Life-course leisure-time physical activity trajectories in relation to health-related behaviors in adulthood: the Cardiovascular Risk in Young Finns study

Non-communicable diseases, such as cardiovascular disease, diabetes mellitus, and several cancers, cause over 70% of deaths worldwide with unhealthy behaviors usually causing these diseases [1]. Physical activity (PA), dietary behavior, smoking and consuming alcohol are traditionally seen as the major four behaviors impacting health [2]. Lately, sleeping and sedentary behavior have been identified as important additional behaviors impacting it as well [3‐6]. Physical inactivity, unhealthy diet, regular smoking, and binge drinking [7], shortened [4] and prolonged sleep duration [3, 4], insomnia [5], and sedentary behavior [6, 8] (e.g., prolonged television viewing [9] and total screen time [10]) have been found to be associated with a higher risk of non-communicable diseases and mortality.

Behavioral risk factors tend to accumulate to same individuals which may have synergistic effects on health [11]. A study examining smoking, alcohol intake, diet, PA, television viewing, and sleep together showed that a combination of multiple health-compromising behaviors was strongly associated with cardiovascular disease mortality and all-cause mortality [12] while a combination of multiple health-enhancing behaviors have been found to decrease the relative risk for all-cause mortality [13]. Thus, promoting the adoption of multiple healthy behaviors is essential for improving public health.

Changing one behavior may translate into changing other behavior as well, especially if both behaviors are health-enhancing or health-risky [14]. Previous evidence suggests that PA might play an important role for adopting to multiple healthy lifestyle factors, e.g. eating healthier [15, 16]. A longitudinal study found that adults who increased their PA also improved their diet when compared to their decreasingly active peers [17], whereas low PA was found to associate with higher risk of sleep problems and incident short sleep time among middle-aged and older adults [18]. Additionally, physical inactivity in adolescence predicted smoking [19] and weekly alcohol intoxication in young adulthood, especially in women [20]. Our previous longitudinal studies showed that persistently physically active adult women watched television less than low-active women [21], leisure-time physical activity (LTPA) evolved in tandem with fruit and vegetable consumption from childhood to adulthood [22], persistent inactivity was associated with smoking [23, 24] whereas, in contrast, persistent higher LTPA was associated with regular alcohol drinking [23].

However, previous research has mainly focused on the association of PA or LTPA with single health behavior. Moreover, longitudinal studies on whether a specific LTPA life-course developmental pathway facilitates the adoption of several other health-enhancing behaviors in adulthood are lacking. This study aimed to investigate whether cross-sectionally measured health-related behaviors in adulthood – including diet, screen time, smoking, binge drinking, sleep difficulties, and sleep duration – differ by LTPA pathways measured from childhood to adulthood in men and women. LTPA was defined as participation in PA and sport by querying intensity, duration and frequency of PA, and participation in organized sports. Occupational PA was not queried. Data-driven trajectory modeling [25] was chosen as the method for identifying LTPA pathways, as trajectories can yield novel information on the complexity of PA behavior [26].

This study examined how diverse pathways (i.e., trajectories) of LTPA from childhood to adulthood were associated with diet, screen time, smoking, binge drinking, sleep difficulties and sleep duration in adulthood in men and women. Higher LTPA was associated with non-smoking and following a healthier diet in both genders and with less frequent sleep difficulties in women. Men whose LTPA increased in young adulthood reported lower screen time than persistently active and low-active men. Thus, those in the persistently, and especially increasingly active, life-course trajectories had accumulated several health-enhancing behaviors. Simultaneously, those in the inactive or low-active life-course trajectories had accumulated health-compromising behaviors. This is a concern from the public health perspective, as the biggest proportion of participants, of both sexes, were in the low-active trajectories. Our results are supported by previous cross-sectional [2, 53] and longitudinal [54, 55] findings showing a clustering of health behaviors at both ends of the lifestyle spectrum: the healthy and unhealthy. Previous findings show the phenomenon being more evident in women than men [38] which was observed also in the current study with generally higher effect sizes observed in women.

The poorest diet was found in the inactive trajectory for women and the low-active trajectory for men. This supports our previous findings [22] and those of another study showing that PA and fruit and vegetable consumption, an indicator of a healthy diet, are accumulated by the same individuals [53]. Of the studied behaviors, dietary behavior in adulthood – especially in women – showed the strongest association with the life-course LTPA trajectories. Several possible mechanisms behind the association have been proposed. For example, positive experiences from exercising may improve self-esteem [56], and body-image [57], and lead to promote individuals’ self-efficacy and motivation to modify dietary habits as well [58‐60]. Also, the similar barriers and motivators experienced for these two behaviors could explain the strong association. A study by Ashton et al. [61] showed that, in young men, motivators for PA and healthy eating were improving physical health, performance and physical appearance while logistic barriers (cost and access) and social factors (e.g., peer influence) were found as barriers for these two behaviours. Generally, women perceive healthy eating as more important [62, 63], experience a higher need for weight control [62], and are more health conscious [64] than men which could explain why the associations between LTPA and diet were stronger in women.

The strongest associations between LTPA trajectories and sleep difficulties were detected in the fully adjusted model for women that included only the three oldest age cohorts (42–49-year-old participants). Sleep difficulties were less prevalent among the increasingly and persistently active women than among their persistently low-active and inactive peers, while no associations were found among men. Generally, poor sleep quality is more prevalent among women than men [65]. In our sample, only 14% of men and up to 20% of women reported severe sleep difficulties in their forties. One explanation for the sex difference is probably the menopausal transition which is related to adverse changes in sleep quality [66, 67]. Previous studies on older adults have reported higher PA to be associated with maintenance of sleep sufficiency [68], lower probability of daytime sleepiness [69], and less sleep difficulties [18]. Among middle-aged adults, the findings are contradictory, some studies reporting an inverse [18, 68] and others no association between PA and sleep quality [69]. When studying adolescents, thus younger populations, no associations between PA trajectories and sleep time were found [70]. These age group-specific associations might reflect the aging-related decline in sleep quality [71] and, on the other hand, the importance of being physically active while ageing as it may improve sleep quality. Older age and female gender [69] both seemed to explain the associations we found between LTPA trajectories and sleep difficulties.

Several studies have reported an inverse association between PA and smoking: for example, smoking [72] and nicotine dependency [73] predicted lower levels of PA. Conversely, inactivity or occasional activity in adolescence predicted a higher prevalence of daily smoking in young adulthood [19]. Both our previous [23, 24] and current results corroborate those findings: LTPA development was inversely associated with smoking in adulthood for both genders – though not as strongly as it was with dietary behavior. Possible explanations for this include the clustering of positive and negative behaviors, and various psychological (e.g., depression), socio-demographic (e.g., education level), or physiological (e.g., lung capacity) factors [74]. Also, people who are physically active usually value physical fitness and strength [75], and are aware that smoking weakens the possibility to improve them.

Somewhat unexpectedly, not only the persistently low-active but also the persistently active men reported higher leisure screen time than their increasingly active counterparts. Earlier studies have found that men watch more sport on television [76], engage more in video gaming [77] and are more sensation-seeking (i.e., willing to engage in novel and intense activities) [78] than women. The persistently active trajectory might include a selected group of men seeking intense activities via participating in sports, viewing sports and playing e-sports or other video games. After full adjustments, no associations between LTPA trajectories and screen time in women were observed. Similarly, LTPA trajectories were not associated with sleeping duration nor binge drinking in either gender which differed from previous findings [18, 20].

The strengths of the current longitudinal study were its six age cohorts and several measurement points during a follow-up of over 30 years, enabling the study of the associations between life-course LTPA trajectories and several other health behaviors in adulthood. In this study, we expanded our previous research on health behaviors by including sleeping behavior, using an overall healthy diet index instead of measuring only fruit and vegetable consumption, and adding screen time to television viewing as sedentary behavior. Finally, instead of regular alcohol consumption we studied binge drinking, since this has been found to have marked effects on cardiovascular disease morbidity and mortality [79].

The study had its limitations. The use of self-reports may bias the results and lead to under- and over-reporting [80‐82]. The proportion of women identified in the persistently active trajectory was small (3.4%): the association between persistent lifelong LTPA and other health behaviors among women should be confirmed in future studies. The measure of screen time did not include the use of electronic mobile devices such as smart phones or tablet computers and therefore does not cover all aspects of screen time. The models were adjusted for previous corresponding behaviors in order to ascertain whether the association between LTPA and a behavior is predicted by LTPA or by the behavior itself. This led to a lower sample size in the fully adjusted models which is why sensitivity analyses were performed in order to detect potential selection bias. Additionally, previous sleeping behavior in 1986 was assessed only from the three oldest age cohorts leading to a selection of older sample when compared to the other fully adjusted models. Moreover, if the study question concerning binge drinking had been phrased separately for both genders allowing the use of a lower threshold value for women, a bigger proportion of women might have been defined as binge drinkers. Even though several covariates were used, the associations may also be affected by unmeasured confounders (e.g., chronic diseases [83, 84], mental health issues [85], temperament [86], transitions and life events [87], social support [88], or occupational status [87]). Conclusions on causality may be biased, as observational studies can include reverse causation. The study sample presents a Finnish population, and therefore, the results are not necessarily generalizable to other populations with, for example, a different socio-economic or ethnic background. Finally, researchers and readers need to acknowledge that trajectory group membership is not certain [89] as it only presents the probability of the participant to follow a trajectory.

In past studies, the previous PA level has been found to predict the future PA level [42, 90]. The current study adds to those findings by highlighting the importance of managing to increase LTPA during the life-course as not only persistent but also increasing activity was associated with several healthy behaviors in adulthood. This is an encouraging message for health promotion: the LTPA level in childhood and adolescence does not necessarily determine the LTPA level later in life, and LTPA initiated even after adolescence may play a role in adoption to other health-promoting behaviors in adulthood. These results may be leveraged as a platform for trajectory group-specific PA counseling with potential ramifications for additional healthy lifestyle choices. Our results also generate hypotheses for future qualitative research on the reasons underlying LTPA behavior, and for intervention studies to ascertain the causal relations behind these associations. Moreover, the constantly developing objective measurements enable collecting data on how PA, sedentary behavior, and sleeping integrate across the whole day (see e.g. [91]). Studies using accelerometers can provide a more comprehensive understanding of the codependency of these so-called time-use behaviors. Future longitudinal studies could identify trajectories/profiles of 24-h time-use behaviors (see e.g. [92]) and study their associations with other lifestyle choices.

The present findings underline the different associations between life-course LTPA trajectories and various other health-related behaviors in adulthood and the sex differences in these associations. While the findings corroborated previous findings on the accumulation of healthy behaviors with persistent activity and of unhealthy behaviors with persistent inactivity and low activity, they also highlighted the meaning of increase in LTPA level. Especially, the study offered insight into the association between persistent LTPA during life-course and increased LTPA after childhood or adolescence with healthy eating and non-smoking in both genders and having less sleep difficulties in middle-aged women. The results suggest that to be physically active in adulthood does not necessarily require the initiation of a physically active lifestyle already in childhood. Promoting LTPA at all ages is important, not only because it may increase the level of PA, but because of its potential associations with healthy lifestyle choices later in life.