because adults often spend more than half of their waking hours at work (Tudor-Locke et al., 2011), and sedentary behavior is characterized by a low energy expenditure, this change has potential effects on obesity (Pate et al., 2008; Church et al., 2011), and, more documentation for changes in body weight from changes in occupational sitting is needed (Hamilton et al., 2007). One prospective study found a positive association between sedentary work and body mass index (BMI) (Hu et al., 2003). However, other prospective studies found no associations ⁎ Corresponding author at: National Research Centre for the Working Environment, Lersø Parkallé 105, DK-2100 Copenhagen Ø, Denmark. E-mail address: [email protected] (D. Eriksen).
http://dx.doi.org/10.1016/j.ypmed.2014.12.038 0091-7435/© 2015 Elsevier Inc. All rights reserved.
between sedentary work and BMI (van Uffelen et al., 2010). The strongest epidemiological evidence for the positive health effects of physical activity is based on studies investigating associations between changes in physical activity and risk for impaired health and diseases (Bravata et al., 2007). However, no previous studies have to our knowledge investigated if changes in sedentary working time are associated with changes in BMI. The main aim of this study was to investigate the association between changes in occupational sitting time and BMI in a representative sample of the working population in Denmark, testing the a priori hypothesis of a positive association between these variables.
Methods Participants/study population This study used data from The Danish Work Environment Cohort Study (DWECS), consisting of a representative sample of the Danish working population aged 18–59 years at entry. In 2005 the survey used a combination of (postalor internet based) questionnaire (90%) and telephone interview (10%) among a representative sample of 20,000 workers with a response rate of 63%. In 2010 the survey was conducted using (postal- or internet based) questionnaire only, and achieved a response rate of 53% of a representative sample of 30,000 workers. This study includes data from the 4732 working respondents who participated in both 2005 and 2010.
2
D. Eriksen et al. / Preventive Medicine 73 (2015) 1–5
Occupational sitting Participants reported their amount of occupational sitting time by answering the question ‘Does your work imply sitting?’ with six response categories (Almost all the time, approximately ¾ of the time, approximately ½ of the time, approximately ¼ of the time, rarely/very little or never). Similar categorical measures of occupational sitting have been used previously (Choi et al., 2010; Stock et al., 2005). Accounting for participants having varying number of working hours, the categories of sitting time were calculated into sitting hours, based on self-reported weekly working hours achieved by response to the questions ‘How many hours a week do you work in your main job, including possible extra hours?’ and ‘How many hours a week do you normally work in your secondary job?’ The total amount of working hours per week was then multiplied by the constant for occupational sitting, corresponding to the previously mentioned categorical occupational sitting time (0.875, 0.750, 0.500, 0.250, 0.125 and 0.000). The change in occupational sitting time from 2005 to 2010 was calculated by subtracting occupational sitting in 2005 from occupational sitting in 2010, and defined on an ordinal scale with five categories of weekly change (hours) (large decrease b−7.5, moderate decrease = −7.5 to b −2.5, no change = −2.5 to 2.5, moderate increase N2.5 to 7.5 and large increase N7.5). Moreover, the baseline weekly occupational sitting hours was defined on an ordinal scale with three categories (low: ≤10, moderate: N 10 to b 25 and high: ≥25).
reported health were excluded (n = 182). Moreover respondents with extreme self-reported weight change from 2005 to 2010 greater than ±95 kg (n = 4), self-reported height difference from 2005 to 2010 greater than ±10 cm (n = 31) and missing values on exposure, outcome or covariates (n = 669) were excluded. Therefore, 3482 respondents remained in these primary statistical analyses. The association between baseline category of occupational sitting time and both baseline BMI and changes in BMI from 2005 to 2010 was analyzed using multiple linear regression models based on the same adjustments as described above (model 1–3). In these analyses, the study population is slightly larger (3544), as those with missing values for occupational sitting time in 2010 are not excluded. Also, sensitivity analyses excluding respondents with self-reported height difference greater than ±5 cm (n = 74) were done. The robustness of the results was examined by additional analyses considering the same regression models as described above based on the study populations not excluding those with extreme values of working hours (n = 3.846 and n = 4.061, respectively). We present the results as differences with 95% confidence intervals. P values were two-sided and considered significant if below 5%. Due to multiple testing in the comparison of the different categories of change in occupational sitting hours, the confidence intervals and p-values in these comparisons were corrected by the Sidák procedure. We analyzed the data using SAS V9.2. (SAS institute Inc., Cary, NC, USA).
Body mass index and sociodemographic variables BMI was calculated as weight/height2 from self-reported measures on height and weight. Height was calculated as mean of self-reported height in 2005 and 2010. BMI was categorized into: Underweight: BMI b 18.5, normal weight: 18.5 ≤ BMI b 25, overweight: 25 ≤ BMI b 30 and obesity: BMI ≥ 30 (WHO, 2006). The change in BMI was calculated as the subtraction of BMI in 2005 from BMI in 2010. Sex and age were obtained from the personal identification code register in Denmark. Age in years was categorized in five (b30, 30– 39, 40–49, 50–59 and ≥60). Socioeconomic status (SES) was defined by selfreported years of vocational training, which is considered a robust measure (Brønnum-Hansen et al., 2004), and provided in four categories (0, b3, 3–4 and N4). Leisure time physical activity and other lifestyle variables Leisure time physical activity in 2005 was self-reported to the question ‘In the past year, which description fits best regarding your leisure time physical activity?’, with four response categories combining intensity, frequency and volume of exercising, previously used in other studies and shown to be strongly related with cardiovascular disease and mortality (Larsson et al., 2012; Holtermann et al., 2012). Smoking was self-reported with 3 response categories (smoker, ex-smoker and never). Job seniority was measured by self-reported years in the same job and categorized (b 5, 5–10, N10–20 and N20). The respondents were classified according to mode of data collection in 2005, being either postal- or internet based questionnaire or telephone interview (Holtermann et al., 2012). Statistical analysis To examine the difference between subgroups defined by the demographic variables the Kruskal–Wallis test was applied for quantitative variables and the Pearson chi-square test for categorical variables. The association between categories of change of occupational sitting time from 2005 to 2010 and change in BMI during the same time interval was analyzed with multiple linear regression models. Adjustments were made for: model 1: age, model 2: model 1 + smoking, baseline BMI, leisure time physical activity, response method, and job seniority and model 3: model 2 + vocational training (SES). All analyses were stratified by gender, because interactions between gender and occupational sitting as well as several of the adjustment variables were found in analyses based on the total sample. If a monotone association between category of change of occupational sitting as a categorical variable and BMI was seen, the log-likelihood test was performed to investigate if category of change of occupational sitting could be included as a quantitative variable in the linear regression model. Due to extreme values in some of the participants working hours, the study population was restricted to respondents working 20–60 h a week (n = 364 excluded). Because disease can influence weight change and bias the results (Hannerz et al., 2004), respondents with bad or very bad self-
Table 1 Characteristics of the study population for each sex separately, numbers (n), percent (%). N = 3482. The Danish Work Environment Cohort (NAK), 2005. Males
Females
(n = 1679, 48.2%)
(n = 1803, 51.8%)
Confounders
n
%
n
%
Age (years) 18–29 30–39 40–49 50–59 ≥60
164 429 587 470 29
9.8 25.6 35.0 28.0 1.7
143 497 693 459 11
7.9 27.6 38.4 25.5 0.6
Baseline BMI Underweight (b18.5) Normal weight (18.5–b25) Overweight (25–b30) Obesity (≥30)
5 781 715 178
0.3 46.5 42.6 10.6
51 1198 416 138
2.8 66.4 23.1 7.7
Physical activity in leisure time Competition level Strenuous Light Inactive
61 437 928 253
3.6 26.0 55.3 15.1
27 360 1182 234
1.5 20.0 65.6 13.0
Smoking status Smoker Ex-smoker Never
430 464 785
25.6 27.6 46.8
457 514 832
25.4 28.5 46.2
Job seniority (years) b5 5–10 N10–20 N20
455 391 441 392
27.1 23.3 26.3 23.4
452 416 514 421
25.1 23.1 28.7 23.4
Questionnaire method Postal or internet based questionnaire Telephone interview
1347 332
80.2 19.8
1493 310
82.8 17.2
SES (years of vocational training) None Short (b3) Medium length (3–4) Long (N4)
184 796 354 345
11.0 47.4 21.1 20.6
149 789 659 206
8.3 43.8 36.6 11.4
Occupational sitting time (hours per week) Low (≤10) Moderate (N10–b25) High (≥25)
607 350 722
36.2 20.9 43.0
683 469 651
37.9 26.0 36.1
D. Eriksen et al. / Preventive Medicine 73 (2015) 1–5
sitting in 2005 and BMI for both genders as in the primary analyses (results not shown).
Table 2 Characteristics of the study population stratified by BMI and sex, number (n) and percent (%). N = 3482. The Danish Work Environment Cohort (NAK), 2005 to 2010. Males (n = 1679; 48.2%) 2005
Discussion
Females (n = 1803; 51.8%)
2010
2005
3
This study is to our knowledge the first to examine the association between five-year changes in occupational sitting time and BMI. The hypothesis of a positive association between changes in sitting time and BMI was confirmed for women, but not for men. The results suggest that for women, change in occupational sitting time is associated with a change in BMI. Specifically, BMI increased with 0.13 for each category of increase in occupational sitting time from 2005 to 2010. In comparison, Pereira found a five year increase in BMI of 0.33 among men and women sitting at work at baseline for 2–3 h/day compared to those sitting 0–1 h/day, but no trend for work sitting and BMI change was found (Pinto Pereira and Power, 2013). The different effect size of sitting time could be due to a different study population (i.e. mid-adulthood British men and women, whereas our study included a wider age interval) or methodological differences like estimation of sitting duration. However, both studies support that occupational sitting is positively associated with BMI among women. The causes for the lacking positive association between occupational sitting and BMI among men are unknown. In principle, the reduced energy expenditure with increasing sitting time ought to lead to comparable increases in BMI among both genders. A potential explanation can be BMI as an outcome. BMI is criticized for being a poor indicator of body composition, because it does not distinguish between fat and muscle mass. Men have higher muscle mass than women, and extensive sedentary behavior is likely to reduce this (Pace et al., 1976; Ferrando et al., 1999) while increasing fat percentage with no significant influence on BMI (Kwasniewska et al., 2014), which may explain the observed lacking association between sitting time and BMI among men. Therefore, better measures of body composition like fat percentage is recommended for future studies on the association between changes in sitting time and body composition. We found no association between baseline occupational sitting time and prospective change in BMI in either men or women working 20– 60 h a week. This is consistent with previous prospective studies (van Uffelen et al., 2010), and may explain the importance of analyzing changes in occupational sitting. In accordance with other studies, our study found high prevalence of high categories of occupational sitting among both genders (Brown et al., 2003; Jans et al., 2007; Mummery et al., 2005). Other studies have found participants do not compensate for much occupational sitting with less leisure time sitting (Chau et al., 2012; Jans et al., 2007). Moreover, work occupies approximately half
2010
BMI
n
%
n
%
n
%
N
%
Underweight (b18.5) Normal weight (18.5–b25) Overweight (25–b30) Obesity (≥30)
5 781 715 178
0.3 46.5 42.6 10.6
4 683 772 220
0.2 40.7 46.0 13.1
51 1198 416 138
2.8 66.4 23.1 7.7
41 1079 493 190
2.3 60.0 27.3 10.5
Results Table 1 presents an overview of baseline characteristics of the study population stratified by sex. Men generally have a higher prevalence of high occupational sitting (≥25 h/week), overweight and obesity than women. Table 2 shows the distribution of the participants in the four BMIcategories in 2005 and 2010, stratified by sex. The prevalence of overweight and obesity increased among both sexes. Table 3 presents the results for men and women based on the linear regression analysis on the association between changes in occupational sitting time and change in BMI. For men, no tendency for an association between category of change in occupational sitting and BMI was observed. For women, a significant positive trend was seen between change in occupational sitting and BMI. Compared to the group with large increase in sitting time (reference), a positive association with BMI was found in the groups with large decrease and moderate decrease in occupational sitting. Correcting the results by the Sidák procedure did not alter this finding. Moreover, a positive association between occupational sitting time as a quantitative variable and BMI was found in women. Each unit of increase in occupational sitting category was associated with an increase in BMI. Change in BMI (CI) = 0.13 (0.06; 0.20), p = 0.0003 in model 3 (results not shown). No associations were found between categories of occupational sitting time in 2005 and both baseline BMI (not shown) and change in BMI from 2005 to 2010 for men and women respectively (Table 4). Sensitivity analyses excluding respondents with self-reported height difference greater than ±5 cm showed similar results (Tables 3 and 4). The additional analyses involving all study participants, independent of numbers of working hours, found a similar pattern both between changes in occupational sitting from 2005 to 2010, and occupational
Table 3 Association between change in occupational sitting time and change in BMI (ΔBMI). N = 3482. The Danish Work Environment Cohort (NAK), 2005 to 2010. Crude (Model 1)a ΔBMI Males Large decrease Moderate decrease No change Moderate increase Large increase Females Large decrease Moderate decrease No change Moderate increase Large increase
CI
Adjusted (Model 2)b P
*
ΔBMI
CI
Further adjusted (Model 3)c P
0.9342 0.04 0.00 −0.04 −0.06 0
−0.27 0.31 −0.22 0.27 −0.27 0.19 −0.27 0.21 Reference
0.7563 0.9750 0.7177 0.6685
−0.46 −0.50 −0.24 −0.06 0
−0.80 −0.12 −0.80 −0.19 −0.50 0.01 −0.36 0.25 Reference
0.0077 0.0016 0.0629 0.7092
*
ΔBMI
CI
P
0.9169 0.04 0.01 −0.04 −0.07 0
−0.22 0.31 −0.25 0.28 −0.26 0.19 −0.34 0.20 Reference
0.7415 0.9137 0.7574 0.6166
−0.46 −0.48 −0.23 −0.08 0
−0.79 −0.12 −0.79 −0.18 0.03 −0.48 −0.38 0.22 Reference
0.0073 0.0018 0.0811 0.0671
0.0038
* 0.8839
0.04 0.02 −0.05 −0.07 0
−0.22 0.31 −0.25 0.28 −0.27 0.18 −0.34 0.20 Reference
0.7446 0.9088 0.6724 0.6013
−0.43 −0.50 −0.25 −0.10 0
−0.77 −0.10 −0.81 −0.20 −0.51 0.00 −0.40 0.20 Reference
0.0108 0.0010 0.0502 0.4961
0.0054
0.0058
ΔBMI: Change in BMI from 2005 to 2010. CI: 95% confidence interval. P: Significance level. Bold writing: P b 0.05. *Overall significance level. Large decrease: b−7.5 h. Moderate decrease: −7.5 h to b−2.5 h. No change: −2.5 h to 2.5 h. Moderate increase: N2.5 h to 7.5 h. Large increase: N7.5 h. a Crude analysis: Adjusted for age. b Adjusted for age + baseline BMI, smoking, leisure time physical activity, questionnaire method, and job seniority. c Adjusted as in model 2 + SES.
4
D. Eriksen et al. / Preventive Medicine 73 (2015) 1–5
Table 4 Association between level of baseline occupational sitting time and change in BMI (ΔBMI), N = 3544. The Danish Work Environment Cohort (NAK), 2005 to 2010. Crude (Model 1)a ΔBMI Males Low Moderate High Females Low Moderate High
CI
Adjusted (Model 2)b P
*
ΔBMI
CI
Further adjusted (Model 3)c P
0.4085 0.11 0.06 0.00
−0.05 0.28 −0.13 0.27 Reference
0.1843 0.4977
0.20 0.09 0.00
0.00 0.40 −0.13 0.31 Reference
0.0496 0.4317
*
ΔBMI
CI
P
0.02 0.03 0.00
−0.16 0.20 −0.17 0.23 Reference
0.8383 0.7742
0.18 0.07 0.00
−0.02 0.39 −0.15 0.29 Reference
0.0823 0.5436
0.5268 0.09 0.06 0.00
−0.07 0.26 −0.13 0.26 Reference
0.2663 0.5237
0.22 0.08 0.00
0.02 0.42 −0.14 0.30 Reference
0.0297 0.4711
0.1442
* 0.9554
0.0893
0.2133
ΔBMI: Change in BMI from 2005 to 2010. CI: 95% confidence interval. P: Significance level. *Overall significance level. Low: ≤10 h. Moderate: N10 h to b25 h. High: ≥25 h. a Crude analysis: Adjusted for age. b Adjusted for age + baseline BMI, smoking, leisure time physical activity, questionnaire method, and job seniority. c Adjusted as in model 2 + SES.
the waking hours for working adults (Tudor-Locke et al., 2011). Therefore, the occupational domain is a relevant setting for intervening on extensive sitting (Chau et al., 2012; Owen et al., 2011). Strengths and limitations The main strength of this study is the prospective study design with a large representative sample of Danish workers. The question on occupational sitting time has been used in earlier studies and have shown acceptable validity and reliability (Stock et al., 2005; Andersen et al., 2000; Holtermann et al., 2012; Choi et al., 2010), and the ability to estimate the weekly number of hours sitting at work strengthens the study (van Uffelen et al., 2010). Additional strength is the ability to investigate the association between changes in occupational sitting time and BMI. We adjusted for baseline BMI and other potential confounders. We did not adjust for baseline sitting time. Given the chosen analytical approach in this study, including the treatment of ordinal variables as quantitative, we consider that changes in sitting time are associated with similar changes in BMI independent of baseline. The main limitation is the use of self-reported measures of occupational sitting time and BMI, which may have been influenced by selfreporting bias. Self-reported sitting time is often overestimated (Clark et al., 2011; Juneau and Potvin, 2010), while self-reported BMI is often underestimated (Nyholm et al., 2007). These issues might have led to underestimation of the association between change in occupational sitting time and BMI. It would be a methodological strength for future cohort studies to collect objective data on these variables (LagerstedOlsen et al., 2013). The survey response rate was 63% in 2005 and 53% in 2010. Additionally the study only included individuals, who responded in both years. This might have led to selection biases, if participants were healthier than non-participants. Potentially this might cause underestimation of the association between changes in occupational sitting time and BMI. Moreover some methodological weaknesses are present, like not being able to estimate the energy expenditure, not having a clear definition of smoking and not having information on commuting physical activity, which may have a significant influence on BMI, even independently of occupational and recreational physical activity. We cannot exclude the risk of reversed causation, since the changes in occupational sitting and associated changes in BMI are measured during the same time span. Future studies should focus on prospective designs using objective measures for occupational and leisure time sitting, and might focus on the potential difference between men and women concerning the association between occupational sitting and overweight/obesity. Conclusion We found a positive association between change in occupational sitting time and BMI from 2005 to 2010 in women. More prospective
studies, with objective measures of occupational sitting time and BMI are recommended to investigate the importance of occupational sitting. Moreover, use of better measures of body composition like fat percentage is recommended for future studies on the association between changes in sitting time and body composition. Conflict of interest statement The authors declare that there are no conflicts of interest
Acknowledgment For help with managing the data in SAS: Data manager Ebbe Villadsen, National Research Center for the Working Environment, Copenhagen, Denmark References Allman-Farinelli, M.A., Chey, T., Merom, D., Bauman, A.E., 2010. Occupational risk of overweight and obesity: an analysis of the Australian Health Survey. J. Occup. Med. Toxicol. 5, 14. Andersen, L.B., Schnohr, P., Schroll, M., Hein, H.O., 2000. All-cause mortality associated with physical activity during leisure time, work, sports, and cycling to work. Arch. Intern. Med. 160, 1621–1628. Bravata, D.M., Smith-Spangler, C., Sundaram, V., et al., 2007. Using pedometers to increase physical activity and improve health. A systematic review. JAMA 298, 2296–2304. Brønnum-Hansen, H., Andersen, O., Kjøller, M., Rasmussen, N., 2004. Social gradient in life expectancy and health expectancy in Denmark. Soz. Praventivmed. 49, 36–41. Brown, W.J., Miller, Y.D., Miller, R., 2003. Sitting time and work patterns as indicators of overweight and obesity in Australian adults. Int. J. Obes. Relat. Metab. Disord. 27, 1340–1346. Chau, J.Y., Van Der Ploeg, H.P., Merom, D., Chey, T., Bauman, A.E., 2012. Cross-sectional associations between occupational and leisure-time sitting, physical activity and obesity in working adults. Prev. Med. 54, 195–200. Choi, B., Schnall, P.L., Yang, H., et al., 2010. Sedentary work, low physical job demand, and obesity in US workers. Am. J. Ind. Med. 53, 1088–1101. Church, T.S., Thomas, D.M., Tudor-Locke, C., et al., 2011. Trends over 5 decades in U.S. occupation-related physical activity and their associations with obesity. PLoS One 6, 1–7. Clark, B., Thorp, A., Winkler, E., et al., 2011. Validity of self-report measures of workplace sitting time and breaks in sitting time. Med. Sci. Sports Exerc. 43, 1907–1912. Ferrando, A.A., Stuart, C.A., Sheffield-Moore, M., Wolfe, R.R., 1999. Inactivity amplifies the catabolic response of skeletal muscle to cortisol. J. Clin. Endocrinol. Metab. 84, 3515–3521. Hamilton, M.T., Hamilton, D.G., Zderic, T.W., 2007. Role of low energy expenditure and sitting in obesity, metabolic syndrome, type 2 diabetes, and cardiovascular disease. Diabetes 56, 2655–2667. Hannerz, H., Albertsen, K., Nielsen, M.L., Tuchsen, F., Burr, H., 2004. Occupational factors and 5-year weight change among men in a Danish national cohort. Health Psychol. 23, 283–288. Haslam, D.W., James, W.P., 2005. Obesity. Lancet 366, 1197–1209. Holtermann, A., Hansen, J.V., Burr, H., Sogaard, K., Sjogaard, G., 2012. The health paradox of occupational and leisure-time physical activity. Br. J. Sports Med. 46, 291–295. Hu, F.B., Li, T.Y., Colditz, G.A., Willett, W.C., Manson, J.E., 2003. Television watching and other sedentary behaviors in relation to risk of obesity and type 2 diabetes mellitus in women. JAMA 289, 1785–1791. Jans, M.P., Proper, K.I., Hildebrandt, V.H., 2007. Sedentary behavior in Dutch workers: differences between occupations and business sectors. Am. J. Prev. Med. 33, 450–454. Juneau, C.E., Potvin, L., 2010. Trends in leisure-, transport-, and work-related physical activity in Canada 1994–2005. Prev. Med. 51, 384–386.
D. Eriksen et al. / Preventive Medicine 73 (2015) 1–5 Kelly, T., Yang, W., Chen, C.S., Reynolds, K., He, J., 2008. Global burden of obesity in 2005 and projections to 2030. Int. J. Obes. (Lond.) 32, 1431–1437. Kwasniewska, M., Jegier, A., Kostka, T., et al., 2014. Long-term effect of different physical activity levels on subclinical atherosclerosis in middle-aged men: a 25-year prospective study. PLoS One 9, e85209. Lagersted-Olsen, J., Korshøj, M., Skotte, J., Carneiro, I.G., Søgaard, K., Holtermann, A., 2014. Comparison of objectively measured and self-reported time spent sitting. Int. J. Sports Med. 35, 534–540. Larsson, C.A., Kroll, L., Bennet, L., Gullberg, B., Rastam, L., Lindblad, U., 2012. Leisure time and occupational physical activity in relation to obesity and insulin resistance: a population-based study from the Skaraborg Project in Sweden. Mummery, W.K., Schofield, G.M., Steele, R., Eakin, E.G., Brown, W.J., 2005. Occupational sitting time and overweight and obesity in Australian workers. Am. J. Prev. Med. 29, 91–97. Nyholm, M., Gullberg, B., Merlo, J., Lundqvist-Persson, C., Råstam, L., Lindblad, U., 2007. The validity of obesity based on self-reported weight and height: implications for population studies. Obesity 15, 197–208. Owen, N., Sugiyama, T., Eakin, E.E., Gardiner, P.A., Tremblay, M.S., Sallis, J.F., 2011. Adults' sedentary behavior determinants and interventions. Am. J. Prev. Med. 41, 189–196.
5
Pace, N., Kodama, A.M., Price, D.C., Grunbaum, B.W., Rahlmann, D.F., Newsom, B.D., 1976. Body composition changes in men and women after 2–3 weeks of bed rest. Life Sci. Space Res. 269–274. Pate, R.R., O'Neill, J.R., Lobelo, F., 2008. The evolving definition of “sedentary”. Exerc. Sport Sci. Rev. 36, 173–178. Pinto Pereira, S.M., Power, C., 2013. Sedentary behaviours in mid-adulthood and subsequent body mass index. PLoS One 8, e65791. Stamatakis, E., Ekelund, U., Wareham, N.J., 2007. Temporal trends in physical activity in England: the Health Survey for England 1991 to 2004. Prev. Med. 45, 416–423. Stock, S., Fernandes, R., Delisle, A., Vézina, N., 2005. Reproducibility and validity of workers' self-reports of physical work demands. Scand. J. Work Environ. Health 6, 409–437. Tudor-Locke, C., Leonardi, C., Johnson, W.D., Katzmarzyk, P.T., 2011. Time spent in physical activity and sedentary behaviors on the working day: the American time use survey. J. Occup. Environ. Med. 53, 1382–1387. Van Uffelen, J.G., Wong, J., Chau, J.Y., et al., 2010. Occupational sitting and health risks: a systematic review. Am. J. Prev. Med. 39, 379–388. Who, 2006. Database on BMI [Online]. Available: http://apps.who.int/bmi/index.jsp? introPage=intro_3.html ([Accessed 21.04. 2013]).
Contents lists available at ScienceDirect
Preventive Medicine journal homepage: www.elsevier.com/locate/ypmed
Sedentary work—Associations between five-year changes in occupational sitting time and body mass index Dorte Eriksen a,⁎, Susanne Rosthøj b, Hermann Burr c, Andreas Holtermann a a b c
National Research Centre for the Working Environment, Copenhagen, Denmark Section of Biostatistics, Department of Public Health, University of Copenhagen, Copenhagen, Denmark Bundesanstalt für Arbeitsschutz und Arbeitsmedizin (BAuA), Germany
a r t i c l e
i n f o
Available online 6 January 2015 Keywords: Public health Work Sedentary behavior Occupational sitting Change in occupational sitting time Body mass index
a b s t r a c t Objective. The aim of this study is to investigate the association between five-year changes in occupational sitting and body mass index (BMI) in working adults. Methods. We analyzed data from The Danish Work Environment Cohort Study (2005 and 2010, n = 3.482). Data on occupational sitting, weight, height and several potential confounders were self-reported. The association between change in occupational sitting (hours) (categorized as large decrease b−7.5, moderate decrease −7.5 to b −2.5, no change −2.5 to 2.5, moderate increase N2.5 to 7.5 and large increase N7.5) and change in BMI was explored by multiple linear regression analyses. Results. 43.0% men and 36.1% women had high occupational sitting time (≥25 h per week) at baseline. 31.8% men and 27.2% women decreased while 30.0% men and 33.0% women increased occupational sitting. The proportion of obese (BMI ≥ 30) increased almost 3% for both genders. BMI changed 0.13 (CI: 0.06; 0.20, p = 0.0003), per category of change in occupational sitting in women, but no association was found in men. Conclusion. In women, there is a positive association between five-year changes in occupational sitting and BMI. © 2015 Elsevier Inc. All rights reserved.
Introduction The prevalence of overweight and obesity has been strongly increasing the last decades (Haslam and James, 2005; Kelly et al., 2008). Because overweight and obesity increase the risk for several diseases such as diabetes, cancer and cardiovascular disease, it is considered to be among the largest challenges for public health worldwide (Haslam and James, 2005; Kelly et al., 2008). A contributing factor for the obesity epidemic could be, that work has changed from mainly involving different forms of physical activity to becoming mainly sedentary for a high proportion of the working population (Stamatakis et al., 2007; Juneau and Potvin, 2010; Allman-Farinelli et al., 2010; Church et al., 2011). Because adults often spend more than half of their waking hours at work (Tudor-Locke et al., 2011), and sedentary behavior is characterized by a low energy expenditure, this change has potential effects on obesity (Pate et al., 2008; Church et al., 2011), and, more documentation for changes in body weight from changes in occupational sitting is needed (Hamilton et al., 2007). One prospective study found a positive association between sedentary work and body mass index (BMI) (Hu et al., 2003). However, other prospective studies found no associations ⁎ Corresponding author at: National Research Centre for the Working Environment, Lersø Parkallé 105, DK-2100 Copenhagen Ø, Denmark. E-mail address: [email protected] (D. Eriksen).
http://dx.doi.org/10.1016/j.ypmed.2014.12.038 0091-7435/© 2015 Elsevier Inc. All rights reserved.
between sedentary work and BMI (van Uffelen et al., 2010). The strongest epidemiological evidence for the positive health effects of physical activity is based on studies investigating associations between changes in physical activity and risk for impaired health and diseases (Bravata et al., 2007). However, no previous studies have to our knowledge investigated if changes in sedentary working time are associated with changes in BMI. The main aim of this study was to investigate the association between changes in occupational sitting time and BMI in a representative sample of the working population in Denmark, testing the a priori hypothesis of a positive association between these variables.
Methods Participants/study population This study used data from The Danish Work Environment Cohort Study (DWECS), consisting of a representative sample of the Danish working population aged 18–59 years at entry. In 2005 the survey used a combination of (postalor internet based) questionnaire (90%) and telephone interview (10%) among a representative sample of 20,000 workers with a response rate of 63%. In 2010 the survey was conducted using (postal- or internet based) questionnaire only, and achieved a response rate of 53% of a representative sample of 30,000 workers. This study includes data from the 4732 working respondents who participated in both 2005 and 2010.
2
D. Eriksen et al. / Preventive Medicine 73 (2015) 1–5
Occupational sitting Participants reported their amount of occupational sitting time by answering the question ‘Does your work imply sitting?’ with six response categories (Almost all the time, approximately ¾ of the time, approximately ½ of the time, approximately ¼ of the time, rarely/very little or never). Similar categorical measures of occupational sitting have been used previously (Choi et al., 2010; Stock et al., 2005). Accounting for participants having varying number of working hours, the categories of sitting time were calculated into sitting hours, based on self-reported weekly working hours achieved by response to the questions ‘How many hours a week do you work in your main job, including possible extra hours?’ and ‘How many hours a week do you normally work in your secondary job?’ The total amount of working hours per week was then multiplied by the constant for occupational sitting, corresponding to the previously mentioned categorical occupational sitting time (0.875, 0.750, 0.500, 0.250, 0.125 and 0.000). The change in occupational sitting time from 2005 to 2010 was calculated by subtracting occupational sitting in 2005 from occupational sitting in 2010, and defined on an ordinal scale with five categories of weekly change (hours) (large decrease b−7.5, moderate decrease = −7.5 to b −2.5, no change = −2.5 to 2.5, moderate increase N2.5 to 7.5 and large increase N7.5). Moreover, the baseline weekly occupational sitting hours was defined on an ordinal scale with three categories (low: ≤10, moderate: N 10 to b 25 and high: ≥25).
reported health were excluded (n = 182). Moreover respondents with extreme self-reported weight change from 2005 to 2010 greater than ±95 kg (n = 4), self-reported height difference from 2005 to 2010 greater than ±10 cm (n = 31) and missing values on exposure, outcome or covariates (n = 669) were excluded. Therefore, 3482 respondents remained in these primary statistical analyses. The association between baseline category of occupational sitting time and both baseline BMI and changes in BMI from 2005 to 2010 was analyzed using multiple linear regression models based on the same adjustments as described above (model 1–3). In these analyses, the study population is slightly larger (3544), as those with missing values for occupational sitting time in 2010 are not excluded. Also, sensitivity analyses excluding respondents with self-reported height difference greater than ±5 cm (n = 74) were done. The robustness of the results was examined by additional analyses considering the same regression models as described above based on the study populations not excluding those with extreme values of working hours (n = 3.846 and n = 4.061, respectively). We present the results as differences with 95% confidence intervals. P values were two-sided and considered significant if below 5%. Due to multiple testing in the comparison of the different categories of change in occupational sitting hours, the confidence intervals and p-values in these comparisons were corrected by the Sidák procedure. We analyzed the data using SAS V9.2. (SAS institute Inc., Cary, NC, USA).
Body mass index and sociodemographic variables BMI was calculated as weight/height2 from self-reported measures on height and weight. Height was calculated as mean of self-reported height in 2005 and 2010. BMI was categorized into: Underweight: BMI b 18.5, normal weight: 18.5 ≤ BMI b 25, overweight: 25 ≤ BMI b 30 and obesity: BMI ≥ 30 (WHO, 2006). The change in BMI was calculated as the subtraction of BMI in 2005 from BMI in 2010. Sex and age were obtained from the personal identification code register in Denmark. Age in years was categorized in five (b30, 30– 39, 40–49, 50–59 and ≥60). Socioeconomic status (SES) was defined by selfreported years of vocational training, which is considered a robust measure (Brønnum-Hansen et al., 2004), and provided in four categories (0, b3, 3–4 and N4). Leisure time physical activity and other lifestyle variables Leisure time physical activity in 2005 was self-reported to the question ‘In the past year, which description fits best regarding your leisure time physical activity?’, with four response categories combining intensity, frequency and volume of exercising, previously used in other studies and shown to be strongly related with cardiovascular disease and mortality (Larsson et al., 2012; Holtermann et al., 2012). Smoking was self-reported with 3 response categories (smoker, ex-smoker and never). Job seniority was measured by self-reported years in the same job and categorized (b 5, 5–10, N10–20 and N20). The respondents were classified according to mode of data collection in 2005, being either postal- or internet based questionnaire or telephone interview (Holtermann et al., 2012). Statistical analysis To examine the difference between subgroups defined by the demographic variables the Kruskal–Wallis test was applied for quantitative variables and the Pearson chi-square test for categorical variables. The association between categories of change of occupational sitting time from 2005 to 2010 and change in BMI during the same time interval was analyzed with multiple linear regression models. Adjustments were made for: model 1: age, model 2: model 1 + smoking, baseline BMI, leisure time physical activity, response method, and job seniority and model 3: model 2 + vocational training (SES). All analyses were stratified by gender, because interactions between gender and occupational sitting as well as several of the adjustment variables were found in analyses based on the total sample. If a monotone association between category of change of occupational sitting as a categorical variable and BMI was seen, the log-likelihood test was performed to investigate if category of change of occupational sitting could be included as a quantitative variable in the linear regression model. Due to extreme values in some of the participants working hours, the study population was restricted to respondents working 20–60 h a week (n = 364 excluded). Because disease can influence weight change and bias the results (Hannerz et al., 2004), respondents with bad or very bad self-
Table 1 Characteristics of the study population for each sex separately, numbers (n), percent (%). N = 3482. The Danish Work Environment Cohort (NAK), 2005. Males
Females
(n = 1679, 48.2%)
(n = 1803, 51.8%)
Confounders
n
%
n
%
Age (years) 18–29 30–39 40–49 50–59 ≥60
164 429 587 470 29
9.8 25.6 35.0 28.0 1.7
143 497 693 459 11
7.9 27.6 38.4 25.5 0.6
Baseline BMI Underweight (b18.5) Normal weight (18.5–b25) Overweight (25–b30) Obesity (≥30)
5 781 715 178
0.3 46.5 42.6 10.6
51 1198 416 138
2.8 66.4 23.1 7.7
Physical activity in leisure time Competition level Strenuous Light Inactive
61 437 928 253
3.6 26.0 55.3 15.1
27 360 1182 234
1.5 20.0 65.6 13.0
Smoking status Smoker Ex-smoker Never
430 464 785
25.6 27.6 46.8
457 514 832
25.4 28.5 46.2
Job seniority (years) b5 5–10 N10–20 N20
455 391 441 392
27.1 23.3 26.3 23.4
452 416 514 421
25.1 23.1 28.7 23.4
Questionnaire method Postal or internet based questionnaire Telephone interview
1347 332
80.2 19.8
1493 310
82.8 17.2
SES (years of vocational training) None Short (b3) Medium length (3–4) Long (N4)
184 796 354 345
11.0 47.4 21.1 20.6
149 789 659 206
8.3 43.8 36.6 11.4
Occupational sitting time (hours per week) Low (≤10) Moderate (N10–b25) High (≥25)
607 350 722
36.2 20.9 43.0
683 469 651
37.9 26.0 36.1
D. Eriksen et al. / Preventive Medicine 73 (2015) 1–5
sitting in 2005 and BMI for both genders as in the primary analyses (results not shown).
Table 2 Characteristics of the study population stratified by BMI and sex, number (n) and percent (%). N = 3482. The Danish Work Environment Cohort (NAK), 2005 to 2010. Males (n = 1679; 48.2%) 2005
Discussion
Females (n = 1803; 51.8%)
2010
2005
3
This study is to our knowledge the first to examine the association between five-year changes in occupational sitting time and BMI. The hypothesis of a positive association between changes in sitting time and BMI was confirmed for women, but not for men. The results suggest that for women, change in occupational sitting time is associated with a change in BMI. Specifically, BMI increased with 0.13 for each category of increase in occupational sitting time from 2005 to 2010. In comparison, Pereira found a five year increase in BMI of 0.33 among men and women sitting at work at baseline for 2–3 h/day compared to those sitting 0–1 h/day, but no trend for work sitting and BMI change was found (Pinto Pereira and Power, 2013). The different effect size of sitting time could be due to a different study population (i.e. mid-adulthood British men and women, whereas our study included a wider age interval) or methodological differences like estimation of sitting duration. However, both studies support that occupational sitting is positively associated with BMI among women. The causes for the lacking positive association between occupational sitting and BMI among men are unknown. In principle, the reduced energy expenditure with increasing sitting time ought to lead to comparable increases in BMI among both genders. A potential explanation can be BMI as an outcome. BMI is criticized for being a poor indicator of body composition, because it does not distinguish between fat and muscle mass. Men have higher muscle mass than women, and extensive sedentary behavior is likely to reduce this (Pace et al., 1976; Ferrando et al., 1999) while increasing fat percentage with no significant influence on BMI (Kwasniewska et al., 2014), which may explain the observed lacking association between sitting time and BMI among men. Therefore, better measures of body composition like fat percentage is recommended for future studies on the association between changes in sitting time and body composition. We found no association between baseline occupational sitting time and prospective change in BMI in either men or women working 20– 60 h a week. This is consistent with previous prospective studies (van Uffelen et al., 2010), and may explain the importance of analyzing changes in occupational sitting. In accordance with other studies, our study found high prevalence of high categories of occupational sitting among both genders (Brown et al., 2003; Jans et al., 2007; Mummery et al., 2005). Other studies have found participants do not compensate for much occupational sitting with less leisure time sitting (Chau et al., 2012; Jans et al., 2007). Moreover, work occupies approximately half
2010
BMI
n
%
n
%
n
%
N
%
Underweight (b18.5) Normal weight (18.5–b25) Overweight (25–b30) Obesity (≥30)
5 781 715 178
0.3 46.5 42.6 10.6
4 683 772 220
0.2 40.7 46.0 13.1
51 1198 416 138
2.8 66.4 23.1 7.7
41 1079 493 190
2.3 60.0 27.3 10.5
Results Table 1 presents an overview of baseline characteristics of the study population stratified by sex. Men generally have a higher prevalence of high occupational sitting (≥25 h/week), overweight and obesity than women. Table 2 shows the distribution of the participants in the four BMIcategories in 2005 and 2010, stratified by sex. The prevalence of overweight and obesity increased among both sexes. Table 3 presents the results for men and women based on the linear regression analysis on the association between changes in occupational sitting time and change in BMI. For men, no tendency for an association between category of change in occupational sitting and BMI was observed. For women, a significant positive trend was seen between change in occupational sitting and BMI. Compared to the group with large increase in sitting time (reference), a positive association with BMI was found in the groups with large decrease and moderate decrease in occupational sitting. Correcting the results by the Sidák procedure did not alter this finding. Moreover, a positive association between occupational sitting time as a quantitative variable and BMI was found in women. Each unit of increase in occupational sitting category was associated with an increase in BMI. Change in BMI (CI) = 0.13 (0.06; 0.20), p = 0.0003 in model 3 (results not shown). No associations were found between categories of occupational sitting time in 2005 and both baseline BMI (not shown) and change in BMI from 2005 to 2010 for men and women respectively (Table 4). Sensitivity analyses excluding respondents with self-reported height difference greater than ±5 cm showed similar results (Tables 3 and 4). The additional analyses involving all study participants, independent of numbers of working hours, found a similar pattern both between changes in occupational sitting from 2005 to 2010, and occupational
Table 3 Association between change in occupational sitting time and change in BMI (ΔBMI). N = 3482. The Danish Work Environment Cohort (NAK), 2005 to 2010. Crude (Model 1)a ΔBMI Males Large decrease Moderate decrease No change Moderate increase Large increase Females Large decrease Moderate decrease No change Moderate increase Large increase
CI
Adjusted (Model 2)b P
*
ΔBMI
CI
Further adjusted (Model 3)c P
0.9342 0.04 0.00 −0.04 −0.06 0
−0.27 0.31 −0.22 0.27 −0.27 0.19 −0.27 0.21 Reference
0.7563 0.9750 0.7177 0.6685
−0.46 −0.50 −0.24 −0.06 0
−0.80 −0.12 −0.80 −0.19 −0.50 0.01 −0.36 0.25 Reference
0.0077 0.0016 0.0629 0.7092
*
ΔBMI
CI
P
0.9169 0.04 0.01 −0.04 −0.07 0
−0.22 0.31 −0.25 0.28 −0.26 0.19 −0.34 0.20 Reference
0.7415 0.9137 0.7574 0.6166
−0.46 −0.48 −0.23 −0.08 0
−0.79 −0.12 −0.79 −0.18 0.03 −0.48 −0.38 0.22 Reference
0.0073 0.0018 0.0811 0.0671
0.0038
* 0.8839
0.04 0.02 −0.05 −0.07 0
−0.22 0.31 −0.25 0.28 −0.27 0.18 −0.34 0.20 Reference
0.7446 0.9088 0.6724 0.6013
−0.43 −0.50 −0.25 −0.10 0
−0.77 −0.10 −0.81 −0.20 −0.51 0.00 −0.40 0.20 Reference
0.0108 0.0010 0.0502 0.4961
0.0054
0.0058
ΔBMI: Change in BMI from 2005 to 2010. CI: 95% confidence interval. P: Significance level. Bold writing: P b 0.05. *Overall significance level. Large decrease: b−7.5 h. Moderate decrease: −7.5 h to b−2.5 h. No change: −2.5 h to 2.5 h. Moderate increase: N2.5 h to 7.5 h. Large increase: N7.5 h. a Crude analysis: Adjusted for age. b Adjusted for age + baseline BMI, smoking, leisure time physical activity, questionnaire method, and job seniority. c Adjusted as in model 2 + SES.
4
D. Eriksen et al. / Preventive Medicine 73 (2015) 1–5
Table 4 Association between level of baseline occupational sitting time and change in BMI (ΔBMI), N = 3544. The Danish Work Environment Cohort (NAK), 2005 to 2010. Crude (Model 1)a ΔBMI Males Low Moderate High Females Low Moderate High
CI
Adjusted (Model 2)b P
*
ΔBMI
CI
Further adjusted (Model 3)c P
0.4085 0.11 0.06 0.00
−0.05 0.28 −0.13 0.27 Reference
0.1843 0.4977
0.20 0.09 0.00
0.00 0.40 −0.13 0.31 Reference
0.0496 0.4317
*
ΔBMI
CI
P
0.02 0.03 0.00
−0.16 0.20 −0.17 0.23 Reference
0.8383 0.7742
0.18 0.07 0.00
−0.02 0.39 −0.15 0.29 Reference
0.0823 0.5436
0.5268 0.09 0.06 0.00
−0.07 0.26 −0.13 0.26 Reference
0.2663 0.5237
0.22 0.08 0.00
0.02 0.42 −0.14 0.30 Reference
0.0297 0.4711
0.1442
* 0.9554
0.0893
0.2133
ΔBMI: Change in BMI from 2005 to 2010. CI: 95% confidence interval. P: Significance level. *Overall significance level. Low: ≤10 h. Moderate: N10 h to b25 h. High: ≥25 h. a Crude analysis: Adjusted for age. b Adjusted for age + baseline BMI, smoking, leisure time physical activity, questionnaire method, and job seniority. c Adjusted as in model 2 + SES.
the waking hours for working adults (Tudor-Locke et al., 2011). Therefore, the occupational domain is a relevant setting for intervening on extensive sitting (Chau et al., 2012; Owen et al., 2011). Strengths and limitations The main strength of this study is the prospective study design with a large representative sample of Danish workers. The question on occupational sitting time has been used in earlier studies and have shown acceptable validity and reliability (Stock et al., 2005; Andersen et al., 2000; Holtermann et al., 2012; Choi et al., 2010), and the ability to estimate the weekly number of hours sitting at work strengthens the study (van Uffelen et al., 2010). Additional strength is the ability to investigate the association between changes in occupational sitting time and BMI. We adjusted for baseline BMI and other potential confounders. We did not adjust for baseline sitting time. Given the chosen analytical approach in this study, including the treatment of ordinal variables as quantitative, we consider that changes in sitting time are associated with similar changes in BMI independent of baseline. The main limitation is the use of self-reported measures of occupational sitting time and BMI, which may have been influenced by selfreporting bias. Self-reported sitting time is often overestimated (Clark et al., 2011; Juneau and Potvin, 2010), while self-reported BMI is often underestimated (Nyholm et al., 2007). These issues might have led to underestimation of the association between change in occupational sitting time and BMI. It would be a methodological strength for future cohort studies to collect objective data on these variables (LagerstedOlsen et al., 2013). The survey response rate was 63% in 2005 and 53% in 2010. Additionally the study only included individuals, who responded in both years. This might have led to selection biases, if participants were healthier than non-participants. Potentially this might cause underestimation of the association between changes in occupational sitting time and BMI. Moreover some methodological weaknesses are present, like not being able to estimate the energy expenditure, not having a clear definition of smoking and not having information on commuting physical activity, which may have a significant influence on BMI, even independently of occupational and recreational physical activity. We cannot exclude the risk of reversed causation, since the changes in occupational sitting and associated changes in BMI are measured during the same time span. Future studies should focus on prospective designs using objective measures for occupational and leisure time sitting, and might focus on the potential difference between men and women concerning the association between occupational sitting and overweight/obesity. Conclusion We found a positive association between change in occupational sitting time and BMI from 2005 to 2010 in women. More prospective
studies, with objective measures of occupational sitting time and BMI are recommended to investigate the importance of occupational sitting. Moreover, use of better measures of body composition like fat percentage is recommended for future studies on the association between changes in sitting time and body composition. Conflict of interest statement The authors declare that there are no conflicts of interest
Acknowledgment For help with managing the data in SAS: Data manager Ebbe Villadsen, National Research Center for the Working Environment, Copenhagen, Denmark References Allman-Farinelli, M.A., Chey, T., Merom, D., Bauman, A.E., 2010. Occupational risk of overweight and obesity: an analysis of the Australian Health Survey. J. Occup. Med. Toxicol. 5, 14. Andersen, L.B., Schnohr, P., Schroll, M., Hein, H.O., 2000. All-cause mortality associated with physical activity during leisure time, work, sports, and cycling to work. Arch. Intern. Med. 160, 1621–1628. Bravata, D.M., Smith-Spangler, C., Sundaram, V., et al., 2007. Using pedometers to increase physical activity and improve health. A systematic review. JAMA 298, 2296–2304. Brønnum-Hansen, H., Andersen, O., Kjøller, M., Rasmussen, N., 2004. Social gradient in life expectancy and health expectancy in Denmark. Soz. Praventivmed. 49, 36–41. Brown, W.J., Miller, Y.D., Miller, R., 2003. Sitting time and work patterns as indicators of overweight and obesity in Australian adults. Int. J. Obes. Relat. Metab. Disord. 27, 1340–1346. Chau, J.Y., Van Der Ploeg, H.P., Merom, D., Chey, T., Bauman, A.E., 2012. Cross-sectional associations between occupational and leisure-time sitting, physical activity and obesity in working adults. Prev. Med. 54, 195–200. Choi, B., Schnall, P.L., Yang, H., et al., 2010. Sedentary work, low physical job demand, and obesity in US workers. Am. J. Ind. Med. 53, 1088–1101. Church, T.S., Thomas, D.M., Tudor-Locke, C., et al., 2011. Trends over 5 decades in U.S. occupation-related physical activity and their associations with obesity. PLoS One 6, 1–7. Clark, B., Thorp, A., Winkler, E., et al., 2011. Validity of self-report measures of workplace sitting time and breaks in sitting time. Med. Sci. Sports Exerc. 43, 1907–1912. Ferrando, A.A., Stuart, C.A., Sheffield-Moore, M., Wolfe, R.R., 1999. Inactivity amplifies the catabolic response of skeletal muscle to cortisol. J. Clin. Endocrinol. Metab. 84, 3515–3521. Hamilton, M.T., Hamilton, D.G., Zderic, T.W., 2007. Role of low energy expenditure and sitting in obesity, metabolic syndrome, type 2 diabetes, and cardiovascular disease. Diabetes 56, 2655–2667. Hannerz, H., Albertsen, K., Nielsen, M.L., Tuchsen, F., Burr, H., 2004. Occupational factors and 5-year weight change among men in a Danish national cohort. Health Psychol. 23, 283–288. Haslam, D.W., James, W.P., 2005. Obesity. Lancet 366, 1197–1209. Holtermann, A., Hansen, J.V., Burr, H., Sogaard, K., Sjogaard, G., 2012. The health paradox of occupational and leisure-time physical activity. Br. J. Sports Med. 46, 291–295. Hu, F.B., Li, T.Y., Colditz, G.A., Willett, W.C., Manson, J.E., 2003. Television watching and other sedentary behaviors in relation to risk of obesity and type 2 diabetes mellitus in women. JAMA 289, 1785–1791. Jans, M.P., Proper, K.I., Hildebrandt, V.H., 2007. Sedentary behavior in Dutch workers: differences between occupations and business sectors. Am. J. Prev. Med. 33, 450–454. Juneau, C.E., Potvin, L., 2010. Trends in leisure-, transport-, and work-related physical activity in Canada 1994–2005. Prev. Med. 51, 384–386.
D. Eriksen et al. / Preventive Medicine 73 (2015) 1–5 Kelly, T., Yang, W., Chen, C.S., Reynolds, K., He, J., 2008. Global burden of obesity in 2005 and projections to 2030. Int. J. Obes. (Lond.) 32, 1431–1437. Kwasniewska, M., Jegier, A., Kostka, T., et al., 2014. Long-term effect of different physical activity levels on subclinical atherosclerosis in middle-aged men: a 25-year prospective study. PLoS One 9, e85209. Lagersted-Olsen, J., Korshøj, M., Skotte, J., Carneiro, I.G., Søgaard, K., Holtermann, A., 2014. Comparison of objectively measured and self-reported time spent sitting. Int. J. Sports Med. 35, 534–540. Larsson, C.A., Kroll, L., Bennet, L., Gullberg, B., Rastam, L., Lindblad, U., 2012. Leisure time and occupational physical activity in relation to obesity and insulin resistance: a population-based study from the Skaraborg Project in Sweden. Mummery, W.K., Schofield, G.M., Steele, R., Eakin, E.G., Brown, W.J., 2005. Occupational sitting time and overweight and obesity in Australian workers. Am. J. Prev. Med. 29, 91–97. Nyholm, M., Gullberg, B., Merlo, J., Lundqvist-Persson, C., Råstam, L., Lindblad, U., 2007. The validity of obesity based on self-reported weight and height: implications for population studies. Obesity 15, 197–208. Owen, N., Sugiyama, T., Eakin, E.E., Gardiner, P.A., Tremblay, M.S., Sallis, J.F., 2011. Adults' sedentary behavior determinants and interventions. Am. J. Prev. Med. 41, 189–196.
5
Pace, N., Kodama, A.M., Price, D.C., Grunbaum, B.W., Rahlmann, D.F., Newsom, B.D., 1976. Body composition changes in men and women after 2–3 weeks of bed rest. Life Sci. Space Res. 269–274. Pate, R.R., O'Neill, J.R., Lobelo, F., 2008. The evolving definition of “sedentary”. Exerc. Sport Sci. Rev. 36, 173–178. Pinto Pereira, S.M., Power, C., 2013. Sedentary behaviours in mid-adulthood and subsequent body mass index. PLoS One 8, e65791. Stamatakis, E., Ekelund, U., Wareham, N.J., 2007. Temporal trends in physical activity in England: the Health Survey for England 1991 to 2004. Prev. Med. 45, 416–423. Stock, S., Fernandes, R., Delisle, A., Vézina, N., 2005. Reproducibility and validity of workers' self-reports of physical work demands. Scand. J. Work Environ. Health 6, 409–437. Tudor-Locke, C., Leonardi, C., Johnson, W.D., Katzmarzyk, P.T., 2011. Time spent in physical activity and sedentary behaviors on the working day: the American time use survey. J. Occup. Environ. Med. 53, 1382–1387. Van Uffelen, J.G., Wong, J., Chau, J.Y., et al., 2010. Occupational sitting and health risks: a systematic review. Am. J. Prev. Med. 39, 379–388. Who, 2006. Database on BMI [Online]. Available: http://apps.who.int/bmi/index.jsp? introPage=intro_3.html ([Accessed 21.04. 2013]).