indoor and outdoor personal exposure to benzene in athens, greece
TRANSCRIPT
www.elsevier.com/locate/scitotenv
Science of the Total Environ
Indoor and outdoor personal exposure to benzene
in Athens, Greece
Christos Chatzis, Evangelos C. AlexopoulosT, Athena Linos
Department of Hygiene and Epidemiology, University of Athens Medical School, Athens, Greece
Received 15 December 2004; accepted 24 January 2005
Available online 15 April 2005
Abstract
Objective: To evaluate the exposure of urban inhabitants to atmospheric benzene in Athens, Greece.
Methods: Fifty non-smoker volunteers from selected occupational groups and their homes were monitored by passive air
samplers for six 5-day periods during a year. An activity diary was completed during each sampling period and relevant data
were collected by a questionnaire at the beginning of the study. Additional data on urban levels on benzene were also available.
Results: Average benzene home and personal levels in six monitoring campaigns varied between 6.0–13.4 and 13.1–24.6 lg/m3, respectively. Urban levels varied between 15.4 and 27.9 lg/m3 with an annual mean of 20.4 lg/m3. Wind speed seems to
determine largely home levels and personal exposure. Proximity to busy road holds also an important influence on indoor
benzene levels. Adjusted for seasonal or climate variation, other significant prognostic factors of personal exposure were home
levels, total time spent outdoors and transportation mean. Time spent outdoors explains the strong relationship between
occupation and personal levels of exposure. Wind had similar effect in clearing indoor and urban pollution in Athens; lessen
personal exposure and home levels about 2–2.5 lg/m3 per 1 m/s increase in speed.
Conclusions: Factors related to climate (use of non-absorbent materials for wall and floor covering and frequent ventilation)
might be one explanation for homes’ high clearing rate. Our exposure pattern, which suggests that outdoors work give the
greater contribution to benzene exposure of Athens citizens, is uncommon in northern towns of Europe. Policy makers have to
take in account these differences in establishing guidelines for ambient benzene exposure.
D 2005 Elsevier B.V. All rights reserved.
Keywords: Benzene; Exposure assessment; Linear mixed model analysis; Health policy
0048-9697/$ - see front matter D 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.scitotenv.2005.01.034
T Corresponding author. Current address. P.O. Box 23 V. Mela St.,
GR 15562 Cholargos, Athens, Greece. Fax: +30 210 5573518.
E-mail addresses: [email protected],
[email protected] (E.C. Alexopoulos).
1. Introduction
Benzene has raised the focus of interest because of
its known genotoxic carcinogenicity even at typical
ambient concentrations (World Health Organization,
1993; Crump, 1994; Karacic et al., 1995). The World
Health Organization stated a risk level between 4.4
ment 349 (2005) 72–80
Table 1
Personal and job characteristics among study population (n=50)
Mean SD n %
Age 36.3 10.9
Sex
Men 33 66
Women 17 34
Total time spending outdoors (h)T 27.3 13.8
Transportation mean
By foot 7 14
By vehicle 35 70
Both 8 16
Use of vehicle in work 28 56
Smoker roommate 15 30
T Duration of each measurement period=108 h.
Table 2
Participants’ house characteristics
n %
Type & floor of house
N4th 16 32
2nd–3rd 13 26
ground–1st floor 14 28
detached house 7 14
City area
Center 42 84
Suburbs 8 16
Recent painting 6 12
Heating mode
Fireplace 5 10
Oil oven 2 4
Natural gas oven 4 8
Central heating 39 78
Proximity to benzene station (b50 m) 3 6
Proximity to busy road (b50 m) 16 32
C. Chatzis et al. / Science of the Total Environment 349 (2005) 72–80 73
and 7.5 myeloid leukaemia cases every million people
exposed continuously to 1 lg/m3 (World Health
Organisation Regional Office for Europe, 2000).
According to a European Union directive entered into
force on December 2000, benzene concentrations in
ambient air should not exceed 5 lg/m3 as a running
annual average, with a long term target of less than 1
lg/m3 (OJ, L 313, 2000).
Previous studies of personal exposure have high-
lighted prominent sources of benzene (Wallace,
1989). Outdoor exposures to benzene are primarily
connected with vehicle emissions, including both
exhaust and evaporative losses of benzene. Studies
have shown that in areas of high traffic density
concentrations in vehicles are typically many-fold
higher than are measured at urban background
locations (Gee and Perry, 1994; European Environ-
ment Agency, 2001). Sources of indoor hydrocarbons
are numerous, including building materials, solvents,
adhesives, and environmental tobacco smoke.
Tobacco smoke is the most prominent source of
indoor pollution, with benzene concentrations increas-
ing by up to 50% compared with homes of non-
smokers (Wallace and Pellizzari, 1987). However, due
to low dilution rates and the amount of time spent
indoors, even minor sources can contribute consid-
erably to exposure. Such sources include several
consumer products which when combined together
can account for up to 20% of the exposure of the total
population to benzene (Wallace et al., 1987).
In Greece only very recently existing ambient air
networks have started monitoring concentrations of
aromatic hydrocarbons in the atmosphere and very
few results are officially available. Research studies
in the field are sparse to give a consistent estimation
of environmental levels (Moschonas et al., 2001;
Bakeas and Siskos, 2002; Petrakis et al., 2003). In
most studies, time integrated measurements at fixed
point monitoring stations at urban background
locations are used for monitoring trends in atmos-
pheric concentrations but also to evaluate the general
pollution climate. Research on personal exposures to
pollutants has shown that the measurements at fixed
point monitoring stations at urban background
locations may not well represent the exposures of
individual members of the general population
(Cocheo et al., 2000). A few studies showed a good
correlation between the estimation of personal
exposure directly by attaching personal samplers to
people, or by fixed point monitors in individual
microenvironments combined with activity data
defining the times spent in each of the microenviron-
ments. In any case little comparative work has been
conducted (Leung and Harrison, 1998; Maitre et al.,
2002; Bono et al., 2003).
Athens, the capital of Greece, is the most densely
populated town in country, located in a latitude of
37.99 and longitude 23.75, spread in an area of 284.6
km2 with an approximate population of 4 million.
Emission of benzene has calculated as 2159.3 tonnes
per year which ranked as one of the highest between
European towns (European Environment Agency,
2001).
Table 3
Personal, home and environmental benzene measurement (Ag/m3 ) in six periods
Periods Personal Home Urban
Median 10th percentile 90th percentile Median 10th percentile 90th percentile Median
Benzene 1st Sep 23.2 14.4 37.5 13.4 5.6 23.2 26.0
(Ag/m3) 2nd Dec 20.2 12.8 29.7 11.2 5.4 22.2 20.0
3rd Feb 17.0 9.7 34.3 10.2 6.5 15.6 18.5
4th Apr 14.3 7.7 28.4 9.0 4.0 15.0 16.1
5th Jun 12.4 5.3 20.3 5.4 2.7 10.5 13.3
6th Sep 16.7 8.7 30.3 7.8 4.3 12.5 18.9
C. Chatzis et al. / Science of the Total Environment 349 (2005) 72–8074
In this study we have used samplers to monitor
volunteers, their homes and urban sites for one year.
Volunteers were non smokers and were selected
from five occupational groups (traffic policemen,
bus drivers, postmen, teachers and students). The
first three spent a lot of time in the streets, so it
was expected to be more heavily exposed compared
to teachers and students who spend more time
indoors.
2. Methods
2.1. Sampling
The whole sampling campaign repeated every 2
months and lasted 1 year (September 1997–Septem-
ber 1998). Each monitoring campaign lasted from
Monday morning to Friday evening and was carried
Table 4
Personal, home and environmental benzene measurements (Ag/m3) and cl
Profession Total Periods
l sd 1st Sep 2nd De
l sd l
Benzene
(Ag/m3)
Teachers 13.4 5.7 20.4 6.6 15.7
Bus drivers 24.7 13.0 32.1 9.0 22.7
Postmen 19.5 9.1 25.2 2.6 19.4
Traffic policemen 22.2 9.4 24.9 11.6 21.5
Students 14.0 7.1 21.0 5.1 22.5
All persons 18.9 10.2 24.6 8.2 20.4
Homes 10.2 6.5 13.4 6.5 13.3
Environmental 20.7 9.3 27.9 11.3 21.9
Temperature (8C) 23.2 13.0
Wind speed (m/s) 0.5 0
Humidity (%) 57 70
Sunshine (h) 9.9 4.2
out placing radial path diffusive samplers—radiello
(Fondazione S Maugeri, Pavia, Italy)—in hundred
sites of Athens. At the same time, fifty volunteers
and their homes were monitored for the same
duration. The campaign included 1200 measurements
and losses did not exceed 6% in any sampling
period. Experimental database composed of 569
environmental data, 280 and 286 personal and home
pollution data, respectively.
For urban monitoring one hundred sampling sites
have been chosen, distributed along the knots of a
multi-scale grid drawn over the town map. The mesh
size was approximately 400 m. The sampling sites
have been divided among 15% and 75% of hot spots
and background zones, respectively (i.e. almost 8
km2 including city center and its peripheral zone).
The other 10% was chosen in a periurban area
(northeast) with scarce traffic and many copses. For
six times each sampling site has been uninterruptedly
imate parameters in six periods (l: mean, sd: standard deviation)
c 3rd Feb 4th Apr 5th Jun 6th Sep
sd l sd l sd l sd l sd
3.3 14.3 4.5 10.3 2.6 7.9 2.5 12.8 3.6
4.4 24.8 10.5 25.5 24.4 18.5 10.6 25.3 7.8
7.0 20.4 9.4 22.0 13.3 13.9 7.2 16.4 8.6
7.0 24.3 12.8 21.9 8.9 16.9 5.5 24.1 9.1
6.1 13.1 3.7 11.1 5.2 6.7 3.1 9.6 3.0
6.0 19.3 9.7 18.1 14.5 13.1 7.9 17.8 9.1
10.4 11.3 4.7 9.5 5.4 6.0 3.0 8.0 3.2
8.8 20.3 7.2 17.3 6.4 15.4 7.2 21.3 8.7
12.5 10.4 23.6 21.8
2.4 2.2 4.0 3.7
81 62 54 69
2.7 7.6 10.2 7.2
C. Chatzis et al. / Science of the Total Environment 349 (2005) 72–80 75
monitored from Monday morning (6–8 a.m.) to
Friday afternoon (6–8 p.m.). Samplers were on site
placed inside its shelter hung to a lamppost about 3
m high. Personnel employed at the beginning of the
campaign for placing the samplers and at its end for
gathering them has been educated and supervised by
the members of the research project (CC and ECA).
The volunteers were selected based on expected level
of exposure. So people who, due to the duties of
their job, spend a lot of time in the street like bus
drivers, policemen and postmen composed the first
group. The group of the lower expected exposure
comprised teachers and students. The selection of
volunteers was based in two criteria: non-smoking
and job located in the city centre. Ten persons of
each of the following groups, teachers, students, bus-
drivers, postmen and policemen were selected. After
the selection of 50 volunteers, personal exposure was
determined directly over the series of six 108 h (4 1/
2 days) sampling periods. The personal sampler used
was attached to the volunteer’s lapel and during
night was set in bedside table. At the end of each
day in each sampling period, volunteers were
instructed to fill in a diary stating their activities
during the sampling period.
period
4th Apr
3rd Feb
2nd Dec
1st Sep
benzene µ
g/m
3 :
108 h
TW
A
30,0
25,0
20,0
15,0
10,0
5,0
0,0
9,010,2
11,2
13,414,3
17,0
20,2
23,2
Fig. 1. Benzene levels (median valu
Home microenvironments monitored were located
in greater Athens basin. Although all volunteers
worked in city centre, about two thirds of all
volunteers’ houses were located outside city centre,
in less urban regions and a few houses were located in
suburbs of Athens. These areas had variant densely
population and traffic density and therefore, were
typically less congested and ambient air concentra-
tions were expected to be lower.
Sampling was performed by a radial symmetry
passive sampler radiello. The sampler works by the
spontaneous transfer of gaseous molecules through a
diffusive barrier. It is composed of a microporous
cylindrical diffusive body and of an absorbing
cartridge, cylindrical also, placed inside the diffu-
sive body and coaxial with it. Once assembled
radiello is exposed and just the day and time of the
exposure beginning and end are needed to know
(Cocheo et al., 1996). Subjects and their homes
were also equipped with radiello to measure the
time weighted average (TWA) concentrations of
benzene in the breathing zone over an 108 h period.
Samplers were desorbed with carbon disulfide,
shaken for 30 min and analysed by gas chromatog-
raphy coupled with mass spectrometry. The method
6th Sep
5th Jun
town
personal exposure
homes
7,8
5,4
16,7
12,4
es) in six monitoring periods.
Table 5
Univariate associations between personal benzene levels (Ag/m3)
and characteristics of study population (regression coefficient (beta),
CI: confidence interval, RC: reference category)
Beta (95% CI) p
Sex
Woman RC
Man 5.14 (1.90, 8.38) 0.003
Age (year) 0.205 (0.053, 0.357) 0.009
Occupation
Students RC
Teachers �0.88 (�5.10, 3.34) 0.674
Postmen 4.37 (0.17, 8.58) 0.042
Policemen 7.99 (3.56, 12.42) 0.001
Bus drivers 9.74 (5.54, 13.94) b10�3
Time spent outdoors (h) 0.324 (0.234, 0.414) b10�3
Transportation mean
By foot RC
Vehicle 7.13 (2.67, 11.58) 0.003
Combination 3.78 (�1.94, 9.50) 0.189
Using vehicle during work
No RC
Yes 5.11 (1.95, 8.28) 0.002
Proximity to benzene station
No RC
Yes 7.11 (0.33, 13.89) 0.040
Home concentration (Ag/m3) 0.537 (0.355, 0.718) b10�3
C. Chatzis et al. / Science of the Total Environment 349 (2005) 72–8076
allows a detection limit as low as 0.2 lg/m3.
Analytical method has been described in details
elsewhere (Cocheo et al., 1996). The overall
reliability of the sampling device has been judged
as excellent by the European Reference Laboratory
for Air Pollution (ERLAP) of the Joint Research
Centre of Ispra and the sampler has been widely
used (Cocheo et al., 2000; Crebelli et al., 2001;
Kouniali et al., 2003).
Table 6Univariate associations between home benzene levels (Ag/m3) and
characteristics of homes (regression coefficient (beta), CI: con-
fidence interval, RC: reference category)
Beta (95% CI) p
Proximity to benzene station
No RC
Yes 3.92 (0.42, 7.42) 0.029
Smoker roommate
No RC
Yes 1.60 (�0.34, 3.54) 0.100
Home location
Suburbs RC
Centre 2.69 (0.30, 5.09) 0.029
Proximity to busy road
No RC
Yes 2.59 (0.74, 4.43) 0.007
3. Statistical analysis
Univariate analyses were performed to examine the
covariates age, sex, roommates’ smoking habit, job
title, time spent outdoors, transportation mean, and
house characteristics (floor, location, recent, painting,
heating mode, proximity to petrol station, and
proximity to busy road). We applied linear mixed-
effects regression models to estimate the significant
prognostic factors of benzene exposure levels by
using maximum likelihood and restricted/residual
maximum likelihood methods. Mixed effects models
had the advantage of adjusting for invariant variables
by fixed-effects models and accounting for individual
differences by random-effects models. In our mixed-
effects models, we treated subjects’ personal and
home characteristics as fixed effects and each subject
as a random effect. Measurement period was used to
identify repeated observations. The type III sum of
squares was used to calculate the effects in the
models. The multivariate linear mixed model included
all variables that contributed significantly to the final
model (Wald statistics, criterion of pb0.05). For each
factor the regression coefficient and 95% confidence
interval (95% CI) were calculated. All statistical
analyses were performed with SPSS 11.0 software.
4. Results
In Tables 1 and 2 personal and home characteristics
of study population are presented. Time spent out-
doors differed markedly between occupational groups.
Teachers and students spent less than 15 h in each
sampling period (108 h) as an annual mean while bus
drivers and traffic policemen exceeded 40 h.
The mean and median concentrations of benzene
measured during the six sampling periods are pre-
sented in Tables 3 and 4 and Fig. 1. As shown in these
tables, average personal levels ranged 13.08–24.63
Ag/m3 and the highest values were observed during
the first two periods, i.e. Autumn and Winter, periods
that wind speed did not exceed 0.5 m/s (Table 4).
Home levels varied between 6.03 and 13.35 Ag/m3,
Students
Traffic policemen
Postmen
Bus drivers
Teachers
All volunteers
Urban level
benze
ne
µg
/m3
30,0
25,0
20,0
15,0
10,0
5,0
0,0
town
personal exposure
homes
8,78,78,3
9,6
8,48,7
12,0
19,118,2
22,1
12,2
17,3
18,5
Fig. 2. Benzene annual average (median) levels.
C. Chatzis et al. / Science of the Total Environment 349 (2005) 72–80 77
resulting in an annual mean of 10.19 Ag/m3 followed
the same trend. Average urban levels varied between
15.4 and 27.9 Ag/m3 resulting in annual mean of 20.4
Ag/m3 higher than 18.85 Ag/m3 of personal exposure
level. Benzene in city center and its peripheral zone
(90% of environmental sampling sites) presented an
Table 7
Multivariate associations between personal benzene levels (Ag/m3) and c
confidence interval, RC: reference category)
MODEL I beta (95% CI)
Transportation mean
Foot RC
Vehicle 4.28 (0.46, 8.11)
Both 2.92 (�1.81, 7.66)
Time spent outdoors (h) 0.278 (0.185, 0.370)
Benzene home levels (Ag/m3) 0.221 (0.040, 0.403)
Period
September 1st 5.19 (2.17, 8.21)
December 2nd 0.45 (�2.80, 3.69)
February 3rd 0.01 (�3.16, 3.18)
April 4th 0.10 (�4.27,�4.48)
June 5th �4.96 (�7.81,�2.12)
September 6th RC
Wind speed (m/s)
Humidity (%)
Sunshine (h)
annual average concentration of 21.6 Ag/m3 (range
16.2–29.2 Ag/m3 ) that far exceeds the limit of 5 Ag/m3, proposed by the European Union guideline.
The lower concentrations during the summer
period are probably due to the fact that car traffic
was considerably reduced. For all six measurements
haracteristics of study population (regression coefficient (beta), CI:
p MODEL II beta (95% CI) p
RC
0.029 4.30 (0.47, 8.13) 0.029
0.218 2.95 (�1.78, 7.69) 0.213
b10�3 0.276 (0.184, 0.368) b10�3
0.018 0.224 (0.043, 0.405) 0.016
0.001
0.785
0.996
0.962
0.001
�2.40 (�3.11,�1.70) b10�3
0.89 (0.52, 1.26) b10�3
3.13 (1.85, 4.41) b10�3
Table 8
Multivariate associations between home benzene levels (Ag/m3) and characteristics of homes (regression coefficient (beta), CI: confidence
interval, RC: reference category)
MODEL I beta (95% CI) p MODEL II beta (95% CI) p
Proximity to busy road RC 0.003 RC 0.003
2.84 (1.06, 4.62) 2.81 (1.01, 4.62)
Wind speed (m/s) �2.14 (�2.47,-1.81) b10�3
Period:
September 1st 5.44 (3.76, 7.13) b10�3
December 2nd 5.12 (2.15, 8.09) 0.001
February 3rd 3.08 (1.77, 4.39) b10�3
April 4th 1.64 (0.17, 3.11) 0.029
June 5th �2.01 (�2.84,�1.18) b10�3
September 6th RC
C. Chatzis et al. / Science of the Total Environment 349 (2005) 72–8078
the wind speed, humidity, duration of sunshine and
temperature were monitored. The average climate
parameters are presented in Table 4.
In Tables 5 and 6 univariate analyses of important
prognostic factors besides seasonal variation are
presented. Bus-drivers were the most heavily exposed
group followed by traffic policemen while volunteers’
houses showed similar levels irrespective of occupa-
tional group (Fig. 2).
Adjusted for seasonal variation only, proximity to
busy road remained as an important factor of benzene
home level while for personal exposure, large con-
tribution hold, in addition to home levels, the time
spent outdoors and transportation mean. Groups who
spent most of work-time outdoors found to be exposed
to higher benzene levels and this explains the strong
relationship between occupation and personal levels of
exposure. Furthermore this specific factor is respon-
sible for leading to non-significance, the relationship
between occupation and personal exposures in multi-
variate analysis. In Tables 7 and 8 two interchangeable
models are presented with seasonal or climate varia-
tion. Wind speed seems to determine largely home
levels and personal exposure Wind had similar effect
in clearing indoor and urban pollution in Athens;
lessen personal exposure and home levels about 2–2.5
Ag/m3 per 1 m/s increase in speed.
5. Discussion
The measured benzene concentrations in the centre
of the city were generally high. They are higher than
values measured at urban sites in other European cities,
similar with concentrations obtained from measure-
ments in the centre of Athens during 1994 and 1996
and, lower than other urban measurements, especially
from big cities of developing countries (Moschonas et
al., 2001; Bakeas and Siskos, 2002; Petrakis et al.,
2003; Cocheo et al., 2000; Leung and Harrison, 1998;
Riveros-Rosas et al., 1997; Bahrami, 2001).
The average pollution level in houses located
within environmental sampling areas was almost half
(45–75%) of urban pollution level came from the same
area (city center and its peripheral zone). The reason
why indoor pollution is much lower than the outdoor
one, even if it reflects its seasonal trend, might be due
to home factors related to climate (use of non
absorbent materials for wall and floor covering)
(Cocheo et al., 2000). Among home factors, only
proximity to busy road was important factor for indoor
benzene levels. Proximity to gasoline station and home
location exhibited a borderline significance but finally
were excluded in multivariate analysis. Probably home
characteristics like frequent ventilation and use of non-
absorbent materials for wall and floor covering in
Greece might also account for the lack of significant
correlations. Opposed to Athens, in poorly ventilated
buildings indoor emission source strength is consid-
ered a more significant influence on benzene concen-
trations than infiltration of outdoor air (Kim et al.,
2002). The fact that personal exposure is lower than
environmental concentrations of benzene is mainly
due to low indoor pollution. Nevertheless, correlations
between personal and home levels were significant but
generally weak, especially in summer compared to
other seasons for the reason that factors outside houses
hold a predominant role in exposure pattern.
C. Chatzis et al. / Science of the Total Environment 349 (2005) 72–80 79
Indoor air pollution and personal exposure were
closely related to measurement period, i.e. seasonal
variation. The nature of the source and meteorology or
the degree of atmospheric dispersion influences
seasonal patterns. The variation between seasons
may be explained in large by differences in climate
parameters. Wind speed seems to determine largely
home levels and personal exposure with comparable
clearing rates. On the other hand, very different rates
of clearing benzene pollution outdoors and indoors
through ventilation, especially in northern towns have
been reported (Cocheo et al., 2000). In our case the
decrease between first and last sampling periods
(Autumn) may also be due to interventions that took
place in transport processes and modernization of bus
fleet. The national inspection programme for the
control of emissions from motor vehicles was initiated
and constantly continued since 1994, consisting of
regular inspections of all private vehicles once per
year and taxis and light trucks twice per year.
Furthermore a project for the full replacement of the
old buses with new ones equipped with anti-pollution
devices has been continued during the study period.
More than 100 buses have been replaced during study
period and in addition to the replacement of buses, the
increase of dedicated bus lanes improved running
conditions. It was anticipated that these measures have
influenced to some extent atmospheric pollution
during the study period especially in city centre.
Adjusted for seasonal or climate variation, signifi-
cant prognostic factors of personal exposure were
home levels and total time spent outdoors. Groups
who spent most of work-time outdoors found to be
exposed to higher benzene levels and this explains the
strong relationship between occupation and personal
levels of exposure. Transportation means was also a
significant prognostic factor of personal benzene
levels. Our results suggest that outdoor works give
the greater contribution to benzene exposure of
Athens citizens.
Our study shows that population exposure
depends mainly on time spent outdoors, indoor
pollution and transportation mean. In northern towns
where an opposite trend in indoor pollution pre-
vailed, outdoors exposure holds a beneficial effect on
personal levels. Our exposure pattern suggests that
strategy of fixed point monitoring is insufficient to
estimate exposure in specific groups of people who
may be subject to higher exposures. Policy makers
have to take into account exposure patterns of
specific groups in establishing guidelines for ambient
benzene exposure.
References
Bahrami AR. Distribution of volatile organic compounds in ambient
air of Tehran. Arch Environ Health 2001;56:380–3.
Bakeas EB, Siskos PA. Volatile hydrocarbons in the atmosphere of
Athens, Greece. Environ Sci Pollut Res Int 2002;9:234–40.
Bono R, Scursatone E, Schiliro T, Gilli G. Ambient air levels and
occupational exposure to benzene, toluene, and xylenes in
northwestern Italy. J Toxicol Environ Health, Part A 2003;66:
519–31.
Cocheo V, Boarreto K, Sacco P. High uptake rates radial diffusive
sampler suitable for both solvent and thermal desorption. Am J
Ind Hyg Assoc 1996;198:79–96.
Cocheo V, Sacco P, Boaretto K, De Saeger E, Ballesta PP, Skov H,
et al. Urban benzene and personal exposure. Nature 2000;
404:141–2.
Crebelli R, Tomei F, Zijno A, Ghittori S, Imbriani M, Gamberale D,
et al. Exposure to benzene in urban workers: environmental and
biological monitoring of traffic police in Rome. Occup Environ
Med 2001;58:165–71.
Crump KS. Risk of benzene-induced leukaemia: a sensitivity
analysis of the pliofilm cohort with additional follow-up and
new exposure estimates. J Toxicol Environ Health 1994;42:
219–42.
European Environment Agency. Air Quality in Larger Cities in the
European Union. Copenhagen7 EEA; 2001 Topic report 3.
Gee IL, Perry R. Vehicle emissions and effect on air quality: indoors
and outdoors. Indoor Environ 1994;3:224–36.
Karacic V, Skeder L, Bosner-Cucancic B, Bogadi-Sare A. Possible
genotoxicity in low level benzene exposure. Am J Ind Med
1995;27:379–88.
Kim YM, Harrad S, Harrison RM. Levels and sources of personal
inhalation exposure to volatile organic compounds. Environ Sci
Technol 2002;36:5405–10.
Kouniali A, Cicolella A, Gonzalez-Flesca N, Dujardin R,
Gehanno JF, Bois FY. Environmental benzene exposure
assessment for parent–child pairs in Rouen, France. Sci Total
Environ 2003;308:73–82.
Leung P, Harrison RM. Evaluation of personal exposure to
monoaromatic hydrocarbons. Occup Environ Med 1998;55:
249–57.
Maitre A, Soulat JM, Masclet P, Stoklov M, Marques M, de
Gaudemaris R. Exposure to carcinogenic air pollutants among
policemen working close to traffic in an urban area. Scand J
Work, Environ 2002;28:402–10.
Moschonas N, Glavas S, Kouimtzis T. C3 to C9 hydrocarbon
measurements in the two largest cities of Greece, Athens and
Thessaloniki Calculation of hydrocarbon emissions by species
Derivation of hydroxyl radical concentrations. Sci Total Environ
2001;271:117–33.
C. Chatzis et al. / Science of the Total Environment 349 (2005) 72–8080
Official Journal of the European Communities L 313, 13/12/2000.
Directive 2000/69/EC Relating to Limit Values for Benzene and
Carbon Monoxide in Ambient Air; 2000. OJ L 313.
Petrakis M, Psiloglou B, Kassomenos PA, Cartalis C. Summertime
measurements of benzene and toluene in Athens using a
differential optical absorption spectroscopy system. J Air Waste
Manage Assoc 2003;53:1052–64.
Riveros-Rosas H, Pfeifer GD, Lyman DR, Pedroza JL, Julian-
Sanchez A, Canales O, et al. Personal exposure to elements in
Mexico City air. Sci Total Environ 1997;198:79–96.
Wallace LA. Major sources of benzene exposure. Environ Health
Perspect 1989;82:165–9.
Wallace LA, Pellizzari ED. Personal air exposures and breath
concentrations of benzene and other volatile hydrocarbons for
smokers and non smokers. Toxicol Lett 1987;35:113–6.
Wallace LA, Pellizzari E, Leader B, Zelon H, Sheldon L. Emissions
of volatile organic compounds from building materials and
consumer products. Atmos Environ 1987;21:385–93.
World Health Organisation Regional Office for Europe. Air quality
guidelines for Europe European series no91. Copenhagen7
WHO regional publications; 2000.
World Health Organization. Environmental Health Criteria 150
Benzene. Geneva7 WHO; 1993.