Article Text

Assessing the causal role of physical activity and leisure sedentary behaviours with chronic obstructive pulmonary disease: a Mendelian randomisation study
  1. Lu Xiao1,
  2. Weina Li1,
  3. Fawei Li2,
  4. Xingjuan Chen1,
  5. Yun Xu1,
  6. Ying Hu3,
  7. Yingkun Fu1 and
  8. Ling Feng1
  1. 1Department of Health Care, China Academy of Chinese Medical Sciences Guang'anmen Hospital, Beijing, China
  2. 2Beijing University of Chinese Medicine, Beijing, China
  3. 3Preventive Treatment Health Management Center, First Teaching Hospital of Tianjin University of Traditional Chinese Medicine (National Clinical Research Center for Chinese Medicine Acupuncture and Moxibustion), Tianjin, China
  1. Correspondence to Dr Yingkun Fu; fyk0317{at}126.com; Professor Ling Feng; flyutong{at}126.com

Abstract

Background Observational studies show that patients with chronic obstructive pulmonary disease (COPD) tend to be sedentary during leisure time. Physical activity (PA) may reduce the risk of COPD, but the causal relationship is unclear. We used a Mendelian randomisation (MR) method to elucidate the association of leisure sedentary behaviours (LSB) and PA with lung function and COPD.

Methods Data on LSB (n=422 218), PA (n=608 595), COPD (n=299 929) and lung function (n=79 055) were obtained from the large-scale genome-wide association study. Causal inference used inverse variance-weighted, MR-Egger and weighted median. Sensitivity analysis was performed to assess heterogeneity and pleiotropy, and radial MR was used to distinguish outliers. The primary outcome was analysed by multifactorial MR adjusted for daily smoking.

Results The inverse variance weighted analysis indicated that increased moderate-to-vigorous PA (MVPA) is associated with higher levels of forced vital capacity (FVC) (beta=0.27, 95% CI 0.12 to 0.42; p=3.51×10–4). For each increment of 2.8 hours in television watching, the odds of COPD were 2.25 times greater (OR=2.25; 95% CI 1.84 to 2.75; p=2.38×10–15). For early-onset COPD, the odds were 2.11 times greater (OR=2.11; 95% CI 1.56 to 2.85; p=1.06×10–6), and for late-onset COPD, the odds were 2.16 times greater (OR=2.16; 95% CI 1.64 to 2.84; p=3.12×10–8). Similarly, the odds of hospitalisation for COPD were 2.02 times greater with increased television watching (OR=2.02; 95% CI 1.59 to 2.55; p=4.68×10–9). Television watching was associated with lower FVC (beta=−0.19, 95% CI −0.28 to −0.10; p=1.54×10–5) and forced expiratory volume in the 1 s (FEV1) (beta=−0.16, 95% CI −0.25 to −0.08; p=1.21×10–4) levels. The results remained significant after adjustment for smoking.

Conclusions Our study suggests a potential association with LSB, particularly television watching, is associated with higher odds of COPD and lower indices of lung function as measured continuously, including FEV1 and FVC. Conversely, an increase in MVPA is associated with higher indices of lung function, particularly reflected in increased FVC levels.

  • COPD epidemiology
  • Pulmonary Disease, Chronic Obstructive
  • Respiratory Function Test

Data availability statement

Data are available in a public, open access repository. All data are from the publicly available GWAS. COPD-related data can be downloaded from https://www.finngen.fi/en; physical activity and lung function-related data can be downloaded from https://www.ebi.ac.uk/gwas/

http://creativecommons.org/licenses/by-nc/4.0/

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

WHAT IS ALREADY KNOWN ON THIS TOPIC

  • Smoking, air pollution, dust inhalation and some hereditary factors are established risk factors of chronic obstructive pulmonary disease (COPD). Individuals with COPD spend more time in sedentary behaviours and less time in moderate-to-vigorous physical activity (PA), but it is unclear if these factors precede (and therefore influence) the occurrence of COPD.

WHAT THIS STUDY ADDS

  • This is the first Mendelian randomisation study on leisure sedentary behaviours and PA and COPD. The study indicates that leisurely sedentary behaviour, particularly television watching, increases the odds of COPD, even after adjusting for smoking factors. Additionally, higher levels of PA are associated with higher indices of lung function, notably reflected in increased forced vital capacity measurements.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

  • Our findings provide direct evidence that leisurely sedentary behaviours increase the odds of COPD. This insight is crucial for enhancing screening and individualised treatment strategies for this disease.

Introduction

Chronic obstructive pulmonary disease (COPD) is now one of the three leading causes of death worldwide, and lower socioeconomic status is associated with an increased risk of death from COPD.1 2 The prevalence of COPD is expected to increase in the coming decades due to rising smoking rates in developing countries and ageing populations in high-income countries.3 4 By 2060, it is anticipated that more than 5.4 million individuals will succumb annually to COPD and its associated conditions.5 Environmental factors, such as smoking and air pollution, are major contributors to COPD, and genetic predispositions, such as mutations in the SERPINA1 gene that cause α-1-antitrypsin deficiency, are also implicated.5

The relationship between lifestyle and COPD is currently unclear. Understanding modifiable risk factors, such as leisure sedentary behaviours (LSB) and physical activity (PA), is crucial for developing personalised treatment approaches and reducing healthcare costs. LSB has established causal associations with type 2 diabetes,6 lung cancer,7 ischaemic stroke8 and COVID-19.9 Sedentary behaviour,10 defined as ‘any waking behaviour characterised by an energy expenditure ≤1.5 METs while in a sitting, lying, or reclining posture’, is adverse respiratory outcomes.11–13 Recent studies indicate a high prevalence of physical inactivity among individuals with chronic airway diseases. This lack of PA is associated with poorer clinical outcomes and a diminished quality of life.14 More than 60% of the day was spent sitting or lying down for people with COPD.15 PA can mitigate the risk of developing COPD.15 While existing studies highlight that individuals with COPD are often more sedentary due to their respiratory condition, the broader implications of sedentary behaviour on respiratory health, particularly among those not diagnosed with COPD, remain less explored. In light of this, our study aims to contribute to the understanding of this area by examining if long-term sedentary behaviour in healthy individuals might be associated with new occurrences of COPD.

Mendelian randomisation (MR) uses genetic variants as instrumental variables (IVs) to infer causal relationships in genetic epidemiology.16 17 This method is based on the principle of random allelic assignment of single nucleotide polymorphisms (SNPs) during gametogenesis, guided by three principles: robust association of SNPs with exposure, independence from confounders, and association with outcome through exposure alone.16 18 MR is particularly useful for studying PA and LSB, providing valuable insights despite methodological limitations and the need for genetic homogeneity in populations.

This study investigated the causal relationship between LSB and PA on COPD and lung function using MR in conjunction with data from the latest genome-wide association study (GWAS).19 GWAS is a research approach used to identify genomic variants that are statistically associated with a risk for a disease or a particular trait.20 The method involves surveying the genomes of many people, looking for genomic variants that occur more frequently in those with a specific disease or trait compared with those without the disease or trait. Once such genomic variants are identified, they are typically used to search for nearby variants that contribute directly to the disease or trait. By incorporating GWAS data, the study was able to identify relevant genetic variants associated with LSB, PA and COPD, which were then analysed by MR to assess causality. This integrative approach provides a framework to elucidate the potential causal associations of LSB and PA with COPD and lung function.

Methods

Study overview

To investigate the causal relationship between LSB and PA in COPD, this study first used a two-sample MR method. Two-sample MR refers to the application of MR methods to summary association results estimated in non-overlapping sets of individuals.21 SNPs selected as IVs for this study were sourced from publicly available GWAS summary data related to exposures and outcomes of interest. In this design, behaviours such as television watching, computer use and driving were considered genetic variant indicators for LSB, while moderate-to-vigorous PA (MVPA) served as genetic variant indicators for PA (online supplemental additional file 2, figure 1). Of particular note, LSB is closely related to smoking, which may jointly influence the risk of COPD. Therefore, our primary analysis included multivariable MR (MVMR). MVMR uses genetic variants associated with multiple related exposures to estimate the effect of each exposure on a single outcome.22 This method allows us to simultaneously account for interactions between LSB and smoking, providing a more comprehensive and robust inference of causality.

Figure 1

Inverse variance weighted results—evaluating the causal relationship between leisure sedentary behaviours, physical activity and COPD. COPD, chronic obstructive pulmonary disease; MVPA, moderate-to-vigorous intensity physical activity.

GWAS summary data for exposures

In this study, self-reported MVPA data were collected from the largest and most recent GWAS dataset, which included 608 595 Europeans.23 The assessment of PA levels was based on participants’ self-reports of engaging in MVPA, including activities like swimming and jogging. A binary classification system was applied to categorise individuals as ‘active’ or ‘inactive’. Participants who reported a minimum of 20 min of MVPA per week were classified as ‘active’. In contrast, those who did not meet this threshold were classified as ‘inactive’. This large data set allowed for a refined understanding of the genetic factors influencing PA behaviours, despite the inherent limitations of self-reported data. Additive genetic models were adjusted for family relatedness, age, age-squared, principal components reflecting population structure, and other study-specific covariates.

The data for LSB were obtained from a GWAS that was conducted on 422 218 participants of European ancestry from the UK Biobank.24 In the first questionnaire, participants were asked how many hours in a typical day they spent watching television, using a computer and driving. These questions were designed to assess participants’ habitual LSB without specifying a particular timeframe. The average age of participants was 57.4 years (±SD 8.0). The average reported daily leisure time spent watching television was 2.8 hours (1.5), computer use was 1.0 hour (1.2) and driving was 0.9 hours (1.0).

GWAS summary data for MVMR

Data on daily smoking habits for this study were obtained from the GWAS and Sequencing Consortium of Alcohol and Nicotine use (GSCAN), which includes approximately 337 000 participants.25 This dataset aims to determine the average number of cigarettes smoked per day by current and former smokers, including all types of cigarettes, such as hand-rolled and factory-made cigarettes. To determine daily cigarette consumption, the contributing studies mainly used direct questioning methods, asking questions such as ‘How many cigarettes do you usually smoke per day?’ or ‘On average, how many cigarettes have you smoked per day in the past?’. GSCAN also provides links to additional data and publications for more in-depth information.26

GWAS summary data for outcomes

GWAS data for COPD

The GWAS data for COPD obtained from the FinnGen consortium27 (https://www.finngen.fi/en) included 16 410 cases and 283 589 controls. This data set included cases with a diagnosis of emphysema. To investigate the effect of PA and LSB on COPD risk in relation to age of onset, analyses were performed on two different subsets within the FinnGen cohort. These subsets were differentiated according to the age of COPD onset: one dataset focused on early-onset COPD (≤65 years, including 7371 cases and 326 794 controls) and the other focused on late-onset COPD (>65 years, including 9334 cases and 326 794 controls). The study also included data on COPD-related hospitalisations involving 12 419 cases and 296 735 controls. Detailed information, including ICD codes for COPD and other specifics, is available in online supplemental additional file 1, table S1.

GWAS data for lung function

Lung function data were obtained from the SpiroMeta Consortium,28 which includes GWAS data for forced expiratory volume in the 1 s (FEV1), forced vital capacity (FVC), the ratio between FEV1 and FVC (FEV1/FVC) and peak expiratory flow (PEF). Data for FEV1, FVC and FEV1/FVC were obtained from 79 055 participants across multiple studies within the SpiroMeta Consortium. Of these participants, PEF data were available for a subset of 24 218 individuals. To avoid overlap with the exposure cohort, we did not use data from the UK Biobank. In this study, lung function was measured before inhaler use and Z-scores were used to assess the association between genetic variants and lung function. The average reported FEV1 was 2.98 L (0.93), FVC was 3.75 L (1.21) and FEV1/FVC was 0.79 (0.1).29 Detailed data on all exposures and outcomes are presented in table 1.

Table 1

Details of GWAS data in the Mendelian randomisation study

Selection of IVs

Several measures were taken to ensure the quality of the SNPs in our study. The IVs satisfied the following conditions: (1) the significance level of the SNP was defined as p<5×10–8 to satisfy the relevant assumption. (2) SNPs in linkage disequilibrium30 (R2≥0.001 and within 10 mb) were removed. (3) SNPs with F-statistic less than 10 were excluded.

In the end, we obtained 16 SNPs for MVPA, 95 SNPs for leisure television watching, 23 SNPs for computer use and 4 SNPs for driving (online supplemental additional file 1, tables S2–S5).

Statistical analyses

Three MR analysis methods were used in the two-sample MR analysis: inverse variance weighted (IVW), MR-Egger and weighted median (WM). IVW was the primary approach, chosen for its consistency when all instruments meet MR assumptions.31 We chose the random effects variant of IVW, which accounts for heterogeneity across IVs.18 32 MR-Egger and WM methods were used as supplementary analyses, with MR-Egger accounting for heterogeneity effects18 and WM providing estimates under the condition that a majority of instruments meet MR assumptions.33 The type 1 error rate was set to 0.0016 to account for multiple comparisons (0.05/32), though p values <0.05 were considered nominally significant.

Sensitivity analyses

In our study, we evaluated the heterogeneity of the IVW model using Cochran’s Q test, with p<0.05 in the Q test indicating heterogeneity. We used the MR-Egger intercept test to monitor directional horizontal pleiotropy. Leave-one-out analyses were performed to assess whether a specific causal association was strongly driven by a single SNP. Scatter plots were used to detect potential bias. Because pleiotropy was present in some results, we conducted radial MR to identify and deal with outlier SNPs that could cause bias.34 IVs used in the two-sample MR analysis, after excluding all outlier SNPs, are detailed in online supplemental additional file 1, table S6. In our MVMR analysis, selected genetic instruments for each exposure were concurrently regressed against the outcome, weighted by the outcome’s inverse variance.

All analyses were performed with R (V.4.3.2) and TwoSampleMR (V.0.5.8). MR estimates are presented as ORs and beta coefficients for categorical and continuous variables, respectively, including 95% CIs for both.

Results

MR estimates

MR analysis of PA and LSB with COPD

Our analysis suggests a possible association between MVPA and reduced odds of hospitalisation for COPD (OR=0.64, 95% CI 0.43 to 0.95; p=0.027) (figure 1). However, this association was not statistically significant after adjustment for multiple comparisons using the Bonferroni correction. Due to the stringent nature of this correction and inconsistent results between different MR methods (online supplemental additional file 1, table S7), these findings should be interpreted with caution. There was no evidence of a causal association between MVPA and early-onset COPD (OR=0.75, 95% CI 0.44 to 1.27; p=0.284), late-onset COPD (OR=0.72, 95% CI 0.46 to 1.14; p=0.156) and COPD (OR=0.70, 95% CI 0.49 to 1.01; p=0.056) (online supplemental additional file 1, table S8).

The IVW analysis suggests a possible association between increased leisure-time television watching and increased odds of COPD (OR=2.25, 95% CI 1.78 to 2.86; p=1.95×10–11), early-onset COPD (OR=2. 30, 95% CI 1.60 to 3.30; p=6.39×10–6), late-onset COPD (OR=2.16, 95% CI 1.64 to 2.84; p=3.12×10–8) and COPD hospitalisation (OR=2.02, 95% CI 1.60 to 2.55; p=4.68×10–9). However, significant heterogeneity was observed in the outcomes for COPD (p=0.0006) and early-onset COPD (p=0.0005) (online supplemental additional file 1, table S7). The MR-Egger intercept test did not indicate the presence of horizontal pleiotropy. After outlier correction using radial MR analysis, the association remained notable for COPD (OR=2.25, 95% CI 1.84 to 2.75; p=2.38×10–15) and early-onset COPD (OR=2.11, 95% CI 1.56 to 2.85; p=1.06×10–6) (online supplemental additional file 1, table S8). The WM analysis yielded similar estimates, with increased odds for COPD (OR=2.40, 95% CI 1.78 to 3.23; p=9.26×10–9), early-onset COPD (OR=1.98, 95% CI 1.27 to 3.07; p=0.002), late-onset COPD (OR=2.03, 95% CI 1.40 to 2.95; p=1.99×10–4) and COPD hospitalisation (OR=1.78, 95% CI 1.30 to 2.44; p=3.35×10–4) (online supplemental additional file 1, table S8). The MR-Egger analysis showed a consistent trend, although not significant, except for early-onset COPD. The IVW analysis also suggests a possible association between driving and increased odds of early-onset COPD, although this finding is of nominal significance and warrants further investigation (OR=4.57, 95% CI 1.27 to 16.40; p=0.02). This should be interpreted with caution given the stricter significance criteria in our analysis. In the MR analysis, we did not find a potential causal association between driving and COPD (OR=1.77, 95% CI 0.38 to 8.29; p=0.468), late-onset COPD (OR=0.89, 95% CI 0.16 to 4.93; p=0.898) and COPD hospitalisation (OR=1.75, 95% CI 0.31 to 9.68; p=0.524). Our analysis found no significant evidence of an association between computer use and increased odds of COPD.

MR analysis of PA and LSB in lung function

The IVW analysis suggested a potential association between MVPA and increased FVC (beta=0.27, 95% CI 0.12 to 0.42; p=3.51×10–4) (figure 2). The WM analysis supported this potential association (beta=0.29, 95% CI 0.10 to 0.48; p=0.003) (online supplemental additional file 1, table S8). For FEV1, the WM analysis suggested a potential association with MVPA as indicated by an increase in FEV1 values (beta=0.32, 95% CI 0.13 to 0.50; p=0.001) (online supplemental additional file 1, table S8), a finding somewhat supported by the IVW analysis (beta=0.22, 95% CI 0.05 to 0.40; p=0.013) (figure 2).

Figure 2

Inverse variance weighted results—assessing the causal relationship between leisure sedentary behaviours, physical activity and lung function. MVPA, moderate-to-vigorous intensity physical activity; FEV1, forced expiratory volume in 1 s; FVC, forced vital capacity; PEF, peak expiratory flow.

The IVW analysis suggests a possible causal relationship between leisure television watching and a decrease in both FVC and FEV1. After excluding outliers, the association remained similar, with FVC (beta=−0.19, 95% CI −0.28 to −0.10; p=1.54×10–5) and FEV1 (beta=−0.16, 95% CI −0.25 to −0.08; p=1.21×10–4). WM analysis also supported this potential association, with FVC (beta=−0.2, 95% CI −0.33 to −0.08; p=0.002) and FEV1 (beta=−0.21, 95% CI −0.33 to −0.08; p=0.001) showing similar patterns. However, no clear evidence of an association between computer use, driving and lung function was found in the MR analysis (online supplemental additional file 1, table S8).

Sensitivity analyses

Extensive sensitivity analyses were performed to assess the consistency of the results obtained with the original statistical approaches. The purpose of this approach was to verify that the results were not unduly influenced by specific assumptions or methodological choices, thereby increasing the reliability of our conclusions. All IV F-statistics were greater than 10 (online supplemental additional file 1, tables S2–S5), indicating no weak instrument bias.33 P values for all MR-Egger intercept tests were greater than 0.05, except for estimates of p<0.05 for computer use and FVC, indicating no horizontal pleiotropy (online supplemental additional file 1, tables S7–S8). However, heterogeneity was observed in some Cochran Q test analyses, including MVPA and early-onset COPD (Q=19.74, p=0.049), television watching and COPD (Q=114.5, p=6.26×10–4), television watching and early-onset COPD (Q=115.3, p=5.24×10–4), television watching and FVC (Q=153.93, p=1.89×10–6), television watching and FEV1 (Q=150.56, p=4.33×10–6), and some outcomes for computer use, driving, COPD and lung function (online supplemental additional file 1, table S7). Radial MR was used to detect outliers (figure 3). After outlier removal, MR analysis was reapplied for causal analysis. In addition, scatter plots showed symmetry (online supplemental additional file 2, figure 2). Leave one out analysis confirmed that no single SNP disproportionately influenced the results (online supplemental additional file 2, figure 3). Therefore, the study results are considered consistent.

Figure 3

Radial MR analysis results; (A) effects of television watching on COPD; (B) effects of television watching on early-onset COPD; (C) effects of television watching on FVC; (D) effects of television watching on FEV1. COPD, chronic obstructive pulmonary disease; FEV1, forced expiratory volume in 1 s; FVC, forced vital capacity; IVW, inverse variance weighted; MR, Mendelian randomisation.

Multivariable MR

Further multivariable analyses were performed on the main positive findings mentioned earlier (online supplemental additional file 1, tables S9–S10). After adjustment for daily smoking, a potential association between television watching and COPD was still observed (OR=1.87, 95% CI 1.46 to 2.4; p=9.19×10–7) (figure 4). This association was evident in both early-onset COPD (OR=1.85, 95% CI 1.29 to 2.66; p=8.68×10–4) and late-onset COPD (OR=1.83, 95% CI 1.35 to 2.46; p=7.85×10–5). In addition, the association between television watching and COPD hospitalisation remained after adjusting for the effect of daily smoking (OR=1.64, 95% CI 1.26 to 2.13; p=2.19×10–4). There was no potential causal relationship between MVPA and COPD, early-onset COPD, late-onset COPD or COPD hospitalisation after adjustment for daily smoking (figure 4).

Figure 4

Multivariable Mendelian randomisation analysis of leisure sedentary behaviours and MVPA with COPD, adjusting for cigarette smoking. COPD, chronic obstructive pulmonary disease; MVPA, moderate-to-vigorous intensity physical activity.

For lung function, a modest association between television watching and FVC remained after adjustment for daily smoking (beta=−0.12, 95% CI −0.23 to −0.003 (three decimal places retained for precision); p=0.045) (figure 5). This should be interpreted with caution given the stricter significance criteria in our analysis. However, we did not find sufficient evidence to suggest a direct association between television watching and FEV1 (beta=−0.09, 95% CI −0.21 to 0.02; p=0.098). The potential causal relationship between MVPA and FEV1 (beta=0.15, 95% CI 0.01 to 0.29; p=0.037) and FVC (beta=0.23, 95% CI 0.09 to 0.36; p=9.24×10–4) remained after adjustment for daily smoking.

Figure 5

Multivariable Mendelian randomisation analysis of leisure sedentary behaviours and MVPA on lung function, adjusting for cigarette smoking. FEV1, forced expiratory volume in 1 s; FVC, forced vital capacity; MVPA, moderate-to-vigorous intensity physical activity.

Discussion

In this study, we used large-scale GWAS data and multiple MR analysis methods to investigate the potential causal relationships between PA and LSB and COPD prevalence and lung function. Our findings indicate an association between prolonged television watching within LSB and higher odds of COPD. Specifically, for each increment of 2.8 hours in television watching, the odds of COPD were 2.25 times greater (OR=2.25; 95% CI 1.84 to 2.75; p=2.38×10–15). For early-onset COPD, the odds were 2.11 times greater (OR=2.11; 95% CI 1.56 to 2.85; p=1.06×10–6), and for late-onset COPD, the odds were 2.16 times greater (OR=2.16; 95% CI 1.64 to 2.84; p=3.12×10–8). Similarly, the odds of hospitalisation for COPD were 2.02 times greater with increased television watching (OR=2.02; 95% CI 1.6 to 2.55; p=4.68×10–9). These relationships persisted even after adjusting for daily smoking. Moreover, an additional 2.8 hours of television watching was significantly associated with lower FEV1 and FVC, showing an average decrease of 148.8 mL and 212.8 mL. Conversely, our analysis suggests a potential association between increased MVPA and higher indices of lung function, reflected by an increase of 302.4 mL in FVC.

Prolonged sedentary behaviour is associated with adverse health outcomes, independent of PA.35 In a prospective cohort study, Waschki et al suggested that objectively measured PA was identified as the strongest predictor of all-cause mortality in patients with COPD.36 A cohort study37 showed that patients who spent more time watching television had higher COPD-related mortality. In a population-based cohort study of 12 283 individuals,38 television watching was negatively associated with FEV1, whereas vigorous leisure time activity was positively associated with FEV1. While some studies have conflicting results, da Silva et al found that spending more time in sedentary activities, such as playing video games, or using a computer, may not impair lung function.39 Several studies have shown that increased PA is associated with improved COPD-related mortality and reduced risk of hospitalisation.40 Sedentary behaviour was significantly associated with an increased risk of COPD,41 which is consistent with our findings. However, because this was a cross-sectional study and the data were collected retrospectively, a causal relationship could not be established. To our knowledge, this is the first MR study to systematically evaluate the causal role of PA and LSB in COPD.

MVPA has a beneficial effect on lung function at the population level.42 A key mechanism is the association between greater overall strength and greater respiratory muscle strength, leading to improved lung function.43 44 Training has been shown to improve the kinetic response of oxygen consumption, carbon dioxide production, minute ventilation and heart rate during moderate exercise.45 Furthermore, exercise training has been observed to significantly reduce high-sensitivity C reactive protein (CRP) levels in patients, indicating a reduction in systemic inflammation.46

COPD is directly related to the prevalence of smoking.47 48 In particular, former smokers have a significantly higher incidence of COPD compared with non-smokers.49 However, there is growing evidence that COPD is characterised by a high proportion of non-smokers in the patient population, as high as 25%–45%.50 This suggests that there are important risk factors for COPD other than smoking. This is consistent with our findings, which suggest a potential causal relationship between television watching and COPD even after adjusting for the important factor of smoking. This finding underscores the importance of considering other lifestyle factors, such as television watching, in addition to smoking when assessing the risk of COPD.

Why does prolonged TV watching increase the risk of COPD? The key lies in the association between sedentary behaviour and inflammation. Research has shown that sedentary behaviour can significantly increase levels of inflammatory markers such as CRP in the body.51 This change can be triggered even by a significant reduction in daily activities over a short period of time. Specifically, in a study involving 22 overweight, pre-diabetic older adults, participants were found to have significantly elevated plasma concentrations of tumour necrosis factor-alpha (TNF-α), interleukin 6 (IL-6) and CRP after a 14-day period of significant activity reduction, which remained above baseline after the recovery period.52 In addition, airway inflammation is positively correlated with sedentary time, and low levels of airway inflammation are associated with high levels of PA.53 54 The progression of COPD is strongly associated with remodelling and stenosis of the small airways, as well as with destruction of the lung parenchyma, pathological changes caused by chronic inflammation in the lung periphery.55 These findings highlight the association between LSB and inflammation and its possible impact on COPD risk, and more in-depth research is needed on this association and its underlying biological mechanisms.

In addition, a meta-analysis of 25 788 participants indicated that sedentary behaviour was positively associated with the development of sarcopenia in the elderly population,56 and that there was a significant association between inflammatory markers such as TNF-α and IL-6, and decreased muscle mass and strength.57 Sarcopenia not only leads to decreased respiratory muscle mass and strength,58 but may also cause a range of respiratory declines including decreased chest wall expandability, decreased respiratory muscle strength, decreased expiratory flow and reduced range of motion of the costovertebral joints, which in turn affects respiratory mechanics during exercise. Despite the lack of direct evidence linking sarcopenia directly to the development of COPD, the prevalence of high-incidence sarcopenia in patients with COPD suggests that,59 60 in addition to traditional risk factors such as cigarette smoking, lifestyle factors (eg, sedentary behaviours) and their triggered changes in systemic inflammation and muscle function pose a complex impact on COPD risk.

This study found no association between computer use and risk of COPD, and driving was only nominally significant in early-onset COPD cases. This could be partly explained by the fact that watching television generally involves less mobility than using a computer or driving, which require hand movements and coordinated body movements, respectively. In addition, television watching is often associated with unhealthy eating habits, such as increased snacking and drinking, which may further influence COPD risk. Recent research supports this link,61 suggesting that such habits, characterised by high consumption of processed foods and sugary beverages, may contribute to systemic inflammation and oxidative stress, potentially exacerbating the risk and severity of COPD.

MR provides a unique perspective for studying the effects of lifetime exposures on health outcomes, with the present study highlighting the potential causal association of LSB and MVPA with lung function and COPD occurrence. However, this study has several limitations. First and foremost, our research sample consisted exclusively of participants of European ancestry, which limits the generalisability of our findings to other ancestry groups. Second, our analysis relied on self-reported data for MVPA and LSB, which may introduce measurement bias. In particular, our study did not include light-intensity PA, which may be positively associated with the prevalence of COPD. In addition, comprehensive sociodemographic information, including important factors such as sex, age, ethnicity and education level, was not available. This lack of data limits the scope of our conclusions. Finally, the GWAS data did not specify types of PA, so further research is needed to determine which activities are most beneficial for lung function.

Conclusion

In conclusion, our study suggests a potential association with LSB, particularly television watching, is associated with higher odds of COPD and lower indices of lung function as measured continuously, including FEV1 and FVC. Conversely, an increase in MVPA is associated with higher indices of lung function, particularly reflected in increased FVC levels.

Data availability statement

Data are available in a public, open access repository. All data are from the publicly available GWAS. COPD-related data can be downloaded from https://www.finngen.fi/en; physical activity and lung function-related data can be downloaded from https://www.ebi.ac.uk/gwas/

Ethics statements

Patient consent for publication

Ethics approval

This study involves human participants but was exempted by the Ethics Committee of Guang’anmen Hospital, China Academy of Traditional Chinese Sciences (2023-105-MC). Participants gave informed consent to participate in the original study before taking part.

Acknowledgments

We want to acknowledge the participants and investigators of the FinnGen study, UK Biobank, Spirometa consortium, GSCAN and GWAS Catalogue.

References

Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

Footnotes

  • LX and WL contributed equally.

  • Collaborators Not applicable.

  • Contributors LX and WL statistically analysed and drafted the manuscript; FL and XC collected the data and created the tables and figures; YH and YX conducted a literature study. YF and LF participated in the design of the study. Additionally, as guarantors, YF and LF accept full responsibility for the work and the conduct of the study, had access to the data, and controlled the decision to publish. They ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

    All authors read and approved the final version.

  • Funding This project received funding from the Science and Technology Innovation Project of China Academy of Chinese Medical Sciences (CI2021A03001).

  • Competing interests None declared.

  • Patient and public involvement In our Mendelian randomisation study, the development of the research question and outcome measures was informed by scientific literature and genetic data analysis, rather than direct input from patients. Given the study's reliance on pre-existing genetic data, patients were not involved in its design, recruitment or conduct. Consequently, direct dissemination of results to study participants is not applicable. Additionally, as our study did not entail patient interventions, the assessment of intervention burdens by patients was not a component of our research. While there were no patient advisers involved due to the nature of our methodology, we extend our gratitude to the contributors of the FinnGen study, UK Biobank, Spirometa consortium, GSCAN and the GWAS Catalogue for their invaluable data contributions.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.