degree of traffic intensity.
Several recent studies show associations between air
pollution and health (Brunekreef and Holgate, 2002).
Science of the Total Environment 3*The personal exposure of 14 children aged 9–12 years to ‘soot’, NOx (NO and NO2) was assessed in Amsterdam between
March and June 2003. Each child's personal exposure was monitored during four repeated 48-h periods. Concurrently, in- and
outdoor NOx measurements were carried out at the school and at the home of each participating child. Measurements were
supplemented by a questionnaire on time activity patterns and possible indoor sources. Flow-controlled battery operated pumps in a
made-to-fit backpack were used to sample personal exposure to ‘soot’, determined from the reflectance of PM2.5 filters. Exposure
to NOx was assessed using Ogawa passive samplers. Children living near busy roads were found to have a 35% higher personal
exposure to ‘soot’ than children living at an urban background location, despite that all children attended the same school that was
located away from busy roads. Smaller contrasts in personal exposure were found for NO (14%), NO2 (15%) and NOx (14%). This
finding supports the use of ‘living near a busy road’ as a measure of exposure in epidemiological studies on the effects of traffic-
related air pollution in children.
© 2006 Elsevier B.V. All rights reserved.
Keywords: Air pollution; Traffic; Personal exposure; Validation; Children; Soot; Nitrogenoxides
1. IntroductionSofie Van Roosbroeck a,⁎, Janine Wichmann b, Nicole A.H. Janssen c, Gerard Hoek a,
Joop H. van Wijnen d, Erik Lebret c, Bert Brunekreef a,e
a Institute for Risk Assessment Sciences, Utrecht University, P.O. Box 80176, 3508 TD Utrecht, The Netherlands
b School of Health Systems and Public Health, University of Pretoria, South Africa
c Center for Environmental Health Research, National Institute for Public Health and the Environment (RIVM), Bilthoven, The Netherlands
d Department of Environmental Medicine, Municipal Health Service Amsterdam (GG&GD), The Netherlands
e Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht, The Netherlands
Received 7 December 2005; received in revised form 10 March 2006; accepted 23 March 2006
Available online 2 May 2006
Abstract
Several recent studies suggest an association between long-term exposure to traffic-related air pollution and health. Most studies
use indicators of exposure such as outdoor air pollution or traffic density on the street of residence. Little information is available
about the validity of these measurements as an estimate of long-term personal exposure to traffic-related air pollution. In this pilot
study, we assessed outdoor and personal exposure to traffic-related air pollution in children living in homes on streets with differentLong-term personal exposure to traffic-related air pollution
among school children, a validation studyCorresponding author. Tel.: +31 30 253 5400; fax: +31 30 253
9499.
E-mail address: s.a.h.vanroosbroeck@iras.uu.nl (S. VanRoosbroeck).
0048-9697/$ - see front matter © 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.scitotenv.2006.03.03468 (2006) 565–573
www.elsevier.com/locate/scitotenvResults of three prospective cohort studies have sug-
gested that long-term exposure to particulate matter

the To(PM) air pollution is associated with increased mortality
from respiratory and cardiovascular disease and lung
cancer (Abbey et al., 1999; Dockery et al., 1993; Pope et
al., 1995). These studies have compared several large
study regions with different ambient air pollution con-
centrations, on the assumption that exposure was uni-
form within each region.
Due to recent reports of a significant variation of
outdoor traffic-related air pollution within cities, a Dutch
cohort study assessed exposure to air pollution on a
smaller spatial scale by taking the proximity to major
roads into account using a geographic information
system (GIS) (Hoek et al., 2002a). Participants who
lived closer to major roads had a significantly increased
risk of death resulting from cardiorespiratory causes
(Hoek et al., 2002a).
Several cross-sectional studies have also shown asso-
ciations between traffic-related air pollution and adverse
health effects (Delfino, 2002). These studies used indi-
cators of exposure, such as traffic density on the street of
residence, distance between the home and busy roads
and/or estimated outdoor concentrations based on such
characteristics. Little information, however, is available
about the validity of these measurements as an esti-
mate of long-term personal exposure to traffic-related air
pollution.
The availability of validation data for short-term ex-
posure studies is far more than that available for long-
term monitoring. Several studies have documented that
the temporal variation in outdoor particulate matter air
pollution is reflected in temporal variation of personal
exposure (Janssen et al., 1999, 2000). However, these
studies do not provide information on the validity of
outdoor air pollution concentrations for long-term expo-
sure studies, which require spatial contrast in average
outdoor air pollution.
A study conducted in the Netherlands at three schools
near freeways with a range of traffic intensities from
45,000 to 150,000 cars/day showed significant differ-
ences in the long-term average personal nitrogen dioxide
(NO2) exposure of school children (Rijnders et al.,
2001). Rijnders et al. found an estimated difference of
8.2 μg/m3 (SE 1.8) between personal NO2 exposure of
the children attending the school with the highest and
lowest traffic intensity; a difference of 46% (Rijnders et
al., 2001). The increase in school outdoor NO2 for these
children was 41%, whereas the difference in home
outdoor NO2 concentration was 28% (Rijnders et al.,
2001). A study by Monn also focused on long-term
exposure and showed highly significant correlations
(R2N0.9) on a city-level between outdoor and personal
566 S. Van Roosbroeck et al. / Science ofannual mean estimates of exposure to NO2 (Monn,2001). However, no long-term studies have involved
personal sampling of the probably more relevant
particulate matter, with a 50% cut off of 2.5 μm in
aerodynamic particle size (PM2.5), and particulate
components such as ‘soot’. This lack of data hinders
the interpretation of epidemiological studies on long-
term air pollution exposures. In this pilot study we
therefore aimed to evaluate the feasibility of personal
monitoring for PM, since these procedures are known to
be highly demanding for the participants. Since
epidemiological studies on long-term effects of traffic-
related air pollution identify children as a sensitive
group, we selected school children as participants in this
study. We assessed personal exposure to traffic-related
air pollutants, PM2.5, ‘soot’, and NOx in locations with
varying degrees of traffic intensity. The overall objective
of this study was to test the validity of traffic-related
characteristics as an estimate for the personal long-term
exposure to traffic-related air pollution, including PM2.5,
‘soot’ and NOx.
2. Methods
2.1. Participant selection
We conducted a pilot study in an urban background
school in Amsterdam.With the cooperation of the school
board, 40 children from grades 7 and 8 (9 to 12 years of
age) were asked to participate in the study. These chil-
dren received an invitation with a cover letter explaining
the purpose of the study. Candidates were asked to return
a participation form and parents had to sign an informed
consent form.
2.2. Study design
Personal exposure to traffic-related air pollution was
monitored 4 times per child in March, April, May and
June of 2003. Children carried a personal PM2.5 sampler
and an Ogawa passive sampler, to provide personal NOx
measurements, continuously for 48 h. Concurrently,
home indoor and outdoor measurements of NOx, and
school indoor and outdoor NOx, and PM2.5 concentra-
tions were collected using identical sampling equipment.
Light absorbance was measured from all PM2.5 filters as
a proxy for ‘soot’ or elemental carbon. Absorption coef-
ficients of PM2.5 filters have been shown to be highly
correlated with measurements of elemental carbon (EC)
or ‘soot’ (Cyrys et al., 2003; N Janssen et al., 2001; NA
Janssen et al., 2000; Kinney et al., 2000). Elemental
carbon or ‘soot’ is a product of incomplete combustion
tal Environment 368 (2006) 565–573and has been found to correlate with diesel exhaust