Benefits for Type 2 Diabetes ofInterrupting Prolonged SittingWith Brief Bouts of Light Walkingor Simple Resistance ActivitiesDiabetes Care 2016;39:964–972 | DOI: 10.2337/dc15-2336

OBJECTIVE

To determine whether interrupting prolonged sitting with brief bouts of light-intensity walking (LW) or simple resistance activities (SRA) improves postprandialcardiometabolic 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 (14men 626 6 years old) underwent the following 8-h conditions on three separatedays (with 6–14 days washout): uninterrupted sitting (control) (SIT), sitting plus3-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 underthe curve (iAUCs) for glucose, insulin, C-peptide, and triglycerides were comparedbetween conditions.

RESULTS

Compared with SIT, both activity-break conditions significantly attenuated iAUCsfor glucose (SIT mean 24.2 mmol · h · L21 [95% CI 20.4–28.0] vs. LW 14.8 [11.0–18.6] and SRA 14.7 [10.9–18.5]), insulin (SIT 3,293 pmol · h · L21 [2,887–3,700] vs.LW 2,104 [1,696–2,511] and SRA 2,066 [1,660–2,473]), and C-peptide (SIT 15,641pmol · h · L21 [14,353–16,929] vs. LW 11,504 [10,209–12,799] and SRA 11,012[9,723–12,301]) (all P < 0.001). The iAUC for triglycerides was significantly atten-uated for SRA (P < 0.001) but not for LW (SIT 4.8 mmol · h · L21 [3.6–6.0] vs. LW 4.0[2.8–5.1] and SRA 2.9 [1.7–4.1]).

CONCLUSIONS

Interrupting prolonged sitting with brief bouts of LW or SRA attenuates acutepostprandial glucose, insulin, C-peptide, and triglyceride responses in adults withT2D. With poor adherence to structured exercise, this approach is potentiallybeneficial and practical.

Lifestyle interventions, including exercise, are the recommended front-line therapyin the management of type 2 diabetes (T2D), even after the commencement ofhypoglycemic agents. Current guidelines stipulate that, in addition to 150min/weekof moderate-vigorous aerobic exercise, individuals with T2D should engage in re-sistance exercises at least 2–3 days/week (1). However, despite the known benefits,particularly for glucose metabolism and insulin sensitivity, meeting prescribed

1Baker IDI Heart andDiabetes Institute,Melbourne,Victoria, Australia2Faculty of Medicine, Nursing and Health Sci-ences, Monash University, Melbourne, Victoria,Australia3Centre of Physical Activity and Nutrition Re-search, School of Exercise and Nutrition Sciences,Deakin University, Burwood, Victoria, Australia4School of Public Health, University of HongKong, Hong Kong, China5Mary MacKillop Institute for Health Research,Australian Catholic University, Melbourne, Victoria,Australia

Corresponding author: Paddy C. Dempsey,paddy.dempsey@bakeridi.edu.au.

Received 28 October 2015 and accepted 13March 2016.

Clinical trial reg. no. ACTRN12613000576729,www.anzctr.org.au.

This article contains Supplementary Data onlineat http://care.diabetesjournals.org/lookup/suppl/doi:10.2337/dc15-2336/-/DC1.

© 2016 by the American Diabetes Association.Readersmayuse this article as longas thework isproperly cited, the use is educational and not forprofit, and the work is not altered.

Paddy C. Dempsey,1,2 Robyn N. Larsen,1

Parneet Sethi,1 Julian W. Sacre,1

Nora E. Straznicky,1 Neale D. Cohen,1

Ester Cerin,1,3,4 Gavin W. Lambert,1,2

Neville Owen,1 Bronwyn A. Kingwell,1

and David W. Dunstan1,3,5

964 Diabetes Care Volume 39, June 2016

CLINCARE/ED

UCATION/N

UTR

ITION/PSY

CHOSO

CIAL

exercise guidelines can be challenging,and many with T2D remain physicallyinactive (2).Furthermore, fewer opportunities

now exist in modern societies for inci-dental (nonexercise) physical activity.Rapidly advancing technological innova-tions in transportation, communications,workplaces, and home entertainmenthave created environments conducive toprolonged periods sittingdsedentary be-haviors, defined as any waking, sitting, orreclining behavior with low-energy ex-penditure (,1.5 METs) (3,4). Sedentarybehaviors are ubiquitous and increasethe risk of T2D, cardiovascular disease,and premature mortality, even whenthe influence of moderate-vigorous orleisure-time physical activity is controlledfor (5,6). Therefore, in addition to increas-ing purposeful exercise, decreasing sittingtime has the potential to reduce the bur-den of T2D.There is observational and experi-

mental evidence that sitting time withbrief interruptions can be associatedwith a more favorable cardiometabolicrisk profile and postprandialmetabolismthan is an equivalent amount of sittingtime accumulated in longer, uninter-rupted bouts (7,8). In overweight/obeseadults at risk for T2D, interrupting sittingtime with 2-min walking bouts every20 min reduced glucose and insulin re-sponses (24–30%), irrespectiveofwhetherthe boutswere of light ormoderate inten-sity (9). However, it is not known whethersuch metabolic benefits extend to thosewith T2D.Moreover, limited consensus exists

on how prolonged sitting time shouldbe interrupted. Experimental studies todate have only examined the utility ofintermittent standing or ambulatorybouts, which may have differing levels ofmetabolic stimulus, practicality, or healthefficacy. As an alternative, resistance-typeactivities use the larger muscle groupsand markedly increase muscle activity(.20-fold compared with sitting [10]),potentially providing a potent stimulusfor increased energy expenditure andfor glucose uptake. They can also beperformed in a fixed position with min-imal disruption to work tasks or leisure-pursuits.We examined, compared with unin-

terrupted sitting (SIT), the effects ofinterrupting sitting timewith brief boutsof light-intensity walking (LW) or simple

resistance activities (SRA) on postpran-dial metabolic responses in adults withT2D.We hypothesized that postprandialblood glucose, insulin, C-peptide, and tri-glyceride levels during sitting wouldbe attenuated by brief intermittentbouts of physical activity, irrespectiveof modality.

RESEARCH DESIGN AND METHODS

Participants

Enrollment Process and Screening (Visit 1)

Overweight/obese (BMI $25 but ,40kg/m2) men and women (aged 35–75years) diagnosed with T2D (diet or met-formin controlled, $3 months’ duration[based on American Diabetes Associationdiagnostic criteria] [11]) were recruitedfrom local community advertisementsand the Baker IDI Diabetes Clinic. All par-ticipantswere required to be inactive (i.e.,currently sitting$5 h/day and not meet-ing physical activity guidelines of $150min/week of moderate-intensity exercisefor .3 months). Other exclusion criteriawere as follows: HbA1c,6.5 or$9%, tak-ing insulin or any other hypoglycemicagents, pregnancy, pre/perimenopausal,current smoker, employment in a nonse-dentaryoccupation,major systemic illness,known physical activity contraindications(including the presence of cardiovasculardisease, unstable angina, or symptoms ofcardiac failure at screening visit), or majorillness or physical problems (acute orchronic) limiting ability to perform thelight-intensity physical activities.

After initial telephone screening, allpotentially eligible participants attendedamedical screening at our laboratory thatincluded the following: anthropometricmeasurements, resting blood pressure,resting 12-lead electrocardiogram, bloodbiochemistry (liver or renal functionand HbA1c), and a physical examinationperformedby the study physician (N.D.C.).Participants also underwent initial orien-tation to the SRA and treadmill walkingduring this visit to ensure the activity in-terventions could be undertaken safelyand consistently.

Study DesignThis randomized crossover trial was un-dertaken at the Baker IDI Heart and Di-abetes Institute between October 2013and November 2014 and was approvedby the Alfred Human Research EthicsCommittee. Eligible participants pro-vided written informed consent and

attended the laboratory on five sepa-rate occasions: medical screening visit,familiarization visit, and three trial con-dition visits in a randomized order: 1)SIT, 2) LW, and 3) SRA.

Randomization and Masking

Trial condition order was randomly as-signed by a third party using computer-generated random numbers and sealedenvelopes (block randomization andbalanced block sizes), stratified by sex.Study personnel were blinded to thecondition order until the night prior tothe first trial condition. Participantswere blinded to trial condition orderup until commencement of the secondtrial visit. Pathology technicians werekept blinded to trial conditions.

Study Protocol and Trial Conditions

Familiarization (Visit 2)

Three to five days prior to the first ex-perimental trial condition (visit three),participants attended a familiarizationvisit and were given further practicewith the SRA and treadmill walking. Par-ticipants were also familiarized with allstudy procedures, including weighedfood diaries and activity records, objec-tive activity monitoring, and require-ments for the restrictive lead-in phaseand fasting prior to each trial condition.Prearranged, standardized text messag-ing or e-mail prompts were used tomax-imize participant compliance.

Indirect Calorimetry

With the aim of characterizing the inter-ventions, indirect calorimetry was com-pleted either before visit three (n = 11)(during familiarization visit) or after visitfive (n = 12) (during a sixth visit) basedon participant availability. Participantsreported to the temperature-controlled(22–248C) laboratory at 0700–0800 hafter a 12-h overnight fast. After partic-ipants voided and were weighed, aTrueOne 2400 metabolic cart (ModelQMC; ParvoMedics, Sandy, UT) wasused to measure VO2, VCO2, and energyexpenditure (kcal zmin21) (based on theWeir equation [12]) over an ;75-minperiod. (See Supplementary Table 1 foradditional data handling and calibra-tions details.) During this time, partici-pants completed a protocol divided intotwo sequential parts (outlined below),each part capturing periods of quiet sit-ting, interspersed with either a 3-minbout of LW or SRA, in a randomized order:

care.diabetesjournals.org Dempsey and Associates 965

Part 1: 20-min sitting quietly → 3-minactivity LW-1/SRA-1 → 10-min sittingquietly→ 3-min LW-1/SRA-1→metaboliccart recalibrated → commence part 2.Part 2: 15-min sitting quietly → 3-minactivity LW-2/SRA-2 → 10-min sittingquietly → 3 min LW-2/SRA-2 → 15-minsitting quietly → protocol finishes.

Trial Conditions (Visits 3–5)

Figure 1 shows the overall study proto-col. Since an acute exercise session mayenhance insulin action for up to 48 h(13), a 6- to 14-day washout periodbetween trial conditions was used toeliminate potential carryover effects.Participants were asked to refrain fromstructured moderate-vigorous physicalactivities (i.e., no physical activity be-yond activities of daily living), caffeine,and alcohol for 48 h prior to each exper-imental condition.During the washout period between

experimental conditions, participantsresumed their habitual diet and physicalactivity patterns. From visit two (famil-iarization visit) until visit five (final trialcondition), participants wore acceler-ometers (GTX3+; ActiGraph, Pensacola,FL) during waking hours for objectivemeasurements of sedentary time andphysical activity. The 1-min epoch ac-tivity data were processed to derive

average sedentary (,100 cpm), light-intensity (100–1,951 cpm), and moderate-

vigorous ($1,952 cpm) activity time on

valid ($10 h) days (14).For minimizing of any potential diet-

induced variability during testing condi-

tions, medication times were standardized

and food intake was strictly controlled

starting from the night before each trial

visit. Meals were standardized between

conditions and were individualized to

meet daily estimated energy require-

ments (Schofield equation [15], 1.5

physical activity factor) and a target

macronutrient profile of 12–15% en-

ergy from protein, 55–58% energy

from carbohydrate, and 29–31% energy

from fat. For accommodation of dietary

preferences, participants were able to

select from a range of meal options,

and each meal provided 33% estimated

energy requirements (mean6 SD 8236124 kcal/meal). Breakfast options in-

cluded bran-based cereal, fruit salad,

ham-and-cheese croissant, and juice.

Lunch options included a salad and

meat bread roll and commercially avail-

able drink. An evening meal pack, con-

sisting of a commercially available drink,

snack, and microwave meal, were also

provided for participants to prepare at

home on the evenings prior to experi-

mental conditions.After a 12-h overnight fast, participants

reported to the laboratory at 0715 h.

After voiding and being weighed, they

remained seated while an indwelling

catheter was inserted into an antecubital

vein and fasting samples collected before

(21 h) and after (0 h) a 60-min steady-

state period (Fig. 1). Each experimental

condition commenced upon starting the

breakfast meal, with the time taken to

consume (,20 min) replicated in subse-

quent conditions. At 3.5 h, participants

consumed lunch (,20 min). Postprandial

blood samples were collected at 30-min

intervals (immediately prior to physical

activity bouts on activity days) over each

7-h condition.Participants consumed water ad libi-

tum during the first trial condition and

were then instructed to replicate the

volume consumed in subsequent trial

conditions. Standardized lavatory visits

incorporated into the protocol mini-

mized unscheduled physical activity;

however, additional lavatory visits

were permitted. Participants complied

with the respective trial condition proto-

cols under direct supervision from re-

search staff.

Figure 1—Study design and protocol for treatment conditions. Participants visited the laboratory on five separate occasions. The 3 trial conditionswere 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. Eachstandardized meal (mean6 SD 822.96 124.3 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.

966 Interrupting Sitting Time and Type 2 Diabetes Diabetes Care Volume 39, June 2016

The three trial conditions were asfollows.SIT. Participants sat upright in a comfort-able chair throughout the experimentalperiod and were instructed to minimizeexcessive movement, only rising fromthe chair to void.LW. Participants rose from the seatedposition every 30 min throughout theexperimental period (except during thelunchmeal) and completed a 3-min boutof LW on a treadmill (zero gradient,3.2 km z h21) and then returned to theseated position. This procedure was un-dertaken on 12 occasions (i.e., 36 mintotal light-intensity activity).SRA. Participants underwent a protocolidentical to that of the LW condition, ex-cept that participants completed 3-minbouts of SRA (total: 36 min) instead ofLW. The 3 min was divided into a total ofnine 20-s movement segments, alternat-ing between body weight half-squats,calf raises, gluteal contractions, andknee raises. The interchange betweenmovements was to provide rest for thecorresponding muscle groups betweeneachmovement segment, allowing for con-tinual muscle activation over the 3-minperiod. To ensure appropriatemovementstandardization, tempo, and correctform, participants mimicked a standard-ized, preprepared video recording (prac-ticed in visits one and two). Range ofmotion (knee/hip angle 45–908 for half-squats/knee raises) was tailored to par-ticipants’ ability, as assessed during visitone.Participants had access to television,

DVDs, books, magazines, and internetservices during the trial conditions. Ac-tivity intensity during the trial condi-tions was monitored using heart ratemonitoring (RS400; Polar Electro Oy,Kempele, Finland) and the Borg rating ofperceivedexertion (RPE) (range:minimum–

maximum 6–20) scale. The mean dif-ferences for heart rate (immediatelypostactivity bout minus preactivity;mean6 SEM) for the LW and SRA activity-break conditions were 17 6 1.2 bpm(range 8–31) and 19 6 1.0 bpm (range10–30) and for mean RPE were 9 6 0.3points (range 7–12) and 106 0.3 points(range 7–13), respectively.

Biochemical AnalysisCode-labeled samples were sent to anindependent National Association ofTesting Authorities (NATA)/The Royal

College of Pathologists of Australasia(RCPA)-accredited laboratory on the dayof testing for the determination of glucose,insulin, and C-peptide levels. Plasma glu-cose (fluoride/oxalate) was measuredusing a hexokinase method. Serum insulinand C-peptide were measured using achemiluminescent microparticle immuno-assay (Architect ci16200; Abbott Diagnos-tics, Santa Clara, CA). Plasma triglycerides(from EDTA tubes, stored at2808C) wereanalyzed using a COBAS Integra 400+ ana-lyzer (Roche Diagnostics, Indianapolis, IN).

Statistical AnalysesCoprimary outcomes were changes innet incremental area under the curve(iAUC) (trapezoidal method) for plasmaglucose and insulin. Sample size calcu-lations were based on a previous trialconducted in our laboratory (9) usingsimilar methodology in overweight/obese adults (24–30% and 23% de-crease in glucose and insulin, respec-tively), an effect similar in magnitudeto that which may be observed after asingle bout of moderate-intensity cy-cling in patients with T2D (16). Therefore,we estimated that 17 paired observa-tions would be needed to achieve 90%power to detect the smallest expectedeffect size (Cohen d = 0.84) in the primaryoutcome variables between the interven-tions (control vs. breaks in sitting, twosided, 5% level). For accommodation ofpotential withdrawals, 24 participantswere randomized. Across all trial condi-tions andparticipants, 3%of outcome val-ues (34 of 1,152 data points) weremissing and treated as such in subse-quent analyses.

After recent recommendations ondata analysis of crossover trials (17),generalized linear mixed models (withrandom intercepts were used to evalu-ate the differential effects of the exper-imental conditions on the selectedoutcomes using Stata 12 (StataCorp LP).All models were adjusted for poten-tial confounders explaining residualoutcome variance (age, sex, and BMI),preprandial values (iAUC only), andperiod effects (treatment order). Sex-by-condition interaction tests wereperformed for each of the iAUC outcomemeasures. Residuals were examined forserial correlation, heteroscedasticity,and normality. Substantial departuresfrom model assumptions were not ob-served. A two-tail probability level of

0.05 was adopted. Data are expressedas mean6 SEM or mean (95% CI) unlessotherwise stated.

RESULTS

Of the 29 participants who attendedscreening, 24 were randomized andcompleted all trials (Fig. 2). There wereno dropouts after randomization. Therewere 14 men and 10 women of mean6SD age 626 6 years with BMI 33.06 3.4kg/m2, HbA1c 7.2 6 0.7% (55.1 6 8.0mmol zmol21), estimated glomerular fil-tration rate 876 8 mL/min per 1.73 m2,total cholesterol 4.36 6 0.83 mmol zL21, fasting triglycerides 1.9 6 0.1mmol z L21, fasting HDL cholesterol1.16 0.3mmol z L21 and LDL cholesterol2.5 6 0.8 mmol z L21, systolic bloodpressure 123 6 14 mmHg, diastolicblood pressure 77 6 9 mmHg, and dia-betes duration6.865.1 years andwithn=23 taking metformin, n = 15 taking statins,and n = 16 taking antihypertensive therapy(included n = 16 taking an ACE inhibitor orangiotensin II receptor blocker, n = 5 a cal-cium channel blocker, n = 11 a thiazidediuretic, and n = 2 a b-blocker). Asidefrom BMI (men 31.5 kg/m2 vs. women35.2 kg/m2, P = 0.0051) and baseline (fast-ing) insulin levels (men 70.6 pmol z L21 vs.women 106.4 pmol z L21, P = 0.0035),there were no significant differencesin sex-related baseline parameters ormedications.

Anthropometric, biochemical, dietary,and accelerometer-derivedphysical activ-ity data before each of the respective trialconditions are presented in Table 1. Apartfrom preprandial C-peptide (adjusted forin statistical models), there were no sig-nificant differences between trials for anyof these measurements.

Based on indirect calorimetry mea-surements (n = 23) (SupplementaryTable 1), compared with 15 min sittingquietly, a bout of LW and SRA increasedmean energy expenditure (kcal zmin21)by 736 5% and 1216 7%, respectively.A bout of SRA, compared with a bout ofLW, elicited a significantly greater in-crease (relative to 15 min sitting quietly)in mean VO2 (0.13 6 0.01 L z min21),VCO2 (0.086 0.01 L zmin21), and energyexpenditure (0.58 6 0.06 kcal z min21)(all P , 0.001); however, the oppositeeffect was observed for respiratory ex-change ratio (VCO2/VO2) (20.026 0.01;P , 0.05).

care.diabetesjournals.org Dempsey and Associates 967

Figure 3 shows mean glucose, insulin,C-peptide, and triglyceride concentrationsduring each of the trial conditions. Net 7-hiAUC during both of the activity-bout con-ditions was significantly (all P , 0.001)attenuated compared with SIT for glu-cose (SIT mean 24.2 mmol z h z L21

[95% CI 20.4–28.0], LW 14.8 [11.0–18.6], and SRA 14.7 [10.9–18.5]), insulin(SIT 3,293 pmol z h z L21 [2,887–3,700],LW 2,104 [1,696–2,511], and SRA 2,066[1,660–2,473]), and C-peptide (SIT15,641 pmol z h z L21 [14,353–16,929],LW 11,504 [10,209–12,799], and SRA11,012 [9,723–12,301]). The iAUC for tri-glycerides was attenuated significantlyfor SRA compared with SIT (P , 0.001)but not for LW (SIT 4.8 mmol z h z L21

[3.6–6.0], LW 4.0 [2.8–5.1], and SRA 2.9[1.7–4.1]). Differences between SRA andLW were only significant for triglyceridelevels (P = 0.048). Meal-specific effects(3.5-h iAUC per meal) are displayed inSupplementary Table 2.A significant sex-by-condition interac-

tion effect was observed for the mean

difference in glucose net iAUC betweenconditions SIT and LW (Supplemen-tary Fig. 1), indicating that, while LWresulted in significantly lowered glucosenet iAUC for both sexes, the magnitudeof the glucose attenuation for LW versusSIT was greater (↓58% vs. ↓26%) inwomen than in men (mean iAUC differ-ence in lowering between women andmen26.8 mmol z h z L21 [95% CI213.46to20.14; P = 0.045]). The sex-by-conditioninteraction for SRA versus SIT trended sim-ilarly (↓53% vs. ↓31%) but was nonsignifi-cant (24.6 mmol z h z L21 [211.17 to1.98]; P = 0.17). No significant sex-by-condition interactions were observed forany other outcomes.

CONCLUSIONS

This study demonstrates, for the firsttime in inactive overweight/obese menand women with T2D, that interruptingprolonged sitting with brief bouts of ei-ther LW or SRA effectively attenuatespostprandial glucose (mean change↓39%), insulin (LW ↓36%, SRA ↓37%),

and C-peptide (LW ↓27%, SRA ↓30%) re-sponses. Despite the novel modality andrelative increase in energy expenditurefor the SRA bouts compared with LW,glucose, insulin, and C-peptide responseswere comparable between the two con-ditions. Triglyceride levels tended to belower for both activity types; however,only the iAUC reduction for SRA (↓40%)was statistically significant.

This study builds on hypotheses gen-erated from epidemiological observa-tional research on the metabolicallybeneficial correlates of breaking up sit-ting time and recent experimental trialsdemonstrating the acutemetabolic ben-efits of interrupting prolonged sittingwith light- (18) and moderate-intensity(9) bouts of ambulation. Our findingsextend upon this work by providingnew insights regarding the potential ef-ficacy of this novel, lifestyle-based treat-ment strategy (interrupting prolongedsitting) in adults with T2D; the potentialefficacy of an alternative, simple, andpractical form of sitting interruption(brief bouts of SRA); and that 3-min lightactivity bouts every 30 min (versus, forexample, ;2 min walking breaks every20 min [9,18], 5-min walking or stand-ing bouts every 30 min [19], or 15-minpostmeal walking bouts [16,20]) mayalso be a useful prescription target.

Our findings are consistent with a re-cent study of similar design in over-weight adults showing reduced glucoseand insulin responses from brief inter-ruptions to sitting, irrespective of activ-ity bout intensity (9). They are alsoconsistent with recent findings in post-menopausal women at high risk of T2D(19), which showed metabolic benefitswith both walking and standing boutsfor 5 min every 30 min. Furthermore,Peddie et al. (18) demonstrated that, inhealthy normal-weight adults, interruptingprolonged sitting with intermittent boutsof walking (1 min and 40 s every 30 min)was more effective than a single 30-minbout of moderate-vigorous walking in re-ducing postprandial glycemia. This is con-sistent with other experimental studies(21–23) suggesting that the manner inwhich physical activity (or sitting time) isaccumulated may differentially influencepostprandial glucose handling. Indeed,van Dijk et al. (16) recently showed inadults with T2D that, compared with pro-longed sitting, both a 45-min boutofmod-erate exercise and three 15-min bouts of

Figure 2—Consolidated Standards of Reporting Trials (CONSORT) flow diagram. Meds.,medications.

968 Interrupting Sitting Time and Type 2 Diabetes Diabetes Care Volume 39, June 2016

light-intensity activity over a day of sittingwere effective in lowering postprandialglucose and insulin responses. Remark-ably, the shorter, more frequent activitybouts used in our study appear to haveresulted in comparable, if not greater,reductions in glucose and insulin re-sponses. This difference in magnitudechange could be due to sex, participantattributes, or methodological differencesor the longer sitting duration used by vanDijk et al. Nonetheless, our findings andthose of van Dijk et al. underscore thepotential metabolic importance of in-creasing intermittent physical activityacross a day of prolonged sitting forglycemic control in patients with T2D.These findings hold clinical and public

health relevance for three key reasons.First, postprandial excursions in glucose,insulin, and triglycerides can trigger oxi-dative stress, elevated inflammatory cyto-kines, reduced nitric oxide bioavailability,and endothelial dysfunction (24,25). Thisdysmetabolic profile represents a direct

and independent risk factor for the devel-opment of diabetic and cardiovascularcomplications in patients with T2D, evenin those receivingoral blood glucosemed-ications (26,27). Second, patients withT2D are more likely to be physically inac-tive (28) and overweight/obese and havereduced exercise tolerance, with uptakeof public health exercise recommenda-tions remaining a persistent challenge atthe population level (29–31). Third, ourfindings provide evidence for two novellifestyle-based treatment strategies (LWand SRA bouts) with a degree of meta-bolic efficacy (postprandial lowering ofglucose and insulin) indirectly comparablewith an acute 45-min bout of moderateexercise followed by a day of prolongedsitting in T2D patients (16). This is an im-portant finding given the intermittent na-ture and low intensity of the activity boutsperformed by our participants.

Responsible mechanisms remain un-clear and will require further study. How-ever, the inherent nature of sitting

clearly does not lend itself to skeletalmuscle contractile activity, increased en-ergy expenditure, or augmented bloodflow/shear stress (32). Therefore, itmay be hypothesized that the reductionin glucose levels during the brief activitybout conditions, consistent with priormentioned studies of similar design,are the result of localized increases incontractile-mediated (insulin-independent)glucose uptake. Moreover, the concur-rent reductions in insulin levelswith phys-ical activity suggest less of a reliance oninsulin-mediated glucose uptake, withconcomitant reductions in C-peptide lev-els (amarker of endogenous insulin secre-tion) further reinforcing this notion.Other possible mediators associatedwith increased postural alterations andlight muscle activity may include hemo-dynamic changes (i.e., increased bloodvolume, tissue perfusion, and capillarypermeability [33–35]), as well as modula-tionof intracellular signaling changes (i.e.,AMPK, translocation/turnover of GLUT4,and calcium-activated proteins [36]), in-creased muscle insulin sensitivity, orchanges in sympathetic nervous systemactivity. It is also possible that the effectsof simply standing up more regularlycould have significantly contributed tothe observed metabolic effects (19), po-tentially via combinations of hemody-namic, hemodilutional, or minimal musclecontractile activity. While these mecha-nistic possibilities could not be ascertainedfrom our study design, they should be ex-amined in future research in populationswith T2D. This could involve longer sta-bilization periods in the standing posi-tion, prior to each activity bout, toaccount for any potential additive meta-bolic effect of the physical activity itself.Moreover, it remains to be determinedwhether different activity bout modali-ties provide unique physiological bene-fits and to what extent they inhibit theadverse consequences of prolonged sit-ting. Future studies should examine theoptimal frequency, duration, and inten-sity of activity bouts to determine thespecific translational potential for themanagement of T2D.

The nonsignificant reduction in tri-glyceride iAUC during the LW conditionis consistent with the intermittent walk-ing (1 min and 40 s every 30 min) condi-tion of Peddie et al. (18) in healthynormal-weight adults, who notably alsodemonstrated a significant lowering in

Table 1—Anthropometric, biochemical, physical activity, and dietary valuesduring the preexperimental period

SIT LW SRA

Weight (kg) 90.4 6 2.1 90.4 6 2.1 90.3 6 2.1

Preprandial levels*

Plasma glucose (mmol z L21) 8.0 6 0.3 8.1 6 0.3 8.1 6 0.3

Serum insulin (pmol z L21) 87.0 6 9.5 83.5 6 9.5 86.0 6 9.5

Serum C-peptide (pmol z L21) 974 6 58 924 6 58§ 984 6 58

Plasma triglycerides (mmol z L21) 1.7 6 0.2 1.9 6 0.2 1.8 6 0.2

Accelerometer data†

Daily wear time (min)Habitual period 822 6 21 877 6 22 841 6 2148-h restricted period 863 6 18 870 6 18 869 6 18

Sedentary time (min/day)Habitual period 559 6 19 570 6 19 535 6 1948-h restricted period 552 6 19 569 6 19 563 6 19

Physical activity time (min/day)Light intensityHabitual period 275 6 14 299 6 14 278 6 1448-h restricted period 306 6 15 296 6 15 302 6 15

Moderate-vigorousHabitual period 7 6 1 8 6 1 9 6 148-h restricted period 5 6 1 5 6 1 5 6 1

Diet‡Total energy intake (kcal/day) 2,069 6 80 2,079 6 80 2,094 6 80Total carbohydrate (energy %) 46.7 6 1.1 46.9 6 1.1 47.1 6 1.1Total fat (energy %) 31.3 6 1.0 31.8 6 1.0 31.9 6 1.0Total protein (energy %) 18.6 6 0.6 17.8 6 0.6 17.8 6 0.6

Data are means6 SEM. *Preprandial values based on average of two time points (21 and 0 h)immediately before the first meal. †Accelerometer data collected during habitual (free-living)days and the 48-h period preceding the trial condition. ‡Dietary intakes were assessed fromweighed/measured food records during the 48-h period before the trial condition, using dietaryanalysis software (FoodWorks; Xyris Software, Highgate Hill, Queensland, Australia). §LWsignificantly different from SIT and SRA (P , 0.05).

care.diabetesjournals.org Dempsey and Associates 969

triglycerides with a single 30-min bout ofmoderate-vigorous walking. However,the significant triglyceride reduction dur-ing the resistance activity condition in ourstudy is a novel finding and a first in pa-tients with T2D. Plausible reasons for thecomparatively lower triglyceride levelswith both walking and resistance activi-ties may be due to our trial condition oractivity bouts being of longer durationthan some previous studies, which isbroadly consistent with studies thathave observed reductions in triglyceridesthe next day after using both intermittentand continuous walking interventions(37,38); the less natural modality (subjec-tive comments from participants notreported), increased activity stimulus (in-tensity and energy expenditure), andlower respiratory exchange ratio duringthe SRA bouts (indicating a relative in-crease in lipid oxidation); differences inthe participants studied (healthy vs.overweight/obese patients with T2D);

or the nutritional composition (i.e.,higher glucose-to-fat ratio) of testmeals.

In light of epidemiological findingsdocumenting sex differences in the as-sociations of television-viewing timewithcardiometabolic biomarkers (39,40),our findings offer some initial experi-mental insights suggesting that (post-menopausal) women with T2D mayderive greater reductions in postprandialglucose thanmen by interrupting their sit-ting timewith LW. Althoughwe suspectedthat the greater glucose reductions maybe related to increases in activity intensity,we did not observe significant increasesin heart rate, oxygen consumption, or per-ceived exertion for women comparedwith men. Differences in adipose andlean body mass or other biological dispar-ities between men and women with T2Dcould be the potential basis for thesefindings. Future studies should continueto elucidate sex-specific effects.

Key strengths of our study include thefocus on both men and women withovert T2D, the well-described and stan-dardized trial condition lead-in periods(as illustrated by minimal variance inconfounder variables such as diet, phys-ical activity, and fasting metabolic levelsduring trial condition lead-in periods)through the use of weighed food re-cords and objectively measured physicalactivity, standardized calorimetric as-sessment of the two modes of activitybout, the strict behavioral supervisionand standardized feeding of a typicalWestern diet during experimental con-ditions (as opposed to less ecologicallyvalid test drinks), full retention of par-ticipants and minimal data loss, and thecollection of regular blood time pointsduring trial visits for more robust timecourse and iAUC calculations.

The present trial also has some limi-tations that future studies could ad-dress. First, the acute nature of the

Figure 3—Fasting and postprandial plasma glucose (A), serum insulin (B), serum C-peptide (C), and plasma triglyceride (D) concentrations measuredduring SIT (○) and sitting interrupted with 3-min LW (,) or SRA (C) bouts. Vertical dashed lines indicate timing of the breakfast (0 h) and lunch (3.5 h)meals. Data are presented as mean6 SEM.

970 Interrupting Sitting Time and Type 2 Diabetes Diabetes Care Volume 39, June 2016

current study precludes extrapolationsabout longer-term exposures to the par-ticular conditions that we examined. It ispresently unknown whether our ap-proach is beneficial over a longer periodof time or whether the putative benefitscan be sustained in ways that have pre-viously been shown in longer-term trialsthat included aerobic exercise (41). Sec-ond, although this well-controlled studyhas offered insights into the metabolicconsequences of prolonged sitting andthe incorporation of alternate modes ofintermittent activity bouts, generaliz-ability to free-living settings is less cer-tain and may not always reflect habitualbehaviors. For example, in theworkplace,other factors such as stress and workloadmay also play a role in glycemic homeo-stasis. Third, the exploratory sex analysesin this study should be interpreted withcaution, given the limited sample size,which may have increased the risk oftype 1 errors. Finally, the standardizedWestern dietary feeding profiles (42)used during this trial, while arguablymore reflective of real-world scenarios,will inevitably vary in daily life settings(e.g., macronutrient profile, glycemic in-dex, meal frequency, and size). Althoughour specific focus was on standardizedmeal responses to the sedentary and ac-tivity patterns, suchdietary variations andtheir interactions with physical activityare an integral piece of the puzzle andwill require further examination.Pragmatically, both activity condi-

tions were easily tolerated and well ac-cepted by our T2D participants (basedon subjective comments collected atthe end of each trial condition [datanot shown]), and it appears the benefi-cial metabolic effects of interruptingprolonged sitting can be achieved withdifferent modes of light-intensity activ-ity. In this regard, both LW and SRAbouts may have application irrespectiveof individual ability or workplace or homecontext encountered. For example, SRAbouts require no specialized equipment,only small amounts of space, and couldbe easily performed behind a work deskor at home with minimal disruption towork tasks or leisure pursuits, whereaslight walking may be more convenientand socially acceptable in certain con-texts. Longer-term engagement in eithermode of activity bout may also elicit dif-fering physiological effects (e.g., in-creased muscle strength, bone density,

and physical function) that are presentlyunknown. Taken together, while thisstudy provides a first piece of experimen-tal evidence on the potential benefits ofinterrupting prolonged sitting in T2Dpatients, further mechanistic studies andinterventions in larger samples in ecolog-ically relevant, free-living, and workplaceenvironments, using a broader range ofparticipants (including premenopausalwomen and patients with less well-controlled T2D [e.g., patients with poorlycontrolled diabetes on insulin or sulfo-nylurea dependent, with b-cell dysfunc-tion and increased risk of experiencinghypoglycemia]), will be informative indeveloping more specific public healthguidelines for the management of T2D.

In conclusion, interrupting prolongedsitting with brief LW or SRA bouts signif-icantly attenuates postprandial glucose,insulin,C-peptide,andtriglycerideresponsesin adults with T2D. With the ubiquity ofsedentary behaviors and the low adher-ence to structured exercise, these twoapproaches are practical strategies thatmay contribute toward reducing the riskof diabetes complications and cardiovas-cular complications. The efficacy andsustainability of our particular approachshould be tested in larger and longer-duration trials, as has been done for aer-obic exercise interventions in the T2Dcontext (41). Nonetheless, our findingscontribute complementary initial exper-imental evidence to further inform theexisting, albeit broad, T2D exercise rec-ommendations to “increase daily move-ment through unstructured activity togain additional health benefits” (1).Thus, in addition to the essential promo-tion of purposeful moderate-vigorousand leisure-timephysical activity, it seemsprudent and nonmaleficent (primum nonnocere) that health care professionalsconsider promoting the message, or pro-viding prescriptive advice to T2D patients,to also regularly interrupt prolonged sit-ting time.

Acknowledgments. The authors are gratefulfor the excellent technical assistance from IanMullis, Hayley Moon, and Donna Vizi (researchnurses); Alaina Natoli (sample analysis); andFrancis Dillon (data cleaning), from BakerIDI Heart and Diabetes Institute. Most impor-tantly, the authors thank the study participantsfor their time and commitment to the studyprotocol; this study would not have been pos-sible without them.

Funding. This research was supported by Na-tional Health and Medical Research Council(NHMRC) project grant 1081734 and the Victo-rian Government Operational Infrastructure Sup-port scheme. P.C.D. is supported by an AustralianPostgraduate Award. J.W.S., G.W.L., N.O., B.A.K.,and D.W.D. are supported by the NHMRC Fellow-ships scheme. E.C. is supportedby anAustralianRe-search Council Future Fellowship (FT140100085).Duality of Interest. No potential conflicts ofinterest relevant to this article were reported.Author Contributions. P.C.D. conceived, de-signed, and conducted the study; analyzedand interpreted data; andwrote themanuscript.R.N.L., P.S., J.W.S., N.E.S., N.D.C., E.C., G.W.L.,N.O., B.A.K., and D.W.D. assisted in the conceptand design of the study and participated incritical revision of the manuscript for intellec-tual content. P.S. and E.C. assisted with datacleaning or management and statistical analy-ses or interpretation. N.D.C. provided clinicalsupport during data collection. All authorsapproved the final version of the manuscript.P.C.D. and D.W.D. are the guarantors of thiswork and, as such, had full access to all the datain the study and take responsibility for theintegrity of the data and the accuracy of thedata analysis.Prior Presentation. Parts of this study werepresented in abstract form at the 62nd AnnualScientific Meeting of the American College ofSportsMedicine, San Diego, CA, 26–30May 2015.

References1. Colberg SR, Albright AL, Blissmer BJ, et al.;American College of Sports Medicine; Ameri-can Diabetes Association. Exercise and type 2diabetes: American College of Sports Medi-cine and the American Diabetes Association:joint position statement. Exercise and type 2diabetes. Med Sci Sports Exerc 2010;42:2282–23032. Zhao G, Ford ES, Li C, Mokdad AH. Compliancewith physical activity recommendations in USadultswithdiabetes.DiabetMed2008;25:221–2273. Matthews CE, Chen KY, Freedson PS, et al.Amount of time spent in sedentary behaviors inthe United States, 2003-2004. Am J Epidemiol2008;167:875–8814. Sedentary Behaviour Research Network. Let-ter to the editor: standardized use of the terms“sedentary” and “sedentary behaviours”. ApplPhysiol Nutr Metab 2012;37:540–5425. Biswas A, Oh PI, Faulkner GE, et al. Sedentarytime and its association with risk for disease in-cidence, mortality, and hospitalization in adults:a systematic review and meta-analysis. Ann In-tern Med 2015;162:123–1326. Dempsey PC, Owen N, Biddle SJ, DunstanDW. Managing sedentary behavior to reducethe risk of diabetes and cardiovascular disease.Curr Diab Rep 2014;14:5227. Healy GN, Dunstan DW, Salmon J, et al.Breaks in sedentary time: beneficial associa-tions with metabolic risk. Diabetes Care 2008;31:661–6668. Healy GN, Matthews CE, Dunstan DW,Winkler EA, Owen N. Sedentary time and cardio-metabolic biomarkers in US adults: NHANES2003-06. Eur Heart J 2011;32:590–5979. Dunstan DW, Kingwell BA, Larsen R, et al.Breaking up prolonged sitting reduces postprandial

care.diabetesjournals.org Dempsey and Associates 971

glucose and insulin responses. Diabetes Care 2012;35:976–98310. Tikkanen O, Haakana P, Pesola AJ, et al.Muscle activity and inactivity periods duringnormal daily life. PLoS One 2013;8:e5222811. American Diabetes Association. Standardsof medical care in diabetes–2013. DiabetesCare 2013;36(Suppl. 1):S11–S6612. Weir JB. New methods for calculating met-abolic rate with special reference to proteinme-tabolism. J Physiol 1949;109:1–913. Mikines KJ, Sonne B, Farrell PA, Tronier B,Galbo H. Effect of physical exercise on sensitiv-ity and responsiveness to insulin in humans. AmJ Physiol 1988;254:E248–E25914. Freedson PS, Melanson E, Sirard J. Calibra-tion of the Computer Science and Applications,Inc. accelerometer. Med Sci Sports Exerc 1998;30:777–78115. Schofield WN. Predicting basal metabolicrate, new standards and review of previouswork. Hum Nutr Clin Nutr 1985;39(Suppl. 1):5–4116. van Dijk JW, Venema M, van Mechelen W,Stehouwer CD, Hartgens F, van Loon LJ. Effect ofmoderate-intensity exercise versus activities ofdaily living on 24-hour blood glucose homeosta-sis in male patients with type 2 diabetes. Diabe-tes Care 2013;36:3448–345317. Kenward MG, Roger JH. The use of baselinecovariates in crossover studies. Biostatistics2010;11:1–1718. Peddie MC, Bone JL, Rehrer NJ, Skeaff CM,Gray AR, Perry TL. Breaking prolonged sittingreduces postprandial glycemia in healthy, normal-weight adults: a randomized crossover trial. Am JClin Nutr 2013;98:358–36619. Henson J, Davies MJ, Bodicoat DH, et al.Breaking up prolonged sitting with standing orwalking attenuates the postprandial metabolicresponse in postmenopausal women: a ran-domized acute study. Diabetes Care 2016;39:130–13820. DiPietro L, Gribok A, Stevens MS, Hamm LF,Rumpler W. Three 15-min bouts of moderatepostmeal walking significantly improves 24-hglycemic control in older people at risk for im-paired glucose tolerance. Diabetes Care 2013;36:3262–326821. Blankenship JM, Granados K, Braun B. Ef-fects of subtracting sitting versus adding exer-cise on glycemic control and variability in

sedentary office workers. Appl Physiol NutrMetab 2014;39:1286–129322. Duvivier BM, Schaper NC, Bremers MA, et al.Minimal intensity physical activity (standing andwalking) of longer duration improves insulin ac-tion and plasma lipids more than shorter periodsof moderate to vigorous exercise (cycling) in sed-entary subjects when energy expenditure is com-parable. PLoS One 2013;8:e5554223. Holmstrup M, Fairchild T, Keslacy S,Weinstock R, Kanaley J. Multiple short boutsof exercise over 12-h period reduce glucose ex-cursions more than an energy-matched singlebout of exercise. Metabolism 2014;63:510–51924. Ceriello A, Hanefeld M, Leiter L, et al.Postprandial glucose regulation and diabeticcomplications. Arch Intern Med 2004;164:2090–209525. O’Keefe JH, Bell DS. Postprandial hyperglyce-mia/hyperlipidemia (postprandial dysmetabolism)is a cardiovascular risk factor. Am J Cardiol 2007;100:899–90426. Praet SF, Manders RJ, Meex RC, et al. Gly-caemic instability is an underestimated prob-lem in type II diabetes. Clin Sci (Lond) 2006;111:119–12627. van Dijk JW, Manders RJ, Hartgens F,Stehouwer CD, Praet SF, van Loon LJ. Postpran-dial hyperglycemia is highly prevalent through-out the day in type 2 diabetes patients. DiabetesRes Clin Pract 2011;93:31–3728. Cichosz SL, Fleischer J, Hoeyem P, et al. Ob-jective measurements of activity patterns inpeople with newly diagnosed type 2 diabetesdemonstrate a sedentary lifestyle. Diabet Med2013;30:1063–106629. Hamer M, Bostock S, Hackett R, Steptoe A.Objectively assessed sedentary time and type 2diabetes mellitus: a case–control study. Diabe-tologia 2013;56:2761–276230. Leenders M, Verdijk LB, van der Hoeven L,et al. Patientswith type 2 diabetes showa greaterdecline in muscle mass, muscle strength, andfunctional capacity with aging. J Am Med DirAssoc 2013;14:585–59231. Praet SF, van Loon LJ. Exercise: the brittlecornerstone of type 2 diabetes treatment. Dia-betologia 2008;51:398–40132. Thosar SS, Bielko SL, Mather KJ, JohnstonJD, Wallace JP. Effect of prolonged sitting andbreaks in sitting time on endothelial function.Med Sci Sports Exerc 2015;47:843–849

33. Hespel P, Vergauwen L, VandenbergheK, Richter EA. Important role of insulin andflow in stimulating glucose uptake in con-tracting skeletal muscle. Diabetes 1995;44:210–21534. Holmang A, Mimura K, Lonnroth P. Involun-tary leg movements affect interstitial nutrientgradients and blood flow in rat skeletal muscle. JAppl Physiol (1985) 2002;92:982–98835. Inyard AC, Chong DG, Klibanov AL, BarrettEJ. Muscle contraction, but not insulin, in-creases microvascular blood volume in thepresence of free fatty acid-induced insulin re-sistance. Diabetes 2009;58:2457–246336. Santos JM, Ribeiro SB, Gaya AR, AppellHJ, Duarte JA. Skeletal muscle pathways ofcontraction-enhanced glucose uptake. Int JSports Med 2008;29:785–79437. Miyashita M, Burns SF, Stensel DJ. Accu-mulating short bouts of brisk walking reducespostprandial plasma triacylglycerol concen-trations and resting blood pressure in healthyyoung men. Am J Clin Nutr 2008;88:1225–123138. Miyashita M, Park JH, Takahashi M, SuzukiK, Stensel D, Nakamura Y. Postprandial lipae-mia: effects of sitting, standing and walking inhealthy normolipidaemic humans. Int J SportsMed 2013;34:21–2739. Dunstan DW, Salmon J, Owen N, et al.;AusDiab Steering Committee. Physical activ-ity and television viewing in relation to riskof undiagnosed abnormal glucose metabo-lism in adults. Diabetes Care 2004;27:2603–260940. Thorp AA, Healy GN, Owen N, et al. Delete-rious associations of sitting time and televisionviewing time with cardiometabolic risk bio-markers: Australian Diabetes, Obesity and Life-style (AusDiab) study 2004-2005. Diabetes Care2010;33:327–33441. Knowler WC, Barrett-Connor E, Fowler SE,et al.; Diabetes Prevention Program ResearchGroup. Reduction in the incidence of type 2 di-abetes with lifestyle intervention or metformin.N Engl J Med 2002;346:393–40342. National Health and Medical ResearchCouncil. Nutrient Reference Values for Australiaand New Zealand, Including Recommended Di-etary Intakes. Canberra, Department of Healthand Ageing, National Health and Medical Re-search Council, 2006

972 Interrupting Sitting Time and Type 2 Diabetes Diabetes Care Volume 39, June 2016