Association Between Health Behaviors and Cardiorespiratory Fitness

Journal of Adolescent Health 53 (2013) 272e279

www.jahonline.org Original article

Association Between Health Behaviors and Cardiorespiratory Fitness in Adolescents: Results From the Cross-Sectional MoMo-Study Eliane Peterhans, M.Sc. a, *, Annette Worth, Dr. paed. b, and Alexander Woll, Dr. phil. c a b c

Department of Sport Sciences, University of Konstanz, Konstanz, Germany University of Education Karlsruhe, Karlsruhe, Germany Department of Sport and Sport Science, Karlsruhe Institute of Technology, Karlsruhe, Germany

Article history: Received May 7, 2012; Accepted February 13, 2013 Keywords: Lifestyle; Fitness; Adolescents; “Motorik-Modul” (MoMo); Physical activity

A B S T R A C T

Purpose: The aim of this study was to analyze the association between adolescent and familial health behavioral factors and cardiorespiratory fitness in German adolescent boys and girls. Methods: This study is based on a large nationwide cross-sectional study and its substudy on physical activity and fitness of children and adolescents (“Motorik-Modul”). For 1,328 adolescents between 11 and 17 years of age, data on cardiorespiratory fitness (Physical working capacity 170, PWC170) and familial and adolescent health behavioral factors were collected. Health behavior was assessed using psychometric questionnaires (socioeconomic status, pubertal stage, daily physical activity, sports-club time, parental physical activity habits, etc.). A hierarchical multiple regression model was used to quantify the association between relative PWC170 values and health behavior. Results: The relationship between adolescents’ health behavioral factors and cardiorespiratory fitness was stronger than the relationship between age, social status, familial health behavior and cardiorespiratory fitness. Familial health behavioral factors explained 4.1% and 2.1% of variance in cardiorespiratory fitness in girls and boys, respectively. Adolescents’ health behavioral factors explained 15.2% of variance in girls and 25.7% of variance in boys. For both girls (b ¼ .273) and boys (b ¼ .400), being normal weight had the greatest effect on relative PWC170 values. Conclusions: The difference in explained variance in cardiorespiratory fitness by familial and adolescents’ health behavioral factors between girls and boys indicates that different predictors for cardiorespiratory fitness are important for girls and boys. Hence, sex specific research and interventions aimed at improving familial and adolescent health behavior may be important. Ó 2013 Society for Adolescent Health and Medicine. All rights reserved.

A range of lifestyle aspects influence health and longevity in adults [1], and adults with several positive health behaviors have a lower relative risk for all-cause mortality than those with no positive health behaviors [1,2]. In addition, low cardiorespiratory fitness is a risk factor for cardiovascular diseases (CVD) in adults and a predictor for CVD mortality and all-cause mortality in both * Address correspondence to: Eliane Peterhans, M.Sc., Department of Sport Sciences, University of Konstanz, Universitätsstrasse 10, 78457 Konstanz, Germany. E-mail address: [email protected] (E. Peterhans).

IMPLICATIONS AND CONTRIBUTION

This nationwide study including 1,328 adolescents between 11 and 17 years of age showed that adolescent and familial health behavioral factors partially predict cardiorespiratory fitness. The investigated variables are not equally important for both sexes. Hence, sex specific research and interventions aimed at improving familial and adolescent health behavior may be important.

women and men [3]. Similarly, in children and adolescents cardiorespiratory fitness is a predictor for CVD risk factors [4], and higher levels of cardiorespiratory fitness are inversely associated with a more favorable metabolic risk profile [5e7]. Some mostly cross-sectional studies on children and adolescents reported associations between single health behaviors, defined as activities performed by individuals that will influence their health [8], bodyweight [9], and cardiorespiratory fitness. Obesity [10], sedentary time such as screen time [11], smoking cigarettes [12], sports club participation [13], and active travel to school [14] are inversely and directly associated with cardiorespiratory

1054-139X/$ e see front matter Ó 2013 Society for Adolescent Health and Medicine. All rights reserved. http://dx.doi.org/10.1016/j.jadohealth.2013.02.011

E. Peterhans et al. / Journal of Adolescent Health 53 (2013) 272e279

fitness. The association between physical activity and cardiorespiratory fitness is controversial [15,16], and little is known about the association between leisure time physical activity and cardiorespiratory fitness. As shown by Bouchard and Shephard [17], fitness is also determined by a person’s social environment, which includes familial health behavioral factors. For instance, adolescents’ fitness level is associated with mothers’ and fathers’ weight [18]. According to social cognitive theory, people learn behaviors by observing the behavior of others [19], and hence the role of familial health behaviors must also be investigated. Further, cardiorespiratory fitness in childhood and adolescence is inversely associated with overweight [10] and the risk of developing cardiovascular disease [7,15] and metabolic syndrome in adulthood [4]. Hence, understanding the impact of multifactorial health behavior on cardiorespiratory fitness is crucial for designing specific programs aimed at preventing cardiovascular disease. However, to date only a few studies have focused on the association between multifactorial health behavior and cardiorespiratory fitness in children and adolescents. Therefore, the aim of this study was to analyze the association between adolescent and familial health behavioral factors and cardiorespiratory fitness in German adolescents. Methods Study design and participants This study was a substudy of the project “Motorik-Modul” (MoMo), which was funded by the German Ministry for Family Affairs, Senior Citizens, Women and Youth, and by the German Federal Ministry of Education and Research. The MoMo Baseline Study is a subsample of the umbrella study German Health Interview and Examination Survey for Children and Adolescents (KiGGS) [20,21]. A total of 17,641 children and adolescents aged 0 to 17 years with primary residence in Germany participated in the representative KiGGS-Study. The study population was selected using a stratified multistage probability sample with three evaluation levels. Level three was the selection of MoMo participants from the KiGGS sample. The studies were approved by the Charité/Universitätsmedizin Berlin Ethics Committee and were been performed according to the Declaration of Helsinki. For the MoMo Basic Study, data on physical activity and physical fitness of 4,529 children and adolescents aged 4 to 17 years from 167 cities across all states of Germany were collected from 2003 to 2006. For the present subsample study, only participants aged between 11 and 17 years with complete data sets were selected (n ¼ 1,328).

273

Participants’ pubertal status was categorized into three different categories (1dprepubertal (Tanner 1), 2dpubertal (Tanner 2e3), 3dpostpubertal (Tanner 4e5)) and dummy-coded with postpubertal stage as reference group. Cardiorespiratory fitness. Cardiorespiratory fitness was assessed using the Physical Working Capacity 170 (PWC170) cycle ergometry test (attained watts at 170 beats/min) on an Ergosana, ERG 911S (Ergosana, Blitz, Germany) bicycle. Initial workload was calculated as .5 watts/kg body mass. The workload was increased incrementally by .5 watts/kg body mass every 2 minutes. Subjects continued this progressive protocol until their heart rate exceeded 190 heartbeats/min for at least 15 seconds, the pedaling rate was less than 50 rpm for at least 20 seconds, or the subject decided to stop because of exhaustion. Heart rate was measured using a chest strap T31 (Polar Electro Oy, Kempele, Finland) immediately before each workload increase. The heart rate signal was transmitted to the bicycle ergometer. Power [watts] at a heart rate of 170 beats per minute (PWC170) were obtained by inter- or extrapolating the measured data in Excel by the same investigator. Data was excluded if the data could not be inter- or extrapolated because of invalid data or if not at least three valid values were achieved. Relative PWC170 was calculated as PWC170 divided by body weight. Socioeconomic status. Data on socioeconomic status was collected in the KiGGS study [20] with a parental questionnaire. Socioeconomic status was calculated separately for both parents. The included items were educational and professional status, as well as the total income of the family household [25]. The higher of the parental scores was assumed for the participant. Participants with single parents were assigned the socioeconomic status of the parent with whom they lived. According to Winkler and Stolzenberg [26], each item (income, educational and professional status) was scored on a scale from 1 to 7 and a sum score was created (range: 3e21) and used to categorize participants: low (3e8), intermediate (9e14), and high (15e21) socioeconomic status [27]. A dichotomous variable low versus intermediate-high was calculated. Health behavior. Information on health behavior was assessed using questionnaires in the umbrella study KiGGS and the subsample study MoMo. Although in KiGGS, a parent questionnaire (completed by one or both parents) and an adolescent questionnaire were used, only an adolescent questionnaire was used in the MoMo study. Data from the questionnaires were dichotomized except for the information about the sports club activity index and the leisure time physical activity index outside of sports clubs. Information about items used and the corresponding questionnaires are presented in Table 1. In addition, information on the coding system for each item is given in Table 1.

Outcome measures Anthropometric measures. Standing height was measured using a stadiometer (Seca, Hamburg, Germany; accuracy .1 cm) and body mass was determined using an electronic scale (Seca, Hamburg, Germany; accuracy .1 kg). Body mass index (BMI) was calculated by dividing body mass by height squared. Prevalence of overweight/obese adolescents was calculated according to the cut-offs proposed by the International Obesity Task Force (IOTF) [22]. This information was dichotomized into normal weight or overweight/obese. In the umbrella study KiGGS, the sexual maturity status was determined by self-assessment using line drawings and a simple explanation of pubic hair [23,24].

Familial health behavior and anthropometric data. Parental BMI was calculated using information on parental weight and height in the parental questionnaire, and parents were classified as normal weight or overweight based on their BMI according to WHO recommendations [28]. Information on parental BMI was categorized into three different categories (both parents normal weight, one parent normal weight, and both parents (if single parent one parent)) overweight and dummy-coded with “both parents (if single parent one parent) overweight” as reference group. Mother’s, father’s and siblings’ physical activity was assessed by the adolescent questionnaire. Information on siblings being regularly physically active was categorized into three different

274


Table 1 Variables examined Items

Independent variable

KiGGS Questionnaire for parents How tall and heavy are you? (Please give information for the parent/s with whom the child lives) KiGGS Questionnaire for adolescents How long do you usually watch TV/videos on average per day? How long do you use a PC or the Internet on average per day? How long do you usually play on a game console on average per day? Do you currently smoke? MoMo questionnaire Is your dad regularly physically active? Is your mom regularly physically active? Are your siblings regularly physically active?

On how many days during the last week were you physically active for at least 60 min/day? On how many days of a normal week were you physically active for at least 60 min/day? How many times per week do you participate in each sport? How long does a training session last? During which months do you participate in each sport? To how many minutes per week do these activities outside of sports clubs usually amount? During which months do you participate in each sport? How do you usually get to school?

a

Both parents are normal weight (BMI 25a) ¼ 1 One parent is normal weight (BMI 25a) ¼ 1 Parents are overweight (BMI > 25a) ¼ 0 (Reference) Screen time 2 hours a day ¼ 1 Screen time > 2 hours a day ¼ 0 Do not smoke ¼ 1 Smoke ¼ 0 Dad is regularly physically active ¼ 1 Dad is not regularly physically active ¼ 0 Mom is regularly physically active ¼ 1 Mom is not regularly physically active ¼ 0 Siblings are regularly physically active ¼ 1 No siblings ¼ 1 Siblings are not regularly physically active ¼ 0 (Reference) WHO activity guidelines achieved (7 days physically active/week) ¼ 1 WHO activity guidelines not achieved (< 7 days physical active/week) ¼ 0

Sports club activity index

Leisure time physical activity index outside of sports clubs

Usually go to school by bicycle ¼ 1 Usually go to school on foot ¼ 1 Others ¼ 0 (Reference)

According to WHO guidelines (2000).

categories and dummy-coded (Table 1). Reliability was assessed in a separate study on 196 participants between 9 and 17 years old [29]. The Kappa coefficients were .83 (mother), .81 (father), .80 (siblings) for a 1-week inter-test interval (p < .001). Adolescent health behavior. Overall physical activity was assessed using a two-item questionnaire developed by Prochaska et al. [30]. The mean of these two items provided information on whether a subject fulfilled the recommended activity level of 60 minutes of physical activity with moderate intensity on 7 days per week [31]. The sports club activity index was calculated in hours based on three items. Subjects provided detailed information on every sports club activity they participated in, including duration (minutes per training), frequency (times per week), and when (month) this sports activity was performed: Indexsports club activity ¼ duration*frequency*number of months h i. h i =12 months 60 min Similarly, the leisure time physical activity index outside of sports clubs was calculated in hours using two items. Subjects provided detailed information about every physical leisure activity outside of sports clubs they participated in, including duration (minutes per week) and when (month) this physical leisure activity was performed: Indexleisure time physical activity ¼ duration*number of months h i. h i =12 months 60 min

Questions about physical activity showed sufficient reliability (between k ¼ .54 and k ¼ .81, mean k ¼ .66 (SD ¼ .19)) on the item level for a one-week inter-test interval [29]. The validity against Actigraph GT1X (Actigraph LLC, Pensacola, FL) and the Previous Day Physical Activity Recall were for overall activity r ¼ .24 (p < .01) and r ¼ .43 (both: p < .01), for sports club activity r ¼ .35 (p < .01) and r ¼ .55 (p < .01), and for leisure time physical activity r ¼ .10 and r ¼ .32 (both: p < .01), respectively. These reliability and validity results were similar to those of other questionnaires for measuring physical activity in adolescents [29]. Screen time was calculated as the sum score of three items. Subjects were asked how long per day they usually watched TV, played on a game console, or used a personal computer (PC). Respondents were presented with five possible answers (never, around 30 minutes, around 1e2 hours, around 3e4 hours, more than 4 hours). To calculate a sum score, the five answers were coded as 0, .5, 1.5, 3.5 and 5 [32]. According to the recommendations of the American Academy of Pediatrics [33], this variable was coded as 1 in subjects with a sum score 2, and as 0 in subjects with a sum score above 2. Commuting to school was assessed by a question with five possible answers (on foot, by bicycle, by bus or train, by car, or by moped or scooter), categorized into three different categories (Table 1) and dummy-coded (Table 1). This item had a Kappa coefficient of .92 (p < .001) for a one-week inter-test interval. Statistical analysis All statistic tests were conducted using SPSS Version 19.0 (IBM Corporation, Armonk, NY). Descriptive statistics were used to characterize the subjects of the sample. An independent t-test


was used to detect differences in relative PWC170 between boys and girls. A hierarchical multiple regression model was used to quantify the association between relative PWC170 values and health behaviors. Independent variables were included blockwise. The first model included age, pubertal status, and parental socioeconomic status. In the second model, familial health behaviors were added, and in the final model, adolescents’ health behaviors were included. The analysis was conducted separately for boys and girls because of significant sex differences in the dependent variable relative PWC170 values. A first model was calculated and the data were subsequently checked for outliers. If a subject’s standardized residual was outside of 2, the value of the externally studentized residual outside of 3, and the value of global measures of influence (DFFITS) outside of 2((kþ1)/n)½, the subject was identified as an outlier and excluded [34]. Overall, seven cases (two girls, five boys) were excluded. As a next step, the hierarchical multiple regression was repeated. The data from excluded subjects (because of incomplete data or identified as an outlier) did not differ significantly from the included subjects in relative PWC170 values, age, and pubertal status. The level of significance for all statistical tests was set a priori to a ¼ .05. Results Relative PWC170 values, height, weight, pubertal status, sports club activity index, leisure time physical activity index,

275

commuting to school on foot, and screen time differed between girls and boys (Table 2). Almost twice as many girls as boys spent less than 2 hours daily in front of a computer screen. Prediction of cardiorespiratory fitness The results of the hierarchical regression analyses for girls and boys are shown in Tables 3 and 4. In girls, familial and adolescent health behavioral factors explained 19.3% of variance in cardiorespiratory fitness. In contrast, in boys, the explained variance for cardiorespiratory fitness by familial and adolescent health behavioral factors was 27.8%. In the first model, socioeconomic status in girls and age in boys significantly predicted cardiorespiratory fitness. In the second model, having normal weight parents was a significant positive predictor for adolescents’ fitness for both sexes. In the third model, adolescents’ health behaviors explained 15.2% of variance in girls and 25.7% of variance in boys. For both girls (b ¼ .273) and boys (b ¼ .400), being normal weight had the strongest association with relative PWC170. In addition, sports club activity index, leisure time physical activity index, and commuting to school by bicycle were positive predictors for cardiorespiratory fitness for both sexes. Adolescents who walked to school did not have higher cardiorespiratory fitness than those who used passive transportation modes. Although normal

Table 2 Characteristics of adolescents with complete data: Baseline “Motorik-Modul” 2003e2006 Variable

Category

Girls (n ¼ 657)

Boys (n ¼ 671)

Prepubertal (Tanner 1) Pubertal (Tanner 2e3) Postpubertal (Tanner 4e5) Low Intermediate and high

14.3 160.9 54.1 1.8 6.2 15.1 78.7 25.9 74.1

(2.0) (8.5) (12.0) (.4) (41) (99) (517) (170) (487)

14.1 165.3 56.9 2.2 8.3 32.3 59.3 23.8 76.2

Both parents are normal weight (BMI 25b) One parent is normal weight (BMI 25b) Father regularly physically active Father not regularly physically active Mother regularly physically active Mother not regularly physically active Siblings regularly physically active No siblings Siblings not regularly physically active

22.5 50.5 36.5 63.5 42.0 58.0 61.0 10.4 28.6

(148) (332) (240) (417) (276) (381) (401) (68) (188)

22.5 49.6 38.2 61.8 42.6 57.4 59.2 13.9 26.9

(151) (333) (256) (415) (286) (385) (397) (93) (181)

Normal weightc Overweightc Physically active on 7 days/week for 60 min Physically active on 2 hours a day Nonsmoker Smoker

81.1 18.9 7.9 92.1 1.6 1.3 21.3 20.5 58.2 59.8 40.2 86.1 13.9

(533) (124) (52) (605) (2.6) (2.1) (140) (135) (382) (393) (264) (566) (91)

78.5 21.5 8.3 91.7 2.3 2.0 14.8 24.6 60.6 35.0 65.0 84.5 15.5

(527) (144) (56) (615) (2.7)* (2.9)* (99)* (165) (407) (235)* (436)* (567) (104)

a

Age (y) Height (cm)a Weight (kg)a Relative PWC170a Pubertal stage

Socioeconomic status Family Body mass index Familial health behaviors

Adolescent Body mass index Physical activity Index (h)a Commuting to school

Screen time Smoking status

All values are given as percent (n), except where noted. * p .05 for sex comparisons. a Mean (standard deviation). b WHO (2000). c Cole et al. (2000).

(1.9) (13)* (15.6)* (.5)* (56) (217)* (398)* (160) (511)

276


Table 3 Results of the hierarchical regression analyses in girls (n ¼ 657) examined the association between familial and adolescent health behaviors and cardiorespiratory fitness (PWC170) Variables

Model 1 Ba

Constant Agec Prepubertal stage (Tanner 1)d Pubertal stage (Tanner 2e3)d SES intermediate to high Family Both parents normal weight (BMI