the bmj |
2015;350:h1111 | doi: 10.1136/bmj.h1111
1Department of Epidemiology,
Harvard School of Public Health,
Boston MA, USA
2Department of Environmental
Health, Harvard School of Public
Health, Boston MA, USA
3Department of Epidemiology,
Johns Hopkins Bloomberg
School of Public Health,
Baltimore MD, USA
4Channing Division of Network
Medicine, Department of
Medicine at Brigham and
Women’s Hospital and Harvard
Medical School, Boston MA,
5Department of Psychiatry,
Brigham and Women’s Hospital
and Harvard Medical School,
Boston MA, USA
Correspondence to: M C Power
Johns Hopkins Universit y,
Phipps 446D, 600 North Wolfe
Street , Baltimore, MD 21287,
Additional material is published
online only. To view please visit
the jour nal online ( http://
dx .d oi .o rg /10.113 6/BMJ.h1111)
Cite this a s: BMJ ;:h
doi:10.1136/b mj .h1111
Accepted: 5 February 2015
The relation between past exposure to ne particulate air
pollution and prevalent anxiety: observational cohort study
Melinda C Power,1, 2, 3 Marianthi-Anna Kioumourtzoglou,2 Jaime E Hart,2, 4 Olivia I Okereke,1, 4, 5
Francine Laden,1, 2, 4 Marc G Weisskopf1, 2
To determine whether higher past exposure to
particulate air pollution is associated with prevalent
high symptoms of anxiety.
Observational cohort study.
Nurses’ Health Study.
71 271 women enrolled in the Nurses’ Health Study
residing throughout the contiguous United States who
had valid estimates on exposure to particulate matter
for at least one exposure period of interest and data on
MAIN OUTCOME MEASURES
Meaningfully high symptoms of anxiety, dened as a
score of 6 points or greater on the phobic anxiety
subscale of the Crown-Crisp index, administered
The 71 271 eligible women were aged between 57 and
85 years (mean 70 years) at the time of assessment of
anxiety symptoms, with a prevalence of high anxiety
symptoms of 15%. Exposure to particulate matter was
characterized using estimated average exposure to
particulate matter <2.5 μm in diameter (PM2.5) and 2.5
to 10 μm in diameter (PM2.5–10) in the one month, three
months, six months, one year, and 15 years prior to
assessment of anxiety symptoms, and residential
distance to the nearest major road two years prior to
assessment. Signicantly increased odds of high
anxiety symptoms were observed with higher exposure
to PM2.5 for multiple averaging periods (for example,
odds ratio per 10 μg/m3 increase in prior one month
average PM2.5: 1.12, 95% condence interval 1.06 to
1.19; in prior 12 month average PM2.5: 1.15, 1.06 to 1.26).
Models including multiple exposure windows
suggested short term averaging periods were more
relevant than long term averaging periods. There was
no association between anxiety and exposure to
PM2.5–10. Residential proximity to major roads was not
related to anxiety symptoms in a dose dependent
Exposure to ne particulate matter (PM2.5) was
associated with high symptoms of anxiety, with more
recent exposures potentially more relevant than more
distant exposures. Research evaluating whether
reductions in exposure to ambient PM2.5 would reduce
the population level burden of clinically relevant
symptoms of anxiety is warranted.
Anxiety disorders, characterized by disruptive fear,
worry, and related behavioral disturbances such as
avoidance or physical sensations of hyperarousal,1 are
the most common type of psychiatric disorder in the
general population.2 Globally, approximately 16% of
people will have an anxiety disorder in their lifetime
and 11% will have experienced an anxiety disorder in
the past year.2 Anxiety disorders are associated with
reduced productivity and increased psychiatric and
non-psychiatric medical care, absenteeism, and risk of
suicide.3 In 2010, anxiety disorders accounted for
approximately 26.8 million disability adjusted life years
worldwide.4 The monetary cost of anxiety disorders is
also substantial; in the United States, the annual direct
cost of anxiety disorders in the 1990s has been esti-
mated to be $42.3bn (£27.3bn; €37.3bn).5 Women have a
higher prevalence of anxiety disorders than men6 and
the onset for most anxiety disorders is commonly in
adolescence or young adulthood. However, the inci-
dence of anxiety disorders remains substantial in mid-
life, and new cases continue to arise into later life,
especially in the case of generalized anxiety disorder.7
Although numerous pharmacologic and non-pharma-
cologic therapies are available, remission is not always
possible. Many people have persistent symptoms
despite use of rst line treatments.8
Given the substantial personal and societal burden
from anxiety and the problem of treatment resistance, it
is imperative to identify modiable risk factors for anxi-
ety disorders and symptoms. One important environ-
mental exposure that may be related to anxiety is air
pollution. Specically, exposure to particulate matter air
pollution may induce or exacerbate anxiety through
increased oxidative stress and systemic inammation9–1 7
or through promotion or aggravation of chronic dis-
ease.18–3 2 Though there is a small set of studies consider-
ing the association between air pollution and mental
health outcomes,33–44 we are aware of only two small
WHAT IS ALREADY KNOWN ON THIS TOPIC
Toxicological work suggests exposure to particulate air pollution may induce or
exacerbate anxiety through increased oxidative stress and systemic inflammation
While a small but growing body of literature suggests an association between air
pollution and mental health outcomes, including anxiety, data on the relation
between exposure to particulate air pollution and anxiety in humans is lacking
WHAT THIS STUDY ADDS
Our study suggests that higher exposure to PM2.5 (particulate matter <2.5 μm in
diameter) especially higher recent exposure, is associated with an increased risk of
high symptoms of anxiety
2doi: 10.1136/bmj.h1111 |
2015;350:h1111 | the bmj
studies that considered anxiety, and neither looked at
total particulate matter. The rst (n=1002) reported that
ozone levels in the prior week were associated with anx-
iety symptoms,33 whereas the second (n=100) reported
that cumulative exposure to airborne manganese was
associated with anxiety symptoms.44 Epidemiologic
research on the relation between exposure to particulate
matter and anxiety is clearly lacking; we evaluated this
association in a large prospective cohort study. Speci-
cally, we hypothesized that higher exposure to particu-
late matter would be associated with a greater risk of
high symptoms of anxiety.
The most biologically relevant period of exposure is
currently unknown. If particulate matter induces anxi-
ety through chronic oxidative stress, inammation, or
induction of chronic disease, long term cumulative
exposure is most likely relevant. If particulate matter
aggravates an existing propensity for anxiety symptoms,
through either aggravation of chronic disease or tran-
sient changes in oxidative stress or inammation, expo-
sures closer to the time of symptom assessment may be
relevant. Therefore, we considered the association
between high anxiety symptoms and exposure to partic-
ulate matter averaged over ve periods prior to the
assessment of anxiety symptoms, specied a priori,
ranging from a measure of long term, cumulative expo-
sure (prior 15 years) to a measure of recent exposure
The Nurses’ Health Study is a prospective cohort study
of women that began in 1976. A total of 121 701 married
registered nurses, ages 30–55, residing in 11 states, were
originally enrolled; at least 10 participants now reside
in each of the 48 continental states. All participants are
mailed follow-up questionnaires every two years, with
a response rate of greater than 90% for each question-
naire.45 As such, we receive updated information on res-
idential address biennially, and we have geocoded all
home addresses 1986 to 2007 within the contiguous
United States to obtain latitude and longitude, allowing
estimation of exposure to particulate air pollution. The
Crown-Crisp index phobic anxiety scale, one of six
scales from the Crown-Crisp experiential index, is a
measure of anxiety symptom levels and was included in
the 1988 and 2004 questionnaires. As our exposure
data were available from 1988 onward (inclusive), we
used data from the 2004 Crown-Crisp index phobic anx-
iety scale as our outcome measure of anxiety. The
nurses’ provided implied informed consent by comple-
tion and return of each questionnaire.
Residential proximity to roadways
Using geographic information soware (ArcGIS, Version
10.2; Esri, CA), we computed distance from the residen-
tial address of each participant in 2002, up to 500 m,
with a street level geocoding match to the nearest US
census feature class code A1 (limited access to primary
roads with dened exits and divided directions of
travel, that is, interstate highways), A2 (primary major,
non-interstate highways and major roads without
access restrictions), or A3 (smaller, secondary roads,
typically with more than two lanes) road segment.
Distance to a major road is a commonly used proxy
fortrac related exposures, including trac related
airpollution (which typically contains a high proportion
of ultrane particles, those <0.1 μm in diameter).46–48
Weclassied distance to the nearest major road a priori
as <50 m, 50 to <200 m, or ≥200 m, based on the
observed pattern of particulate concentrations with
increasing distance47–4 9 and the distribution of roadway
proximity in our sample.
Particulate matter air pollution
We used spatiotemporal prediction models yielding
monthly estimates of exposure to particulate matter
<10 μm (PM10) and <2.5 μm (PM2.5 or ne particulate
matter) in aerodynamic diameter from January 1988
onward at the residential address with at least a zip
code level geocoding match for each participant to
derive multiple exposure metrics for each participant.
These models cover the contiguous United States and
are extensions of previously described models covering
a more limited area.50–52 Data used in these models
included nationwide monitor data, geographic data
(for example, distance to major roadway, population
density, elevation, proportion of urban land use, point
or area source emissions), and meteorological data (for
example, temperature, wind speed, precipitation,
barometric pressure). As nationwide PM2.5 monitor
data were unavailable prior to 1999, our pre-1999 esti-
mates of PM2.5 exposure were derived from a model that
estimates the predicted ratio of PM2.5 to PM10 between
1988 and 1999; to get PM2.5 predictions we combined
the results of this model with estimates from the PM10
model. We derived estimates of exposure to coarse
(PM2. 5–1 0) particulate matter by taking the dierence
between PM10 and PM2.5 estimates. For the current anal-
yses we used these models to derive measures of aver-
age exposure to PM2.5 and PM2.5 –1 0, at the residential
address of each participant for several exposure peri-
ods, including the average exposure between 1 January
1988 and 31 December 2003, and over the 1, 3, 6, and 12
calendar months prior to the participant’s 2004 ques-
tionnaire cycle return date (that is, for a questionnaire
returned in July 2004, we use the average exposure in
June 2004 for the one month averaging period).
The Crown-Crisp index phobic anxiety scale consists of
eight self rated questions about fearfulness and desire
for avoidance of common situations or environments
(that is, having “unreasonable fear of enclosed spaces”,
being “scared of heights”, disliking “going out alone”,
feeling “panicky in crowds”, feeling “more relaxed
indoors”, feeling “uneasy traveling on buses or trains”)
and tendency to worry (that is, “about getting some
incurable illness”, worrying “unduly when relatives are
late coming home”); total possible index scores range
from 0–16 points, with higher scores indicating more
anxiety.53 We required complete data on all eight items
3the bmj |
2015;350:h1111 | doi: 10.1136/bmj.h1111
to compute a total score. The Crown-Crisp index phobic
anxiety scale has been shown to dierentiate between
people with general anxiety or phobias from those with
other psychiatric conditions and healthy comparison
participants53 54 and has been used in population based
research.55–61 For primary analyses, we dichotomized
Crown-Crisp index phobic anxiety scale scores from
2004 and considered those with a score of 6 points or
more to have high symptoms of anxiety, as prior work
suggests that this cut-o represents a clinically import-
ant threshold.58 61
Covariates included in all models were selected a priori
because they were thought to be potential confounders
or proxies for potential confounders (for example, socio-
economic status) and include calendar month of ques-
tionnaire return (categorical month), educational
attainment (RN, BA, MA, or PhD), husband’s educational
attainment (≤12 years, 12–16 years, >16 years, not applica-
ble, missing), age, age squared, married or has a partner
(yes/no), employment status (yes/no), physical activity
(<12, 12 to 30, >30 metabolic equivalent task hours per
week), three residential census tract level characteristics
(percent white race/ethnicity, percent of adults without a
high school diploma, and median home value; in
fourths), region of residence (north east, south, midwest,
west), residence within a metropolitan statistical area
(yes/no), and social support62 (low, low-medium,
medium, high social networks). Many covariates were
assessed at multiple cycles; we used the value at the 1988
or closest available questionnaire when considering the
1988–2003 averaging period and the 2002 or closest
available questionnaire in models considering roadway
proximity or particulate matter exposures within the year
prior to the 2004 questionnaire return. With the excep-
tion of month of questionnaire return, we used missing
indicators when covariate data were missing for more
than 2% of our sample and replaced missing data with
median or mode values when covariate data were miss-
ing in less than 2% of our sample.
For each model we restricted our analytical sample to
people with 2004 Crown-Crisp index phobic anxiety
scale scores and relevant exposure data. We used sepa-
rate logistic regression models to estimate the associa-
tion between each exposure and high anxiety symptoms
(Crown-Crisp index phobic anxiety scale score ≥6). For
models considering exposure to particulate matter, we
evaluated the shape of the dose-response curve using
penalized splines and report analyses using both hs
of exposure and linear terms. As exposures to particu-
late matter are correlated across averaging periods (see
supplementary table e1), it is challenging to determine
which exposure periods are most relevant when multi-
ple periods appear associated with anxiety. Therefore,
we also considered mutually adjusted models including
either 1988–2003 and past one month or past 12 month
and past one month exposures to particulate matter
parameterized using penalized splines to tackle
whether long term or short term exposures were more
relevant when we observed an association between
anxiety and multiple averaging periods. To avoid the
potential for dierences in the variability of metrics for
exposure to particulate matter across the two averaging
periods to inuence the ndings, we used z score trans-
formations of each of the particulate matter exposures
(that is, one month, 12 months, and 1988–2003) in the
mutually adjusted models.
We conducted several sensitivity analyses to examine
the robustness of our primary ndings, including use of
alternate categorizations for roadway proximity (<50 m,
50–200 m, and >200 m from A1 or A2 roadways;
<100 m, 100–300 m, and >300 m from either A1, A2, or
A3 roadways or A1 or A2 only roadways; and as a contin-
uous variable for distance from A1, A2, or A3 roadways
using a linear or spline parameterization within the
range of 0 to 500 m); additional adjustment for individ-
ual level covariates oen correlated with anxiety symp-
toms but which were not expected to be confounders,
including physical functioning63 (high, low), self rated
health (excellent or very good, poor to average), num-
ber of major medical comorbidities (≥3, <3), alcohol
consumption (non-drinker, <3, 3–6, >7 alcoholic drinks
per week), body mass index (normal, overweight,
obese), and smoking status (never, former, current);
restriction to non-movers (to reduce misclassication of
exposure measures given some participants changed
addresses but exact move dates are unknown); restric-
tion to those who returned the questionnaire within
three months of the initial mailing (to reduce misclassi-
cation of short term exposure measures, which are
based on the return date for the questionnaire); restric-
tion to those living in a metropolitan area, dened
using rural-urban commuting codes64 (to reduce poten-
tial confounding by urban versus rural environments);
restriction to non-Hispanic white participants (96.7% of
the sample, to allay concerns about confounding by
race); use of negative binomial regression, which con-
siders Crown-Crisp index phobic anxiety scale scores as
count data and is similar to, but more appropriate and
generally more conservative than Poisson regression
when dealing with over-dispersed count data; and use
of an alternate case denition with improved sensitivity
but less specicity where we considered all people with
a Crown-Crisp index phobic anxiety scale score of 6 or
more and/or self report of use of anti-anxiety or antide-
pressant medications on the 2004 questionnaire to
have high anxiety symptoms.
We used multiplicative interaction terms and likelihood
ratio tests to evaluate evidence for eect modication
by several factors. These were residence within a metro-
politan statistical area (yes/no) and United States cen-
sus region (north east, south, midwest, west), as
particulate matter composition may vary spatially;
prevalent reactive airway disease (chronic obstructive
pulmonary disease or asthma), atrial brillation, heart
4doi: 10.1136/bmj.h1111 |
2015;350:h1111 | the bmj
failure, or multiple major medical conditions at the time
of anxiety assessment (yes/no), as particulate matter
may lead to anxiety through aggravation of symptoms
of common medical conditions; age (over or under 65 at
the time of anxiety assessment), as anxiety incidence
and prevalence change with age; and 1988 Crown-Crisp
index phobic anxiety scale score (0–1, 2–5, ≥6), given
that high anxiety symptoms may have been present
prior to the 2004 assessment. To limit the number of
tests, we evaluated eect modication only when pri-
mary analyses indicated a main eect, and then only for
the averaging period we judged to have the strongest
association. We made no other adjustments to account
for multiple comparisons. We report 95% condence
intervals and consider a P value <0.05 to be statistically
signicant. All analyses were conducted using SAS,
Version 9.3 or R, Version 3.0.1.
Sample sizes diered across analyses, based on avail-
ability of valid estimates on exposure (n=63 677 for
roadway proximity analyses, n=69 966 for 1988–2003
average analyses of exposure to particulate matter, and
n=71 271 for all other analyses). Among the largest
group (n=71 271), at the time of completion of the
Crown-Crisp index phobic anxiety scale the women in
our sample were on average aged 70 (SD 7, range 57–85)
years, 16% (n=11 320) of them reported current use of
antidepressants and/or anti-anxiety medications, and
the prevalence of high anxiety symptoms (Crown-Crisp
index phobic anxiety scale ≥6) was 15% (n=10 818) (g 1).
Table 1 shows the socioeconomic characteristics of this
sample. Of the 63 677 with valid estimates of 2002 resi-
dential roadway proximity, distance to the nearest
major road was >200 m for 59.0% (n=37 545), 50–200 m
for 26.4% (n=16 802), and <50 m for 14.7% (n=1120). In
line with temporal trends, mean estimates for exposure
to PM2.5 and PM2 .5 –10 were highest for the 1988–2003 expo-
sure period (Table 2 and supplementary table e1). For
example, the mean (SD) of exposures to PM2.5 –10 particu-
late matter was 9.0 μg/m3 (SD 4.1) in 1988–2003 com-
pared with 7.3 (4.8 μg/m3) for the one month averaging
period. Similarly, the mean (SD) of exposures to PM2.5
particulate matter was 13.8 (2.8 μg/m3) in 1988–2003
Crown-Crisp index score
Fig | Distribution of Crown-Crisp index phobic anxiety scale
scores among eligible participants of Nurses’ Health Study
Table | Socioeconomic characteristics from or
nearest available Nurses’ Health Study questionnaire
Characteristics No (%) of women (n = )
Educational attainment :
Registered nurse 44 907 (63.0)
Bachelors degree 13 368 (18.8)
Master s degree or PhD 66 07 (9.3)
Missing 6389 (9.0)
High school degree or less 24 664 (34.6)
College degree 16 321 (22.9)
P rofessional or gr aduate school
13 978 (19.6)
Not applicable 49 7 7 (7.0 )
Missing 11 331 (15.9)
No current life partner 20 521 (28.8)
Current life partner 49 855 (70.0)
Missing 895 (1.3)
Currently not employed outside
46 617 (65.4)
Currently employed outside the
23 891 (33.5)
Missing 763 (1.1)
<12 MET S/we ek 32 161 (45.1)
12 to <30 METS/week 27 582 (38.7)
≥30 METS/week 11 290 (15.8)
Missing 238 (0.3)
Percent of census tr act, white race/ethnicity:
<85% 15 627 (21.9)
85% to <94% 18 858 (26.5)
94% to <97% 15 992 (22.4)
≥97% 20 789 (29.2)
Missing 5 (0.0)
Percent of census tr act, adult residents without a high school
<5% 8911 (12 .5)
5% to <10% 19 211 (27.0)
10% to <15% 18 037 (25.3)
≥15% 25 107 (35.2)
Missing 5 (0.0)
Median home value ($):
<95 000 16 934 (23.8)
95 000 to <135 000 18 143 (25.5)
135 000 to <210 000 19 959 (28.0)
≥210 000 16 139 (22.7)
Missing 96 (0.0)
Region of residence:
North east 35 040 (49.2)
Midwest 12 355 (17.3)
West 10 199 (14.3)
South 13 677 (19.2)
Residence within a metropolitan statistical district:
Yes 64 648 (90.7)
No 662 3 (9.3)
Social support (Berkman-Syme index):
Low 4264 (6.0)
Low-medium 20 148 (28.3)
Medium 10 283 (14.4)
High 28 310 (39.7)
Missing 826 6 (11.6)
$1.00 (£0.65; €0.88).
5the bmj |
2015;350:h1111 | doi: 10.1136/bmj.h1111
compared with 12.7 (4.2 μg/m3) for the one month aver-
Residential proximity to roadways
Nurses who lived 50 to 200 m from the nearest major
road were more likely to have increased Crown-Crisp
index phobic anxiety scale scores than those living
>200 m away (adjusted odds ratio 1.06, 95% con-
dence interval 1.01 to 1.12; P=0.03). However, there was
no evidence of a dose-response pattern, as those living
within 50 m of the nearest major road did not have
increased odds (adjusted odds ratio 1.01, 0.95 to 1.08;
P=0.74). Findings of all sensitivity analyses were simi-
lar or more uniformly null (see supplementary table e2
and gure e1).
We observed associations between higher PM2.5 and
high anxiety across several averaging periods. Given
evidence for slightly non-linear dose-response patterns
in some averaging periods (see supplementary gure
e2), we report associations with both hs of exposure
(g 2) and per 10 μg/m3 increase in exposure (table 2).
Notably, while associations were similar across 1, 3, 6,
and 12 month averaging periods, associations for the
1988–2003 averaging period were weaker than for the
shorter averaging periods. All sensitivity analyses were
reasonably consistent with our primary models (see
supplementary tables e4 to e10). Mutually adjusted
models suggest that these associations were primarily
driven by an association between anxiety and shorter
averaging periods (g 3). There was little evidence to
support an association between high anxiety and expo-
sure to PM2. 5–1 0 in either our primary (see supplemen-
tary table e3 and gure e3) or our sensitivity analyses
(see supplementary tables e4 to e10). We did not
observe signicant eect modication of the associa-
tion with one month PM2.5 by any of the proposed vari-
ables (all likelihood ratio test P>0.16).
Table | Odds ratio (% condence interval) for g/m increase in exposure to PM.
over multiple averaging periods and high symptoms of anxiety in participants of Nurses’
Period Mean (SD) PM. (g/m) Odds ratio* (% CI) P value
1 month 12.74 (4.18) 1.12 (1.0 6 to 1.19) 0.0001
3 months 12.13 (3.40 ) 1.13 (1.0 6 to 1.2 1) 0.0004
6 months 11.59 (2. 77) 1.14 (1.05 t o 1.23) 0.002
12 months 11.3 8 (2.6 0) 1.15 (1.06 to 1. 25) 0.001
198 8–2 0 03 13. 75 (2. 82) 1.09 (1.0 1 to 1.18) 0.03
PM2.5=particulate matter <2.5 μm in diameter.
*Adjuste d for month of ques tionnaire retur n, nurse’s educ ation, husband ’s education, age, age squared,
whethe r the nurse has a partne r, employment sta tus, physical activ ity, percent of re sidential censu s tract that
is white, percent of r esidential cens us tract adult s who lack a high sc hool educatio n, median home val ue of
residential census trac t, geographic regio n, residence wit hin a metropoli tan statistical area , and social
Fihs of PM
Odds ratio (95% CI)
Odds ratio (95% CI)
Odds ratio (95% CI)
Odds ratio (95% CI)
Odds ratio (95% CI)
Fig | Adjusted* odds ratios (% condence intervals) between particulate matter <. m in diameter (PM.)
considered using hs of exposure, over multiple averaging periods and high symptoms of anxiety in Nurses’ Health
Study. *Adjusted for month of questionnaire return, nurse’s education, husband’s education, age, age squared, whether
the nurse has a partner, employment status, physical activity, percent of residential census tract that is white, percent of
residential census tract adults who lack a high school education, median home value of residential census tract,
geographic region, residence within a metropolitan statistical area, and social support
6doi: 10.1136/bmj.h1111 |
2015;350:h1111 | the bmj
Our data support an association between exposure to
particulate matter of <2.5 μm in diameter (PM2.5) but not
2.5 to 10 μm in diameter (PM2.5 –1 0) or proximity to road-
ways, and high symptoms of anxiety. The association
between PM2.5 and high anxiety seems primarily driven
by a relation with shorter term average exposures to
PM2.5. There is little evidence to suggest dierences in
this association by demographic, geographic, or health
Limitations and strengths of this study
Our study has some limitations. We were unable to con-
sider the clinical diagnosis of specic anxiety disorders.
However, prior epidemiologic work suggests that Crown-
Crisp index scores are a valid16 17 and clinically relevant
measure, as they are associated with accelerated
aging,58 ischemic heart disease,56 and sudden cardiac
death.55 59 65 We were unable to consider the association
between anxiety and uctuations in PM2.5 over periods
of less than one month, or short term uctuations in
ultrane particulate matter (residential distance to a
major road as a proxy measure is necessarily a longer
term indicator). It is possible that the association we
observed with PM2.5 is attributable, in whole or in part,
to a correlation between PM2.5 and another exposure.
Ambient ozone and noise are unlikely, although still
possible, candidates given relatively weak correlations
with PM2.5.66 68 Similarly, we cannot preclude a contribu-
tion of other pollutants that share sources with PM2.5 (for
example, nitrogen dioxide, sulphur dioxide).67 Our mod-
els provide predictions of exposure at each participant’s
residential address. Given lack of information on the
activity pattern of each participant, this could lead to bias
due to misclassication. Nevertheless, any such bias is
expected to be towards the null69 and so would not
account for the observed associations with PM2.5. Fur-
thermore, as environmental regulations set acceptable
exposure limits based on outdoor measures, we believe
that this exposure is most relevant from the public
health perspective. Our study considered only women; it
is possible that our results may not be generalizable to
men. Similarly, the women in our study were relatively
old. Given that advancing age is related to lower physio-
logic reserve,70 it is possible that our results would not
generalize to younger age groups.
Our study also has several strengths. While dis-
cussed previously as a limitation, our focus on anxiety
symptoms is also a strength. Our data suggest a short
term, potentially reversible relation between exposures
to particulate matter and severity of anxiety symptoms,
which may not have been identied if we had focused
exclusively on anxiety disorders. Although the relevant
exposure period was unknown, we considered multi-
ple averaging periods of exposure; this ultimately sug-
gested that short term exposure to PM2.5 may be the
PM2.5, z score prior 1 month (μg/m3)
-4 -2 0 2 46
PM2.5, z score 1988-2003 (μg/m3
-4 -2 0 2
PM2.5, z score prior 1 month (μg/m3)
-4 -2 0246
PM2.5, z score prior 12 months (μg/m3
-4 -2 024
Fig | Odds ratio (% condence interval) exposure to particulate matter <. m in diameter (PM.) on high symptoms
of anxiety in Nurses’ Health Study when multiple averaging periods are included in same model, parameterized using
splines. Results of a model including both z transformed prior month and z transformed prior one month PM.
exposure (panel A), and including both the z transformed prior years (–) and z transformed prior one month
PM. exposure (panel B). The th centile was chosen as the reference level for the corresponding exposure of interest.
Pvalues indicate strength of evidence for an association (versus no association) for exposure to particulate matter, over
its entire range, on high anxiety
7the bmj |
2015;350:h1111 | doi: 10.1136/bmj.h1111
most relevant exposure. We had access to a large
prospective cohort, allowing adequate power to detect
modest but meaningful associations. Attrition was
small and any potential selection bias due to informa-
tive drop-out would be expected to be a downward
bias, and so our estimate of an adverse association
with PM2.5 may be an underestimate. We were able to
adjust for many socioeconomic and sociodemographic
factors, which we thought to be the strongest potential
confounders. Our results were robust to multiple sensi-
Comparison to other studies and discussion of
To our knowledge this is the rst study to consider the
association between exposure to particulate matter and
anxiety. However, our ndings are consistent with two
prior studies of other air pollutants and anxiety,33 44 as
well as work suggesting associations between air pollu-
tion and other related, but distinct, mental health out-
comes, including depressive symptoms,34 35 psychiatric
emergency,36–38 emergency room visits for depression or
suicide,39–42 and reported suicide.43
Exposure to particulate matter could induce or
exacerbate anxiety through increased oxidative stress
and inammation or through inducing or aggravating
major medical conditions. Inammation and oxida-
tive stress have been hypothesized to contribute to the
incidence and severity of anxiety.9 10 Several toxico-
logical studies have shown that oxidative stress7 1–75 or
systemic inflammation75 76 induces anxiety-like
behaviors in mice and rats. These results are consis-
tent with cross sectional associations between anxi-
ety symptoms and inflammatory markers in
people,77–7 9 as well as epidemiologic ndings linking
C reactive protein, an inammatory marker, to gener-
alized anxiety disorder in patients with stable coro-
nary heart disease.80 Inhaled particulate matter may
therefore contribute to anxiety through induction of
systemic11–1 7 or brain based8 1–8 3 oxidative stress and
inammation. Alternatively, anxiety may occur as a
result of a respiratory or cardiac medical condition.
Reduced lung function, reactive airway diseases such
as asthma and chronic obstructive pulmonary dis-
ease, atrial brillation, and congestive heart failure
are associated with an increased prevalence of anxi-
ety symptoms or disorders.18–2 3 These associations are
likely mediated by fear and misinterpretation of
symptoms, although an impact of the stress of dealing
with major medical conditions or a purely physiolog-
ical reaction to oxygenation changes associated with
dysfunctional breathing and/or heart function may
also contribute.84–89 As particulate matter has been
linked to multiple medical conditions and aggrava-
tion of symptoms,24–32 particulate air pollution may
also contribute to anxiety through this alternative
mechanism. While our ndings are consistent with
the oxidative stress/inflammatory mechanistic
hypothesis, our data do not support the hypothesis
that particles promote anxiety through induction or
aggravation of medical conditions, as there was no
dierence in the association by whether or not people
had major medical comorbidities. The reported asso-
ciation with PM2.5, but not PM2. 5–1 0 may be related to
size related dierences in toxicity, which are likely a
function of dierences in lung penetrability, surface
area, and composition by particle size.90–94
Anxiety is a common and costly disorder. Our data
support an association between exposure to PM2.5, a
common environmental exposure, and high symp-
toms of anxiety. If conrmed, our ndings may have
policy and clinical implications, as it is possible that
reductions in exposure to PM2.5, through changes to
regulations or individual behavior, may help reduce
anxiety symptoms. Future work directly evaluating
this possibility is warranted.
We thank Peter James for his help with the Rural Urban Commuting
Contributors: All authors made substantial contributions to the
conception and design (MCP, MGW, MAK, JEH, OIO, and FL), acquisition
of the data (FL, JEH), or analysis and interpretation (MCP, MGW, MAK,
JEH, OIO, and FL). MCP draed the article and all other authors revised
it critically for important intellectual content. MCP is guarantor. All
authors had full access to all of the data in the study and can take
responsibility for the integrity of the data and the accuracy of the data
Funding: This work was supported by grants from the National
Institute of Environmental Health Sciences (R21 ES019982, R01
ES017017). MCP was supported by a training grant from the
National Institute of Aging (T32 AG027668). The funding agencies
had no role in the design and conduct of the study; collection,
management, analysis, and interpretation of the data; preparation,
review, or approval of the manuscript; or decision to submit the
manuscript for publication. The authors were independent of the
Competing interests: All authors have completed the ICMJE uniform
disclosure form at www.icmje.org/coi_disclosure.pdf (available on
request from the corresponding author) and declare: no support from
any organization for the submitted work; no nancial relationships
with any organizations that might have an interest in the submitted
work in the previous three years; no other relationships or activities
that could appear to have influenced the submitted work.
Ethical approval: This study was approved by the institutional review
board of the Brigham and Women’s Hospital and the human subjects
committee of the Harvard School of Public Health.
Data sharing: The statistical code is available from the corresponding
author at firstname.lastname@example.org.
Transparency: The lead author (MCP) arms that the manuscript is
an honest, accurate, and transparent account of the study being
reported; that no important aspects of the study have been omitted;
and that any discrepancies from the study as planned (and, if relevant,
registered) have been explained.
This is an Open Acces s article dist ributed in accordance with the
Creati ve Commons Attr ibution Non Comm ercial (CC BY-NC 4.0) license,
which per mits others to distribute, remix, adapt, build upon this wor k
non-commercially, and licens e their derivat ive works on di erent
terms, provided the orig inal work is proper ly cited and the use is
non-commercial. See: http://creativecommons.org/licenses/
by - nc /4 .0 /.
1 American Psychiatric Association. Diagnostic and statistical manual of
mental disorders, h edition (DSM-V). APA, 2013.
2 Kessler RC, Aguilar-Gaxiola S, Alonso J, Chatterji S, Lee S, Ormel J, et al.
The global burden of mental disorders: an update from the WHO
World Mental Health (WMH) surveys. Epidemiol Psichiatr Soc
3 Lepine JP. The epidemiology of anxiety disorders: prevalence and
societal costs. J Clin Psychiatry 2002;63:4–8.
4 Whiteford HA, Degenhardt L, Rehm J, Baxter AJ, Ferrari AJ, Erskine HE,
et al. Global burden of disease attributable to mental and substance
use disorders: ndings from the global burden of disease study 2010.
8doi: 10.1136/bmj.h1111 |
2015;350:h1111 | the bmj
5 Greenberg PE, Sisitsky T, Kessler RC, Finkelstein SN, Berndt ER,
Davidson JR, et al. The economic burden of anxiety disorders in the
1990s. J Clin Psychiatry 1999;60:427–35.
6 McLean CP, Asnaani A, Litz BT, Hofmann SG. Gender dierences in
anxiety disorders: prevalence, course of illness, comorbidity and
burden of illness. J Psychiatr Res 2011;45:1027–35.
7 Kessler RC, Berglund P, Demler O, Jin R, Merikangas KR, Walters EE.
Lifetime prevalence and age-of-onset distributions of DSM-IV
disorders in the national comorbidity survey replication. Arch Gen
8 Lanouette NM, Stein MB. Advances in the management of
treatment-resistant anxiety disorders. Focus (Am Psychiatr Publ)
9 Salim S, Chugh G, Asghar M. Inflammation in anxiety. Adv Protein
Chem Struct Biol 2012;88:1–25.
10 Hou R, Baldwin DS. A neuroimmunological perspective on anxiety
disorders. Hum Psychopharmacol 2012;27:6–14.
11 Sørensen M, Daneshvar B, Hansen M, Dragsted LO, Hertel O, Knudsen L,
et al. Personal PM2.5 exposure and markers of oxidative stress in
blood. Environ Health Perspect 2003;111:161–5.
12 Dagher Z, Garçon G, Billet S, Gosset P, Ledoux F, Courcot D, et al.
Activation of dierent pathways of apoptosis by air pollution
particulate matter (PM2.5) in human epithelial lung cells (l132) in
culture. Toxicology 2006;225:12–24.
13 Kim JY, Mukherjee S, Ngo L, Christiani DC. Urinary 8-hydroxy-2’-
deoxyguanosine as a biomarker of oxidative DNA damage in workers
exposed to ne particulates. Environ Health Perspect
14 Hovatta I, Tennant RS, Helton R, Marr RA, Singer O, Redwine JM, et al.
Glyoxalase 1 and glutathione reductase 1 regulate anxiety in mice.
15 Brook RD, Rajagopalan S, Pope CA, Brook JR, Bhatnagar A, Diez-Roux AV,
et al. Particulate matter air pollution and cardiovascular disease: an
update to the scientic statement from the American Heart
Association. Circulation 2010;121:2331–78.
16 Dubowsky SD, Suh H, Schwartz J, Coull BA, Gold DR. Diabetes, obesity,
and hypertension may enhance associations between air pollution
and markers of systemic inflammation. Environ Health Perspect
17 Bind MA, Baccarelli A, Zanobetti A, Tarantini L, Suh H, Vokonas P, et al.
Air pollution and markers of coagulation, inflammation, and
endothelial function: associations and epigene-environment
interactions in an elderly cohort. Epidemiology 2012;23:332–40.
18 Spitzer C, Gläser S, Grabe HJ, Ewert R, Barnow S, Felix SB, et al. Mental
health problems, obstructive lung disease and lung function: ndings
from the general population. J Psychosom Res 2011;71:174–9.
19 Scott KM, Von Kor M, Ormel J, Zhang M-y, Bruaerts R, Alonso J, et al.
Mental disorders among adults with asthma: Results from the world
mental health survey. Gen Hosp Psychiatry 2007;29:123–33.
20 Brenes GA. Anxiety and chronic obstructive pulmonary disease:
Prevalence, impact, and treatment. Psychosom Med
21 Yohannes AM, Willgoss TG, Baldwin RC, Connolly MJ. Depression and
anxiety in chronic heart failure and chronic obstructive pulmonary
disease: Prevalence, relevance, clinical implications and management
principles. Int J Geriatr Psychiatry 2010;25:1209–21.
22 Thrall G, Lip GYH, Carroll D, Lane D. Depression, anxiety, and quality of
life in patients with atrial brillation. CHEST 2007;132:1259–64.
23 Lane DA, Langman CM, Lip GYH, Nouwen A. Illness perceptions,
aective response, and health-related quality of life in patients with
atrial brillation. J Psychosom Res 2009;66:203–10.
24 Ling SH, van Eeden SF. Particulate matter air pollution exposure: Role
in the development and exacerbation of chronic obstructive
pulmonary disease. Int J Chron Obstruct Pulmon Dis 2009;4:233–43.
25 Faustini A, Stafoggia M, Cappai G, Forastiere F. Short-term eects of
air pollution in a cohort of patients with chronic obstructive
pulmonary disease. Epidemiology 2012;23:861–79.
26 Delno RJ. Epidemiologic evidence for asthma and exposure to air
toxics: Linkages between occupational, indoor, and community air
pollution research. Environ Health Perspect 2002;110:573.
27 Künzli N, Bridevaux P-O, Liu L-JS, Garcia-Esteban R, Schindler C,
Gerbase MW, et al. Trac-related air pollution correlates with
adult-onset asthma among never-smokers. Thorax 2009;64:664–70.
28 Link MS, Luttmann-Gibson H, Schwartz J, Mittleman MA, Wessler B,
Gold DR, et al. Acute exposure to air pollution triggers atrial
brillation. J Am Coll Cardiol 2013;62:816–25.
29 Pope CA. Review: Epidemiological basis for particulate air pollution
health standards. Aerosol Sci Technol 2000;32:4–14.
30 Rich DQ, Mittleman MA, Link MS, Schwartz J, Luttmann-Gibson H,
Catalano PJ, et al. Increased risk of paroxysmal atrial brillation
episodes associated with acute increases in ambient air pollution.
Environ Health Perspect 2006;114:120–3.
31 Kesavachandran C, Pangtey BS, Bihari V, Fareed M, Pathak MK,
Srivastava AK, et al. Particulate matter concentration in ambient air
and its eects on lung functions among residents in the national
capital region, India. Environ Monit Assess 2013;185:1265–72.
32 Shah ASV, Langrish JP, Nair H, McAllister DA, Hunter AL, Donaldson K,
et al. Global association of air pollution and heart failure: A systematic
review and meta-analysis. Lancet;382:1039–48.
33 Evans GW, Colome SD, Shearer DF. Psychological reactions to air
pollution. Environ Res 1988;45:1–15.
34 Lim YH, Kim H, Kim JH, Bae S, Park HY, Hong YC. Air pollution and
symptoms of depression in elderly adults. Environ Health Perspect
35 Jacobs SV, Evans GW, Catalano R, Dooley D. Air pollution and
depressive symptomatology: Exploratory analyses of intervening
psychosocial factors. Popul Environ 1984;7:260–72.
36 Briere J, Downes A, Spensley J. Summer in the city: Urban weather
conditions and psychiatric emergency-room visits. J Abnorm Psychol
37 Rotton J, Frey J. Psychological costs of air pollution: Atmospheric
conditions, seasonal trends, and psychiatric emergencies. Popul
38 Strahilevitz M, Strahilevitz A, Miller JE. Air pollutants and the admission
rate of psychiatric patients. Am J Psychiatry 1979;136:205–7.
39 Szyszkowicz M, Willey JB, Grafstein E, Rowe BH, Colman I. Air pollution
and emergency department visits for suicide attempts in Vancouver,
Canada. Environ Health Insights 2010;4:79–86.
40 Szyszkowicz M, Rowe BH, Colman I. Air pollution and daily emergency
department visits for depression. Int J Occup Med Environ Health
41 Szyszkowicz M. Air pollution and emergency department visits for
depression in Edmonton, Canada. Int J Occup Med Environ Health
42 Kim C, Jung SH, Kang DR, Kim HC, Moon KT, Hur NW, et al. Ambient
particulate matter as a risk factor for suicide. Am J Psychiatry
43 Biermann T, Stilianakis N, Bleich S, Thürauf N, Kornhuber J, Reulbach U.
The hypothesis of an impact of ozone on the occurrence of completed
and attempted suicides. Med Hypotheses 2009;72:338–41.
44 Bowler RM, Harris M, Gocheva V, Wilson K, Kim Y, Davis SI, et al.
Anxiety aecting parkinsonian outcome and motor eciency in adults
of an Ohio community with environmental airborne manganese
exposure. Int J Hyg Environ Health 2012;215:393–405.
45 Colditz GA, Manson JE, Hankinson SE. The Nurses’ Health Study:
20-year contribution to the understanding of health among women.
JWomens Health 1997;6:49–62.
46 Ran U, Schikowski T, Sugiri D, Krutmann J, Kramer U. Long-term
exposure to trac-related particulate matter impairs cognitive
function in the elderly. Environ Res 2009;109:1004–11.
47 Zhu Y, Hinds WC, Kim S, Sioutas C. Concentration and size distribution
of ultrane particles near a major highway. J Air Waste Manag Assoc
48 Sioutas C, Delno RJ, Singh M. Exposure assessment for atmospheric
ultrane particles (UFPS) and implications in epidemiologic research.
Environ Health Perspect 2005;113:947–55.
49 Karner AA, Eisinger DS, Niemeier DA. Near-roadway air quality:
Synthesizing the ndings from real-world data. Environ Sci Technol
50 Yanosky JD, Paciorek CJ, Suh HH. Predicting chronic ne and coarse
particulate exposures using spatiotemporal models for the
northeastern and midwestern United States. Environ Health Perspect
51 Yanosky JD, Paciorek CJ, Schwartz J, Laden F, Puett R, Suh HH.
Spatio-temporal modeling of chronic PM10 exposure for the Nurses’
Health Study. Atmos Environ 2008;42:4047–62.
52 Yanosky J, Paciorek C, Laden F, Hart J, Puett R, Liao D, et al.
Spatio-temporal modeling of particulate air pollution in the
conterminous united states using geographic and meteorological
predictors. Environ Health 2014;13:63.
53 Crown S, Crisp AH. A short clinical diagnostic self-rating scale for
psychoneurotic patients. Br J Psychiatry 1966;112:917–23.
54 Mavissakalian M, Michelson L. The Middlesex Hospital Questionnaire:
A validity study with american psychiatric patients. Br J Psychiatry
55 Albert CM, Chae CU, Rexrode KM, Manson JE, Kawachi I. Phobic
anxiety and risk of coronary heart disease and sudden cardiac death
among women. Circulation 2005;111:480–7.
56 Haines AP, Imeson JD, Meade TW. Phobic anxiety and ischaemic heart
disease. Br Med J (Clin Res Ed) 1987;295:297–9.
57 Kawachi I, Sparrow D, Vokonas PS, Weiss ST. Decreased heart rate
variability in men with phobic anxiety (data from the Normative Aging
Study). Am J Cardiol 1995;75:882–5.
58 Okereke OI, Prescott J, Wong JY, Han J, Rexrode KM, De Vivo I. High
phobic anxiety is related to lower leukocyte telomere length in
women. PLoS One 2012;7:e40516.
59 Watkins LL, Blumenthal JA, Babyak MA, Davidson JR, McCants CB, Jr.,
O’Connor C, et al. Phobic anxiety and increased risk of mortality in
coronary heart disease. Psychosom Med 2010;72:664–71.
60 Weisskopf MG, Chen H, Schwarzschild MA, Kawachi I, Ascherio A.
Prospective study of phobic anxiety and risk of parkinson’s disease.
Mov Disord 2003;18:646–51.
No commercial reuse: See rights and reprints http://www.bmj.com/permissions Subscribe: http://www.bmj.com/subscribe
61 McGrath M, Kawachi I, Ascherio A, Colditz GA, Hunter DJ, De Vivo I.
Association between catechol-o-methyltransferase and phobic
anxiety. Am J Psychiatry 2004;161:1703–5.
62 Berkman LF, Syme SL. Social networks, host resistance, and
mortality:A nine-year follow-up study of Alameda County residents.
Am J Epidemiol 1979;109:186–204.
63 Ware JE, Jr., Sherbourne CD. The MOS 36-item short-form health
survey (SF-36). I. Conceptual framework and item selection.
64 United States Department of Agriculture. Documentation: 2010
rural-urban commuting area (RUCA) codes. June 2, 2014. http://www.
65 Kawachi I, Colditz GA, Ascherio A, Rimm EB, Giovannucci E, Stampfer MJ,
et al. Prospective study of phobic anxiety and risk of coronary heart
disease in men. Circulation 1994;89:1992–7.
66 Bell ML, Kim JY, Dominici F. Potential confounding of particulate matter
on the short-term association between ozone and mortality in multisite
time-series studies. Environ Health Perspect 2007;115:1591–5.
67 Brunekreef B, Holgate ST. Air pollution and health. Lancet
68 Babisch W, Wolf K, Petz M, Heinrich J, Cyrys J, Peters A. Associations
between trac noise, particulate air pollution, hypertension, and
isolated systolic hypertension in adults: The KORA Study. Environ
Health Perspect 2014;122:492–8.
69 Kioumourtzoglou MA, Spiegelman D, Szpiro AA, Sheppard L, Kaufman
JD, Yanosky JD, et al. Exposure measurement error in PM2.5 health
eects studies: A pooled analysis of eight personal exposure
validation studies. Environ Health 2014;13:2.
70 Buchner DM, Wagner EH. Preventing frail health. Clin Geriatr Med
71 Ng F, Berk M, Dean O, Bush AI. Oxidative stress in psychiatric
disorders: Evidence base and therapeutic implications. Int J
72 de Oliveira MR, Silvestrin RB, Mello e Souza T, Moreira JCF. Oxidative
stress in the hippocampus, anxiety-like behavior and decreased
locomotory and exploratory activity of adult rats: Eects of sub acute
vitamin A supplementation at therapeutic doses. Neurotoxicology
73 Salim S, Sarraj N, Taneja M, Saha K, Tejada-Simon MV, Chugh G.
Moderate treadmill exercise prevents oxidative stress-induced
anxiety-like behavior in rats. Behav Brain Res 2010;208:545–52.
74 Salim S, Asghar M, Chugh G, Taneja M, Xia Z, Saha K. Oxidative stress:
A potential recipe for anxiety, hypertension and insulin resistance.
Brain Res 2010;1359:178–85.
75 Patki G, Solanki N, Atrooz F, Allam F, Salim S. Depression, anxiety-like
behavior and memory impairment are associated with increased
oxidative stress and inflammation in a rat model of social stress. Brain
76 Engler H, Doenlen R, Engler A, Riether C, Prager G, Niemi M-B, et al.
Acute amygdaloid response to systemic inflammation. Brain Behav
77 Pitsavos C, Panagiotakos DB, Papageorgiou C, Tsetsekou E, Soldatos C,
Stefanadis C. Anxiety in relation to inflammation and coagulation
markers, among healthy adults: The Attica Study. Atherosclerosis
78 Liukkonen T, Räsänen P, Jokelainen J, Leinonen M, Järvelin MR,
Meyer-Rochow VB, et al. The association between anxiety and
c-reactive protein (CRP) levels: Results from the Northern Finland
1966 Birth Cohort Study. Eur Psychiatry 2011;26:363–9.
79 Vogelzangs N, Beekman ATF, de Jonge P, Penninx BWJH. Anxiety
disorders and inflammation in a large adult cohort. Transl Psychiatry
80 Bankier B, Barajas J, Martinez-Rumayor A, Januzzi JL. Association
between c-reactive protein and generalized anxiety disorder in stable
coronary heart disease patients. Eur Heart J 2008;29:2212–7.
81 Campbell A, Oldham M, Becaria A, Bondy SC, Meacher D, Sioutas C,
etal. Particulate matter in polluted air may increase biomarkers of
inflammation in mouse brain. Neurotoxicology 2005;26:133–40.
82 Calderon-Garciduenas L, Maronpot RR, Torres-Jardon R, Henriquez-
Roldan C, Schoonhoven R, Acuna-Ayala H, et al. DNA damage in nasal
and brain tissues of canines exposed to air pollutants is associated
with evidence of chronic brain inflammation and neurodegeneration.
Toxicol Pathol 2003;31:524–38.
83 MohanKumar SM, Campbell A, Block M, Veronesi B. Particulate
matter, oxidative stress and neurotoxicity. Neurotoxicology
84 Mikkelsen RL, Middelboe T, Pisinger C, Stage KB. Anxiety and
depression in patients with chronic obstructive pulmonary disease
(COPD). A review. Nord J Psychiatry 2004;58:65–70.
85 Livermore N, Sharpe L, McKenzie D. Panic attacks and panic disorder
in chronic obstructive pulmonary disease: A cognitive behavioral
perspective. Respir Med 2010;104:1246–53.
86 Griez E, Schruers K. Experimental pathophysiology of panic.
JPsychosom Res 1998;45:493–503.
87 Gehi AK, Sears S, Goli N, Walker TJ, Chung E, Schwartz J, et al.
Psychopathology and symptoms of atrial brillation: Implications for
therapy. J Cardiovasc Electrophysiol 2012;23:473–8.
88 Katon W, Lin EHB, Kroenke K. The association of depression and
anxiety with medical symptom burden in patients with chronic
medical illness. Gen Hosp Psychiatry 2007;29:147–55.
89 Roy-Byrne PP, Davidson KW, Kessler RC, Asmundson GJG, Goodwin
RD, Kubzansky L, et al. Anxiety disorders and comorbid medical
illness. Gen Hosp Psychiatry 2008;30:208–25.
90 Araujo JA, Nel AE. Particulate matter and atherosclerosis: Role of
particle size, composition and oxidative stress. Part Fibre Toxicol
91 Diociaiuti M, Balduzzi M, De Berardis B, Cattani G, Stacchini G,
Ziemacki G, et al. The two PM2.5 (ne) and PM2.5–10 (coarse)
fractions: Evidence of dierent biological activity. Environ Res
92 Kelly FJ. Oxidative stress: Its role in air pollution and adverse health
eects. Occup Environ Med 2003;60:612–6.
93 Li N, Sioutas C, Cho A, Schmitz D, Misra C, Sempf J, et al. Ultrane
particulate pollutants induce oxidative stress and mitochondrial
damage. Environ Health Perspect 2003;111:455–60.
94 Pakkanen TA, Loukkola K, Korhonen CH, Aurela M, Mäkelä T, Hillamo RE,
et al. Sources and chemical composition of atmospheric ne and
coarse particles in the Helsinki area. Atmos Environ
© BMJ Publishing Group Ltd 2015