STRENGTHS AND LIMITATIONS OF THIS STUDY
Using ordered categorical outcomes can truly reflect the disease situation and the expose–disease association.
Quantifying physical activity with metabolic equivalent helped accurately assess the health effects of habitual physical activity (HPA).
This large-scale multicentre study helped reduce selection bias and allowed for corresponding subgroup analyses.
It is difficult to establish causal links between HPA and depression and anxiety with the cross-sectional design.
Whether the findings can be applied to the juvenile or senile elderly needs to more evidence.
Globally, it was estimated that 782 million people suffered mental illnesses in 2017, accounting for 10.7% of all those with medical conditions1; the COVID-19 pandemic has exacerbated many of the determinants of poor mental health.2–4 Depression and anxiety are two leading causes of mental health concerns threatening human health,5 and their disease burden is increasing year by year.5–7 Depression and anxiety are also related to premature mortality from suicide and other related disease.8 9 Depression and anxiety will continue to affect individuals, the healthcare system and the whole society if they are left unprevented.
A large number of factors, such as overweight, gender, changes in health status, physical disabilities, chronic pain, childhood health status, lack of social contact, unhealthy lifestyle and negative event stimulation, were reported to influence the occurrence of depression and anxiety.10–12 Increasing evidence is highlighting the benefits of lifestyle modification in chronic diseases and mental health prevention.12–15 Several systematic reviews and meta-analyses have strengthened the effectiveness of physical activity or daily exercise in managing depression or anxiety.16–18 However, from the perspective of aetiological prevention, the association of physical activity with anxiety did not reach a conclusive result; some studies have found that physical activity or exercise can reduce the incidence of anxiety,19 while others reported no association.20 21 Gender was found to play a significant role in influencing anxiety and depression levels, and a lack of exercise might exacerbate the symptoms in both genders when compared with physically active individuals.22 In addition, most studies treat depression or anxiety as a dichotomous variable rather than as an ordered categorical variable; using dichotomous outcome classification that ignores disease severity may exaggerate or reduce associations with exposure. Therefore, the severity of the disease and gender should be considered when conducting relevant research.
Therefore, this study aimed to explore the association between habitual physical activity (HPA) and the risk of depression and anxiety among Chinese adults, with the consideration of the disease severity and gender effect. We hypothesised that HPA might reduce the risk of depression and anxiety.
The participants in this cross-sectional study were from the nationwide survey of the psychology and behaviour of Chinese residents in 2022.23–25 Briefly, a total of 30 505 participants were recruited by using a multistage sampling method from 23 provinces, 5 autonomous regions and 4 municipalities between June 2022 and August 2022. The inclusion criteria of this study were Chinese permanent residents aged between 20 and 75 years old, capable of reading and answering electronic questionnaires, and not quarantined for COVID-19 infection or anything else. Those who had Parkinson’s disease or Alzheimer’s disease, emotional disorder, mobility problems, malignant tumours, stroke, fracture, Chronic obstructive pulmonary disease/asthma, and chronic renal and liver disease were excluded. Those who were pregnant or lactating women were also excluded. Finally, 19 798 participants were included for further analysis. This present study was a secondary analysis of survey data from the nationwide survey of the psychology and behaviour of Chinese residents in 2022.
Patient and public involvement
Sample size estimation
To ensure the statistical power, the sample size was prior estimated. Current studies have reported the effect of physical activity on preventing depression (OR=0.78)26 and anxiety (OR=0.87).19 The prevalence rate of depression and anxiety was about 26.9% and 21.8%.27 Considering the following conditions of alpha being 0.05 and beta being 0.10, about 3600 samples were needed. This present study included 19 798 adult participants, meeting the statistical requirement.
Depression and anxiety assessment
Depression and anxiety were measured using the nine-item Patient Health Questionnaire (PHQ-9) and seven-item Generalised Anxiety Disorder scale (GAD-7), both have been found to have validity and repeatability in the Chinese population.28 29 The PHQ-9 is a 9-item scale, using a 4-point Likert scale ranging from 0 (not at all) to 3 (nearly every day), with a total score between 0 and 27. The levels of depression were classified as follows: normal (0–4), mild (5–9), moderate (10–14) and severe.15–28 The GAD-7 comprised 7 items using a 4-point Likert scale ranging from 0 (not at all) to 3 (nearly every day), with a total score between 0 and 21. The levels of anxiety were classified as follows: normal (0–4), mild (5–9), moderate (10–14) and severe.15–21 29
The HPA was assessed at baseline survey using the International Physical Activity Questionnaire-short form (IPAQ-SF) by using a self-administered approach. The Chinese version of IPAQ-SF was found to have good validation and reliability.30 First, participants were asked to report the frequency of PHA performed in the past week with examples of activities under three intensity categories: light (eg, fast running, ball games, aerobics, fast cycling), moderate (eg, Carrying goods, moderate speed cycling, brisk walking) and vigorous (eg, walking). According to Ainsworth’s compendium of physical activities, a metabolic equivalent (MET, 1 MET=1kcal/hour/kg) value was assigned to each intensity category as follows: 2.5 METs for light, 4.5 METs for moderate and 7.5 METs for vigorous.31 Second, participants were asked to estimate the duration usually spent on HPA every day in the past week. The volume of HPA (MET-hours/week) was calculated by multiplying the intensity (METs) by frequency (day/week) and then by duration (hours/week). According to the guidelines on physical activity by the WHO, to attain substantial health benefits, adults should perform at least 150 min of moderate-intensity or 75 min vigorous-intensity aerobic physical activity per week, or an equivalent combination of both should be performed32; this means that conducting activity with at least 10 MET-hours/week is suggested to reach the minimum level of the recommended standard.32 Considering cut-points of 2.5 MET-hours/week (a quarter of recommended level), 10.0 MET-hours/week (recommended level) and 20.0 MET-hours/week (double the recommended level), HPA was classified into the following categories: inactive (no HPA or HPA<2.5 MET-hours/week), low (HPA 2.5 to <10.0 MET-hours/week), moderate (HPA 10.0 to <20.0 MET-hours/week) and high (HPA≥20.0 MET-hours/week).
A questionnaire was used to collect information on age (years), gender (male, female), marital status (single, married, others), education (primary school or lower, middle school, college or above), smoking (never, ever, current), alcohol drinking (never, ever, current), employment status (student, regular job, freelancer, unemployed, retirement). The well-being of each participant was estimated by using the Chinese version of the 5-item WHO Well-Being Index, which consists of 5 items with a 6-point Likert-type scale ranging from 0 (at no time) to 5 (all of the time).33 In addition, weight, height and chronic disease status were also required to report. Weight and height were measured to calculate body mass index (kg/m2). The subject who self-reported physician-diagnosed hypertension, diabetes or dyslipidaemia was defined as having hypertension, diabetes or dyslipidaemia.
All statistical analyses were performed by using R V.4.0.1 (R Development Core Team, Vienna, Austria); the tests were two tailed, and the p<0.05 was considered statistically significant. Frequency and percentage were used to describe categorical variables, and a χ2 test was performed to compare the distribution difference. For continuous variables, normally distributed variables were described by mean and SD and a t-test or a one-way analysis of variance was performed to examine the difference; meanwhile, non-normal distributed variables are described by median and IQR, and a Wilcoxon rank-sum test or a Kruskal-Walli’s test was performed.
Crude and adjusted ORs with their 95% CIs were estimated by using ordinal logistic regression. The linear exposure–response relationship was examined by putting the median of each tertile of exposure as a continuous variable into the model. Stratified analysis was done by gender and age. The multiplicative interaction of HPA with age and gender was examined respectively by using the likelihood ratio test, with a comparison of the likelihood scores of the two models with and without the interaction terms. To test the robustness of the results, a sensitivity analysis was conducted by excluding participants with diabetes, hypertension and dyslipidaemia.
Of the 19 798 participants, 15 897 (80.30%), 3378 (17.06%) and 523 (2.64%) were classified as normal, mild anxiety and moderate anxiety; meanwhile, 10 884 (54.98%), 7515 (39.75%), 1207 (6.10%) and 192 (0.97%) were classified as normal, mild depression, moderate depression and severe depression (table 1). The prevalence of depression in females is slightly higher than in males (p<0.05), whereas the prevalence of anxiety is comparable between genders. The median (IQR) of HPA was 39.71 (80.23) MET-hours/week for both genders, 46.56 (94.04) MET-hours/week for male and 35.35 (69.49) MET-hours/week for female. A significant difference between genders was observed in most confounders (p<0.05) with the exception of Well-being and social economics status.
The distribution of these covariables by anxiety and depression is shown in online supplemental tables S1 and S2. The median (IQR) of HPA for non-depression, mild, moderate and severe depression was 37.24 (72.26) MET-hours/week, 42.00 (85.75) MET-hours/week, 48.14 (114.25) MET-hours/week and 71.03 (175.7) MET-hours/week, respectively, with a significant distribution difference (p<0.001). The median (IQR) of HPA for non-anxiety, mild and moderate anxiety was 37.71 (73.98), 50.85 (107.62) MET-hours/week and 41.65 (127.91) MET-hours/week, respectively, with a significant distribution difference (p<0.001). A significant difference among different levels of depression and anxiety was observed in most confounders (p<0.05), with the exception of whether to have diseases of diabetes, hypertriton and dyslipidaemia.
As shown in table 2, compared with inactive participants, depression risk in individuals reporting low, median and high volume HPA was reduced by 23% (OR 0.77, 95% CI 0.66 to 0.89), 21% (OR 0.79, 95% CI 0.68 to 0.90) and 13% (OR 0.87, 95% CI 0.78 to 0.98), respectively, after adjustment for confounder; however, no exposure-response trend was observed. Stratified analysis by gender and age yielded similar results. As shown in table 3, no significant association of HPA with anxiety risk was observed in all participants, males and females after adjustment for confounder. The exposure-response trend was also not observed. Multiplicative interactions of HPA with gender or age on the risk of depression and anxiety were not observed (p>0.05). Sensitivity analysis by excluding subjects with a history of diabetes, hypertension and dyslipidaemia yielded similar results (table 4).
This large nationwide cross-sectional study displayed that a higher level of HPA was associated with a lower risk of depression, while no significant association was observed between HPA and the risk of anxiety. To our best knowledge, this study is the first study to consider the severity of depression and anxiety when exploring their association with HPA.
This study found that low, median and high volume HPA were all associated with reduced risk of depression, and stratified and sensitivity analyses found similar results, indicating the robustness of our finding. A systematic review and meta-analysis with 15 cohort studies showed similar protective effects and that the maximum benefit was obtained from those meeting the recommended guidelines by either moderate or vigorous-intensity activities.26 Recently, a cross-sectional study from National Health and Nutrition Examination Survey (NHANES) also revealed the protective effect of physical activity against depression.34 Additionally, several epidemiological studies among the Chinese also revealed a similar beneficial role of physical activity in decreasing depression risk,35–38 while sedentary behaviour was positively associated with depression in a dose-response manner.39 The possible mechanism might be explained by the reaction of the endocannabinoid system to acute neuroendocrine stimulation and inflammation and changes in neural architecture.40 In addition, conducting HPA could increase the frequency of social interactions, strengthen self-efficacy and improve body image, which leads to better mental health41; HPA might also promote appetite and sleep, which in turn promotes rejuvenation and physical recovery and ultimately helps maintain good mental health.42
This study did find no significant association between HPA and anxiety risk. Consistently, several meta-analyses also reported that anxiety risk was not significantly associated with physical activity,20 21 though physical activity was found to be effective for managing anxiety symptoms.17 18 In addition, a study found that gender would influence anxiety levels, and a lack of exercise exacerbates these measures in both genders when compared with physically active individuals.22 However, our study did not find such an effect in different genders or age groups. This distinct might be partly due to the small sample size in Skalidou’s study,22 and partly due to the difference in social environment. The extent of social support, home care and antiepidemic measures for people in different countries may contribute to differences in anxiety risk during epidemics.43 44
The prevalence of depression and anxiety was 45.02% and 19.7% in this study, which was relatively high. However, we found that most people in this current study had mild cases, only a small proportion had moderate to severe depression and had moderate anxiety. These indicated that our results were credible. This was comparable to the results from other studies during the same period when taking into account the impact of the COVID-19 pandemic.22 27 45 46
This study has some advantages. First, the different levels of depression and anxiety were assessed with PHQ-9 and GAD-7. We took into account the severity of depression and anxiety when estimating the effect caused by HPA by using the ordinal logistic regression model. Compared with depression and anxiety as dichotomous outcome variables, ordered categorical outcomes can truly reflect the disease situation and the exposure–disease association. Second, we assessed the overall level of HPA based on MET, which helps to accurately assess the health effects of HPA. Third, this was a large nationwide survey using a multistage sampling approach. The subjects of the study covered nearly all provinces, autonomous regions and municipalities, and covered people with different stages of age. The results, to some extent, could reflect the national profile, help reduce selection bias and allow for corresponding subgroup analyses.
However, the limitations must be considered. First, HPA was assessed by using a validated questionnaire, which might lead to recall bias. Second, it is difficult to establish causal links between HPA and depression and anxiety with the cross-sectional design. Nevertheless, evidence about the effect of PA on depression and anxiety has been reported in several pooled analyses26 and Mendelian randomisation studies,47 indicating that reverse causality was unlikely in this study. Third, this study only recruited participants aged between 20 and 75 years old, whether the findings can be applied to the juvenile or senile elderly needs more evidence. Thus, prospective studies with rigorous design should be conducted to further address these issues.
In conclusion, this nationwide study found that participation in HPA, whether low, median or high volume, was associated with a lower risk of depression compared with inactive participation, while no significant association was observed between HPA and anxiety. Future studies with longitudinal design are needed.
Data availability statement
Data are available on reasonable request. SL was involved in the collection of the data and received written authorisation to use the data for paper publication. The data used to support the findings of this study are available from the corresponding author or SL on request.
Patient consent for publication
This study involves human participants and ethics approval for the nationwide survey of the psychology and behaviour of Chinese residents in 2022 was obtained from the Second Xiangya Hospital, Central South University Clinical Research Ethics Committee (2022-K050). The survey was performed in line with the Declaration of Helsinki and all participants provided informed consent. This present study was a secondary analysis of survey data from the nationwide survey of the psychology and behaviour of Chinese residents in 2022. Participants gave informed consent to participate in the study before taking part.
The authors would like to thank epidemiologists, investigators and participants for their support in the 2022 nationwide survey of the psychology and behaviour of Chinese residents.
This post was originally published on https://bmjopen.bmj.com