Physical exercise in adults and mental health status

Physical exercise in adults and mental health status

Journal of Psychosomatic Research 71 (2011) 342–348 Contents lists available at ScienceDirect Journal of Psychosomatic Research Physical exercise i...

217KB Sizes 0 Downloads 64 Views

Journal of Psychosomatic Research 71 (2011) 342–348

Contents lists available at ScienceDirect

Journal of Psychosomatic Research

Physical exercise in adults and mental health status Findings from the Netherlands Mental Health Survey and Incidence Study (NEMESIS) Margreet ten Have a,⁎, Ron de Graaf a, Karin Monshouwer a,b a b

Netherlands Institute of Mental Health and Addiction, Utrecht, the Netherlands Department of Social Science, University of Utrecht, the Netherlands

a r t i c l e

i n f o

Article history: Received 29 November 2010 Received in Revised form 29 March 2011 Accepted 5 April 2011 Keywords: Affective disorders Anxiety Physical exercise Population surveys Substance use disorders

a b s t r a c t Objective: To establish associations between physical exercise during leisure time and prevalence, incidence and course of mental disorders. Method: Data were derived from the Netherlands Mental Health Survey and Incidence Study, a 3-wave cohort study in a representative sample (N = 7,076) of Dutch adults. Mental disorders were assessed with the Composite International Diagnostic Interview. Physical activity was established by the number of hours per week people spent on taking physical exercise. Results: Physical exercise was negatively associated with presence and first-onset of mood and anxiety disorders after adjustment for confounders. Evidence for a dose–response relationship between exercise levels and mental health was not found. Among those with mental disorder at baseline, exercise participants were more likely to recover from their illness (OR = 1.47) compared to their counterparts who did not take exercise. Conclusion: Physical exercise is beneficial to mental health, but it remains uncertain whether this association truly reflects a causal effect of exercise. © 2011 Elsevier Inc. All rights reserved.

Introduction In the last two decades relatively few large-scale population-based studies have examined the relationship between physical activity and mental health. The cross-sectional studies conducted thus far have well established that regular physical activity is negatively associated with symptoms of anxiety or depression [1–9]. This relationship is found among adolescents as well as adults, among men as well as women, and seems to remain after adjustment for possible confounders. We found only one large-scale study that failed to establish a positive cross-sectional association between physical activity and emotional wellbeing after adjustment for confounders [10]. It should be noted that the focus of the majority of these studies was rather limited, because they only looked at the association between physical activity and symptoms of anxiety or depression. Other mental health symptoms or symptoms that met criteria for a mental disorder were not studied. Only three studies examined whether physical exercise is negatively related to substance misuse,

Abbreviations: NEMESIS, Netherlands Mental Health Survey and Incidence Study; CIDI, Composite International Diagnostic Interview; T0, first wave; T1, second wave; T2, third wave; OR, odds ratio; CI, confidence interval; Adj, adjusted. ⁎ Corresponding author at: Netherlands Institute of Mental Health and Addiction, P.O. Box 725, 3500 AS Utrecht, the Netherlands. Tel.: + 31 30 297 11 00; fax: + 31 30 297 11 11. E-mail address: [email protected] (M. ten Have). 0022-3999/$ – see front matter © 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.jpsychores.2011.04.001

aggressive behaviour and withdrawn behaviour, besides symptoms of anxiety and depression [1,4,7]. The findings of longitudinal studies on this topic are somewhat more ambiguous. Nine studies found that regular physical activity at baseline is associated with fewer symptoms of anxiety or depression at follow-up [11–19]. Three other studies found a longitudinal association in men but not in women [20–22], and five studies did not find evidence for a longitudinal association at all [23–27]. Possible reasons for not finding a significant association between physical activity and mental health status are: the duration of the follow-up period was longer (more than 5 years versus a shorter follow-up period); mental health problems were measured more stringently (using diagnostic criteria versus symptom scales); the relationship was expressed differently (using analyses of congruence of change versus otherwise); the relationship was assessed in a specific subpopulation (adolescents or elderly versus adults); and the analyses were adjusted for different sets of confounders. Again, the focus of these longitudinal studies was rather limited, because the majority only looked at the association between physical activity at baseline and the presence of symptoms of anxiety or depression at follow-up. Other mental health symptoms or symptoms that met criteria for a mental disorder were not linked to physical activity levels. Only one longitudinal study examined the association between physical activity and the first-onset of a broad range of mental disorders [13]. It found that regular physical activity is associated with a substantially lower incidence of anxiety disorders

M. ten Have et al. / Journal of Psychosomatic Research 71 (2011) 342–348

and dysthymia, but not of major depressive disorder among adolescents and young adults. The other longitudinal studies focused on cumulative incidence rates of mental health problems, and as a result, the direction of any associations could be established less clearly than was possible in the study of Strohle and colleagues [13]. A longitudinal design offers the possibility of studying determinants of the course of mental disorders. To date, none of the population-based studies have focused on whether physical activity is associated with a shorter course of a variety of mental disorders. That is a missed opportunity, because if such associations exist, new prevention activities and treatment programmes for mental health problems could be developed. Thus, there are still a lot of unsolved issues concerning the role physical activity can play in the prevention and promotion of mental health: is physical activity negatively associated with diagnostically assessed mental disorders? Does a lack of physical activity actually precede the first-onset of mental disorders? Given a mental disorder, do people who are physically active get better more quickly than their counterparts who are not physically active? Is there a dose–response relationship between level of physical activity and mental health? And, is there reason to believe that the association between physical activity and mental health is different for men and women? In this paper, we will address these issues using data from the Netherlands Mental Health Survey and Incidence Study (NEMESIS), a prospective cohort study in a large sample representative of the Dutch adult general population. A major advantage of NEMESIS is that mental disorders are assessed with a reliable and valid diagnostic instrument. In NEMESIS, physical activity was established according to the number of hours per week people spent on engaging in physical exercise/sport. In the following, we will therefore refer to physical exercise during leisure time and not physical activity in general. Aims of the study are to address the following research questions: 1) Is physical exercise during leisure time negatively associated with the prevalence of mental disorders? 2) Is physical exercise during leisure time negatively associated with the incidence of mental disorders? 3) Is physical exercise during leisure time positively associated with a favourable course of mental disorders? Method Sample Data were derived from the Netherlands Mental Health Survey and Incidence Study (NEMESIS). Methods have been reported elsewhere [28]. Briefly, NEMESIS is a prospective, cohort study of the Dutch general population, aged 18 to 64 years conducted in three waves (1996, 1997, 1999). It is based on a multistage, stratified, random sampling procedure of households, with one respondent randomly chosen in each household and interviewed face-to-face. In the first wave (T0), 7,076 persons were interviewed (response rate 69.7%). They accurately reflected the population in terms of gender, civil status and urbanicity [28]. The nonresponders comprised mainly refusals (23.6%) and to a lesser extent non-contacts (6.7%). The second wave (T1) included 5618 respondents (79.4% of T0 subjects) and the third wave (T2) 4796 respondents (67.8% of T0 subjects). After adjustment for sociodemographic characteristics, a 12-month disorder at T0 only slightly increased the probability of loss to followup between T0 and T1 as well as between T0 and T2 (OR = 1.20, CI = 1.04–1.38; OR = 1.29, CI = 1.15–1.46) [29]. Sample attrition was also slightly associated with a lack of any physical exercise at T1 (adj. OR = 1.14, CI = 1.01–1.28), but not at T2. However, the associations between physical exercise and prevalence of any mental disorder at baseline among the responders and dropouts were similar. NEMESIS was conducted with the approval of the ethics committee of the Netherlands Institute of Mental Health and Addiction, Utrecht, the Netherlands. Respondents provided informed consent

343

according to the prevailing Dutch law of 1996 after having been informed about the aims of the study. Diagnostic instrument The Composite International Diagnostic Interview (CIDI), computerized version 1.1 [30] was used to determine DSM-III-R diagnoses. The CIDI is a fully structured interview developed by the World Health Organization [31] and designed for use by trained non-clinicians. Research has demonstrated acceptable reliability [32] and validity [33] for virtually all CIDI diagnoses. Mental disorders In this paper the following DSM-III-R diagnoses were included: mood disorders (depression, dysthymia, bipolar disorder), anxiety disorders (panic disorder, agoraphobia, social phobia, simple phobia, generalized anxiety disorder, obsessive-compulsive disorder), and substance use disorders (alcohol or drug abuse and dependence). Eating disorders and schizophrenia were not investigated here, because of the very small numbers of diagnoses in NEMESIS. Other included variables were: any mental disorder and two or more mental disorders; both were based on all the above-cited disorders assessed in NEMESIS. The prevalence rates of mental disorders were assessed in the 12 months preceding T0. The incidence rates were based on information from all three waves. People with an incident mental disorder were defined as the ones in whom a mental disorder was assessed for the first time in their lives in the three years between T0 and T2. This means that these rates were based on the population at risk at baseline (those without a particular lifetime mental disorder at T0) and those who developed that particular disorder in the subsequent two waves. To give an example: Of the 3998 respondents without a lifetime major depressive disorder at baseline, the so called population at risk, 6.1% developed a major depressive disorder for the first time in their life during one of the follow-up assessments. This implies that the incidence rates are based on different populations at risk (see Table 3). The recovery rates were based on different populations at baseline (see table 4). Among the respondents with a particular 12-months disorder at baseline, the percentage who did not report that particular mental disorder in the subsequent two waves was calculated. To give an example: of the 266 respondents with a 12-months major depressive disorder at baseline 62.7% recovered from that disorder, or stated differently did not meet the diagnostic criteria assessed for major depressive disorder during one of the follow-up assessments. Measurement of physical exercise Physical activity was assessed at baseline with the question: ‘How many hours per week have you engaged in physical exercise/sport lately?’ It was stressed that engaging in passive pursuits like fishing and chess did not count. Measurement of potential confounders Sociodemographic characteristics These, assessed at T0, include: gender, age, education, partner status, and employment status. Somatic illnesses Presence of one or more of 31 somatic illnesses treated or monitored by a physician in the 12 months before T0 [34]. Methodological studies have documented moderate-to-good concordance between self-reports and medical records [35].

344

M. ten Have et al. / Journal of Psychosomatic Research 71 (2011) 342–348

Fig. 1. Hours per week recently spent on physical exercise among Dutch adults (N = 7076). Mean number of hours per week adults spend on exercise: 2.3 (2.2 - 2.4).

Statistical analysis The data were weighted to ensure they were representative of the national population. Robust standard errors were calculated by using the first-order Taylor-series linearization method, as implemented in the STATA 9.1 program, in order to obtain correct 95% confidence intervals and p-values [36]. First, summary statistics (percentages, means) were used to describe the number of hours per week the respondents have engaged in physical exercise recently (Fig. 1). Second, they were divided into three groups: those who have recently spent 0 hours, 1–3 h, and ≥ 4 h per week on physical exercise. This classification was based on a proportional division of exercise participants over the last two categories. Additional analyses showed that other classifications (0, 1, and ≥2 h per week; or 0, 1–2, and ≥ 3 h per week) did not result in different conclusions (results not shown). Third, the three groups with different levels of physical exercise were characterised according to the above-cited sociodemographic characteristics and the presence of any somatic illness (Table 1). To address research questions 1 and 2, ‘Is physical exercise during leisure time negatively associated with the prevalence and incidence of mental disorders?’ logistic regression analyses were performed,

controlling for gender, age, education, partner status, employment status, and presence of any somatic illness (Tables 2 and 3). In both series of analyses the reference group consisted of those who have spent 0 h per week on physical exercise recently. In answering these research questions, different exercise levels were distinguished to investigate if a dose–response relationship between physical exercise and mental health existed. In these analyses, the reference group consisted of moderately active participants (1–3 h per week). To address research question 3, ‘Is physical exercise during leisure time positively associated with a favourable course of mental disorders?’ logistic regression analyses were performed again, controlling for the same possible confounders (Table 4). Among the respondents with a mental disorder at T0, two groups were distinguished here: those with a disorder who have taken exercise recently and those who haven't. The reason was that the number of respondents with a mental disorder was too small to allow for a further differentiation among the exercise participants. Two-tailed testing procedures were used with 0.05 alpha levels in all analyses except the tests for interaction effects, where an alpha of 0.001 was used. This was decided because of the large number of calculated interaction effects of gender on the associations between physical exercise and the prevalence, incidence and course of mental disorders. Results At the first interview, 45% of the respondents reported that they have not taken exercise lately (Fig. 1). 32% took exercise or played an active sport on average 1–3 h per week and 23% 4 h or more. Adults who took exercise were more likely to be men, of a younger age (18–24 years), to have more education (especially higher professional education, university), to live without a partner, and to be in paid employment (Table 1). If women, those aged 35–54 years, and those living with a partner were more likely to take exercise, they were more likely to engage in it for 1–3 h per week than ≥4 h per week compared to men, subjects aged 18–24 years, and those not living with a partner. For the lower educated and those without a paid employment, the opposite was seen: they engaged less frequently in exercise or sport; but if they did take exercise, they were just as likely to engage in it for 1–3 h per week as ≥4 h per week compared to their counterparts who had more education and a paid employment. Adults who took exercise were also more likely not to have a somatic illness. Table 2 shows that after adjustment for the influence of sociodemographic characteristics and somatic illness, physical exercise was negatively associated with the

Table 1 Sociodemographic characteristics and presence of any somatic illness associated with the number of hours per week recently spent by adults on physical exercise, in percentages (N = 7076). Number of hours per week adults spend on exercise

Gender Male Female Age (in years) 18–24 25–34 35–44 45–54 55–64 Education Primary education Lower secondary education Higher secondary education Higher professional education, university Partner status Living with partner Not living with partner Employment status Paid employment No paid employment Somatic illness Yes No

0h

1–3 h

≥4 h

%

%

%

%

50.6 49.4

42.7 47.4

28.8 34.7

14.1 26.1 24.1 20.9 14.8

39.8 43.4 44.1 50.0 47.5

6.3 36.5 29.3 27.9

Chi-square

P value

28.5 17.9

113.5

b 0.001

29.8 32.8 35.3 31.1 26.5

30.4 23.8 20.6 18.9 26.0

77.9

b 0.001

65.2 53.5 41.8 32.9

16.3 26.8 32.0 41.2

18.5 19.7 26.2 25.9

290.4

b 0.001

68.4 31.6

45.6 43.7

33.3 28.3

21.1 28.0

44.5

b 0.001

69.4 30.6

42.9 50.0

33.9 26.6

23.2 23.4

42.2

b 0.001

40.7 59.3

49.4 42.0

28.8 33.7

21.8 24.3

38.9

b 0.001

M. ten Have et al. / Journal of Psychosomatic Research 71 (2011) 342–348

345

Table 2 Prevalence of mental disorders associated with the number of hours per week adults recently spent on physical exercise, in adjusted odds ratios (adj. OR) and 95% confidence intervals (CI) and inspection of a dose–response relationship (N = 7076). Prevalence of mental disorders

Number of hours per week adults spend on physical exercise

%

Adj. OR

≥4 h

1–3 h

Any mood disorder Major depressive disorder Dysthymia Bipolar disorder Any anxiety disorder Panic disorder Agoraphobia Social phobia Specific phobia Generalised anxiety disorder Obsessive compulsive disorder Any substance use disorder Alcohol abuse Alcohol dependence Drug abuse Drug dependence Any mental disorder Two or more mental disorders

b

(95% CI) 0.63 (0.51–0.79) 0.63 (0.49–0.81) 0.60 (0.41–0.89) 0.56 (0.31–1.03) 0.67 (0.56–0.81) 0.48 (0.31–0.74) 0.50 (0.30–0.83) 0.53 (0.40–0.70) 0.80 (0.63–1.01) 0.61 (0.33–1.12) 0.93 (0.36–2.41) 0.78 (0.61–0.99) 1.03 (0.74–1.43) 0.58 (0.40–0.83) 0.98 (0.42–2.28) 0.64 (0.29–1.40) 0.73 (0.63–0.84) 0.54 (0.43–0.68)

7.7 5.7 2.5 1.1 12.4 2.2 1.5 4.8 7.1 1.2 0.4 8.9 4.6 3.7 0.5 0.8 23.3 7.8

P valuea

Adj. OR b

Dose–response relationship?

(95% CI) 0.58 (0.45–0.76) 0.57 (0.43–0.77) 0.49 (0.31–0.78) 0.83 (0.46–1.50) 0.66 (0.54–0.81) 0.46 (0.28–0.75) 0.87 (0.53–1.45) 0.57 (0.41–0.78) 0.67 (0.50–0.88) 0.78 (0.42–1.45) 0.86 (0.33–2.25) 0.75 (0.58–0.98) 0.93 (0.65–1.34) 0.62 (0.42–0.90) 0.33 (0.10–1.09) 1.10 (0.55–2.22) 0.69 (0.59–0.82) 0.56 (0.43–0.74)

NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS

The results are adjusted for differences in gender, age, education, partner status, employment status, and presence of a somatic illness. a The adjusted ORs between both categories of exercise participants are compared here. b The reference category is 0 h per week spent on exercise.

prevalence of any mood disorder, any anxiety disorder, and any substance use disorder. Examination of the specific disorders demonstrated that this holds for two of the three mood disorders, four of the six anxiety disorders, and alcohol dependence. To give an example: adults who spent 1–3 h per week on physical exercise at baseline had a lower risk for panic disorder in the preceding 12 months compared to adults who did not take exercise at baseline. The odds ratios between both groups of exercise participants did not differ. This means that adults who engaged in exercise for 1–3 h per week at baseline equally often reported mental disorders in the preceding 12 months as those who took exercise more often. None of the above described associations between physical exercise and mental disorders differed significantly between men and women (data not shown). Table 3 shows that after adjustment for the influence of sociodemographic characteristics and somatic illness, physical exercise was negatively associated with the first-onset of any mood disorder and any anxiety disorder, but not with substance use disorder. Examination of the specific disorders demonstrated that this holds for major depressive disorder and five of the six anxiety disorders. The odds ratios between both groups of exercise participants did not differ. Adults who engaged in exercise for 1–3 h per week equally often developed mental disorders

over a three-year period as those who engaged more often. The only exception was seen for social phobia. Taking exercise for 1–3 h per week was associated with a significantly lower incidence of social phobia than exercising more regularly. Pre-existing psychopathology was not a confounder in these analyses. Even after adjustment for the presence of any lifetime anxiety and/or substance use disorder at baseline, the odds ratio of developing any incident mood disorder among those taking exercise for 1–3 h per week was similar to the one presented in Table 3: 0.63 (0.44–0.90). After additionally controlling for any lifetime mood and/or substance use disorder at baseline, the odds ratio of developing any incident anxiety disorder among those taking exercise for 1–3 h per week was also equal to the one presented in Table 3: 0.57 (0.41–0.81). The above-mentioned associations between physical exercise and the first-onset of mental disorders did not significantly differ between men and women (data not shown). Table 4 shows that after adjustment for confounders, the odds ratio of exercise participants with a mental disorder in the 12 months preceding baseline to recover from their mental illness over a three-year period was 1.5 compared to nonparticipants with a mental disorder. This did also hold for any anxiety disorder even

Table 3 The incidence of mental disorders associated with the number of hours per week adults have recently spent on physical exercise, in adjusted odds ratios (adj. OR) and 95% confidence intervals (CI) and inspection of a dose–response relationship (N = 4796). Population at risk

Incidence of mental disorders

Number of hours per week adults spend on exercise

N

%

Adj. OR

≥4 h

1–3 h

Any mood disorder Major depressive disorder Dysthymia Bipolar disorder Any anxiety disorder Panic disorder Agoraphobia Social phobia Specific phobia Generalised anxiety disorder Obsessive compulsive disorder Any substance use disorder Alcohol disorder Drug disorder Any mental disorder

3821 3998 4432 4697 3815 4605 4630 4444 4316 4537 4756 4036 4101 4678 2874

5.6 6.1 1.2 0.5 5.7 1.4 1.5 1.6 4.1 1.8 0.3 4.2 4.2 0.8 11.7

P valuea

Adj. OR b

(95% CI) 0.62 (0.43–0.89) 0.63 (0.45–0.88) 0.52 (0.26–1.06) 0.77 (0.26–2.27) 0.56 (0.40–0.79) 0.46 (0.25–0.84) 0.45 (0.25–0.80) 0.35 (0.18–0.67) 0.52 (0.36–0.77) 0.44 (0.24–0.79) 0.82 (0.25–2.72) 0.86 (0.51–1.43) 0.89 (0.52–1.52) 0.50 (0.19–1.30) 0.68 (0.49–0.95)

b

(95% CI) 0.79 (0.53–1.18) 0.79 (0.55–1.13) 0.72 (0.32–1.63) 0.31 (0.06–1.59) 0.71 (0.48–1.07) 0.26 (0.10–0.66) 0.50 (0.24–1.04) 1.06 (0.60–1.88) 0.72 (0.47–1.11) 0.65 (0.32–1.32) 0.77 (0.20–2.95) 0.81 (0.48–1.34) 1.04 (0.62–1.75) 0.30 (0.09–1.03) 0.94 (0.67–1.30)

The results are adjusted for differences in gender, age, education, partner status, employment status, and presence of a somatic illness. a The adjusted ORs between both categories of exercise participants are compared here. b The reference category is 0 h per week spent on exercise.

Dose–response relationship?

NS NS NS NS NS NS NS 0.004 NS NS NS NS NS NS NS

346

M. ten Have et al. / Journal of Psychosomatic Research 71 (2011) 342–348

Table 4 The course of mental disorders associated with the number of hours per week adults have recently spent on physical exercise, in adjusted odds ratios (adj. OR) and 95% confidence intervals (CI) (N = 4796). Population at baseline

Recovery from mental disorder

N

%

Adj. OR (95% CI)a

347 266 105 46 568 105 64 207 324 44 22 312 287 40 991

54.3 62.7 74.2 45.3 54.0 70.3 74.2 70.6 61.6 96.0 75.7 52.0 55.0 35.7 45.1

1.22 1.35 0.93 0.32 1.58 1.38 1.16 1.57 1.58

12-months disorder at baseline

Any mood disorder Major depressive disorder Dysthymia Bipolar disorder Any anxiety disorder Panic disorder Agoraphobia Social phobia Specific phobia Generalised anxiety disorder Obsessive compulsive disorder Any substance use disorder Alcohol disorder Drug disorder Any mental disorder

No mental disorder during the 3-year follow-up period Exercise participants

(0.76–1.95) (0.76–2.40) (0.32–2.69) (0.07–1.47) (1.09–2.29) (0.48–3.98) (0.29–4.61) (0.77–3.21) (0.94–2.65)

b

2.65 1.19 1.03 0.33 1.47

(0.14–49.36) (0.69–2.04) (0.59–1.80) (0.06–1.90) (1.11–1.96)

The results are adjusted for differences in gender, age, education, partner status, employment status, and presence of a somatic illness. a The reference category exists of non-exercise participants; adults who spent 0 h per week on physical activity at baseline. b The absolute numbers were too small to generate a meaningful OR. when pre-existing comorbid mood disorders were taken into account, but not for any mood or substance use disorder. These associations between physical exercise and the course of mental disorders did not significantly differ between men and woman (data not shown).

weakened the effects of physical activity on mental health in the prospective analyses. A third explanation is that of reverse causality. However, we did not investigate whether the presence of a mental disorder actually resulted in a change in physical activity level.

Discussion

Some physical exercise is better than none, but more physical exercise is not necessarily better than some We did not find evidence for a dose–response relationship between physical exercise levels and mental health. This is in line with some other population studies [6,11,18,37], and two reviews [38,39]. None of the associations between physical exercise levels and the prevalence, incidence and course of mental disorders differed between men and women. The findings of other studies are ambiguous. Some studies found interaction effects of gender [19– 23,40], while others did not [1,5,10]. It is difficult to explain these differences, because the studies differed to a great extent in research design, target population and measuring instruments.

Key findings Physical exercise is associated with good mental health Physical exercise during leisure time was associated with a lower prevalence of mental disorders, especially mood and anxiety disorders. This is in line with the majority of cross-sectional studies, which established a negative association between physical activity and symptoms of anxiety or depression (see e.g. 1). Physical exercise during leisure time was also associated with a lower first-onset of mental disorders, especially mood and anxiety disorders. Additional analyses revealed that these relations were not confounded by pre-existing psychopathology. This is in line with the findings of the only other longitudinal study performed on this subject that focused on first-incidences of mental disorders [13]. Strohle and colleagues also found that regular physical activity was associated with a substantially lower incidence of any mental disorder, and any anxiety disorder in particular. However, they did not find a negative association between physical activity and the first-onset of mood disorders, and major depressive disorder in particular. That could have been the result of using another diagnostic instrument (a short form of the CIDI) among a different population (adolescents) than in NEMESIS. Physical exercise during leisure time was also associated with a better course of mental disorders, especially anxiety disorders. As far as we know, such a relationship has not been the focus of population studies before. Additional analyses revealed that the higher recovery rate from anxiety disorders among exercise participants could not be explained by a lower prevalence rate of pre-existing comorbid mood disorders. A lot of the associations found in Table 2 were toned down in the prospective analyses described in Tables 3 and 4. One probable explanation is that the prospective analyses were based on a lower number of respondents. Another possible explanation is that physical activity was assessed at baseline, but that changes in activity level could have occurred during follow-up. This would probably have

Almost half of all adults do not actively engage in physical exercise The average number of hours per week Dutch adults spent on physical exercise during leisure time was 2.3 h. This figure was not influenced by seasonal effects (people are usually more active in the summer than in the winter), because the interviews took place throughout a whole year. Despite the fact that the reported exercise levels are difficult to compare with other studies because of differences in design, methods and target population, they seem comparable with the exercise levels reported in another Dutch study [41]. Exercise participants have on average a higher socioeconomic status Adults who took exercise were more likely to have more education. This is in line with another Dutch study [41]. It does not imply that those with a lower education were less physically active in general. Other physical activities, such as general daily activities and occupational activities were not included in the present study. However, other studies show that education is also negatively associated with physical activity in general [1,13,42]. Exercise participants less frequently report somatic illnesses Adults without a somatic illness more often take exercise. That could be the result of fewer physical impairments or an ongoing

M. ten Have et al. / Journal of Psychosomatic Research 71 (2011) 342–348

physically active lifestyle that protected them from a lot of physical ailments. What preceded physical exercise or what resulted from it, could not be inferred from our analyses. If physical exercise mainly prevents the onset of somatic conditions, then adjustment for these conditions would have led to attenuation of the associations found. Strengths and limitations A significant advantage of NEMESIS is that mental disorders were assessed with a reliable and valid diagnostic instrument in a cohort of adults who were monitored over a period of several years. To date, population-based, longitudinal studies with a substantial number of respondents such as NEMESIS are still scarce. However, this study also has some limitations. Even though the sample was broadly representative of the Dutch adult population, at least three categories were underrepresented: people with an insufficient mastery of Dutch, those with no fixed address, and the long-term institutionalized population. It is difficult to infer how this may have affected the results of the study. It is likely to have led to mitigation, not inflation, of the associations found. People with a mental disorder and those lacking physical exercise were more likely to be lost to follow-up. Since the associations between physical exercise and the prevalence of any mental disorder at baseline among the responders and dropouts were similar, sample attrition seems unlikely to have affected the results in any major way. Physical activity was measured with one question on the number of hours per week people spent on taking exercise. Future research should also document the frequency and intensity of the exercise as well as the nature of the activity (team or individual sports, walking, swimming etc.). This information is useful for revealing the mechanisms and pathways underlying the association between physical exercise during leisure time and mental health. Then it may for example be found that this association is caused by both neurobiological mechanisms (e.g. increased serotonin synthesis) and psychosocial processes (e.g. increased self-esteem, social support, distraction from worrisome thoughts). Besides, future research should also document general daily activities and occupational activities and should not confine the measurement of physical activity to physical exercise during leisure time alone. In this way it can be made clear whether the associations found are also true for certain situations or in different circumstances (doing housework, leisure pursuits, work). The associations found between physical exercise and the firstonset of mental disorders are not by definition causal. Despite the fact that physical exercise was measured at an earlier time point than the onset of mental disorders, non-measured variables could have influenced both exercise levels and the development of mental disorders, such as genetic vulnerability [24]. This paper focused on the association between physical exercise and mental health in the general population. This does not mean that the findings can be generalized to a treatment setting. Conclusions Physical exercise is beneficial to one's mental health, but it remains uncertain whether this association truly reflects a causal or physical effect of exercise. Randomized controlled trials (RCTs) may shed more light on this matter. A recent, systematic review of RCTs showed that exercise seems to improve depressive symptoms in adults with a depressive disorder [43], but that methodologically robust trials are lacking to establish the effect of exercise in reducing depression as well as other mental health problems in the general adult population. Despite this lack of knowledge, our findings support the policies of most Western governments to promote appropriate physical activity. Engaging in

347

sport is inexpensive, moderate exercise has few negative side effects and may even generate positive side (mental health) effects.

Acknowledgments This research was funded by the Ministry of Health, Welfare and Sport.

References [1] Monshouwer K, Ten Have M, Van Poppel M, Kemper H, Vollebergh W. Low physical activity in adolescence is associated with increased risk for mental health problems. Med Sport 2009;13:74–81. [2] Muhsen K, Lipsitz J, Garty-Sandalon N, Gross R, Green MS. Correlates of generalized anxiety disorder: independent of co-morbidity with depression. Findings from the first Israeli National Health Interview Survey (2003–2004). Soc Psychiatry Psychiatr Epidemiol 2008;43:898–904. [3] Ussher MH, Owen CG, Cook DG, Whincup PH. The relationship between physical activity, sedentary behaviour and psychological wellbeing among adolescents. Soc Psychiatry Psychiatr Epidemiol 2007;42:851–6. [4] Tao FB, Xu ML, Kim SD, Sun Y, Su PY, Huang K. Physical activity might not be the protective factor for health risk behaviours and psychopathological symptoms in adolescents. J Pediatr Child Health 2007;43:762–7. [5] De Moor MHM, Beem AL, Stubbe JH, Boomsma DI, De Geus EJC. Regular exercise, anxiety, depression and personality: a population-based study. Prev Med 2006;42: 273–9. [6] Taylor MK, Pietrobon R, Pan D, Huff M, Higgins LD. Healthy People 2010 Physical Activity Guidelines and psychological symptoms: evidence from a large nationwide database. J Phys Act Health 2004;1:114–30. [7] Goodwin RD. Association between physical activity and mental disorders among adults in the United States. Prev Med 2003;36:698–703. [8] Kirkcaldy BD, Shephard RJ, Siefen RG. The relationship between physical activity and self-image and problem behaviour among adolescents. Soc Psychiatry Psychiatr Epidemiol 2002;37:544–50. [9] Steptoe A, Wardle J, Fuller R, Holte A, Justo J, Sanderman R, et al. Leisure-time physical exercise: prevalence, attitudinal correlates, and behavioral correlates among young Europeans from 21 countries. Prev Med 1997;26:845–54. [10] Allison KR, Adlaf EM, Irving HM, Hatch JL, Smith TF, Dwyer JJM, et al. Relationship of vigorous physical activity to psychologic distress among adolescents. J Adolesc Health 2005;37:164–6. [11] Sanchez-Villegas A, Ara I, Guillen-Grima F, Bes-Rastrollo M, Varo-Cenarruzabeitia JJ, Martinez-Gonzalez MA. Physical activity, sedentary index, and mental disorders in the SUN Cohort Study. Med Sci Sports Exerc 2008;40:827–34. [12] Beard JR, Heathcote K, Brooks R, Earnest A, Kelly B. Predictors of mental disorders and their outcome in a community based cohort. Soc Psychiatry Psychiatr Epidemiol 2007;42:623–30. [13] Strohle A, Hofler M, Pfister H, Muller A-G, Hoyer J, Wittchen H-U, et al. Physical activity and prevalence and incidence of mental disorders in adolescents and young adults. Psychol Med 2007;37:1657–66. [14] Brown WJ, Ford JH, Burton NW, Marshall AL, Dobson AJ. Prospective study of physical activity and depressive symptoms in middle-aged women. Am J Prev Med 2005;29:265–72. [15] Motl RW, Birnbaum AS, Kubik MY, Dishman RK. Naturally occurring changes in physical activity are inversely related to depressive symptoms during early adolescence. Psychosom Med 2004;66:336–42. [16] Lee C, Russell A. Effects of physical activity on emotional well-being among older Australian women. Cross-sectional and longitudinal analyses. J Psychosom Res 2003;54:155–60. [17] Strawbridge WJ, Deleger S, Roberts RE, Kaplan GA. Physical activity reduces the risk of subsequent depression for older adults. Am J Epidemiol 2002;156:328–34. [18] Camacho TC, Roberts RE, Lazarus NB, Kaplan GA, Cohen RD. Physical activity and depression: evidence from the Alameda County Study. Am J Epidemiol 1991;134: 220–31. [19] Farmer ME, Locke BZ, Moscicki EK, Dannenberg AL, Larson DB, Radloff LS. Physical activity and depressive symptoms: the NHANES I epidemiologic follow-up study. Am J Epidemiol 1988;128:1340–51. [20] Sagatun A, Sogaard AJ, Bjertness E, Selmer R, Heyerdahl S. The association between weekly hours of physical activity and mental health: a three-year follow-up study of 15–16-year-old students in the city of Oslo, Norway. BMC Public Health 2007;7: 155. [21] Wendel-Vos GCW, Schuit AJ, Tijhuis MAR, Kromhout D. Leisure time physical activity and health-related quality of life: cross-sectional and longitudinal associations. Qual Life Res 2004;13:667–77. [22] Bhui K, Fletcher A. Common mood and anxiety states: gender differences in the protective effect of physical activity. Soc Psychiatry Psychiatr Epidemiol 2000;35: 28–35. [23] Wiles NJ, Jones GT, Haase AM, Lawlor DA, Macfarlane GJ, Lewis G. Physical activity and emotional problems among adolescents. A longitudinal study. Soc Psychiatry Psychiatr Epidemiol 2008;43:765–72. [24] De Moor MHM, Boomsma DI, Stubbe JH, Willemsen G, De Geus EJC. Testing causality in the association between regular exercise and symptoms of anxiety and depression. Arch Gen Psychiatry 2008;65:897–905.

348

M. ten Have et al. / Journal of Psychosomatic Research 71 (2011) 342–348

[25] Kritz-Silverstein D, Barrett-Connor E, Corbeau C. Cross-sectional and prospective study of exercise and depressed mood in the elderly. The Rancho Bernardo Study. Am J Epidemiol 2001;153:596–603. [26] Cooper-Patrick L, Ford DE, Mead LA, Chang PP, Klag MJ. Exercise and depression in midlife: a prospective study. Am J Public Health 1997;87:670–3. [27] Weyerer S. Physical inactivity and depression in the community. Evidence from the Upper Bavarian Field study. Int J Sports Med 1992;13:492–6. [28] Bijl RV, Zessen van G, Ravelli A. The Netherlands Mental Health Survey and Incidence Study (NEMESIS): objectives and design. Soc Psychiatry Psychiatr Epidemiol 1998;33:581–6. [29] De Graaf R, Bijl RV, Smit F, Ravelli A, Vollebergh WAM. Psychiatric and sociodemographic predictors of attrition in a longitudinal study: the Netherlands Mental Health Survey and Incidence Study (NEMESIS). Am J Epidemiol 2000;52:1039–47. [30] Smeets RMW, Dingemans PMAJ. Composite International Diagnostic Interview (CIDI), Version 1.1. Geneva: World Health Organization; 1993. [31] World Health Organization. Composite International Diagnostic Interview (CIDI), Version 1.0. Geneva: World Health Organization; 1990. [32] Wittchen H-U, Robins LN, Cottler LB, Sartorius N, Burke JD, Regier DA, et al. Crosscultural feasibility, reliability and sources of variance in the Composite International Diagnostic Interview (CIDI). Br J Psychiatry 1991;159:645–53. [33] Wittchen H-U. Reliability and validity studies of the WHO-Composite International Diagnostic Interview (CIDI): a critical review. Int J Psychiatr Res 1994;28:57–84. [34] Van 't Land H, Verdurmen J, ten Have M, van Dorsselaer S, Beekman A, de Graaf R. The association between arthritis and psychiatric disorders; results from a longitudinal population-based study. J Psychosom Res 2010;68:187–93.

[35] Baker M, Stabile M, Deri C. What do self-reported, objective measures of health measure? J Hum Resour 2001;39:1067–93. [36] Skinner CJ, Holt D, Smith TMF. Analysis of Complex Surveys. Chichester: Wiley; 1989. [37] Stephens T. Physical activity and mental health in the United States and Canada: evidence from four population surveys. Prev Med 1988;17:35–47. [38] Larun L, Nordheim LV, Ekeland E, Hagen KB, Heian F. Exercise in prevention and treatment of anxiety and depression among children and young people. Cochrane Database Syst Rev 2006, doi:10.1002/14651858.CD004691.pub2 Art. No.: CD004691. [39] Scully D, Kremer J, Meade MM, Graham R, Dudgeon K. Physical exercise and psychological well being: a critical review. Br J Sports Med 1998;32:111–20. [40] Faulkner GEJ, Adlaf EM, Irving HM, Allison KR, Dwyer JJM, Goodman J. The relationship between vigorous physical activity and juvenile delinquency: a mediating role for self-esteem. J Behav Med 2007;30:155–63. [41] Breedveld K, Kamphuis C, Tiessen-Raaphorst A. Rapportage sport 2008 [Report engaging in physical exercise/sport 2008]. Den Haag: Sociaal en Cultureel Planbureau; 2008. [42] Hildebrandt VH, Ooijendijk WTM, Hopman-Rock M. Trendrapport bewegen en gezondheid 2006/2007 [Trend report physical activity and health 2006/2007]. Leiden: TNO Kwaliteit van Leven; 2008. [43] Mead GE, Morley W, Campbell P, Greig CA, McMurdo M, Lawlor DA. Exercise for depression. Cochrane Database Syst Rev 2008, doi:10.1002/14651858.CD004366.pub3 Art. No.: CD004366.