Background
Method
Search strategy
Inclusion/exclusion criteria
Data extraction and quality assessment procedures
Study | Country | Study design | Population | N (% of females) | Mean age (SD) | SB measure/indicator | Anxiety measure | Positive, negative or non-significant elation-ship between anxiety and the type of SB | Main effects/findings | Quality score (%) |
---|---|---|---|---|---|---|---|---|---|---|
Asfour et al., 2016 [59] | USA | Cross-sectional | Adolescents; (general population) | 575 (45%) | 13.8 ± 0.64 years | Time spent engaging in SB was calculated using five questions that inquired about time spent watching TV, playing video games, text messaging, Internet use and time spent on the telephone | Internalizing symptoms subscale of 112-item The Youth Self-Report [60] | 0 for ST | Increased SB were associated with higher levels of externalizing symptoms (β = 2.03, SE = 0.32, p < .001), but not internalizing symptoms (β = 0.93, SE = 0.57, p = .1) | 98 |
Asztalos et al., 2015 [61]a | Belgium | Cross-sectional | Adults (general population) | 4344 (52%) | 43.55 ± 11.05 years | Self-report - International Physical Activity Questionnaire (IPAQ) [62] | 10-items for anxiety symptoms of Symptom Check List (SCL)(e.g. [63]) | + for sitting time | Sitting were positively associated with anxiety both in model adjusted for demographic (gender, age, education) (β = 0.037; p = .018) as well as adjusted for demographics and MVPA (β = 0.033; p = .038) | 95 |
Bampton et al., 2015 [64]a | Canada | Cross-sectional | Older adults (≥ 55 yr) (general population) | 358 (66%) | 66.5 ± 8.0 years | Total and Domain Specific Measure of Sitting [5] | Generalized Anxiety Disorder scale (GAD-2) [65] | + for sitting time (+ for high sedentary time/low resistance group) | Low sedentary time/high RT group reported lower anxiety symptoms compared to high sedentary time/low RT group (Mdiff = − 0.67, p = .001). The low sedentary time/low RT group reported lower anxiety symptoms than high sedentary time/low RT group (Mdiff = − 0.57, p = .018) | 81 |
Cao et al., 2011 [66] | China | Cross-sectional | Adolescents (general population) | 5003 (48%) | 13.13 ± 0.97 years | Self-report to an open-ended question: how many hours per day, on average, the participants spent on the sedentary activities outside school hours on a usual weekday, as well as a weekend day (TV viewing, computer usage) [6]. ST was categorized as ≤2 h/d and > 2 h/d | 41-item Screen for Child Anxiety Related Emotional Disorders (SCARED) [67] | + for ST | High screen time (> 2 h/day) was a risk factor for anxiety symptoms both in crude model (OR = 1.39, 95% CI: [1.22, 1.60], p < .001) as well as in adjusted for gender, grade, family type, perceived socioeconomic status, BMI, fruit and vegetable or fizzy drinks intake (OR = 1.36, 95% CI: [1.18, 1.57], p < .001) | 95 |
de Wit et al., 2011 [68] | Netherlands | Cross-sectional | Adults (general population) | 2353 (65.45%) | 41.2 ± 13.0 years | Self-report - the daily number of hours a person spent on watching TV or computer use in leisure time | Composite International Diagnostic Interview (CIDI, WHO ver. 2.1) [69] | + for TV viewing; 0 for computer use | Anxiety was related to spending more time watching TV (β = 0.051, p < .05; β = 0.103, p < .001, respectively) but not to time of computer use (β = 0.001, p = .971; β = 0.036, p = .155, respectively) in leisure time | 82 |
de Wit et al., 2015 [70] | Netherlands, Austria, Belgium, Ireland, Italy, Poland, Spain, UK, Denmark | Cross-sectional | Adult pregnant women (general population) | 98 (100%) | 31.6 ± 5.8 years | Actigraph GT3X, GT1M or Actitrainer accelerometer [activity were calculated as time spent sedentary (< 100 cpm)] | Pregnancy-related worries were measured with the 13-item Cambridge Worry Scale (CWS) [71] | 0 for ASB | Pregnancy-related worries were not significantly associated with sedentary behaviors | 100 |
Dillon et al., 2018 [72] | Ireland | Cross-sectional | Adults (general population) | 397 (54%) | 59.6 ± 5.5 | A triaxial, GENEActiv accelerometer (ActivInsights Ltd., Kimbolton, Cambridgeshire, UK) | Anxiety subscale of the 14-item Hospital Anxiety and Depression Scale [73] | + for ASB | Participants with moderate to severe anxiety symptoms had significantly more minutes of SB daily than those with low levels/ no symptoms of anxiety (p = .04) | 100 |
Edwards & Loprinzi, 2016 [41] | USA | RCT | Adults (general population) | 39 (59%) | Control: 22.08 ± 2.75; Intervention: 21.69 ± 2.71 | Overall Anxiety Severity and Impairment Scale (OASIS) [75] | + for ASB | A statistically significant time x group interaction effect for OASIS scores F(1,37) = 11.13, p = .002). Mean and SE OASIS scores were significantly higher after the one-week sedentary behaviors-inducing intervention (M = 5.35, SE = 0 .86) compared to scores from before the intervention (M = 3.88, SE = 0.69) which means that an increase was observed in anxiety levels when participants increased their SB | 89 | |
Feng et al., 2014 [76]a | China | Cross-sectional | Young adults (general population) | 1106 (43%) | 18.90 ± 0.9 years | Self-report ST measured with one item: ‘How many hours per day do you spend on computer, including internet use, watching TV/video programs and playing games on a usual weekday and weekend day, respectively?’ The ST was categorized as ≤2 h/d and > 2 h/d | Self-rating anxiety scale (SAS) [77] | 0 for ST | No statistically significant associations were found between ST and anxiety both for ≤2 h/day (OR = 1.52, 95% CI: [0.87, 2.64], p > .05) as well as for > 2 h/day ST levels | 93 |
Gaskin et al., 2016 [78] | Australia | Cross-sectional | Prostate cancer survivors (chronic illness) | 98 with complete data; 49 with no complete data | 67.3 ± 8.0 with complete data; 62.1 ± 8.6 with no complete data | Hip-mounted ActiGraph GT1 M accelerometer (Pensacola, FL) units | Memorial Anxiety Scale for Prostate Cancer (MAX-PC) – three subcales (prostate cancer anxiety, prostate-specific antigen anxiety, fear of recurrence) and a total anxiety scale [79] | 0 for ASB | Prostate cancer anxiety (B = 0.01, 95% CI: [− 0.03, 0.04], p = .78), prostate-specific antigen anxiety (B = 0.00, 95% CI: [− 0.00, 0.00], p = .96), fear of recurrence (B = − 0.01, 95% CI: [− 0.02, 0.01], p = .44) and a total anxiety (B = − 0.00, 95% CI: [− 0.05, 0.04], p = .94) were not significantly associated with SB | 91 |
Gibson et al., 2017 [80]a | United Kingdom | Cross-sectional | Adults (general population) | 42 (55%) | 38.0 ± 11.5 | ActivPAL mini, an inclinometer-based activity monitor | Anxiety subscale of The 14-item Hospital Anxiety and Depression Scale [73] | + for ASB + for sitting time | Those with < 8 h of SB per day on weekdays had significantly lower levels of anxiety compared with those who were sitting > 8 h or > 10 h per day. The main effect for weekday sitting time on anxiety (F(1, 41) = 3.05, p = .040, η2 = 0.18) | 77 |
Gunnell et al., 2016 [81]a | Canada | Longitudinal | Adolescents (general population) | 1160 (61%) - Time 1 | 13.54 ± 1.12 years | Self-report questionnaire - 6-items querying how many hours per day subjects typically engaged in TV viewing/video game playing/computer use). The first 3 items assess ST during weekdays and the last 3 items inquired about weekend days | 10-items Multidimensional Anxiety Scale for Children-10 (MASC-10) [82] | 0 for ST | Initial symptoms of anxiety and ST did not predict change in ST and anxiety, respectively | 95 |
Hiles et al., 2017 [83] | Netherlands | Longitudinal | Adults (general population) | 2932 (66%) at baseline | 41.9 ± 13.1 | Self-report single question – sedentary behavior as an average hours of sitting on a weekday. | 21-item Beck Anxiety Inventory (BAI; [84]) | 0 for anxiety- > sitting time change | Anxiety at the baseline did not predict SB at a 2-year follow up (B = 0.02, p = .561) | 100 |
Janney et al., 2013 [85]a | USA | RCT | Adults with a diagnosis of schizophrenia/schizoaffective disorder with BMI > 27 (chronic illness) | 46 (63%) | 45.6 ± 9.8 years | ActiGraph AM-7164 accelerometer (ActiGraph, Ft. Walton Beach, FL). Sedentary was established as ≤100 cpm | PANSS (one item for anxiety) [86] | 0 for ASB | No association was observed between SB and PANSS psychiatric symptoms (PANSS or subscale: p ≥ .59). There were no significant associations for the PANSS questions asking about anxiety (rs = .22, p = .15 for sedentary minutes and rs = .15, p = .32 for percentage of sedentary time) | 89 |
Kovess-Mastefy et al., 2015 [87] | German, Netherlands, Lithuania, Romania, Bulgaria, Turkey | Cross-sectional | Schoolchildren (general population) | 3184 | 8.72 years | Parents were asked how long their child spends playing video games on average during the week. Low video game use was defined as 0–60 min per week; moderate use was defined as 61–300 min, and high use was > 300 min. | GAD indexes of self-reported mental health computerized cartoon-like assessment tool ‘Dominic Interactive’ for children [88] | 0 for video game playing | Playing video games (1–5, and 5+ vs. 1 or less h) was not associated with GAD (OR = 1.08, 95% CI: [0.69, 1.7]; OR = 0.95, 95% CI: [0.53, 1.69], p > .05, respectively) | 91 |
Kroeders et al., 2013 [89]a | Australia | Cross-sectional | Stroke patients (chronic illness) | 19 (53%) | 66.2 ± 19.3 years | PAL2 electronic device - dual axis accelerometer | Anxiety subscale from Irritability, Depression and Anxiety (IDA) scale [90] | - for sitting time | Patients with anxiety symptoms compared with those without symptoms tended to spend more time lying (mean 64% vs. 43%), less time sitting (mean 33% vs. 51%), and less time standing or walking (mean 2% vs. 6%). The difference between these groups in time spent lying bordered on significance (t(17) = − 2.0; p = .06) | 76 |
Liu et al., 2016 [91]a | China | Cross-sectional | Secondary school pupils/adolescents (general population) | 13,659 (49%) | 15.18 ± 1.89 years | The Youth Risk Behavior Survey (YRBS) questionnaire [92] ‘How many hours do you watch television or play VG/CU (including activities such as Nintendo, Game box, Xbox, computer games, and the internet) on a typical school day?’ The ST was categorized as: non-ST (0 h/day), occasional ST (> 2 h/day), moderate ST (> 1 to ≤2 h/day), high ST (> 2 h/day) | + for TV viewing; + for VG/CU time | More than 2 h per school day of TV watching was associated with higher risk of anxiety in boys (OR = 1.43, 95% CI: [1.05, 1.95]) compared with no TV exposure. School day with high VG/CU time (> 2 h) was associated with higher risks of anxiety in boys (OR = 1.40, 95% CI: [1.061.86]) compared with no VG/CU | 98 | |
Maras et al., 2015 [94]a | Canada | Cross-sectional | Adolescents (general population) | 2482 (58%) | 14.10 ± 1.57 years | The Leisure-Time Sedentary Activities 6-item questionnaire measured how many hours per day subjects typically engage in: TV viewing/video game playing, computer use). The first 3 items assess screen-based activities during a typical weekday and the last 3 items screen time accrued on a typical weekend day. | The Multidimensional Anxiety Scale for Children-10 (MASC-10) | + for ST (hours per day of TV viewing; recreational computer use; video games); + for video game playing; 0 for TV viewing; 0 for computer | Duration of screen time (β = 0.07, p < .001) and VG playing (β = 0.11, p < .001) was associated with severity of anxiety. | 93 |
Mesquita et al., 2017 [95] | Netherlands | Prospective observational study | Adults with COPD before/after pulmonary rehabilitation (chronic illness) | 90 (40%) | 67.0 ± 8.0 | CIRO activity monitor (CAM or the MOX Activity Monitor (Maastricht Instruments B.V in Maastricht, the Netherlands) | Anxiety subscale of the 14-item Hospital Anxiety and Depression Scale [73] | 0 for anxiety- > ASB change | Baseline anxiety levels were unrelated to changes in minutes of SB (pre-post rehabilitation, Spearman R = −.08 | 95 |
Opdenacker, & Boen, 2008 [96]a | Belgium | Longitudinal | Adults (general population) | 66 | 2 groups, aged M = 38.8 ± 11.4 years; and 39.9 ± 9.9 years | Self-report - International Physical Activity Questionnaire (IPAQ) [62] | The Spielberger state-trait anxiety inventory (STAI) [21] | 0 for sitting time | Sitting time was not associated with anxiety (r = 0.46, p = .623). | 85 |
Padmapriya et al., 2016 [97]a | Singapore | Cross-sectional | Pregnant women (general population) | 257 with state anxiety symptoms; 270 with trait anxiety symptoms | 29.5 ± 5.7 with state anxiety symptoms; 29.5 ± 5.6 with trait anxiety symptoms | Two categories of self-reported total sitting time per day and TV viewing time per day: < 7 h of total sitting time per day and < 2 h of TV viewing time per day were identified as lower tertiles; ≥7 h of total sitting per day and ≥ 2 h of TV viewing time per day were defined as higher total sitting time and TV time | The Spielberger state-trait anxiety inventory (STAI) | + for TV viewing | Women with higher TV viewing time had higher state anxiety compared to women with lower TV viewing time in crude analysis (OR = 1.56, 95% CI: [1.14, 2.14], p = .006) | 95 |
Park et al., 2017 [98]a | United Kingdom | Cross-sectional | Older adults (residents from assisted living facilities), (chronic illness) | 87 | 77.5 ± 8.2 | Accelerometers (GT3X+, WGT3X-BT; ActiGraph (Pensacola, FL, USA) | Anxiety subscale of the 14-item Hospital Anxiety and Depression Scale [73] | - for ASB | SB was negatively associated with anxiety (r = −.39, p < .01) | 95 |
Rebar et al., 2014 [99]a | Australia | Cross-sectional | Adults (general population) | 1104 (55%) | 58 (range 48–66) | 10-item Workforce Sitting Questionnaire [100] | Anxiety subscale from 21-item Depression, Anxiety, and Stress Scale (DASS-21) [101] | + for overall sitting time + for computer sitting 0 for leisure/work/TV sitting | Overall sitting time (b = 0.03, p < .05) and computer use (b = 0.03, p < .05) were significantly associated with anxiety. Leisure (b = 0.01, p > .05), work (b = 0.02, p > .05), and TV (b = 0.00, p > .05) were not associated with anxiety | 100 |
Sanchez-Villegas et al., 2008 [102] | Spain, USA | Longitudinal | Adults (general population) | 10,381 | 27 years | Self-report sedentary index: hours per week spent on watching television and/or using computer. | Self-reported anxiety: ‘Have you ever been diagnosed of anxiety by a health professional?’ - classified as an incident case of anxiety | 0 for ST | There was no significant association between the sedentary index and anxiety risk (p = .17) | 89 |
Straker et al., 2013 [103]a | Australia | Longitudinal | Adolescents (general population) | 643 (54%) | 14.0 ± 0.2 years | Screen Based Media – Self-report recall electronic diary/questionnaire MARCA - clusters: C1. instrumental computer gamers; C2. multi-modal e-gamers; C3. computer e-gamers [104] | Internalizing symptoms index of 112-item the Youth Self-Report [59] | + for game playing | C1 males reported less internalizing behavioral problems than C2 (difference − 1.7, 95% CI: [− 3.5, 0.1], p = .065) or C3 males (difference − 2.6, 95% CI: [− 4.9, − 0.3], p = .027) | 91 |
Teychenne & Hinkley, 2016 [105]a | Australia | Cross-sectional | Adult women with children aged 2–5 years (general population) | 575 (100%) | 37.18 ± 4.62 years | Self-report –of women’s own activities, including time spent on TV/DVD/video viewing, computer/e-games/hand held device use on a typical weekday and weekend day | 7-items related to symptoms of anxiety experienced in the past week: a subscale (HADS-A) of the Hospital Anxiety and Depression Scale (HADS) [106] | + for computer /device use + for total ST; 0 for TV viewing | TV viewing was not associated with anxiety symptoms (B = 0.109, p = .188) but computer/device use (B = 0.212, p = .011) and overall screen time (B = 0.109, p = .025) were positively associated with heightened anxiety symptoms in models adjusted for key demographic and behavioral covariates (including MVPA) | 100 |
Uijtdewilligen et al., 2011 [107] | Netherlands | Longitudinal | Adolescents (general population) | 217 (58%) | M: 13.0 ± 0.6 years; F: 12.9 ± 0.6 years | Acti-Graph accelerometers (Model GT1M, ActiGraph, LLC, Fort Walton Beach, FL) | Facilitating anxiety index of Achievement Motivation Test (AMT) [108] | + for ASB | In males, a higher score on facilitating anxiety (B = 5.13, 95% CI: [0.08, 10.19], p < .05) was associated with more minutes spent sedentary in adulthood. | 88 |
Vallance et al., 2015 [109]a | Canada; Australia | Cross-sectional | Adults (general population) | 197 (overall 180–45%) | 64.3 ± 10.3 | Acitgraph GT3X+ accelerometer (Actigraph, LLC, Pensacola, FL) | Spielberger’s State Anxiety Inventory (STAI) - 10 items | 0 for ASB | No significant associations emerged for sedentary time and psychological health outcomes (including anxiety) [Wilks’ λ = 0.956, F(9382.3) = 0.788, p = .628] | 93 |
van Roekel et al., 2016 [110]a | Netherlands | Cross-sectional | Adults treated for stage I–III colorectal cancer (chronic illness) | 145 (37.2%) | 70.0 ± 8.7 years | The triaxial MOX activity monitor (MMOXX1, upgraded version of the CAM monitor) | Anxiety subscale of the 14-item Hospital Anxiety and Depression Scale [73] | 0 for ASB (0 for sedentary per 1 h/day) | Substituting sedentary time with physical activity was not significantly associated with lower anxiety (β = −0.7, 95% CI: [− 1.7, 0.3]) | 95 |
Vancampfort et al., 2018 [111] | China, Ghana, India, Mexico, Russia, South Africa | Cross-sectional | Adults (general population) | 42,469 (50.1%) | 43.8 ± 14.4 | Self-report sitting time – total time usually spent (expressed in minutes per day) sitting or reclining including at work, at home, getting to and from places, or with friends (e.g., sitting at a desk, sitting with friends, travelling in car, bus, train, reading, playing cards or watching television) | Self-reported anxiety by the question ‘Overall in the past 30 days, how much of a problem did you have with worry or anxiety’ with response alternatives: none, mild, moderate, severe, extreme. Those who answered severe or extreme were considered to have anxiety | + for sitting time | Anxiety was significantly associated with higher mean time spent sitting (b = 24.16, 95% CI: [6.95, 41.38], p < .01) | 100 |
Wu et al., 2015 [112] | China | Cross-sectional | Adults (general population) | 4747 (58.4%) - 16.3% with anxiety | 19.2 ± 1.41 years | Self-report screen time measured with one item: ‘How many hours per day do you spend on the computer (including playing video or computer games or computer for something) and watching TV/video programs on a usual weekday and weekend day, respectively?’ ST was categorized as ≤2 h/d and > 2 h/d | Self-rating anxiety scale (SAS) [77] | + for ST | High screen time > 2 h/day (OR = 1.38, 95% CI: [1.15, 1.65], p < .001) was significantly positively associated with anxiety in crude model as well as in a model adjusted for gender, age, residential background, BMI, perceived family economy and perceived study burden (OR = 1.49, 95% CI: [1.24, 1.79], p < 0.001) | 98 |
Coding
Methods of data synthesis and data analysis
Results
Search results
A synthesis of findings from studies included into the systematic review
Sample/Category of SB | Accelerometry measured sedentary behaviors) | Total sitting time | Total screen time | TV viewing | Computer use/internet use/video game playing | Proportion of significant positive associations across studies (regardless of the category of SB) | Proportion of significant positive associations: Adults vs. children/adolescents (regardless of the category of SB) | Proportion of significant positive associations: general population vs. with chronic illness (regardless of the category of SB) |
---|---|---|---|---|---|---|---|---|
Adults | ||||||||
General population | Adults; general population: 14 (61%) in 23 associations | Total adults: 14 (48%) in 29 associations | Total for general population (children/adolescents vs. adults) 21 (60%) in 35 associations | |||||
With a chronic mental or physical illness | - (Kroeders et al., 2013) [89] | |||||||
Children and adolescents (aged < 18) | ||||||||
General population | + (Uijtdewilligen et al., 2011) [107] | Children/Adolescents; general population: 7 (58%) in 12 associations | Total children/adolescents: 7 (58%) in 12 associations | |||||
With a chronic mental or physical illness | ||||||||
Proportion of significant positive associations across studies (regardless age/ health status) | 4 (36%) of 11 associations | 6 (67%) of 9 associations | 4 (50%) in 8 associations | 3 (50%) of 6 associations | 5 (63%) in 8 associations | 21 (51%) in 41 associations |
The meta-analytic synthesis of findings
Estimate of the average effect | Range of correlation coefficients retrieved from original studies | 95% CI for the estimate of the average effect |
N
|
K
| Heterogeneity | Test for moderating effects | |||||
---|---|---|---|---|---|---|---|---|---|---|---|
Q
|
I
2
%
|
Tau
|
Tau
2
|
Q
B
|
P
| ||||||
Overall effect | .093 | .01; .46 | [.05; .13] | 26,204b | 17 | 77.04 p < .001 | 79.23 | .06 | .003 | ||
Moderators effects for overall effect and levels of respective moderators | |||||||||||
Age group | 2.97 | .085 | |||||||||
Children/adolescents | .05 | .01; .17 | [−.01; .11] | 17,873 | 5 | ||||||
Adults (over 18) | .12 | .02; .46 | [.06; .17] | 7868 | 7 | ||||||
Health status | 0.05 | .820 | |||||||||
Adults with a chronic physical or mental illness | .16 | .03; .39 | [.03; .30] | 463 | 5 | ||||||
Adults from the general population | .15 | .02; .46 | [.05; .23] | 6990 | 6 | ||||||
The type of measurement of sedentary behaviors | 1.08 | .299 | |||||||||
Objective | .14 | .03; .39 | [.04; .24] | 505 | 6 | ||||||
Self-report | .08 | .01; .46 | [.04; .12] | 25,699 | 11 | ||||||
The type of SB | 2.21 | .137 | |||||||||
Total sitting time | .12 | .02; .46 | [.06; .19] | 7298 | 5 | ||||||
Total screen time | .06 | .01; .17 | [.00; 11] | 18,401 | 6 | ||||||
The sub-type of screen use-related behaviors | 5.04a | .080 | |||||||||
Computer using |
.12
| .10; .14 | [.05; .18] | 2183 | 2 | ||||||
Computer/video games playing |
.02
| .02; .03 | [−.03; .08] | 15,896 | 2 | ||||||
TV viewing | .05 | −.001; .10 | [.01; .09] | 16,475 | 4 |