Abstract
Purpose:
To investigate associations between accelerometer-determined sedentary time (ST) in prolonged (≥30 minutes) and non-prolonged (<30 minutes) bouts with physical activity energy expenditure (PAEE) from doubly-labeled water (DLW). Additionally, associations between ST and body mass index (BMI) and waist circumference were examined.
Methods:
Data from n=736 women and n=655 men aged 43–82 years were analyzed. Participants wore the Actigraph GT3X for seven days on two occasions approximately 6 months apart, and the average of the measurements was used. PAEE was estimated by subtracting resting metabolic rate and the thermic effect of food from DLW estimates of total daily energy expenditure. Cross-sectional associations were analyzed using isotemporal substitution modelling.
Results:
Re-allocation of prolonged ST to non-prolonged was not associated with increased PAEE and only significantly associated with lower BMI (β = −0.57 kg/m2; 95% CI: −0.94, −0.20) and waist circumference (β = −1.61 cm; 95% CI: −2.61, −0.60) in men. Replacing either type of ST with light or moderate-to-vigorous physical activity was significantly associated with higher PAEE, and lower BMI and waist circumference in both women and men.
Conclusion:
Limiting time spent sedentary as well as decreasing ST accumulated in prolonged bouts may have beneficial effects on BMI and waist circumference. Replacing any type of ST with activities of light or higher intensity may also have a substantial impact on PAEE.
Keywords: Accelerometer, Body weight, Doubly-labeled water, Physical activity, Sedentary lifestyle, Waist circumference
Introduction
A large part of our waking day is spent sedentary, a behavior defined as engaging in activities while in a sitting or reclining posture and resulting in low energy expenditure (≤1.5 METs, metabolic equivalents) (1). Objectively-measured sedentary time (ST) is unfavorably associated with metabolic markers (2). When replacing sitting with standing in workplace interventions, distinct improvements in cardiometabolic risk factors have been reported (3,4). Evidence is emerging on the detrimental effects of ST accumulated in prolonged bouts and potential benefits of breaking up ST (5–9).
With hours in a day fixed, participating in one activity means not participating in another. Nevertheless, most previous studies have not looked at time spent in activities of different intensities as interdependent. By using isotemporal substitution modeling (10) it is possible to examine how increasing time spent in one activity while reducing time spent in another activity is associated with an outcome of interest. For example, replacing one hour of self-reported sitting time with purposeful exercise has been prospectively associated with reduced mortality (11). In studies using accelerometers, replacing ST with light-intensity physical activity (LPA) or moderate-to-vigorous physical activity (MVPA) has shown favorable associations with mortality, metabolic markers, waist circumference and body mass index (BMI) (12–17). Replacing prolonged ST (bouts of ST ≥30 min), with non-prolonged ST (bouts of ST <30 min), has also been associated with lower BMI and waist circumference (18,19).
Compared to being sedentary, standing or walking results in increased skeletal muscle activity (20), which results in more energy expended. Thus, findings that suggest replacing ST with LPA or MVPA is beneficially associated with BMI and waist circumference may be explained by increased energy expenditure. It is, however, unknown if there is a difference in energy expenditure between prolonged or non-prolonged ST.
We aimed to study the associations between accelerometer determined ST in prolonged (≥30 minutes) and non-prolonged (<30 minutes) bouts with physical activity energy expenditure (PAEE) from doubly-labeled water (DLW) in a large sample of women and men. We additionally examined associations between both types of ST and self-reported BMI and waist circumference. We used the isotemporal substitution approach to take into account the interdependency between time spent at different intensity levels of physical activity.
Methods
Study population
In 2010–2013, two validation studies of the food frequency questionnaire (FFQ) and the physical activity questionnaire (PAQ) used in the Nurses’ Health Study (NHS) (21), Nurses’ Health Study II (NHSII) (22) and the Health Professionals Follow-up Study (HPFS) (23), were conducted. The Women’s Lifestyle Validation Study (WLVS) included a subset of women from NHS and NHSII and data collection was performed in 2010–2012. The Men’s Lifestyle Validation Study (MLVS) data collection was performed in 2011–2013 and included a subset of men from HPFS and an additional subset of men from the local community who were members of the Harvard Pilgrim Health Care insurance plan. The participants in all three cohorts respond to biennial questionnaires and participants in WLVS and MLVS were a random sample of NHS/NHS II, and HPFS participants who completed the 2006/2007 FFQ and had previously provided blood in these cohorts. Eligibility criteria for participation included availability of broadband internet and no expected changes in habitual dietary intake or physical activity patterns. Individuals with a history of coronary heart disease, stroke, cancer, or major neurological disease were excluded. In total, 796 women and 671 men consented to participate.
The Institutional Review Board of the Harvard T.H. Chan School of Public Health and Brigham and Women’s Hospital approved the study.
Data Collection Procedures
As described in detail previously (24), data collected in WLVS and MLVS included: two administrations of the PAQ used in the large cohorts, one doubly-labeled water (DLW) measurement and two accelerometer (ActiGraph GT3X, Actigraph Corporation, Pensacola, FL) assessments. In brief, the two accelerometer measurements occurred approximately six months apart, and one of the accelerometer measurements was designed to occur during the same three-month window as the DLW measurement A detailed description and a figure of the WLVS and MLVS time line has been published previously (24). A sub-sample of both women (n=90) and men (n=104) had a repeat DLW measurement which allowed for assessment of within-person variability. For the purpose of this study, information on physical activity from the PAQs was not included. The total data collection period for each individual was approximately 12 months.
Accelerometers and a wear time diary were mailed to participants who were instructed to wear the monitor on the hip for seven days during all waking time, except while bathing or swimming. The accelerometer was not worn during bed time. Participants then mailed the accelerometer and diary back to the study center.
For DLW measurements, participants were mailed a dose of bottled DLW a few days before they underwent the protocol. Standard doses of 1.5g DLW per kg body weight were used. Participants provided a total of six 10 ml urine samples. Two samples were collected before intake of the DLW dose, two post-dose and two more after 10 to 14 days.
Data from the first DLW measurement and the daily average of valid days from the two accelerometer measurements were analyzed in the present study. Among the 796 women and the 671 men enrolled in WLVS and MLVS, respectively, only those with at least one accelerometer measurement with at least 4 days with ≥10 hours of wear per day were included (n=775 women and n=667 men). Further, women and men with missing DLW data (n=27 and n=7), a reported BMI <18.5 or ≥50 kg/m2 (n=9 and n=1) or missing information on waist circumference (n=3 and n=4) were excluded. In total, 736 women and 655 men were included in final analysis.
Accelerometer derived variables
For accelerometer measurements, an epoch length of 1 second was set. To increase the monitor sensitivity in the low frequency range, the low frequency extension was enabled (25). For this study, activity counts from triaxial vector magnitude were utilized, accelerometer measurements have been described in more detail previously (24). Non-wear time was determined using standard methods (26) and defined as ≥60 consecutive minutes with zero counts allowing for limited movement during up to two minutes with <200 counts/minute (27). Daily wear time was assessed by subtracting non-wear time from 24 hours and days with ≥10 hours of registered wear time were considered valid.
Sedentary behavior (<1.5 METs) was defined as consecutive minutes with <200 counts/minute (26) based on intensity alone. Minutes with 200–2689 counts were classified as light intensity (1.5–2.99 METs) and minutes of ≥2690 counts were classified as activity of moderate-to-vigorous intensity (≥3 METs) (28). Prolonged and non-prolonged ST was defined separately as bouts of ST ≥30 minutes or <30 minutes, respectively. Minutes per day spent in prolonged and non-prolonged ST, LPA, and MVPA were summarized and mean daily time spent in each intensity category was calculated across valid wear days.
DLW derived variables
Total daily energy expenditure (TDEE, kcal) was determined using isotope ratio mass spectrometry to analyze urine specimens for deuterium and oxygen-18 (29,30). PAEE (kcal) was assessed by subtracting resting metabolic rate (RMR) and the thermic effect from food from TDEE. RMR was calculated based on sex, age, weight and height using the Mifflin-St Jeor equation (31). A constant of 10% of TDEE was used to account for the thermic effect of food.
Anthropometric variables and covariate assessment
Body weight (kg) was self-reported at the time of the DLW assessment and used to account for differences in energy expenditure due to body size in models. Height and waist circumference were self-reported on the first PAQ. BMI (kg/m2) was derived from self-reported weight (kg) and height (m). In addition, the Alternative Healthy Eating Index (AHEI) was calculated using data from the FFQ to allow adjustment for diet in analysis of BMI and waist circumference. In brief, the AHEI is based on adherence to recommended levels of eleven different dietary components and has been described in detail elsewhere (32).
Statistical analysis
All analyses were run separately for women and men. Mean values and standard deviations (SD) were calculated to examine characteristics by quartiles of prolonged ST. Differences in characteristics between quartiles were assessed using one-way ANOVA.
Spearman correlation coefficients were calculated to assess the associations between accelerometer derived variables and PAEE, BMI, and waist circumference. Partial Spearman correlations were adjusted for age and MVPA as well as the AHEI (in analyses of BMI and waist circumference), and body weight (in analysis of PAEE). In addition, to account for random within-person variability in PAEE (33), de-attenuated Spearman correlations (34) were calculated for PAEE using information from the repeated DLW measurements in a sub-sample of participants (n=90 women and n=104 men).
Associations of accelerometer derived variables with PAEE, BMI and waist circumference were also examined using multivariable linear regression and three different models for each outcome: 1) Single activity models that modeled each activity separately, adjusted for confounders (models of PAEE were adjusted for age and body weight, while models of BMI and waist circumference were adjusted for age and the AHEI), 2) Partition models that examined each activity adjusted for every other activity and confounders from single activity models, and 3) Isotemporal substitution models (10) that examined the effect of substituting one activity for another by adjusting for time spent in each activity (except the activity which is being “replaced”), confounders and accelerometer wear time. The resulting regression coefficients represent the estimated difference in mean PAEE, BMI or waist circumference if the mean time of the omitted activity was reduced by 60 minutes and replaced by a 60-minute increase in the included activities when total time is fixed (isotemporal).
All analyses were performed using SAS statistical software, version 9.3 (SAS Institute Inc, Cary, North Carolina, United States). The significance level was set at p <0.05.
Results
Baseline characteristics of participants by quartiles of prolonged ST are presented in Table 1. The mean (SD) age, BMI, waist circumference and PAEE among women and men, respectively, were: 63.0 (9.4) and 68.1 (7.8) years, 26.6 (5.3) and 26.2 (3.7) kg/m2, 89.8 (13.2) and 97.3 (10.0) cm, and 718 (234) and 912 (329) kcal. Women and men in the highest quartile of prolonged ST were older, had a higher weight, BMI and waist circumference, as well as lower TDEE and PAEE compared to the lower quartiles. There was no difference in total wear time between quartiles, intuitively, participants in the highest quartile acquired more total ST and less time in LPA or MVPA.
Table 1.
Baseline Characteristics of Women in WLVS and Men in MLVS by Quartiles of Prolonged (≥30 minute bouts) Sedentary Time
| Women (n=736) | Q1 (n=184) | Q2 (n=184) | Q3 (n=184) | Q4 (n=184) | |
|---|---|---|---|---|---|
| mean (SD) | mean (SD) | mean (SD) | mean (SD) | P1 | |
| Age (at DLW measurement) | 59.8 (8.5) | 61.0 (8.7) | 64.5 (9.5) | 66.6 (9.2) | <.0001 |
| Weight (kg) | 67.8 (13.2) | 70.6 (14.5) | 71.9 (15.4) | 75.3 (16.9) | <.0001 |
| Height (cm) | 163.9 (7.0) | 164.5 (6.9) | 163.3 (6.8) | 163.3 (6.7) | .22 |
| BMI (kg/m2) | 25.2 (4.4) | 26.1 (5.2) | 26.9 (5.0) | 28.2 (6.1) | <.0001 |
| Waist circumference (cm) | 86.1 (11.1) | 88.6 (13.2) | 91.0 (12.9) | 93.6 (14.5) | <.0001 |
| Total daily energy expenditure (kcal/day) | 2,296 (349) | 2,233 (343) | 2,142 (370) | 2,126 (332) | <.0001 |
| Physical activity energy expenditure (kcal/day) | 823 (239) | 741 (234) | 672 (210) | 634 (209) | <.0001 |
| Resting Energy Expenditure2 (kcal/day) | 1,243 (168) | 1,269 (175) | 1,256 (201) | 1,279 (197) | .27 |
| Alternative Healthy Eating Index | 60.6 (12.4) | 58.8 (11.9) | 59.9 (11.8) | 58.2 (12.0) | .24 |
| Accelerometer derived variables3 | |||||
| Total wear time (h/day) | 15.1 (1.0) | 15.1 (1.0) | 15.0 (1.1) | 15.0 (1.1) | .27 |
| Prolonged sedentary time, ≥30 min bouts (h/day) | 0.8 (0.2) | 1.5 (0.2) | 2.1 (0.2) | 3.3 (0.8) | <.0001 |
| Non-prolonged sedentary time, <30 min bouts (h/day) | 5.9 (1.0) | 6.3 (0.9) | 6.1 (0.9) | 5.9 (0.9) | .0003 |
| Light physical activity (h/day) | 7.5 (1.1) | 6.7 (1.0) | 6.3 (0.9) | 5.3 (1.1) | <.0001 |
| Moderate-to-Vigorous physical activity (h/day) | 0.9 (0.5) | 0.7 (0.4) | 0.6 (0.4) | 0.5 (0.3) | <.0001 |
| Men (n=655) | Q1 (n=163) | Q2 (n=164) | Q3 (n=164) | Q4 (n=164) | |
|---|---|---|---|---|---|
| mean (SD) | mean (SD) | mean (SD) | mean (SD) | P1 | |
| Age (at DLW measurement) | 66.3 (8.4) | 67.4 (7.8) | 69.4 (6.3) | 69.1 (8.0) | <.0005 |
| Weight (kg) | 77.9 (10.8) | 80.5 (10.4) | 82.0 (11.8) | 87.2 (15.1) | <.0001 |
| Height (cm) | 176.3 (6.7) | 176.8 (6.5) | 176.3 (5.9) | 178.0 (6.8) | .055 |
| BMI (kg/m2) | 25.1 (3.0) | 25.8 (3.1) | 26.4 (3.4) | 27.5 (4.5) | <.0001 |
| Waist circumference (cm) | 92.8 (8.2) | 96.6 (8.4) | 97.9 (9.3) | 102.0 (11.6) | <.0001 |
| Total daily energy expenditure (kcal/day) | 2,915 (458) | 2,795 (416) | 2,730 (408) | 2,674 (380) | <.0001 |
| Physical activity energy expenditure (kcal/day) | 1,069 (334) | 938 (326) | 877 (312) | 763 (267) | <.0001 |
| Resting Energy Expenditure2 (kcal/day) | 1,555 (151) | 1,578 (136) | 1,580 (148) | 1,644 (178) | <.0001 |
| Alternative Healthy Eating Index | 60.8 (13.5) | 61.7 (11.2) | 59.3 (13.3) | 58.5 (13.0) | .11 |
| Accelerometer derived variables3 | |||||
| Total wear time (h/day) | 15.1 (1.0) | 15.1 (1.1) | 15.2 (1.2) | 15.0 (1.1) | .67 |
| Prolonged sedentary time, ≥30 min bouts (h/day) | 1.1 (0.3) | 1.9 (0.2) | 2.7 (0.2) | 4.0 (0.9) | <.0001 |
| Non-prolonged sedentary time, <30 min bouts (h/day) | 6.0 (0.9) | 6.2 (1.0) | 6.2 (1.0) | 5.8 (0.9) | <.0001 |
| Light physical activity (h/day) | 6.9 (1.0) | 6.0 (0.9) | 5.5 (0.9) | 4.6 (0.9) | <.0001 |
| Moderate-to-Vigorous physical activity (h/day) | 1.1 (0.5) | 0.8 (0.5) | 0.7 (0.4) | 0.7 (0.4) | <.0001 |
One-way ANOVA
Calculated based on sex, age, weight and height using the Mifflin-St Jeor equation
Triaxial thresholds: Sedentary time <200 counts/min: Light PA = 200 – 2659 counts/min; MVPA ≥2690 counts/min. Non-wear time was defined as ≥60 consecutive minutes with zero counts allowing for limited movement during up to two minutes with <200 counts/min
Both women and men spent the majority of their measured waking time sedentary; on average 7.9 (1.3) and 8.5 (1.4) h/day, corresponding to 52 and 56%, respectively. Most ST was accumulated in non-prolonged bouts with only 1.9 (1.0) and 2.4 (1.2) h/day, corresponding to 24 and 28% of the total ST in women and men, respectively, accumulated in prolonged bouts ≥30 minutes. The average number of prolonged sedentary bouts per day was 2.8 (1.1) for women and 3.4 (1.3) for men. Women had less prolonged and non-prolonged ST compared to men and spent more time in LPA (6.5 (1.3) vs. 5.8 (1.3) h/day, corresponding to 43 vs. 38%), but less time in MVPA (0.67 (0.42) vs. 0.83 (0.48) h/day, 4.5 vs. 5.5%).
Prolonged ST was significantly and more strongly correlated with PAEE (negative), BMI (positive) and waist circumference (positive), than non-prolonged ST among both women and men (Table 2). Multivariable-adjusted correlations between prolonged ST and outcomes were also significant with coefficients of −0.21 and −0.27 for PAEE, 0.16 and 0.18 for BMI and 0.15 and 0.26 for waist circumference, for women and men, respectively. For PAEE, multivariable-adjusted de-attenuated correlations were slightly stronger; −0.26 and −0.36. For non-prolonged ST, correlations with all outcomes were weak but statistically significant in unadjusted models for women. Following adjustment for covariates, only the correlation with PAEE remained significant with r = −0.10 in the multivariable-adjusted model and r = −0.15 in the multivariable-adjusted de-attenuated model. There was no significant correlation between non-prolonged ST and any of the outcomes in men.
Table 2.
Spearman Correlation Coefficients Between Prolonged (≥30 minute bouts) and Non-prolonged (<30 minute bouts) Sedentary Time and Physical Activity Energy Expenditure (PAEE, kcal), Body Mass Index (BMI, kg/m2), and Waist Circumference (cm) Among Women and Men in the WLVS and MLVS, respectively.
| PAEE1 | BMI2 | Waist circumference2 | ||||
|---|---|---|---|---|---|---|
| Women | r | p | r | p | r | p |
| Sedentary time (h/day) | ||||||
| Prolonged | ||||||
| Unadjusted | −0.32 | <.0001 | 0.23 | <.0001 | 0.25 | <.0001 |
| Multivariable adjusted | −0.21 | <.0001 | 0.16 | <.0001 | 0.15 | <.0001 |
| De-attenuated3 multivariable adjusted | −0.26 | |||||
| (95% CI) | (−0.36 to −0.17) | |||||
| Non-prolonged | ||||||
| Unadjusted | −0.09 | .02 | 0.10 | .005 | 0.07 | .048 |
| Multivariable adjusted | −0.10 | .01 | 0.05 | .19 | 0.03 | .43 |
| De-attenuated3 multivariable adjusted | −0.15 | |||||
| (95% CI) | (−0.23 to −0.06) | |||||
| Men | ||||||
|---|---|---|---|---|---|---|
| Sedentary time (h/day) | ||||||
| Prolonged | ||||||
| Unadjusted | −0.35 | <.0001 | 0.24 | <.0001 | 0.32 | <.0001 |
| Multivariable adjusted | −0.27 | <.0001 | 0.18 | <.0001 | 0.26 | <.0001 |
| De-attenuated3 multivariable adjusted | −0.36 | |||||
| (95% CI) | (−0.45 to −0.27) | |||||
| Non-prolonged | ||||||
| Unadjusted | −0.01 | .81 | 0.05 | .24 | 0.03 | .40 |
| Multivariable adjusted | −0.01 | .80 | 0.005 | .90 | 0.01 | .77 |
| De-attenuated3 multivariable adjusted | −0.07 | |||||
| (95% CI) | (−0.16 to 0.03) |
Covariates in multivariable adjusted models: age (years), body weight (kg), and Moderate-to-Vigorous physical activity (MVPA, min/day), n=736 (women) and n=655 (men)
Covariates in multivariable adjusted models: age, the Alternative Healthy Eating Index, and MVPA, n=715 (women) and n=639 (men)
Measurement error correction based on repeat DLW measurement in women (n=90) and men (n=104)
Table 3 shows results from single activity and partition models. In single activity models, a negative association was seen between prolonged ST and PAEE, while positive associations were seen for BMI and waist circumference, all association were significant both among women and men. In partition models, prolonged ST was also significantly associated with higher BMI among both women and men as well as higher waist circumference among men only, but not associated with PAEE. Non-prolonged ST was significantly associated with lower PAEE in both women and men and higher BMI and waist circumference among women only in single activity models, but not associated with any of the outcomes in partition models. In single activity and partition models, significant positive associations were seen between LPA and MVPA with PAEE, while negative associations were seen with BMI, and waist circumference among both women and men.
Table 3.
Associations from single activity and partition models between 60 min per day spent performing prolonged (≥30 minute bouts) and non-prolonged (<30 minute bouts) sedentary time, light physical activity (LPA) and moderate-to-vigorous physical activity (MVPA), with physical activity energy expenditure (PAEE), body mass index (BMI), and waist circumference among women and men.
| Prolonged Sedentary | Non-prolonged Sedentary | LPA | MVPA | |||||
|---|---|---|---|---|---|---|---|---|
| β | 95% CI | β | 95% CI | β | 95% CI | β | 95% CI | |
| PAEE, kcal/day | ||||||||
| Women | ||||||||
| Single activity model1 | −72 | −90, −55 | −40 | −58, −22 | 74 | 62, 87 | 227 | 186, 268 |
| Partition model2 | 3.4 | −19, 25 | −3.1 | −20, 14 | 65 | 48, 82 | 191 | 150, 232 |
| Men | ||||||||
| Single activity model1 | −97 | −117, −77 | −33 | −58, −7.9 | 102 | 83, 120 | 254 | 203, 304 |
| Partition model2 | −27 | −55, 2.0 | −9.0 | −33, 15 | 69 | 43, 95 | 196 | 145, 247 |
| BMI, kg/m2 | ||||||||
| Women | ||||||||
| Single activity model3 | 1.47 | 1.07, 1.86 | 0.72 | 0.30, 1.13 | −1.16 | −1.45, −0.86 | −3.49 | −4.48, −2.51 |
| Partition model4 | 0.61 | 0.05, 1.17 | 0.27 | −0.17, 0.71 | −0.63 | −1.05, −0.21 | −2.39 | −3.43, −1.35 |
| Men | ||||||||
| Single activity model3 | 0.91 | 0.68, 1.13 | −0.02 | −0.31, 0.26 | −0.81 | −1.02, −0.59 | −1.80 | −2.40, −1.21 |
| Partition model4 | 0.42 | 0.07, 0.76 | −0.15 | −0.44, 0.15 | −0.43 | −0.74, −0.11 | −1.26 | −1.88, −0.63 |
| Waist circumference, cm | ||||||||
| Women | ||||||||
| Single activity model3 | 3.47 | 2.50, 4.44 | 1.39 | 0.36, 2.41 | −2.73 | −3.46, −2.00 | −8.93 | −11.35, −6.51 |
| Partition model4 | 1.17 | −0.21, 2.54 | 0.17 | −0.92, 1.26 | −1.65 | −2.69, −0.61 | −6.66 | −9.23, −4.09 |
| Men | ||||||||
| Single activity model3 | 2.78 | 2.17, 3.39 | 0.11 | −0.68, 0.90 | −2.51 | −3.09, −1.93 | −4.63 | −6.27, −3.00 |
| Partition model4 | 1.47 | 0.54, 2.41 | −0.13 | −0.93, 0.67 | −1.29 | −2.14, −0.43 | −2.69 | −4.38, −1.00 |
Covariates: age (years) and body weight (kg), n=736 (women) and n=655 (men)
Covariates: age, body weight, prolonged sedentary time (h/day), non-prolonged sedentary time (h/day), LPA (h/day) and MVPA (h/day), n=736 (women) and n=655 (men)
Covariates: age and the Alternative Healthy Eating Index (AHEI), n=715 (women) and n=639 (men)
Covariates: age, the AHEI, prolonged sedentary time, non-prolonged sedentary time, LPA, and MVPA, n=715 (women) and n=639 (men)
In isotemporal substitution models, replacing one hour of prolonged ST with one hour of non-prolonged ST was not associated with PAEE (kcal) in women (β = −7 (95% CI: −30, 17)) or men (β = −18 (95% CI: −13, 48)) (Figure 1). Further, replacing prolonged with non-prolonged ST was not associated with BMI or waist circumference in women but was significantly associated with lower BMI (−0.57 kg/m2) and lower waist circumference (−1.61 cm) in men. Reallocating one hour of prolonged ST to LPA was significantly associated with higher PAEE (61 and 96 kcal), lower BMI (−1.24 and −0.85 kg/m2), and lower waist circumference (−2.81 and −2.76 cm) in both women and men, respectively. Even stronger associations were seen when prolonged ST was substituted for MVPA; PAEE was higher (187 and 222 kcal), BMI was lower (−3.00 and −1.68 kg/m2), and waist circumference was lower (−7.83 and −4.16 cm) in both women and men, respectively.
Figure 1a-c.
Results from multivariable adjusted isotemporal substitution models of difference in mean outcome of a) physical activity energy expenditure (PAEE, kcal/day), b) body mass index (BMI, kg/m2), and c) waist circumference (cm), per one-hour replacement of prolonged (≥30 minute bouts) and non-prolonged (<30 minute bouts) sedentary time (ST), light physical activity (LPA), moderate-to-vigorous physical activity (MVPA) among women (―) and men (– – –) in the WLVS and MLVS-study, respectively. All models included all physical activity variables (except the activity which is being replaced) and accelerometer wear time, and additionally, models for PAEE were adjusted for age and body weight, and models for BMI and waist circumference were adjusted for age and the AHEI.
Replacing one hour of non-prolonged ST with LPA was significantly associated with higher PAEE (68 and 78 kcal), lower BMI (−0.90 and −0.28 kg/m2), and lower waist circumference (−1.82 and −1.16 cm) among women and men, respectively. Similar to results for prolonged ST, stronger associations were seen for all outcomes when replacing non-prolonged ST with MVPA. Finally, replacing one hour of LPA with MVPA was associated with significantly higher PAEE (126 and 127 kcal) and lower BMI (−1.76 and −0.83 kg/m2) among both women and men as well as lower waist circumference (−5.01 cm) among women only.
Discussion
This study investigated associations between prolonged and non-prolonged ST measured objectively using accelerometers and DLW determined PAEE as well as self-reported BMI and waist circumference. Replacing prolonged ST with non-prolonged was significantly associated with lower BMI and waist circumference in men but not women and was not associated with PAEE in either group. Replacement of either type of ST with light or moderate-to-vigorous physical activity was associated with higher PAEE, and lower BMI and waist circumference in both men and women.
Recently, studies showing benefits of breaking up ST on both metabolic markers and indicators of adiposity have emerged (5–9), implying that prolonged and non-prolonged ST may have different effects on health. Although still an emerging method, the isotemporal substitution approach has been used to study replacement of ST with higher intensity activities in relation to metabolic markers and indicators of adiposity (13,15–19,37) as well as mortality (11,12). Using this approach, Healy et al. (18) and Gupta et al. (19) differentiated between prolonged (≥30 minute bouts) and non-prolonged (<30 minute bouts) ST to examine associations with cardiometabolic risk markers and adiposity indicators.
In a study among adults with type 2 diabetes, Healy et al. (18) showed that replacing 30 minutes of prolonged ST with non-prolonged ST was significantly associated with a lower BMI. Further, replacing prolonged ST with LPA was associated with both lower BMI and waist circumference. Gupta et al. (19) investigated similar associations replacing prolonged sedentary time with time accumulated in brief sedentary bouts, and showed even stronger associations with lower BMI and waist circumference. These results are similar to our findings of replacing prolonged sedentary time with non-prolonged sedentary time among men. However, while point estimates for the same time reallocation among women in our study indicated both lower BMI and waist circumference, these were not statistically significant. Similar to Healy et al. (18) we also found that replacing one hour of either prolonged or non-prolonged ST with LPA was associated with significantly lower BMI and waist circumference.
Interestingly, Healy et al. (18) did not report any significant associations when replacing prolonged or non-prolonged ST with MVPA. Nonetheless, similar to our study, Gupta et al. (19) showed significantly lower BMI and waist circumference when replacing 30 minutes of prolonged ST with MVPA. In the current study, lower BMI and waist circumference were also seen when re-allocating non-prolonged ST to MVPA. Replacing ST with MVPA seemed to have a larger impact on all outcome variables compared to replacing ST with LPA. Nevertheless, focusing on the impact of LPA on BMI and waist circumference as well as PAEE may be a more feasible strategy to reduce ST in the general population. The difference between our results and those of Healy et al. may be explained by the larger sample size and higher levels of MVPA measured in our study. Results from other cross-sectional studies also suggest improvements in several metabolic biomarkers and lower BMI (14, 16, 17) and waist circumference (13, 16) when ST is replaced by MVPA.
Along with adding to the existing literature on benefits of replacing prolonged ST with activities of higher intensity on adiposity indicators, this study also examined associations with DLW measured PAEE. Standing or walking results in increased skeletal muscle activity compared to sitting or being in a reclined position (20). Additionally, breaking up ST with spontaneous movement and intermittent light intensity activity during sedentary behaviors, e.g. fidgeting, may result in substantial increases in energy expenditure (21). Intermittent muscle contractions may also have an effect on energy expenditure. However, our results showed no difference in PAEE when replacing prolonged ST with non-prolonged ST. As expected, PAEE was significantly higher when replacing one hour of ST, prolonged or non-prolonged, with LPA or MVPA. Although our results are cross-sectional, the differences in energy expended when substituting activities of lower intensity for higher intensity activities may partially explain the associations seen for BMI and waist circumference. Higher PAEE may result in less weight gain; substituting television viewing with activities of higher intensity has been associated with decreased weight gain over 6-years of follow up (10).
The cross-sectional design of our study is a limitation and prevents us from making inferences regarding causality of associations seen between ST and PAEE or indicators of adiposity. The 30-minute cut-point used to differentiate between prolonged and non-prolonged ST is somewhat arbitrary but breaking up ST every 30 minutes has been shown to have beneficial cardiometabolic effects in laboratory based settings (35). This cut-point has also been used by others (18,19,36). Reverse causation cannot be ruled out and having a high BMI/waist circumference may lead to low levels of physical activity and more time spent sedentary (37). However, Helajärvi et al. (38) looked at changes in television viewing time in relation to BMI and waist circumference longitudinally, and showed that the levels of television time were temporally antecedent to both outcomes. Individuals reporting a high amount of television viewing at all of the time points had more than twice the increase in BMI and waist circumference during follow-up compared to those reporting constant low levels. Nevertheless, future prospective studies using objective measures of sedentary behaviors are needed to assess the direction of this association. It is also important to acknowledge that estimates of substitution in the present study are model-based and do not reflect actual substitution.
Notable strengths of our study include a large sample of both men and women of a wide age range. Nevertheless, the study population consisted predominantly of white individuals of higher socioeconomic status and may therefore not be representative of the general population. However, the amount of time spent sedentary in the present study was similar (18,19) or slightly lower (14,15,39) than previous reports. The proportion of total ST spent in prolonged bouts was also similar to prior studies (18,39), although some studies have reported higher percentages (9,19,36). Assessment of additional lifestyle factors allowed for statistical adjustment although residual or unmeasured confounding cannot be ruled out. An additional limitation is that BMI and waist circumference were self-reported by study participants. However, participants in the WLVS and MLVS studies are experienced in self-reporting these variables during follow-up assessments in NHS, NHSII and the HPFS. Further, while the AHEI allows for adjustment of diet quality it does not take energy intake into account. Nevertheless, greater adherence to the AHEI has been associated with both lower BMI and more exercise (32) although we cannot rule out residual confounding from energy intake.
Objective assessment of ST and time spent at different intensity levels using accelerometers is a strength of our study. However, while the precision in measuring MVPA is high, measuring ST using the Actigraph GT3X has been shown to be less precise due to lack of information regarding posture, resulting in the inability to differentiate between sitting and standing (40,41). This is a limitation as these two behaviors may have different effects on health. Reallocating sitting to standing or stepping has been associated with lower BMI and waist circumference (42) and replacing sitting with standing in workplace interventions has been associated with improved HDL (3) and postprandial glucose excursion (4). The misclassification of time standing still as sedentary may therefore have contributed to smaller differences seen when reallocating ST to activities of higher intensity in our study. Standing has also been associated with increased skeletal muscle activity compared to a seated posture (20), which may have a more direct implication for associations between ST and PAEE. Additionally, it is possible that some periods of prolonged ST were misclassified as non-wear time. Consequently, when the threshold of ≥10 hours/day of accelerometer wear time to account for a valid day was applied, days with high amounts of prolonged ST may have been excluded from analysis, potentially contributing to more conservative effect estimates for prolonged ST. Another factor that may contribute to lower correlations is the fact that accelerometer and DLW measurements did not occur during the same time. However, collecting both measures concurrently overestimates the true long term correlation between these methods, which is usually the focus of epidemiologic studies with disease outcomes (43). An additional strength is repeated measurements of energy expenditure using DLW which allowed for adjustments of within-person variation in analysis of PAEE.
In conclusion, re-allocations of prolonged ST to non-prolonged ST was not associated with higher PAEE and was only significantly associated with lower BMI and waist circumference in men, but not women. Replacing ST, both prolonged and non-prolonged, with either LPA or MVPA was associated with higher PAEE, and lower BMI and waist circumference. Although cross-sectional, our findings provide preliminary evidence regarding the beneficial effects of limiting time spent sedentary as well as decreasing the number of prolonged sedentary bouts on BMI and waist circumference. Furthermore, these results may be important for future public health messages as for some individuals, particularly older adults, it may be more feasible to replace ST with LPA, rather than MVPA. Replacing ST with activities of light or higher intensity may have a substantial impact on PAEE and could be an important factor in weight maintenance. Nevertheless, prospective studies, which allow examination of the direction of associations, utilizing accelerometers are needed.
Acknowledgements
This study was supported National Institute of Health grants UM1- CA186107, UM1-CA176726, UM1-CA167552, P01-CA055075–18S1, U01-CA152904, and CA55075. Stephanie E. Bonn was funded by the Thorne-Holst foundation, and the Gålö foundation.
The authors thank Kate Clowry and Sean Sinnott for their help in coordination and data management of the WLVS and MLVS as well as the study staff for their devotion to detail and data quality.
Footnotes
Conflicts of interest
The authors declare no conflict of interest. The results of the present study are presented clearly, honestly, and without fabrication, falsification, or inappropriate data manipulation. The results of this study does not constitute endorsement by ACSM.
References
- 1.Tremblay MS, Aubert S, Barnes JD, et al. ; SBRN Terminology Consensus Project Participants. Sedentary Behavior Research Network (SBRN) – Terminology Consensus Project process and outcome. Int J Behav Nutr Phys Act. 2017;14(1):75. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Brocklebank LA, Falconer CL, Page AS, et al. Accelerometer-measured sedentary time and cardiometabolic biomarkers: A systematic review. Prev Med. 2015;76:92–102. [DOI] [PubMed] [Google Scholar]
- 3.Alkhajah TA, Reeves MM, Eakin EG, et al. Sit-stand workstations: a pilot intervention to reduce office sitting time. Am J Prev Med. 2012;43(3):298–303. [DOI] [PubMed] [Google Scholar]
- 4.Buckley JP, Mellor DD, Morris M, et al. Standing-based office work shows encouraging signs of attenuating post-prandial glycaemic excursion. Occup Environ Med. 2014;71(2):109–111. [DOI] [PubMed] [Google Scholar]
- 5.Healy GN, Matthews CE, Dunstan DW, et al. Sedentary time and cardio-metabolic biomarkers in US adults: NHANES 2003–06. Eur Heart J. 2011;32(5):590–597. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Healy GN, Dunstan DW, Salmon J, et al. Breaks in sedentary time: beneficial associations with metabolic risk. Diabetes Care. 2008;31(4):661–666. [DOI] [PubMed] [Google Scholar]
- 7.Dunstan DW, Kingwell BA, Larsen R, et al. Breaking up prolonged sitting reduces postprandial glucose and insulin responses. Diabetes Care. 2012;35(5):976–983. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Henson J, Yates T, Biddle SJ, et al. Associations of objectively measured sedentary behaviour and physical activity with markers of cardiometabolic health. Diabetologia. 2013;56(5):1012–1020. [DOI] [PubMed] [Google Scholar]
- 9.Jefferis BJ, Parsons TJ, Sartini C, et al. Does duration of physical activity bouts matter for adiposity and metabolic syndrome? A cross-sectional study of older British men. Int J Behav Nutr Phys Act. 2016;13:36. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Mekary RA, Willett WC, Hu FB, et al. Isotemporal substitution paradigm for physical activity epidemiology and weight change. Am J Epidemiol. 2009;170(4):519–527. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Matthews CE, Moore SC, Sampson J, et al. Mortality Benefits for Replacing Sitting Time with Different Physical Activities. Med Sci Sports Exerc. 2015;47(9):1833–1840. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Matthews CE, Keadle SK, Troiano RP, et al. Accelerometer-measured dose-response for physical activity, sedentary time, and mortality in US adults. Am J Clin Nutr. 2016;104(5):1424–1432. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Buman MP, Winkler EA, Kurka JM, et al. Reallocating time to sleep, sedentary behaviors, or active behaviors: associations with cardiovascular disease risk biomarkers, NHANES 2005–2006. Am J Epidemiol. 2014;179(3):323–334. [DOI] [PubMed] [Google Scholar]
- 14.Hamer M, Stamatakis E, Steptoe A. Effects of substituting sedentary time with physical activity on metabolic risk. Med Sci Sports Exerc. 2014;46(10):1946–1950. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Chastin SF, Palarea-Albaladejo J, Dontje ML, et al. Combined Effects of Time Spent in Physical Activity, Sedentary Behaviors and Sleep on Obesity and Cardio-Metabolic Health Markers: A Novel Compositional Data Analysis Approach. PLoS One. 2015;10(10):e0139984. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Knaeps S, De Baere S, Bourgois J, Mertens E, Charlier R, Lefevre J. Substituting Sedentary Time With Light and Moderate to Vigorous Physical Activity is Associated with Better Cardiometabolic Health. J Phys Act Health. 2018. March 1;15(3):197–203. [DOI] [PubMed] [Google Scholar]
- 17.Phillips CM, Dillon CB, Perry IJ. Replacement of Sedentary Time with Physical Activity: Effect on Lipoproteins. Med Sci Sports Exerc. 2018. >may;50(5):967–976. [DOI] [PubMed] [Google Scholar]
- 18.Healy GN, Winkler EA, Brakenridge CL, et al. Accelerometer-derived sedentary and physical activity time in overweight/obese adults with type 2 diabetes: cross-sectional associations with cardiometabolic biomarkers. PLoS One. 2015;10(3):e0119140. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Gupta N, Heiden M, Aadahl M, et al. What Is the Effect on Obesity Indicators from Replacing Prolonged Sedentary Time with Brief Sedentary Bouts, Standing and Different Types of Physical Activity during Working Days? A Cross-Sectional Accelerometer-Based Study among Blue-Collar Workers. PLoS One. 2016;11(5):e0154935. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Tikkanen O, Haakana P, Pesola AJ, et al. Muscle activity and inactivity periods during normal daily life. PLoS One. 2013;8(1):e52228. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Belanger CF, Hennekens CH, Rosner B, et al. The nurses’ health study. Am J Nurs. 1978;78:1039–1040. [PubMed] [Google Scholar]
- 22.Solomon CG, Willett WC, Carey VJ, et al. A prospective study of pregravid determinants of gestational diabetes mellitus. JAMA. 1997;278:1078–1083. [PubMed] [Google Scholar]
- 23.Grobbee DE, Rimm EB, Giovannucci E, et al. Coffee, caffeine, and cardiovascular disease in men. NEJM. 1990;323:1026–1032. [DOI] [PubMed] [Google Scholar]
- 24.Chomistek AK, Yuan C, Matthews CE, et al. Physical Activity Assessment with the ActiGraph GT3X and Doubly Labeled Water. Med Sci Sport Exerc. 2017;49(9):1935–1944. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Wanner M, Martin BW, Meier F, et al. Effects of filter choice in GT3X accelerometer assessments of free-living activity. Med Sci Sport Exerc. 2013;45(1):170–177. [DOI] [PubMed] [Google Scholar]
- 26.Tudor-Locke C, Camhi SM, Troiano RP. A catalog of rules, variables, and definitions applied to accelerometer data in the National Health and Nutrition Examination Survey, 2003–2006. Prev Chronic Dis. 2012;9:E113. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Aguilar-Farías N, Brown WJ, Peeters GM. ActiGraph GT3X+ cut-points for identifying sedentary behaviour in older adults in free-living environments. J Sci Med Sport. 2014;17(3):293–299. [DOI] [PubMed] [Google Scholar]
- 28.Sasaki JE, John D, Freedson PS. Validation and comparison of ActiGraph activity monitors. J Sci Med Sport. 2011;14(5):411–416. [DOI] [PubMed] [Google Scholar]
- 29.Schoeller DA, Luke AH. Rapid 18O analysis of CO2 samples by continuous-flow isotope ratio mass spectrometry. J Mass Spectrom. 1997;32(12):1332–1336. [DOI] [PubMed] [Google Scholar]
- 30.Schoeller DA, Colligan AS, Shriver T, et al. Use of an automated chromium reduction system for hydrogen isotope ratio analysis of physiological fluids applied to doubly labeled water analysis. J Mass Spectrom. 2000;35(9):1128–1132. [DOI] [PubMed] [Google Scholar]
- 31.Mifflin MD, St Jeor ST, Hill LA, et al. A new predictive equation for resting energy expenditure in healthy individuals. Am J Clin Nutr. 1990;51(2):241–247. [DOI] [PubMed] [Google Scholar]
- 32.Chiuve SE, Fung TT, Rimm EB, et al. Alternative dietary indices both strongly predict risk of chronic disease. J Nutr. 2012;142(6):1009–1018. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Schoeller DA, Hnilicka JM. Reliability of the doubly labeled water method for the measurement of total daily energy expenditure in free-living subjects. J Nutr. 1996;126(1):348–354. [PubMed] [Google Scholar]
- 34.Rosner B, Willett WC. Interval estimates for correlation coefficients corrected for within-person variation: implications for study design and hypothesis testing. Am J Epidemiol. 1988;127:377–386. [DOI] [PubMed] [Google Scholar]
- 35.Peddie MC, Bone JL, Rehrer NJ, et al. Breaking prolonged sitting reduces postprandial glycemia in healthy, normal-weight adults: a randomized crossover trial. Am J Clin Nutr. 2013;98(2):358–366. [DOI] [PubMed] [Google Scholar]
- 36.Diaz KM, Howard VJ, Hutto B, et al. Patterns of Sedentary Behavior in US Middle-Age and Older Adults: The REGARDS Study. Med Sci Sports Exerc. 2016;48(3):430–438. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Pedisic Z, Grunseit A, Ding D, et al. High sitting time or obesity: Which came first? Bidirectional association in a longitudinal study of 31,787 Australian adults. Obesity. 2014;22(10):2126–2130. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Helajärvi H, Rosenström T, Pahkala K,et al. Exploring causality between TV viewing and weight change in young and middle-aged adults. The Cardiovascular Risk in Young Finns study. PLoS One. 2014;9(7):e101860. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Shiroma EJ, Freedson PS, Trost SG, et al. Patterns of accelerometer-assessed sedentary behavior in older women. JAMA. 2013;310(23):2562–2563. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Kozey-Keadle S, Libertine A, Lyden K, et al. Validation of wearable monitors for assessing sedentary behavior. Med Sci Sports Exerc. 2011;43(8):1561–1567. [DOI] [PubMed] [Google Scholar]
- 41.Lyden K, Kozey-Keadle SL, Staudenmayer JW, et al. Validity of two wearable monitors to estimate breaks from sedentary time. Med Sci Sports Exerc. 2012;44(11):2243–2252. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Healy GN, Winkler EA, Owen N, et al. Replacing sitting time with standing or stepping: associations with cardio-metabolic risk biomarkers. Eur Heart J. 2015;36(39):2643–2649. [DOI] [PubMed] [Google Scholar]
- 43.Willett WC, Lenart EB. Chapter 6: Reproducibility and Validity of Food Frequency Questionnaires In: Willett WC, editor. Nutritional Epidemiology, 3rd edition. New York: Oxford University Press, 2013 [Google Scholar]