The characteristics of the study population in midlife and means and standard deviations of temperament traits (TCI scores) across the clusters at the 31 years in males and females are presented in Table 1 and in Additional file 1. In this study, the participants who continued to the 46-year follow-up study and returned PA questionnaires were more often female than male (56% vs. 44%; p < .046). Males who continued to the 46-year follow up study perceived their health more often as very good or good than fair, poor or very poor (66% vs. 45%; p = .040). Females had higher mean (SD) scores than males in novelty seeking (20.7 [6.0] vs. 19.5 [5.9]; p < .001), harm avoidance (15.1 [5.9] vs. 12.8 [6.0]; p < .001), and reward dependence (16.1 [3.4] vs. 13.2 [3.7]; p < .001). Conversely, males scored higher than females in persistence (4.6 [1.8] vs. 4.1 [1.7]; p < .001). Males reported higher MVPA (minutes in a week) than females at both time points (Table 1). Both genders showed a slight increase in MVPA time between the two time points.
[insert Table 1 here]
3.1 Cross-sectional associations between temperament clusters and leisure-time physical activity at ages 31 and 46
At the age of 31 years, TCI and PA questionnaires were returned by 3,084 (1,359 male and 1,725 female) participants. The TCI persistent, overactive, dependent, and passive clusters included 28%, 22%, 29%, and 21% of the male participants, and 26%, 24%, 28%, and 22% of the female participants, respectively. Individuals in the passive cluster perceived their health more often as fair, poor, or very poor than individuals in other clusters (Male persistent 26%, overactive 22%, dependent 30%, and passive 44%, χ2[3, N = 1357] = 40.33, p < .001; Female persistent 28%, overactive 26%, dependent 31% and passive 43%, χ2[3, N = 1721] = 31.14 p < .001). At the age of 46 years, TCI and PA questionnaires were returned by 2,985 (1,303 male and 1,682 female) participants. The TCI persistent, overactive, dependent, and passive clusters included 26%, 24%, 28%, and 22% of the male participants, and 26%, 25%, 27%, and 22% of the female participants, respectively.
The self-reported MVPA levels among males and females in the four TCI clusters at the age of 31 and 46 years are presented in Table 2. The MVPA levels differed between the TCI clusters, with the passive participants having the lowest PA level in both time points. Post hoc analysis revealed significant difference in MVPA level between the overactive and passive clusters in both genders in both time points. Post hoc comparisons also showed that the females in the persistent and overactive clusters reported higher MVPA levels than those in the dependent and passive clusters.
At the age of 31 years, the rates of belonging to the group of engaging more than 150 minutes of MVPA in a week were 25%, 34%, 24%, and 22% among males (χ2[3, N = 1359] = 12.60, p = .006) and 24%, 28%, 16%, and 15% among females (χ2[3, N = 1725] = 29.29, p < .001) in the persistent, overactive, dependent, and passive clusters, respectively. At the age of 46 years, the rates of belonging to the group of more than 150 minutes of MVPA in a week were 32%, 31%, 28%, and 26% among males (χ2[3, N = 1303] = 3.73, p = .292) and 37%, 33%, 25%, and 22% among females (χ2[3, N = 1682] = 30.04, p < .001) in the persistent, overactive, dependent, and passive clusters, respectively.
[insert Table 2 here]
3.2 Associations between temperament clusters and changes in leisure-time physical activity from age 31 to age 46
The frequencies of TCI clusters and longitudinal changes in MVPA levels from age 31 to age 46 are presented in Table 3. More than half of the participants belonged to the group of engaging less than 150 minutes of MVPA in a week between young adulthood and midlife.
Table 4 shows the associations between TCI clusters and longitudinal changes in MVPA levels according to the logistic regression analysis. A stable high MVPA level from young adulthood to midlife was more common among males (OR = 1.76, 95% CI: 1.19 − 2.59, p = .004) and females (OR = 2.12, 95% CI: 1.50 − 3.01, p < .001) in the overactive cluster and among females (OR = 1.75, 95% CI: 1.23 − 2.49, p = .002) in the persistent cluster. It was less common among females (OR = 0.44, 95% CI: 0.28 − 0.68, p < .001) in the dependent cluster and among females (OR = 0.42, 95% CI: 0.26 − 0.68, p < .001) in the passive cluster. These associations remained statistically significant after adjusting for presence and severity of anxiety and depressive symptoms at the age of 31 years, perceived health at the age of 31 years, level of education, smoking status, alcohol consumption, employment status, marital status, employment status and perceived health at the age of 46.
A decreasing MVPA level from young adulthood to midlife was more common among males (OR = 1.76, 95% CI: 1.22 − 2.55, p = .003) in the overactive cluster. An increasing MVPA level was more common in females (OR = 1.64, 95% CI: 1.25 − 2.26, p < .001) in the persistent cluster and less common in passive cluster (OR = 0.65, 95% CI: 0.47 − 0.89, p = .007). These associations remained statistically significant after adjusting for presence and severity of anxiety and depressive symptoms at the age of 31 years, perceived health at the age of 31 years, level of education, smoking status, alcohol consumption, employment status, marital status, and perceived health at the age of 46 (Table 4). A decreasing MVPA level from young adulthood to midlife was less common among males in the passive cluster (OR = 0.64, 95% CI: 0.42 − 0.98, p = .039), however this associations did not remain significant after the model was adjusted for presence and severity of anxiety and depressive symptoms at the age of 31 years, perceived health at the age of 31 years, level of education, smoking status, alcohol consumption, employment status, marital status, and perceived health at the age of 46.
[insert Table 3 here]
[insert Table 4 here]
3.3 Associations between temperament traits and leisure-time physical activity at ages 31 and 46
The correlations between baseline temperament traits (TCI scores) and MVPA levels at the ages of 31 and 46 years are shown in Additional file 2. Lower harm avoidance scores were associated with longer MVPA times in both genders at both time points. Higher reward dependence scores were associated with longer MVPA times among males at the ages of 31 (rho = 0.07; p = .014) and 46 (rho = 0.11; p < .001). Higher persistence scores were associated with longer MVPA times in both genders at the ages of 31 and 46. Higher novelty seeking scores were associated with longer MVPA times at the age of 31 only in females (rho = 0.07; p = .006).
The associations between TCI scores and longitudinal changes in leisure-time MVPA level from the age of 31 to the age of 46 years according to the logistic regression analysis are presented in Additional file 3. Higher novelty seeking scores were associated with decreasing MVPA levels among males (OR = 1.25, 95% CI: 1.06 − 1.47, p = .009) and remained significant after adjusting for presence and severity of anxiety and depression at the age of 31 years, perceived health at the age of 31 years, level of education, smoking status, alcohol consumption, employment status, marital status, and perceived health at the age of 46.
Lower harm avoidance scores were associated with decreasing MVPA levels in both genders and remained after adjusting for above mentioned covariates. Lower harm avoidance scores associated with stability of high MVPA levels in both genders and remained significant after adjusting for covariates in females. Higher reward dependence scores were associated with increasing MVPA levels among males (OR = 1.25, 95% CI: 1.11 − 1.44, p = .007) and remained significant after adjusting for covariates. Higher persistence scores were associated with increasing MVPA levels among females (OR = 1.26, 95% CI: 1.11 − 1.44, p < .001), and with stability of high MVPA levels among males (OR = 1.38, 95% CI: 1.16 − 1.65, p < .001) and females (OR = 1.36, 95% CI: 1.15 − 1.61, p < .001), but also with decreasing MVPA levels among females (OR = 1.27, 95% CI: 1.08 − 1.50, p = .004). The mean TCI scores across the PA groups according to changes in leisure-time MVPA from the age of 31 and to the age of 46 years are presented in Additional file Table 4.