Respiratory Health Effects of Airborne Particulate Matter: The Role of Particle Size, Composition, and Oxidative Potential-The RAPTES Project.

Abstract

Background: Specific characteristics of particulate matter (PM) responsible for associations with respiratory health observed in epidemiological studies are not well established. High correlations among, and differential measurement errors of, individual components contribute to this uncertainty.Objectives: We investigated which characteristics of PM have the most consistent associations with acute changes in respiratory function in healthy volunteers.Methods: We used a semiexperimental design to accurately assess exposure. We increased exposure contrast and reduced correlations among PM characteristics by exposing volunteers at five different locations: an underground train station, two traffic sites, a farm, and an urban background site. Each of the 31 participants was exposed for 5 hr while exercising intermittently, three to seven times at different locations during March-October 2009. We measured PM10, PM2.5, particle number concentrations (PNC), absorbance, elemental/organic carbon, trace metals, secondary inorganic components, endotoxin content, gaseous pollutants, and PM oxidative potential. Lung function [FEV1 (forced expiratory volume in 1 sec), FVC (forced vital capacity), FEF25-75 (forced expiratory flow at 25-75% of vital capacity), and PEF (peak expiratory flow)] and fractional exhaled nitric oxide (FENO) were measured before and at three time points after exposure. Data were analyzed with mixed linear regression.Results: An interquartile increase in PNC (33,000 particles/cm3) was associated with an 11% [95% confidence interval (CI): 5, 17%] and 12% (95% CI: 6, 17%) FENO increase over baseline immediately and at 2 hr postexposure, respectively. A 7% (95% CI: 0.5, 14%) increase persisted until the following morning. These associations were robust and insensitive to adjustment for other pollutants. Similarly consistent associations were seen between FVC and FEV1 with PNC, NO2 (nitrogen dioxide), and NOx (nitrogen oxides).Conclusions: Changes in PNC, NO2, and NOx were associated with evidence of acute airway inflammation (i.e., FENO) and impaired lung function. PM mass concentration and PM10 oxidative potential were not predictive of the observed acute responses.

Full text

Page 1

Environmental Health Perspectives • volume 120 | number 8 | August 2012
1183
Research
Positive associations between airborne
particulate matter (PM) and respiratory
health have been observed in epidemiological
studies (Brunekreef and Holgate 2002; Pope
and Dockery 2006). In most studies, effects
were linked to PM10 and PM2.5 (particulate
matter < 10 µm and 2.5 µm in aero dynamic
diameter, respectively). Fewer studies have
reported health effects associated with exposure
to coarse (PM2.5–10; Brunekreef and Forsberg
2005) and ultrafine (PM0.1; Ibald-Mulli et al.
2002) particles. Depending on sources, there is
a significant hetero geneity in PM composition,
which is reflected in in vitro and in vivo
toxicological studies (Valavanidis et al. 2008).
Current knowledge does not allow precise
quantifi ca tion of the health effects of individual
PM components or of PM emissions from
different sources [Brunekreef 2010; World
Health Organization (WHO) 2007], although
various PM charac teris tics, such as surface area
of particles, transition metal content, surface
absorbed organic components, and biological
products (endotoxin), have been proposed. A
measure of oxidative potential (OP) of PM has
gained attention as a more integrative measure
of biological response (Ayres et al. 2008). OP
is an attractive measure because it integrates the
effects of multiple individual PM components
on health. There is currently, however, very
limited evidence in epidemiological studies that
the OP of PM predicts health effects better than
individual components (Ayres et al. 2008).
Disentangling the independent health
effects of individual PM charac teris tics in epi-
demiological studies is often limited by high
correlations between air pollution compo-
nents (Brunekreef 2010). Different degrees
of measure ment error for different air pollu-
tion components related to charac terizing
exposure at a central monitoring location is
another problem because more consistent
associations tend to be found with air pollu-
tion components with less measurement error
(Zeger et al. 2000). Controlled experimental
exposure studies in laboratory settings cannot
wholly represent the complexity of ambi-
ent PM exposures and are largely restricted
to individual air pollutants or defined mix-
tures (e.g., diesel engine exhaust). Moreover,
experi mental concentrations are often higher
than those encountered in real-world situa-
tions, and the concentration levels used are
constant rather than (highly) variable.
Building on recommendations of a recent
WHO workshop (WHO 2007), we addressed
these uncertainties using a semi experi mental
design as part of the RAPTES project (Risk
of Airborne Particles: a Toxicological–
Epidemiological hybrid Study). We studied
health effects of short-term exposure of healthy
volunteers to ambient PM at real-world loca-
tions with well-established differences in PM
charac teris tics (Strak et al. 2011). The aim
of the study was to assess the independent
contribution of specific PM charac teris tics to
various health outcomes. Here, we focus on
acute changes in respiratory health parameters.
We hypothesized that PM10 OP would have a
stronger and more consistent relation ship with
airway inflammation and lung function than
other measured PM charac teris tics because
oxidative stress is an important mechanism of
PM health effects.
Methods
Study design. We exposed healthy human
volun teers to ambient PM at five locations
with different PM charac teris tics. A detailed
Address correspondence to B. Brunekreef, Institute
for Risk Assessment Sciences, Utrecht University,
P.O. Box 80178 3508TD, Utrecht, Netherlands.
Telephone: 31 30 2539494. Fax: 31 30 2539499.
E-mail: B.Brunekreef@uu.nl
Supplemental Material is available online (http://
dx.doi.org/10.1289/ehp.1104389).
We thank all study participants. We also thank
J. Boere, P. Fokkens, D. Leseman, L. van den Burg,
V. Huijgen, M. Kleintjes, M. Meijerink, and
J. Musters for their excellent support in data collection;
E. van Otterloo for his help with participant recruit-
ment; and M. Groothoff for medical super vision.
The RAPTES (Risk of Airborne Particles: a
Toxicological–Epidemiological hybrid Study) project
was funded by the National Institute for Public Health
and the Environment (RIVM) Strategic Research
Program (S630002).
The authors declare they have no actual or potential
competing financial interests.
Received 23 August 2011; accepted 2 May 2012.
Respiratory Health Effects of Airborne Particulate Matter: The Role of Particle
Size, Composition, and Oxidative Potential—The RAPTES Project
Maciej Strak,1,2 Nicole A.H. Janssen,1 Krystal J. Godri,3,4 Ilse Gosens,1 Ian S. Mudway,3 Flemming R. Cassee,1
Erik Lebret,1,2 Frank J. Kelly,3 Roy M. Harrison,4,5 Bert Brunekreef,2,6 Maaike Steenhof,2 and Gerard Hoek2
1National Institute for Public Health and the Environment (RIVM), Bilthoven, Netherlands; 2Institute for Risk Assessment Sciences (IRAS),
Utrecht University, Utrecht, Netherlands; 3MRC-HPA Centre for Environmental Health, King’s College London, London, United Kingdom;
4Division of Environmental Health and Risk Management, University of Birmingham, Birmingham, United Kingdom; 5Center of
Excellence in Environmental Studies, King Abdulaziz University, Jeddah, Saudi Arabia; 6Julius Centre for Health Sciences and Primary
Care, University Medical Centre Utrecht, Utrecht, Netherlands
Background: Specific charac teris tics of particulate matter (PM) responsible for associations with
respiratory health observed in epidemiological studies are not well established. High correlations
among, and differential measurement errors of, individual components contribute to this uncertainty.
oBjectives: We investigated which charac teris tics of PM have the most consistent associations with
acute changes in respiratory function in healthy volunteers.
Methods: We used a semi experimental design to accurately assess exposure. We increased exposure
contrast and reduced correlations among PM charac teris tics by exposing volunteers at five different
locations: an under ground train station, two traffic sites, a farm, and an urban background site. Each
of the 31 participants was exposed for 5 hr while exercising intermittently, three to seven times at dif-
ferent locations during March–October 2009. We measured PM10, PM2.5, particle number concentra-
tions (PNC), absorbance, elemental/organic carbon, trace metals, secondary inorganic components,
endotoxin content, gaseous pollutants, and PM oxidative potential. Lung function [FEV1 (forced expi-
ratory volume in 1 sec), FVC (forced vital capacity), FEF25–75 (forced expiratory flow at 25–75% of vital
capacity), and PEF (peak expira tory flow)] and fractional exhaled nitric oxide (FENO) were measured
before and at three time points after exposure. Data were analyzed with mixed linear regression.
results: An interquartile increase in PNC (33,000 particles/cm3) was associated with an 11%
[95% confidence interval (CI): 5, 17%] and 12% (95% CI: 6, 17%) FENO increase over baseline
immediately and at 2 hr post exposure, respectively. A 7% (95% CI: 0.5, 14%) increase persisted
until the following morning. These associations were robust and insensitive to adjustment for other
pollutants. Similarly consistent associations were seen between FVC and FEV1 with PNC, NO2
(nitrogen dioxide), and NOx (nitrogen oxides).
conclusions: Changes in PNC, NO2, and NOx were associated with evidence of acute airway
inflammation (i.e., FENO) and impaired lung function. PM mass concentration and PM10 oxidative
potential were not predictive of the observed acute responses.
key words: air pollution, experimental exposure, FENO, FEV1, FVC, oxidative potential, PM,
ultrafine particles. Environ Health Perspect 120:1183–1189 (2012). http://dx.doi.org/10.1289/
ehp.1104389 [Online 2 May 2012]

Page 2

Strak et al.
1184
volume 120 | number 8 | August 2012 • Environmental Health Perspectives
charac teriza tion of PM air pollution was
performed on-site. Preexposure and post-
exposure measurements were made to assess
respiratory health.
We used a semi experi mental rather than a
pure observational design to reduce exposure
measurement error. Our design with multiple
sampling days at multiple locations used tem-
poral and spatial variability to increase contrast
in the measured PM charac teris tics and reduce
correlations. Briefly, we selected five loca-
tions in the Netherlands with different source
charac teris tics to increase exposure contrasts
and reduce correlations between PM charac-
teris tics (Strak et al. 2011). The locations were
an under ground train station, a continuous
traffic location, a stop-and-go traffic location, a
farm, and an urban background site. None of
the locations were > 70 km away from the col-
lection point located at the Utrecht University
campus where all pre exposure and some post-
exposure health measurements were made.
We aimed to recruit 32 volunteers, each
undertaking 7 study visits. Each participant
had to visit all five locations once, with the
two remaining visits assigned randomly to a
location. We planned 30 sampling days dur-
ing March through October 2009, one site
per day, with 8 participants exposed during
each visit. Results of a previous screening
phase showed much higher concentrations at
the under ground site compared with the out-
door sites, which also strongly influenced cor-
relations between specific air pollutants (Strak
et al. 2011). In order to separate the health
associations at the under ground site and the
four outdoor sites, we scheduled 9 visits at the
under ground and 21 at the remaining four
sites. Due to practical constraints, we finally
included 31 volunteers, who were measured
for an average of 5.5 sampling days (range,
3–7 days). Twenty-six partici pants visited the
under ground site at least once, and 13 visited
all five sites at least once. In summary, we
obtained 45 observations at the under ground
site and 28–37 at the other sites.
To avoid potential carryover effects from
previous exposures, an individual’s visits to the
sites were separated by ≥ 14 days. Exposure of
each participant was started between 0900 and
0930 hours and lasted for 5 hr. The partici-
pants cycled for 20 min on a stationary bicycle
each hour. We selected a 5-hr exposure period
with intermittent exercise in order to increase
the contrast between pre exposure and post-
exposure to ambient air pollution. Moreover,
we hypothesized that longer exposure would
result in a clearer health response required to
study the independent associations with differ-
ent pollutants. To keep the dose similar among
the participants, before the study began we
determined for each partici pant the heart rate
corresponding to a minute ventilation rate of
20 L/min/body surface area (square meters)
(Zuurbier et al. 2009); we then instructed
partici pants to monitor their heart rates as they
cycled in order to maintain the desired ventila-
tion rate. Exercise may affect the measured
respiratory variables, but with each participant
cycling at a consistent minute ventilation rate
each sampling day, bias in associations between
the fluctuating air pollution and health end
points is likely small. We measured lung func-
tion [i.e., forced vital capacity (FVC), forced
expiratory volume in 1 sec (FEV1), forced expi-
ratory flow between 25% and 75% of FVC
(FEF25–75), peak expiratory flow (PEF)], mea-
sured the fraction of nitric oxide in exhaled air
(FENO) as an indicator of airway inflamma-
tion, and recorded respiratory symptoms of
the participants—all respiratory health indica-
tors that are widely used. These health param-
eters were measured before exposure (t = 0;
collection point), before exposure (t = 2; sam-
pling location), immediately after exposure
(t = 7; sampling location), 2 hr after exposure
(t = 9; collection point), and the next morn-
ing (t = 25; collection point). The measure-
ments at t = 2 were performed to investigate
the effect of transport between the collection
point and the sampling location. To avoid day-
of-week effects, we obtained all measurements
on Monday through Thursday of each week.
Study population. We recruited 31 healthy,
young, non smoking participants from among
Utrecht University students living on campus
to mini mize exposure to traffic-related air pol-
lution when traveling to the collection point.
The participants completed an online
screening questionnaire. Exclusion criteria
included smoking or living in a household
with a smoker; lifetime diagnosis of asthma
or chronic obstructive pulmonary disease; or
history of cardio vascular disease, diabetes mel-
litus, or pregnancy. Before the study, each
participant was examined by a physician and
obtained medical clearance for participation.
The study was approved by the ethics com-
mittee at University Medical Center Utrecht.
Written informed consent was provided by
all participants.
Exposure measurements. The methods
for measuring air pollution on-site during
each day of partici pants’ exposure have been
described elsewhere [Strak et al. 2011; see also
Supplemental Material, p. 2 (http://dx.doi.
org/10.1289/ehp.1104389)]. Briefly, we mea-
sured PM10 and PM2.5, and we determined
the absorbance of PM2.5 samples and endo-
toxin content of PM10 samples. We made
real-time measurements of particle number
concentration (PNC) and the gaseous pol-
lutants ozone (O3), nitrogen dioxide (NO2),
and (nitrogen oxides (NOx). We measured
the concentrations of elemental carbon (EC),
organic carbon (OC), trace metals, nitrate, and
sulfate in PM2.5–10 and PM2.5 samples. OP
was measured in three fractions—PM2.5–10,
PM0.18–2.5, and PM0.18—and assessed in vitro
by measuring anti oxidant depletion of ascor-
bate (OPAA) and reduced glutathione (OPGSH)
(Godri et al. 2010).
We equipped a minibus with a custom-
made cabin air filter to minimize exposure
during transport of participants between the
collection point and the sampling locations.
To estimate traffic-related air pollution during
transport, we measured the PNC in the mini-
bus during each commute.
Clinical measurements. FENO was mea-
sured with a Niox Mino monitor (Aerocrine,
Solna, Sweden), an instrument that complies
with recommendations from the American
Thoracic Society/European Respiratory Society
(ATS/ERS 2005). The instrument was also
used to measure ambient-air nitric oxide levels
where FENO measurements were performed.
Lung function parameters (FVC, FEV1,
FEF25–75, PEF) were measured with an
EasyOne electronic spirometer (ndd, Zürich,
Switzerland), which meets ATS/ERS spirom-
etry standards (Miller et al. 2005). Each par-
ticipant performed at least three maneuvers
supervised by one of the eight technicians
operating the device. The best value from the
technically correct maneuvers was selected
according to the maximum value method of
the European Respiratory Society (Quanjer
et al. 1993). After each sampling day, two
syringe checks using a calibrated 3-L syringe
were recorded to monitor the accuracy of the
device, which had to be within 3%.
At each time point, a short questionnaire
was administered to subjectively grade respira-
tory symptoms that were present (e.g., cough,
congestion/rhinorrhea, wheeze): symptoms
were scaled from 0 (no complaints) to 3
(severe complaints).
Before the morning health measurements,
the participants completed a questionnaire
reporting additional exposure to traffic- or
workplace-related air pollution, medication
use, and so forth during the preceding 24 hr.
Data analyses. We analyzed the associa-
tions between air pollution concentrations
during exposure at the sampling locations
and the difference in lung function and FENO
between post exposure (t = 7, t = 9, t = 25)
and pre exposure (t = 0) for each sampling
day using mixed linear regression. We used
mixed models to account for the influence of
repeated observations per participant (using
compound symmetry of the residuals). We did
not include sampling location in the analy-
sis because it was not signifi cantly associated
with the outcome after including exposure
and potential confounders. We used the 5-hr
average concentrations of air pollutants mea-
sured at the locations as independent variables.
For OP and trace metals, the data from the
individual PM size fractions were aggregated.
We first specified single-pollutant models.

Page 3

PM characteristics and acute changes in respiratory function
Environmental Health Perspectives • volume 120 | number 8 | August 2012
1185
Then, to study the individual associations of
different pollutants, we specified two-pollutant
models with all possible combinations of mea-
sured pollutants. Here, we primarily report
and discuss the results of the two-pollutant
models. Models in which two pollutants had
a Spearman’s rank correlation coefficient (rS)
> 0.7 were considered highly correlated and
were not interpreted. We defined a large num-
ber of models, so we focused on the consis-
tency of significant associations and not on
single isolated significant associations. Because
there was a substantial difference in some
exposure parameters between outdoor loca-
tions and the under ground location, we also
analyzed the data separately for the outdoor
locations and the under ground location. To
offset potential confounders, we adjusted for
temperature and relative humidity measured
at the location during sampling, the season in
which the sampling day occurred (before or
after the start date of the calendar summer),
an indicator variable for low/high grasses and
birch pollen counts, and an indicator variable
for respiratory infection. Pollen counts were
obtained from the station in Leiden, located
within 50 km of our sampling locations. We
selected grass and birch pollen as highly aller-
genic indicator pollen, with good spatial cor-
relation with pollen counts at another station.
Because the distribution was highly skewed,
we included pollen as a low/high variable
(Brunekreef et al. 2000). For the lung func-
tion measurements, we investigated possible
technician and instrument effects, but neither
had an effect.
Sensitivity analysis assessed participants
who a) did not report nasal allergies, b) were
not former smokers, and c) did not take anti-
inflammatory medication in the 24-hr period
before the start of the sampling day. We
assessed the impact of influential values on
the regression results by comparing effect esti-
mates with and without the 1% of observa-
tions with the highest Cook’s distance value.
Effect estimates and their confidence
intervals (CI) are presented as percentage
increases over a study population mean of
the baseline (t = 0) values. We express these
values as percentage increases per change in
interquartile ranges (IQR) for the outdoor
locations. We calculated IQRs for endotoxin
concentrations for all locations, but without
the levels measured at the farm location. We
used these IQRs in the analysis of the com-
plete data set and the outdoor-only data set
to allow direct comparison of effect estimates.
Statistical signifi cance was defined as p < 0.05
and borderline significance as p < 0.10. All
data analy ses were carried out using SAS, ver-
sion 9.2 (SAS Institute Inc., Cary, NC, USA).
Results
A total of 170 observations were obtained
from 31 participants (Table 1) who were
exposed at least three and at most seven
times. Each participant visited the under-
ground train station at least once. We did
not analyze the data from the respiratory
symptoms questionnaires because only
questions about congestion/rhinorrhea and
cough reported > 15% changes in scores over
the sampling days.
Exposure measurements. The measure-
ments at the under ground train station
showed substantially higher concentrations
of nearly every PM charac teris tic, especially
levels of coarse PM, iron (Fe), copper (Cu)
and the sum of OPAA and OPGSH (OPTOTAL)
[see Supplemental Material, Table S1 (http://
dx.doi.org/10.1289/ehp.1104389)]. For exam-
ple, the mean concentration of coarse PM was
252 µg/m3 at the station and 13 µg/m3 at the
outdoor sites. PNC was the highest at the
continuous traffic site at 66,500 particles/cm3,
and substantially increased levels of endotoxin
were measured at the farm. Variability of con-
centrations was large at the outdoor sites but
more limited within the under ground-only
data set. Therefore, we will not report under-
ground-only air pollution effect estimates.
Correlations between air pollution con-
centrations are shown in Table 2. PM10 and
PM2.5 were highly correlated with each other,
as well as with absorbance, EC, OC, trace
metals, and OPTOTAL, but not with PNC.
The high correlations decreased considerably
after we excluded the measurements from
the under ground train station. However,
as a result of the exclusion, we observed a
substantial increase in correlations between
Table 1. Population characteristics and baseline
(t = 0) FENO and lung function.
Characteristic
Age (years)
Sex [n (%)]
Female
Male
Nasal allergy [n (%)]a
Former smoker [n (%)]
Body mass index (kg/m2)
FENO (ppb)b
FEV1 (L)c
FVC (L)c
FEF25–75 (L/sec)c
PEF (L/sec)c
Unless otherwise stated, values are mean (range).
aIncludes hay fever. bN = 151–169. cN = 165–170.
Value
22 (19–26)
21 (68)
10 (32)
5 (16)
3 (10)
22.3 (17.0–32.0)
15.9 (5–61)
3.86 (2.57–5.51)
4.68 (2.73–6.70)
3.94 (2.06–6.43)
8.71 (5.38–14.68)
Table 2. Spearman’s rank correlation coefficients (rS) between PM characteristics.
PM10 PM2.5 PM2.5–10
PNC Absa
EC(C)
PM10
0.940.820.220.74
PM2.5
0.88 0.670.150.68
PM2.5–10
0.55 0.22 0.210.71
PNC0.190.070.15
0.65
Absa
0.370.22 0.310.84
EC(C) 0.280.170.260.770.73
EC(F) 0.250.13 0.190.860.96
OC(C) 0.520.390.57–0.060.00 –0.04 –0.13
OC(F) 0.59
0.72 0.06–0.20 0.05 –0.26 –0.07
Fe(tot) 0.240.040.27 0.900.83
Fe(sol)–0.05 –0.11 –0.010.860.65
Cu(tot)0.280.120.260.820.76
Cu(sol)0.550.41 0.37 0.710.85
Ni(tot)
0.40 0.27 0.49–0.090.11 –0.11 –0.01
Ni(sol) –0.01 –0.060.00 0.460.35
V(tot) 0.140.19 –0.050.200.19
V(sol)0.040.070.00 0.19 0.14
Endo0.22 0.220.22 –0.37 –0.30 –0.49 –0.31
NO3–a
0.560.74 –0.10–0.26 –0.12 –0.05 –0.21
SO42–a
0.500.72 –0.12–0.140.08 –0.07
OPAA
0.750.790.320.280.51
OPGSH
0.54 0.480.440.120.27
OPTOTAL
0.730.730.400.22 0.42
O3
–0.21 –0.15 –0.18 –0.35 –0.57 –0.33 –0.57
NO2
0.490.450.280.560.74
NOx
0.320.210.250.750.87
Abbreviations: Abs, absorbance; C, coarse PM fraction; Endo, endotoxin; F, fine PM fraction; sol, water-soluble metal extraction; tot, total. Values in the light-blue area represent correlations in the outdoor-only data set.
aMeasured in PM2.5.
EC(F)
0.69
0.64
0.67
0.67
0.98
0.89
OC(C) OC(F) Fe(tot) Fe(sol) Cu(tot) Cu(sol) Ni(tot) Ni(sol) V(tot) V(sol) Endo NO3–aSO42–aOPAAOPGSHOPTOTAL
0.76 0.73 0.700.44 0.700.800.760.00
0.68 0.79 0.62 0.430.650.74 0.71–0.04
0.78 0.460.70 0.46 0.690.730.79 0.01
0.00 –0.040.620.650.600.56 0.070.47
0.480.49 0.92 0.750.890.920.64 0.19
0.45 0.360.88 0.750.92 0.930.540.25
0.42 0.43 0.920.74 0.890.900.600.24
0.460.49 0.27 0.54 0.57 0.62 –0.08
0.08 0.370.20 0.370.530.62–0.18
0.07 –0.22 0.780.96 0.88 0.57 0.15
0.66 –0.23 –0.27 0.80 0.790.740.400.31
0.77 0.12 –0.230.930.69 0.92 0.530.24
0.800.150.090.780.550.82 0.63 0.19
0.22 0.37 –0.10 –0.11 –0.160.13 0.07
0.46 –0.22 –0.370.270.490.360.260.11
0.29 –0.18 –0.180.040.060.210.220.160.75
0.24 –0.17 –0.300.00 0.11 0.130.130.16 0.81
0.400.13 –0.52 –0.45 –0.49 –0.420.200.08
0.110.64 –0.22 –0.27 –0.09 0.11 0.18–0.13
0.050.120.54 –0.32 –0.29 –0.190.100.330.20
0.45 0.400.380.26 0.030.350.56 0.240.07
0.28 0.50 –0.010.14 –0.150.350.470.150.02
0.350.560.24 0.23 –0.110.380.590.28–0.07
0.07 –0.06 –0.18 –0.14 –0.26 –0.35 –0.20–0.54 –0.52 –0.48 –0.21 –0.03
0.670.060.260.520.34 0.520.710.280.28
0.87 –0.110.010.700.530.660.720.140.47
O3
NO2
0.26
0.21
0.14
0.50
0.39
0.27
0.36
NOx
0.37
0.31
0.35
0.70
0.70
0.61
0.67
0.15
0.21
0.54
0.56
0.57
0.64
0.24
0.37
0.35
0.21
0.70
0.66
0.68
0.60
0.88
0.66 –0.25
0.67 –0.26
0.58 –0.27
0.25
0.66 –0.19 –0.01 –0.30
0.77 –0.06 –0.07 –0.24
0.71 –0.15 –0.02 –0.35
0.43 –0.25
0.37 –0.35
0.62 –0.26 –0.07 –0.31
0.48 –0.07 –0.15 –0.38
0.67 –0.17 –0.05 –0.24
0.66 –0.16
0.67 –0.20
0.440.74
0.20
0.96
0.050.14
0.200.06
0.490.39
0.470.36
0.210.15
0.37 0.26 –0.02
0.36
0.34
0.37 –0.29
0.07
0.18
–0.15
–0.04
–0.44
–0.16
–0.36
–0.41
–0.38
–0.22
0.08
–0.55
–0.44
–0.46
–0.31
–0.24
0.30
–0.16
0.51 –0.15 –0.22
0.10 0.27
0.670.03 –0.19
–0.10 –0.22
0.67
0.340.54
0.540.88
–0.47 –0.46 –0.21
0.320.69
0.130.54
0.88
0.91
0.73
0.36
0.80
0.79
0.79
0.70
0.66
0.72
0.51
0.76
0.82
0.72
0.11
0.79
0.82
0.79
0.79
0.32
0.73
0.77
0.73
0.77
0.48
0.69
0.43
0.75
0.80
0.68
0.11
0.70
0.89
0.88
0.77
0.35
0.78
0.80
0.76
0.77
0.60
0.72
0.45
0.76
0.83
0.73
0.04
0.76
–0.20
0.30
–0.05
–0.16
0.95
0.92
–0.67
–0.65
–0.65
–0.37
–0.81
–0.71
–0.81
–0.48 –0.01
–0.50
–0.67
–0.51
–0.70
–0.73
–0.67
–0.37
–0.80
–0.01
–0.32 –0.15 –0.19
0.170.27 –0.17
0.120.29 –0.07
–0.780.32
–0.690.14
–0.720.28
–0.36 –0.57
–0.62
–0.680.91
0.17 –0.32 –0.27

0.77
0.59 –0.07
0.220.36 0.19
0.27
0.08
0.24
0.34
0.11
0.26
0.16
0.23
0.77
0.59
0.82
0.83
0.81
0.00 –0.14
0.40 –0.10
0.06
0.27 –0.11
–0.07

0.02
0.33
0.00
0.13
0.43
0.47
0.42
0.03
0.15
–0.05 0.38

0.66
0.57
0.10
0.42
0.47
0.39
0.50
0.820.47
0.35
0.41

0.80
–0.27
0.58
0.42
0.60
0.71
0.36
0.39
0.27 –0.19
0.35 –0.23 –0.05
0.260.28
0.21
0.65

Page 4

Strak et al.
1186
volume 120 | number 8 | August 2012 • Environmental Health Perspectives
PNC and EC or absorbance; the correlations
between PNC and PM10 or PM2.5 remained.
O3 showed a strong negative correlation
with several PM charac teris tics. Overall, the
correlations between several PM charac-
teris tics were sufficiently low to investigate
their independent associations with health
end points.
Associations between air pollution and
FENO. Single-pollutant models. We observed
significant associations with a range of pol-
lutants including PNC, NOx, absorbance,
Table 4. Two-pollutant models of associations between air pollution exposure and percentage changes (postexposure – preexposure) in FVC immediately after
exposure (all sites).
IQR
13.50
11.54
8.23
32,906
3.49
4.35
0.40
1.82
0.79
895.10
32.09
57.96
8.65
3.53
1.82
2.04
1.94
0.19
5.19
2.99
19.08
15.53
38.71
9.74
10.54
28.05
Adjustment pollutant
PM10
0.02
0.90**
–0.64**
–1.26**
–1.37**
–1.36**
0.24
0.61
–0.24
–0.02
0.00
–0.01
0.01
0.10
–0.98**
–0.24
–0.56
0.00
–0.06
–0.10
0.03
0.00
0.03
2.58**
–2.19**
–1.50**
PM2.5
–0.34*
0.08
–0.21*
–1.26**
–1.39**
–1.46**
–0.04
0.50
–0.32
–0.03*
–0.06
–0.04
0.00
–0.02
–1.02**
–0.29**
–0.59
0.00
–0.08
–0.09
0.01
–0.02
0.00
2.53**
–2.27**
–1.63**
PM2.5–10
0.70**
0.60**
0.01
–1.26**
–1.11*
–1.06
0.32*
0.67*
–0.18
0.02
0.02
0.02
0.01
0.18
–0.95**
–0.17
–0.55
0.00
–0.05
–0.10
0.04
0.02
0.05
2.24**
–2.13**
–1.40**
PNC
0.03
0.11
0.03
–1.19**
0.27
0.31
0.07
0.55*
–0.05
0.00
0.16
0.01
0.01
0.05
–0.62*
0.00
–0.53
0.00
–0.35
–0.28
0.01
0.01
0.01
0.07
–1.26
–0.06
Absa
0.20**
0.50**
0.14*
–1.47**
–0.10
1.12
0.37**
0.79**
0.00
0.03**
0.12
0.04*
0.01
0.18**
–0.83**
0.00
–0.52
0.00
–0.08
–0.15
0.05**
0.03
0.05**
1.08
–2.31**
–1.64**
EC(F)
0.18**
0.47**
0.12
–1.47**
–1.10
–0.10
0.28**
0.73**
–0.01
0.03*
0.10
0.03*
0.01
0.17*
–0.83**
0.00
–0.52
0.00
–0.08
–0.15
0.05**
0.02
0.04*
1.01
–2.30**
–1.61**
EC(C)
–0.16
0.15
–0.20*
–1.34**
–1.84**
–1.56**
0.04
0.55
–0.24
–0.03
–0.08
–0.05
0.00
0.01
–1.01**
–0.21*
–0.55
0.00
–0.04
–0.07
0.02
–0.03
–0.01
2.32**
–2.22**
–1.86**
OC(F)
–0.02
0.00
–0.03
–1.26**
–0.45
–0.43
–0.02
0.50
–0.12
0.00
–0.18
–0.01
–0.01
–0.01
–0.98**
–0.10
–0.52
0.01
–0.07
–0.13
0.00
–0.01
–0.01
0.80**
–2.25**
–1.27**
OC(C)
0.05
0.17
0.03
–1.19**
–0.10
–0.09
0.08
0.52
–0.04
0.01
0.06
0.01
0.01
0.09
–0.83**
–0.02
–0.53
0.00
–0.04
–0.11
0.03**
0.03**
0.03**
0.43
–1.81**
–0.92
Fe(tot)
0.17
0.57*
–0.13
–1.27**
–1.38**
–1.41*
0.30
0.63
–0.17
0.00
0.01
0.01
0.01
0.16
–0.99**
–0.24
–0.55
0.00
–0.04
–0.08
0.04
0.01
0.03
2.20**
–2.13**
–1.44**
Fe(sol)
0.02
0.09
0.01
–1.31**
–0.19
–0.17
0.05
0.67*
–0.06
0.00
0.04
0.00
0.01
0.04
–0.86**
–0.03
–0.53
0.00
–0.05
–0.10
0.02
0.00
0.01
0.49
–1.93**
–1.22**
Cu(tot)
0.04
0.47
–0.06
–1.27**
–1.15*
–1.11*
0.32
0.67
–0.18
–0.01
–0.01
0.00
0.01
0.12
–1.02**
–0.19
–0.55
0.00
–0.04
–0.08
0.04
0.00
0.03
1.68**
–2.11**
–1.53**
Cu(sol)
0.00
0.07
–0.01
–1.25**
–0.32
–0.28
0.04
0.73
–0.16
0.00
–0.14
0.00
0.01
0.03
–0.95**
–0.06
–0.54
0.00
–0.07
–0.10
0.01
0.00
0.01
0.70*
–1.99**
–1.52**
Ni(tot)
–0.05
0.11
–0.10
–1.22**
–0.83*
–0.89*
0.03
0.51
–0.27
–0.01
–0.03
–0.01
0.00
0.04
–1.01**
–0.21*
–0.56
0.00
–0.01
–0.05
0.03
0.00
0.02
1.38**
–2.05**
–1.38**
Ni(sol)
0.05
0.16*
0.04
–1.01*
0.10
0.12
0.09*
0.67**
0.10
0.01
0.15
0.01
0.01
0.09*
–0.77**
0.14
–0.07
0.00
–0.09
–0.20
0.02
0.01
0.02
0.07
–1.50**
–0.63
PM10
PM2.5
PM2.5–10
PNC
Absa
EC(F)
EC(C)
OC(F)
OC(C)
Fe(tot)
Fe(sol)
Cu(tot)
Cu(sol)
Ni(tot)
Ni(sol)
V(tot)
V(sol)
Endo
NO3–a
SO42–a
OPAA
OPGSH
OPTOTAL
O3
NO2
NOx
Abbreviations: Abs, absorbance; C, coarse PM fraction; Endo, endotoxin; F, fine PM fraction; sol, water-soluble metal extraction; tot, total. Values in light blue boxes indicate rS > 0.7. Values in each row represent effect
estimates for the pollutants in two-pollutant models; values in dark blue boxes forming a diagonal are effect estimates in a single-pollutant model. All models were adjusted for temperature, relative humidity, season,
pollen counts, respiratory infection, and adjustment pollutant. Estimates are percentage increases above population-average baseline expressed per IQR of outdoor-sites (N = 170), except for all models including OP
(N = 153) and all models including EC(C), OC(C), and trace metals (N = 166).
aMeasured in PM2.5. *p < 0.10, **p < 0.05.
Table 3. Two-pollutant models of associations between air pollution exposure and percentage changes (postexposure – preexposure) in FENO immediately after
exposure (all sites).
IQR
13.50
11.54
8.23
32,906
3.49
4.35
0.40
1.82
0.79
895.10
32.09
57.96
8.65
3.53
1.82
2.04
1.94
0.19
5.19
2.99
19.08
15.53
38.71
9.74
10.54
28.05
Adjustment pollutant
EC(C)
0.40
0.10
0.41
11.28**11.57**
11.51**
11.80**
0.12
–2.90–1.10
–0.38
0.02
0.22
–0.12
–0.07
–0.37
1.14
–0.09
2.74
–0.07–0.08*
–2.09–2.09
–2.09–2.05
–0.15
–0.07
–0.17
–2.12–2.90
6.54
5.49
PM10
0.09
–1.43
1.02
11.28**
10.74**
12.75**
–0.41
–2.81
–0.79
–0.10
0.18
–0.18
–0.05
–0.71
1.05
–0.38
2.65
–0.07
–2.12
–1.99
–0.16
–0.12
–0.20
–2.16
6.87
5.31
PM2.5
0.68
0.17
0.41
11.26**
8.76**
10.61**
0.07
–2.48
–0.53
0.02
0.25
–0.08
–0.04
–0.33
1.15
–0.12
2.70
–0.07
–2.21
–2.02
–0.12
–0.03
–0.09
–2.72
6.98
5.49
PM2.5–10
–1.00
–0.86
0.10
11.30**
10.68**
12.58**
–0.46
–2.91
–0.88
–0.15
0.15
–0.17
–0.05
–0.83
1.00
–0.47
2.62
–0.07
–2.08
–1.97
–0.16
–0.17
–0.22
–1.64
6.81
5.18
PNC
–0.02
–0.02
–0.02
11.24**
–0.55
–0.21
–0.11
–1.92
0.28
–0.01
–0.66
–0.02
–0.04
–0.02
–0.79
–0.11
2.22
–0.01
0.46
–0.44
–0.03
–0.02
–0.03
1.56
–7.40
–5.77
Absa
–1.27**
–2.43**
–1.17**
11.80**
2.41
18.26
–1.91**
–4.58*
–1.42
–0.24**
–0.68
–0.29**
–0.10
–1.11**
–0.06
–1.45*
2.44
–0.05
–1.68
–1.54
–0.38**
–0.31**
–0.45**
10.61**
4.42
2.00
EC(F)
–1.35**
–2.64**
–1.23**
11.44**
–13.81
2.92*
–1.70**
–4.23*
–1.58
–0.26**
–0.62
–0.30**
–0.09
–1.25**
–0.08
–1.55**
2.38
–0.05
–1.59
–1.43
–0.37**
–0.27**
–0.40**
9.52**
4.12
1.52
OC(F)
0.28
0.57
0.28
OC(C)
0.22
0.38
0.22
11.07**
3.42*
4.19**
0.18
–1.04
0.12
0.02
0.41
0.02
–0.01
0.05
1.43
0.18
3.15
–0.07
–2.31
–2.15
–0.04
–0.03
–0.04
–1.61
6.53
4.65
Fe(tot)
0.73
–0.03
0.99
11.17**
12.39**
15.33**
–0.03
–2.44
–0.48
0.01
0.23
–0.24
–0.04
–0.67
1.12
–0.27
2.71
–0.07
–2.06
–2.09
–0.13
–0.06
–0.14
–2.16
6.49
5.00
Fe(sol)
0.09
0.15
0.09
11.56**
2.90
3.36*
0.08
–2.38
–0.02
0.01
0.40
0.00
–0.07
–0.01
1.18
0.07
2.86
–0.07
–2.09
–2.12
–0.02
0.00
–0.01
–1.09
6.46
4.87
Cu(tot)
0.93
1.03
0.80
11.23**
11.21**
13.09**
0.88
–2.31
–0.18
0.18
0.39
0.01
–0.04
–0.18
1.32
0.15
2.78
–0.07
–2.09
–2.14
–0.07
0.01
–0.04
–2.96
6.71
5.58
Cu(sol)
0.23
0.44
0.23
11.30**
3.95**
4.32**
0.38
–2.05
0.25
0.03
1.39
0.03
0.00
0.11
1.59
0.21
2.95
–0.08
–2.11
–2.15
0.02
0.04
0.04
–2.06
6.87
6.23
Ni(tot)
0.58
0.73
0.63
11.06**
7.07**
8.84**
0.45
–1.72
–0.03
0.08
0.42
0.04
–0.02
0.05
1.38
0.25
2.82
–0.07
–2.12
–2.23
–0.06
–0.01
–0.04
–2.47
6.62
5.15
Ni(sol)
0.07
0.12
0.08
11.30**
2.37
2.88
0.07
–1.24
–0.13
0.01
0.25
0.00
–0.02
–0.01
1.34
–0.06
2.79
–0.06
–2.03
–2.04
0.00
0.00
0.00
–0.90
6.28
4.19
PM10
PM2.5
PM2.5–10
PNC
Absa
EC(F)
EC(C)
OC(F)
OC(C)
Fe(tot)
Fe(sol)
Cu(tot)
Cu(sol)
Ni(tot)
Ni(sol)
V(tot)
V(sol)
Endo
NO3–a
SO42–a
OPAA
OPGSH
OPTOTAL
O3
NO2
NOx
Abbreviations: Abs, absorbance; C, coarse PM fraction; Endo, endotoxin; F, fine PM fraction; sol, water-soluble metal extraction; tot, total. Values in light blue boxes indicate rS > 0.7. Values in each row represent
effect estimates for the pollutants in two-pollutant models; values in dark blue boxes forming a diagonal are effect estimates in a single-pollutant model. All models were adjusted for temperature, relative humidity,
season, pollen counts, respiratory infection, and adjustment pollutant. Estimates are percentage increases above population-average baseline expressed per IQR of outdoor-sites (N = 170), except for all models
including OP (N = 153) and all models including EC(C), OC(C), and trace metals (N = 166).
aMeasured in PM2.5. *p < 0.10. **p < 0.05.
4.45**
4.87**
0.42
0.33
0.04
1.21
0.05
0.04
0.21
1.71
0.33
2.73
0.08
0.10
0.11
7.94*
5.95*

Page 5

PM characteristics and acute changes in respiratory function
Environmental Health Perspectives • volume 120 | number 8 | August 2012
1187
fine-fraction (F) EC [EC(F)], OC(F), OP, Fe,
Cu, vanadium (V), and water-soluble nickel
(Ni) [see Supplemental Material, p. 3 (http://
dx.doi.org/10.1289/ehp.1104389); see also
Strak et al. 2012, Table R1.] [Tables in the
web-based additional supplement (Strak et al.
2012) are denoted with an “R” to distinguish
them from tables in the Supplemental Material
(denoted with “S”); information presented in
the web-based additional supplement has not
been peer reviewed.]
Two-pollutant models. Immediately after
exposure, the associations for PNC at 11.2%
(95% CI: 5.5, 17.0%) remained unchanged
after adjustment for all other pollutants
(Table 3; PNC row) and cancelled out the
other significant association of EC(F) seen
in the univariate model (Table 3; PNC col-
umn). For the outdoor locations, PNC and
Fe showed the most consistent associations in
the two-pollutant models [see Supplemental
Material, Table S2 (http://dx.doi.org/10.1289/
ehp.1104389)]. We could not study individual
associations with PNC and Fe because they
were highly correlated.
Findings for 2 hr after exposure were very
similar (see Strak et al. 2012, Table R4). PNC
was most consistently associated with FENO
in the full data set. In the outdoor-only mod-
els, the consistent significant associations of
PNC, coarse-fraction (C) EC [EC(C)], water-
soluble Cu, and total Fe could not be further
disentangled (see Strak et al. 2012, Table R5).
The morning after exposure, associations
were weaker than at previous time points (see
Strak et al. 2012, Table R6). The most consis-
tent associations were found for water-soluble
Ni and V, which remained similar after adjust-
ing for all the other pollutants. However, those
associations were driven by one single influen-
tial observation and decreased and became
non significant after excluding this observation.
The water-soluble fractions of Ni and V were
too highly correlated to study their individual
associations with FENO. PNC associations at
7.3% (95% CI: 0.5, 14.1%) were less stable
than at the first two time points, with mod-
estly reduced non significant effect estimates
after adjusting for water-soluble Ni and NOx.
The associations with water-soluble Ni and V
were not present in the outdoor-only data set,
even though their concentrations were not
increased at the under ground location (see
Strak et al. 2012, Table R7).
Associations between air pollution and
lung function. Single-pollutant models. We
observed significant associations of FVC
and FEV1 with a range of pollutants includ-
ing NOx, PNC, absorbance, EC, Fe, Cu, and
water-soluble Ni. None of the exposure param-
eters were associated with changes in PEF and
FEF25–75 at any time point [see Supplemental
Material , pp. 3–4 (http://dx.doi.org/10.1289/
ehp.1104389); see also Strak et al. 2012,
Tables R2, R3].
Two-pollutant models. Immediately
after exposure, fairly consistent associations
with FVC were found for PNC, NO2, and
water-soluble Ni (Table 4). In the outdoor-
only models, we observed the most consistent
associations for PNC and NOx, respectively,
at –1.3% (95% CI: –2.4, –0.3%) and –2.4%
Table 4. continued
Adjustment pollutant
SO42–a
OPAA
0.02 –0.08
0.08–0.01
0.01–0.10
–1.32** –1.40**
–0.13 –1.13**
–0.14 –1.08**
0.04–0.02
0.50
–0.05 –0.55**
0.00–0.02
0.04 –0.13
0.00–0.02
0.01
0.04 –0.08
–0.82** –0.93**
–0.02–0.16
–0.52 –1.34**
0.00
0.06 –0.19
–0.12 –0.07
0.01
0.00 –0.03
0.01–0.07
0.34
–1.80**–2.33**
–0.90*–1.51**
V(tot)
0.13*
0.39**
0.08
–1.19**
–0.11
–0.08
0.16*
0.71**
–0.01
0.02
0.07
0.02
0.01
0.16*
–1.29**
–0.02
–0.58
0.00
–0.05
–0.11
0.03*
0.02
0.03
0.91
–2.02**
–0.99*
V(sol)
0.03
0.10
0.02
–1.18**
–0.09
–0.07
0.04
0.46
0.01
0.00
0.03
0.00
0.01
0.04
–0.73
0.02
–0.53
0.00
0.03
–0.05
0.01
0.01
0.01
0.25
–1.67**
–0.77
Endo
0.02
0.09
0.02
–1.31**
–0.08
–0.07
0.05
0.58*
–0.04
0.00
0.06
0.00
0.01
0.04
–0.78**
–0.02
–0.53
0.00
–0.06
–0.11
0.01
0.01
0.01
0.34
–2.06**
–1.01*
NO3–a
0.02
0.08
0.01
–1.41**
–0.11
–0.11
0.04
0.50
–0.04
0.00
0.04
0.00
0.01
0.04
–0.78**
–0.02
–0.54
0.00
–0.06
–0.16
0.01
0.01
0.01
0.34
–1.81**
–0.91*
OPGSH
0.00
0.20
–0.06
–1.36**
–0.82
–0.66
0.17
0.79*
–0.69**
0.00
0.00
0.00
0.00
0.01
–1.02**
–0.15
–1.40**
0.00
–0.18
–0.08
0.04
0.00
0.09
1.61**
–2.27**
–1.34**
OPTOTAL
–0.07
0.11
–0.11
–1.39**
–1.13**
–0.98*
0.07
0.68
–0.64**
–0.01
–0.06
–0.02
0.00
–0.05
–0.98**
–0.17
–1.37**
0.00
–0.19
–0.07
0.08
–0.06
0.01
1.86**
–2.32**
–1.45**
O3
0.26**
0.64**
0.21**
–1.15**
0.62
0.64
0.33**
0.85**
0.09
0.03**
0.16
0.04**
0.02
0.20**
–0.74*
0.14
–0.47
0.00
–0.05
–0.12
0.06**
0.04**
0.06**
0.34
–2.17**
–0.97
NO2
0.06
0.17*
0.04
–0.60
0.32
0.37
0.09*
0.72**
0.00
0.01
0.14
0.01
0.01
0.07
–0.54
0.05
–0.18
–0.01
–0.02
–0.03
0.02*
0.01
0.02*
–0.27
–1.82** –2.31**
0.47
NOx
0.07*
0.22**
0.06
–1.15*
0.47
0.52
0.14**
0.71**
0.06
0.01
0.23
0.02*
0.02*
0.10*
–0.66*
0.03
–0.43
0.00
–0.10
–0.13
0.03**
0.02
0.02*
–0.07
PM10
PM2.5
PM2.5–10
PNC
Absa
EC(F)
EC(C)
OC(F)
OC(C)
Fe(tot)
Fe(sol)
Cu(tot)
Cu(sol)
Ni(tot)
Ni(sol)
V(tot)
V(sol)
Endo
NO3–a
SO42–a
OPAA
OPGSH
OPTOTAL
O3
NO2
NOx
0.56
0.00
0.00
0.01
1.78**
–0.89*
Table 3. continued
Adjustment pollutant
SO42–a
OPAA
0.060.56
0.120.98
0.060.52
11.02**14.66**
2.09 10.10**
2.5411.16**
0.07
–1.10–2.88
0.00
0.01
0.32
0.00
0.00–0.01
–0.06
0.64
0.03
3.14
–0.07–0.07
–1.07 –4.08*
–2.05 –2.86
–0.01
0.01
0.00
–1.25–2.88
7.33
4.40
V(tot)
0.28
0.34
0.30
11.17**
6.60**
7.91**
0.17
–1.68
–0.20
0.03
0.33
0.00
–0.02
–0.10
1.54
0.13
2.84
–0.07
–2.09
–2.13
0.00
0.02
0.01
–2.65
6.71
4.82
V(sol)
0.08
0.15
0.08
10.94**
2.27
2.76
0.10
–0.66
–0.30
0.01
0.41
0.01
–0.01
0.02
0.06
0.00
2.84
–0.07
–2.66
–2.26
0.01
0.01
0.01
–0.88
5.93
3.98
Endo
0.00
–0.03
0.01
10.93**
1.69
2.18
–0.04
–2.21
0.06
0.00
–0.18
–0.01
–0.04
–0.08
0.92
–0.05
2.75
–0.07
–1.93
–1.94
–0.02
–0.01
–0.02
–0.04
4.89
2.95
NO3–a
0.10
0.25
0.09
11.55**
2.16
2.63
0.11
–1.04
0.52
0.01
0.37
0.01
0.00
–0.01
1.09
0.11
4.09
–0.07
–2.11
–1.32
0.01
0.02
0.02
–1.37
7.03
4.33
OPGSH
0.54
0.42
0.68
14.58**
10.51**
10.39**
0.46
–4.11
1.37
0.05
0.36
0.00
–0.04
0.15
1.21
–0.08
2.35
–0.06
–4.13*
–2.83
–0.09
0.01
–0.18
–2.15
6.43
5.47
OPTOTAL
0.69
0.83
0.70
14.63**
11.53**
11.85**
0.79
–3.63
1.37
0.08
0.49
0.04
–0.02
0.27
1.26
0.03
2.42
–0.07
–4.10*
–2.84
–0.19
0.15
0.01
–2.71
6.51
5.68
O3
NO2
0.00
–0.05
0.01
14.66** 14.85**
1.62
2.12
0.00
–1.98
–0.01
0.00
0.07
–0.01
–0.02
–0.03
0.50
–0.06
1.67
–0.05
–2.19
–2.26
–0.02
–0.01
–0.01
0.92
6.88
1.63
NOx
–0.08
–0.27
–0.06
PM10
PM2.5
PM2.5–10
PNC
Absa
EC(F)
EC(C)
OC(F)
OC(C)
Fe(tot)
Fe(sol)
Cu(tot)
Cu(sol)
Ni(tot)
Ni(sol)
V(tot)
V(sol)
Endo
NO3–a
SO42–a
OPAA
OPGSH
OPTOTAL
O3
NO2
NOx
–0.11
–0.42
–0.04
12.04**
9.58**
9.96**
–0.15
–2.77
–0.40
–0.02
0.14
–0.04
–0.04
–0.23
0.95
–0.35
2.54
–0.07
–2.17
–2.06
–0.07
–0.03
–0.07
–1.23
8.02
6.13
1.73
2.35
–0.17
–2.37
–0.35
–0.01
–0.33
–0.03
–0.06
–0.16
0.52
–0.14
1.98
–0.05
–1.89
–1.91
–0.05
–0.03
–0.05
1.40
5.03
4.65
0.71
1.28
0.07
0.66
0.06
0.34
1.29
0.11
2.47
0.01
0.08
0.20
6.56
5.82

Page 6

Strak et al.
1188
volume 120 | number 8 | August 2012 • Environmental Health Perspectives
(95% CI: –3.8, –1.0%), which decreased
and became non significant after adjusting
for O3 (and PNC after adjusting for NO2)
[see Supplemental Material, Table S4 (http://
dx.doi.org/10.1289/ehp.1104389)]. A posi-
tive association with O3 in the single-pollutant
models was consistent. For FEV1, the only
consistent pattern was a positive association
with OC(F) in the complete data set (see Strak
et al. 2012, Tables R12–R13).
Two hours after exposure, the most
consistent associations for FVC were found
with NO2 and NOx, respectively, at –1.5%
(95% CI: –2.8, –0.3%) and –1.1% (95% CI:
–2.1, –0.1%), although the latter decreased
somewhat after adjusting for PNC (see Strak
et al. 2012, Table R8). We did not interpret
models with NO2 and NOx further because
NO2 is a part of NOx. In the outdoor-only
models, the significant associations with NOx
disappeared after adjusting for O3 and EC(C)
(see Strak et al. 2012, Table R9). The positive
association with O3 was weaker after adjusting
for NOx. The association of NO2 with FEV1
at –1.6% (95% CI: –2.8, –0.3%) observed
in single-pollutant models remained stron-
ger than other associations in two-pollutant
models (see Strak et al. 2012, Table R14). In
the outdoor-only data set, NOx was most con-
sistently associated with FEV1 (see Strak et al.
2012, Table R15).
The morning after exposure, NO2 and
NOx were consistently associated with a drop
in FVC (see Strak et al. 2012, Table R10)
with the effect estimates of –1.9% (95%
CI: –3.2, –0.6%) and –1.4% (95% CI:
–2.4, –0.4%), respectively. In the outdoor
models, the associations of NOx with FVC
were insensitive to adjusting for other pol-
lutants except for O3 (see Strak et al. 2012,
Table R11), for which there was a fairly con-
sistent positive association. For FEV1, NOx
had a fairly consistent nega tive association (see
Strak et al. 2012, Table R16), whereas sulfate
had a fairly consistent positive association. In
the outdoor data set, the fairly consistent asso-
ciation with NOx remained (see Strak et al.
2012, Table R17) with an effect estimate of
–1.3% (96% CI: –2.5, –0.2%).
Additional analyses. Exposure of partici-
pants to PNC during transport to and from
the sampling sites was not associated with
changes in respiratory health and did not affect
associations with the experimental exposures.
Pollen counts were the only variables not mea-
sured on-site. Analyses with and without pollen
counts resulted in similar associations.
Exclusion of the three former smokers
and the five participants with reported nasal
allergy showed similar associations between
air pollutants and FENO, FVC, and FEV1.
Exclusion of the 12 observations with anti-
inflammatory medication did not change the
effect estimates.
Discussion
We investigated acute respiratory health
effects in a panel of healthy, young volunteers
semi experi mentally exposed to ambient air
pollution with contrasting PM charac teris-
tics. We found associations of PNC, NO2,
NOx, absorbance, EC, and trace metals with
changes in FENO, FVC, and FEV1 immedi-
ately after, 2 hr after, and the morning after
exposure. The most consistent associations in
two- pollutant models were between PNC and
FENO and between NO2/NOx and lung func-
tion. Changes in those parameters were not
consistently related to PM mass concentration,
sulfate/nitrate content, or OP of particles.
We used a semi experi mental design to
study the independent associations between
respiratory function and a large number of
PM charac teris tics. That design allowed us
to define two-pollutant models to investigate
independent associations of single pollutants
with fewer problems than observational studies.
Because we observed participants in a semi-
experi mental setting, exposure measurement
error was largely due to instrumental errors;
therefore, issues such as representativeness of
outdoor central monitoring for personal expo-
sure do not affect our study. Instrumental pre-
cision of measurements was between 5% and
10%, which is very low compared with the range
of measured concentrations. Furthermore, PNC
and NOx were not more precisely measured
than the other components [with the possible
exception of total Cu and Ni; see Supplemental
Material, Table S3 (http://dx.doi.org/10.1289/
ehp.1104389)]. Difference in instrumental pre-
cision is thus an unlikely explanation for their
stronger associations. Correlations between
PM charac teris tics were reduced by performing
repeated measurements at multiple locations
with different source charac teris tics, although
some correlations remained too high to interpret
in two-pollutant models.
In the present study, we observed a consis-
tent association between PNC and increased
FENO immediately after and 2 hr after expo-
sure. This association was insensitive to adjust-
ment by any of the 25 other pollutants. PNC
was not highly correlated with other pollutants
in the whole data set. In the outdoor-only
models, we could not disentangle the associa-
tions with PNC, absorbance, EC, Fe, and Cu
because they were highly correlated. However,
the latter components were dominant at the
under ground site but their associations were
insignificant in the data set including under-
ground, and the PNC association was consis-
tently strong in both the whole data set and
the outdoor-only data set, thus suggesting that
the associations with absorbance, EC, Fe, and
Cu for the outdoor sites are likely explained by
those with PNC.
PNC especially reflects the ultrafine par-
ticles for which respiratory health effects have
been documented in previous epidemiological
and controlled exposure studies (Ibald-Mulli
et al. 2002). The consistency of PNC effects
is notable when we take into account the low
correlation of FENO with FVC and FEV1.
Our results for lung function are contrary
to findings of two experimental studies with
higher PNC concentrations than in our study
(Larsson et al. 2007; Samet et al. 2009). The
shorter duration of these studies (2 hr) may
explain the lack of a lung function response. In
a study in Arnhem, the Netherlands, Zuurbier
et al. (2011) reported a 5% increase in FENO
in response to an 18,000-particles/cm3 (IQR)
increase in PNC measured 6 hr after a 2-hr
exposure in a car or bus. This increase corre-
sponds to a 9% increase in FENO if expressed
per our IQRs, roughly comparable with our
findings. Similarly, Strak et al. (2009) reported
a 13% increase in FENO in cyclists after a 1-hr
commute, as expressed per the IQRs in our
study. McCreanor et al. (2007) reported that a
2-hr walk near heavy diesel traffic resulted in an
approximately 4% increase in FENO expressed
per our IQR.
Similarly consistent associations were
observed between NO2 and NOx and lung
function parameters 2 hr after and the morn-
ing after the exposure. PNC was most con-
sistently related to FVC immediately after
exposure. There is still a debate whether the
associations observed between respiratory
health and NO2 at the concentrations cur-
rently found in western European countries are
due to direct effects of NO2 or other PM com-
ponents co-varying with NO2 (WHO 2006).
In the present study, we measured a detailed
set of PM charac teris tics including PNC as a
proxy for ultrafine particles and metals, but we
still observed associations with NO2.
For some components that were higher
at the under ground station than at other
locations, we observed associations only in
the outdoor-only models. This likely argues
against a causal role, although saturation of
biological parameters after exposure to very
high air pollution concentrations could also
provide an explanation.
Exposure to PM mass, irrespective of the
size fraction, was not associated with acute
respiratory health changes. The lack of asso-
ciation between PM10 and PM2.5 and acute
changes in respiratory function is consistent
with the results from two Dutch studies on
bicycle commuting (Strak et al. 2009; Zuurbier
et al. 2011) in which increases in FENO were
not associated with mass-based PM metrics but
PNC and absorbance were.
In contrast to our hypothesis, OP did
not display a strong and consistent relation-
ship with acute respiratory health effects. In
the single-pollutant models, associations of
OP metrics with increased FENO 2 hr after
exposure were evident in the outdoor-only

Page 7

PM characteristics and acute changes in respiratory function
Environmental Health Perspectives • volume 120 | number 8 | August 2012
1189
models, whereas they disappeared in the two-
pollutant models. In our primary analyses, we
used the OP of aggregated PM fractions. A
secondary investigation of OP of the small-
est available fraction (PM0.18) did not show
consistent associations with the measured
respiratory parameters, nor did it reduce the
associations observed for PNC and NOx.
Our characterization of OP focused on PM
and did not include oxidative properties of
other co-pollutants, such as O3 and NO2.
However, in the two-pollutant models with
those co-pollutants, we did not observe con-
sistent associations with OP. Our measure-
ment of AA or GSH depletion is one of the
existing methods of PM OP determination,
and it is possible that other methods would
show associations that we did not observe.
The assay we used examined only the intrinsic
potential of PM to drive oxidation reactions
in an acellular model—reflecting its redox-
active transition metals and quinone content.
Upon inter action with airway cells, PM can
elicit oxidative stress through alternative path-
ways; therefore, this assay can account for
only a fraction of in vivo PM activity.
Conclusions
Changes in PNC, NO2, and NOx were asso-
ciated with evidence of acute airway inflam-
mation (FENO) and impaired lung function.
These associations were robust and insensitive
to adjustment for other pollutants. PM mass
concentration and PM10 OP were not predic-
tive of the observed acute responses.
RefeRences
ATS/ERS (American Thoracic Society, European Respiratory
Society). 2005. ATS/ERS recommendations for standard-
ized procedures for the online and offline measurement
of exhaled lower respiratory nitric oxide and nasal nitric
oxide, 2005. Am J Respir Crit Care Med 171:912–930.
Ayres JG, Borm P, Cassee FR, Castranova V, Donaldson K, Ghio A,
et al. 2008. Evaluating the toxicity of airborne particulate
matter and nanoparticles by measuring oxidative stress
potential—a workshop report and consensus statement.
Inhal Toxicol 20:75–99.
Brunekreef B. 2010. The color of smoke. Epidemiology 21:903–904.
Brunekreef B, Forsberg B. 2005. Epidemiological evidence of
effects of coarse airborne particles on health. Eur Respir J
26:309–318.
Brunekreef B, Hoek G, Fischer P, Spieksma FTM. 2000. Relation
between airborne pollen concentrations and daily
cardiovascular and respiratory-disease mortality. Lancet
355:1517–1518.
Brunekreef B, Holgate ST. 2002. Air pollution and health. Lancet
360:1233–1242.
Godri KJ, Duggan ST, Fuller GW, Baker T, Green D, Kelly FJ, et al.
2010. Particulate matter oxidative potential from waste trans-
fer station activity. Environ Health Perspect 118:493–498.
Ibald-Mulli A, Wichmann H-E, Kreyling W, Peters A. 2002.
Epidemiological evidence on health effects of ultrafine
particles. J Aerosol Med 15:189–201.
Larsson B-M, Sehlstedt M, Grunewald J, Sköld CM, Lundin A,
Blomberg A, et al. 2007. Road tunnel air pollution induces
bronchoalveolar inflammation in healthy subjects. Eur
Respir J 29:699–705.
McCreanor J, Cullinan P, Nieuwenhuijsen MJ, Stewart-Evans J,
Malliarou E, Jarup L, et al. 2007. Respiratory effects of
exposure to diesel traffic in persons with asthma. N Engl J
Med 357:2348–2358.
Miller MR, Hankinson J, Brusasco V, Burgos F, Casaburi R,
Coates A, et al. 2005. Standardisation of spirometry. Eur
Respir J 26:319–338.
Pope CA III, Dockery DW. 2006. Health effects of fine particu late
air pollution: lines that connect. J Air Waste Manag Assoc
56:709–742.
Quanjer PH, Tammeling GJ, Cotes JE, Pedersen OF, Peslin R,
Yernault JC. 1993. Lung volumes and forced ventilatory
flows. Report Working Party Standardization of Lung
Function Tests, European Community for Steel and Coal.
Official statement of the European Respiratory Society.
Eur Respir J Suppl 16:5–40.
Samet JM, Rappold A, Graff D, Cascio WE, Berntsen JH,
Huang Y-CT, et al. 2009. Concentrated ambient ultrafine
particle exposure induces cardiac changes in young healthy
volunteers. Am J Respir Crit Care Med 179:1034–1042.
Strak M, Boogaard H, Meliefste K, Oldenwening M, Zuurbier M,
Brunekreef B, et al. 2009. Respiratory health effects of
ultrafine and fine particle exposure in cyclists. Occup
Environ Med 67:118–124.
Strak M, Janssen NA, Godri KJ, Gosens I, Mudway IS, Cassee FR,
et al. 2012. Additional Supplementary Material: Respiratory
Health Effects of Airborne Particulate Matter: The Role of
Particle Size, Composition and Oxidative Potential - The
RAPTES Project. Available: http://www.raptes.nl/rsc/pdf/
strak-2012-raptes-ehp.pdf [accessed 5 July 2012].
Strak M, Steenhof M, Godri KJ, Gosens I, Mudway IS, Cassee FR,
et al. 2011. Variation in charac teris tics of ambient partic-
ulate matter at eight locations in the Netherlands - the
RAPTES project. Atmos Environ 45:4442–4453.
Valavanidis A, Fiotakis K, Vlachogianni T. 2008. Airborne particu-
late matter and human health: toxicological assessment
and importance of size and composition of particles for oxi-
dative damage and carcinogenic mechanisms. J Environ
Sci Health C Environ Carcinog Ecotoxicol Rev 26:339–362.
WHO (World Health Organization). 2006. Air Quality Guidelines.
Global Update 2005. Particulate Matter, Ozone, Nitrogen
Dioxide, and Sulfur Dioxide. Copenhagen:WHO Regional
Office for Europe. Available: http://www.euro.who.
int/__data/assets/pdf_file/0005/78638/E90038.pdf [accessed
6 July 2012].
WHO (World Health Organization). 2007. Health Relevance of
Particulate Matter from Various Sources. Available: http://
www.euro.who.int/__data/assets/pdf_file/0007/78658/
E90672.pdf [accessed 6 July 2012].
Zeger SL, Thomas D, Dominici F, Samet JM, Schwartz J,
Dockery D, et al. 2000. Exposure measurement error in time-
series studies of air pollution: concepts and consequences.
Environ Health Perspect 108:419–426.
Zuurbier M, Hoek G, Oldenwening M, Meliefste K, van den Hazel P,
Brunekreef B. 2011. Respiratory effects of commuters’ expo-
sure to air pollution in traffic. Epidemiology 22:219–227.
Zuurbier M, Hoek G, van den Hazel P, Brunekreef B. 2009. Minute
ventilation of cyclists, car and bus passengers: an experi-
mental study. Environ Health 8:48; doi:10.1186/1476-069X-
8-48 [Online 27 October 2009].