Impact of ambient air pollution on physical activity among adults: a systematic review and meta-analysis

Abstract

Aims:

This study systematically reviewed literature regarding the impact of ambient air pollution on physical activity among children and adults.

Methods:

Keyword and reference search was conducted in PubMed and Web of Science to systematically identify articles meeting all of the following criteria – study designs: interventions or experiments, retrospective or prospective cohort studies, cross-sectional studies, and case-control studies; subjects: adults; exposures: specific air pollutants and overall air quality; outcomes: physical activity and sedentary behaviour; article types: peer-reviewed publications; and language: articles written in English. Meta-analysis was performed to estimate the pooled effect size of ambient PM_2.5 air pollution on physical inactivity.

Results:

Seven studies met the inclusion criteria. Among them, six were conducted in the United States, and one was conducted in the United Kingdom. Six adopted a cross-sectional study design, and one used a prospective cohort design. Six had a sample size larger than 10,000. Specific air pollutants assessed included PM_2.5, PM₁₀, O₃, and NO_x, whereas two studies focused on overall air quality. All studies found air pollution level to be negatively associated with physical activity and positively associated with leisure-time physical inactivity. Study participants, and particularly those with respiratory disease, self-reported a reduction in outdoor activities to mitigate the detrimental impact of air pollution. Meta-analysis revealed a one unit (μg/m³) increase in ambient PM_2.5 concentration to be associated with an increase in the odds of physical inactivity by 1.1% (odds ratio = 1.011; 95% confidence interval = 1.001, 1.021; p-value < .001) among US adults.

Conclusions:

Existing literature in general suggested that air pollution discouraged physical activity. Current literature predominantly adopted a cross-sectional design and focused on the United States. Future studies are warranted to implement a longitudinal study design and evaluate the impact of air pollution on physical activity in heavily polluted developing countries.

Keywords

air pollution exercise sedentary lifestyle review meta-analysis

Introduction

The fossil fuel economy, population growth and industrialisation have driven the air pollution worldwide to an unprecedented level.^1–5 There has been mounting evidence linking elevated air pollution level to various adverse health outcomes, including cardiovascular disease, stroke, lung cancer, and respiratory disease such as asthma.⁶ The US Environmental Protection Agency (EPA) has identified and regulated the emission levels of six major air pollutants in accordance with human health-based and/or environmentally based scientific criteria, including carbon monoxide (CO), lead, nitrogen oxides (NO_x), ground-level ozone (O₃), particulate matter (PM), and sulphur oxides.⁷ Common measures for PM include PM₁₀ and PM_2.5, denoting a mixture of solid and liquid particles suspended in the air that are less than 10 and 2.5 µm in diameter, respectively. PM₁₀ and PM_2.5 can be effectively inhaled and deposited in the airway, and sometimes bloodstream, causing health problems.^8,9

Although the adverse effects of air pollution on health outcomes have been extensively documented,⁶ much less is known regarding its impact on people’s health behaviour. Physical inactivity is a leading risk factor for morbidity and mortality worldwide.^10–12 A vast majority of children and adults in both developed and developing countries fall short of the guidelines-recommended physical activity level.¹³ Ambient air pollution may discourage people from engaging in regular physical activity through several mechanisms. Exposure to air pollutants has been linked to decreased lung function, elevated blood pressure, and other cardiovascular and respiratory symptoms,^14–17 resulting in impaired exercise capacity and performance.^18–20 Smog appearance could discourage people from engaging in outdoor activities.²¹ Media alerts of air qualities to inform the public about harmful air pollution may alter people’s decisions on physical activity.^22,23

Promoting physical activity serves as a key strategy to maintain and improve population health. However, air pollution is a profound side product of economic growth and social evolution, and the war against air pollution is expected to be long-lasting in many developing nations. Minimising the negative impact of air pollution on physical activity and identifying innovative means to promote physical activity under compromised air quality calls for evidence-based research. Findings from this review can be informative to policy makers and stakeholders in an effort to design and implement effective interventions that mitigate the negative influence of ambient air pollution on people’s health behaviour.

This study, to our knowledge, is the first that systematically reviewed existing literature regarding the impact of ambient air pollution on physical activity among children and adults. It synthesised scientific evidence linking specific air pollutant and/or overall air quality to modifications in health behaviour, and quantified its relationship using meta-analysis. It also identified the limitations and gaps in this field that warranted future research.

Methods

Systematic review and meta-analysis procedures were conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses.²⁴

Study selection criteria

Studies that met all of the following criteria were included in the review: (1) study designs: interventions or experiments, retrospective or prospective cohort studies, cross-sectional studies, and case-control studies; (2) subjects: adults; (3) exposures: specific air pollutants (e.g., PM₁₀, PM_2.5, O₃, and NO_x) and overall air quality; (4) outcomes: physical activity and sedentary behaviour; (5) article types: peer-reviewed publications; (6) time window of search: from the inception of an electronic bibliographic database to 28^th February 2017; and (7) language: articles written in English.

Studies that met any of the following criteria were excluded from the review: (1) Studies that focused on children aged 0–17 years; (2) case reports; (3) studies that incorporated no outcome pertaining to physical activity and sedentary behaviour; (4) articles not written in English; or (5) letters, editorials, study/review protocols, or review articles.

Search strategy

Keyword search was performed in two electronic bibliographic databases: PubMed and Web of Science. The search algorithm included all possible combinations of keywords from the following two groups: (1) ‘air pollutant’, ‘air pollutants’, ‘air pollution’, and ‘air quality’, and (2) ‘physical activity’, ‘physical activities’, ‘physical inactivity’, ‘sedentary behavior’, ‘sedentary behaviors’, ‘sedentary lifestyle’, ‘exercise’, ‘exercises’, ‘motor activity’, ‘motor activities’, ‘sport’, ‘sports’, ‘physical fitness’, and ‘physical exertion’. Medical subject headings of ‘air pollutants’, ‘air pollution’, ‘exercise’, and ‘sports’ were used in PubMed searching. Titles and abstracts of the articles identified through the keyword search were screened against the study selection criteria. Potentially relevant articles were retrieved for evaluation of the full text. S.Z. and M.J. independently conducted title and abstract screening and identified potentially relevant articles for full-text review. Inter-rater agreement was assessed using the Cohen’s kappa (κ = 0.77). Discrepancies were resolved through discussion. Subsequently, S.Z. and M.J. independently conducted full-text review, and R.A., S.Z., and M.J. jointly decided the final pool of articles included in the review.

A reference list search (i.e. backward reference search) and cited reference search (i.e. forward reference search) were conducted based on the full-text articles meeting the study selection criteria that were identified from the keyword search. Articles identified from the backward and forward reference search were further screened and evaluated using the same study selection criteria. Reference searches were repeated on all newly identified articles until no additional relevant article was found.

Data extraction and preparation

A standardised data extraction form was used to collect the following methodological and outcome variables from each included study: first author, year of publication, country, sample size, age range, proportion of males, health/disease status of sample, data source, air pollutants, air pollution/quality level, measures of physical activity and/or sedentary behaviour, statistical model, and key results.

Meta-analysis

Meta-analysis was performed to estimate the pooled effect size of ambient PM_2.5 air pollution on physical inactivity among US adults aged 18 years and above. No other exposure (i.e. air pollutant) and outcome were incorporated in the meta-analysis because no two studies reported the same exposure and outcome (except for the three studies that examined the impact of ambient PM_2.5 concentration on physical inactivity). Study heterogeneity was assessed using the I² index. The level of heterogeneity represented by the I² index was interpreted as modest (I² ≤ 25%), moderate (25% < I² ≤ 50%), substantial (50% < I² ≤ 75%), or considerable (I² > 75%). A fixed-effect model would be estimated when modest to moderate heterogeneity was present, and a random-effect model would be estimated when substantial to considerable heterogeneity was present. Publication bias was assessed by the Begg’s and Egger’s tests. All statistical analyses were conducted using the Stata 14.2 SE version (StataCorp, College Station, TX). Specific STATA commands included ‘metan’ and ‘metabias’. All analyses used two-sided tests, and p-values less than .05 were considered statistically significant.

Study quality assessment

We used the National Institutes of Health’s Quality Assessment Tool for Observational Cohort and Cross-Sectional Studies to assess the quality of each included study.²⁵ This assessment tool rates each study based on 14 criteria. For each criterion, a score of one was assigned if ‘yes’ was the response, whereas a score of zero was assigned otherwise (i.e. an answer of ‘no’, ‘not applicable’, ‘not reported’, or ‘cannot determine’). A study-specific global score, ranging from zero to 14, was calculated by summing up scores across all criteria. Study quality assessment helped measure strength of scientific evidence, but was not used to determine the inclusion of studies.

Results

Study selection

Figure 1 shows the study selection flow chart. We identified a total of 1,202 articles through keyword and reference search, including 864 articles from PubMed and 338 articles from Web of Science. After removing duplicates, 1,095 articles underwent title and abstract screening, in which 1,084 articles were excluded. The remaining 11 articles were reviewed in full text against the study selection criteria. Of these, five articles were excluded. Reasons for exclusion included: three articles were commentaries rather than original studies,^23,26,27 one article was not written in English,²⁸ and one article exclusively focused on children.²⁹ An additional relevant article was identified from forward and backward reference search.³⁰ In total, seven articles were included in the review.^{21,22,30–34}

Basic characteristics of the included studies

Table 1 summarises basic characteristics of the seven studies included in the review. Six were conducted in the United States, and the remaining one was conducted in the United Kingdom. All studies were published within the past decade (two in 2009, two in 2012, and one each in 2014, 2015, and 2016). Six of the seven studies adopted a cross-sectional study design, and only one adopted a prospective cohort study design.³¹ Sample sizes were generally large but varied substantially across studies. Except for the prospective cohort study that had a sample size of 73,³¹ all the six cross-sectional studies had a sample size larger than 10,000. The two studies of the largest scale were An and Xiang³² and Roberts et al.,²¹ with a sample size of over 2 million and 0.3 million, respectively. Five studies focused on the general adult population aged 18 years and above, and one study exclusively examined older adults. Males and females were largely equally distributed in all studies. The prospective cohort study exclusively focused on patients with chronic obstructive pulmonary disease (COPD),³¹ whereas none of the six cross-sectional studies examined disease-specific population. The Behavioral Risk Factor Surveillance System (BRFSS) was the most common data source used by four of the seven included studies.^21,22,30,32 The BRFSS is an annually repeated cross-sectional telephone-based survey that collects state data about US residents regarding their health-related risk behaviours, chronic health conditions, and use of preventive services.³⁶ The alternative waves (i.e. years) of the BRFSS data used included 2001,^21,22,30 and 2003–2011.³² Besides the BRFSS, the other national health survey data used was the National Health and Nutrition Examination Survey (NHANES).³⁴ Specific air pollutants assessed by the included studies comprised PM_2.5, ^21,30,32,33 PM₁₀,^21,31 O₃,^21,31,33 and NO_x.³³ Two studies did not examine any specific air pollutant but focused on overall air quality: Wells et al.³⁴ examined self-perceived ambient air quality, and Wen et al.²² investigated a composite air quality index (AQI) reported in media alerts. A variety of statistical tests and models were applied across the studies, including Wilcoxon test, analysis of variance (ANOVA), analysis of covariance (ANCOVA), multiple linear regression, multiple logistic regression, population-attributable fraction, inverse distance weighting, empirical Bayesian kriging, generalised estimating equation, and multilevel logistic regression. Four of the seven included studies adjusted for some individual sociodemographics in statistical analysis, and one of them in addition adjusted for some weather-related characteristics (e.g. daytime temperature, precipitation, and sunlight).

Table 1

Basic characteristics of the studies included in the review

Study ID	First author (year)	Country	Study design	Sample Size	Age (years)	Male (%)	Health status	Air pollution level	Data source
1	Alahmari (2016)	UK	Prospective cohort study	73	71.1 ± 8.7	69.86	Chronic obstructive pulmonary disease (COPD) patients	PM₁₀ and O₃ daily mean concentrations were 17 µg/m³ and 27 µg/m³ respectively.	Daily data for atmospheric PM₁₀ and O₃ were obtained for Bloomsbury Square, Central London from the Air Quality Information Archive databases. Seventy-three stable COPD patients recorded on daily diary cards worsening of respiratory symptoms, peak expiratory flow rate, hours spent outside the home and the number of steps taken per day.
2	An (2015)	US	Cross-sectional study	2,381,292	18 and above	50.54		Mean and standard deviation of PM_2.5 concentration across county and study period (2003–2011) were 11.90 µg/m³ and 3.30 µg/m³, respectively.	Air pollution data came from US Environmental ProtectionAgency (EPA) Air Quality System. Individual-level data came from the Behavioral Risk Factor Surveillance System (BRFSS) 2003–2011 surveys.
3	Roberts (2014)	US	Cross-sectional study	329,628	18 and above	51.71		The mean of PM_2.5, PM₁₀ and O₃ annual mean concentrations were 9.50 µg/m³ (SD = 1.80), 19.52 µg/m³ (SD = 4.16), and 45 ppb (SD = 4.48) for the BRFSS counties, respectively.	Pollutant data came from US Environmental ProtectionAgency (EPA) Air Quality System. Individual-level data came from the Behavioral Risk Factor Surveillance System (BRFSS) 2011 surveys.
4	Hankey (2012)	US	Cross-sectional study	30,007	38 (21–54)	50		Annual-average air pollution exposure for the survey population averaged 49 μg/m³ for NO₂ (IQR = 41–60 μg/m³), 99 μg/m³ for O₃ (86–112 μg/m³, annual average of 1-hr daily maximums), and 22 μg/m³ for PM_2.5 (20–24 μg/m³).	Estimates for air pollution exposure were based on monitoring data (US Environmental Protection Agency (EPA) 2010) for PM_2.5, O₃ and NO_x in 2001. Physical inactivity data came from 2001 Post-Census Regional Travel Survey.
5	Wells (2012)	US	Cross-sectional study	10,898	46.9 (95% CI = 46.3, 47.6)	48.64	27% with respiratory and/or cardiovascular diseases	None	Physical activity data came from the National Health and Nutrition Examination Survey (NHANES) 2007–2010.

Table 2 summarises the outcome measures and key results of the seven studies included in the review. Outcome measures included time spent on physical activity per week,³³ leisure-time physical inactivity (defined as non-participation in any physical activity or exercise other than one’s regular job during the past month),^21,30,32 change in physical activity due to air quality concerns in the past 12 months,^22,34 and number of steps taken per day.³¹ Except for the step count that was measured by a pedometer, all other outcome measures were based on self-report (e.g. multiple-item physical activity questionnaires, daily diary cards, and single interview question). Two key results from the included studies were identified. First, air pollution level in general was found to be negatively associated with physical activity and positively associated with leisure-time physical inactivity. All three studies that assessed leisure-time physical inactivity in relation to ambient PM_2.5 and/or PM₁₀ concentration using data from the BRFSS confirmed such a positive relationship.^21,30,32 Ambient O₃ but not PM₁₀ concentration was found to be associated with reduced time spent outside home per week and daily step count among patients with COPD.³¹ Second, people were found to mitigate the impact of an elevated air pollution level by engaging in less outdoor activities. Over a 10^th of survey participants reported that they spent less time outdoors given self-perceived undesirable air quality.²² People with respiratory disease tended to be more sensitive to change in air quality. The prevalence of change in outdoor activity due to media alerts of AQI was found to be 31% among adults with lifetime asthma and 16% without asthma.²²

Table 2

Measures, statistical methods, and outcomes of the studies included in the review

Study ID	Measurement		Key results	Statistical model
Study ID	Pollutant	Physical activity/inactivity	Key results	Statistical model
1	PM10, O3	Hours spent outside home; number of steps taken per day (measured by pedometer)	(1) Relationship between PM₁₀ (μg/m³) and daily step count: regression coefficient = –5.4, 95% CI = −12.2, 1.3, p-value = 0.112;(2) Relationship between O₃ (μg/m³) and daily step count: regression coefficient = –8.0, 95% CI= −13.5, −2.4, p-value = .005;(3) Relationship between PM₁₀ (μg/m³) and hours spent outdoors: regression coefficient = 2.3 × 10⁻³, 95% CI = −6.5 × 10⁻³, 1.8 × 10⁻³, p-value = .275;(4) Relationship between O₃ (μg/m³) and hours spent outdoors: regression coefficient = –9.9 × 10⁻³, 95% CI = −14.2 × 10⁻³, −5.6 × 10⁻³, p-value < .001	Generalized estimating equation
2	PM2.5	22.83% of the 2003–2011 BRFSS sample did not participate in any physical activity during the past month.	One unit (μg/m³) increase in county monthly average PM_2.5 concentration was found to be associated with an increase in the odds of physical inactivity by 0.46% (OR = 1.0046, 95% CI = 1.0034, 1.0059).	Multilevel logistic regression
3	PM2.5, PM10, O3	24.5% of the 2011 BRFSS sample did not participate in any physical activity during the past month.	(1) A 2.4% relative increase in the odds of physical inactivity per mg/m³ increase of PM_2.5 exposure among the obese respondents (OR = 1.02, 95% CI = 1.00,1.05);(2) An increasing concentration of PM₁₀ among the normal weight respondents was also associated with higher odds of inactivity (odds ratio (OR) = 1.01, 95% confidence interval (CI) = 1.00, 1.02).	Inverse distance weighting, empirical Bayesian kriging
4	PM2.5, O3, NOx	Self-reported physical activity levels averaged 77 min/week. Most (83.5%) survey participants reported being physically inactive (0 min/week), 5.6% reported being insufficiently active (1–150 min/week), and 10.9% reported being active (>150 min/week).	Between-neighbourhood differences in the estimated risk for ischaemic heart disease (IHD) mortality from air pollution were comparable in magnitude (9 more risk for IHD deaths/100,000/year for PM_2.5 and 3 fewer IHD deaths for O₃ in high- vs. low-walkability neighbourhoods), suggesting that population health benefits from increased physical activity in high-walkability neighbourhoods may be offset by adverse effects of air pollution exposure.	Population-attributable fraction
5	Self-perceived ambient air quality	Activity change was determined from the question ‘During the past 12 months, when you thought or were informed air quality was bad, did you do anything differently?’ NHANES respondents could answer ‘yes”, “no” or “never thought/not informed about air quality”.	A total of 1305 (12.0%, 95% CI = 10.9, 13.1) individuals responded that they did something differently due to bad air quality. Among those who reported changing their activity, the most commonly reported change was to spend less time outdoors.	Analysis of variance (ANOVA)
6	Air quality index (AQI): a standardized number and color-coded index that helps people understand the quality of air each day and the associated health risks	Four questions in the 2005 BRFSS: (1) “Please think of the past 12 months. How many times did you reduce or change your outdoor activity level because you thought the air quality was bad or was affecting how well you felt?”; (2) “The government routinely collects information on air quality that may be distributed by local radio, TV and newspapers to help inform the public about air pollution levels. Have you ever heard or read about the air quality index or air quality alerts where you live?”; (3) “Please think of the past 12 months. How many times did you reduce or change your outdoor activity level based on the air quality index or air quality alerts?”; and (4) “Has a doctor, nurse, or other health professional ever told you to reduce your outdoor activity level when the air quality is bad?”	The prevalence of change in outdoor activity due to media alerts of AQI was 31% among adults with lifetime asthma and 16% without asthma. The prevalence of outdoor activity change increased to 75% among those with lifetime asthma and to 68% without asthma, when the combined the effects of media alerts and individual perception were examined. The odds of activity change based on the media alerts was 2.30 (OR = 2.16, 95% CI = 1.61, 2.90) among those with lifetime asthma and 1.72 (OR = 1.72, 95% CI = 1.50, 1.98) without asthma, compared to those unaware of media alerts, after adjustment for demographic variables and covariates.	Wilcoxon test, logistic regression
7	PM2.5	24.9% of the 2001 BRFSS sample did not participate in any physical activity during the past month.	A 10-unit (μg/m3) increase in county annual average PM2.5 concentration was found to be associated with an increase in the odds of physical inactivity by 16% (OR = 1.16, 95% CI = 1.06, 1.27).	Linear regression, analysis of covariance (ANCOVA)

Meta-analysis of the impact of air pollution on physical inactivity

Meta-analysis was conducted to estimate the pooled effect size of ambient PM_2.5 air pollution on physical inactivity among US adults, based on the three studies utilizing data from multiple waves of the BRFSS.^21,30,32 Figure 2 shows the forest plot generated from the meta-analysis. A random-effect model was performed given an I² = 67.5%. A one unit (μg/m³) increase in ambient PM_2.5 concentration was found to be associated with an increase in the odds of physical inactivity by 1.1% (odds ratio = 1.011; 95% confidence interval = 1.001, 1.021; p-value < .001) among US adults. No publication bias was identified as neither of the Egger’s tests or Begg’s tests was statistically significant.

Study quality assessment

Table 3 reports criterion-specific and global ratings from the study quality assessment. The included studies on average scored 10.4 out of 14, with a range from 10 to 13. The major limitations of most studies included in the review pertained to cross-sectional study design, and lack of adjustment for characteristics of neighbourhood built and social environment.

Table 3

Quality assessment of the studies included in the review

Criteria Study ID	1	2	3	4	5	6	7
1. Was the research question or objective in this paper clearly stated?	Yes	Yes	Yes	Yes	Yes	Yes	Yes
2. Was the study population clearly specified and defined?	Yes	Yes	Yes	Yes	Yes	Yes	Yes
3. Was the participation rate of eligible persons at least 50%?	Yes	Yes	Yes	Yes	Yes	Yes	Yes
4. Were all the subjects selected or recruited from the same or similar populations (including the same time period)? Were inclusion and exclusion criteria for being in the study pre-specified and applied uniformly to all participants?	Yes	Yes	Yes	Yes	Yes	Yes	Yes
5. Was a sample size justification, power description, or variance and effect estimates provided?	Yes	Yes	Yes	Yes	Yes	Yes	Yes
6. For the analyses in this paper, were the exposure(s) of interest measured prior to the outcome(s) being measured?	No	No	No	No	No	No	No
7. Was the timeframe sufficient so that one could reasonably expect to see an association between exposure and outcome if it existed?	Yes	No	No	No	No	No	No
8. For exposures that can vary in amount or level, did the study examine different levels of the exposure as related to the outcome (e.g., categories of exposure, or exposure measured as continuous variable)?	Yes	Yes	Yes	Yes	Yes	Yes	Yes
9. Were the exposure measures (independent variables) clearly defined, valid, reliable, and implemented consistently across all study participants?	Yes	Yes	Yes	Yes	Yes	Yes	Yes
10. Was the exposure(s) assessed more than once over time?	Yes	Yes	Yes	Yes	Yes	Yes	Yes
11. Were the outcome measures (dependent variables) clearly defined, valid, reliable, and implemented consistently across all study participants?	Yes	Yes	Yes	Yes	Yes	Yes	Yes
12. Were the outcome assessors blinded to the exposure status of participants?	Yes	Yes	Yes	Yes	Yes	Yes	Yes
13. Was loss to follow-up after baseline 20% or less?	Yes	No	No	No	No	No	No
14. Were key potential confounding variables measured and adjusted statistically for their impact on the relationship between exposure(s) and outcome(s)?	Yes	No	No	No	No	No	No
Sum score	13	10	10	10	10	10	10

Discussion

We systematically reviewed and identified seven studies that assessed ambient air pollution in relation to physical activity among adults. Six were conducted in the United States, and all were published within the past decade. Six adopted a cross-sectional study design, and one used a prospective cohort study design. Six had a sample size larger than 10,000. Specific air pollutants assessed included PM_2.5, PM₁₀, O₃, and NO_x, whereas two studies focused on overall air quality. All studies found air pollution level to be negatively associated with physical activity and positively associated with leisure-time physical inactivity. Study participants, and those with respiratory disease in particular, self-reported a reduction in outdoor activities to mitigate the detrimental impact of air pollution. Meta-analysis found a one unit (μg/m³) increase in ambient PM_2.5 concentration to be associated with an increase in the odds of physical inactivity by 1.1% among US adults.

Ambient air pollution has long been a leading public health concern.⁶ Existing literature predominantly investigated its impacts on disease development,⁶ whereas its influence on health behaviour received much less attention. Scientific evidence regarding the deterrent effect of air pollution on physical activity started to accumulate during the past decade. This study serves as the first known attempt to systematically review and quantitatively synthesise the relevant literature. The strengths of the included studies pertained to large sample size and geographical coverage, reliable measures on ambient air pollutants, and adequate application of statistical modelling. Nevertheless, a few major limitations remain. Only one of the seven studies included in the review adopted a longitudinal study design,³¹ whereas all the others remained cross-sectional, which were subjective to confounding and self-selection bias (e.g. physically less active individuals disproportionally lived in areas with higher air pollution level). Studies using data from the BRFSS were not able to differentiate outdoor from indoor physical activities due to its questionnaire wording.^21,22,30,32 It is possible that a proportion of survey participants mediated the influence of air pollution by switching from outdoor to indoor exercises, but such coping mechanisms were unidentifiable owing to data unavailability. Despite the efforts to adjust for environmental covariates such as weather that could impact physical activity, no study included in the review incorporated measures on built environment such as street connectivity/walkability (except for Hankey et al.³³) and availability/proximity of park and recreational facilities,^35,36 or social environment such as neighbourhood safety and social capital and cohesion.^37,38 Six out of seven studies included in the review were conducted in the United States,^{21,22,30,32–34} whereas no study to date examines ambient air pollution in relation to physical activity in developing countries such as China and India, which have been profoundly impacted by record high air pollution levels. Effect quantification of air pollution on physical activity was limited to one specific air pollutant (PM_2.5), as no other studies shared the same exposure and outcome measure.

Air pollution might affect people’s health behaviour through its detrimental impact on various acute and chronic illnesses such as respiratory and cardiovascular diseases.^14–17 Future studies are warranted to examine this important mechanism. In addition, the impact of air pollution on physical activity might be heterogeneous across population subgroups. For example, people with preexisting conditions such as asthma might be more sensitive to air quality and thus restrict their physical activity more than their healthy counterparts in response to an alarming level of air pollution. However, none of the seven included studies examined the differential relationship between air pollution and physical activity by preexisting disease status.

The relationship between air pollution and physical activity may well not be linear. The modest impact of PM_2.5 concentration on leisure-time physical inactivity identified in the meta-analysis is contingent upon the nationwide low PM_2.5 level in the United States (county monthly average ranged from 2.15 µg/m³ to 39.67 µg/m³ during 2003–2011). However, similar relationship between PM_2.5 air pollution and physical inactivity may not hold for highly polluted areas of the world where PM_2.5 levels were at least tenfold higher. This calls for primary data collection and urgent evaluation in heavily polluted developing economies such as China and India.^4,5

Heavy air pollution has gradually become a norm that urban residents in major developing countries such as China and India have to live with.^4,5 Individuals may modify their health behaviours to better adapt to the extended days under severe air pollution and minimise its adverse impact. Switching from outdoor physical activities (e.g. walking on sidewalks) to indoor exercises (e.g. walking on a treadmill) could partially offset the disruption of air pollution. However, this behavioural adaptation is contingent upon at least two essential conditions – change in social norm, and accessibility and affordability of indoor exercise facilities in close proximity (e.g. 5 min of walking distance). Improving the availability of exercise facilities in communities has become a national policy, as highlighted in China’s Twelfth Five-Year Plan (2011–2015) for the Sport Industry.³⁹ However, its protective effect on preserving physical activity under severe air pollution, if any, remains to be examined.

Under what air pollution level would the health benefit of outdoor physical activity outweigh the detrimental impact of air pollution? Recent research starts to examine this question, but is so far only based on data collected from developed countries. Andersen et al.⁴⁰ found that exposure to high levels of traffic-related air pollution did not attenuate the beneficial effects of physical activity on mortality in Denmark. Similar findings were reported in a Netherlands-based study comparing the health benefits and risks of outdoor cycling.⁴¹ However, the applicability of these findings to heavily polluted developing countries remains unknown. Moreover, the cost-benefit ratio of outdoor exercises under air pollution is likely to vary substantially across age and health/disease status.

Conclusion

This systematical review identified consistent evidence linking an elevated ambient air pollution level to reduced physical activity and increased leisure-time physical inactivity. Adult study participants, and particularly those with respiratory disease, mitigated the detrimental impact of air pollution by reducing their outdoor activities. Meta-analysis found a one unit (μg/m³) increase in ambient PM_2.5 concentration to be associated with an increase in the odds of physical inactivity by 1.1% among US adults. Current literature predominantly adopted a cross-sectional study design and focused on the United States. Future studies are warranted to implement a longitudinal study design and evaluate the impact of air pollution on physical activity in heavily polluted developing countries.

Footnotes

Acknowledgements

Ruopeng An and Sheng Zhang contributed equally.

Conflict of Interest

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Ethical Approval

This study is a systematic review that involves no human or animal subjects. Therefore, no ethnical approval was pursued.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Faergeman

Climate change, disease and public health. Ugeskrift for Laeger 2008; 170(35): 2667–8.

Minos

Butzlaff

Demmler

et al . Economic growth, climate change, and obesity. Current Obesity Reports 2016; 5(4): 441–8.

Carlisle

Sharp

NC.

Exercise and outdoor ambient air pollution. British Journal of Sports Medicine 2001; 35(4): 214–22.

Greenstone

Hanna

Environmental regulations, air and water pollution, and infant mortality in India. American Economic Review 2014; 104(10): 3038–72.

Rohde

Muller

RA.

Air pollution in China: Mapping of concentrations and sources. PLoS ONE 2015; 10(8): e0135749.

World Health Organization. Ambient (Outdoor) Air Quality and Health. Available online at: http://www.who.int/mediacentre/factsheets/fs313/en/ (2016, Last accessed 13^th February 2017).

United States Environmental Protection Agency. Criteria Air Pollutants. Available online at: https://www.epa.gov/criteria-air-pollutants (2017, Last accessed 18^th February 2017).

United States Environmental Protection Agency. Evaluation of Air Exposures in the North Birmingham Area-Final Air PHA. Available online at: https://www.epa.gov/sites/production/files/2014-06/documents/air-pha-fs-n-birm-244624a-final-5-14-14.pdf (2014, Last accessed 15^th February 2017).

Xing

Shi

et al . The impact of PM2.5 on the human respiratory system. Journal of Thoracic Disease 2016; 8(1): E69–74.

10.

Arsenault

Rana

Lemieux

et al . Physical inactivity, abdominal obesity and risk of coronary heart disease in apparently healthy men and women. International Journal of Obesity 2010; 34(2): 340–7.

11.

Kruk

Health and economic costs of physical inactivity. Asian Pacific Journal of Cancer Prevention 2014; 15(18): 7499–503.

12.

Lee

Shiroma

Lobelo

et al . Effect of physical inactivity on major non-communicable diseases worldwide: An analysis of burden of disease and life expectancy. Lancet 2012; 380(9838): 219–29.

13.

World Health Organization. Physical Inactivity: A Global Public Health Problem. Available online at: http://www.who.int/dietphysicalactivity/factsheet_inactivity/en/ (2017, Last accessed 5^th February 2017).

14.

Auchincloss

Diez Roux

Dvonch

et al . Associations between recent exposure to ambient fine particulate matter and blood pressure in the Multi-Ethnic Study of Atherosclerosis (MESA). Environmental Health Perspectives 2008; 116(4): 486–91.

15.

Brook

Rajagopalan

Pope

et al . Particulate matter air pollution and cardiovascular disease: An update to the scientific statement from the American Heart Association. Circulation 2010; 121(21): 2331–78.

16.

Cakmak

Dales

Leech

et al . The influence of air pollution on cardiovascular and pulmonary function and exercise capacity: Canadian Health Measures Survey (CHMS). Environmental Research 2011; 111(8): 1309–12.

17.

Miller

Siscovick

Sheppard

et al . Long-term exposure to air pollution and incidence of cardiovascular events in women. New England Journal of Medicine 2007; 356(5): 447–58.

18.

Cutrufello

Rundell

Smoliga

et al . Inhaled whole exhaust and its effect on exercise performance and vascular function. Inhalation Toxicology 2011; 23(11): 658–67.

19.

Marr

Ely

MR.

Effect of air pollution on marathon running performance. Medicine & Science in Sports & Exercise 2010; 42(3): 585–91.

20.

Rundell

Caviston

Ultrafine and fine particulate matter inhalation decreases exercise performance in healthy subjects. Journal of Strength & Conditioning Research 2008; 22(1): 2–5.

21.

Roberts

Voss

Knight

The association of ambient air pollution and physical inactivity in the United States. PloS ONE 2014; 9(3): e90143.

22.

Wen

Balluz

Mokdad

Association between media alerts of air quality index and change of outdoor activity among adult asthma in six states, BRFSS, 2005. Journal of Community Health 2009; 34(1): 40–6.

23.

Liu

Lü

et al . Ambient air pollution in China poses a multifaceted health threat to outdoor physical activity. Journal of Epidemiology & Community Health 2015; 69(3): 201–4.

24.

PRISMA. Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA). Available online at: http://www.prisma-statement.org/ (2015, Last accessed 9^th February 2017).

25.

National Institutes of Health National Heart Lung and Blood Institute. Quality Assessment Tool for Observational Cohort and Cross-Sectional Studies. Available online at: https://www.nhlbi.nih.gov/health-pro/guidelines/in-develop/cardiovascular-risk-reduction/tools/cohort (2014, Last accessed 11^th February 2017).

26.

Hartig

Mitchell

de Vries

et al . Nature and health. Annual Review of Public Health 2014; 35(1): 207–28.

27.

Lü

Liang

Feng

et al . Air pollution exposure and physical activity in China: Current knowledge, public health implications, and future research needs. International Journal of Environmental Research & Public Health 2015; 12(11): 14887–97.

28.

Ust’Iantsev

Kulikov

VG.

A decrease in the motor activity of children under the influence of technogenic air pollution. Gigiena i Sanitariia 1999; 2: 26–8.

29.

Villarreal-Calderón

Acuña

Villarreal-Calderón

et al . Assessment of physical education time and after-school outdoor time in elementary and middle school students in south Mexico City: The dilemma between physical fitness and the adverse health effects of outdoor pollutant exposure. Archives of Environmental Health 2002; 57(5): 450–60.

30.

Wen

Balluz

Shire

et al . Association of self–reported leisure–time physical inactivity with particulate matter 2.5 air pollution. Journal of Environmental Health 2009; 72(1): 40–4.

31.

Alahmari

Mackay

Patel

et al . Influence of weather and atmospheric pollution on physical activity in patients with COPD. Respiratory Research 2015; 16(1): 71.

32.

Xiang

Ambient fine particulate matter air pollution and leisure-time physical inactivity among US adults. Public Health 2015; 129(12): 1637–44.

33.

Hankey

Marshall

Brauer

Health impacts of the built environment: Within-urban variability in physical inactivity, air pollution, and ischemic heart disease mortality. Environmental Health Perspectives 2012; 120(2): 247–53.

34.

Wells

Dearborn

Jackson

LW.

Activity change in response to bad air quality, National Health and Nutrition Examination Survey, 2007–2010. PLoS ONE 2012; 7(11): e50526.

35.

Xiang

Mapping the prevalence of physical inactivity in U.S. States, 1984–2015. PLoS ONE 2016; 11(12): e0168175.

36.

Sallis

Floyd

Rodríguez

et al . The role of built environments in physical activity, obesity, and CVD. Future of Children 2006; 16(1): 89–108.

37.

Carver

Timperio

Crawford

. Playing it safe: The influence of neighbourhood safety on children’s physical activity. A review. Health & Place 2008; 14(2): 217–27.

38.

Mcneill

Kreuter

Subramanian

SV.

Social environment and physical activity: A review of concepts and evidence. Social Science & Medicine 2006; 63(4): 1011–22.

39.

General Administration of Sport of China. China’s Twelfth Five-Year Plan (2011–2015) for Building the Public Sport Infrastructure (in Chinese). Available online at: http://www.sport.gov.cn/n16/n33193/n33208/n33448/n33793/3425606.html (2016, Last accessed 15^th February 2017).

40.

Andersen

Nazelle

Mendez

et al . A study of the combined effects of physical activity and air pollution on mortality in elderly urban residents: The Danish Diet, Cancer, and Health Cohort. Environmental Health Perspectives 2015; 123(6): 557–63.

41.

Hartog

Boogaard

Nijland

et al . Do the health benefits of cycling outweigh the risks? Environmental Health Perspectives 2010; 118(8): 1109–16.