This population based longitudinal study explored differences in physical activity patterns between adolescence and young adulthood were associated with cardiovascular disease (CVD) risk factors and mental health in young adulthood. To our knowledge no other study of this size has previously examined tracking of PA associated with CVD risk factors and mental health.
To characterize participants, we compared four physical activity patterns; active maintainers (AMs), adopters, relapsers and inactive maintainers (IMs), observed over a ten-year period. In adolescence (baseline) and in young adulthood (follow-up) there were no significant differences (mean) among these physical activity groups and overweight/obesity measured as BMI. This is in accordance with studies on the relationship between physical activity and obesity measured as BMI, who reported that the relationship may be more strongly related to cardiorespiratory fitness or screen activities (watching television, computer use) than physical activity [30, 31]. But the obesity-related indicator, waist circumference (WC) was lower among AM males than relapsers and IMs. This is supported by Lakerveld et al., who found that abdominal obesity was associated with reduced physical activity over a period of five years . Analyses stratified by gender showed that the association between PA and WC was only significant in males. An inverse relationship between physical activity and overweight/obesity is well known [33, 34], although few studies have examined males and females separately.
We found no association between PA behaviour and WC, comparing adopters and IMs + relapsers, even in an unadjusted analysis. Overweight/obesity in adulthood has been associated with decreased probability of being physically active . The lack of association between the adopters and WC in our study, contrasts with the assumption that improvement in physical activity during this time period reduces subsequent obesity and CVD risk .
Physical inactivity is a well-known CVD risk factor in adolescence , and patterns of physical activity maintenance may be protective in longitudinal studies. This was confirmed when we compared the two extreme groups (IMs compared to AMs); AMs had a much better CVD risk profile in adulthood than the IMs, both in the unadjusted and adjusted analyses, and supporting other studies reporting that physical activity and physical fitness may protect against CVD risk . We found a significant gender difference in the relationship between PA and CVD risk factors. Resting heart rate was lower for AMs in both sexes, but AM males had a considerably smaller waist circumference (WC), lower TC (triglyceride concentration) and higher HDL cholesterol (HDL-C) than the IMs. The gender difference could be anticipated for TC, where previous population studies have found a higher TC in men compared women in the first fifty years . There is also some evidence that a high level of TC is a significant independent risk factor for CVD for both genders . HDL-C is an independent predictor of CVD, both in males and females, but females may have a lower HDL-C than males . Our results show an opposite gender difference, female IMs did not have a higher risk for low HDL-C than AMs. However, it is still important to focus on TC and HDL-C level, in both sexes among those who are not physically active.
The PA behaviour and CVD risk may be confounded by obesity, which we know is associated with higher triglyceride levels and cholesterol. But our descriptive data (Table 1) indicated no significant differences in BMI between our four physical activity patterns. This lack of difference in BMI between the different PA patterns is interesting, especially when the patterns did show differences for CVD risks. Additional analyses examined how obesity affected the associations, by adjusting for obesity measured as BMI, both at baseline and follow-up (data not shown). These analyses did not attenuate the results presented in the tables, indicating that physical activity patterns and subsequent CVD risk are likely to be independent of obesity.
Several cross-sectional studies have shown a positive effect by physical activity on mental health in general, and in particularly on self-perception and self-esteem . Physical activity has also been recommended as a tool in therapy for depression and anxiety , but information on how different physical activity behaviours from adolescence to young adulthood affects mental health and satisfaction with life in adulthood is sparse. Our longitudinal data indicate that the AMs had better life satisfactions and mental health status than inactive maintainers. In addition, there were gender differences, with female AMs having a lower likelihood of feeling nervous/restless and being troubled by anxiety, compared to inactive maintainers. We did not see this in males, but the odds ratios indicate the same trend also for males. This is in accordance with previous studies, where physically active adults’ reports fewer symptoms of anxiety than physically inactive persons . Some longitudinal studies have also found negative associations between sedentary behaviour and mental health, while sedentary behaviour as TV viewing was associated with increased odds of mental distress .
Physical activity is important for maintaining good health, and physically inactive people have a higher incidence of cardiovascular disease . In addition to its preventive effect, physical activity is also recommended in treatment of several chronic diseases . It could therefore be expected that those who increased their physical activity from adolescence to adulthood (adopters) might have a lowered CVD risk and better mental health than those who stayed inactive (IMs) or relapsed to lower physical activity. Our data indicated, however, that the adopters had no metabolic or mental health advantages compared to IMs and relapsers. This is rather surprising, because we would expect that increased physical activity would have a positive effect on risk factors. One explanation could be that their PA increase was minor and occurred late in the measured period. According to the literature , we expected that the change of PA would take place in adolescence (years zero to four in the ten-year period) but our data on adopters and relapsers did not support this. In additional analyses we also compared these two groups (adopters and relapsers) according to when they changed the PA in this period. They did not differ in CVD risk factors. The higher cut point for physical activity did not change our findings.
The unexpected findings on adopters compared to inactive maintainers and relapsers (Table 4–5), became more apparent when we compared adopters with AMs. The active maintainers had a better CVD risk profile in adulthood than the adopters, allowing us to combine the adopters, IMs and relapsers in one group. AMs differed from the other groups concerning association with CVD risk factors. We also revealed the same tendency on the likelihood of physical activity patterns associated with mental health and satisfaction with life. A physiologically plausible explanation could be that adopters altered their physical activity pattern late in the period, and we therefore cannot distinguish them from IMs. Our additional analysis shows, however that it mainly changed early in the adolescence in our 10-year follow-up. The adopter group was also smaller than the other groups, which could also be an explanation for the unexpected absence of positive outcome of adopting PA. But, in the descriptive data the adopter group is quite similar to the IMs and relapsers (Table 1).
Our longitudinal data indicate that AMs are more likely to have better mental health than IMs, relapsers and adopters. This also applies perceived health and satisfaction with life, where AMs had higher odds for subjective good health and are more satisfied with their life compared to the other physical activity patterns.
The main strength of the present study is the ten-year follow-up from adolescence to young adulthood in a representative population-based sample. The study is also unique looking at ten-year physical activity patterns, the change of physical activity, assessing physical activity over several time points, and its impact on subsequent cardio-metabolic risk factors and mental health. However, we acknowledge there are some limitations. Physical activity is measured using validated questions, but relies on self-report, rather than objective measures. These physical activity questions have shown high reliability and acceptable validity [13, 26], and dichotomisation as “active” and “inactive” provides good information on physical activity patterns. Another limitation could be that PA behavior might have changed very early in the ten-year period studied, obscuring the classification in some of the participants. Measuring physical activity four years after baseline, thus being able to see when the change of PA behaviour occurred both for adopters and relapsers (see Results section), revealed that nearly the same number of people changed their PA early and late in this ten-year period. We thus believe that this limitation is of minor importance.
Not having basal metabolic measurements in adolescence is a limitation because some of the participants could have CVD risk or mental distress at baseline, independent of their physical activity. Another limitation to our study may be the low participation rate at follow-up. HUNT Studies are based on repeated cross sectional studies, but as HUNT is a study of a total population longitudinal studies, as the follow-up from Young-HUNT1 to HUNT3, are possible. Although the participation rate in Young-HUNT 1 was high, the participation rate in HUNT 3 in the age group 20–29 years was low (31.5%). Many people in this age group had moved out of the county for further education or work and were not eligible for in vitiation to the HUNT 3 survey. Of the invited (5353 people), 42% of the women and 30% of the men participated. There were no significant differences between Young-HUNT1 participants who also participated at HUNT3 (follow-up) and those who did not regarding mean BMI, systolic and diastolic blood pressure, heart rate and physical activity. We therefore have strong reasons to believe that there are no major selection effects on physical activity or health behaviors between the two groups.