Adolescent predictors of objectively measured physical activity and sedentary behaviour at age 42: the Amsterdam Growth and Health Longitudinal Study (AGAHLS)

Background This study investigated the associations of physical characteristics and personality in adolescence with physical activity and sedentary behaviour in adulthood. Findings Physical characteristics (i.e. objectively measured BMI, sum of skin folds, MOPER test battery performance), and personality (i.e. self-reported inadequacy, social inadequacy, rigidity, self-sufficiency/recalcitrance, dominance, achievement motivation, facilitating anxiety, debilitating anxiety, and social desirability) were assessed in 217 adolescent boys (Mean 13.0, SD 0.6) and girls (Mean 12.9, SD 0.6). Twenty-nine years later, at the age of 42, their physical activity and sedentary behaviour were assessed by means of accelerometry. Boys who scored lower on self-sufficiency/recalcitrance and higher on facilitating anxiety spent more time sedentary in adulthood. Girls with a superior standing high jump performance, and a lower score on social desirability spent more time sedentary in adulthood. In contrast with sedentary behaviour, physical activity at age 42 year could not be predicted by physical characteristics or personality in adolescence. Conclusions Sedentary behaviour in adulthood was partly explained by physical characteristics and/or personality in adolescence. Thus, our results suggest that it may be possible to identify people who are at risk of becoming sedentary at a rather young age.


Introduction
Physical activity and sedentary behaviour are generally accepted as being two distinct classes of behaviour, which have been shown to be independently associated with energy expenditure, body weight, and metabolic factors [1][2][3].
Although a substantial body of literature has focused on drivers for adopting an inactive lifestyle the majority of these studies are cross-sectional [4]. The association between age, education, self-efficacy and physical activity has been frequently investigated, whereas other factors such as personality and fitness have been rarely examined [4]. Research on sedentary behaviour is rapidly growing [5,6], however, evidence on its determinants is relatively scarce.
Considering the above, research investigating under reported determinants of physical activity and sedentary behaviour in a prospective design and by means of objective measurement instruments is of great importance [4][5][6]. Therefore, the present study aims to extend existing knowledge by investigating which physical characteristics and personality in adolescence are longitudinally associated with objectively measured physical activity and sedentary behaviour in adulthood.

Participants and procedures
We used data collected at wave 1 (1976/77) and wave 10 (2006) of the Amsterdam Growth and Health Longitudinal Study (AGAHLS). This longitudinal study started in 1976/1977 monitoring growth, health, and lifestyle in more than 600 boys and girls aged 13 years. The study rationale, recruitment procedures and protocol have been reported in detail elsewhere [7]. We included participants with physical characteristics and/or personality data at wave 1 and data on physical activity and/or sedentary time at wave 10, resulting in a sample of 217 participants (33% of the baseline sample; 42% male). Compared to those with complete data, participants without wave 10 data had a significant higher BMI, performed better in the arm pull test and scored higher on social desirability at baseline. The AGAHLS was approved by the medical ethics committee of the VU University Medical Center, Amsterdam, The Netherlands. All subjects gave their written informed consent [7].

Measurements in adolescence 2.2.1 Physical characteristics
Body height and weight were measured using a Harpenden digital readout, wall-mounted or portable stadiometer (Holtain, UK), and a spring balance (Van Vucht, the Netherlands), and BMI (kg/m 2 ) was calculated. The sum of four skin folds (biceps, triceps, subscapular and supraliliac) was used as indicator of body fatness and measured with a Harpenden calliper (Holtain, UK) [8].
Aerobic fitness was assessed by measuring the maximal oxygen uptake (VO 2 max) while running on a treadmill (Quinton 18-45, USA). During the entire run, the expired air was analysed on O 2 and CO 2 by the Ergoanalyzer (Jaeger, the Netherlands), and subsequently expressed in VO 2 max (ml·min·kg -2/3 ) relative to the individuals' body weight [9].
Muscular fitness, i.e. the respondents' strength, speed, flexibility and endurance capacity was measured by means of the MOPER test battery including 8 different tests. The MOPER components are described in table 1 [10]. Validity and reliability of the MOPER tests have been shown to be acceptable in children [11,12].

Personality
Personality traits were assessed using the youth versions of the Dutch Personality Inventory (DPI) [13], and the Achievement Motivation Test (AMT) [14]. The DPI assessed the participant's inadequacy, social inadequacy, rigidity, self-sufficiency/recalcitrance, and dominance. The AMT assessed the participants' achievement motivation, facilitating anxiety, debilitating anxiety, and social desirability. Psychometric properties of the DPI and AMT are presented in table 2.

Objectively measured physical activity levels and sedentary time in adulthood
Physical activity was objectively measured using Acti-Graph accelerometers (Model GT1M, ActiGraph, LLC, Fort Walton Beach, FL). At age 42, 345 participants were instructed to wear an accelerometer attached to a provided waist belt, for eight consecutive days during waking hours but not during water activities. The accelerometers were set to record acceleration and movement frequency at 60-second epochs. Data were considered eligible for analyses if the respondent had worn the accelerometer for at least one day for ≥ 500 minutes per day. From the accelerometer data we computed two scores: physical activity (counts/min), and time spent sedentary (min/day) [15].
In total, 104 participants (30%) did not provide Acti-Graph data. Subjects with and without ActiGraph data were reasonably equal in terms of self-reported physical activity and sedentary behaviour. Of the remaining 241 participants, 12 (5%) wore the accelerometer for < 500 minutes per day and were thus excluded from analyses. Those participants recorded significantly less counts per minute, less sedentary time and less wearing days.

Statistical analyses
We conducted all analyses for males and females separately. We used multiple regression analyses to investigate the associations of physical characteristics and personality in adolescence with physical activity (counts/  Table 3 presents descriptive data of the participants during adolescence and adulthood. Multivariate regression analyses revealed no significant associations between physical characteristics and/or personality in adolescence and physical activity in adulthood (data not shown). Regarding sedentariness, in males, a lower score on self-sufficient/recalcitrant and a higher score on facilitating anxiety was associated with more minutes spent sedentary in adulthood. In females, a superior standing high jump performance, and a lower score on social desirability were associated with more minutes spent sedentary at age 42 (Table 4).

Discussion
Previous studies suggest that people with an 'easy going' personality practice healthier lifestyles [16,17]. However, we found that male subjects who possessed more self-  sufficiency/recalcitrance were less sedentary as adults.
Individuals with a more self-sufficient/recalcitrant personality, characterised by higher levels of rebellion and hostility [13], might be more restless and volatile and thus engage in less sedentary behaviour. Males who scored higher on facilitating anxiety, characterised by impulsivity and sensation/stimulation seeking, spent more time being sedentary in adulthood. Also, a superior standing high jump performance in girls was associated with more sedentary time in adulthood. Previous studies found that similar personality and physical characteristics were positively associated with physical activity; i.e. extravert and conscientious people were more physically active [18], and sufficient levels of muscular fitness were predictors of physical activity at a later age [19,20]. In our study these characteristics predicted sedentary time as well. This supports the assumption that physical activity and sedentary behaviour are two different types of behaviour [1][2][3], and that people who are sufficiently physically active can be highly sedentary at the same time. Since little evidence on determinants of sedentary behaviour is available, more prospective research needs to be conducted to confirm our findings and establish the mechanisms causing these relationships.
To the best of our knowledge, up to now the association between social desirability and sedentary time has only been explored by Jago and colleagues [21]. Although Jago and colleagues examined a slightly different study sample (10 to 14-year old Boy Scouts) with different measures (self-reported sedentary time instead of accelerometry) using a cross-sectional design, a comparable inverse association between social desirability and sedentariness was found. A possible explanation for this association might be that people with a less social desirable nature care less about prevailing norms in society and therefore participate less in social desirable behaviour. Currently much attention is paid to initiatives trying to increase people's physical activity level and decrease their time spent sedentary [22]. People who score low on social desirability may be less likely to participate in such initiatives.

Limitations
Several limitations are noteworthy. First, participants were rather active as compared to the general Dutch population [23], which may be explained by the participants' relatively high educational background [24]. Therefore, the current results may not be generalisable to the Dutch population. Second, our study sample significantly differed from the baseline sample and from subjects who did not provide ActiGraph data at age 42 which may have biased our results Third, accelerometry is not a gold standard for measuring physical activity nor sedentary time. Although accelerometry provides real time data storage, it does not provide qualitative information on the type of activity. Besides, accelerometry underestimates some activities, such as cycling. Since cycling is a common method of transportation in the Netherlands, underestimation of physical activity may have occurred. Therefore, our findings should be interpreted with caution.

Conclusion
Sedentary behaviour in adulthood was partly explained by physical characteristics and/or personality in adolescence. Our findings need to be confirmed in other studies. LU performed the statistical analyses, interpreted the data and drafted the manuscript. AS participated in the design of the study, contributed to the analyses and interpretation of data and provided critical revision of the manuscript. JT participated in the fund raising, conception and design of the current study, provided statistical expertise and critical revision of the manuscript, and participated in the conception, design and data acquisition of AGAHLS. LK participated in the fund raising, conception and design of the study, and provided critical revision of the manuscript. WM provided critical revision of the manuscript, and participated in the conception, design, fund raising and data acquisition of AGAHLS. MC participated in the design of the study, contributed to the analyses and interpretation of data and provided critical revision of the manuscript. All authors read and approved the final manuscript.

Competing interests
The authors declare that they have no competing interests.