Benefits for Type 2 Diabetes of Interrupting Prolonged Sitting With

964 Diabetes Care Volume 39, June 2016

Paddy C. Dempsey,1,2 Robyn N. Larsen,1 Benefits for Type 2 Diabetes of Parneet Sethi,1 Julian W. Sacre,1 Nora E. Straznicky,1 Neale D. Cohen,1 Interrupting Prolonged Sitting Ester Cerin,1,3,4 Gavin W. Lambert,1,2 Neville Owen,1 Bronwyn A. Kingwell,1 With Brief Bouts of Light Walking and David W. Dunstan1,3,5 or Simple Resistance Activities Diabetes Care 2016;39:964–972 | DOI: 10.2337/dc15-2336 CLIN CARE/EDUCATION/NUTRITION/PSYCHOSOCIAL

OBJECTIVE To determine whether interrupting prolonged sitting with brief bouts of light- intensity walking (LW) or simple resistance activities (SRA) improves postprandial cardiometabolic risk markers in adults with type 2 diabetes (T2D).

RESEARCH DESIGN AND METHODS In a randomized crossover trial, 24 inactive overweight/obese adults with T2D (14 men 62 6 6 years old) underwent the following 8-h conditions on three separate days (with 6–14 days washout): uninterrupted sitting (control) (SIT), sitting plus 3-min bouts of LW (3.2 km · h21) every 30 min, and sitting plus 3-min bouts of SRA (half-squats, calf raises, gluteal contractions, and knee raises) every 30 min. Stan- dardized meals were consumed during each condition. Incremental areas under the curve (iAUCs) for glucose, insulin, C-peptide, and triglycerides were compared between conditions.

RESULTS 1Baker IDI Heart and Diabetes Institute, Melbourne, fi Victoria, Australia Compared with SIT, both activity-break conditions signi cantly attenuated iAUCs 2 · · 21 – – Faculty of Medicine, Nursing and Health Sci- for glucose (SIT mean 24.2 mmol h L [95% CI 20.4 28.0] vs. LW 14.8 [11.0 ences, Monash University, Melbourne, Victoria, 21 18.6] and SRA 14.7 [10.9–18.5]), insulin (SIT 3,293 pmol · h · L [2,887–3,700] vs. Australia LW 2,104 [1,696–2,511] and SRA 2,066 [1,660–2,473]), and C-peptide (SIT 15,641 3Centre of Physical Activity and Nutrition Re- pmol · h · L21 [14,353–16,929] vs. LW 11,504 [10,209–12,799] and SRA 11,012 search, School of Exercise and Nutrition Sciences, – P < fi Deakin University, Burwood, Victoria, Australia [9,723 12,301]) (all 0.001). The iAUC for triglycerides was signi cantly atten- 4School of Public Health, University of Hong 21 uated for SRA (P < 0.001) but not for LW (SIT 4.8 mmol · h · L [3.6–6.0] vs. LW 4.0 Kong, Hong Kong, China [2.8–5.1] and SRA 2.9 [1.7–4.1]). 5Mary MacKillop Institute for Health Research, Australian Catholic University, Melbourne, Victoria, CONCLUSIONS Australia Interrupting prolonged sitting with brief bouts of LW or SRA attenuates acute Corresponding author: Paddy C. Dempsey, [email protected]. postprandial glucose, insulin, C-peptide, and triglyceride responses in adults with Received 28 October 2015 and accepted 13 T2D. With poor adherence to structured exercise, this approach is potentially March 2016. beneficial and practical. Clinical trial reg. no. ACTRN12613000576729, www.anzctr.org.au. Lifestyle interventions, including exercise, are the recommended front-line therapy This article contains Supplementary Data online in the management of type 2 diabetes (T2D), even after the commencement of at http://care.diabetesjournals.org/lookup/ suppl/doi:10.2337/dc15-2336/-/DC1. hypoglycemic agents. Current guidelines stipulate that, in addition to 150 min/week © 2016 by the American Diabetes Association. of moderate-vigorous aerobic exercise, individuals with T2D should engage in re- Readers may use this article as long as the work is sistance exercises at least 2–3 days/week (1). However, despite the known benefits, properly cited, the use is educational and not for particularly for glucose metabolism and insulin sensitivity, meeting prescribed profit, and the work is not altered. care.diabetesjournals.org Dempsey and Associates 965

exercise guidelines can be challenging, resistance activities (SRA) on postpran- attended the laboratory on five sepa- and many with T2D remain physically dial metabolic responses in adults with rate occasions: medical screening visit, inactive (2). T2D. We hypothesized that postprandial familiarization visit, and three trial con- Furthermore, fewer opportunities blood glucose, insulin, C-peptide, and tri- dition visits in a randomized order: 1) now exist in modern societies for inci- glyceride levels during sitting would SIT, 2) LW, and 3)SRA. dental (nonexercise) physical activity. be attenuated by brief intermittent Rapidly advancing technological innova- bouts of physical activity, irrespective Randomization and Masking Trial condition order was randomly as- tions in transportation, communications, of modality. signed by a third party using computer- workplaces, and home entertainment generated random numbers and sealed have created environments conducive to RESEARCH DESIGN AND METHODS envelopes (block randomization and prolonged periods sittingdsedentary be- Participants balanced block sizes), stratified by sex. haviors, defined as any waking, sitting, or Enrollment Process and Screening (Visit 1) Study personnel were blinded to the reclining behavior with low-energy ex- $ , Overweight/obese (BMI 25 but 40 condition order until the night prior to penditure (,1.5 METs) (3,4). Sedentary 2 – kg/m ) men and women (aged 35 75 the first trial condition. Participants behaviors are ubiquitous and increase years) diagnosed with T2D (diet or met- were blinded to trial condition order the risk of T2D, cardiovascular disease, $ ’ formin controlled, 3 months duration up until commencement of the second and premature mortality, even when [based on American Diabetes Association trial visit. Pathology technicians were the influence of moderate-vigorous or diagnostic criteria] [11]) were recruited kept blinded to trial conditions. leisure-time physical activity is controlled from local community advertisements for (5,6). Therefore, in addition to increas- and the Baker IDI Diabetes Clinic. All par- Study Protocol and Trial Conditions ing purposeful exercise, decreasing sitting ticipants were required to be inactive (i.e., Familiarization (Visit 2) time has the potential to reduce the bur- currently sitting $5 h/day and not meet- Three to five days prior to the first ex- den of T2D. ing physical activity guidelines of $150 perimental trial condition (visit three), There is observational and experi- min/week of moderate-intensity exercise participants attended a familiarization mental evidence that sitting time with for .3 months). Other exclusion criteria visit and were given further practice brief interruptions can be associated , $ were as follows: HbA1c 6.5 or 9%, tak- with the SRA and treadmill walking. Par- with a more favorable cardiometabolic ing insulin or any other hypoglycemic ticipants were also familiarized with all risk profile and postprandial metabolism agents, pregnancy, pre/perimenopausal, study procedures, including weighed than is an equivalent amount of sitting current smoker, employment in a nonse- food diaries and activity records, objec- time accumulated in longer, uninter- dentary occupation, major systemic illness, tive activity monitoring, and require- rupted bouts (7,8). In overweight/obese known physical activity contraindications ments for the restrictive lead-in phase adults at risk for T2D, interrupting sitting (including the presence of cardiovascular and fasting prior to each trial condition. time with 2-min walking bouts every disease, unstable angina, or symptoms of Prearranged, standardized text messag- 20 min reduced glucose and insulin re- cardiac failure at screening visit), or major ing or e-mail prompts were used to max- sponses (24–30%), irrespective of whether illnessorphysicalproblems(acuteor imize participant compliance. the bouts were of light or moderate inten- chronic) limiting ability to perform the sity (9). However, it is not known whether light-intensity physical activities. Indirect Calorimetry such metabolic benefits extend to those After initial telephone screening, all With the aim of characterizing the inter- with T2D. potentially eligible participants attended ventions, indirect calorimetry was com- Moreover, limited consensus exists a medical screening at our laboratory that pleted either before visit three (n =11) on how prolonged sitting time should included the following: anthropometric (during familiarization visit) or after visit be interrupted. Experimental studies to measurements, resting blood pressure, five (n = 12) (during a sixth visit) based date have only examined the utility of resting 12-lead electrocardiogram, blood on participant availability. Participants intermittent standing or ambulatory biochemistry (liver or renal function reported to the temperature-controlled bouts, which may have differing levels of (22–248C) laboratory at 0700–0800 h and HbA1c), and a physical examination metabolic stimulus, practicality, or health performed by the study physician (N.D.C.). after a 12-h overnight fast. After partic- efficacy. As an alternative, resistance-type Participants also underwent initial orien- ipants voided and were weighed, a activities use the larger muscle groups tation to the SRA and treadmill walking TrueOne 2400 metabolic cart (Model and markedly increase muscle activity during this visit to ensure the activity in- QMC; ParvoMedics, Sandy, UT) was . ( 20-fold compared with sitting [10]), terventions could be undertaken safely used to measure VO2,VCO2, and energy 21 potentially providing a potent stimulus and consistently. expenditure (kcal z min )(basedonthe for increased energy expenditure and Weir equation [12]) over an ;75-min for glucose uptake. They can also be Study Design period. (See Supplementary Table 1 for performed in a fixed position with min- This randomized crossover trial was un- additional data handling and calibra- imal disruption to work tasks or leisure- dertaken at the Baker IDI Heart and Di- tions details.) During this time, partici- pursuits. abetes Institute between October 2013 pants completed a protocol divided into We examined, compared with unin- and November 2014 and was approved two sequential parts (outlined below), terrupted sitting (SIT), the effects of by the Alfred Human Research Ethics each part capturing periods of quiet sit- interrupting sitting time with brief bouts Committee. Eligible participants pro- ting, interspersed with either a 3-min of light-intensity walking (LW) or simple vided written informed consent and bout of LW or SRA, in a randomized order: 966 Interrupting Sitting Time and Type 2 Diabetes Diabetes Care Volume 39, June 2016

Part 1: 20-min sitting quietly → 3-min average sedentary (,100 cpm), light- home on the evenings prior to experi- activity LW-1/SRA-1 → 10-min sitting intensity (100–1,951 cpm), and moderate- mental conditions. quietly → 3-min LW-1/SRA-1 → metabolic vigorous ($1,952 cpm) activity time on After a 12-h overnight fast, participants cart recalibrated → commence part 2. valid ($10 h) days (14). reported to the laboratory at 0715 h. Part 2: 15-min sitting quietly → 3-min For minimizing of any potential diet- After voiding and being weighed, they activity LW-2/SRA-2 → 10-min sitting induced variability during testing condi- remained seated while an indwelling → → quietly 3 min LW-2/SRA-2 15-min tions, medication times were standardized catheter was inserted into an antecubital → fi sitting quietly protocol nishes. and food intake was strictly controlled vein and fasting samples collected before starting from the night before each trial (21 h) and after (0 h) a 60-min steady- Trial Conditions (Visits 3–5) visit. Meals were standardized between state period (Fig. 1). Each experimental Figure 1 shows the overall study proto- conditions and were individualized to col. Since an acute exercise session may condition commenced upon starting the enhance insulin action for up to 48 h meet daily estimated energy require- breakfast meal, with the time taken to fi (13), a 6- to 14-day washout period ments (Scho eldequation[15],1.5 consume (,20 min) replicated in subse- between trial conditions was used to physical activity factor) and a target quent conditions. At 3.5 h, participants eliminate potential carryover effects. macronutrient profile of 12–15% en- consumed lunch (,20 min). Postprandial Participants were asked to refrain from ergy from protein, 55–58% energy blood samples were collected at 30-min structured moderate-vigorous physical from carbohydrate, and 29–31% energy intervals (immediately prior to physical activities (i.e., no physical activity be- from fat. For accommodation of dietary activity bouts on activity days) over each yond activities of daily living), caffeine, preferences, participants were able to 7-h condition. and alcohol for 48 h prior to each exper- select from a range of meal options, Participants consumed water ad libi- imental condition. and each meal provided 33% estimated tum during the first trial condition and During the washout period between energy requirements (mean 6 SD 823 6 were then instructed to replicate the experimental conditions, participants 124 kcal/meal). Breakfast options in- volume consumed in subsequent trial resumed their habitual diet and physical cluded bran-based cereal, fruit salad, conditions. Standardized lavatory visits activity patterns. From visit two (famil- ham-and-cheese croissant, and juice. incorporated into the protocol mini- iarization visit) until visit five (final trial condition), participants wore acceler- Lunch options included a salad and mized unscheduled physical activity; ometers (GTX3+; ActiGraph, Pensacola, meat bread roll and commercially avail- however, additional lavatory visits FL) during waking hours for objective able drink. An evening meal pack, con- were permitted. Participants complied measurements of sedentary time and sisting of a commercially available drink, with the respective trial condition proto- physical activity. The 1-min epoch ac- snack, and microwave meal, were also cols under direct supervision from re- tivity data were processed to derive provided for participants to prepare at search staff.

Figure 1—Study design and protocol for treatment conditions. Participants visited the laboratory on five separate occasions. The 3 trial conditions were completed in a randomized order separated by a 6- to 14-day washout. Blood was collected half-hourly, 2 min prior to each activity bout. Each standardized meal (mean 6 SD 822.9 6 124.3 kcal/meal) constituted 33% of participants’ daily estimated energy requirements (Schofield equation [15], physical activity factor 1.5) with a target macronutrient profile of 55–58% energy from carbohydrate, 12–15% energy from protein, and 29–31% energy from fat. care.diabetesjournals.org Dempsey and Associates 967

The three trial conditions were as College of Pathologists of Australasia 0.05 was adopted. Data are expressed follows. (RCPA)-accredited laboratory on the day as mean 6 SEM or mean (95% CI) unless SIT. Participants sat upright in a comfort- of testing for the determination of glucose, otherwise stated. able chair throughout the experimental insulin, and C-peptide levels. Plasma glu- period and were instructed to minimize cose (fluoride/oxalate) was measured RESULTS excessive movement, only rising from using a hexokinase method. Serum insulin Of the 29 participants who attended the chair to void. and C-peptide were measured using a screening, 24 were randomized and LW. Participants rose from the seated chemiluminescent microparticle immuno- completed all trials (Fig. 2). There were position every 30 min throughout the assay (Architect ci16200; Abbott Diagnos- no dropouts after randomization. There experimental period (except during the tics, Santa Clara, CA). Plasma triglycerides were 14 men and 10 women of mean 6 lunch meal) and completed a 3-min bout (from EDTA tubes, stored at 2808C) were SD age 62 6 6 years with BMI 33.0 6 3.4 of LW on a treadmill (zero gradient, analyzed using a COBAS Integra 400+ ana- 2 21 kg/m ,HbA1c 7.2 6 0.7% (55.1 6 8.0 3.2 km z h ) and then returned to the lyzer (Roche Diagnostics, Indianapolis, IN). 2 mmol z mol 1), estimated glomerular fil- seated position. This procedure was un- tration rate 87 6 8 mL/min per 1.73 m2, dertaken on 12 occasions (i.e., 36 min Statistical Analyses total cholesterol 4.36 6 0.83 mmol z total light-intensity activity). Coprimary outcomes were changes in 2 L 1, fasting triglycerides 1.9 6 0.1 SRA. Participants underwent a protocol net incremental area under the curve mmol z L21, fasting HDL cholesterol identical to that of the LW condition, ex- (iAUC) (trapezoidal method) for plasma 1.1 6 0.3 mmol z L21 and LDL cholesterol cept that participants completed 3-min glucose and insulin. Sample size calcu- 2.5 6 0.8 mmol z L21, systolic blood bouts of SRA (total: 36 min) instead of lations were based on a previous trial pressure 123 6 14 mmHg, diastolic LW. The 3 min was divided into a total of conducted in our laboratory (9) using blood pressure 77 6 9 mmHg, and dia- nine 20-s movement segments, alternat- similar methodology in overweight/ betes duration 6.8 6 5.1 years and with n = ing between body weight half-squats, obese adults (24–30% and 23% de- calf raises, gluteal contractions, and crease in glucose and insulin, respec- 23 taking metformin, n = 15 taking statins, knee raises. The interchange between tively), an effect similar in magnitude and n = 16 taking antihypertensive therapy movements was to provide rest for the to that which may be observed after a (included n =16takinganACEinhibitoror corresponding muscle groups between single bout of moderate-intensity cy- angiotensin II receptor blocker, n =5acal- each movement segment, allowing for con- cling in patients with T2D (16). Therefore, cium channel blocker, n = 11 a thiazide tinual muscle activation over the 3-min diuretic, and n =2ab-blocker). Aside we estimated that 17 paired observa- 2 period. To ensure appropriate movement from BMI (men 31.5 kg/m vs. women tions would be needed to achieve 90% 2 standardization, tempo, and correct power to detect the smallest expected 35.2 kg/m , P = 0.0051) and baseline (fast- z 21 form, participants mimicked a standard- effect size (Cohen d = 0.84) in the primary ing) insulin levels (men 70.6 pmol L vs. z 21 ized, preprepared video recording (prac- outcome variables between the interven- women 106.4 pmol L , P = 0.0035), fi ticed in visits one and two). Range of tions (control vs. breaks in sitting, two there were no signi cant differences motion (knee/hip angle 45–908 for half- sided, 5% level). For accommodation of in sex-related baseline parameters or squats/knee raises) was tailored to par- potential withdrawals, 24 participants medications. ticipants’ ability, as assessed during visit were randomized. Across all trial condi- Anthropometric, biochemical, dietary, one. tions and participants, 3% of outcome val- and accelerometer-derived physical activ- Participants had access to television, ues (34 of 1,152 data points) were ity data before each of the respective trial DVDs, books, magazines, and internet missing and treated as such in subse- conditions are presented in Table 1. Apart services during the trial conditions. Ac- quent analyses. from preprandial C-peptide (adjusted for tivity intensity during the trial condi- After recent recommendations on in statistical models), there were no sig- tions was monitored using heart rate data analysis of crossover trials (17), nificant differences between trials for any monitoring (RS400; Polar Electro Oy, generalized linear mixed models (with of these measurements. Kempele, Finland) and the Borg rating of random intercepts were used to evalu- Based on indirect calorimetry mea- perceived exertion (RPE) (range: minimum– ate the differential effects of the exper- surements (n = 23) (Supplementary maximum 6–20) scale. The mean dif- imental conditions on the selected Table 1), compared with 15 min sitting ferences for heart rate (immediately outcomes using Stata 12 (StataCorp LP). quietly, a bout of LW and SRA increased 21 postactivity bout minus preactivity; All models were adjusted for poten- mean energy expenditure (kcal z min ) mean 6 SEM) for the LW and SRA activity- tial confounders explaining residual by 73 6 5% and 121 6 7%, respectively. break conditions were 17 6 1.2 bpm outcome variance (age, sex, and BMI), A bout of SRA, compared with a bout of (range 8–31) and 19 6 1.0 bpm (range preprandial values (iAUC only), and LW, elicited a significantly greater in- 10–30) and for mean RPE were 9 6 0.3 period effects (treatment order). Sex- crease (relative to 15 min sitting quietly) – 6 21 points (range 7 12) and 10 0.3 points by-condition interaction tests were in mean VO2 (0.13 6 0.01 L z min ), – 21 (range 7 13), respectively. performed for each of the iAUC outcome VCO2 (0.08 6 0.01 L z min ), and energy 2 measures. Residuals were examined for expenditure (0.58 6 0.06 kcal z min 1) Biochemical Analysis serial correlation, heteroscedasticity, (all P , 0.001); however, the opposite Code-labeled samples were sent to an and normality. Substantial departures effect was observed for respiratory ex- independent National Association of from model assumptions were not ob- change ratio (VCO2/VO2)(20.02 6 0.01; Testing Authorities (NATA)/The Royal served. A two-tail probability level of P , 0.05). 968 Interrupting Sitting Time and Type 2 Diabetes Diabetes Care Volume 39, June 2016

and C-peptide (LW ↓27%, SRA ↓30%) responses. Despite the novel modality and relative increase in energy expenditure for the SRA bouts compared with LW, glucose, insulin, and C-peptide responses were comparable between the two conditions. Triglyceride levels tended to be lower for both activity types; however, only the iAUC reduction for SRA (↓40%) was statistically significant. This study builds on hypotheses generated from epidemiological observational research on the metabolically beneficial correlates of breaking up sitting time and recent experimental trials demonstrating the acute metabolic benefits of interrupting prolonged sitting with light- (18) and moderate-intensity (9) bouts of ambulation. Our findings extend upon this work by providing new insights regarding the potential efficacy of this novel, lifestyle-based treatment strategy (interrupting prolonged sitting) in adults with T2D; the potential efficacy of an alternative, simple, and practical form of sitting interruption (brief bouts of SRA); and that 3-min light activity bouts every 30 min (versus, for example, ;2 min walking breaks every 20 min [9,18], 5-min walking or standing bouts every 30 min [19], or 15-min Figure 2—Consolidated Standards of Reporting Trials (CONSORT) flow diagram. Meds., medications. postmeal walking bouts [16,20]) may also be a useful prescription target. Our findings are consistent with a re- Figure 3 shows mean glucose, insulin, difference in glucose net iAUC between cent study of similar design in over- C-peptide, and triglyceride concentrations conditions SIT and LW (Supplemen- weight adults showing reduced glucose during each of the trial conditions. Net 7-h tary Fig. 1), indicating that, while LW and insulin responses from brief inter- iAUC during both of the activity-bout con- resulted in significantly lowered glucose ruptions to sitting, irrespective of activ- ditions was significantly (all P , 0.001) net iAUC for both sexes, the magnitude ity bout intensity (9). They are also attenuated compared with SIT for glu- of the glucose attenuation for LW versus consistent with recent findings in post- 21 cose (SIT mean 24.2 mmol z h z L SIT was greater (↓58% vs. ↓26%) in menopausal women at high risk of T2D [95% CI 20.4–28.0], LW 14.8 [11.0– women than in men (mean iAUC differ- (19), which showed metabolic benefits 18.6], and SRA 14.7 [10.9–18.5]), insulin ence in lowering between women and with both walking and standing bouts 21 21 (SIT 3,293 pmol z h z L [2,887–3,700], men 26.8 mmol z h z L [95% CI 213.46 for 5 min every 30 min. Furthermore, LW 2,104 [1,696–2,511], and SRA 2,066 to 20.14; P = 0.045]). The sex-by-condition Peddie et al. (18) demonstrated that, in [1,660–2,473]), and C-peptide (SIT interaction for SRA versus SIT trended sim- healthy normal-weight adults, interrupting 2 15,641 pmol z h z L 1 [14,353–16,929], ilarly (↓53% vs. ↓31%) but was nonsignifi- prolonged sitting with intermittent bouts 21 LW 11,504 [10,209–12,799], and SRA cant (24.6 mmol z h z L [211.17 to of walking (1 min and 40 s every 30 min) 11,012 [9,723–12,301]). The iAUC for tri- 1.98]; P = 0.17). No significant sex-by- was more effective than a single 30-min glycerides was attenuated significantly condition interactions were observed for bout of moderate-vigorous walking in re- for SRA compared with SIT (P , 0.001) any other outcomes. ducing postprandial glycemia. This is con- 2 but not for LW (SIT 4.8 mmol z h z L 1 sistent with other experimental studies [3.6–6.0], LW 4.0 [2.8–5.1], and SRA 2.9 CONCLUSIONS (21–23) suggesting that the manner in [1.7–4.1]). Differences between SRA and This study demonstrates, for the first which physical activity (or sitting time) is LW were only significant for triglyceride time in inactive overweight/obese men accumulated may differentially influence levels (P = 0.048). Meal-specific effects and women with T2D, that interrupting postprandial glucose handling. Indeed, (3.5-h iAUC per meal) are displayed in prolonged sitting with brief bouts of ei- van Dijk et al. (16) recently showed in Supplementary Table 2. ther LW or SRA effectively attenuates adults with T2D that, compared with pro- Asignificant sex-by-condition interac- postprandial glucose (mean change longed sitting, both a 45-min bout of mod- tion effect was observed for the mean ↓39%), insulin (LW ↓36%, SRA ↓37%), erate exercise and three 15-min bouts of care.diabetesjournals.org Dempsey and Associates 969

Table 1—Anthropometric, biochemical, physical activity, and dietary values clearly does not lend itself to skeletal during the preexperimental period muscle contractile activity, increased en- SIT LW SRA ergy expenditure, or augmented blood flow/shear stress (32). Therefore, it Weight (kg) 90.4 6 2.1 90.4 6 2.1 90.3 6 2.1 may be hypothesized that the reduction Preprandial levels* in glucose levels during the brief activity z 21 6 6 6 Plasma glucose (mmol L ) 8.0 0.3 8.1 0.3 8.1 0.3 bout conditions, consistent with prior z 21 6 6 6 Serum insulin (pmol L )87.09.5 83.5 9.5 86.0 9.5 mentioned studies of similar design, z 21 6 6 6 Serum C-peptide (pmol L )97458 924 58§ 984 58 are the result of localized increases in 21 Plasma triglycerides (mmol z L ) 1.7 6 0.2 1.9 6 0.2 1.8 6 0.2 contractile-mediated (insulin-independent) Accelerometer data† glucose uptake. Moreover, the concur- Daily wear time (min) rent reductions in insulin levels with phys- 6 6 6 Habitual period 822 21 877 22 841 21 ical activity suggest less of a reliance on 48-h restricted period 863 6 18 870 6 18 869 6 18 insulin-mediated glucose uptake, with Sedentary time (min/day) concomitant reductions in C-peptide lev- Habitual period 559 6 19 570 6 19 535 6 19 48-h restricted period 552 6 19 569 6 19 563 6 19 els (a marker of endogenous insulin secre- Physical activity time (min/day) tion) further reinforcing this notion. Light intensity Other possible mediators associated Habitual period 275 6 14 299 6 14 278 6 14 with increased postural alterations and 48-h restricted period 306 6 15 296 6 15 302 6 15 light muscle activity may include hemo- Moderate-vigorous dynamic changes (i.e., increased blood Habitual period 7 6 186 196 1 volume, tissue perfusion, and capillary 6 6 6 48-h restricted period 5 15151 permeability [33–35]), as well as modula- ‡ Diet tion of intracellular signaling changes (i.e., Total energy intake (kcal/day) 2,069 6 80 2,079 6 80 2,094 6 80 Total carbohydrate (energy %) 46.7 6 1.1 46.9 6 1.1 47.1 6 1.1 AMPK, translocation/turnover of GLUT4, Total fat (energy %) 31.3 6 1.0 31.8 6 1.0 31.9 6 1.0 and calcium-activated proteins [36]), in- Total protein (energy %) 18.6 6 0.6 17.8 6 0.6 17.8 6 0.6 creased muscle insulin sensitivity, or Data are means 6 SEM. *Preprandial values based on average of two time points (21 and 0 h) changes in sympathetic nervous system immediately before the first meal. †Accelerometer data collected during habitual (free-living) activity. It is also possible that the effects days and the 48-h period preceding the trial condition. ‡Dietary intakes were assessed from of simply standing up more regularly weighed/measured food records during the 48-h period before the trial condition, using dietary could have significantly contributed to analysis software (FoodWorks; Xyris Software, Highgate Hill, Queensland, Australia). §LW significantly different from SIT and SRA (P , 0.05). the observed metabolic effects (19), potentially via combinations of hemody- namic, hemodilutional, or minimal muscle contractile activity. While these mecha- light-intensity activity over a day of sitting and independent risk factor for the devel- nistic possibilities could not be ascertained were effective in lowering postprandial opment of diabetic and cardiovascular from our study design, they should be ex- glucose and insulin responses. Remark- complications in patients with T2D, even amined in future research in populations ably, the shorter, more frequent activity in those receiving oral blood glucose med- with T2D. This could involve longer sta- bouts used in our study appear to have ications (26,27). Second, patients with bilization periods in the standing posi- resulted in comparable, if not greater, T2D are more likely to be physically inac- tion, prior to each activity bout, to reductions in glucose and insulin re- tive (28) and overweight/obese and have account for any potential additive meta- sponses. This difference in magnitude reduced exercise tolerance, with uptake bolic effect of the physical activity itself. change could be due to sex, participant of public health exercise recommenda- Moreover, it remains to be determined attributes, or methodological differences tions remaining a persistent challenge at whether different activity bout modali- or the longer sitting duration used by van the population level (29–31). Third, our ties provide unique physiological bene- Dijk et al. Nonetheless, our findings and findings provide evidence for two novel fits and to what extent they inhibit the those of van Dijk et al. underscore the lifestyle-based treatment strategies (LW adverse consequences of prolonged sit- potential metabolic importance of in- and SRA bouts) with a degree of meta- ting. Future studies should examine the creasing intermittent physical activity bolic efficacy (postprandial lowering of optimal frequency, duration, and inten- across a day of prolonged sitting for glucose and insulin) indirectly comparable sity of activity bouts to determine the glycemic control in patients with T2D. with an acute 45-min bout of moderate specific translational potential for the These findings hold clinical and public exercise followed by a day of prolonged management of T2D. health relevance for three key reasons. sitting in T2D patients (16). This is an im- The nonsignificant reduction in tri- First, postprandial excursions in glucose, portant finding given the intermittent na- glyceride iAUC during the LW condition insulin, and triglycerides can trigger oxi- ture and low intensity of the activity bouts is consistent with the intermittent walk- dative stress, elevated inflammatory cyto- performed by our participants. ing (1 min and 40 s every 30 min) condi- kines, reduced nitric oxide bioavailability, Responsible mechanisms remain un- tion of Peddie et al. (18) in healthy and endothelial dysfunction (24,25). This clear and will require further study. How- normal-weight adults, who notably also dysmetabolic profile represents a direct ever, the inherent nature of sitting demonstrated a significant lowering in 970 Interrupting Sitting Time and Type 2 Diabetes Diabetes Care Volume 39, June 2016

Figure 3—Fasting and postprandial plasma glucose (A), serum insulin (B), serum C-peptide (C), and plasma triglyceride (D) concentrations measured during SIT (○) and sitting interrupted with 3-min LW (,)orSRA(C) bouts. Vertical dashed lines indicate timing of the breakfast (0 h) and lunch (3.5 h) meals. Data are presented as mean 6 SEM.

triglycerides with a single 30-min bout of or the nutritional composition (i.e., Key strengths of our study include the moderate-vigorous walking. However, higher glucose-to-fat ratio) of test focusonbothmenandwomenwith the significant triglyceride reduction dur- meals. overt T2D, the well-described and stan- ing the resistance activity condition in our In light of epidemiological findings dardized trial condition lead-in periods study is a novel finding and a first in pa- documenting sex differences in the as- (as illustrated by minimal variance in tients with T2D. Plausible reasons for the sociations of television-viewing time with confounder variables such as diet, phys- comparatively lower triglyceride levels cardiometabolic biomarkers (39,40), ical activity, and fasting metabolic levels with both walking and resistance activi- our findings offer some initial experi- during trial condition lead-in periods) ties may be due to our trial condition or mental insights suggesting that (post- through the use of weighed food re- activity bouts being of longer duration menopausal) women with T2D may cords and objectively measured physical than some previous studies, which is derive greater reductions in postprandial activity, standardized calorimetric as- broadly consistent with studies that glucose than men by interrupting their sit- sessment of the two modes of activity have observed reductions in triglycerides ting time with LW. Although we suspected bout, the strict behavioral supervision the next day after using both intermittent that the greater glucose reductions may and standardized feeding of a typical and continuous walking interventions be related to increases in activity intensity, Western diet during experimental con- (37,38); the less natural modality (subjec- we did not observe significant increases ditions (as opposed to less ecologically tive comments from participants not in heart rate, oxygen consumption, or per- valid test drinks), full retention of par- reported), increased activity stimulus (in- ceived exertion for women compared ticipants and minimal data loss, and the tensity and energy expenditure), and with men. Differences in adipose and collection of regular blood time points lower respiratory exchange ratio during lean body mass or other biological dispar- during trial visits for more robust time the SRA bouts (indicating a relative in- ities between men and women with T2D course and iAUC calculations. crease in lipid oxidation); differences in could be the potential basis for these The present trial also has some limi- the participants studied (healthy vs. findings. Future studies should continue tations that future studies could ad- overweight/obese patients with T2D); to elucidate sex-specific effects. dress. First, the acute nature of the care.diabetesjournals.org Dempsey and Associates 971

current study precludes extrapolations and physical function) that are presently Funding. This research was supported by Na- about longer-term exposures to the par- unknown. Taken together, while this tional Health and Medical Research Council fi (NHMRC) project grant 1081734 and the Victo- ticular conditions that we examined. It is study provides a rst piece of experimen- rian Government Operational Infrastructure Sup- presently unknown whether our ap- tal evidence on the potential benefits of port scheme. P.C.D. is supported by an Australian proach is beneficial over a longer period interrupting prolonged sitting in T2D Postgraduate Award. J.W.S., G.W.L., N.O., B.A.K., of time or whether the putative benefits patients, further mechanistic studies and and D.W.D. are supported by the NHMRC Fellow- can be sustained in ways that have pre- interventions in larger samples in ecolog- ships scheme. E.C. is supported by an Australian Re- search Council Future Fellowship (FT140100085). viously been shown in longer-term trials ically relevant, free-living, and workplace Duality of Interest. No potential conflicts of that included aerobic exercise (41). Sec- environments, using a broader range of interest relevant to this article were reported. ond, although this well-controlled study participants (including premenopausal Author Contributions. P.C.D. conceived, de- has offered insights into the metabolic women and patients with less well- signed, and conducted the study; analyzed consequences of prolonged sitting and and interpreted data; and wrote the manuscript. controlled T2D [e.g., patients with poorly R.N.L., P.S., J.W.S., N.E.S., N.D.C., E.C., G.W.L., the incorporation of alternate modes of controlled diabetes on insulin or sulfo- N.O., B.A.K., and D.W.D. assisted in the concept intermittent activity bouts, generaliz- nylurea dependent, with b-cell dysfunc- and design of the study and participated in ability to free-living settings is less cer- tion and increased risk of experiencing critical revision of the manuscript for intellec- tain and may not always reflect habitual hypoglycemia]), will be informative in tual content. P.S. and E.C. assisted with data behaviors. For example, in the workplace, fi cleaning or management and statistical analy- developing more speci c public health ses or interpretation. N.D.C. provided clinical other factors such as stress and workload guidelines for the management of T2D. support during data collection. All authors may also play a role in glycemic homeo- In conclusion, interrupting prolonged approved the final version of the manuscript. stasis. Third, the exploratory sex analyses sitting with brief LW or SRA bouts signif- P.C.D. and D.W.D. are the guarantors of this in this study should be interpreted with work and, as such, had full access to all the data icantly attenuates postprandial glucose, in the study and take responsibility for the caution, given the limited sample size, insulin, C-peptide, and triglyceride responses integrity of the data and the accuracy of the which may have increased the risk of in adults with T2D. With the ubiquity of data analysis. type 1 errors. Finally, the standardized sedentary behaviors and the low adher- Prior Presentation. Parts of this study were Western dietary feeding profiles (42) ence to structured exercise, these two presented in abstract form at the 62nd Annual Scientific Meeting of the American College of used during this trial, while arguably approaches are practical strategies that – fl Sports Medicine, San Diego, CA, 26 30 May 2015. more re ective of real-world scenarios, may contribute toward reducing the risk will inevitably vary in daily life settings of diabetes complications and cardiovas- References (e.g., macronutrient profile, glycemic in- cular complications. The efficacy and 1. Colberg SR, Albright AL, Blissmer BJ, et al.; dex, meal frequency, and size). Although sustainability of our particular approach American College of Sports Medicine; Ameri- fi our speci c focus was on standardized should be tested in larger and longer- can Diabetes Association. Exercise and type 2 meal responses to the sedentary and ac- duration trials, as has been done for aer- diabetes: American College of Sports Medi- tivity patterns, such dietary variations and cine and the American Diabetes Association: obic exercise interventions in the T2D joint position statement. Exercise and type 2 their interactions with physical activity context (41). Nonetheless, our findings diabetes. Med Sci Sports Exerc 2010;42:2282– are an integral piece of the puzzle and contribute complementary initial exper- 2303 will require further examination. imental evidence to further inform the 2. Zhao G, Ford ES, Li C, Mokdad AH. Compliance Pragmatically, both activity condi- with physical activity recommendations in US existing, albeit broad, T2D exercise rec- – tions were easily tolerated and well ac- adults with diabetes. Diabet Med 2008;25:221 227 ommendations to “increase daily move- 3. Matthews CE, Chen KY, Freedson PS, et al. cepted by our T2D participants (based ment through unstructured activity to Amount of time spent in sedentary behaviors in on subjective comments collected at gain additional health benefits” (1). the United States, 2003-2004. Am J Epidemiol the end of each trial condition [data 2008;167:875–881 Thus, in addition to the essential promo- not shown]), and it appears the benefi- 4. Sedentary Behaviour Research Network. Let- tion of purposeful moderate-vigorous cial metabolic effects of interrupting ter to the editor: standardized use of the terms and leisure-time physical activity, it seems “sedentary” and “sedentary behaviours”.Appl prolonged sitting can be achieved with prudent and nonmaleficent (primum non Physiol Nutr Metab 2012;37:540–542 different modes of light-intensity activ- nocere) that health care professionals 5. Biswas A, Oh PI, Faulkner GE, et al. Sedentary ity. In this regard, both LW and SRA time and its association with risk for disease in- consider promoting the message, or pro- bouts may have application irrespective cidence, mortality, and hospitalization in adults: viding prescriptive advice to T2D patients, of individual ability or workplace or home a systematic review and meta-analysis. Ann In- to also regularly interrupt prolonged sit- tern Med 2015;162:123–132 context encountered. For example, SRA ting time. 6. Dempsey PC, Owen N, Biddle SJ, Dunstan bouts require no specialized equipment, DW. Managing sedentary behavior to reduce only small amounts of space, and could the risk of diabetes and cardiovascular disease. be easily performed behind a work desk Curr Diab Rep 2014;14:522 Acknowledgments. The authors are grateful 7. Healy GN, Dunstan DW, Salmon J, et al. or at home with minimal disruption to for the excellent technical assistance from Ian Breaks in sedentary time: beneficial associa- work tasks or leisure pursuits, whereas Mullis, Hayley Moon, and Donna Vizi (research tions with metabolic risk. Diabetes Care 2008; light walking may be more convenient nurses); Alaina Natoli (sample analysis); and 31:661–666 and socially acceptable in certain con- Francis Dillon (data cleaning), from Baker 8. Healy GN, Matthews CE, Dunstan DW, texts. Longer-term engagement in either IDI Heart and Diabetes Institute. Most impor- Winkler EA, Owen N. Sedentary time and cardio- tantly, the authors thank the study participants metabolic biomarkers in US adults: NHANES mode of activity bout may also elicit dif- for their time and commitment to the study 2003-06. Eur Heart J 2011;32:590–597 fering physiological effects (e.g., in- protocol; this study would not have been pos- 9.DunstanDW,KingwellBA,LarsenR,etal. creased muscle strength, bone density, sible without them. Breaking up prolonged sitting reduces postprandial 972 Interrupting Sitting Time and Type 2 Diabetes Diabetes Care Volume 39, June 2016

glucose and insulin responses. Diabetes Care 2012; sedentary office workers. Appl Physiol Nutr 33. Hespel P, Vergauwen L, Vandenberghe 35:976–983 Metab 2014;39:1286–1293 K, Richter EA. Important role of insulin and 10. Tikkanen O, Haakana P, Pesola AJ, et al. 22. Duvivier BM, Schaper NC, Bremers MA, et al. flow in stimulating glucose uptake in con- Muscle activity and inactivity periods during Minimal intensity physical activity (standing and tracting skeletal muscle. Diabetes 1995;44: normal daily life. PLoS One 2013;8:e52228 walking) of longer duration improves insulin ac- 210–215 11. American Diabetes Association. Standards tion and plasma lipids more than shorter periods 34. Holmang¨ A, Mimura K, Lonnroth¨ P. Involun- of medical care in diabetes–2013. Diabetes of moderate to vigorous exercise (cycling) in sed- tary leg movements affect interstitial nutrient Care 2013;36(Suppl. 1):S11–S66 entary subjects when energy expenditure is com- gradients and blood flow in rat skeletal muscle. J 12. Weir JB. New methods for calculating met- parable. PLoS One 2013;8:e55542 Appl Physiol (1985) 2002;92:982–988 abolic rate with special reference to protein me- 23. Holmstrup M, Fairchild T, Keslacy S, 35. Inyard AC, Chong DG, Klibanov AL, Barrett tabolism. J Physiol 1949;109:1–9 Weinstock R, Kanaley J. Multiple short bouts EJ. Muscle contraction, but not insulin, in- 13. Mikines KJ, Sonne B, Farrell PA, Tronier B, of exercise over 12-h period reduce glucose ex- creases microvascular blood volume in the Galbo H. Effect of physical exercise on sensitiv- cursions more than an energy-matched single presence of free fatty acid-induced insulin re- ity and responsiveness to insulin in humans. Am bout of exercise. Metabolism 2014;63:510–519 sistance. Diabetes 2009;58:2457–2463 J Physiol 1988;254:E248–E259 24. Ceriello A, Hanefeld M, Leiter L, et al. 36. Santos JM, Ribeiro SB, Gaya AR, Appell 14. Freedson PS, Melanson E, Sirard J. Calibra- Postprandial glucose regulation and diabetic HJ, Duarte JA. Skeletal muscle pathways of tion of the Computer Science and Applications, complications. Arch Intern Med 2004;164: contraction-enhanced glucose uptake. Int J Inc. accelerometer. Med Sci Sports Exerc 1998; 2090–2095 Sports Med 2008;29:785–794 30:777–781 25. O’Keefe JH, Bell DS. Postprandial hyperglyce- 37. Miyashita M, Burns SF, Stensel DJ. Accu- 15. Schofield WN. Predicting basal metabolic mia/hyperlipidemia (postprandial dysmetabolism) mulating short bouts of brisk walking reduces rate, new standards and review of previous is a cardiovascular risk factor. Am J Cardiol 2007; postprandial plasma triacylglycerol concen- work. Hum Nutr Clin Nutr 1985;39(Suppl. 1): 100:899–904 trations and resting blood pressure in healthy 5–41 26. Praet SF, Manders RJ, Meex RC, et al. Gly- young men. Am J Clin Nutr 2008;88:1225– 16. van Dijk JW, Venema M, van Mechelen W, caemic instability is an underestimated prob- 1231 Stehouwer CD, Hartgens F, van Loon LJ. Effect of lem in type II diabetes. Clin Sci (Lond) 2006; 38. Miyashita M, Park JH, Takahashi M, Suzuki moderate-intensity exercise versus activities of 111:119–126 K, Stensel D, Nakamura Y. Postprandial lipae- daily living on 24-hour blood glucose homeosta- 27. van Dijk JW, Manders RJ, Hartgens F, mia: effects of sitting, standing and walking in sis in male patients with type 2 diabetes. Diabe- Stehouwer CD, Praet SF, van Loon LJ. Postpran- healthy normolipidaemic humans. Int J Sports tes Care 2013;36:3448–3453 dial hyperglycemia is highly prevalent through- Med 2013;34:21–27 17. Kenward MG, Roger JH. The use of baseline out the day in type 2 diabetes patients. Diabetes 39.DunstanDW,SalmonJ,OwenN,etal.; covariates in crossover studies. Biostatistics Res Clin Pract 2011;93:31–37 AusDiab Steering Committee. Physical activ- 2010;11:1–17 28. Cichosz SL, Fleischer J, Hoeyem P, et al. Ob- ityandtelevisionviewinginrelationtorisk 18. Peddie MC, Bone JL, Rehrer NJ, Skeaff CM, jective measurements of activity patterns in of undiagnosed abnormal glucose metabo- Gray AR, Perry TL. Breaking prolonged sitting people with newly diagnosed type 2 diabetes lism in adults. Diabetes Care 2004;27:2603– reduces postprandial glycemia in healthy, normal- demonstrate a sedentary lifestyle. Diabet Med 2609 weight adults: a randomized crossover trial. Am J 2013;30:1063–1066 40. Thorp AA, Healy GN, Owen N, et al. Delete- Clin Nutr 2013;98:358–366 29. Hamer M, Bostock S, Hackett R, Steptoe A. rious associations of sitting time and television 19. Henson J, Davies MJ, Bodicoat DH, et al. Objectively assessed sedentary time and type 2 viewing time with cardiometabolic risk bio- Breaking up prolonged sitting with standing or diabetes mellitus: a case–control study. Diabe- markers: Australian Diabetes, Obesity and Life- walking attenuates the postprandial metabolic tologia 2013;56:2761–2762 style (AusDiab) study 2004-2005. Diabetes Care response in postmenopausal women: a ran- 30. Leenders M, Verdijk LB, van der Hoeven L, 2010;33:327–334 domized acute study. Diabetes Care 2016;39: et al. Patients with type 2 diabetes show a greater 41. Knowler WC, Barrett-Connor E, Fowler SE, 130–138 decline in muscle mass, muscle strength, and et al.; Diabetes Prevention Program Research 20. DiPietro L, Gribok A, Stevens MS, Hamm LF, functional capacity with aging. J Am Med Dir Group. Reduction in the incidence of type 2 di- Rumpler W. Three 15-min bouts of moderate Assoc 2013;14:585–592 abetes with lifestyle intervention or metformin. postmeal walking significantly improves 24-h 31. Praet SF, van Loon LJ. Exercise: the brittle N Engl J Med 2002;346:393–403 glycemic control in older people at risk for im- cornerstone of type 2 diabetes treatment. Dia- 42. National Health and Medical Research paired glucose tolerance. Diabetes Care 2013; betologia 2008;51:398–401 Council. Nutrient Reference Values for Australia 36:3262–3268 32. Thosar SS, Bielko SL, Mather KJ, Johnston and New Zealand, Including Recommended Di- 21. Blankenship JM, Granados K, Braun B. Ef- JD, Wallace JP. Effect of prolonged sitting and etary Intakes. Canberra, Department of Health fects of subtracting sitting versus adding exer- breaks in sitting time on endothelial function. and Ageing, National Health and Medical Re- cise on glycemic control and variability in Med Sci Sports Exerc 2015;47:843–849 search Council, 2006