YPMED-04520; No of Pages 4

Preventive Medicine xxx (2016) xxx–xxx

Contents lists available at ScienceDirect

Preventive Medicine

journal homepage: www.elsevier.com/locate/ypmed

Brief Original Report

Can air pollution negate the health benefits of cycling and walking?
Marko Tainio a,⁎, Audrey J. de Nazelle b, Thomas Götschi c, Sonja Kahlmeier c, David Rojas-Rueda d,e,f,
Mark J. Nieuwenhuijsen d,e,f, Thiago Hérick de Sá g, Paul Kelly h, James Woodcock a
a
UKCRC Centre for Diet and Activity Research, MRC Epidemiology Unit, University of Cambridge School of Clinical Medicine, Institute of Metabolic Science, Cambridge, UK
b
Centre for Environmental Policy, Imperial College London, London, UK
c
Physical Activity and Health Unit, Epidemiology, Biostatistics and Prevention Institute, University of Zurich, Zurich, Switzerland
d
Center for Research in Environmental Epidemiology (CREAL), Barcelona, Spain
e
Universitat Pompeu Fabra (UPF), Barcelona, Spain
f
Centro de Investigación Biomédica en Red de Epidemiología y Salud Pública (CIBERESP), Madrid, Spain
g
Centre for Epidemiological Research in Nutrition and Health, School of Public Health, University of São Paulo, São Paulo, Brazil
h
Physical Activity for Health Research Centre (PAHRC), University of Edinburgh, UK

a r t i c l e i n f o a b s t r a c t

Article history: Active travel (cycling, walking) is beneficial for the health due to increased physical activity (PA). However, active
Received 9 October 2015 travel may increase the intake of air pollution, leading to negative health consequences. We examined the risk–
Received in revised 28 January 2016 benefit balance between active travel related PA and exposure to air pollution across a range of air pollution and
Accepted 1 February 2016
PA scenarios.
Available online xxxx
The health effects of active travel and air pollution were estimated through changes in all-cause mortality for dif-
Keywords:
ferent levels of active travel and air pollution. Air pollution exposure was estimated through changes in back-
Physical activity ground concentrations of fine particulate matter (PM2.5), ranging from 5 to 200 μg/m3. For active travel
Air pollution exposure, we estimated cycling and walking from 0 up to 16 h per day, respectively. These refer to long-term av-
Bicycling erage levels of active travel and PM2.5 exposure.
Walking For the global average urban background PM2.5 concentration (22 μg/m3) benefits of PA by far outweigh risks
Mortality from air pollution even under the most extreme levels of active travel. In areas with PM2.5 concentrations of
Health Impact Assessment 100 μg/m3, harms would exceed benefits after 1 h 30 min of cycling per day or more than 10 h of walking per
Risk–Benefit Assessment
day. If the counterfactual was driving, rather than staying at home, the benefits of PA would exceed harms
from air pollution up to 3 h 30 min of cycling per day. The results were sensitive to dose–response function
(DRF) assumptions for PM2.5 and PA.
PA benefits of active travel outweighed the harm caused by air pollution in all but the most extreme air pollution
concentrations.
© 2016 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY license
(http://creativecommons.org/licenses/by/4.0/).

Introduction exposure for low to moderate levels of air pollution, whereas the bene-
fits of PA increase curvy-linearly with increasing dose (Kelly et al., 2014;
Several health impact modelling (HIM) studies have estimated the World Health Organization, 2014). Thus, at a certain level of background
health benefits and risks of active travel (cycling, walking) in different air pollution and of active travel, risks could outweigh benefits, which
geographical areas (Mueller et al., 2015; Doorley et al., 2015). In most would directly imply that, from a public health perspective, active travel
of these studies, the health benefits due to physical activity (PA) from could not be always recommended.
increased active travel are significantly larger than the health risks In this study we compare the health risks of air pollution with the
caused by increases in exposure to air pollution. PA-related health benefits from active travel across a wide range of pos-
Most of the existing active travel HIM studies have been carried out sible air pollution concentrations and active travel levels. We use two
in cities in high income countries with relatively low air pollution levels thresholds to compare PA benefits and air pollution risks (Fig. 1): At
(Mueller et al., 2015; Doorley et al., 2015). This raises the question on the “tipping point” an incremental increase in active travel will no longer
the risk–benefit balance in highly polluted environments. Health risks lead to an increase in health benefits (i.e. max. benefits have been
of air pollution are usually thought to increase linearly with increased reached). Increasing active travel even more could lead to the “break-
even point”, where risk from air pollution starts outweighing the bene-
⁎ Corresponding author. fits of PA (i.e. there are no longer net benefits, compared to not engaging
E-mail address: [email protected] (M. Tainio). in active travel).

http://dx.doi.org/10.1016/j.ypmed.2016.02.002
0091-7435/© 2016 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

Please cite this article as: Tainio, M., et al., Can air pollution negate the health benefits of cycling and walking?, Prev. Med. (2016), http://
dx.doi.org/10.1016/j.ypmed.2016.02.002
2 M. Tainio et al. / Preventive Medicine xxx (2016) xxx–xxx

1.5

1.4

Relative risk of all-cause mortality

PM2.5 background level: 50µg/m3
1.3

1.2
Increase in
1.1 risk due to
AP

1
Risk
Break-even point:
0.9 reduction
beyond this, additional PA will cause
due to PA
adverse health effects
0.8
Tipping point:
beyond this, additional PA will not lead to higher health benefits
0.7
0 30 60 90 120 150 180 210 240 270 300 330 360 390 420 450 480 510 540 570 600
Cycling (min./day)

Fig. 1. Illustration of tipping point and break-even point as measured by the relative risk (RR) for all-cause mortality (ACM) combining the effects of air pollution (at 50 μg/m3 PM2.5) and
physical activity (cycling).

Methods supplementary file for the sensitivity analysis with counterfactual sce-
narios where cycling time would replace motorised transport time.
Our approach followed a general active travel HIM method (Mueller The ventilation rates differences whilst at sleep, rest, cycling and walk-
et al., 2015; Doorley et al., 2015). Air pollution exposures due to active ing were taken into account when converting exposure to inhaled dose.
travel were quantified by estimating the differences in the inhaled For sleep, rest, walking and cycling we used ventilation rates of 0.27,
dose of fine particulate matter (PM2.5) air pollution. We selected PM2.5 0.61, 1.37 and 2.55, respectively (de Nazelle et al., 2009; Johnson,
because it is a commonly used indicator of air pollution in active travel 2002). The sleep time was assumed to be 8 h in all scenarios and the
HIM studies (Mueller et al., 2015; Doorley et al., 2015), and because of resting time was 16 h minus the time for active travel.
the large health burden caused by PM2.5 (GBD 2013 Risk Factors For the PM2.5 DRF we used a relative risk (RR) value of 1.07 per
Collaborators et al., 2015). For both air pollution and PA we used all- 10 μg/m3 change in exposure (World Health Organization, 2014). We
cause mortality as the health outcome because there is strong evidence assumed that DRF is linear from zero to maximum inhaled dose. As a
for its association with both long-term exposure to PM2.5 (Héroux et al., sensitivity analysis we used non-linear integrated risk function from
2015) and long-term PA behaviour (Kelly et al., 2014). Burnett et al. (2014) (see supplementary material for details).
The reduction in all-cause mortality from active travel was estimat- The model used for all calculations is provided in Lumina Decision
ed by converting the time spent cycling or walking to metabolically Systems Analytica format in supplementary file 2 (readable with
equivalent of task (MET) and calculating the risk reduction using Analytica Free 101, http://www.lumina.com/products/free101/), and a
dose–response functions (DRFs) adapted from Kelly et al.'s3 meta- simplified model containing the main results is provided in Microsoft
analysis. From the different DRFs reported in Kelly et al. (2014) we Excel format in supplementary file 3.
chose the one with the “0.50 power transformation” as a compromise
between linear and extremely non-linear DRFs. Non-linearity in a DRF Results
means that the health benefits of increased active travel would level
out sooner and a tipping point would be reached earlier than with The tipping point and break-even point for different average cycling
more linear DRFs. See supplementary material for the sensitivity analy- times and background PM2.5 concentrations are shown in Fig. 2. For half
sis with different DRFs. To convert cycling and walking time to PA we an hour of cycling every day, the background PM2.5 concentration
used the values of 4.0 METs for walking and 6.8 METs for cycling, would need to be 95 μg/m3 to reach the tipping point. In the WHO
based on the Compendium of Physical Activities (Ainsworth et al., Ambient Air Pollution Database less than 1% of cities have PM2.5 annual
2011). The walking and cycling levels used in this study are assumed concentrations above that level (World Health Organization (WHO),
to reflect long-term average behaviour. 2014). The break-even point for half an hour of cycling every day was
The health risks of PM2.5 were estimated by converting background at 160 μg/m3 (Fig. 2). For half an hour of walking the tipping point
PM2.5 concentrations to travel mode specific exposure concentrations, and break-even point appear at a background concentration level
and by taking into account ventilation rate whilst being active. For back- above 200 μg/m3 (Fig. S3, supplementary file). For the average urban
ground PM2.5 we used values between 5 and 200 μg/m3 with 5 μg/m3 background PM2.5 concentration (22 μg/m3) in the WHO database, the
intervals. We also estimated tipping points and break-even points for tipping point would only be reached after 7 h of cycling and 16 h of
the average and most polluted cities in each region included in the walking per day.
World Health Organization (WHO) Ambient Air Pollution Database Tables S2 and S3 (supplementary file) show the tipping point for cy-
(World Health Organization (WHO), 2014), which contains measured cling and walking, respectively, in different regions of the world. In the
and estimated background PM2.5 concentrations for 1622 cities around most polluted city in the database (Delhi, India, background concentra-
the world. tion of 153 μg/m3), the tipping and break-even points were 30 and
The mode specific exposure concentrations were estimated by mul- 45 min of cycling per day, respectively (Table S2, supplementary file).
tiplying background PM2.5 concentration by 2.0 for cycling or 1.1 for In most global regions the tipping points for the most polluted cities
walking, based on a review of studies (Kahlmeier et al., 2014). The coun- (44 μg/m3 to 153 μg/m3) varied between 30 and 120 min per day for cy-
terfactual scenario for the time spent cycling or walking was assumed to cling, and 90 min to 6 h 15 min per day for walking (Table S3, supple-
be staying at home (i.e. in background concentration of PM2.5). See mentary material).

Please cite this article as: Tainio, M., et al., Can air pollution negate the health benefits of cycling and walking?, Prev. Med. (2016), http://
dx.doi.org/10.1016/j.ypmed.2016.02.002
 M. Tainio et al. / Preventive Medicine xxx (2016) xxx–xxx 3

cause mortality and have, thus, not taken into account the morbidity
impact.
For the health risks of air pollution we only estimated the increased
risk during cycling and walking, not the overall health risk from every-
day air pollution. Air pollution causes a large burden of diseases all over
the world (Burnett et al., 2014) and reducing air pollution levels would
provide additional health benefits. Since transport is an important
source of air pollution in urban areas, mode shifts from motorised trans-
port to active travel would not only improve health in active travellers,
but also help to reduce air pollution exposures for the whole population
(Johan de Hartog et al., 2010).
The results are sensitive to assumptions of the linearity of dose–re-
sponse relationships between active travel-related PA and health bene-
fits, and between PM2.5 and adverse health effects. With linear DRFs for
PA the benefits always exceeded the risks at all levels of PM2.5 concen-
trations. Evidence for a linear DRF for high PM2.5 concentrations is
small and, for example, the Global Burden of Disease study applied
non-linear, disease specific DRFs for PM2.5 (Burnett et al., 2014). If the
risks of PM2.5 level out after PM2.5 concentrations over 100 μg/m3, the
health benefits of PA would always exceed the risks of PM2.5.
It should also be taken into account that the results are based on gen-
Fig. 2. Tipping and break-even points for different levels of cycling (red dashed line and
erally representative values without detailed information on local con-
blue solid line, respectively) (minutes per day, x-axis) and for different background
PM2.5 concentrations (y-axis). Green lines represent the average and 99th percentile ditions, or from the background PA and disease history of individuals.
background PM2.5 concentrations in World Health Organization (WHO) Ambient Air For individuals highly active in non-transport domains the benefits
Pollution Database (World Health Organization (WHO), 2014). from active travel will be smaller, and vice versa.

When we assumed that time spend cycling would replace time driv- Conclusions
ing a car, benefits always exceeded the risks in the background air pol-
lution concentrations below 80 µg/m3, a concentration exceeded in The benefits from active travel generally outweigh health risks from
only 2% of cities (World Health Organization (WHO), 2014). Other sen- air pollution and therefore should be further encouraged. When
sitivity analyses showed that the results are sensitive to the shape of the weighing long-term health benefits from PA against possible risks
DRF functions. With the linear DRF for active travel the break-even point from increased exposure to air pollution, our calculations show that
would be reached with background PM2.5 concentrations of 170 μg/m3 promoting cycling and walking is justified in the vast majority of set-
regardless of the active travel time (Fig. S4, supplementary material); a tings, and only in a small number of cities with the highest PM2.5 con-
level not currently found in any of the cities in the WHO air pollution da- centration in the world cycling could lead to increase in risk.
tabase (World Health Organization (WHO), 2014). With the most
curved DRF (0.25 power) the PM2.5 concentration where harms exceed Author contributions
benefits for 1 h of cycling per day would drop from 150 μg/m3 to 130 μg/
m3 (Fig. S4, supplementary material), a level currently found only in 9 MT made the calculations and drafted the first version of the manu-
cities (World Health Organization (WHO), 2014). With a non-linear script. AJN, TG, MJN, SK, THS, DRR, PK and JW participated in designing
DRF for PM2.5 the break-even point was not reached in any background the scope of the study. AJN and TG helped to clarify the message of the
PM2.5 concentration when using “power 0.50” DRF for cycling and walk- study. All authors contributed to the writing of this paper. All authors
ing. Other input value modifications had small or insignificant impact to approved the final version to be submitted for consideration of
the results. publication.

Discussions Conflict of interest statement

The authors declare that there are no conflicts of interests.
This study indicates that, practically, air pollution risks will not ne-
gate the health benefits of active travel in urban areas in the vast major- Transparency document
ity of settings worldwide. Even in areas with high background PM2.5
concentrations, such as 100 μg/m3, up to 1 h 15 min of cycling and The Transparency document associated with this article can be
10 h 30 min of walking per day will lead to net reduction in all-cause found, in the online version.
mortality (Fig. S5, supplementary material). This result is supported
by epidemiological studies that have found the statistically significant Acknowledgments
protective effects of PA even in high air pollution environments
(Matthews et al., 2007; Andersen et al., 2015). However, a small minor- MT and JW: The work was undertaken by the Centre for Diet and Ac-
ity engaging in unusually high levels of active travel (i.e. bike messen- tivity Research (CEDAR), a UKCRC Public Health Research Centre of Ex-
gers) in extremely polluted environments may be exposed to air cellence. Funding from the British Heart Foundation, Cancer Research
pollution such that it negates the benefits of PA. UK, Economic and Social Research Council, Medical Research Council,
Some considerations of the limitations and the strengths of our the National Institute for Health Research, and the Wellcome Trust,
study need to be applied when generalising these findings. under the auspices of the UK Clinical Research Collaboration, is grateful-
In this analysis we took into account only the long-term health con- ly acknowledged.
sequences of regular PA and chronic exposure to PM2.5. Impacts of AJN, DRR, MJN, SK and TG: The work was supported by the project
short-term air pollution episodes, where concentrations significantly Physical Activity through Sustainable Transportation Approaches
exceed the average air pollution levels for a few days, may induce addi- (PASTAs) funded by the European Union's Seventh Framework Program
tional short term health effects. We have also only worked with all- under EC-GA no. 602624-2 (FP7-HEALTH-2013-INNOVATION-1). The

Please cite this article as: Tainio, M., et al., Can air pollution negate the health benefits of cycling and walking?, Prev. Med. (2016), http://
dx.doi.org/10.1016/j.ypmed.2016.02.002
4 M. Tainio et al. / Preventive Medicine xxx (2016) xxx–xxx

sponsors had no role in the study design; in the collection, analysis, and Héroux, M.-E., Anderson, H.R., Atkinson, R., et al., 2015. Quantifying the health impacts of
ambient air pollutants: recommendations of a WHO/Europe project. Int. J. Public
interpretation of data; in the writing of the report; and in the decision to Health 619–627 http://dx.doi.org/10.1007/s00038-015-0690-y.
submit the article for publication. Johan de Hartog, J., Boogaard, H., Nijland, H., Hoek, G., 2010. Do the health benefits of cy-
JW is supported by an MRC Population Health Scientist fellowship. cling outweigh the risks? Environ. Health Perspect. 118 (8), 1109–1116. http://dx.
doi.org/10.1289/ehp.0901747.
THS is supported by the Brazilian Science without Borders Scheme Johnson, T., 2002. A guide to selected algorithms, distribution, and databases used
(process number: 200358/2014-6) and the Sao Paulo Research Founda- in exposure models developed by the office of air quality planning and standards.
tion (process number: 2012/08565-4). U.S. Environmental Protection Agency. Available at: http://www2.epa.gov/sites/
production/files/2013-08/documents/report052202.pdf.
Kahlmeier, S., Schweizer, C., Rojas-Rueda, D., Nieuwenhuijsen, M., Nazelle, A. de, Bode, O.,
References 2014. Development of the Health Economic Assessment Tools (HEAT) for Walking
and Cycling — Meeting Background Document. Bonn, Germany.
Ainsworth, B.E., Haskell, W.L., Herrmann, S.D., et al., 2011. 2011 compendium of physical Kelly, P., Kahlmeier, S., Götschi, T., et al., 2014. Systematic review and meta-analysis of re-
activities: a second update of codes and MET values. Med. Sci. Sports Exerc. 43 (8), duction in all-cause mortality from walking and cycling and shape of dose response
1575–1581. http://dx.doi.org/10.1249/MSS.0b013e31821ece12. relationship. Int. J. Behav. Nutr. Phys. Act. 11 (1), 132. http://dx.doi.org/10.1186/
Andersen, Z.J., de Nazelle, A., Mendez, M.A., et al., 2015. A study of the combined effects of s12966-014-0132-x.
physical activity and air pollution on mortality in elderly urban residents: the Danish Matthews, C.E., Jurj, A.L., Shu, X.O., et al., 2007. Influence of exercise, walking, cycling,
diet, cancer, and health cohort. Environ. Health Perspect. 123 (6), 557–563. http://dx. and overall nonexercise physical activity on mortality in Chinese women. Am.
doi.org/10.1289/ehp.1408698. J. Epidemiol. 165 (12), 1343–1350. http://dx.doi.org/10.1093/aje/kwm088.
Burnett, R.T., Pope, C.A., Ezzati, M., et al., 2014. An integrated risk function for estimating Mueller, N., Rojas-Rueda, D., Cole-Hunter, T., et al., 2015. Health impact assessment of ac-
the global burden of disease attributable to ambient fine particulate matter exposure. tive transportation: A systematic review. Prev. Med. (Baltim) http://dx.doi.org/10.
Environ. Health Perspect. 122 (4), 397–403. http://dx.doi.org/10.1289/ehp.1307049. 1016/j.ypmed.2015.04.010.
de Nazelle, A., Rodríguez, D.A., Crawford-Brown, D., 2009. The built environment and World Health Organization (WHO), 2014. Ambient air pollution database. Available at:
health: impacts of pedestrian-friendly designs on air pollution exposure. Sci. Total http://apps.who.int/gho/data/view.main.AMBIENTCITY2014?lang=en Accessed Au-
Environ. 407 (8), 2525–2535. http://dx.doi.org/10.1016/j.scitotenv.2009.01.006. gust 10, 2015.
Doorley, R., Pakrashi, V., Ghosh, B., July 2015. Quantifying the health impacts of active World Health Organization, 2014. WHO expert meeting: methods and tools for assessing
travel: assessment of methodologies. Transp. Rev. 2015, 1–24. http://dx.doi.org/10. the health risks of air pollution at local, national and international level. Meeting re-
1080/01441647.2015.1037378. port. Copenhagen, Denmark. Available at: http://www.euro.who.int/__data/assets/
GBD 2013 Risk Factors CollaboratorsMH, Forouzanfar, Alexander, L., et al., 2015. Global, pdf_file/0010/263629/WHO-Expert-Meeting-Methods-and-tools-for-assessing-the-
regional, and national comparative risk assessment of 79 behavioural, environmental health-risks-of-air-pollution-at-local,-national-and-international-level.pdf.
and occupational, and metabolic risks or clusters of risks in 188 countries, 1990–
2013: a systematic analysis for the Global Burden of Disease Study 2013.
Lancet (London, England) 6736 (15), 1990–2013. http://dx.doi.org/10.1016/S0140-
6736(15)00128–2.

Please cite this article as: Tainio, M., et al., Can air pollution negate the health benefits of cycling and walking?, Prev. Med. (2016), http://
dx.doi.org/10.1016/j.ypmed.2016.02.002