Objectively Measured Physical Activity and Sedentary Behavior in Successful Weight Loss Maintainers
Funding agencies: This work was supported by grants from the National Institutes of Health (NIH K23 DK078913, P30 DK048520, and NIH UL1 TR001082), as well as from the American Heart Association (AHA 16PRE29660012).
Disclosure: KL is currently a consultant for PAL Technologies, the company that manufactures the activPAL device. JOH and HRW report stock options from Retrofit and are partners in Shakabuku, LLC, companies that provide weight management services, outside the submitted work. JOH and HRW have been issued a patent on the “Energy Gap.” HRW accepts personal fees as an advisory board member for Atkins, a low-carbohydrate weight loss program. The other authors declared no conflict of interest.
Author contributions: VAC, HRW, and JOH conceived of and designed the study. VAC, HRW, and JOH obtained funding. VAC wrote the protocol and acquired the data. KL cleaned and identified additional activity metrics with the activPAL data. ZP and DMO performed the statistical analysis. DMO, KL, ZP, JOH, ELM, and VAC interpreted the data. DMO and VAC drafted the manuscript. DMO generated tables and figures. All authors were involved in writing and revising the manuscript and approved the final version of the manuscript.
Abstract
Objective
The objective of this study was to compare patterns of objectively measured moderate-to-vigorous physical activity (MVPA, ≥ 3.00 metabolic equivalents [METs]), light-intensity physical activity (LPA, 1.50-2.99 METs), and sedentary behavior (SB, < 1.50 METs) in successful weight loss maintainers (WLMs), normal weight controls (NC), and controls with overweight/obesity (OC).
Methods
Participants (18-65 y) were recruited in three groups: WLM (maintaining ≥ 13.6-kg weight loss for ≥ 1 year, n = 30), NC (BMI matched to current BMI of WLM, n = 33), and OC (BMI matched to pre–weight loss BMI of WLM, n = 27). All participants wore the activPAL for 1 week.
Results
Compared with OC and NC, WLM spent more awake time in total MVPA (WLM: 9.6 ± 3.9%, NC: 7.1 ± 2.1%, OC: 5.9 ± 2.0%; P < 0.01) and more time in sustained (≥ 10 min) bouts of MVPA (WLM: 39 ± 33, NC: 17 ± 14, OC: 9 ± 11 min/d; P < 0.01). Compared with OC, WLM and NC spent more awake time in LPA (WLM: 29.6 ± 7.9%, NC: 29.1 ± 8.3%, OC: 24.8 ± 6.7%; P = 0.04) and less awake time sedentary (WLM: 60.8 ± 9.3%, NC: 63.8 ± 9.5%, OC: 69.3 ± 7.5%; P < 0.01).
Conclusions
Results provide additional data supporting the important role of MVPA in weight loss maintenance and suggest notable differences in LPA and SB between normal weight individuals and those with overweight/obesity. Increasing LPA and/or decreasing SB may be additional potential targets for weight management interventions.
Introduction
High levels of moderate-to-vigorous physical activity (MVPA) are strongly associated with successful long-term weight loss maintenance (1-4), and current guidelines recommend 300 min/wk of moderate-intensity activity (or 150 min/wk of vigorous-intensity activity) to prevent weight gain and sustain weight loss (5, 6). However, the role of light-intensity physical activity (LPA) and sedentary behavior (SB) in weight management is less clear.
LPA comprises activities that expend low levels of energy (1.50-2.99 metabolic equivalents [METs]). A recent study of adults with overweight/obesity in a behavioral weight loss intervention found that increases in objectively measured LPA were associated with improved short-term (6 mo) and long-term (18 mo) weight loss (7), suggesting an important role for LPA in weight loss maintenance that deserves further exploration.
SB comprises activities that expend very little energy (i.e., < 1.50 METs) and is typically associated with sitting, reclining, or lying during waking hours (8). On average, Americans spend ∼55% of awake time (7.7 h/d) engaged in SB (9). SB is associated with several poor health outcomes, including an increased risk of obesity (10-12) and weight gain (11, 13). However, the association between objectively measured SB and weight loss maintenance has not been evaluated.
Much of what is known about long-term weight loss maintenance comes from data collected from the National Weight Control Registry (NWCR), a prospective cohort study established in 1994. Entry criteria include maintenance of a ≥ 13.6-kg weight loss for ≥ 1 year (14). NWCR members engage in high levels of MVPA (15, 16) and spend a minimal amount of time watching television, one of the most common types of SB (17). However, patterns of LPA and SB in individuals successful at long-term weight loss maintenance have never been objectively measured or compared with the patterns of individuals of normal body weight or with overweight/obesity.
The primary aim of this study was to compare free-living patterns of objectively measured MVPA, LPA, and SB in successful weight loss maintainers (WLM) to the patterns of normal weight controls (NC) and controls with overweight/obesity (OC). We hypothesized that WLM would have higher levels of MVPA and LPA and lower levels of SB, as compared with both OC and NC.
Methods
Participants
This case control study was conducted at the University of Colorado Anschutz Medical Campus and was approved by the Colorado Multiple Institutional Review Board. Participants were studied between October 2009 and August 2012 to assess physical activity (PA) over a range of seasons.
Participants were recruited through campus flyers and email announcements. To enhance recruitment of WLM, a recruitment letter was also sent to NWCR members in the Denver metro area. Interested individuals underwent preliminary telephone screening to determine eligibility for one of three subject groups: WLM (maintaining ≥ 13.6-kg [30-lb] weight loss for ≥ 1 year, current BMI: 18-28 kg/m2), NC (BMI: 18-28 kg/m2) with no history of overweight/obesity, and OC (BMI > 27 kg/m2). A nested subject selection procedure achieved similar group means for age, gender, and the desired BMI matching.
Individuals meeting preliminary eligibility criteria were invited to attend an in-person screening visit. After providing informed written consent, a health history and physical examination were completed. Individuals were excluded if they had any physical or medical condition that restricted PA (including having diabetes, cardiovascular disease, cancer, or significant musculoskeletal, neurologic, or psychiatric disorders), had undergone bariatric surgery, were smokers, were not weight stable (> 5-kg fluctuation in body weight over past 6 mo), were taking weight loss medications or other medications known to affect appetite or metabolism, or were pregnant or lactating. Weight was measured with a calibrated digital scale (to the nearest 0.2 lb), and height was measured with a wall-mounted stadiometer (to the nearest 0.1 cm). Waist circumference was measured using a tape measure over the iliac crests. Eligible individuals were scheduled for a 1-week PA monitoring period.
Assessment of SB and PA
PA patterns were assessed by using the activPAL activity monitor (PAL Technologies, Glasgow, Scotland). The activPAL is a small (23 × 43 × 5 mm) and lightweight (10 g) device that uses accelerometer-derived information about thigh position to estimate time spent sitting/lying, standing, and stepping. The device is attached to the anterior thigh and is waterproofed by wrapping it in a nitrile sleeve, allowing for 24-hour measurement. Participants were asked to wear the device continuously for 7 days. Data were considered valid and were used for analysis if the device was worn for > 10 h/d on ≥ 4 days (including ≥ 2 weekdays and ≥ 1 weekend day).
The time-stamped “event” data file from the activPAL software was used to determine time spent sitting/lying, standing, and stepping per day. The activPAL has been validated for use in adults to distinguish between sitting/lying, standing, and stepping activities (18-21). The software uses a linear regression algorithm to (1) assign MET values for sitting/lying events (1.20 METs) and standing events (1.50 METs) and (2) assign MET values for stepping events by using a cadence-based algorithm. Several previous studies have validated the use of step cadence against indirect calorimetry to distinguish LPA and MVPA. This methodology has been reviewed in detail by Tudor-Locke and Rowe, who reported a very high correlation (r = 0.93) between mean step cadence and PA intensity assessed with indirect calorimetry on the basis of five treadmill/overground walking studies (22). Stepping events were categorized as LPA (1.50-2.99 METs) and MVPA (≥ 3.00 METs) by using 75 steps/min as a threshold (75 steps/min = 3.00 METs). Lyden et al. (23) recently reported very high accuracy for activPAL estimates of time in PA intensity category using this methodology. To control for differences in sleep time by expressing data in terms of the percentage of awake time, we visually inspected the event data file to estimate time into bed at night and time out of bed in the morning by using methods described previously (24).
A customized R program (www.r-project.org) was used to convert the event data file to a second-by-second data file to estimate additional metrics of SB (e.g., breaks in sedentary time, average duration of sedentary bouts) and time in PA intensity category (LPA and MVPA). MET hours were computed by multiplying the MET value for each activity by the duration it was performed (in hours). Daily MET hours were summed to compute total MET hours per day. Total MVPA (min/d) was computed as the sum of time spent in MVPA (stepping events ≥ 3.00 METs). To assess whether levels of MVPA met the 2008 Physical Activity Guidelines for Americans (6), which suggest that activity be accumulated in bouts ≥ 10 minutes, we calculated “guideline MVPA” as the sum of minutes in bouts lasting at least 10 minutes wherein > 80% of the entire bout was spent in ≥ 3.00 METs (5). Guideline MVPA was calculated as (a) average minutes per day (total min/d spent in guideline MVPA bouts), (b) average minutes per week (average min/d in guideline MVPA multiplied by 7), and (c) number of discrete guideline MVPA bouts per day. LPA (min/d) was computed from standing and stepping events with a MET value between 1.50 and 2.99 METs. The following metrics of SB were computed during time spent awake: total sedentary time (total time spent in sitting/lying events), total breaks in sedentary time (number of times a sitting/lying event was followed by a standing or stepping event), sedentary break rate (total number of breaks per total sedentary time in hours), time (min/d) in sedentary bouts ≥ 30 and ≥ 60 minutes, and number of discrete sedentary bouts ≥ 30 and ≥ 60 min/d.
Statistical analysis
Statistical analyses were performed with SAS (version 9.4, SAS Institute Inc.), with the type I error rate fixed at 0.05. Fisher's exact tests were used to compare categorical demographic characteristics across subject groups. The Shapiro-Wilk test was used to determine normality of outcome measures. If the Shapiro-Wilk test P < 0.05, data transformations were used. A square root transformation was used for minutes per day in guideline MVPA, percentage of hours awake worn in LPA, total breaks in sedentary time, minutes per day in sedentary bouts ≥ 60 minutes, and number of sedentary bouts ≥ 30 minutes. A log transformation was used for guideline MVPA bouts, minutes per week in guideline MVPA, and break rate. A Kruskal-Wallis test was used to compare the number of weekend days and weekdays the device was worn across subject groups. For all other variables, one-way analysis of variance (ANOVA) was used to examine the null hypothesis that samples in subject groups were drawn from populations with the same mean values. The omnibus F test P value was reported, followed by between-subject group comparisons. Results are presented as mean ± SD, unless otherwise stated. PA behavior was compared both in absolute minutes per day and as percentage of awake time (to account for differences in sleep time). A Pearson's correlation coefficient was used to examine correlations between total MVPA, LPA, sedentary time, and steps within each subject group. Results were not corrected for multiple comparisons because we were not concerned with the universal null hypothesis that subject groups were identical on all variables (25). There was no a priori power analysis for the outcomes variables in this secondary analysis. A sensitivity analysis was performed to ensure that seasonality during time of PA assessment did not impact between-group differences in PA patterns.
Results
Study enrollment and subject characteristics
One hundred fourteen participants were screened, and one hundred six completed the study (Figure 1). Fourteen participants were missing data because of device malfunction or incorrect device placement, and two participants did not meet valid monitoring period criteria (Figure 1), resulting in a final sample size of 90 participants (30 WLM, 33 NC, and 27 OC). Groups were similar in regard to age, sex, and ethnicity (Table 1). The current BMI of WLM (23.7 ± 2.4 kg/m2) was well matched to that of NC (22.7 ± 2.0 kg/m2, P = 0.22). The current BMI of OC (33.4 ± 5.1 kg/m2) was well matched to the pre-weight loss maximum BMI of WLM (32.8 ± 4.9 kg/m2, P = 0.84). WLM were maintaining a weight loss of 26.3 ± 11.6 kg for 9.5 ± 10.2 years.
Study enrollment.
| WLM (n = 30) | NC (n = 33) | OC (n = 27) | Overall P value | |
|---|---|---|---|---|
| Age (y), mean ± SD | 44.8 ± 11.8 | 46.8 ± 13.8 | 47.1 ± 11.0 | 0.74 |
| Anthropometric measures, mean ± SD | ||||
| Weight (kg) | 68.1 ± 11.0b | 64.1 ± 11.0b | 93.2 ± 18.7 | < 0.01 |
| Height (cm) | 169.0 ± 9.1 | 167.4 ± 9.8 | 166.6 ± 8.7 | 0.61 |
| BMI (kg/m2) | 23.7 ± 2.4b | 22.7 ± 2.0b | 33.4 ± 5.1 | < 0.01 |
| Waist circumference (cm) | 83.1 ± 7.3b | 82.6 ± 7.1b | 112.2 ± 39.7 | < 0.01 |
| Maximum weight (kg)d | 93.7 ± 15.9c | 68.5 ± 11.8b | 99.5 ± 21.9 | < 0.01 |
| Minimum weight (kg)e | 63.3 ± 11.3c | 57.2 ± 9.9b | 65.4 ± 15.3 | 0.03 |
| Maximum BMI (kg/m2) | 32.8 ± 4.9b, c | 24.3 ± 2.1b | 35.6 ± 6.1 | < 0.01 |
| Maximum weight ever lost (kg) | 26.3 ± 11.6b, c | 6.1 ± 5.1b | 12.2 ± 8.2 | < 0.01 |
| Weight loss maintenance duration (y) | 9.5 ± 10.2 | N/A | N/A | N/A |
| Sex (male), n (%) | 8 (27%) | 9 (27%) | 4 (15%) | 0.47 |
| Ethnicity, n (%) | 0.13 | |||
| Hispanic/Latino | 0 (0%) | 4 (12%) | 3 (11%) | |
| Not Hispanic/Latino | 30 (100%) | 29 (88%) | 24 (89%) | |
| Race, n (%) | 0.03 | |||
| White | 30 (100%) | 29 (88%) | 23 (85%) | |
| Black/African American | 0 (0%) | 1 (3%) | 4 (15%) | |
| Asian | 0 (0%) | 2 (6%) | 0 (0%) | |
| Not reported | 0 (0%) | 1 (3%) | 0 (0%) | |
- a Fisher's exact test was used for categorical variables; continuous variables were analyzed by using one-way ANOVA. Significant P values (α < 0.05) are indicated in bold.
- b Significantly different from OC (P < 0.05).
- c Significantly different from NC (P < 0.05).
- d Excluding pregnancy.
- e After age 18 and excluding illness.
- N/A, not applicable.
Wear time and sleep parameters
Median number of weekend days (2 d) and weekdays (4 d) worn, mean sleep time (8 h), and mean wake time (16 h) were similar across groups. There was a nonsignificant trend for longer sleep time in OC as compared with WLM and NC (Table 2). Thus, PA behavior was expressed both in absolute minutes per day and as a percentage of awake time. WLM were more likely to be assessed in the spring/summer versus fall/winter seasons, as compared with NC and OC (Supporting Information Table S1). There was no difference in PA behavior within subject group between the spring/summer versus fall/winter seasons (Supporting Information Table S2). In addition, when seasonality was added to the ANOVA model for total MVPA, guideline MVPA, LPA, sedentary time, and steps, results did not change (Supporting Information Table S3).
| WLM (n = 30) | NC (n = 33) | OC (n = 27) | P value, omnibus F test | P value, WLM:NC | P value, WLM:OC | P value, NC:OC | |
|---|---|---|---|---|---|---|---|
| Wear time and sleep | |||||||
| Sleep time (h/d) | 7.7 ± 1.5 | 7.9 ± 1.1 | 8.3 ± 1.3 | 0.21 | 0.49 | 0.08 | 0.25 |
| Wake time (h/d) | 16.3 ± 1.5 | 16.1 ± 1.1 | 15.7 ± 1.3 | 0.21 | 0.49 | 0.08 | 0.25 |
| MVPA and LPA | |||||||
| Total MVPA (min/d) | 95 ± 40 | 69 ± 20 | 56 ± 20 | < 0.01 | < 0.01 | < 0.01 | 0.07 |
| Total MVPA (% h awake worn) | 9.6 ± 3.9 | 7.1 ± 2.1 | 5.9 ± 2.0 | < 0.01 | < 0.01 | < 0.01 | 0.09 |
| Guideline MVPA (min/d)b | 39 ± 33 | 17 ± 14 | 9 ± 11 | < 0.01 | < 0.01 | < 0.01 | 0.04 |
| Guideline MVPA (min/wk) | 272 ± 234 | 117 ± 100 | 63 ± 76 | < 0.01 | 0.01 | < 0.01 | 0.18 |
| Guideline MVPA bouts (count/d)c | 1.4 ± 1.2 | 0.7 ± 0.6 | 0.4 ± 0.5 | < 0.01 | 0.01 | < 0.01 | 0.12 |
| MET hours per day | 26.2 ± 3.3 | 24.7 ± 2.0 | 23.3 ± 2.2 | < 0.01 | 0.02 | < 0.01 | 0.03 |
| LPA (min/d) | 290 ± 82 | 281 ± 85 | 234 ± 68 | 0.02 | 0.68 | 0.01 | 0.02 |
| LPA (% h awake worn)b | 29.6 ± 7.9 | 29.1 ± 8.3 | 24.8 ± 6.7 | 0.04 | 0.79 | 0.02 | 0.03 |
| Sedentary behavior | |||||||
| Sedentary time (min/d) | 596 ± 105 | 617 ± 102 | 654 ± 79 | 0.08 | 0.40 | 0.03 | 0.14 |
| Sedentary time (% of h awake worn) | 60.8 ± 9.3 | 63.8 ± 9.5 | 69.3 ± 7.5 | < 0.01 | 0.19 | < 0.01 | 0.02 |
| Sedentary bouts ≥ 30 min (min/d) | 300 ± 99 | 309 ± 111 | 346 ± 93 | 0.20 | 0.73 | 0.09 | 0.16 |
| Sedentary bouts ≥ 30 min (count/d)b | 5.3 ± 1.8 | 5.3 ± 1.5 | 6.1 ± 1.4 | 0.08 | 0.86 | 0.04 | 0.06 |
| Sedentary bouts ≥ 60 min (min/d)b | 143 ± 68 | 154 ± 89 | 169 ± 80 | 0.50 | 0.78 | 0.26 | 0.34 |
| Sedentary bouts ≥ 60 min (count/d) | 1.5 ± 0.7 | 1.6 ± 0.7 | 1.8 ± 0.8 | 0.24 | 0.88 | 0.13 | 0.16 |
| Sedentary breaks (count/d)b | 57.6 ± 13.1 | 58.0 ± 11.0 | 52.1 ± 10.9 | 0.11 | 0.84 | 0.09 | 0.05 |
| Sedentary break rate (break count/sedentary time in h)c | 6.08 ± 2.21 | 5.74 ± 1.17 | 4.84 ± 1.14 | 0.01 | 0.75 | 0.01 | 0.01 |
| Postural allocation and steps | |||||||
| Standing time (min/d) | 249 ± 73 | 245 ± 78 | 205 ± 63 | 0.04 | 0.82 | 0.02 | 0.03 |
| Stepping time (min/d) | 135 ± 48 | 105 ± 28 | 85 ± 25 | < 0.01 | < 0.01 | < 0.01 | 0.03 |
| Steps (count/d) | 12,256 ± 5,095 | 9,047 ± 2,703 | 7,072 ± 2,346 | < 0.01 | < 0.01 | < 0.01 | 0.04 |
- a Results are from one-way ANOVA. Significant P values (α < 0.05) are indicated in bold.
- b Results are from one-way ANOVA using a square root transformation, but data are presented using untransformed mean ± SD.
- c Results are from one-way ANOVA using a log transformation, but data are presented using untransformed mean ± SD.
MVPA
Compared with NC and OC, WLM spent more time in total and guideline MVPA, engaged in more bouts of guideline MVPA, accumulated more MET hours of PA (Table 2), and spent a greater percentage of time awake in MVPA (Figure 2). Extrapolating to a 1-week period for comparison to current PA guidelines, WLM engaged in more minutes per week of guideline MVPA than NC and OC (WLM: 272 ± 234, NC: 117 ± 100, OC: 63 ± 76 min/wk; Table 2).
Proportion of awake time spent in SB, LPA, and total MVPA across subject groups. Results from one-way ANOVA are reported; total n = 90 (WLM: 30, NC: 33, OC: 27). % Time awake in LPA results are from one-way ANOVA using a square root transformation, but data are presented using untransformed mean ± SD.
LPA
WLM and NC spent more total time in LPA (WLM: 290 ± 82, NC: 281 ± 85, OC: 234 ± 68 min/d; Table 2) and spent a greater percentage of time awake in LPA than OC (∼30% vs. ∼25%, respectively) (Figure 2).
SB
Both WLM and NC spent a lesser percentage of time awake sedentary (Table 2, Figure 2) and had a higher sedentary break rate than OC (Table 2). There was a trend for a between-group difference across subject groups in total sedentary time (WLM: 596 ± 105, NC: 617 ± 102, OC: 654 ± 79 min/d) and in the number of sedentary bouts ≥ 30 min/d, with WLM and NC having a lower number of bouts ≥ 30 min/d than OC (Table 2). There were no between-group differences in any other SB metrics.
Postural allocation and steps
Compared with OC, both WLM and NC spent more minutes standing per day. Compared with both NC and OC, WLM spent more time stepping and achieved more average steps per day (Table 2).
Correlations between PA and sedentary time categories
Sedentary time and LPA were negatively correlated within all subject groups. Steps and total MVPA were strongly correlated positively within all subject groups. Within NC, but not WLM or OC, LPA was positively (and sedentary time was negatively) correlated with total MVPA and steps (Table 3). In all groups combined, LPA was positively correlated (and sedentary time was negatively correlated) with total MVPA (LPA: r = 0.36 vs. sedentary time: r = −0.36) and steps (LPA: r = 0.40 vs. sedentary time: r = −0.39) and negatively correlated with sedentary time (r = −0.69), with all correlations significant at P < 0.05 (data not shown).
| Total MVPA (min/d) | LPA (min/d) | Sedentary time (min/d) | Steps (count/d) | |
|---|---|---|---|---|
| WLM | ||||
| Total MVPA (min/d) | 1.00 | 0.23 (P = 0.23) | −0.21 (P = 0.28) | 0.99 (P < 0.01) |
| LPA (min/d) | 0.23 (P = 0.23) | 1.00 | −0.63 (P < 0.01) | 0.26 (P = 0.16) |
| Sedentary time (min/d) | −0.21 (P = 0.28) | −0.63 (P < 0.01) | 1.00 | −0.25 (0.17) |
| Steps (count/d) | 0.99 (P < 0.01) | 0.26 (P = 0.16) | −0.25 (0.17) | 1.00 |
| NC | ||||
| Total MVPA (min/d) | 1.00 | 0.517 (P < 0.01) | −0.54 (P < 0.01) | 0.99 (P < 0.01) |
| LPA (min/d) | 0.52 (P < 0.01) | 1.00 | −0.77 (P < 0.01) | 0.49 (P < 0.01) |
| Sedentary time (min/d) | −0.54 (P < 0.01) | −0.766 (P < 0.01) | 1.00 | −0.53 (P < 0.01) |
| Steps (count/d) | 0.99 (P < 0.01) | 0.49 (P < 0.01) | −0.53 (P < 0.01) | 1.00 |
| OC | ||||
| Total MVPA (min/d) | 1.00 | 0.33 (P = 0.09) | −0.23 (P = 0.24) | 0.99 (P < 0.01) |
| LPA (min/d) | 0.33 (P = 0.09) | 1.00 | −0.53 (P < 0.01) | 0.36 (P = 0.07) |
| Sedentary time (min/d) | −0.23 (P = 0.24) | −0.53 (P < 0.01) | 1.00 | −0.27 (P = 0.17) |
| Steps (count/d) | 0.99 (P < 0.01) | 0.36 (P = 0.07) | −0.27 (P = 0.17) | 1.00 |
- a Results are from Pearson correlation, with significant r and P values (α < 0.05) indicated in bold.
Discussion
This study is the first to objectively assess multiple components of free-living activity behavior (MVPA, LPA, SB, postural allocation, and steps) in successful WLM, as compared with individuals with normal body weight (NC) and individuals with overweight/obesity (OC). Consistent with previous studies, our results suggest that successful WLM achieve higher levels of MVPA as compared with non–weight-reduced individuals (NC and OC). We also showed that individuals of normal body weight (NC and WLM) achieve higher levels of LPA and lower levels of SB than individuals with overweight/obesity (OC), suggesting that increasing LPA and/or decreasing SB may be additional potential targets for weight management interventions.
WLM engaged in twice the minutes per day of guideline MVPA when compared with NC and four times the minutes per day of guideline MVPA when compared with OC. Differences between subject groups in total MVPA appear to have been driven primarily by differences in minutes of guideline MVPA, as we observed previously (15). Results from the current study confirm results from previous observational studies (15, 16, 26-29) and randomized trials reporting on the association between MVPA and long-term (18 mo) weight loss (7, 30-32). Furthermore, these results are consistent with those from a previous study by our group (15) that compared MVPA (assessed with the RT3 accelerometer) in a sample of WLM (n = 26) with MVPA in NC (n = 30) and OC (n = 34) studied from 2004 to 2006. Time spent in guideline MVPA by subject group observed in that study (15) showed a similar pattern to the results we observed in the current study using the activPAL, with WLM spending the most time of the three groups in guideline MVPA, followed by NC, and then OC. Taken together, these results suggest WLM may require higher levels of guideline MVPA to maintain a normal body weight after weight loss than nonreduced individuals, as has been previously suggested (15, 16, 31).
We also observed higher levels of LPA and lower levels of SB in individuals maintaining a normal body weight (both NC and WLM) than in individuals with overweight/obesity (OC), which may have important implications for the prevention of weight gain (primary prevention) or the prevention of weight regain after weight loss (secondary prevention). On average, WLM engaged in ∼56 min/d more LPA and ∼58 min/d less sedentary time than OC. NC engaged in ∼47 min/d more LPA and ∼38 min/d less sedentary time than OC. There was an inverse correlation between time spent sedentary and time spent in LPA. Individuals with higher LPA had lower levels of sedentary time (overall r = −0.69, P < 0.01), and this correlation was qualitatively similar to (although not as strong as) correlations reported in Healy et al. (r = −0.96) (33).
Our results regarding SB are consistent with prior observational data (11, 12, 34), which suggest a potential role for reducing SB to support body weight regulation. For example, a cross-sectional study of 1,422 NWCR members provided evidence that avoidance of watching television (the most common type of SB) was associated with successful weight loss maintenance, independent of self-reported MVPA (17). Our results regarding LPA are novel, as there is currently limited epidemiological evidence on the association between LPA and body weight. To our knowledge, only one other study has evaluated the association between objectively measured LPA and weight loss maintenance. Jakicic et al. (7) examined data from 260 women with overweight/obesity enrolled in an 18-month behavioral weight loss program and found that those who achieved 10% weight loss at 18 months demonstrated greater increases in objectively measured LPA. Although these findings are intriguing, it is not clear whether higher levels of LPA and lower levels of SB are protective against weight gain or whether weight gain causes LPA to decrease and SB to increase.
WLM accumulated the most steps per day and spent the most time stepping, followed by NC, and then OC, suggesting that achieving a high number of steps per day may play an important role in the prevention of weight gain and weight regain after weight loss. Our results are consistent with those of Nakata et al. (35), who found that individuals in the highest quartile of weight loss 2 years after a 6-month intervention significantly increased their step count (by ∼2,607 steps/d) when compared with the lowest quartile of weight loss. In addition, WLM and NC spent ∼1.2 times the minutes per day standing that OC spent, suggesting that increasing time spent standing may help individuals to maintain a normal body weight. Furthermore, in the Nurses' Health Study (11), time spent standing or walking at home was associated with a 23% lower risk of obesity (P < 0.01). Breaking up sedentary time by increasing time spent standing and stepping may be an important additional recommendation for individuals seeking weight management. As Healy et al. (36) suggest, even activities as minimal as standing, rather than sitting, have been shown to result in substantial increases in total daily energy expenditure.
It is important to recognize that relative differences in MVPA between groups in our study were more dramatic than relative differences in LPA or SB, suggesting that the observed differences in MVPA are likely to play a larger role in weight loss maintenance than differences in LPA. For example, when comparing WLM with OC, the relative between-group differences in total MVPA (∼70%) were much higher than the relative between-group differences in LPA (∼24%) or sedentary time (∼9%). The estimated energetic effect of the additional 56 min/d of total LPA observed in WLM as compared with OC (56 min/d × LPA MET range of 1.50-2.99 METs = 84-167 MET min/d) is generally less than the estimated energetic effect of the additional 39 minutes of total MVPA observed in WLM as compared with OC (39 min × MVPA MET range of 3.0-10.0 METs = 117-390 MET min/d). However, the calculated ranges demonstrate that the exact energetic benefits depend on the mean intensity of the LPA and MVPA performed.
Although the energetic benefits of increasing MVPA are greater than an equivalent increase in duration of LPA, these benefits will not be accrued unless individuals actually achieve and sustain higher levels of MVPA. Increasing MVPA is a goal that is difficult for many adults with overweight/obesity to achieve, even when supported by a behavioral weight loss program (31, 32). Interventions to increase LPA and/or decrease SB may be more practical and universal across different settings (including home and the workplace) and may be more acceptable. Thus, interventions to increase LPA and/or decrease SB may ultimately promote greater energetic benefits than interventions designed to increase MVPA because of a greater level of adherence. Increasing LPA and/or decreasing SB may provide additional health benefits and contribute to weight management because of the cumulative energetic effects of LPA accrued throughout the day, as well as because of the metabolic benefits that may occur with reductions in sedentary time (36). A recent study concluded that in a group of sedentary adults, minimal-intensity PA (standing and walking) of longer duration improves insulin action and plasma lipids more than shorter periods of MVPA when energy expenditure is comparable (37), indicating that LPA may play a larger role in metabolic health than previously thought. An uncontrolled, observational study found that in participants with overweight/obesity who completed a behavioral weight loss program (mean ± SD weight loss of 6.7 ± 8.7 kg), enrollment in a 6-month fitness program focused on increasing both light-intensity (gardening, light housework) and moderate-intensity PA (walking) limited weight regain (38), indicating that the addition of LPA to MVPA recommendations may help promote long-term weight loss maintenance. Owen et al. noted that, “Every minute of sedentary time replaced with LPA would expend 1 additional kilocalorie (calculated assuming 1.50 vs. 2.30 METs for a person weighing 72 kg),” revealing the potential for several health improvements with a recommendation such as this (36, 39). As suggested first by Healy et al. (40), replacing SB with LPA may be a successful, additional approach to improving weight management. A practicable recommendation may be to replace 1 hour of SB with 1 hour of LPA every day. However, well-designed, prospective, interventional trials are needed to evaluate the effectiveness of interventions to increase LPA and/or decrease SB (in addition to or instead of meeting current guidelines for MVPA) during weight loss or weight loss maintenance to better understand the effectiveness of these recommendations as a weight management strategy.
Our study has some limitations. Because of the observational study design, we were unable to assess causality. Increasing MVPA and LPA and decreasing SB may be behavioral strategies that aid in maintaining a normal body weight. However, reverse causality may exist; as people gain weight, they may become less active (and more sedentary) because activity becomes more difficult. It is possible that the intensity of some activities was misclassified by the activPAL cadence-based algorithm. Our study results may not be generalizable, as our study population was relatively small and predominantly female (77%), white (91%), and non-Hispanic/Latino (92%). However, 24-hour objective assessment of multiple components of free-living activity allowed us to detect important distinctions in patterns of activity among successful WLM, normal weight individuals, and individuals with overweight/obesity, which is a novel aspect of our study.
Conclusion
We observed significantly higher levels of MVPA in WLM than in non–weight-reduced individuals (NC and OC), providing additional objective data to suggest that weight-reduced individuals may require a greater level of MVPA to maintain a normal body weight than individuals not maintaining a weight loss. Our results also suggest that individuals of a normal body weight (WLM and NC) engage in significantly more LPA and less SB than their counterparts with overweight/obesity (OC). Although the energetic benefits of increasing MVPA are likely greater than those of increasing LPA, current intervention strategies to increase MVPA are only modestly successful. Increasing LPA and/or decreasing SB may be additional potential targets for weight management interventions and should be tested in prospective studies.
Acknowledgments
We would like to thank all of our study subjects for their participation and Dr. Rena Wing for providing access to the NWCR to assist with recruitment of study participants.
