Content » Vol 46, Issue 4

Original report

Differences in fat loss in response to physical activity among severely obese men and women

Eivind Aadland, PhD1, Randi Jepsen, MSc1, John Roger Andersen, PhD1,2 and Sigmund Alfred Anderssen, PhD3,4

From the 1Sogn og Fjordane University College, Faculty of Health Studies, 2Department of Surgery, Førde Central Hospital, Førde, 3Norwegian School of Sport Sciences, Department of Sports Medicine, Oslo and 4Sogn og Fjordane University College, Faculty of Teacher Education and Sports, Sogndal, Norway

BACKGROUND: Direct measurement of physical activity may be critical to revealing its relationship with the change in fat mass over time. The aim of this study was to determine the relationship between accelerometer-determined change in physical activity and change in fat mass in severely obese men and women.

METHODS: A total of 49 severely obese men (n = 12) and women (n = 37) (mean age 42.8 years (standard deviation (SD) 9.6); mean body mass index 42.0 kg/m2 (SD 5.9) participated in a 1-year lifestyle intervention with 4 measurements of physical activity (Actigraph GT1M accelerometer), energy intake (180-item food frequency questionnaire), and body composition. Associations were determined using linear regression analyses.

RESULTS: In the total group change in both duration and intensity of physical activity were independently related to change in fat mass (partial r = –0.38 to –0.31, p = 0.001–0.007) after adjustment for baseline body weight, gender and change in energy intake. A gender-specific effect of change in physical activity duration was found for change in fat mass (p for change in physical activity duration*gender < 0.001), where the association was significant in men (partial r = –0.53, p < 0.001), but not in women (partial r = –0.17, p = 0.149).

CONCLUSION: Increased intensity of physical activity was associated with loss of fat mass in severely obese men and women, whereas physical activity duration was associated with fat mass loss in men.

Key words: exercise; accelerometer; obesity; intervention studies; gender; sex.

J Rehabil Med 2014: 46: 00–00

Correspondence address: Eivind Aadland, Sogn og Fjordane University College, Faculty of Health Studies, Box 523, NO-6803 Førde, Norway. E-mail: eivind.aadland@hisf.no

Accepted Nov 18, 2013; Epub ahead of print Feb 4, 2014

Introduction

Physical activity (PA) is the most important means of increasing energy expenditure. However, very modest weight loss is generally found to result from increased PA level (1, 2). It is likely that lack of compliance with prescribed PA cause both variation at an individual level and underestimation of the effectiveness of PA as a weight loss therapy on a group level (3, 4). Thus, accurate measurement of PA is critical in order to separate the question of whether the intervention was implemented successfully from that of whether the change in PA affects body weight (BW). Owing to the great variability between self-reported and objectively measured PA (5), PA must be measured directly. Previous studies have shown stronger relationships with BW and fat mass (FM) loss by using objective measures of PA (6) compared with self-reported PA (7, 8).

Men frequently achieve greater BW and FM losses in response to lifestyle interventions than do women (9–12). The apparent violation of the energy balance equation in women led Gleim (13) to conclude that it is “not prudent to tell a woman that she will lose weight if she exercises”, unless she “is committed to many hours of exercise per day”. However, this conclusion might not apply unless the actual PA level in such interventions is clearly determined. Nevertheless, evidence for a gender-specific effect of PA duration is suggested by 2 previous studies using self-reported PA (11, 14). We therefore performed a 1-year lifestyle intervention study in severely obese subjects with 4 measurements of energy intake (EI) and accelerometer-determined PA. The main aim was to determine the correlation between changes in PA and changes in BW and FM. A secondary aim was to test for gender-specific effects. We hypothesized an association between PA and loss of BW and FM, and that PA would lead to greater reductions in these variables in men than in women.

Patients and methods

Subjects

A total of 49 severely obese subjects participated in a lifestyle intervention at the Red Cross Haugland Rehabilitation Center (RCHRC) in Norway between February 2010 and October 2011. Inclusion criteria were: age 18–60 years, and body mass index (BMI) ≥ 40 kg/m2 with or without comorbidities, or ≥ 35 with comorbidities. Exclusion criteria were: pregnancy; heart disease; drug or alcohol abuse; previous bariatric surgery; and mental disorders or physical impairments that could reduce the subject’s ability to comply with the programme. Written informed consent was obtained from each subject before inclusion in the study. The study met the standards of the Declaration of Helsinki and was approved by the Regional Committee for Medical Research Ethics.

Intervention programme

The intervention consisted of 3 intermittent inpatient periods (the first was 6 weeks’ duration; the second and third were 3 weeks’ duration) over the course of 1 year. The time-line of the study was an inpatient period from baseline to 6 weeks, a home period of approximately 14 weeks, an inpatient period from week 20 to 23, a home period of approximately 27 weeks, and an inpatient period from week 50 to 53.

An interdisciplinary team of health professionals (dietician, nurse, physician, physiotherapist and exercise specialist) was responsible for the programme at RCHRC, which had 3 main components: diet, PA and cognitive behaviour therapy. Regarding cognitive behaviour therapy, subjects participated in weekly group-sessions targeting how to take control over their eating (minimize use of food for purposes other than nutrition) and cope with barriers to PA in various situations. In addition, individual sessions were offered if needed. Regarding PA and diet, both theoretical and practical sessions were incorporated. The main goal regarding diet was to introduce a healthy diet and eating pattern to provide each subject with tools to improve the quality of their diet in the long-term. Thus, each subject followed a high-fibre, low-fat, reduced-energy meal plan based on the Nordic Nutrition Recommendations (15), which included 3 main meals and 2–3 snacks each day. Some meals were served, others were prepared by the subjects during group lessons on how to prepare healthy meals. As such, no specific diet, meal replacements or severe energy-restriction was applied.

Regarding PA, subjects participated in a supervised and structured exercise programme consisting of 20–30 min brisk walking before breakfast and 2 × 45–60-min exercise sessions (e.g. swimming, walking, aerobics, dancing, ball games or strength training) performed individually or in groups 5 days per week at the RCHRC. Occasionally (approximately once a week), subjects performed a 1–2 h scheduled hike. All exercise sessions were led by a physiotherapist and/or an exercise specialist. No specific target regarding intensity was specified; however, intensity varied from moderate (continuous swimming or walking) to high (interval training or ball games). In addition, subjects were encouraged to perform PA on their own during the inpatient periods. Together with a team member, each subject developed a plan for PA to be performed at home, adjusted to the subject’s home situation and based on each subject’s opportunities and preferences. A standard programme was omitted during the periods at home, as the main goal of this plan was to develop a programme that would be sustainable in the long-term relative to each subject’s physical capacity, overall heath and context of living. The subjects were not offered any systematic follow-up at home, but they could contact the RCHRC or their general practitioner if needed. Thus, neither supervised exercise, nor any standard exercise programme was performed between the in-patient periods.

Procedures

BW and FM were measured every 3 weeks during the inpatient periods, which yielded a total of 7 measurements (baseline and weeks 3, 6, 20, 23, 50 and 53). PA level and diet were measured at 4 time-points. The first measurement was taken 1 month prior to the first inpatient period (baseline); the second was obtained during the first inpatient period (week 4/6), and the 2 last measurements were performed at home approximately 1 month before inpatient periods number 2 (week 16) and 3 (week 46), respectively.

BW and body composition were measured with subjects in the fasted state in light clothes in the morning after voiding. Both measures were reported to the nearest 0.1 kg using bioelectrical impedance analysis (BC 420 S MA, Tanita Corp, Tokyo, Japan). Notably, Tanita instruments provide valid estimates of changes in FM in obese subjects when compared with criterion-measured total body water (16, 17). In the present study the reliability of the FM measurements was assessed by performing measurements over 5 consecutive days in 15 subjects. A standard error of 0.47 kg was found.

PA was measured using an Actigraph GT1M accelerometer (Actigraph, Fort Walton Beach, FL, USA) and analysed with the ActiLife software v. 5.3. Subjects were instructed to wear the accelerometer over a 7-day period, except during water activities or while sleeping. The accelerometer was attached in the mid-axillary line of the right hip at the height of the umbilicus. A wear-time of ≥ 10 h/day for ≥ 4 days was used as the criterion for a valid measure, whereas periods of ≥ 60 min (allowing for ≤ 2 min of non-zero counts) were defined as non-wear time (18, 19). PA was reported as PA duration and PA intensity. Duration of PA was reported as minutes of moderate to vigorous PA per day (MVPA/day) (≥ 3 metabolic equivalents (METs)) using a previously defined cut-point of ≥ 852 counts/min in these severely obese subjects (20). The applied cut-point is slightly higher than that reported previously owing to the application of a slightly different model. The combined measure of MVPA/day was used because very few subjects were found to perform vigorous PA (≥ 6 METs; ≥ 5,200 counts/min). Self-reported duration of swimming, bicycling and resistance training were added to the accelerometer-determined minutes of MVPA/day, because these modes of activity are poorly captured by the accelerometer and comprised a large part of the programme at RCHRC.

Intensity of PA was reported as the mean counts/min in bouts of MVPA. The time spent in bouts of MVPA was defined as consecutive time in MVPA of ≥ 10 min in duration allowing for a decrease of ≤ 2 min below the cut-point. Diet was assessed using a validated 180-item food frequency questionnaire (21). Subjects were asked to report the average diet consumed over the last year at baseline, to report the diet over the last month for measurement number 2 and to report the diet over the last 3 months for measurements number 3–4. The questionnaires were analysed through computer scanning and manually checked for any items added at the Department of Nutrition, The University of Oslo, Norway.

Statistical analysis

The data on PA duration, PA intensity, EI, BW, FM and lean mass (LM) are presented as the mean values (standard deviation (SD)). The independent samples t-test was used to test for baseline differences between the genders, and the Man-Whitney test was used for comparing the groups on percentage PA assessed by self-report and in the attrition analyses.

A linear mixed model based on restricted maximum likelihood estimation was used in all analyses for changes (∆) over time. Distribution of observations was verified for each variable prior to analysis. Actual values were used to analyse changes over time in the combined group of men and women (observations were normally distributed). Post-hoc analyses were performed using the Sidak post-hoc test comparing all time-points with baseline. Differences between genders were tested using the ranks of the ∆-scores from baseline (because data were skewed within each gender group) by including time and gender in the model.

The association between ∆PA duration, ∆PA intensity and ∆EI (independent variables) and ∆BW, ∆FM and ∆LM (dependent variables) were analysed using ∆-scores between time-points (∆y1 = y2–y1; ∆x1 = x2–x1, etc.) (22). The standard repeated linear mixed model using actual values was omitted because the interpretation of the regression coefficients in such a model is difficult, owing to the mixing of longitudinal (within-subject) changes and the cross-sectional (between-subject) differences (22). Six regression models were run for each outcome variable: models 1–3 tested the bivariate relationships; model 4 contained ∆PA duration, ∆PA intensity and ∆EI; model 5 was similar to model 4, but also included gender and baseline BW; in model 6 we added interactions between ∆PA duration, ∆PA intensity, ∆EI and gender (∆PA duration*gender; ∆PA intensity*gender; ∆EI*gender) to model 5. To determine gender-specific associations the gender variable were re-coded (0 and 1 vs 1 and 0 for men and women, respectively) and the analyses repeated. Residuals were normally distributed in all models and no variables were collinear. The figure shows ∆PA duration vs residual ∆FM based on model 6, after exclusion of PA duration and insignificant interactions.

For body composition the differences between the measurements at baseline and week 6 (∆1), between week 6 and week 20 (∆2) and between week 20 and week 50 (∆3) were used in the analyses of associations (4 time-points). Because time-points were unequally spaced, ∆BW, ∆FM and ∆LM were scaled to changes per week to render the regression coefficients meaningful.

All analyses were performed using SPSS version 19.0 (SPSS Inc., Chicago, USA). A p-value ≤ 0.05 indicated statistically significant findings.

Results

Subject characteristics and attrition analysis

The characteristics of the subjects are shown in Table I. Eleven subjects (4 men and 7 women) were lost to follow-up during the intervention; 1 due to pregnancy, 4 dropped out of the programme (1 underwent bariatric surgery; 1 having reached his weight goal, and 2 for unknown reasons); 6 withdrew from the study (4 having problems with the study protocol and 2 for unknown reasons). In addition, a total of 16 observations of PA were excluded due to a failure to fulfil the validity criteria (2, 4, 5 and 5 at baseline, weeks 4, 16 and 46, respectively). Final sample sizes at each time-point are shown in Tables II and III. The subjects lost to follow-up were heavier (median 140.5 kg, 95% confidence interval (CI) 118.1–145.6 vs median 116.7 kg, 95% CI 113.5–124.2, p = 0.042) and exhibited reduced PA durations (median 74.9, 95% CI 33.0–92.7 vs median 98.3, 95% CI 89.6–108.0 min of MVPA/day, p = 0.045) compared with the study sample. The early responses (changes observed to the 6-week time-point) did not differ between the groups (p = 0.071–0.734).

Table I. Subject characteristics

Total

(n = 49)

Mean (SD) [n]

Men

(n = 12)

Mean (SD) [n]

Women

(n = 37)

Mean (SD) [n]

p-value for gender differences

Age

42.8 (9.6)

42.3 (8.2)

43.5 (10.1)

0.843

Height, cm

171.9 (9.2)

183.2 (8.2)

168.3 (6.0)

< 0.001

Weight, kg

123.9 (18.6)

135.2 (16.5)

120.2 (18.0)

0.014

WC, cm

128.3 (13.0)

131.6 (10.8)

127.2 (13.6)

0.321

BMI, kg/m2

42.0 (5.9)

40.4 (5.4)

42.5 (6.1)

0.305

Fat mass, kg

58.2 (11.7)

51.7 (9.4)

60.3 (11.7)

0.026

Lean mass, kg

65.5 (13.0)

83.5 (10.2)

59.8 (7.4)

< 0.001

VO2max, l/min

3.28 (0.71) [37]

4.22 (0.59) [9]

2.97 (0.41) [28]

< 0.001

VO2max, ml/kg/min

27.4 (5.4) [37]

32.7 (5.2) [9]

25.6 (4.3) [28]

< 0.001

WC: waist circumference; BMI: body mass index; VO2max: maximal oxygen consumption; SD: standard deviation.

Changes in physical activity level and energy intake

The duration of PA increased by 74.8% from baseline to 6 weeks (p < 0.001) (Table II). Changes did not differ from baseline at week 16 (5.8% increase, p = 0.782) or week 46 (15.6% increase, p = 0.215). Men increased their PA duration more than women (93.5%, 50.0% and 48.7% increases from baseline to weeks 6, 16 and 46, respectively, in men vs corresponding numbers in women of 73.7%, 2.2% and 1.8%, p = 0.032). The intensity of PA increased by 22.8% (p < 0.001) from baseline to week 4 and remained increased (22.5%, p = 0.001) at week 16. The change at week 46 was not different from baseline values (4.1% increase, p = 0.147). There were no gender difference (p = 0.709). At baseline and weeks 16 and 46, median duration of MVPA from self-report were 0.0–7.3% of total MVPA. The corresponding number at week 4 was 25.8%. The numbers were similar between genders (p > 0.108).

Table II. Change in duration and intensity of physical activity (PA) and energy intake (EI) during the 1-year intervention

Baseline

Mean (SD)

Week 4/6

Mean (SD)

Week 16

Mean (SD)

Week 46

Mean (SD)

p-value for effect of time

PA duration

94.2 (42.5)

164.7 (31.4)**

99.7 (33.0)

108.9 (38.3)

< 0.001

Men

75.2 (37.8)

161.0 (28.7)

107.7 (34.9)

113.4 (30.6)

Women

100.1 (42.7)

166.0 (32.6)

97.0 (32.5)

107.6 (40.9)

PA intensity

2,389 (742)

2,935 (738)**

2,870 (1018)**

2,620 (828)

< 0.001

Men

2,829 (1168)

3,303 (676)

3,260 (1484)

2,803 (682)

Women

2,286 (586)

2,809 (726)

2,736 (794)

2,564 (873)

EI

11.3 (4.3)

8.9 (2.7)**

9.1 (2.9)*

9.7 (3.0)

0.006

Men

11.1 (4.2)

9.5 (3.2)

9.8 (3.6)

9.6 (3.5)

Women

11.3 (4.4)

8.7 (2.6)

8.7 (2.6)

9.7 (2.9)

*p < 0.05; **p < 0.01.

Units are minutes of MVPA/day for PA duration, counts/min in bouts of MVPA for PA intensity and counts/min for total PA level; PA: physical activity; MVPA: moderate to vigorous physical activity; n for PA: total (men) = 42 (10), 43 (11), 36 (9) and 30 (7) at baseline, weeks 4, 16 and 46, respectively; n for EI: total (men) = 40 (9), 47 (12), 37 (11) and 34 (8) at baseline, weeks 4, 16 and 46, respectively.

SD: standard deviation.

EI was significantly decreased both at 6 weeks (15.9%, p = 0.002) and 16 weeks (16.8%, p = 0.016), whereas the change at week 46 (15.0% decrease) was not significant (p = 0.097). The changes in EI were similar in men and women (p = 0.291).

Changes in body weight and body composition

BW and FM decreased over time, and all time-points were significantly different from baseline (p < 0.001) (Table III). Men lost more weight (p = 0.015) and FM (p = 0.029) than women. Change in LM was only different from baseline at weeks 20 and 23 (p = 0.001) and week 53 (p = 0.026) and was similar in men and women (p = 0.182).

Table III. Change in body weight, fat mass and lean mass during the 1-year intervention

Baseline

Mean (SD)

Week 3

Mean (SD)

Week 6

Mean (SD)

Week 20

Mean (SD)

Week 23

Mean (SD)

Week 50

Mean (SD)

Week 53

Mean (SD)

p for change over time

Body weight, kg

123.9 (18.6)

120.2 (17.7)**

117.7 (17.5)**

113.1 (17.9)**

111.4 (17.9)**

114.6 (18.4)**

112.5 (18.1)**

< 0.001

Men

135.2 (16.5)

131.1 (15.7)

128.1 (15.9)

123.1 (16.4)

122.3 (16.9)

125.6 (15.0)

122.8 (15.4)

Women

120.2 (18.0)

116.6 (17.0)

114.2 (16.8)

110.1 (17.5)

108.2 (17.1)

110.8 (18.2)

109.0 (17.9)

Fat mass, kg

58.2 (11.7)

55.3 (12.4)**

53.0 (12.5)**

51.0 (12.1)**

49.5 (12.3)**

51.2 (11.9)**

50.0 (12.6)**

< 0.001

Men

51.7 (9.4)

47.6 (10.7)

44.6 (11.5)

44.2 (11.2)

41.9 (11.2)

45.5 (7.9)

43.4 (11.2)

Women

60.3 (11.7))

57.9 (11.9)

55.8 (11.7)

53.1 (11.8)

51.8 (11.8)

53.1 (12.5)

52.3 (12.5)

Lean mass, kg

65.6 (13.0

64.9 (12.7)

64.7 (12.9)

62.1 (11.4)**

61.9 (12.1)**

63.4 (12.3)

62.5 (11.9)*

< 0.001

Men

83.5 (10.2)

83.4 (8.6)

83.5 (8.4)

78.9 (7.1)

80.4 (7.8)

80.1 (9.2)

79.5 (7.1)

Women

59.8 (7.4)

58.7 (6.0)

58.4 (6.1)

57.1 (6.5)

56.3 (6.1)

57.6 (6.7)

56.7 (6.2)

*p < 0.05; **p < 0.01 compared with baseline.

Total n (men) = 49 (11), 48 (11), 48 (11), 43 (10), 43, (10), 35 (9) and 35 (9) at baseline, and weeks 3, 6, 20, 23, 50 and 53, respectively.

SD: standard deviation.

Associations between changes in physical activity, energy intake and body composition

Associations between behaviour changes and body composition are shown in Table IV. Changes in PA duration, PA intensity and EI were significantly related to ∆BW and ∆FM in the bivariate analyses. Whereas ∆PA duration and ∆PA intensity remained significantly related to ∆BW and ∆FM in the multivariable model (model 5), ∆EI was no longer related to ∆FM or ∆BW (p > 0.080). We found significant interactions (model 6) between gender and ∆PA duration for ∆FM (Fig. 1) and ∆LM (differences in slopes (women vs men) were 0.0064 (95% CI 0.0029–0.0098) kg/min MVPA/week for ∆FM and –0.0035 (95% CI –0.0062 to –0.0008) kg/min MVPA/week for ∆LM). The relationships between PA duration and ∆FM (partial r = –0.53, p < 0.001) and ∆LM (partial r = 0.31, p = 0.010) were significant in men, but not in women (∆FM: partial r = –0.17, p = 0.149; ∆LM: partial r = 0.08, p = 0.524)

Table IV. Associations (partial r) between change in behaviour (∆PA and ∆EI) and body composition (∆BW, ∆FM and ∆LM)

Model

Covariate

BW

FM

LM

Partial r

(p)

Partial r

(p)

Partial r

(p)

1

∆PA duration

–0.57

(< 0.001)

–0.60

(< 0.001)

0.05

(0.624)

2

∆PA intensity

–0.38

(< 0.001)

–0.35

(0.001)

–0.03

(0.793)

3

∆EI

0.37

(< 0.001)

0.35

(< 0.001)

0.06

(0.548)

4

∆PA duration

–0.34

(0.003)

–0.39

(0.001)

0.08

(0.522)

∆PA intensity

–0.29

(0.012)

–0.32

(0.005)

0.04

(0.736)

∆EI

0.23

(0.052)

0.16

(0.174)

0.10

(0.419)

5

Gender

–0.02

(0.873)

0.05

(0.701)

–0.08

(0.478)

BW

–0.07

(0.536)

0.03

(0.811)

–0.13

(0.261)

∆PA duration

–0.34

(0.003)

–0.38

(0.001)

0.06

(0.589)

∆PA intensity

–0.29

(0.013)

–0.31

(0.007)

0.03

(0.826)

∆EI

0.21

(0.080)

0.17

(0.164)

0.06

(0.610)

6

Gender

–0.01

(0.954)

0.04

(0.720)

–0.07

(0.592)

BW

–0.07

(0.557)

0.06

(0.599)

–0.18

(0.138)

∆PA duration

–0.32

(0.007)

–0.53

(< 0.001)

0.31

(0.010)

∆PA intensity

–0.17

(0.158)

–0.21

(0.087)

0.02

(0.827)

∆EI

–0.01

(0.928)

0.11

(0.358)

–0.16

(0.196)

∆PA duration*gender

0.18

(0.136)

0.41

(< 0.001)

–0.30

(0.011)

∆PA intensity*gender

–0.07

(0.591)

–0.04

(0.730)

–0.04

(0.755)

∆EI*gender

0.09

(0.443)

–0.02

(0.866)

0.16

(0.200)

Model 1–3 are bivariate; model 4 includes all behaviour changes; model 5 are adjusted for body weight and gender; model 6 adds interactions between gender and behaviour changes to model 5. Gender: 0 = men, 1 = women. n = 77 observations.

PA: physical activity; EI: energy intake; BW: body weight; FM: fat mass; LM: lean mass.

1887Fig1.tif

Fig. 1. Scatter-plot showing the relationship between the change in physical activity (PA) duration and change in residual fat mass (FM) in men and women after adjustment for gender, body weight, change in PA intensity, and change in energy intake. Solid circles and solid line = men; open squares and dotted line = women. Number of observations = 77. MVPA: moderate to vigorous PA; 95% CI: 95% confidence interval.

Discussion

There are 3 main findings from this study. First, PA showed a dose–response relationship with change in BW and FM in severely obese subjects. Secondly, PA duration induced a gender-specific effect on FM; men had greater losses for every minute of change in MVPA compared with women. Thirdly, whereas the men increased their free-living PA duration, women did not.

Although lifestyle interventions have been shown to lead to weight reduction in severely obese subjects (9, 14, 23, 24), there are few reports on the relationship between PA and weight reduction (14, 23). Although bariatric surgery is an effective weight-loss modality, the present study shows that severely obese subjects have the potential to lose weight if behaviour changes are incorporated, and demonstrates that a lifestyle intervention can be a viable option for some patients. Moreover, there is irrefutable evidence of the effectiveness of regular PA in the primary and secondary prevention of a range of chronic diseases, as well as premature death (25).

To the best of our knowledge, we are the first to present results showing a gender-specific FM change in response to change in PA duration using an objective measure of PA. However, two previous studies have shown similar results using self-reported PA (11, 14). Hollis et al. (11) demonstrated a significant interaction between gender and PA duration in the prediction of BW change in a large multicentre randomized controlled trial (n = 1,685) with a 6-month follow-up period. Furthermore, we have recently shown similar results in another small sample of severely obese subjects undergoing lifestyle treatment at RCHRC (14). In that study, the duration of PA, recorded with training diaries, explained 55.8% and 5.6% of BW loss for men and women, respectively. Thus, a gender-specific effect of PA duration seems to be a consistent finding in this severely obese population, as findings are similar whether PA is measured by self-report or by accelerometry.

Several hypotheses have been proposed to explain the greater weight loss response to PA in men compared with women (13, 26). First, an obvious cause is that men, because of their greater BW, spend more energy for every minute of PA than women. Evidence from the Midwest Exercise Trial suggests that differences found between genders when exercise is prescribed in terms of PA duration (10), disappears when men and women are matched on energy expenditure (27). However, we found a highly significant gender-specific effect of PA duration after adjustment for both BW and ∆PA intensity (i.e. given an equal BW and PA intensity in both gender groups). Thus, the present results suggest that PA duration might affect FM differently for men and women beyond differences in PA energy expenditure.

Secondly, fat distribution differs between men and women (28) and gender-specific fat distribution and fat tissue characteristics may explain the findings. Greater intra-abdominal FM in conjunction with a possibly greater adrenergic response to PA (29) in men may favour FM loss in men compared with women. Similarly, Kuk & Ross (30) showed that men lost more intra-abdominal fat and that women lost more lower-body subcutaneous fat for a given BW loss; however, total fat losses was similar between genders.

Thirdly, clustering or interaction between a host of factors may make weight loss extremely difficult for some individuals (26), especially women, as indicated by their more restrictive energy conservation (31, 32). In response to PA, women have been shown to compensate by increasing the EI corresponding to approximately 33% of the PA energy expenditure in the short-term, whereas no compensation is seen in men (31, 32). However, this is unlikely to explain our findings, as we adjusted for ∆EI during the intervention.

Fourthly, the gender-differences in slope for ∆FM could be due to an impaired within-subject precision of the body composition measurements in women over time, for example due to the menstrual cycle, which we did not assess. Alternatively, the clear gender-specific effect found using ∆FM vs ∆BW shows the importance of measuring body composition in weight loss trials, and confirms bioelectrical impedance analysis as a valid measurement tool.

We have no explanation for the interaction effect for ∆LM. It could be that men performed more strength training than women, as strength training can conserve LM in weight loss trials (33). However, few subjects performed resistance training, and we could not find any evidence for a gender difference. This is in line with a previous study at RCHRC (14).

In both genders there were increases in PA during the first in-patient period, but this level of PA seemed difficult for subjects to maintain at home, especially for women. This might be expected because simple behaviour change might be quite easily achieved outside one’s normal context of living and away from daily obligations, compared with the complex task of changing one’s way of living in the home environment. Close follow-up in the long term, e.g. weekly or biweekly contact during the periods at home (34), would probably be important to maintain an active lifestyle throughout the intervention.

It is well known that compliance with lifestyle interventions varies, which points to a crucial aspect regarding the reporting of such study results. We argue that measuring the actual behaviour changes is of critical importance to accurately identify which question one is answering; one of effectiveness or efficacy (35). The present study answered both questions; in terms of PA the implementation of the intervention was very moderately successful (the intervention had moderate effectiveness); however, PA proved to be important in terms of BW and FM loss (the efficacy of PA to promote weight loss was modest to good), at least in men. This is an important distinction that could potentially clarify some of the “mysterious” weight loss outcomes in response to lifestyle interventions.

Strengths and limitations

The present study has two main strengths. First, changes in PA were measured objectively using an accelerometer. Due to the highly inconsistent relationships found between self-reported and objectively measured PA (5), we believe that this is an important improvement over many previous weight-loss studies. Secondly, the behaviour changes were measured at multiple time-points during the 1-year intervention, showing the patterns of change over time. This allowed us to explore the effectiveness and efficacy of the intervention (35).

An important limitation of the present study was the relatively small sample size and the fact that men comprised only approximately a quarter of the sample, which limited the power to detect interactions and the ability to perform statistical analysis within each gender group. Although we detected significant gender-interactions (∆PA duration*gender) for ∆FM and ∆LM, inclusion of a larger and balanced sample could leave alternative results. Secondly, EI was assessed by self-report and might be flawed by measurement error (36, 37). The results indicated that EI was under-reported in both genders (results not shown), as expected (36). Although the reporting bias might be relatively stable over time (38) (i.e. leaving valid observations for changes over time), an increased underestimation over time might be promoted by the ongoing intervention due to increased social desirability (37). Thirdly, accelerometers have certain limitations. Importantly, they do not capture some commonly performed modes of PA, such as swimming, bicycling or strength training. Therefore, measurements of PA were corrected for self-reported duration of these modes of activity. Moreover, the attachment of the accelero­meters can be challenging for severely obese subjects, and tilting of the instrument is known to reduce the level of counts (39). Together with great variability in gait patterns between subjects, greater measurement variability may be expected in this population than would have been found in other populations. Finally, our results are based on associations and are vulnerable for residual confounding. Thus, causality may not be inferred from our findings.

Further research

Further research is needed to target possible differences between genders in the weight loss field, as previously recommended by others (40), as this might facilitate the development of tailored lifestyle intervention programmes. Furthermore, we recommend the objective measurement of PA level in weight loss trials. Finally, research should target strategies to implement and maintain behaviour changes in free-living obese subjects.

Conclusion

We found that increased duration and intensity of PA were associated with BW and FM loss in severely obese men and women undergoing a 1-year lifestyle intervention. Thus, we conclude that PA can improve weight status, if actually carried out. However, increased PA duration induced greater FM loss in the men compared with the women. Although this finding must be interpreted with caution due to the small sample size, it indicates that men may have a higher probability of succeeding in lifestyle interventions for obesity.

Acknowledgements

The authors thank the staff and participants at the Red Cross Haugland Rehabilitation Centre for their excellent collaboration and cooperation during the collection of data.

The study was funded by The Western Norway Regional Health Authority

The authors declare no conflicts of interest.

References

1.Wu T, Gao X, Chen M, van Dam RM. Long-term effectiveness of diet-plus-exercise interventions vs. diet-only interventions for weight loss: a meta-analysis. Obes Rev 2009; 10: 313–323.

2. Franz MJ, VanWormer JJ, Crain AL, Boucher JL, Histon T, Caplan W, et al. Weight-loss outcomes: a systematic review and meta-analysis of weight-loss clinical trials with a minimum 1-year follow-up. J Am Diet Assoc 2007; 107: 1755–1767.

3. Jakicic JM, Marcus BH, Gallagher KI, Napolitano M, Lang W. Effect of exercise duration and intensity on weight loss in overweight, sedentary women: a randomized trial. JAMA 2003; 290: 1323–1330.

4. Tate DF, Jeffery RW, Sherwood NE, Nancy E, Wing RR. Long-term weight losses associated with prescription of higher physical activity goals. Are higher levels of physical activity protective against weight regain? Am J Clinl Nutr 2007; 85: 954–959.

5. Prince SA, Adamo KB, Hamel ME, Meghan E, Hardt J, Gorber SC, Tremblay M. A comparison of direct versus self-report measures for assessing physical activity in adults: a systematic review. Int J Behav Nutr Phys Act 2008; 5: 56.

6. Colley RC, Hills AP, O’Moore-Sullivan TM, Hickman IJ, Prins JB, Byrne NM. Variability in adherence to an unsupervised exercise prescription in obese women. Int J Obes 2008; 32: 837–844.

7. Jakicic JM, Wing RR, Winters-Hart C. Relationship of physical activity to eating behaviors and weight loss in women. Med Sci Sports Exerc 2002; 34: 1653–1659.

8. Dunn CL, Hannan PJ, Jeffery RW, Sherwood NE, Pronk NP, Boyle R. The comparative and cumulative effects of a dietary restriction and exercise on weight loss. Int J Obes 2006; 30: 112–121.

9. Anderson JW, Grant L, Gotthelf L, Stifler LTP. Weight loss and long-term follow-up of severely obese individuals treated with an intense behavioral program. Int J Obes 2007; 31: 488–493.

10. Donnelly JE, Hill JO, Jacobsen DJ, Potteiger J, Sullivan DK, Johnson SL, et al. Effects of a 16-month randomized controlled exercise trial on body weight and composition in young, overweight men and women – the midwest excercise trial. Arch Intern Med 2003; 163: 1343–1350.

11. Hollis JF, Gullion CM, Stevens VJ, Brantley PJ, Appel LJ, Ard JD, et al. Weight loss during the intensive intervention phase of the Weight-Loss Maintenance trial. Am J Prev Med 2008; 35: 118–126.

12. Wadden TA, West DS, Neiberg RH, Wing RR, Ryan DH, Johnson KC, et al. One-year weight losses in the Look AHEAD study: factors associated with success. Obesity 2009; 17: 713–722.

13. Gleim G. Exercise is not an effective weight loss modality in women. J Am Coll Nutr 1993; 12: 363–367.

14. Aadland E, Robertson L. Physical activity is associated with weight loss and increased cardiorespiratory fitness in severely obese men and women undergoing lifestyle treatment. J Obes 2012; 2012: 810594.

15. Becker W, Lyhne N, Pedersen AN, Aro A, Fogelholm M, Phosdottir I, et al. Nordic Nutrition Recommendations 2004 – integrating nutrition and physical activity. Scand J Nutr 2004; 48: 178–187.

16. Jebb SA, Siervo M, Murgatroyd PR, Evans S, Fruebeck G, Prentice AM. Validity of the leg-to-leg bioimpedance to estimate changes in body fat during weight loss and regain in overweight women: a comparison with multi-compartment models. Int J Obes 2007; 31: 756–762.

17.Strain GW, Wang J, Gagner M, Pomp A, Inabnet WB, Heymsfield SB. Bioimpedance for severe obesity: comparing research methods for total body water and resting energy expenditure. Obesity 2008; 16: 1953–1956.

18. Trost SG, McIver KL, Pate RR. Conducting accelerometer-based activity assessments in field-based research. Med Sci Sports Exerc 2005; 37: S531–S543.

19.Sirard JR, Forsyth A, Oakes JM, Schmitz KH. Accelerometer test-retest reliability by data processing algorithms: results from the twin cities walking study. J Phys Act Health 2011; 8: 668–674.

20. Aadland E, Anderssen SA. Treadmill calibration of the Actigraph GT1M in young-to-middle-aged obese-to-severely obese subjects. J Obes 2012; 2012: 318176.

21. Andersen LF, Solvoll K, Johansson LR, Salminen I, Aro A, Drevon CA. Evaluation of a food frequency questionnaire with weighed records, fatty acids, and alpha-tocopherol in adipose tissue and serum. Am J Epidemiol 1999; 150: 75–87.

22. Twisk JWR. Applied longitudinal data analysis for epidemiology. New York: Cambridge University Press; 2003.

23. Maffiuletti NA, Agosti F, Marinone PG, Silvestri G, Lafortuna CL, Sartorio A. Changes in body composition, physical performance and cardiovascular risk factors after a 3-week integrated body weight reduction program and after 1-y follow-up in severely obese men and women. Eur J Clin Nutr 2005; 59: 685–694.

24.Goodpaster BH, DeLany JP, Otto AD, Kuller L, Vockley J, South-Paul JE, Thomas SB, et al. Effects of diet and physical activity interventions on weight loss and cardiometabolic risk factors in severely obese adults: a randomized trial. JAMA 2010; 304: 1795–1802.

25. Warburton DER, Nicol CW, Bredin SSD. Health benefits of physical activity: the evidence. CMAJ 2006; 174: 801–809.

26. Boutcher SH, Dunn SL. Factors that may impede the weight loss response to exercise-based interventions. Obes Rev 2009; 10: 671–680.

27. Donnelly JE, Honas JJ, Smith BK, Mayo MS, Gibson CA, Sullivan DK, et al. Aerobic exercise alone results in clinically significant weight loss for men and women: midwest exercise trial 2. Obesity 2013; 21: E219–E228.

28. McMurray RG, Hackney AC. Interactions of metabolic hormones, adipose tissue and exercise. Sports Med 2005; 35: 393–412.

29. Zouhal H, Jacob C, Delamarche P, Gratas-Delamarche A. Catecholamines and the effects of exercise, training and gender. Sports Med 2008; 38: 401–423.

30.Kuk JL, Ross R. Influence of sex on total and regional fat loss in overweight and obese men and women. Int J Obes 2009; 33: 629–634.

31. Stubbs RJ, Sepp A, Hughes DA, Johnstone AM, Horgan GW, King N, et al. The effect of graded levels of exercise on energy intake and balance in free-living men, consuming their normal diet. Eur J Clin Nutr 2002; 56: 129–140.

32. Stubbs RJ, Sepp A, Hughes DA, Johnstone AM, King N, Horgan GW, et al. The effect of graded levels of exercise on energy intake and balance in free-living women. Int J Obes Relat Metab Disord 2002; 26: 866–869.

33. Stiegler P, Cunliffe A. The role of diet and exercise for the maintenance of fat-free mass and resting metabolic rate during weight loss. Sports Med 2006; 36: 239–262.

34. Middleton KMR, Patidar SM, Perri MG. The impact of extended care on the long-term maintenance of weight loss: a systematic review and meta-analysis. Obes Rev 2012; 13: 509–517.

35. Shadish W, Cook T, Campbell D. Experimental and quasi-experimental designs for general causal inference. Belmont: Wadsworth Cengage Learning; 2002.

36. Hill RJ, Davies PS. The validity of self-reported energy intake as determined using the doubly labelled water technique. Br J Nutr 2001; 85: 415–430.

37.Schoeller DA, Thomas D, Archer E, Heymsfield SB, Blair SN, Goran MI, et al. Self-report-based estimates of energy intake offer an inadequate basis for scientific conclusions. Am J Clin Nutr 2013; 97: 1413–1415.

38. Black AE, Cole TJ. Biased over- or under-reporting is characteristic of individuals whether over time or by different assessment methods. J Am Diet Assoc 2001; 101: 70–80.

39. Metcalf BS, Curnow JSH, Evans C, Voss LD, Wilkin TJ. Technical reliability of the CSA activity monitor: the EarlyBird Study. Med Sci Sports Exerc 2002; 34: 1533–1537.

40. Lovejoy JC, Sainsbury A, Stock Conference 2008 Working Group. Sex differences in obesity and the regulation of energy homeostasis. Obes Rev 2009; 10: 154–167.

Comments

Do you want to comment on this paper? The comments will show up here and if appropriate the comments will also separately be forwarded to the authors. You need to login/create an account to comment on articles. Click here to login/create an account.
Advertisement