A Systematic Review of Standing and Treadmill Desks in the Workplace

Preventive Medicine 70 (2015) 50–58

Contents lists available at ScienceDirect

Preventive Medicine

journal homepage: www.elsevier.com/locate/ypmed

Review A systematic review of standing and treadmill desks in the workplace

Brittany T. MacEwen a,DanyJ.MacDonaldb, Jamie F. Burr a,⁎ a Department of Applied Human Sciences, Human Performance and Health Research Laboratory, University of Prince Edward Island, Charlottetown, Canada b Department of Applied Human Sciences, Sport Psychology Research Centre, University of Prince Edward Island, Charlottetown, Canada article info abstract

Available online 28 November 2014 Objectives. Standing and treadmill desks are intended to reduce the amount of time spent sitting in today's otherwise sedentary office. Proponents of these desks suggest that health benefits may be acquired as standing Keywords: desk use discourages long periods of sitting, which has been identified as an independent health risk factor. Sedentary lifestyle Our objectives were thus to analyze the evidence for standing and treadmill desk use in relation to physiological Exercise (chronic disease prevention and management) and psychological (worker productivity, well-being) outcomes. Work Methods. A computer-assisted systematic search of Medline, PubMed, PsycINFO, SPORTDiscus, CINAHL, Chronic illness Productivity CENTRAL, and EMBASE databases was employed to identify all relevant articles related to standing and treadmill desk use. Results. Treadmill desks led to the greatest improvement in physiological outcomes including postprandial glucose, HDL cholesterol, and anthropometrics, while standing desk use was associated with few physiological changes. Standing and treadmill desks both showed mixed results for improving psychological well-being with little impact on work performance. Discussion. Standing and treadmill desks show some utility for breaking up sitting time and potentially improving select components of health. At present; however, there exist substantial evidence gaps to compre- hensively evaluate the utility of each type of desk to enhance health benefits by reducing sedentary time. © 2014 Elsevier Inc. All rights reserved.

Contents

Introduction...... 51 Methods...... 51 Literaturesearch...... 51 Inclusioncriteria...... 51 Articleretrievalprocess ...... 52 Dataextractionandmethodologicalqualityassessment...... 52 Results...... 52 Physiologicaloutcomes...... 52 Baselinemeasures...... 52 Traditionalclinicalcardiometabolicriskfactors...... 54 Anthropometrics...... 54 Psychologicaloutcomes ...... 54 Workperformance...... 54 Moodstates...... 55 Cognitivefunction...... 55 Discussion...... 56 Physiological...... 56 Psychological...... 56 Limitationsandfuturedirections...... 56 Conclusion...... 57 Conﬂictofintereststatement ...... 57

⁎ Corresponding author. Rm. 327, Health Sciences Building, 550 University Ave., University of Prince Edward Island, Charlottetown C1A 4P3, PE, Canada. Fax: +1 902 628 4367. E-mail address: [email protected] (J.F. Burr).

Acknowledgments...... 57 AppendixA. Supplementarydata...... 57 References...... 57

Introduction workstations. It has been shown that both desks are practical for the workplace and will be used if available (Grunseit et al., 2012; It has long been understood that a physically active lifestyle is im- Thompson et al., 2008). Consistent with the contention that education portant in overall health and well-being along with reducing the risk is an important component to use, Wilks et al. (2006) found that partic- for chronic diseases (Warburton et al., 2006). In order to observe health ipants who received instructions on the use and benefits of sit–stand benefits, including reduced risk for chronic diseases, the Global Physical workstations used them more frequently than those who did not re- Activity (PA) Guidelines currently recommend a minimum of 150 min ceive any instructions. Additionally, desks with electronically adjustable of moderate to vigorous PA per week for adults aged 18–64 years tables (for height adjustment from sitting to standing) resulted in more (World Health Organization, 2010), which have been adopted by most frequent usage than those needing manual adjustments (Wilks et al., developed nations. However, it has been conservatively estimated that 2006; Grunseit et al., 2012). These results suggest that standing and only 15% of Canadians currently meet these guidelines with the average treadmill desks are feasible in the workplace if properly implemented. Canadian adult spending 69% or 9.5 of their waking hours engaging However, the extent of health benefit outcomes from the use of these in sedentary behaviors (Colley et al., 2011). Owing to increases in desks and whether their use should reasonably be expected to decrease non-physically demanding occupational tasks, among other factors, a the incidence or progression of chronic disease are unclear. Therefore, majority of current sedentary behavior is associated with the workplace the objective of this systematic review is to examine the current litera- where a large proportion of workers spend the day sitting (Juneau and ture investigating the use of standing and treadmill desks to understand Potvin, 2010). how these interventions can be used in the prevention or treatment of Increased sedentary behavior is significantly associated with an common chronic diseases including obesity, diabetes and cardiovascu- elevated risk of diabetes, cardiovascular disease, and all-cause mortality lar disease. Additionally, cognitive function, workplace performance, (Wilmot et al., 2012). More specifically, time spent sitting is strongly job satisfaction, mood states and quality of life are important psycholog- associated with increased rates of the metabolic syndrome, type-2 ical outcomes that may influence a person's performance at work and diabetes mellitus, and obesity (Hamilton et al., 2007). Distinct health their overall well-being. Therefore, this review further considered outcomes observed between sitting, non-exercise PA, and exercise psychological variables and their resulting impacts of standing desk or suggest that sitting has an independent association with overall treadmill desk use. health and mortality that must be considered independently of other PA (Katzmarzyk et al., 2009; Hamilton et al., 2007). Positive Methods correlations and dose–response relationships have been shown be- Literature search tween sitting time and mortality, even in individuals who are otherwise physically active. For example, Katzmarzyk et al. (2009) found A computer assisted database search of Medline, PubMed, PsychINFO, that the group of individuals who spent the highest amount of time SPORTDiscus, CINAHL, CENTRAL (Cochrane Central Register of Controlled Trials) fi sitting had a signi cantly higher risk of mortality than did the reference and EMBASE up to June 2013 was conducted to find English language studies group, regardless of their level of PA. This suggests that compensation investigating sitting, standing, or treadmill walking at work with physiological for time spent sitting cannot be achieved by meeting or even exceeding or psychological outcomes. Search words (outlined in Supplementary Table 1) the current PA guidelines (Katzmarzyk et al., 2009), thus addressing the included variations of possible interventions and terms related to the physiolog- cause of the sedentary behavior itself is of increased importance. Further- ical and psychological variables of interest. Reference lists of articles retrieved more, Katzmarzyk et al.'s (2009) group showed that sitting time and mor- were manually checked for additional articles. tality were associated independent of body mass index (BMI), thus this effect was not simply the result of the typical complications resulting Inclusion criteria from overweight or obesity. However, the highest mortality rates were Peer-reviewed studies published in academic journals which involved a observed for obese individuals who spent most of their time sitting, standing or treadmill walking intervention (standing desk or treadmill desk) suggesting that prolonged sitting may be most detrimental to obese compared to regular seated desk work or investigations that compared sitting individuals. to either standing or treadmill walking at work were eligible for inclusion. The overall effectiveness of current workplace interventions to Studies with participants of working age (N18) and of any health status were reduce sitting time was recently shown by Chau et al. (2010) to lack included. To be eligible for inclusion, the study must have evaluated at least strong supporting evidence. This suggests that it is difficult for workers one relevant physiological or psychological outcome listed in Table 1.Aspecific to incorporate non-exercise PA into their workday and thus novel interventions to off-set sitting must be explored. One novel intervention to decrease sitting in the workplace utilizes a workstation wherein the Table 1 user does not sit in a standard chair, but rather stands or walks using a Physiological and psychological outcomes of interest. specially designed “standing desk” or “treadmill desk” instead. Standing Physiological outcomes –Cardiovascular disease (cholesterol, blood lipids, or treadmill walking while working is intended to reduce sitting time blood pressure, heart rate, chronic venous insufficiency, and encourage non-exercise PA in the workplace and thus improve varicose veins, deep vein thrombosis) health. Anecdotally, many users claim that standing at work also –Diabetes (insulin, glucose) – increases their energy levels and consequent motivation for leisure Obesity (energy expenditure, weight loss, BMI, waist/hip circumference) time PA as well, but this remains speculative. Psychological outcomes –Cognitive function To date, numerous studies have found that these desks are effective –Job satisfaction in reducing the amount of time spent sitting during the workday –Mood states (Straker et al., 2009; Grunseit et al., 2012). Multiple studies have further –Productivity (workplace performance) –Quality of life examined the feasibility and usage of standing desks and treadmill 52 B.T. MacEwen et al. / Preventive Medicine 70 (2015) 50–58 focus was given to diseases of cardiometabolic origin, as they represent Data extraction and methodological quality assessment important public health issues, and have specific links to PA, or a lack thereof (Burr et al., 2010, 2012; Katzmarzyk et al., 2009). Specific risk markers for Extracted information included year of publication, study design, objectives, each disease were identified as they represent important measures and predic- participant characteristics (age, sex, ethnicity, and health status), sample size, tors of disease. intervention (standing or treadmill walking), FITT (frequency, intensity, time It is equally important to investigate the psychological implications of and type), outcomes measured, assessments used, results and conclusions. changes in work environment. If an intervention is beneficial physiologically Assessment of the methodologies used in the studies was achieved by using but aversive psychologically, or vice versa, then its utility must be considered a modified version of the Downs and Black (1998) checklist. The modified to reach an optimal balance of overall health. Cognitive function, workplace version excluded three questions as they were not relevant to the present performance, job satisfaction, mood states, and quality of life are important psy- study. Combined means and standard deviations between studies were calcu- chological outcomes that must be considered as they play important roles in lated using given values, or from estimated values interpreted from tables or influencing work performance and overall well-being. figures. Table 2 outlines the included studies and data extracted. Exclusion criteria included published supplements, abstracts, reports, reviews, opinion articles, commentaries, magazine articles, book chapters, and Results presentations. Only studies comparing sitting to standing or treadmill walking specifically were included, those comparing sitting with general PA were excluded. A total of 23 studies were included in the review, 19 of which were quasi-experimental studies lacking appropriate randomization or control and 4 randomized controlled trials. Consistent with the outcomes of inter- Article retrieval process est presented in Table 1, data are presented across the two dimensions of physiological and psychological outcomes. The initial database search returned a total of 6369 results, which was reduced to a total of 4702 articles after duplicates were removed. These articles Physiological outcomes were first screened based on title and abstract, leading to the collection of 73 apparently relevant titles. Of the 73 relevant titles, 25 abstracts were considered appropriate for further review. Additional manual reference searches yielded Baseline measures another 8 studies that were retrieved and evaluated against the inclusion criteria. Of the 33 captured studies, 10 were culled, resulting in 23 articles that Standing desk. A total of 118 participants were included in 5 different fit full inclusion criteria (Fig. 1). studies measuring energy expenditure, heart rate (HR) or blood pressure

Fig. 1. Flowchart of the article selection process based on inclusion/exclusion criteria. B.T. MacEwen et al. / Preventive Medicine 70 (2015) 50–58 53

Table 2 Characteristics of studies included in the review.

Authors Intervention Length of #of Health status Outcome measures Methodological intervention participants score (/24)

Alderman et al. (2014) Treadmill workstation Single 66 Healthy Heart rate, cognitive function, typing 16 assessment performance, reading comprehension Alkahjah et al. (2012) Sit–stand workstation 3 months 32 Not specified Fasting blood lipids, glucose, weight, BMI 16 Beers et al. (2008) Standing Single 24 Non-obese Heart rate, energy expenditure, fatigue, comfort, 13 assessment (BMI b 30) typing performance Cox et al. (2011) Active workstation Single 31 Healthy Heart rate, blood pressure, energy expenditure, 14 (standing assessment work performance (speech quality) and treadmill walking) Drury et al. (2008) Standing Single 12 Not specified Perceived workload, performance 16 assessment Dunstan et al. (2012) Sitting with light–moderate Single 19 Overweight/obese Postprandial glucose and insulin 21 treadmill walking breaks assessment (BMI N 25) Ebara et al. (2008) Sit–stand workstation Single 24 Healthy Sleepiness, work performance, arousal level 17 assessment Edelson and Danoff Treadmill workstation Single 5 Not specified Typing performance, stress, arousal 14 (1989) assessment Funk et al. (2012) Treadmill workstation Single 24 Healthy Typing performance 15 assessment Hasegawa et al. (2001) Sit–stand Single 16 Not specified Work performance, fatigue 13 assessment Hedge and Ray (2004) Standing height adjustable 4–6 weeks 54 Not specified Work performance 16 desk Husemann et al. (2009) Sit–stand workstation 5 days 60 Healthy Mood states and office behaviors, fatigue, 16 typing performance John et al. (2009) Treadmill workstation Single 20 Not specified Cognitive function, typing performance 15 assessment John et al. (2011) Treadmill workstation 9 months 12 BMI N 28 Body composition, heart rate, blood pressure, 16 weight, waist and hip circumference, blood lipids, glucose, insulin, glycosylated hemoglobin Koepp et al. (2013) Treadmill desk 12 months 36 10 lean; 15 Body composition, blood lipids, glucose, waist 17 overweight; circumference, energy expenditure, work 11 obese performance Levine and Miller Walk and work desk Single 15 Obese (BMI N 30) Energy expenditure 13 (2007) assessment Miyashita et al. Active workstation 2 days 15 Healthy Blood lipids, insulin, glucose 14 (2013) (standing and treadmill walking) Ohlinger et al. (2011) Active workstation Single 50 b150 kg Cognitive function, motor performance 13 (standing assessment and treadmill walking) Pronk et al. (2012) Sit–stand desk 7 weeks 34 Healthy Mood states, office behaviors, fatigue 18 Reiff et al. (2012) Standing desk Single 20 Healthy Energy expenditure 16 assessment Speck and Schmitz Standing Single 14 Sedentary with Energy expenditure 10 (2011) assessment BMI N 20 Straker et al. (2009) Active workstation Single 30 Healthy Heart rate, typing performance, keyboard and 15 assessment mouse performance Thompson and Levine Treadmill desk Single 11 Not specified Energy expenditure, typing performance 14 (2011) assessment

(BP) during a single assessment of the standing posture (Beers et al., Treadmill desk. Energy expenditure, HR and BP were also measured for 2008; Cox et al., 2011; Levine and Miller, 2007; Reiff et al., 2012; treadmill desk use with 201 individuals participating in 7 separate stud- Speck and Schmitz, 2011). Of the 4 studies that measured energy ies (Alderman et al., 2014; Cox et al., 2011; Levine, Miller, 2007; John expenditure using indirect calorimetry, 3 found that a standing pos- et al., 2011; Koepp et al., 2013; Straker et al., 2009; Thompson and ture resulted in significant increases varying from 4.1 kcal/h to 20.4 Levine, 2011). Four studies measured energy expenditure using indirect kcal/h in each study (Beers et al., 2008; Cox et al., 2011; Reiff et al., calorimetry and 3 found a significant increase averaging 100.04 ± 2012). The single study that did not find a difference in energy ex- 51.9 kcal while walking at 1–1.1 mph, during a single assessment of penditure had a very low rating of methodological quality owing the treadmill desk (Cox et al., 2011; Levine and Miller, 2007; primarily to a lack of detail in the manuscript (Speck and Schmitz, Thompson and Levine, 2011). Alternatively, Koepp et al. (2013) found 2011). HR was measured in 3 studies and showed a significant in- no significant change in gross energy expenditure over a 1 year period crease while standing of 8.4 ± 4.8 beats/min (bpm) on average of treadmill desk use. HR was measured in 4 studies with two reporting (Beers et al., 2008; Straker et al., 2009; Cox et al., 2011). BP was asignificant increase averaging 12.32 ± 9.28 bpm while using the measured in only one study and also showed a significant increase treadmill desk at a speed of 1.6 km/h (Alderman et al., 2014; Straker in systolic BP from 123 ± 3 mm Hg to 124 ± 3 mm Hg while stand- et al., 2009; Cox et al., 2011). In contrast, John et al. (2011) measured ing, as well as an increase in diastolic BP from 76 ± 3 mm Hg to resting HR and BP and did not find any difference between active and 82 ± 3 mm Hg while standing (Cox et al., 2011). sedentary conditions. Only one other study measured BP and observed 54 B.T. MacEwen et al. / Preventive Medicine 70 (2015) 50–58 a significant increase of systolic BP from 124 ± 3 mm Hg to 129 ± changes in body composition using BMI and waist and hip circumference 3 mm Hg and diastolic BP increased from 76 ± 3 mm Hg to 80 ± in relation to the comparison group (14 participants). 3 mm Hg while using a treadmill workstation (Cox et al., 2011). Treadmill desk. Anthropometric measures were investigated in 3 studies Traditional clinical cardiometabolic risk factors that included a total of 63 participants using treadmill desks (John et al., 2011; Koepp et al., 2013; Levine and Miller, 2007). All 3 studies included Standing desk. Only one study investigated traditional cardiometabolic body mass and composition in their measurements. In a 12 month in- risk factors with the use of a sit–stand workstation. Alkhajah et al. tervention, Koepp et al. (2013) found that 65% of their participants (2012) included a total of 32 participants (18 intervention and 14 com- lost an average of 3.4 ± 5.4 kg of fat mass, while John et al. (2011) parison) in their study. Measurements of fasting blood lipids, glucose, showed a non-significant trend toward weight loss but did not find total cholesterol, HDL and triglycerides were taken at baseline, 1 week, any differences in body fat and mass following a 9 month intervention. and 3 months. The study found a significant increase in HDL cholesterol Both investigations reported that obese participants lost more weight of 0.26 mmol/L between intervention and comparison groups. Howev- than non-obese participants. Levine and Miller (2007) extrapolated er, there were no significant differences in the other cardiometabolic further data to suggest that replacing sitting with a treadmill desk for risk factors. 2–3 h each workday could result in a weight loss of 20–30 kg/year in obese individuals. Treadmill desk. A treadmill desk intervention in place of a traditional sit- Waist circumference was measured by John et al. (2011) and Levine ting desk was investigated in two studies with a combined total of 48 and Miller (2007) with both groups finding significant decreases in participants (John et al., 2011; Koepp et al., 2013). One study included waist circumference. A summary of physiological changes is presented 12 overweight or obese office workers (based on BMI), 8 of whom in Table 3. were pre-hypertensive or hypertensive at baseline, with high levels of LDL, total cholesterol and triglycerides. These investigators also in- Psychological outcomes cluded glycosylated hemoglobin and insulin. The participants used the treadmill desk in place of a traditional sitting desk for 9 months (John Work performance et al., 2011). The other study used the treadmill desk for 1 year and involved 36 participants of whom 10 were lean, 15 were overweight and Standing desk. Workplace performance while using a sit–stand worksta- 11 were obese (Koepp et al., 2013). Both studies measured glucose, lipid tion was measured in a total of 220 participants in 7 separate studies panels, waist circumference, and body composition. (Beers et al., 2008; Drury et al., 2008; Ebara et al., 2008; Hasegawa In the shorter 9 month intervention (John et al., 2011), a significant et al., 2001; Hedge and Ray, 2004; Husemann et al., 2009; Straker reduction in total cholesterol from 123 ± 26 mg/dL to 115 ± 36 mg/dL et al., 2009). Ebara et al. (2008) found that performance on a typing and reductions of LDL from 206 ± 31 mg/dL to 195 ± 36 mg/dL were task was not impacted with sit–stand workstation use of less than observed. However, similar reductions were not evident in the longer 150 min. However, typing performance decreases were observed above intervention (Koepp et al., 2013). This may be due to a baseline effect 150 min of use per day. Conversely, no significant changes in typing per- as the participants had high levels of LDL and total cholesterol at base- formance were found between sitting and standing postures in three line. A significant increase in HDL from 55 ± 20 mg/dL to 60 ± 23 studies (Drury et al., 2008; Beers et al., 2008; Straker et al., 2009), which mg/dL was demonstrated by Koepp et al. (2013) with similar increases tested computer task performance over 40 min (Drury et al., 2008), observed in another study involving the use of a sit–stand workstation typing ability over 20 min (Beers et al., 2008), and typing and mouse per- (Alkhajah et al., 2012). Within the pre-hypertensive or hypertensive formance over 3 min (Straker et al., 2009). participants approximately 33% reduced their blood pressures to nor- Husemann et al. (2009) assessed typing abilities while using a sit– mal levels (John et al., 2011). Although no differences were found in standworkstationfor4hperdayoverafive day period. The study ob- postprandial or fasting glucose and insulin, there was a significant reserved a non-significant trend toward reduced efficiency in the number duction in glycosylated hemoglobin which is a more accurate measure of characters typed per minute and no difference in errors per minute or of long-term glycemic control (John et al., 2011). overall error rate while standing. The study concluded that data entry Dunstan et al. (2012) demonstrated a decreased net glucose was not influenced by the sit–stand intervention. Hedge and Ray response to a standardized test drink after a 2 min treadmill walking (2004) used an electronic height adjustable work surface over a break every 10 min during prolonged sitting in 19 obese participants. 4–6 week test period and investigated changes in work patterns for Lower intensity of PA resulted in a 24.1% decrease in net glucose daily keyboard and mouse use. The study reported no changes in response while moderate intensity treadmill walking showed a 29.6% work patterns for daily keyboard or mouse use and reported improve- decrease. The study concluded that short interruptions to sitting signif- ments in comfort for keyboard, mouse, chair and workstation use icantly reduced postprandial glucose and insulin responses regardless of while using the height adjustable desk. the intensity of the PA, which in turn encourages increased insulin sen- The final study found that a change of posture reduced perceived sitivity and reduced insulin secretion consistent with preservation of workload. However, the longer the task time, the lower the effects of pancreatic β-cell function. Similarly, postprandial glucose was reduced changed posture. Additionally, maintaining a standing posture for a for treadmill desk use but not for standing posture in a separate study long time was determined to be more harmful than a combination of with 15 healthy participants (Miyashita et al., 2013); however, insulin postures (Hasegawa et al., 2001). showed no significant difference in either condition. These different results may be infl uenced by the health status of the participants, as Treadmill desk. Studies investigating performance and cognitive differ- one population was composed of obese participants and the other ences between sitting and treadmill walking have shown mixed results. group was of a healthy weight. Eight quasi-experimental studies included a total of 242 participants (Alderman et al., 2014; Funk et al., 2012; John et al., 2009; Koepp Anthropometrics et al., 2013; Ohlinger et al., 2011; Straker et al., 2009; Edelson and Danoff, 1989; Thompson and Levine, 2011). Five studies measured typ- Standing desk. Only one study specifically investigated alterations in an- ing performance while using a treadmill workstation (Funk et al., 2012; thropometric measures using a standing desk intervention (Alkhajah Straker et al., 2009; John et al., 2009; Edelson and Danoff, 1989; et al., 2012). The 18 participants in the intervention group showed a re- Thompson and Levine, 2011). Consistent results were evident in two duction in weight over a 3 month period but did not show any significant of the studies which found that walking on a treadmill at 1 mph B.T. MacEwen et al. / Preventive Medicine 70 (2015) 50–58 55

Table 3 Summary of physiological ﬁndings of studies examining the effects of standing and treadmill desks.

Outcome measures Standing Treadmill walking

# of studies Outcomes # of studies Outcomes

Baseline measures Increased energy expenditure 4 − 4+ Increased resting heart rate 0 n/a 1 0 Increased heart rate 3 ++ 3 ++ Decreased resting diastolic blood pressure 0 n/a 1 0 Decreased resting systolic Blood pressure 0 n/a 1 0 Increased diastolic blood pressure 1 − 1+ Increased systolic blood pressure 1 + 1 ++

Traditional cardiometabolic risk factors Lowered total fasting cholesterol (3 months) 1 / 0 n/a Lowered total cholesterol (9–12 months) 0 n/a 2 − Increased HDL 1 + 1 + Lowered LDL (9–12 months) 0 n/a 2 − Lowered TG 1 0 1 0 Improved fasting glucose 1 0 1 0 Improved postprandial glucose 1 0 2 ++ Improved glucose 0 n/a 1 0 Improved A1C 0 n/a 2 − Improved glycosylated insulin 0 n/a 1 0 Improved postprandial insulin 0 n/a 2 +

Anthropometrics Weight loss 1 − 2+ Lowered BMI 0 n/a 1 / Decreased waist circumference 0 n/a 2 ++

Very strong: ++; strong: +; weak: −; inconclusive: /; no change: 0; n/a = not applicable.

(Thompson and Levine, 2011) or a self-selected speed (Edelson and standing intervention, vigor and total mood disturbance returned to Danoff, 1989) did not result in decreased performance. Two studies baseline and self-esteem was reduced below baseline. However, found that treadmill walking at relatively slow (1.3–1.6 km/h) or fast Husemann et al. (2009) assessed present psychological state, mood, (3.2 km/h) speed resulted in the largest decreases in typing and alertness, tiredness and calmness or agitation and did not find any dif- mouse performance (Funk et al., 2012; Straker et al., 2009). These stud- ference in mood scores between control and intervention groups over ies suggest that a self-selected speed between 1.6 and 3.2 km/h is ideal a one week period. Hedge and Ray (2004) found that comfort increased in optimizing typing and mouse performance. Overall, the studies over time with sit–stand workstation use while less favorable results concluded that the use of a treadmill desk results in a small to medium were found by Beers et al. (2008) and Drury et al. (2008) whose 24 decrease in short-term performance (Straker et al., 2009)andthat and 12 respective participants gave 1) sitting a higher comfort rating treadmill walking decreases performance on tests of typing and than standing, 2) fatigue and discomfort a higher rating in the standing mouse proficiency and lowers performance on tasks of fine motor posture (Beers et al., 2008) and 3) a decreased rating of well-being in an movement (John et al., 2009). upright posture (Drury et al., 2008). These differences may be due to General task performance was measured by two studies (Koepp measures of prolonged sitting and standing collected separately rather et al., 2013; Ohlinger et al., 2011). Koepp et al. (2013) investigated than investigating a combination of postures. Subjective sleepiness overall work performance, quality of work and quality of interactions was measured in 24 participants by Ebara et al. (2008) and was found with co-workers. Findings showed a minor loss in work performance to increase with time under all conditions of sit–stand workstation over the first 3–5 months but found that performance after one year use, thus contributing little to the evaluation of standing desk use. How- exceeded baseline. This indicates a learning effect, which should likely ever, another study (Hasegawa et al., 2001) found that dullness and be considered when interpreting the results of other more acute inter- drowsiness were reduced with changes of posture. ventional studies. The study concluded that work performance was not significantly different between the active and sedentary conditions Treadmill desk. One small study investigated mood states for 5 partici- (Koepp et al., 2013). Olhinger et al. (2011) found that treadmill walking pants using a treadmill desk (Edelson and Danoff, 1989). The study at 1.6 km/h had a significant decrease in scores of the Digital Finger found significantly decreased stress levels, improved arousal levels Tapping Test of motor abilities between sitting, standing and treadmill and fewer physical complaints while using the treadmill workstation walking but not between sitting and standing. over five 2 h sessions during a 2 week period. Further work in this area is warranted. Mood states Cognitive function Standing desk. Mood states were investigated for standing desk use in a total of 224 participants in 7 different studies (Beers et al., 2008; Drury Treadmill desk. Various components of cognitive function while using et al., 2008; Ebara et al., 2008; Hasegawa et al., 2001; Hedge and Ray, the treadmill workstation were measured in a total of 167 participants 2004; Husemann et al., 2009; Pronk et al., 2012). Significant improve- in 4 studies (Alderman et al., 2014; Cox et al., 2011; John et al., 2009; ments in fatigue, vigor, tension, confusion, depression, and total mood Ohlinger et al., 2011). Alderman et al. (2014) found that treadmill walk- disturbance were found after a 5 week sit–stand desk intervention ing at a self-selected speed does not lower cognitive functions such as with 34 participants (Pronk et al., 2012). Participants stated that they information processing and task performance. John et al. (2009) found felt more comfortable, energized, healthier, focused, productive, happi- that treadmill walking at 1 mph influences math solving abilities but er, and less stressed after using the workstation. After withdrawing the does not alter performance on a Stroop test or reading. Olhinger et al. 56 B.T. MacEwen et al. / Preventive Medicine 70 (2015) 50–58

Table 4 Summary of psychological ﬁndings of studies examining the effects of standing and treadmill desks.

Outcome measures Standing Treadmill walking

# of studies Outcomes # of studies Outcomes

No change in work performance 7 ++ 7 − Improved mood states 7 / 1 − No change in cognitive function 0 n/a 4 ++ Job satisfaction 0 n/a 0 n/a Quality of life 0 n/a 0 n/a

Very strong: ++; strong: +; weak: −; inconclusive: /; no change: 0; n/a = not applicable.

(2011) found no significant differences between conditions on tests of Mixed results were shown for glucose management and standing cognitive function and concluded that low-intensity treadmill walking desk use; however, treadmill desk use was shown to result in reductions negatively affects simple motor task performance but does not impact of both glycosylated hemoglobin and postprandial glucose. Increased short-term auditory verbal memory and divided attention. Cox et al. insulin sensitivity and reduced insulin secretion are consistent with (2011) measured speech quality while using a treadmill desk and these results suggesting preservation of pancreatic β cell function found that walking at a speed of 1.6 km/h did not compromise speech resulting in improved glycemic control and therefore a reduced risk of quality. diabetic complications and possibly overall prevalence (Burr et al., Additionally, low intensity arousal observed during treadmill walking 2010; Riddell and Burr, 2011). did not cause deficits in higher order thinking (information processing Overall, the health benefits of standing and treadmill desks seem to speed, executive abilities, selective attention and the ability to inhibit ha- be more influential for obese individuals than non-obese individuals. bitual responses) (Alderman et al., 2014) while improved psychological However, even with small benefits standing and treadmill desks are measures, such as decreased stress, were shown to be possible advan- shown to be an important alternative to prolonged sitting for obese tages of treadmill workstation use (Edelson and Danoff, 1989). A summa- and non-obese persons. ry of psychological changes is presented in Table 4. Psychological Discussion Standing allowed performance to stay steady and showed no de- Physiological crease in work performance over time. It was shown that a change of posture was useful in reducing perceived workload. However, the lon- Both standing and treadmill desks showed some promise of an ger the task time, the lower the effects of changed posture. Substituting ability for improving health outcomes with regular use. However, prolonged standing for prolonged sitting results in a greater feeling of the use of a treadmill desk typically resulted in greater physiological discomfort as a likely result of discomfort in the legs, back or shoulders improvements than the use of a standing desk. This may be the result due to prolonged standing. With a sit –stand workstation, the individual of the treadmill desk interventions tending to be of longer duration would be able to adjust their posture at the onset of discomfort and thus than the standing desk interventions. However, given the increased ef- reduce the total amount of discomfort resulting from prolonged time in fort of treadmill walking compared to standing, greater physiological the same posture. These results suggest that changing posture through- improvement can be expected with treadmill desk use. out the day while using a sit–stand workstation may lead to reductions Both standing and treadmill desks increased energy expenditure, in work stress, discomfort and psychological strain, while contributing measured through indirect calorimetry, compared to sitting, suggesting not only to increases in productivity but also to greater worker satisfac- that these desks may be useful in maintaining energy balance. Obese tion and quality of life. participants lost more weight than did non-obese participants, likely Although the treadmill desk did have a positive impact on health owing to a baseline effect or because posture change affected a greater outcomes, a decrease in typing and mouse performance was shown absolute work in the obese, but in any event the evidence on whole in- with its use. The magnitude of the decrease in performance seems to de- dicates that active desks may be a useful weight management tool for pend on the speed at which one is walking. A treadmill speed between both obese and non-obese individuals. 1.6 km/h and 3.2 km/h appears to be ideal to minimize decreases in typ- Even in the absence of alterations in body composition, it is now well ing and mouse performance; however, individual differences must be recognized that important changes in health can occur with low load PA taken into consideration, and the long term stability of these changes doses that would not be expected to alter fitness. The treadmill desk still remains to be conclusively demonstrated. Performance may be was shown to reduce LDL and total cholesterol in a sample of obese compromised with the initial installation of a treadmill workstation individuals with high levels of blood lipids. Additionally, two studies but improvements exceeding baseline were shown suggesting that (one standing and one treadmill) found a significant increase in HDL. learning and acclimatization to the new workstation may play an As decreases in total cholesterol and LDL and increases in HDL are tradi- important role in returning performance to an acceptable level. Given tional risk factors for cardiovascular disease (Kannel et al., 1976), these an ideal speed and appropriate acclimatization for each individual, the results along with the acute increases in heart rate and sustained reduc- treadmill desk would allow the worker to continue their daily tasks tions in blood pressure in pre-hypertensive or hypertensive individuals without compromising long term typing and mouse performance. suggest that treadmill desks may be useful in attenuating the cardiovas- There were no differences in cognitive processes while standing or cular risk factors. treadmill walking suggesting that these activities do not alter thought Treadmill desk use was effective in reducing waist circumference in processes important to performing desk work and thus have no detri- both obese and non-obese individuals. As a surrogate for visceral fat, a mental impact on the quality of work being produced. high waist circumference is a risk factor for various chronic conditions including cardiovascular disease, diabetes, and the metabolic syndrome Limitations and future directions (Burr et al., 2010; Riddell and Burr, 2011). Reductions in waist circumference with treadmill desk use thus represent a potentially important This study recognizes several limitations. Only English language reduction in cardiometabolic risk factors. studies were included which may limit the data extracted and result B.T. MacEwen et al. / Preventive Medicine 70 (2015) 50–58 57 in exclusion of some relevant literature. Differences in assessment tech- research is necessary in order to determine the appropriateness of niques, setting, participants, length of interventions, and measured out- these desks with respect to enhancing health benefits by reducing comes made it challenging to compare results and reduced the ability to sedentary time. generalize the results to a specific population. This review did not in- volve a meta-analysis and therefore, the statistical significance of com- Conflict of interest statement bined results could not be determined. The authors declare that there are no conflicts of interests. Additionally, adverse outcomes unrelated to the chronic diseases of interest were not addressed. For example, musculoskeletal conditions Acknowledgments which may be influenced by standing or treadmill walking for extended periods of time were not addressed due to an overwhelming body of lit- There are no sources of funding to declare. erature that would be more appropriately discussed in a separate systematic review. Appendix A. Supplementary data Since standing and treadmill desks are a relatively novel intervention, there is currently a limited evidence base and additional research Supplementary data to this article can be found online at http://dx. is still required. As society is currently suffering from a pandemic of doi.org/10.1016/j.ypmed.2014.11.011. chronic disease, further studies need to focus specifically on the possible impacts these desks may have on chronic disease. While a handful of References studies measured energy expenditure specifically in obese participants, no studies focused specifically on chronic disease prevention or reduc- Alderman, B.L., Olson, R.L., Mattina, D.M., 2014. Cognitive function during low-intensity walking: a test of the treadmill workstation. J Phys Act Health 11 (4), 752–758. tion as a primary outcome. The majority of studies involved relatively Alkhajah, T.A., Reeves, M.M., Eakin, E.G., Winkler, E.A.H., Owen, N., Healy, G.N., 2012. Sit– healthy participants and could therefore not compare these results to stand workstations: a pilot intervention to reduce office sitting time. Am. J. Prev. Med. those already suffering from a chronic disease. Other than those includ- 43 (3), 298–303. fi Beers, E.A., Roemmich, J.N., Epstein, L.H., Horvath, P.J., 2008. Increasing passive energy ex- ing obese participants, there were no studies speci cally focusing on penditure during clerical work. Eur. J. Appl. Physiol. 103 (3), 353–360. participants with specific chronic diseases, or high risk populations. Burr, J.F., Rowan, C.P., Jamnik, V.K., Riddell, M.C., 2010. The role of physical activity in type For this reason, more variability in participants is necessary in order to 2 diabetes prevention: physiological and practical perspectives. Phys. SportsMed. 38 – accurately assess the benefits. (1), 72 82. Burr, J., Goodman, J., Thomas, S., Jamnik, V., Gledhill, N., 2012. Addressing the cardiomet- There is limited evidence to make conclusions regarding the physio- abolic risk of working in physically demanding occupations. Curr. Cardiovasc. Risk logical outcomes of sit–stand workstations. Given that the sit–stand Rep. 6 (4), 347–354. workstation seems to provide the best advantage psychologically, Chau, J.Y., der Ploeg, H., van Uffelen, J., Wong, J., Riphagen, I., Healy, G.N., Gilson, N.D., Dunstan, D.W., Bauman, A.E., Owen, N., Brown, W.J., 2010. Are workplace interven- more research is warranted in regard to their ability to illicit a physio- tions to reduce sitting effective? A systematic review. Prev. Med. 51 (5), 352–356. logical response. Colley, R.C., Garriguet, D., Janssen, I., Craig, C.L., Clarke, J., Tremblay, M.S., 2011. Physical ac- The majority of the studies reviewed involved only a single assess- tivity of Canadian adults: accelerometer results from the 2007 to 2009 Canadian Health Measures Survey, Health Reports/Statistics Canada, Canadian Centre For ment of the intervention. With chronic disease treatment and prevention Health Information=Rapports Sur La Santé/Statistique Canada. Cent. Can. Inf. Sur La in mind, it is necessary to conduct longitudinal studies to understand the Santé 22 (1), 7–14. long term impacts and possible side effects of standing desk and treadmill Cox, R.H., Guth, J., Siekemeyer, L., Kellems, B., Brehm, S.B., Ohlinger, C.M., 2011. Metabolic cost and speech quality while using an active workstation. J. Phys. Act. Health 8 (3), desk use. Further research should consist of randomized controlled trials 332–339. addressing questions such as how these desks may influence specific Downs, S.H., Black, N., 1998. The Feasibility of Creating a Checklist for the Assessment of health conditions, if changes return to baseline after time even with con- the Methodological Quality Both of Randomised and Non-randomised Studies of Health Care Interventions. tinued use, and what impact does discontinuing the use of these desks Drury, C.G., Hsiao, Y.L., Joseph, C., Joshi, S., Lapp, J., Pennathur, P.R., 2008. Posture and per- have on the individual? formance: sitting vs. standing for security screening. Ergonomics 51 (3), 290–307. Additionally, none of the reviewed studies investigated job satis- Dunstan, D.W., Kingwell, B.A., Larsen, R., Healy, G.N., Cerin, E., Hamilton, M.T., Shaw, J.E., faction and quality of life as a primary outcome for either standing or Bertovic, D.A., Zimmet, P.Z., Salmon, J., Owen, N., 2012. Breaking up prolonged sitting reduces postprandial glucose and insulin responses. Diabetes Care 35 (5), 976–983. treadmill desks. Ebara, T., Kubo, T., Inoue, T., Murasaki, G., Takeyama, H., Sato, T., Suzumura, H., Niwa, S., Takanishi, T., Tachi, N., Itani, T., 2008. Effects of adjustable sit–stand VDT workstations Conclusion on workers' musculoskeletal discomfort, alertness and performance. Ind. Health 46 (5), 497–505. Edelson, N., Danoff, J., 1989. Walking on an electric treadmill while performing office Based on the empirical evidence of current literature, this review work. ACM SIGCHI Bull. 21 (1), 72–77. concludes that standing and treadmill desks are potentially useful in Funk, R.E., Taylor, M.L., Creekmur, C.C., Ohlinger, C.M., Cox, R.H., Berg, W.P., 2012. Effect of fl walking speed on typing performance using an active workstation. Percept. Motor reducing workplace sedentariness while having a positive in uence Skills 115 (1), 309–318. on workplace stress and overall mood. The treadmill desk provides Grunseit, A., Chau, J., Van der Ploeg, H., Bauman, A., 2012. Thinking on your feet: a quali- the greatest physiological improvements and is most beneficial for tative evaluation of an installation of sit–stand desks in a medium-sized workplace. J. Sci. Med. Sport 15, S195–S196. overweight and obese participants. However, the use of a treadmill Hamilton, M.T., Hamilton, D.G., Zderic, T.W., 2007. Role of low energy expenditure and desk results in larger decreases in work productivity and motor abilities sitting in obesity, metabolic syndrome, type 2 diabetes, and cardiovascular disease. than the standing desk. Diabetes 56 (11), 2655–2667. Hasegawa, T., Inoue, K., Tsutsue, O., Kumashiro, M., 2001. Effects of a sit–stand schedule Standing desk use does not elicit the same physiological impact as on a light repetitive task. Int. J. Ind. Ergon. 28 (3–4), 219–224. the treadmill desk but does result in the least change in productivity Hedge, A., Ray, E., 2004. Effects of an electronic height-adjustable worksurface on comput- and motor abilities. Of the standing desks, a sit–stand desk seems to er worker musculoskeletal discomfort and productivity. Proc. Hum. Factors Ergon. – provide the most benefit allowing the employee to adjust their desks Soc. Annu. Meet. 48 (8), 1091 1095. Husemann, B., Von Mach, C.Y., Borsotto, D., Zepf, K.I., Scharnbacher, J., 2009. Comparisons throughout the day. The standing-only desk could potentially result in of musculoskeletal complaints and data entry between a sitting and a sit–stand work- additional complications with musculoskeletal conditions and feelings station paradigm. Hum. Factors 51 (3), 310–320. of fatigue and discomfort. John, D., Bassett, D., Thompson, D., Fairbrother, J., Baldwin, D., 2009. Effect of using a treadmill workstation on performance of simulated office work tasks. J. Phys. Act. Overall, current evidence suggests that both standing and treadmill Health 6 (5), 617–624. desks may be effective in improving overall health considering both John, D., Thompson, D.L., Raynor, H., Bielak, K., Rider, B., Bassett, D.R., 2011. Treadmill physiological and mental health components. However, at present workstations: a worksite physical activity intervention in overweight and obese office workers. J. Phys. Act. Health 8 (8), 1034–1043. there still exist substantial gaps in the research to fully comprehend Juneau, C., Potvin, L., 2010. Trends in leisure-, transport-, and work-related physical activity the utility of each type of desk to promote health. Therefore additional in Canada 1994–2005. Prev. Med. 51 (5), 384–386. 58 B.T. MacEwen et al. / Preventive Medicine 70 (2015) 50–58

Kannel, W.B., McGee, D., Gordon, T., 1976. A general cardiovascular risk profile: the Speck, R.M., Schmitz, K.H., 2011. Energy expenditure comparison: a pilot study of stand- Framingham Study. Am. J. Cardiol. 38 (1), 46–51. ing instead of sitting at work for obesity prevention. Prev. Med. Int. J. Devoted Prac. Katzmarzyk, P.T., Church, T.S., Craig, C.L., Bouchard, C., 2009. Sitting time and mortality Theory 52 (3–4), 283–284. from all causes, cardiovascular disease, and cancer. Med. Sci. Sports Exerc. 41 (5), Straker, L., Levine, J., Campbell, A., 2009. The effects of walking and cycling computer 998–1005. workstations on keyboard and mouse performance. Hum. Factors 51 (6), 831–844. Koepp, G.A., Manohar, C.U., McCrady-Spitzer, S., Ben-Ner, A., Hamann, D.J., Runge, C.F., Thompson, W.G., Levine, J.A., 2011. Productivity of transcriptionists using a treadmill Levine, J.A., 2013. Treadmill desks: a 1-year prospective trial. Obesity (Silver Spring) desk. Work 40 (4), 473–477. 21 (4), 705–711. Thompson, W.G., Foster, R.C., Eide, D.S., Levine, J.A., 2008. Feasibility of a walking work- Levine, J.A., Miller, J.M., 2007. The energy expenditure of using a “walk-and-work” desk station to increase daily walking. Br. J. Sports Med. 42 (3), 225–228. for office workers with obesity. Br. J. Sports Med. 41 (9), 558–561. Warburton, D.E.R., Nicol, C.W., Bredin, S.S.D., 2006. Health benefits of physical activity: the Miyashita, M., Park, J., Takahashi, M., Suzuki^1,, K., Stensel, D., Nakamura, Y., 2013. evidence. Can. Med. Assoc. J. 174 (6), 801–809. Postprandial lipaemia: effects of sitting, standing and walking in healthy Wilks, S., Mortimer, M., Nylén, P., 2006. The introduction of sit–stand worktables; aspects normolipidaemic humans. Int. J. Sports Med. 34 (1), 21–27. of attitudes, compliance and satisfaction. Appl. Ergon. 37 (3), 359–365. Ohlinger, C.M., Horn, T.S., Berg, W.P., Cox, R.H., 2011. The effect of active workstation Wilmot, E., Edwardson, C., Achana, F., Davies, M., Gorely, T., Gray, L., Khunti, K., Yates, T., use on measures of cognition, attention, and motor skill. J. Phys. Act. Health 8 (1), Biddle, S., 2012. Sedentary time in adults and the association with diabetes, 119–125. cardiovascular disease and death: systematic review and meta-analysis. Diabetologia Pronk, N.P., Katz, A.S., Lowry, M., Payfer, J.R., 2012. Reducing occupational sitting time and 55 (11), 2895–2905. improving worker health: the Take-a-Stand Project, 2011. Prev. Chron. Dis. 9, E154. World Health Organization, 2010. Global recommendations on physical activity Reiff, C., Marlatt, K., Dengel, D.R., 2012. Difference in caloric expenditure in sitting versus for health. Retrieved from http://www.who.int/dietphysicalactivity/publications/ standing desks. J. Phys. Act. Health 9 (7), 1009–1011. 9789241599979/en/. Riddell, M.C., Burr, J., 2011. Evidence-based risk assessment and recommendations for physical activity clearance: diabetes mellitus and related comorbidities. Appl. Physiol. Nutr.Metab.36(Suppl.1),S154–S189.