Literature search

We reviewed all studies reporting the impact of occupational sedentary behavior on mental health. The search was conducted on the PubMed, Embase, Cochrane, and Psycinfo databases with the following keywords: “sedentary behavior” and (“mental health” or stress or depression or anxiety) and (occupation or work). The search was conducted up to August 2024 (details for the search strategy used within each database are available in S1 File).

Eligibility (inclusion and exclusion) criteria

To be included in the systematic review, studies had to be peer-reviewed and to report both data on occupational sedentary behavior and on mental health in workers. More specifically, articles had to report either sedentary time at work or total sedentary behavior (including sedentary time at work). When a risk (relative risk, hazard ratio, or odds ratio) of mental health issues was described in workers exposed to occupational sedentary behavior, studies were further included in the meta-analysis. We excluded population other than adult workers, in a normal range of age – thus we excluded putative studies on children as well as studies on elderly. We also excluded articles not written in English. Furthermore, reference lists of publications meeting the inclusion criteria were manually searched to identify possible further articles not retrieved from electronic search with our keywords.

Search strategies and selection process

Two authors (Hijrah Nasir and Reza Bagheri) conducted the literature searches, reviewed the abstracts, decided the suitability of the articles for inclusion and extracted the data, and selected the included papers. A third author (Frédéric Dutheil) reviewed the articles where consensus on suitability was debated. Finally, all authors reviewed eligible articles. We followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (S1 Table) [42].

Data extraction

The collected data consist of author’s name, publication year, title, study design, country, objectives and outcomes of included articles, inclusion and exclusion criteria, sample size, main characteristics of individuals (age, sex, education, occupation, BMI, physical activity level, and marital status), level of sedentary exposure (as reported in articles, i.e., higher than a cut-off of hours of sitting per day or per week chosen by authors), characteristics of mental health issues (type, i.e., psychological distress, stress, anxiety, depression, deep sadness, burn out syndrome; and intensity of symptoms, i.e., low, intermediate, severe), and risk of mental health (all risks were reported with the following characteristics: crude or adjusted, variables used for adjustment, and type of risk – hazard ratio, odds ratio, relative risk, coefficient beta). Data extraction was conducted using Excel.

Risk of bias and certainty of evidence

In order to verify the quality of the study, we used the Scottish Intercollegiate Guidelines Network (SIGN) checklist for cohort studies to evaluate the main causes of bias through 14 items with four possible answers (yes, no, can’t say or not applicable) [43]. Furthermore, we used the Newcastle-Ottawa Scale (NOS) to evaluate the risk of bias of included observational studies [44]. The NOS consists of three dimensions – selection, comparability, and outcome – for a total of nine items scored one point each and providing a maximum score of 9 (S2 File) [44]. Scores at the SIGN and NOS checklist were converted into a percentage (one point was given per criteria fulfilled, divided by the sum of applicable criteria). Besides, we also used the Grading of Recommendations Assessment, Development and Evaluation (GRADE) guidelines as the evidence synthesis for the review (S2 Table) [45].

Statistical analysis

The statistical analyses were conducted using STATA® software (v15, StataCorp, College Station, USA). Quantitative variables were described as mean and standard deviation, and qualitative (categorical) variables were presented in the form of number and associated percentage. We computed random effects meta-analyses (DerSimonian and Laird approach) when data could be pooled to calculate the risk of mental health issues in workers exposed to occupational sedentary behavior [46]. We stratified meta-analyses on type of mental health issues, i.e., psychological distress, stress, anxiety and/or depression, deep sadness, and burn out syndrome. We described our results by calculating the hazard ratio, i.e., the risk of mental health issues for workers exposed to occupational sedentary behavior – either sedentary time at work or total sedentary behavior (including sedentary time at work) –, and its associated 95% confidence intervals (95 CI). Hazard ratio were centered at one if the prevalence of mental health issues in workers exposed to occupational sedentary behavior did not differ from the prevalence of mental health issues in workers not exposed to occupational sedentary behavior. Hazard Ratio >1 with the lower limit of 95 CI also >1 denoted an increased risk of mental health issues. Hazard Ratio <1 with the upper limit of 95 CI also <1 reflected a decreased risk of mental health issues for workers with occupational sedentary behavior. We computed a main meta-analysis based on the risk of both intermediate and severe mental health issues, and taking into account all risks mentioned in each study (crude and adjusted). Then, we computed five sensitivity meta-analyses. First, we repeated aforementioned meta-analysis with all risk listed in all studies (crude and all adjusted risks) but we took into account only the highest level of symptoms for each included study (severe mental health issues). Then, to preclude a bias of selection of data, we used a method that we previously published and computed four meta-analyses with the inclusion of only one risk per study (crude/less adjusted or fully/most adjusted, and pessimistic, i.e., the lowest risk or optimistic model, i.e., the highest risk) [47]. Therefore, the six meta-analyses were: all risks of intermediate + severe mental health issues, all risks of severe mental health issues, crude/ pessimistic risks, fully adjusted/ pessimistic risks, crude/ optimistic risks, and fully adjusted/ optimistic risks. We further repeated again those six meta-analyses using only the studies that specifically assessed a risk associated with occupational sitting time (sensitivity analyses). Heterogeneity between studies was evaluated by examining forest plots, 95% confidence intervals (95 CI) and I-squared (I²). I² is a measure of heterogeneity that ranges from 0 to 100%. Heterogeneity is low under 25%, modest between 25 and 50%, and high above 50%. In order to verify the strength of our results, we repeated aforementioned meta-analyses after exclusion of outliers from meta-funnels. Finally, when we get enough data (>5 studies), we computed meta-regressions to compare the risk between each type of mental health issue (psychological distress, stress, anxiety, depression, deep sadness, vs burnout syndrome) and to identify the association between the risk of mental health issue and sociodemographic (age, sex, body mass index – BMI, marital status, education) and lifestyle behavior (physical activity). The results were presented in 95% confidence intervals (95 CI) and corresponding p-values. Statistical significance was marked by p-value ≤0.05.

Risk of bias and certainty of evidence

According to SIGN cohort checklist, the methodological quality of the included studies was good with a mean score of 68.3 ± 8.9%, ranging from 58.3% [5,48] to 78.6% [49]. There was an unclear risk of bias for the assessment of outcomes and exposures, as well as for confusion bias (Fig 2). Using the NOS checklist, the overall quality of studies was 66.7 ± 11.9%, with three studies had a score greater than two third [5,20,49]. All included articles mentioned ethical approval, except one study [50]. According to GRADE checklist, the reliability of evidence was scored high for our main outcome and moderate for most secondary outcomes (S2 Table).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Risk of bias summary using the Scottish Intercollegiate Guidelines Network (SIGN) checklist.

We used the Scottish Intercollegiate Guidelines Network (SIGN) grid for cohort studies that consist of 14 items, evaluating the main causes of bias through 4 possible answers (yes, no, can’t say or not applicable).

https://doi.org/10.1371/journal.pone.0328678.g002

Aims and outcomes of included studies

The main objective of all included studies was to examine the associations between occupational sitting and mental health in employed adults, independent of leisure-time physical activity, except one study that had for main objective to examine the associations between leisure and occupational sitting and mental health [12].

Study designs

Four studies were cross-sectional [12,21,48,50], one was a prospective cohort [49], one study used both a cross sectional and prospective design [5], and one was retrospective [20]. These studies were located in Asia (Japan [49], and India [21]), Australia [50], Europe (Sweden [12] and UK [5]), and South America (Brazil [48] and Mexico [20]).

Recruitment, inclusion, and exclusion criteria

All studies recruited workers from targeted population. Four studies recruited workers from public sectors: government agency [5,50] and military police officers [48] and health-care workers [20]; and three studies recruited workers from private companies: healthcare services [12], remote working information and technology [21], and manufacturing and transportation company [49]. Age was an inclusion criteria in five studies: > 18 years [20], > 20 years [49], age 23–35 years [21], age 21–56 years [48], age 35–55 years [5]. Some studies added the following inclusion criteria: comprehension of Japanese [49], having >1 year of experience in the IT industry and a similar designation in the company for the past 1 year, with a Facebook account and at least login once in 2 days [21]. The two other studies did not clearly define the inclusion criteria [48,50]. Only three studies reported exclusion criteria [21,48,49]. Two of them excluded workers with major depressive episode or being officially absent due to mental health problems in the last 12 months [49], and having an alteration of mental or psychological health [21]. One study also excluded workers with a chronic disease or taking drugs, pregnant or who delivered in the past month, and those who planned to shift the job in the next two months [21]. One study excluded workers who took medical leave [48].

Population

Sample size ranged from 80 [21] to 23,644 [12]. A total 40,314 individuals were included in the study.

Age was on average 35 years old (95 CI 24.0 to 46.1 years old), ranging from 25.7 [21] to 47 years old [50] using the mean age reported in four studies [12,20,21,50]. Two studies reported age category [48,49], one study reported mean age for men and women [50], and one study reported a range (min-max) [5].

Gender was reported in all studies. The mean proportion of men was 65.5% (95 CI 55.0 to 76.0) ranging from 38.2 [20] to 83.5% [48].

Education was reported in four studies [12,21,49,50]. The mean proportion of people with education higher than high school were 67.0% (95 CI 30.0 to 100.4%), ranging from 52.0% [50] to 100% [21].

BMI was reported in five studies [12,20,21,48,50] using category: overweight (25–30 kg/m²) and obese (>30 kg/m²) [12,21,50], all workers with a BMI > 25 kg/m² [48] or >40 kg/m² [20]. Including the three studies with the same cut-off [12,21,50], the percentage of obese people (>30 kg/m²) was 23% (14–33%), ranging from 6 [20] to 65% [48]. Two studies also reported the average BMI [21,50].

Marital status was reported in four studies [20,21,49,50]. Proportion of married participants was 56% (34–78%), ranging from 19 [21] to 81% [50].

Physical activity was auto-reported in four studies [12,21,49,50] through questionnaires. Three studies reported prevalence, i.e., percentage of workers meeting daily physical activity recommendations [21], doing never/ sometimes/ one-two/ 3–5/ or ≥6 times per week [12], or doing low, moderate, or high intensity of physical activity [49]. One study reported the average leisure time physical activity [50]. However, in our analysis, we did not take physical activity as an influencing variable due to insufficient data.

Other descriptive variables were seldomly reported precluding further analyses, i.e., middle or upper class [48], occupation [20], permanent or temporary contract [20], full- or part-time [50], seniority [48], shiftwork [20], working hours [5,21,48,49], telework [21], psychosocial risk factors [49,50], smoking [12,49], alcohol [49], personal medical history (such as pulmonary disease, hypertension, diabetes) [20], use of pain medication [12], and sleep duration [21].

Instruments used to assess sedentary behavior at work

Four studies collected data on sedentary behavior using The International Physical Activity Questionnaire-Short Form (IPAQ) [20,48,50], two studies used Workers’ Living Activity Time Questionnaire (WLAT-Q) [21,49], one study used Self-Administered Physical Activity Questionnaire [5], and one study used Swedish Health Profile Assessment (HPA) survey [5]. All studies used different cut-off for sedentary behavior at work. One study defined “high sedentary” if sitting >8 h/day [33]. Two studies defined “high sedentary” if sitting >6 h/day [20,21]. One study categorized sedentary behavior into low (<3 h/day), moderate (3–6 h/day) and high sitting (>6 h/day) [50]. One study defined “high sedentary” if sitting >9.5 h/day [49]. One study categorized total sitting into “low” (<33 h/week), “medium” (33–50 h/week) and “high” sitting (51–135 h/week) [5]. One study used the percentage of sitting that were categorized into <25%, 25–49%, 50–74%, ≥ 75% of time, and almost always sedentary [5]. One study only defined sedentary behavior as insufficiently active or sedentary [48]. In our meta-analysis, when a study reported more than two categories of sedentary behavior at work, we used the risks corresponding to the highest level of sedentary behavior at work.

Instruments used to assess mental health

Mental health data was evaluated using questionnaires. Two studies [20,21] used the Depression Anxiety Stress Scale - 21 (DASS21). One study [5] used the General Health Questionnaire (GHQ-30), the Short Form-36 Mental Component Summary (SF-36 MCS), as well as the Center for Epidemiologic Studies Depression Scale (CES-D). One study [50] used the Kessler Psychological Distress scale (K10). One study used a simple question: ‘I experience worry, depressed mood or anxiety...’ with five possibilities of answer (very often, often, sometimes, rarely, and never) [12]. One study [48] used the Perceived stress scale (PSS-10), the Youth Risk Behavior Surveillance System (YRBSS) for suicidal ideation, the Hospital and Anxiety Depression Scale (HADS), as well as the Maslach Burnout Inventory (MBI). One study used the Composite International Diagnostic Interview version 3.0 (CIDI 3.0) [49]. Among all those assessments, included studies only calculated a risk for psychological distress [5,50], depression [12,20,21,49], anxiety [12,20,21], stress [20,21], deep sadness [48], and burnout [48].

Narrative synthesis on studies included only in the systematic review

Sakakibara et al. [51] demonstrated an association between higher occupational sedentary behavior and poorer mental health in blue and white-collars, but there were no risk nor data to calculate a risk for mental health issues. Hanna et al. found that the risk of “sitting too much” was increased in workers with depressed mood [52]. Similarly, Sung et al. showed that the risk of sedentary behavior increased in workers with high stress compared to those with low stress [53]. But both studies did not assess the risk in an opposite way, i.e., the risk of mental health issues in sedentary workers compared to less-sedentary workers. Muniswamy et al. showed the positive effects of a social media-based intervention on workers, by reducing their occupational sedentary behavior and their level of stress and anxiety [54]. Xu et al. showed that shift workers had a higher risk of depression and anxiety compared to non-shift workers by taking into account sedentary time [55]. Those studies did not assess the risk of occupational sedentary behavior on mental health.

Meta-analysis on the risk of mental health

Using the global model with all risks both for intermediate and severe levels of mental health issues, we found that occupational (only occupational in 4 studies and total, i.e., occupational and leisure in 3 studies) sedentary behaviour was associated with an increased the risk of mental health issues by 34% (OR=1.34, 95 CI 1.18 to 1.49). Stratification by type of mental health issues showed an increased risks for all mental health issues that were reported in several studies. Due to sedentary behavior, the risk increased by 11% (1.11, 1.00 to 1.21) for psychological distress, 44% (1.44, 1.15 to 1.72) for depression, 83% (1.83, 1.20 to 2.46) for anxiety, 43% (1.43, 1.31 to 1.54) for depression and anxiety, and 64% (1.64, 1.20 to 2.08) for stress. Mental health issues reported in only study were not significant, i.e., for deep sadness (1.85, 0.70 to 3.00) despite being close to significance for burnout (2.49, 0.99 to 4.00) (Figs 3 and S1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Summary of meta-analyses on the risk of mental health issues due to occupational sedentary behavior (only occupational in 4 studies and total, i.e., occupational and leisure in 3 studies).

Blue lozenges represent the overall risk of mental health after exposure to occupational sedentary behavior. Grey lozenges represent risk for each type of mental health issue after exposure to occupational sedentary behavior. The length of each horizontal line around the blue/grey lozenges represents their 95% confidence interval (95 CI). The black solid vertical line represents the null risk for mental health (with a value of 1). When horizontal lines cross the null vertical line, the risk for mental health is not significant. I-squared (%): percentage of heterogeneity between studies for each meta-analysis; Weight (%): Weight of each study for each meta-analysis.

https://doi.org/10.1371/journal.pone.0328678.g003

Sensitivity analyses by type of risks

All sensitivity analyses on severe mental health issues were also significant, whatever the model: + 35% (12–58%) using all risks, 39% (15–63%) using fully adjusted pessimistic model, + 36% (13–59%) using crude or less adjusted pessimistic model, 85% (27–143%) using fully adjusted optimistic model, + 85% (26–143%) using crude or less adjusted optimistic model. Sensitivity analyses stratified by type of mental health issues also showed similar results, with the same types of mental health issues being significant in all models, i.e., depression, anxiety, depression and anxiety, and stress – except in the pessimistic models where depression, anxiety, and stress were no longer or less significant (Figs 3 and S1). The overall risks of mental health issue are marked with a high heterogeneity in all aforementioned meta-analysis, ranging from 71 to 97% (Figs 3 and S1). The high heterogeneity is also shown by the numerous studies that were outside the base of the meta-funnel, hindering further sensitivity analysis (S2 Fig).

Sensitivity analyses using only occupational sedentary behavior

Repetition of all aforementioned meta-analyses with only the four studies that specifically assessed a risk associated with occupational sitting time, all results remained significant. Occupational sedentary behavior was associated with an increased risk of mental health issues using the global model with all risks both for intermediate and severe levels of mental health issues (+17% of risk; OR=1.17, 95 CI 1.07 to 1.26), as well as for meta-analyses on severe mental health issues: 46% (27–66%) using fully adjusted pessimistic model, + 43% (26–60%) using crude or less adjusted pessimistic model, 46% (27–66%) using fully adjusted optimistic model, + 43% (26–60%) using crude or less adjusted optimistic model, but not anymore significant for model taking all risk of severe mental health: + 12% (−10% to +30%) (Fig 4).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Summary of meta-analyses on the risk of mental health issues due to occupational sedentary behavior, using only the four studies that specifically assessed a risk associated with occupational sitting time.

Blue lozenges represent the overall risk of mental health after exposure to occupational sedentary behavior. Grey lozenges represent risk for each type of mental health issue after exposure to occupational sedentary behavior. The length of each horizontal line around the blue/grey lozenges represents their 95% confidence interval (95 CI). The black solid vertical line represents the null risk for mental health (with a value of 1). When horizontal lines cross the null vertical line, the risk for mental health is not significant. I-squared (%): percentage of heterogeneity between studies for each meta-analysis; Weight (%): Weight of each study for each meta-analysis.

https://doi.org/10.1371/journal.pone.0328678.g004

Meta-regressions, i.e., putative influencing factors

We only computed meta-regressions on meta-analyses with the seven studies (insufficient data otherwise). Considering the type of mental health issues, the risk is higher for anxiety in the crude optimistic model (versus psychological distress and depression), and the risk is higher for anxiety than psychological distress in the two models with all risks.

In the pessimistic models, when exposed to occupational sedentary behavior, older age increased the risk of severe mental health issues by 25% (95 CI 5–44%) per 10-year for fully adjusted odds ratio and by 27% (12–43%) per 10-year for crude odds ratio. On the opposite, higher level of education decreased the risk of severe mental health issues due to sedentary behavior at work: −8% (−13 to −2%) for fully adjusted odds ratio and −8% (−13 to −3%) for crude odds ratio, per 10%-high-education (Fig 5). Those results were not significant in all other models. Sex, BMI, and marital status did not influence our results.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Meta-regressions, i.e., variables influencing the risk of mental health issues, depending on groups: 1) risk of mental health (both intermediate and severe), 2) risk of severe mental health (all risks), 3) risk of severe mental health pessimistic model (crude and fully adjusted), 4) risk of severe mental health optimistic model (crude and fully adjusted).

The effect of each variable on the risk of mental health issues is represented by a dot on a horizontal line in the forest-plot. The dots represent the coefficient for each variable, and the length of each line around the dots represent their 95% confidence interval (95 CI). The black solid vertical line represents the null estimate (with a value of 0). Horizontal lines that cross the null vertical line represent non-significant variables on the risk of mental health issues.

https://doi.org/10.1371/journal.pone.0328678.g005

Occupational sedentary behavior and mental health risk

Our main finding was that occupational sedentary behavior was associated with an increased risk of mental health, consistently in all models. Study design did not significantly influence the results. All cross-sectional studies included in our meta-analysis showed an association of prolonged sitting time at work with mental health. Furthermore, two out of three longitudinal studies included showed an association of prolonged sitting time at work with mental health [20,34]. Result remains consistent even after adjusting with the level of physical activity, job stressors, and other covariates [49]. However, other studies reported contradictory results. Hagger-Johnson found that mental disorder symptoms were driven by sitting time at work in a cross sectional, but not in longitudinal analysis [5]. They found prior symptoms of mental disorder were linked to longer sitting in leisure time. Other study also did not find the association between sedentary time and depressive symptoms, anxiety or cognitive performance after a follow-up period of 5.7 years [56]. Additionally, a 13-year cohort study demonstrated differential effects between passive (e.g., watching TV) and active (e.g., reading, office work) sedentary behaviors on major depressive disorder indicating that mentally active sedentary behaviors were protective factor. Interestingly, these effects are largely independent of physical activity levels [57]. A possible explanation may be that mentally active tasks, such as working and sitting in a meeting that happened in the working environment is usually linked to self-reliance, achievement, and sense of autonomy, as well as active participation and engagement in working environment [58]. On the other side, mentally passive sedentary behavior, such as TV-viewing was associated with minimal cognitive effort [57]. Similar to the “physical activity paradox” the fact that leisure time physical activity is beneficial for health while occupational physical activity can be detrimental [59], further studies should also compare the effect of sedentary time at work with the effects of leisure sedentary time on mental health. Sedentary time at work might be less detrimental than leisure sedentary time because it is related to mentally active tasks – while leisure sedentary time is passive tasks (watching TV, etc.). We can therefore have a sedentary paradox (occupational sedentary time less bad) opposite to the physical activity paradox (occupational physical activity worst). In the literature, no studies concomitantly assessed all types of mental health issues that could be linked with occupational sedentary behavior, but they rather focused on a dedicated pathology. Our meta-analyses included psychological distress, depression, anxiety, stress, deep sadness, and burnout, giving a global overview, but lack of comparisons between pathologies. Further studies should deeply explore the effect the of occupational sedentary behavior on each type of mental health.

Occupational factors and mental health risk

Mental health issues associated with occupation have been well-defined in many studies. In our meta-analysis, we identified various types of occupations, including police officers, healthcare workers, caregivers, corporate employees, and civil servants. However, insufficient data precluded comparisons between occupational groups. Nevertheless, a study found that white-collar workers who spent more than 3 h/day had a higher risk of high perceived stress, while no impact was found on blue-collar workers [25]. This result was consistent with a study included in our review revealing the association between long sedentary behavior and poor mental health in white-collar workers, but this association was not observed in blue-collar workers [51]. It can be explained by the fact that white-collar workers usually have higher psychological demands due to their high position that correlated with self-perceived work stress [60]. Besides, prolonged sitting time was also associated with lower work engagement among white collar-workers [51]. This finding may also be due to lower sedentary levels in blue-collars compared to white-collars, that eventually did not affect their mental health [51]. Interestingly, we also included one study focusing on remote workers [21]. A recent study in Japan published that individuals who worked in office spent higher energy expenditure and higher time sitting compared to those who work from home [61]. A randomized controlled trial aiming to intervene physical and mental well-being of the remote workers in India found that 95% of our participants did not meet the PA recommendation and 60% of them experienced mild to moderate anxiety, stress, and depression [21]. The intervention was only able to have low to moderate effect in their PA and sedentary level as well as a slight increase on stress, anxiety, and depression score. Thus, the effect of working at home on mental health are disputable as demonstrated in a systematic review, with some positive effect on stress due to family and work integration and less commuting, but also some negative effect on stress because of unclear work-home boundaries, social isolation, extended working hours, poor organizational support, as well as conflicts of work-family [62]. Seniority was also a protective factor against anxiety for workers as shown in a study [48]. Shift workers have a higher risk of depression and anxiety, that could be mediated by modifiable factors such as sedentary time, smoking, sleep duration, and BMI [55].

Sociodemographics and mental health risk

Despite inconsistent results in our meta-analysis, age may increase the risk of mental health issue for workers with high sedentary behavior at work. In a study included in our meta-analysis, older police officers had a twice greater risk of deep sadness possibly due to the cumulative effect of all stressors, violence, and psychological traumatisms experienced during their careers [48]. In contrary, a study reported that younger age was associated with greater psychological distress alongside with higher occupational sedentary and less leisure time physical activity [50]. The fact that detrimental effect of sedentary behavior on mental health may vary depending on age should be further studied. In occupational context, the relation between age and increased risk of mental health issues could be explained by a sense of job insecurity, job strain, and low job satisfaction, although occupational sedentary behavior was not assessed [63]. Despite the absence of effect in our study, sex could also be an influencing factor on the risk of mental health due to occupational sedentary behavior. Some studies showed that women had higher levels of stress, anxiety, and depression compared to men [20,21], possibly linked with shared responsibility inequities in parenting and household tasks [64]. A study in police officers also showed sex-specific response to traumatic situations and stressful events with higher self-blame and denials in women [48,65]. Additionally, our result showed that higher education was a protective factor reducing the risk of mental health when exposed to sedentary behavior at work. A study assessing sitting time in older office workers showed that lower education levels were significantly related to higher perceived stress and more depressive symptoms [63] suggesting that psychological health is improved by education [66]. In the literature, while lower education was linked with prolonged sedentary behavior [67,68], workers with higher education were more likely to perceive their work as more stressful [25]. Yet no effect found in our study, higher BMI was associated with depression and anxiety through several psychological processes such as weight discrimination [69,70], body dissatisfaction [71], emotional eating, poor emotion regulation, and low exercise confidence [72]. Although no effects were found in our study, the relation between stress and marital status was always conflicting, where married individuals or couples being associated with lower [22,23] or greater psychological distress [50]. A modulator of this relation could be sex, marriage being a protective health factor mostly for men [73,74]. Even if insufficient data precluded to search for an influence of physical activity on mental health, low levels of physical activity were frequently associated in the literature with poor mental health [5,12,50] and lower productivity [21]. Interestingly, a meta-analysis found that physical activity at work was positively associated with mental health as well as mental ill-health [75]. Physical activity intervention generally improve quality of life, mood, and functional capacity [76]. In total, there are numerous articles that assessed the influence of sociodemographic on mental health or on sedentary behavior (in leisure time mostly and some at work), but there is a need for studies assessing the influence of sociodemographic on the impact of sedentary behavior on mental health, with a particular attention for distinguishing leisure time and occupational sedentary behavior, that may yield to different effect similar to the physical activity paradox [59].

Strength and limitations

To the best of our knowledge, this is the first meta-analysis to quantitatively examine the association between occupational sedentary behavior and several mental health issues. We used a statistical method by computing four meta-analyses with the inclusion of only one risk per study (crude/less adjusted or fully/most adjusted, and pessimistic, i.e., the lowest risk or optimistic model, i.e., the highest risk) to reduce a bias of data selection. However, this study has several limitations. The number of included studies was small; however, we included a total sample size of more than 40,000 workers. Our meta-analysis inherits the bias from each included study, particularly bias of exposure and bias of outcome assessment. Studies included sitting time and mental health issues used mainly self-reported questionnaires, which may limit the quality of data. On the other hand, it also makes the evaluation reproducible. Most studies used self-questionnaire without supervision leading to bias such as incomplete information, non-disclosure or skipping questions due to social desirability bias or lack of self-awareness [77], that is frequently found in private or sensitive questions such as mental health condition [78]. The questionnaire used also varied between studies, making comparisons between studies more difficult. Additionally, most included studies were cross-sectional precluding to establish a cause-effect relation. Lack of longitudinal data did not permit to study long-term implications of sedentary behavior in the follow-up of mental health. Our population may seem heterogeneous but it is also in favor of the generalizability of our results. Moreover, we cover four continents except Africa, as it is unfortunately often the case in research [79]. Despite included studies used various cut-off both for sedentary behavior, we computed several sensitivity analyses using different cut-off. We conducted sensitivity analyses that yielded consistent results across various study designs. In our comprehensive meta-analysis model, some studies appeared multiple times with different data for the same outcome based on different adjustment models. Consequently, due to the lack of a clear consensus and despite the practice of including more effect sizes within each meta-analysis than the number of constituent studies, as seen in several articles published in high-ranked journals [80,81], the weighting of studies might have required adjustment [82]. This necessity supports the use of optimistic and pessimistic models, where each included article is represented by only one set of data. Further study can also involve device-based measures of sedentary behavior to determine the sitting time more accurately [83] and adjustments with physical activity and sleep levels. Recent studies have shown that physical activity, sedentary behavior and sleep are co-dependent, moreover in case of poor mental health, it should be considered together [84]. Despite these limitations, our meta-analysis is noteworthy to understand the effect of occupational sedentary behavior that may increase the risk mental health indicating the urgent need for mitigation at work.