Skip to main content
Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Potential causal association between leisure sedentary behaviors, physical activity and musculoskeletal health: A Mendelian randomization study

Abstract

Background

Increasing evidence shows that leisure sedentary behaviors (LSB) and physical activity (PA) are associated with various musculoskeletal disorders. However, the causality between LSB/PA and musculoskeletal health remained unknown. In this study, we aimed to evaluate the causal relationships between LSB/PA and lower back pain (LBP), intervertebral disc disorder (IVDD), rheumatoid arthritis (RA), and bone mineral density (BMD) by using a two-sample Mendelian randomization method.

Methods

The exposure data were obtained from large-scale genome-wide association studies (GWAS), including the PA dataset (self-reported PA, n  =  377,234; accelerometer-assessed PA, n  = 91,084) and LSB dataset (n  =  422,218). The outcome data were derived from the FinnGen LBP dataset (n  =  248,528), FinnGen IVDD dataset (n  =  256,896), BMD GWAS dataset (n  =  56,284), and RA GWAS dataset (n  =  58,284). The causal relationships were estimated with inverse variance weighted (IVW), MR-Egger, and weighted median methods. Sensitivity analyses were performed with Cochran’s Q test, MR-Egger intercept test, and leave-one-out analysis to estimate the robustness of our findings.

Results

Genetically predicted leisure television watching increased the risk of LBP (OR = 1.68, 95% CI 1.41 to 2.01; P = 8.23×10−9) and IVDD (OR  =  1.62, 95% CI 1.37 to 1.91; P  =  2.13 × 10−8). In addition, this study revealed a potential causal relationship between computer use and a reduced risk of IVDD (OR  = 0.60, 95% CI 0.42 to 0.86; P  =  0.005) and RA (OR  = 0.28, 95% CI 0.13 to 0.60; P  =  0.001).

Conclusions

Our results suggest that leisure television watching is a risk factor for LBP and IVDD, whereas leisure computer use may act as a protective factor against IVDD and RA. These findings emphasized the importance of distinguishing between different sedentary behaviors in musculoskeletal disease studies.

Introduction

Musculoskeletal disorders, including lower back pain (LBP), intervertebral disc disorder (IVDD), rheumatoid arthritis (RA), and osteoporosis, are a major cause of morbidity and disability globally, imposing a heavy economic burden on society [14]. IVDD is a common degenerative disorder that is frequently associated with lower back pain [5, 6]. Osteoporosis, a systemic musculoskeletal disease characterized by decreased bone mass and degeneration of bone tissue microarchitecture, has been recognized as one of the most pressing public health problems [7]. Osteoporosis is diagnosed by measuring bone mineral density (BMD) [8]. RA is an autoimmune disease characterized by synovial inflammation as well as bone and cartilage destruction [9, 10]. A key step in the prevention of musculoskeletal disorders is the identification of possible risk factors and protective factors, especially those that can be intervened.

Leisure sedentary behaviors (LSB) refers to any waking behavior characterized by energy expenditure ≤1.5 metabolic equivalents in a reclining or sitting posture [11]. Studies have previously shown that sedentary behaviors are associated with increased risk of all-cause mortality, cardiovascular disease, and musculoskeletal disorders [1215]. Physical activity (PA) is the musculoskeletal movement that results in energy expenditure [16]. There is some evidence that exercise is effective for musculoskeletal pain and confers unique benefits to the musculoskeletal system [17, 18].

Observational studies have suggested that sedentary behaviors may increase the risk of LBP [14, 15] and IVDD [19]. Sedentary behavior and physical inactivity may be associated with poor health-related outcomes (e.g., low functionality and worse disease symptoms) in RA [20]. And studies have indicated that physical activity may reduce the risk of LBP [14], IVDD [21], and osteoporosis [22]. Physical activity interventions for LBP are commonly proposed to work through postural changes and muscle activation [23]. In addition, from a biopsychosocial perspective, LBP involves combinations of different social, psychological, lifestyle, and physical factors [24]. PA is a complex and multidimensional behavior [25]. PA has various positive biopsychosocial effects that are beneficial for the prevention of musculoskeletal system disorders [26].

However, the causal relationship between LSB/PA and musculoskeletal disorder risk is not clear. Existing evidence, based on observational studies, is unable to entirely rule out the possibility of biases, including confounding factors and reverse causality. Besides, randomized controlled studies on this topic are impractical and unethical to constrain participants’ physical activity. Mendelian randomization (MR) is a method of using genetic variants associated with the specific exposures of interest to study the causal relationships of modifiable exposures (i.e., possible risk factors) on outcomes [27]. Because single nucleotide polymorphisms (SNPs) are allocated randomly at conception, MR analysis could effectively exclude confounding factors and reverse causality [28]. MR utilizes genetic variations as instrumental variables (IVs) to evaluate the associations between exposures and outcomes.

Given the uncertainty regarding the causal effects of LSB/PA on musculoskeletal disorders, we applied the MR design to assess the potential causal associations. Overall, this MR analysis explores the potential causal relationship between LSB/PA and musculoskeletal health (LBP, IVDD, RA, and BMD).

Methods

Study design

In this study, we conducted a two-sample MR analysis to investigate the causal associations of PA/LSB with LBP, IVDD, BMD, and RA by using genome-wide association studies (GWAS) summary statistics. To obtain unbiased causal effects, this MR analysis needs to satisfy three assumptions: (i) the genetic variants are robustly related to the exposure of interest; (ii) genetic variants are not associated with potential confounders; and (iii) genetic variants affect the outcomes merely through the exposure of interest. The study design is shown in Fig 1. No separate ethical approval was required in this study.

Data sources for sedentary behaviors and physical activity

The genetic instruments for sedentary behaviors were obtained from the latest summary-level GWAS of 422,218 European-ancestry individuals from the UK Biobank [29]. The UK Biobank cohort is a large prospective cohort of 500,000 participants recruited in the United Kingdom between 2006 and 2010 [30]. Leisure sedentary behaviors included three phenotypes: hours of television watching per day, hours of computer use per day (excluding computer use at work), and hours of driving per day. The number of driving-related SNPs was insufficient, with a risk of weak instrument bias; therefore, it was not included in our study.

The summary statistics for physical activity were derived from a recent GWAS, which was performed in the UK Biobank [31].This GWAS, consisting of over 377,000 European-ancestry participants, evaluated three measures based on self-report and two measures based on wrist-worn accelerometry data. The self-reported moderate to vigorous physical activity (MVPA) was calculated by taking the sum of total minutes/week of moderate PA multiplied by 4 and the total number of vigorous PA minutes/week multiplied by 8, corresponding to their metabolic equivalents [31]. We used two phenotypes of PA, including self-reported MVPA and accelerometer-assessed PA (average acceleration).

Data source for LBP, IVDD, BMD, and RA

The summary results of LBP and IVDD were obtained from the FinnGen consortium. On June 1, 2022, FinnGen Data (R7) was released to the public. The dataset contains LBP (21,140 cases and 227,388 controls) and IVDD (29,508 cases and 227,388 controls).

Summarized data of total body BMD was derived from a large meta-analysis of GWAS comprised of 56,284 European-ancestry individuals [8]. Total body BMD (g/cm2) was measured by dual-energy X-ray absorptiometry according to standard manufacturer protocols. Moreover, total body BMD was adjusted for age, weight, height, genomic principal components, and additional study-specific covariates (e.g., recruiting center). The GWAS summary data for total body BMD is obtained from the IEU Open GWAS database [32]. The GWAS ID for BMD is ebi-a-GCST005348.

The summary-level data for RA were available from a large-scale meta-analysis of GWAS, comprising over 100,000 participants of European and Asian ancestries. To reduce population-stratification bias, we extracted summarized data for RA from 58,284 participants of European ancestry (14,361 RA cases and 43,923 controls). A detailed description of the study can be found elsewhere [33]. Summary GWAS data for RA were obtained from the IEU Open GWAS database [32]. The GWAS ID for RA is ieu-a-832. Details of all GWASs included in this study are shown in Table 1.

thumbnail
Table 1. Details of the GWASs included in the Mendelian randomization.

https://doi.org/10.1371/journal.pone.0283014.t001

Selection of genetic instruments

In this study, we used four sets of genetic instruments indicating LSB and PA, including (1) index SNPs representing television watching (listed in S1 Table in S1 File), (2) index SNPs representing computer use (listed in S2 Table in S1 File), (3) index SNPs representing MVPA (listed in S3 Table in S1 File), and (4) index SNPs representing accelerometer-assessed PA (listed in S4 Table in S1 File).

Genetic instrumental variables were selected by the following criteria: (1) a GWAS-correlated P-value of 5×10−8; (2) removal of SNPs in linkage disequilibrium (LD, R2 ≥ 0.001 and within 10 mb); and (3) removal of SNPs with F statistic less than 10. Instrumental variables with F-statistics <10 were regarded as weak genetic instruments [34].

Mendelian randomization analyses

Three MR methods, including inverse-variance weighted (IVW), weighted median (WM), and MR‐Egger, were conducted to evaluate the causal effects of PA/LSB on LBP, IVDD, BMD, and RA. These MR approaches have different underlying assumptions about horizontal pleiotropy. IVW, which assumes that instrumental variables can affect the outcome only via the exposure, was used as the main method. Weighted median and MR-Egger methods were implemented as complements to the IVW. The weighted median method provides consistent estimates when more than 50% of the information comes from valid IVs [35]. MR-Egger assumes that the variant-exposure association is independent of the pleiotropic effect of the genetic variant [36].

For important estimates, MR‐pleiotropy residual sum and outlier (MR‐PRESSO) method was performed to correct horizontal pleiotropy through outlier removal [37]. The MR-Egger intercept, Cochran’s Q test, funnel pot, and leave-one-out analyses were used to verify the robustness of our results [38]. Specifically, we assessed horizontal pleiotropy according to the intercept term derived from MR-Egger regression (P < 0.05 was considered the presence of pleiotropy). And P<0.05 of Cochran’s Q-test indicated that heterogeneity was detected. We also performed leave-one-out analysis, in which each exposure-related SNP was removed in turn to repeat the IVW analysis, in order to determine whether the causal association was driven by a single SNP.

Statistics

All analyses were performed using TwoSampleMR (version 0.5.6) in R (version 4.2.1) packages. To account for multiple testing in our analyses, a Bonferroni-corrected P-value was set as 0.05/16 (0.003). P < 0.003 was regarded as statistically significant, and P < 0.05 was considered nominally significant. MR estimates were shown as odds ratios (OR) with 95% confidence intervals (CI) or β estimates with 95% CI. In addition, we calculated the statistical power (https://shiny.cnsgenomics.com/mRnd/).

Results

MR estimates

Among the tested LSB phenotypes, IVW analysis indicated that leisure television viewing increased the risk of LBP (OR = 1.68, 95% CI 1.41 to 2.01; P = 8.23×10−9). Similar causal estimates were gained from WM analysis (OR  =  1.54, 95% CI 1.21 to 1.95; P  =  0.0004). The result from MR-Egger analysis showed a consistent but nonsignificant direction (OR  =  2.31, 95% CI 1.00 to 5.34; P  =  0.055) (Fig 2A). Additionally, a significant association between computer use and LBP was observed using IVW analysis, whereas the causal estimates from MR-Egger analysis displayed directional inconsistency (Fig 2A).

thumbnail
Fig 2.

Causal effects for PA/LSB on lower back pain (A), intervertebral disc disorder (B), bone mineral density (C) and rheumatoid arthritis (D). Summarized Mendelian randomization (MR) effect sizes were derived from the inverse variance weighted (IVW), MR-Egger and weighted median methods. PA, physical activity; MVPA, moderate to vigorous physical activity.

https://doi.org/10.1371/journal.pone.0283014.g002

Television viewing was significantly associated with an increased risk of IVDD with the IVW method (OR  =  1.62, 95% CI 1.37 to 1.91; P  =  2.13 × 10−8). Meanwhile, similar risk estimates were obtained using the WM method (OR = 1.56, 95% CI = 1.27 to 1.93, P = 3.19 × 10−5). MR-Egger also showed similar causal estimates, although the association was not statistically significant (Fig 2B). In addition, a nominally significant association between computer use and IVDD was observed using IVW analysis (OR  = 0.60, 95% CI 0.42 to 0.86; P  =  0.005). WM and MR-Egger showed a consistent but nonsignificant association (Fig 2B).

Among the LSB phenotypes (television watching and computer use), we did not observe evidence of causal associations between genetically predicted sedentary behaviors and total body BMD (Fig 2C).

Moreover, a negative causality between genetically proxied computer use and RA was detected using IVW analysis (OR  = 0.28, 95% CI 0.13 to 0.60; P  =  0.001). WM analysis showed a nominally significant result (OR  =  0.26, 95% CI 0.10 to 0.67; P  =  0.006) and MR-Egger presented a consistent but nonsignificant result (Fig 2D). Additionally, causal estimates of the relationship between television watching and RA displayed directional inconsistency in three MR methods (Fig 2D).

Furthermore, among the tested PA phenotypes (MVPA and accelerometer assessed PA), we did not observe evidence of causal associations between genetically proxied PA phenotypes and musculoskeletal disorders (LBP, IVDD, BMD, and RA) (Fig 2A–2D). In addition, details on MR estimates of the causal relationships between PA/LSB and these musculoskeletal disorders are shown in S5 Table in S1 File. We also calculated the statistical power for important estimates. A power of over 80% was observed. Statistical power is presented in S6 Table in S1 File.

Sensitivity analyses

The Cochran’s Q test and the MR-Egger intercept test were conducted to evaluate the robustness of the results (Table 2). The P-values of the MR-Egger intercept tests were all > 0.05, suggesting that there was no horizontal pleiotropy. However, heterogeneity was detected in the Cochran’s Q test except for estimates of MVPA on RA (P = 0.093), accelerometer assessed PA on RA (P = 0.068), television watching on BMD (P = 0.128), computer use on LBP (P = 0.330), and computer use on RA (P = 0.081). Although heterogeneity was observed in some results, it did not invalidate the MR estimates as random-effect IVW in the current study. Furthermore, the Egger intercept did not detect any pleiotropy, suggesting that no pleiotropy bias was introduced into the MR estimates despite heterogeneity (Fig 3). Funnel plots, forest plots, and leave-one-out analyses are shown in S1-S8 Figs in S1 File. The results were not biased by a single SNP, indicating that estimates were not violated.

thumbnail
Fig 3.

Scatter plots from genetically predicted television watching on lower back pain (A), television watching on intervertebral disc disorder (B), computer use on intervertebral disc disorder (C) and computer use on rheumatoid arthritis (D).

https://doi.org/10.1371/journal.pone.0283014.g003

thumbnail
Table 2. Sensitivity analysis of the causal effects of LSB/PA on musculoskeletal disorders.

https://doi.org/10.1371/journal.pone.0283014.t002

Discussion

In the current study, we evaluated the casual associations between genetically predicted LSB/PA phenotypes and different musculoskeletal disorders (LBP, IVDD, BMD, and RA) by using three MR methods. Our results suggest that leisure television watching is a risk factor for LBP and IVDD, whereas leisure computer use may be a protective factor against IVDD and RA. In addition, we found little evidence for the association between PA phenotypes and musculoskeletal disorders.

Observational studies have suggested that sustained sedentary behaviors are negatively associated with musculoskeletal health [15, 39]. Lack of exercise is a main cause of most chronic diseases, including chronic musculoskeletal disorders [40]. However, direct evidence for the causal relationships between LSB/PA and various musculoskeletal disorders is still lacking. Compared with large-scale prospective clinical trials requiring long-term observation, the MR study revealed the potential causal relationship between LSB/PA and different musculoskeletal disorders in a time- and cost-saving manner.

Television watching is the main sedentary behavior associated with leisure time [41]. Compared with other leisure sedentary traits, television watching was a typical sedentary behavior, leading to an energy surplus through lower energy expenditure and excessive energy intake (especially snacks) [42]. Studies have shown that weight gain associated with an energy surplus is a risk factor for LBP and IVDD [43, 44]. Consistent with these findings, our MR results indicated a potential causal association of television watching with LBP and IVDD. Intriguingly, leisure computer use was negatively associated with IVDD and RA. These results indicated a biological heterogeneity behind different sedentary behaviors. In addition, MR-Egger intercept tests detected no horizontal pleiotropy. Sensitivity analyses suggested the robustness of our results. Moreover, the causal inference had high reliability because the statistical power of the IVW method for important estimates was above the threshold of 80%.

Additionally, our in-depth analyses provided several potential explanations for the causal relationships between sedentary behaviors and LBP, IVDD, and RA. First, television watching is generally considered ‘mentally passive’ behavior, whereas computer use is usually considered ‘mentally active’ behavior [45, 46]. Mentally passive sedentary behavior (television watching) was always accompanied by poor physical and mental health (such as depression) [45, 46]. A meta-analysis showed that symptoms of depression increased the risk of developing LBP [47]. And depression is strongly associated with RA [48]. A recent study illustrated the bidirectional communication between the central nervous system and chronic peripheral inflammation, which indicated the pathophysiologic interactions between neuropsychiatric diseases and RA, highlighting the existence of the joint-brain axis [9]. Therefore, mental health disorders might serve as a key intermediate factor in the LSB-RA pathway. Second, sitting behavior is associated with LBP [49]. Patients with LBP exhibited a more static sitting behavior. Compared with static posture, dynamic sitting behavior is thought to reduce spinal load and muscle fatigue [50, 51]. Television viewing is a more immersive and less reflective form of leisure entertainment than computer use [52]. Television watchers tend to show a more static sitting behavior, which is a risk factor of LBP and IVDD. Third, older adults are more likely to watch television as a leisure entertainment, whereas young people tend to spend more time on computer use [52]. Aging has been considered an important cause of musculoskeletal disorders, especially LBP and IVDD [5, 53, 54]. Our study indicates that public health initiatives targeting LSB can lower the risk of LBP and IVDD, thereby gaining potential health benefits. The major audience for television watching, especially elderly adults, should be assisted in reducing television viewing time and static sitting behavior.

This MR study has several strengths. The main strength of our study is the two-sample MR design, which reduces unobserved confounding and reverse causality. Additionally, multiple sensitivity analyses were performed to test the validity of the MR assumptions, minimizing the possibility of biased results. IVW was used as the main method, with higher statistical power than other MR methods, especially MR-Egger [52, 55]. Thus, it is not surprising that the MR-Egger results with low statistical power had nonsignificant P values and wider confidence compared to IVW. Our study strengthened the requirement for consistent beta direction in all MR methods.

This MR study has several limitations. First, the findings cannot be immediately applied to other ethnic groups with different cultures and lifestyles, as the participants in the datasets were of European ancestry. Second, it is difficult to completely exclude pleiotropy because the functional biological roles of these genetic variants are not fully understood. Horizontal pleiotropy affects the stability of MR results, but vertical pleiotropy, where exposure acts on outcome through other factors along the same causal pathway, is acceptable. Third, our findings show the potential causal relationships between different sedentary behaviors and musculoskeletal disorders, but the potential mechanisms warrant further study.

Conclusions

Our results suggest that television viewing is causally related to an increased risk of LBP and IVDD, while computer use may be causally associated with a decreased risk of IVDD and RA. Our findings highlight the importance of distinguishing between different domains of sedentary behaviors.

Acknowledgments

We wish to acknowledge all participants and investigators who contributed to the GWAS data.

References

  1. 1. Hoy DG, Smith E, Cross M, Sanchez-Riera L, Buchbinder R, Blyth FM, et al. The global burden of musculoskeletal conditions for 2010: an overview of methods. Ann Rheum Dis. 2014;73(6):982–9. pmid:24550172.
  2. 2. Roodenrijs N, Welsing P, van der Goes MC, Tekstra J, Lafeber F, Jacobs J, et al. Healthcare utilization and economic burden of difficult-to-treat rheumatoid arthritis: a cost-of-illness study. Rheumatology (Oxford). 2021;60(10):4681–90. pmid:33502493.
  3. 3. Clynes MA, Harvey NC, Curtis EM, Fuggle NR, Dennison EM, Cooper C. The epidemiology of osteoporosis. Br Med Bull. 2020;133(1):105–17. pmid:32282039.
  4. 4. March L, Smith EU, Hoy DG, Cross MJ, Sanchez-Riera L, Blyth F, et al. Burden of disability due to musculoskeletal (MSK) disorders. Best Pract Res Clin Rheumatol. 2014;28(3):353–66. pmid:25481420.
  5. 5. Francisco V, Pino J, González-Gay MÁ, Lago F, Karppinen J, Tervonen O, et al. A new immunometabolic perspective of intervertebral disc degeneration. Nat Rev Rheumatol. 2022;18(1):47–60. pmid:34845360.
  6. 6. Clark S, Horton R. Low back pain: a major global challenge. Lancet. 2018;391(10137):2302. pmid:29573869.
  7. 7. Wang L, Zhang C, Liang H, Zhou N, Huang T, Zhao Z, et al. Polyunsaturated Fatty Acids Level and Bone Mineral Density: A Two-Sample Mendelian Randomization Study. Front Endocrinol (Lausanne). 2022;13:858851. pmid:35872990.
  8. 8. Medina-Gomez C, Kemp JP, Trajanoska K, Luan J, Chesi A, Ahluwalia TS, et al. Life-Course Genome-wide Association Study Meta-analysis of Total Body BMD and Assessment of Age-Specific Effects. Am J Hum Genet. 2018;102(1):88–102. pmid:29304378.
  9. 9. Süß P, Rothe T, Hoffmann A, Schlachetzki J, Winkler J. The Joint-Brain Axis: Insights From Rheumatoid Arthritis on the Crosstalk Between Chronic Peripheral Inflammation and the Brain. Front Immunol. 2020;11:612104. pmid:33362800.
  10. 10. Zheng LX, Li KX, Hong FF, Yang SL. Pain and bone damage in rheumatoid arthritis: role of leukotriene B4. Clin Exp Rheumatol. 2019;37(5):872–8. pmid:30943138.
  11. 11. Tremblay MS, Aubert S, Barnes JD, Saunders TJ, Carson V, Latimer-Cheung AE, et al. Sedentary Behavior Research Network (SBRN)—Terminology Consensus Project process and outcome. Int J Behav Nutr Phys Act. 2017;14(1):75. pmid:28599680.
  12. 12. Basterra-Gortari FJ, Bes-Rastrollo M, Gea A, Núñez-Córdoba JM, Toledo E, Martínez-González MÁ. Television viewing, computer use, time driving and all-cause mortality: the SUN cohort. J Am Heart Assoc. 2014;3(3):e864. pmid:24965030.
  13. 13. Celis-Morales CA, Lyall DM, Steell L, Gray SR, Iliodromiti S, Anderson J, et al. Associations of discretionary screen time with mortality, cardiovascular disease and cancer are attenuated by strength, fitness and physical activity: findings from the UK Biobank study. Bmc Med. 2018;16(1):77. pmid:29792209.
  14. 14. Hanna F, Daas RN, El-Shareif TJ, Al-Marridi HH, Al-Rojoub ZM, Adegboye OA. The Relationship Between Sedentary Behavior, Back Pain, and Psychosocial Correlates Among University Employees. Front Public Health. 2019;7:80. pmid:31024881.
  15. 15. Dzakpasu F, Carver A, Brakenridge CJ, Cicuttini F, Urquhart DM, Owen N, et al. Musculoskeletal pain and sedentary behaviour in occupational and non-occupational settings: a systematic review with meta-analysis. Int J Behav Nutr Phys Act. 2021;18(1):159. pmid:34895248.
  16. 16. Rausch OA, Niedermann K, Braun J, Adams J, Brodin N, Dagfinrud H, et al. 2018 EULAR recommendations for physical activity in people with inflammatory arthritis and osteoarthritis. Ann Rheum Dis. 2018;77(9):1251–60. pmid:29997112.
  17. 17. Kristensen J, Franklyn-Miller A. Resistance training in musculoskeletal rehabilitation: a systematic review. Br J Sports Med. 2012;46(10):719–26. pmid:21791457.
  18. 18. Maestroni L, Read P, Bishop C, Papadopoulos K, Suchomel TJ, Comfort P, et al. The Benefits of Strength Training on Musculoskeletal System Health: Practical Applications for Interdisciplinary Care. Sports Med. 2020;50(8):1431–50. pmid:32564299.
  19. 19. Teichtahl AJ, Urquhart DM, Wang Y, Wluka AE, O’Sullivan R, Jones G, et al. Physical inactivity is associated with narrower lumbar intervertebral discs, high fat content of paraspinal muscles and low back pain and disability. Arthritis Res Ther. 2015;17(1):114. pmid:25947906.
  20. 20. Sun L, Zhu J, Ling Y, Mi S, Li Y, Wang T, et al. Physical activity and the risk of rheumatoid arthritis: evidence from meta-analysis and Mendelian randomization. Int J Epidemiol. 2021;50(5):1593–603. pmid:33760079.
  21. 21. Bowden JA, Bowden AE, Wang H, Hager RL, LeCheminant JD, Mitchell UH. In vivo correlates between daily physical activity and intervertebral disc health. J Orthop Res. 2018;36(5):1313–23. pmid:28976592.
  22. 22. Pinheiro MB, Oliveira J, Bauman A, Fairhall N, Kwok W, Sherrington C. Evidence on physical activity and osteoporosis prevention for people aged 65+ years: a systematic review to inform the WHO guidelines on physical activity and sedentary behaviour. Int J Behav Nutr Phys Act. 2020;17(1):150. pmid:33239014.
  23. 23. Shiri R, Falah-Hassani K. Does leisure time physical activity protect against low back pain? Systematic review and meta-analysis of 36 prospective cohort studies. Br J Sports Med. 2017;51(19):1410–8. pmid:28615218.
  24. 24. Kamper SJ, Apeldoorn AT, Chiarotto A, Smeets RJEM, Ostelo RWJG, Guzman J, et al. Multidisciplinary biopsychosocial rehabilitation for chronic low back pain: Cochrane systematic review and meta-analysis. BMJ: British Medical Journal. 2015;350:h444. http://doi.org/10.1136/bmj.h444
  25. 25. Pettee GK, Morrow JJ, Woolsey AL. Framework for physical activity as a complex and multidimensional behavior. J Phys Act Health. 2012;9 Suppl 1:S11–8. pmid:22287443.
  26. 26. John JM, Haug V, Thiel A. Physical Activity Behavior from a Transdisciplinary Biopsychosocial Perspective: a Scoping Review. Sports Med Open. 2020;6(1):49. pmid:33068179.
  27. 27. Richmond RC, Davey SG. Mendelian Randomization: Concepts and Scope. Cold Spring Harb Perspect Med. 2022;12(1). pmid:34426474.
  28. 28. Davey SG, Hemani G. Mendelian randomization: genetic anchors for causal inference in epidemiological studies. Hum Mol Genet. 2014;23(R1):R89–98. pmid:25064373.
  29. 29. van de Vegte YJ, Said MA, Rienstra M, van der Harst P, Verweij N. Genome-wide association studies and Mendelian randomization analyses for leisure sedentary behaviours. Nat Commun. 2020;11(1):1770. pmid:32317632.
  30. 30. Fry A, Littlejohns TJ, Sudlow C, Doherty N, Adamska L, Sprosen T, et al. Comparison of Sociodemographic and Health-Related Characteristics of UK Biobank Participants With Those of the General Population. Am J Epidemiol. 2017;186(9):1026–34. pmid:28641372.
  31. 31. Klimentidis YC, Raichlen DA, Bea J, Garcia DO, Wineinger NE, Mandarino LJ, et al. Genome-wide association study of habitual physical activity in over 377,000 UK Biobank participants identifies multiple variants including CADM2 and APOE. Int J Obes (Lond). 2018;42(6):1161–76. pmid:29899525.
  32. 32. Elsworth B, Lyon M, Alexander T, Liu Y, Matthews P, Hallett J, et al. The MRC IEU OpenGWAS data infrastructure. bioRxiv. 2020:2020–8. http://doi.org/10.1101/2020.08.10.244293
  33. 33. Okada Y, Wu D, Trynka G, Raj T, Terao C, Ikari K, et al. Genetics of rheumatoid arthritis contributes to biology and drug discovery. Nature. 2014;506(7488):376–81. pmid:24390342.
  34. 34. Burgess S, Thompson SG. Avoiding bias from weak instruments in Mendelian randomization studies. Int J Epidemiol. 2011;40(3):755–64. pmid:21414999.
  35. 35. Bowden J, Davey SG, Haycock PC, Burgess S. Consistent Estimation in Mendelian Randomization with Some Invalid Instruments Using a Weighted Median Estimator. Genet Epidemiol. 2016;40(4):304–14. pmid:27061298.
  36. 36. Bowden J, Davey SG, Burgess S. Mendelian randomization with invalid instruments: effect estimation and bias detection through Egger regression. Int J Epidemiol. 2015;44(2):512–25. pmid:26050253.
  37. 37. Ong JS, MacGregor S. Implementing MR-PRESSO and GCTA-GSMR for pleiotropy assessment in Mendelian randomization studies from a practitioner’s perspective. Genet Epidemiol. 2019;43(6):609–16. pmid:31045282.
  38. 38. Chen X, Kong J, Pan J, Huang K, Zhou W, Diao X, et al. Kidney damage causally affects the brain cortical structure: A Mendelian randomization study. Ebiomedicine. 2021;72:103592. pmid:34619639.
  39. 39. Alzahrani H, Alshehri MA, Alzhrani M, Alshehri YS, Al AW. The association between sedentary behavior and low back pain in adults: a systematic review and meta-analysis of longitudinal studies. Peerj. 2022;10:e13127. pmid:35391924.
  40. 40. Booth FW, Roberts CK, Laye MJ. Lack of exercise is a major cause of chronic diseases. Compr Physiol. 2012;2(2):1143–211. pmid:23798298.
  41. 41. Clark BK, Healy GN, Winkler EA, Gardiner PA, Sugiyama T, Dunstan DW, et al. Relationship of television time with accelerometer-derived sedentary time: NHANES. Med Sci Sports Exerc. 2011;43(5):822–8. pmid:20980928.
  42. 42. Otten JJ, Jones KE, Littenberg B, Harvey-Berino J. Effects of television viewing reduction on energy intake and expenditure in overweight and obese adults: a randomized controlled trial. Arch Intern Med. 2009;169(22):2109–15. pmid:20008695.
  43. 43. Liuke M, Solovieva S, Lamminen A, Luoma K, Leino-Arjas P, Luukkonen R, et al. Disc degeneration of the lumbar spine in relation to overweight. Int J Obes (Lond). 2005;29(8):903–8. pmid:15917859.
  44. 44. Samartzis D, Karppinen J, Chan D, Luk KD, Cheung KM. The association of lumbar intervertebral disc degeneration on magnetic resonance imaging with body mass index in overweight and obese adults: a population-based study. Arthritis Rheum. 2012;64(5):1488–96. pmid:22287295.
  45. 45. He Q, Bennett AN, Fan B, Han X, Liu J, Wu K, et al. Assessment of Bidirectional Relationships between Leisure Sedentary Behaviors and Neuropsychiatric Disorders: A Two-Sample Mendelian Randomization Study. Genes (Basel). 2022;13(6). pmid:35741723.
  46. 46. Hallgren M, Dunstan DW, Owen N. Passive Versus Mentally Active Sedentary Behaviors and Depression. Exerc Sport Sci Rev. 2020;48(1). pmid:31663866
  47. 47. Pinheiro MB, Ferreira ML, Refshauge K, Ordoñana JR, Machado GC, Prado LR, et al. Symptoms of Depression and Risk of New Episodes of Low Back Pain: A Systematic Review and Meta-Analysis. Arthritis Care Res (Hoboken). 2015;67(11):1591–603. pmid:25989342.
  48. 48. Nerurkar L, Siebert S, McInnes IB, Cavanagh J. Rheumatoid arthritis and depression: an inflammatory perspective. Lancet Psychiat. 2019;6(2):164–73. pmid:30366684.
  49. 49. Bontrup C, Taylor WR, Fliesser M, Visscher R, Green T, Wippert PM, et al. Low back pain and its relationship with sitting behaviour among sedentary office workers. Appl Ergon. 2019;81:102894. pmid:31422243.
  50. 50. Davis KG, Kotowski SE. Postural variability: an effective way to reduce musculoskeletal discomfort in office work. Hum Factors. 2014;56(7):1249–61. pmid:25490805.
  51. 51. van Dieën JH, de Looze MP, Hermans V. Effects of dynamic office chairs on trunk kinematics, trunk extensor EMG and spinal shrinkage. Ergonomics. 2001;44(7):739–50. pmid:11437206.
  52. 52. Chen X, Hong X, Gao W, Luo S, Cai J, Liu G, et al. Causal relationship between physical activity, leisure sedentary behaviors and COVID-19 risk: a Mendelian randomization study. J Transl Med. 2022;20(1):216. pmid:35562752.
  53. 53. Wang F, Cai F, Shi R, Wang XH, Wu XT. Aging and age related stresses: a senescence mechanism of intervertebral dis degeneration. Osteoarthritis Cartilage. 2016;24(3):398–408. pmid:26455958.
  54. 54. Buchbinder R, van Tulder M, Öberg B, Costa LM, Woolf A, Schoene M, et al. Low back pain: a call for action. Lancet. 2018;391(10137):2384–8. pmid:29573871.
  55. 55. Lin Z, Deng Y, Pan W. Combining the strengths of inverse-variance weighting and Egger regression in Mendelian randomization using a mixture of regressions model. Plos Genet. 2021;17(11):e1009922. pmid:34793444.