1479-5868-2-6 1479-5868 Research <p>Estimated physical activity in Bavaria, Germany, and its implications for obesity risk: Results from the BVS-II Study</p> Schaller Nina ninaschaller@web.de Seiler Henrike hseiler@wzw.tum.de Himmerich Stephanie himmerich@wlh.wi.tum.de Karg Georg karg@wlh.wi.tum.de Gedrich Kurt gedrich@wlh.wi.tum.de Wolfram Günther gwolfram@wzw.tum.de Linseisen Jakob j.linseisen@dkfz-heidelberg.de

Human Nutrition and Cancer Prevention, Technical University of Munich, Alte Akademie 16, 85350 Freising-Weihenstephan, Germany

Consumer Economics, Technical University of Munich, Weihenstephaner Steig 17, 85350 Freising-Weihenstephan, Germany

Department of Food and Nutrition, Technical University of Munich, Alte Akademie 16, 85350 Freising-Weihenstephan, Germany

Division of Clinical Epidemiology, German Cancer Research Centre, Im Neuenheimer Feld 280, 69120 Heidelberg, Germany

 International Journal of Behavioral Nutrition and Physical Activity 1479-5868 2005 2 1 6 http://www.ijbnpa.org/content/2/1/6 15943862 10.1186/1479-5868-2-6 26 11 2004 08 6 2005 08 6 2005 2005 Schaller et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

Background

Adequate physical activity (PA) is considered as a key factor in the fight against the obesity epidemic. Therefore, detailed description of the actual PA and its components in the population is necessary. Additionally, this study aims to investigate the association between PA and obesity risk in a representative population sample in Bavaria, Germany.

Methods

Data from 893 participants (age 13–80 years) of the Bavarian Food Consumption Survey II (BVS II) were used. In each participant, three computer-based 24-hour recalls were conducted by telephone assessing type and duration of PA in the domains occupation, sports, other strenuous leisure time activities (of mostly moderate intensity) as well as TV/PC use in leisure time and duration of sleeping. After assigning metabolic equivalents (METs) to each activity, estimates of energy expenditure (MET*h) and total daily PA level (PALest.) were calculated. In a subgroup of adults (n = 568) with anthropometric measurements logistic regression models were used to quantify the impact of PA on obesity risk.

Results

Estimated average PA in women and men was 38.5 ± 5.0 and 40.6 ± 9.3 MET*h/d, respectively, corresponding to PALest. values of 1.66 ± 0.22 and 1.75 ± 0.40. Obese subjects showed lower energy expenditure in the categories sports, occupation, and sleeping, while the time spent with TV/PC during leisure time was highest. This is confirmed in logistic regression analyses revealing a statistically significant association between obesity and TV/PC use during leisure time, while sports activity was inversely related to obesity risk. Overall, less than 1/3 of the study participants reached the recommended PAL of ≥ 1.75. Subjects within the recommended range of PA had an about 60 % (odds ratio = 0.43; 95% CI: 0.21–0.85) reduced risk of obesity as compared to inactive subjects with a PALest. <1.5.

Conclusion

Based on the results of short-term PA patterns, a major part of the Bavarian adult population does not reach the recommendations (PAL>1.75; moderate PA of > 30 min/d). Despite the limitations of the study design, the existing associations between sports activity, TV/PC use and obesity risk in this population give further support to the recommendation of increasing sports activity and reducing sedentary behaviour in order to prevent rising rates of obesity.

Background

Globally, there are more than 1 billion overweight adults, at least 300 million of them obese. These alarming facts published by the World Health Organisation (WHO) 1 demonstrate that obesity has reached epidemic dimension in developed as well as in developing countries. Consequences on health range from several non-fatal but debilitating disorders that reduce quality of life to increased risk of premature death because of serious chronic diseases. Besides genetic factors and food consumption patterns exceeding the individual energy need, a sedentary lifestyle with lack of physical activity (PA) is one of the key causes 2. The relationship between obesity, PA and chronic diseases is close and several epidemiological studies could show that regular PA can prevent from obesity and related chronic diseases, such as type-2 diabetes, cardiovascular disease, hypertension, stroke, cancers of different sites, osteoporosis, and contribute to maintain mental health 13. Thus, PA promotes health and well-being and has also enormous economic benefits considering the health care costs that could be attributed to obesity. However, the question of the adequate dose of exercise is still a matter of debate 456.

In order to provide a solid basis for obesity prevention strategies detailed knowledge of PA patterns in the target population is necessary. Therefore, we assessed short-term PA and sedentary behaviour of the Bavarian population by means of three unannounced 24-h recalls. Different activity domains contributing to total daily energy expenditure are described and their impact on obesity risk is quantified. Additionally, PA estimates in the Bavarian population are compared with current recommendations to prevent obesity and promote well-being and health.

Methods

Study Design

The Bavarian Nutrition Survey II (BVS II) is designed as a representative study of the Bavarian population to investigate dietary habits and PA. From September 2002 until June 2003, 1050 subjects aged 13–80 years were recruited by a three-stage random route sampling procedure from the German-speaking Bavarian population. This recruitment procedure included the selection of 42 communities as so-called sampling points (stratified by county and community characteristics), a random walk (every third household) with a given start address, and a random selection of one household member who meets the selection criteria. At baseline, subjects' characteristics, lifestyle, socio-economic and health status were assessed by means of a computerized face-to-face interview. Within the following two weeks, participants were contacted by telephone on two workdays and one weekend day for recalling their dietary intake as well as PA on the day before. Within six weeks after recruitment, all adult study subjects (=18 years) were invited to their nearest health office for blood sampling and standardized anthropometric measurements.

Participation rate in the whole study was 71 % (n = 1050). All adults that completed at least one 24-h dietary recall (n = 879) were invited to the health offices; from 65 % (n = 568) of those approached blood samples and anthropometric measurements could be obtained. For the present evaluation, 893 subjects who completed at least two 24-h activity-recalls were included. Within this group standardized anthropometric measurements were available from 552 subjects (61.8 %). All participants gave their written informed consent. The study was approved by the local ethical committee.

Assessment of Physical Activity

According to a method described and validated by Matthews et al. 14, information on the short-term PA of each subject was collected by means of three unannounced computer-assisted telephone interviews. Trained interviewers asked the study participants to recall the exact type and time spent in activities of the following 5 categories during the last 24 hours: occupation, sports, other strenuous leisure time activities (LTPAstrenuous), TV or PC use in leisure time and sleeping. In the categories sports and LTPAstrenuous, the interviewers used a list of common activities laid on the screen in order to give examples to the participants and to fasten the interview process. Different types of walking (including walking for pleasure) were attributed to the category 'sports' since this type of activity is very important in older age; the category LTPAstrenuous included mainly leisure time PAs of moderate and vigorous intensity, such as different types of gardening, homemaking and household activities, or child caring. Although the wording of the question ('strenuous') may imply vigorous activities only, we actually assessed mainly activities of moderate intensity by means of this question (see results).

Based on the results of their validation study, Matthews et al. 14 concluded that a series of three unannounced 24-h PA recalls provides an assessment of PA comparable to other short-term PA assessments that utilize activity monitors (Actillume monitoring) or the Baecke questionnaire. Deattenuated Pearson correlation coefficients between results from the 24-h recalls and the Baecke questionnaire ranged from 0.34–0.68 (p < 0.01). A correlation coefficient of 0.64 (p < 0.01) was reported for the association between 24-h recall results (total MET*h/d) and the Actillume measures (counts*min-1 *d-1). They assessed four intensities of activity (light, moderate, vigorous, and very vigorous) in each of three activity domains (household, occupational, leisure-time) as well as sleeping time, and assigned 1.5 MET for light, 4.0 MET for moderate, and 6.0 MET for vigorous activities 14. In our study, we more precisely assessed the time and type of activity spent in different PA categories and assigned individual MET values; however, except for TV/PC use, we did not actively assess the time spent with light activities during leisure time.

As described in the compendium of physical activities by Ainsworth et al. 78, multiples of the metabolic equivalent (METs) were used to estimate the relative intensity of each reported activity with one MET equal to the standard for resting energy expenditure (roughly 3.5 ml of oxygen consumed per kilogram of body weight/min) for the average adult. According to the assigned MET-values, all self-reported activities were classified as light (< 3 METs), moderate (3–6 METs) or vigorous (>6 METs) 48.

The MET-values of occupational activities were determined by a combination of self-reported work-intensity (ranging from mainly sitting to laborious physical workload or actually not working) and respective job-title. When a description of activities was missing or the provided information unclear standardized mean MET-values were assigned. In particular, if job activities of students and retired persons were reported that could not be classified, a MET-value of 1.85 representing light work was assumed to be applicable. Type and intensity of the activity of homemakers was also difficult to evaluate; only for this group all reported strenuous activities belonging to the area of household activities were considered as being included in occupational household work and, therefore, not attributed to LTPAstrenuous. To acknowledge homemakers' activities as full occupation, we filled up the reported working time to at least 8 hours of work per weekday for all homemakers under 65 years. An intensity level of 2.5 METs representing "multiple household tasks all at once, light effort" 8 was assigned.

Energy expenditure estimates (MET*h) independent from body weight were calculated by multiplying the reported duration of any activity (h) by respective intensity (MET) 78. By summing up all activities, participants' daily MET*h were obtained for the different activity domains, e.g. sports-MET*h per day. In order to estimate a total daily PA score, it was necessary to introduce a new activity domain, called non-reported PA during leisure time (LTPAnon-reported), according to a method described by Norman et al. 910. The difference between 24 hours per day and the total duration of self-reported activity/inactivity was considered as LTPAnon-reported. These unknown activities were multiplied by an estimated MET-value of 1.75, which is between the suggested values of 1.5 MET 14 and 2.0 MET 910. The intensity factor corresponds to the mean of sitting (1.5 MET) and light home and self-care activities (2.0 MET) 78. Since our study participants mentioned also several light activities under the category LTPAstrenuous – which were multiplied with the most exact MET value given by Aintsworth et al. – we tried not to overestimate the remaining non-reported time.

The single recalls were weighted for weekday or weekend day to calculate a subject's total daily short-term PA and its components. We also estimated the participants' short-term PA level (PALest.) by dividing the individual total daily PA score (MET*h/d = kcal/(kg body weight*d) =~ 1 kcal/(kg b.w.*min) 78) by the minimum score of 23.2 MET*h/d (assumption of 8 hours of sleep × 0.9 MET and 16 h being awake, but resting × 1.0 MET) 11. Since 23.2 MET*h should reflect resting metabolic rate (RMR) expressed in units of MET*h, the resulting ratio gives the multiple of RMR 11, similar to the PAL value. However, it has to be emphasized that the calculated PALest. values are of limited precision as compared to the PAL values mainly derived by means of the doubly labelled water method.

Case definition

To assess the prevalence of overweight and obesity, the subjects' body mass index (BMI) was calculated as measured weight divided by the square of measured height (kg/m2). Self-reported figures were used for subjects who did not undergo anthropometric measurements. Following the WHO-guidelines 12 participants were classified into six categories as being underweight (<18.5 kg/m2), normalweight (18.5-<25 kg/m2), overweight (25-<30 kg/m2), obese grade I (30-<35 kg/m2), obese grade II (35-<40 kg/m2) and obese grade III (≥ 40 kg/m2). All obese subjects (n = 144) with BMI ≥ 30 kg/m2 were considered as cases and all other study participants served as controls in the logistic regression analyses.

Statistical Analysis

The given descriptive results were weighted to correct for the deviation of the study group from the distribution of gender, age, and living area in the underlying Bavarian population. Since the PA data were not normally distributed, median and interquartile range are presented. Comparisons between gender and BMI groups were made by means of the Mann-Whitney U test. In order to examine the association between PA and obesity risk, logistic regression models were used. Risk calculations were conducted only for the subgroup with standardized measurement of weight and height. Additionally, subjects with an energy intake below 80% of the estimated basal metabolic rate (BMR, calculated by WHO-equations 13) were excluded from risk estimations because of an increased likelihood of misreporting of PA. Thus, risk evaluation was conducted in a subgroup of 507 subjects. The activity estimates (MET*h/d) over each activity-domain as well as the total daily activity (MET*h/d and PALest, respectively) were divided into four groups according to the distribution in the entire study population or by predefined cut points. Odds ratios (OR) and corresponding 95% confidence intervals (CI) are given for models adjusted for sex, age (< 18 y, 18-<30 y, 30-<40 y, 40-<50 y, 50-<65 y, ≥ 65 y), energy intake (kcal/100/d), smoking (never, former, current) and socio-economic status (low, low-medium, medium, medium-high, high). Categorization of socio-economic status is based on the value of three characteristics on a point-scale including household net income, educational level of the one who is being interviewed and career position of the principal earner. Tests on trend were calculated using the quartile-based PA scores as a continuous variable as well as using the continuous variables (in MET*h/d). All statistical analyses were performed by means of the SPSS 11.0 software package (SPSS Inc., Chicago, USA).

Results

Baseline characteristics and prevalence of obesity

Baseline characteristics of the study participants are summarized in Table 1. Significant gender differences existed for BMI groups, socioeconomic status, employment level, smoking habits and marital status; also anthropometric measures as well as basal metabolic rate (BMR) and energy intake differed by gender. The proportion of obese subjects in the whole sample (n = 893) was estimated to 17.1% in women and 16.1% in men. Excluding subjects with self-reported weight and height, the prevalence of obesity was even higher with 19.6% in women and 20.4% in men (overall 20.0%).

<p>Table 1</p>

Baseline characteristics of the study participants1.

Total (n = 893)

Women (n = 528)

Men (n = 365)

p-value 2

n

%

n

%

n

%

Age (years)

0.713

<18

48

7.4

20

6.4

28

8.6

18-<30

99

13.5

65

14.0

34

12.9

30-<40

196

21.0

125

20.2

71

21.8

40-<50

182

19.0

119

19.7

63

18.2

50-<65

228

23.8

129

23.5

99

24.2

≥65

140

15.2

70

16.2

70

14.1

Body mass index (kg/m2)

<0.001

underweight (<18.5)

35

4.2

25

5.9

10

2.4

normal (18.5-<25)

402

44.9

265

49.3

137

40.0

overweight (25-<30)

312

34.1

154

27.4

158

41.5

obese (≥30)

144

16.6

84

17.1

60

16.1

grade I (30-<35)

99

11.2

54

10.5

45

12.0

grade II (35-<40)

31

3.6

17

3.3

14

3.8

grade III (≥40)

14

1.9

13

3.5

1

0.2

Socioeconomic status

0.001

low

133

13.6

81

14.3

52

12.9

low-medium

230

25.5

129

25.7

101

25.4

medium

262

29.3

163

29.9

99

28.7

medium-high

178

21.2

118

24.0

60

18.2

high

90

10.3

37

6.2

53

14.8

Employment

<0.001

employed

429

48.1

241

41.2

188

55.6

homemaker

152

13.9

151

26.5

1

0.2

student/articled

78

12.3

35

10.3

43

14.6

unemployed/other

36

4.1

14

2.2

22

6.2

retired

198

21.5

87

19.7

111

23.4

Smoking status

<0.001

never

473

52.3

320

61.2

153

42.6

former

183

21.0

83

16.0

100

26.3

current

236

26.7

124

22.6

112

31.1

missing Data

1

0.1

1

0.2

0

0.0

Marital status

<0.001

single

176

21.6

82

17.1

94

26.6

married/cohabiting

578

67.0

337

65.8

241

68.4

divorced/widowed

138

11.3

108

17.1

30

5.0

missing data

1

0.0

1

0.0

0

0.0

mean ± SD

Height (cm)

169.6 ± 9.1

164.0 ± 6.7

175.5 ± 7.3

<0.001

Weight (kg)

74.3 ± 15.6

68.1 ± 13.9

81.0 ± 14.6

<0.001

BMI (kg/m2)

25.8 ± 5.2

25.5 ± 5.8

26.2 ± 4.3

0.038

BMR (kcal)

1601 ± 253

1419 ± 132

1801 ± 197

<0.001

Energy intake (kcal)3

2001 ± 667

1704 ± 529

2326 ± 652

<0.001

1weighted for deviation from the underlying Bavarian population (sex, age, region)

2Chi-Square test or independent-samples t-test for gender differences

33 × 24-hour dietary recall

Estimated Physical Activity

Estimates of PA by activity domain (MET*h/d) and intensity are given in Table 2, including also the corresponding duration of activities (h/d). Men as compared to women showed significantly higher values in total scores of sports activity, TV/PC use and total daily activity, while women reported a significantly longer sleeping time per day. This is also reflected in results by intensity sub-groups with men spending more time in PA with moderate or vigorous intensity. The most important intensity subgroup was occupational PA of light intensity showing the highest mean energy expenditure for both men and women. Non-reported time of PA in the 24-hour recalls was higher in women than in men. Total daily PA was estimated to 37.35 (5.58) MET*h/d (median, interquartile range) in women and 37.92 (8.80) in men, corresponding to PALest. values of 1.61 (0.24) and 1.63 (0.38), respectively.

<p>Table 2</p>

Estimated physical activity (h/d and MET*h/d) by sex, type, and intensity of activity1.

Type and intensity* of activity/inactivity

Total (n = 893)

Women (n = 528)

Men (n = 365)

p-value2 (MET*h/d)

h/d

MET*h/d

h/d

MET*h/d

h/d

MET*h/d

Median (Interquatile range)

Occupation

total

2.86 (5.71)

5.36 (14.00)

2.86 (5.71)

5.14 (14.29)

2.95 (6.07)

6.83 (12.60)

0.799

light

82.6% #

68.1%

89.9%

82.4%

75.8%

55.2%

moderate

13.2%

21.5%

8.8%

13.6%

17.7%

28.7%

vigorous

4.0%

10.4%

1.7%

4.0%

6.5%

16.2%

Sports

total

0.12 (0.63)

0.48 (3.34)

0.08 (0.53)

0.38 (2.83)

0.14 (0.73)

0.59 (4.50)

0.028

moderate

57.5%

43.1%

66.7%

52.9%

51.1%

36.4%

vigorous

40.0%

56.9%

33.3%

46.6%

48.9%

63.6%

LTPAstrenuous3

total

0.00 (0.54)

0.00 (1.71)

0.00 (0.57)

0.00 (2.10)

0.00 (0.36)

0.00 (1.55)

0.893

light

5.0%

3.2%

5.4%

4.4%

4.8%

2.2%

moderate

92.5%

92.4%

91.95

90.4%

95.2%

94.0%

vigorous

2.5%

4.5%

2.7%

5.2%

2.4%

3.8%

TV/PCleisure time4

1.64 (1.82)

1.64 (1.82)

1.38 (1.52)

1.38 (1.52)

2.00 (1.96)

2.00 (1.96)

<0.001

Sleeping5

7.43 (1.39)

6.69 (1.25)

7.56 (1.37)

6.80 (1.23)

7.31 (1.36)

6.58 (1.22)

0.016

LTPAnon-reported6

10.93 (4.04)

19.12 (7.08)

11.26 (4.17)

19.71 (7.30)

10.61 (4.08)

18.56 (7.14)

<0.001

Total daily activity score

MET*h

37.52 (7.18)

37.35 (5.58)

37.92 (8.80)

<0.001

PALest.

1.62 (0.31)

1.61 (0.24)

1.63 (0.38)

<0.001

1 Weighted for deviation from the underlying Bavarian population (sex, age, region)

2 Mann-Whitney U-test

3 LTPAstrenuous = strenuous leisure time physical activity

4 TV/PCleisure time (1.0 MET)

5 Sleeping (0.9 MET)

6 LTPAnon-reported = non-reported leisure time physical activity (1.75 METs)

* Light (<3 METs), moderate (3–6 METs), vigorous (>6 METs)

# Percentage of mean h/d and mean MET*h/d, respectively, of the corresponding activity domain

Table 3 shows the results for the estimated PA by type and intensity level in different BMI categories. Obese subjects reported less participation in occupational (women) and sports (men) activities but performed more LTPAstrenuous than non-obese women and men. On the contrary, the time spent with TV/PC use during leisure time was highest in overweight and obese subjects. Sleeping time was shortest among obese women while underweight subjects slept most. Total daily activity scores were lowest in obese and underweight subjects, thus, the difference between obese and non-obese subjects did not reach statistical significance.

<p>Table 3</p>

Estimated physical activity (MET*h/d) by weight class (BMI), type and, intensity of activity.

Underweight (<18.5 kg/m2)

Normalweight (18.5-<25)

Overweight (25-<30)

Obese (≥30)

p-value

Type and intensity* of activity/inactivity

Women (n=528)

difference2 obese vs. others)

n = 25

n = 265

n = 154

n = 84

Median (Interquatile range)

Occupation

total

2.93 (9.76)

5.79 (14.29)

5.35 (14.29)

2.41 (11.68)

0.033

light

55.8% #

83.8%

81.6%

89.7%

moderate

18.8%

12.2%

16.8%

10.3%

vigorous

25.3%

3.9%

1.8%

0%

Sports

total

1.62 (3.24)

0.63 (3.12)

0.00 (2.30)

0.00 (1.73)

0.105

moderate

39.9%

47.0%

54.8%

80.4%

vigorous

60.1%

53.0%

43.9%

19.6%

LTPAstrenuous3

total

0.00 (0.00)

0.00 (1.61)

0.00 (3.21)

0.00 (2.41)

0.026

light

0.0%

5.3%

2.2%

5.4%

moderate

100.0%

87.6%

89.9%

94.6%

vigorous

0 %

7.1%

7.3%

0%

TV/PCleisure time4

1.21 (1.11)

1.11 (1.33)

1.69 (2.04)

1.78 (1.61)

0.004

Sleeping5

7.90 (2.07)

6.80 (1.24)

6.75 (1.03)

6.75 (1.25)

0.332

LTPAnon-reported6

20.77 (8.20)

19.26 (7.72)

19.49 (7.45)

19.89 (6.66)

0.325

Total daily activity score

MET*h/d

36.56 (7.08)

37.78 (6.00)

37.20 (5.97)

36.70 (4.95)

0.083

PALest.

1.58 (0.31)

1.63 (0.26)

1.60 (0.26)

1.58 (0.21)

0.083

Men (n = 365)

n = 10

n = 137

n = 158

n = 60

Median (Interquatile range)

Occupation

total

3.47 (6.51)

7.48 (13.68)

5.09 (12.45)

3.01 (12.28)

0.098

light

100.0%

60.6%

46.0%

61.4%

moderate

0%

18.2%

46.5%

10.7%

vigorous

0%

21.3%

7.5%

28.1%

Sports

total

7.52 (15.67)

1.29 (7.86)

0.39 (3.21)

0.00 (2.05)

0.006

moderate

13.2%

33.0%

48.5%

28.6%

vigorous

86.9%

67.0%

51.0%

71.4%

LTPAstrenuous3

total

0.00 (0.00)

0.00 (0.54)

0.00 (2.86)

0.29 (4.39)

0.007

light

0%

1.9%

3.1%

0.7%

moderate

100.0%

87.5%

94.6%

98.7%

vigorous

0%

11.5%

2.2%

0.7%

TV/PCleisure time4

1.46 (2.53)

1.77 (1.62)

2.07 (1.97)

2.46 (1.92)

0.001

Sleeping5

8.21 (1.63)

6.62 (1.13)

6.56 (1.28)

6.48 (0.90)

0.040

LTPAnon-reported6

15.80 (6.43)

18.10 (7.45)

18.66 (6.95)

19.50 (7.84)

0.413

Total daily activity score

MET*h/d

37.54 (11.56)

38.94 (8.81)

37.19 (8.42)

37.42 (8.92)

0.087

PALest.

1.62 (0.50)

1.68 (0.38)

1.60 (0.36)

1.61 (0.38)

0.087

1 Weighted for deviation from the underlying Bavarian population (sex, age, region)

2 Mann-Whitney U-test

3 LTPAstrenuous = strenuous leisure time physical activity

4 TV/PCleisure time (1.0 MET)

5 Sleeping (0.9 MET)

6 LTPAnon-reported = non-reported leisure time physical activity (1.75 METs),

* Light (<3 METs), moderate (3–6 METs), vigorous (>6 METs)

# Percentage of mean MET*h/d of the corresponding activity domain