Health risk assessment on human exposed to environmental polycyclic aromatic hydrocarbons pollution sources

Health risk assessment on human exposed to environmental

polycyclic aromatic hydrocarbons pollution sources

Szu-Chich Chen, Chung-Min Liao*

Ecotoxicological Modeling Center, Department of Bioenvironmental Systems Engineering, National Taiwan University, Taipei, Taiwan 10617 ROC Received 16 April 2005; received in revised form 15 August 2005; accepted 29 August 2005

Available online 22 November 2005

Abstract

To assess how the human exposure to environmental carcinogenic polycyclic aromatic hydrocarbons (PAHs) pollution sources generated from industrial, traffic and rural settings, we present a probabilistic risk model, appraised with reported empirical data. A probabilistic risk assessment framework is integrated with the potency equivalence factors (PEFs), age group-specific occupancy probability and the incremental lifetime cancer risk (ILCR) approaches to quantitatively estimate the exposure risk for three age groups of adults, children, and infants. The benzo[a]pyrene equivalents based PAH concentrations in rural, traffic, and industrial areas associated with age group-specific occupancy probability at different environmental settings are used to calculate daily exposure level through inhalation and dermal contact pathways. Risk analysis indicates that the inhalation-ILCR and dermal contact-ILCR values for adults follow a lognormal distribution with geometric mean 1.04  10 4and 3.85  10 5and geometric standard deviation 2.10 and 2.75, respectively, indicating high potential cancer risk; whereas for the infants the risk values are less than 10 6, indicating no significant cancer risk. Sensitivity analysis indicates that the input variables of cancer slope factor and daily inhalation exposure level have the greater impact than that of body weight on the inhalation-ILCR; whereas for the dermal-ILCR, particle-bound PAH-to-skin adherence factor and daily dermal exposure level have the significant influence than that of body weight.

D 2005 Elsevier B.V. All rights reserved.

Keywords: Polycyclic aromatic hydrocarbons; Benzo[a]pyrene; Potency equivalency factor; Health risk assessment; Environmental pollution

1. Introduction

Human cancer causes of skin, lungs, and bladder have always been associated with polycyclic aromatic hydrocarbons (PAHs) (Boffetta et al., 1997). The rela-tionship between cancer and the environment is largely conditioned by investigations involving PAH exposures (Armstrong et al., 2004). Several individual PAHs such as benzo[a]pyrene (B[a]P), chrysene, indeno[1,2,3-c,d]pyrene, and benzo[b]fluoranthene have produced

carcinogenic, mutagenic, and genotoxic effects in ani-mal experiments (Thyssen et al., 1981; Deutsch-Wenzel et al., 1983).Somers et al. (2002, 2004)have also found out that air pollution enriched with PAHs has been shown to include heritable (paternal germ-line) muta-tions in mice.

More recently, PAHs have been associated with elevated levels of DNA adducts (PAH-DNA adducts) and P53 mutations in persons who smoke or are ex-posed to PAH in the workplace and ambient air ( Alex-androv et al., 2002; Gaspari et al., 2003).Perera et al. (2002)has also indicated that airborne PAHs have been implicated in human reproductive effects, PAH-DNA adducts in newborns as well as preterm birth and

0048-9697/$ - see front matterD 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.scitotenv.2005.08.047

* Corresponding author. Tel.: +886 2 2363 4512; fax: +886 2 2362 6433.

E-mail address: cmliao@ntu.edu.tw (C.-M. Liao).

intrauterine growth restriction. Although exposure to environmental PAHs has been based on the assumption that inhalation was the primary route (Venkataraman and Raymond, 1998), dermal contact is increasingly taken into account (ATSDR, 1990; Tsai et al., 2001). The risk associated with human exposure to atmo-spheric PAHs is highest in cities, considering the den-sity of population, increasing vehicular traffic, and scarce dispersion of the atmospheric pollutants. In Tai-wan region, several significant contributor to PAHs sources had been sampled such as stationary industrial combustion: steel and iron industries (Yang et al., 2002) with a mean total-PAHs concentrations measured to be 1020 Ag m 3, traffic vehicles exhaust: motorcycle (Yang et al., 2005) and highway toll station (Tsai et al., 2004) with a mean total-PAHs concentrations ranged from 8280–12300 ng m 3. Fang et al. (2004a,b,c) indicated that mean total PAHs levels at industrial, urban, and rural areas in central Taiwan region ranged from 1232–1650, 700–1740, and 610– 831 ng m 3, respectively.

However, occurrence of PAHs in ambient air causes particular concern due to the continuous nature of exposures and the size of populations at risk, especially in urban, suburban and industrial areas. Therefore, in light of the mutagenicity, carcinogenicity and ubiquity of some PAHs in the atmosphere, the setting of air quality standards and guidelines to limit human expo-sure should be of primary concern for public health policy. However, it is complicated to derive scientifi-cally the standards or guidelines based on relatively limited empirical data, owing to the difficulties in inter-preting heterogeneous experimental and epidemiologi-cal findings (Tsai et al., 2001).

Currently, no information on the human cancer risk assessment for specific-age groups related to environ-mental PAHs pollution sources is available in Taiwan region. Furthermore, identification of the most danger-ous environmental PAHs and their mode of action in producing specific health effects remain uncertain and difficult to quantify the exposure risk precisely. To reduce the potential risk of harmful negative health consequences and problems, we suggest that risk anal-yses be taken seriously to characterize the impact of

PAHs in the human environment. Fang et al.

(2004a,b,c) have been reported the relevant measure-ments of PAHs in industrial, urban, and rural areas and these measurements were sufficiently to motivate the most serious kind of concern for personal exposure to environmental carcinogenic PAHs. We were stimulated to develop a probabilistic risk assessment framework to evaluate the carcinogenic risk from personal exposure

to environmental PAHs in industrial, urban, and rural areas.

The objective of this study is to develop an integrat-ed environmental exposure risk model that incorporates inhalation and dermal contact from atmospheric PAHs existing in industrial, urban, and rural areas in Taiwan region. The risk model is intended to apply to general Taiwanese population. A detailed sensitivity and uncer-tainty analyses are conducted, and the results are used to identify critical input variables requiring further study. We employed a quantitative risk assessment method used for PAHs based on B[a]P equivalent concentration from animal studies. Several studies have also documented the impact that the outdoor air (Ando et al., 1996; Li and Ro, 2000; Naumova et al., 2002), and more specifically traffic (Kingham et al., 2000; Fisher et al., 2000), has on the quality of indoor air. We were extended our risk analysis to include PAH levels in indoor air in light of these findings.

2. Materials and methods 2.1. Data reanalysis

There is relatively little empirical data in environ-mental PAHs sources because the experiments are com-plicated. Accordingly, we must rely on data reanalysis technique together with whatever empirical data is available. Here, we adopted three published empirical data in total PAHs concentrations and particle-bound PAHs concentrations sampled from industrial, traffic, and rural areas in Taiwan region.

Fang et al. (2004a) selected two sampling sites situated at Taichung Industry Park and suburban area (Tunghai University) in central Taiwan to measure 21 individual PAH concentrations with 48 h consecutive sampling (n = 23) from August 2002 to July 2003 in that the meteorological conditions such as temperature, relative humidity, wind speed, and atmospheric pres-sure were also monitored. Taichung Industry Park represents the typical large-scale multiple industry with more than 800 factories. The sampling site is located on the roof of a pharmaceutical factory (18 m in height), situated at the center of Taichung Industry Park. Suburban area (Tunghai University) is a back-ground site to compare with the neighbor Industrial Park. The sampling heights were set in the range of 1–1.5 m above ground level to simulate the breathing zone. The instruments were calibrated using at least five standard concentrations covering the concentration of interest for ambient air work. The calibration curve had significant correlated with the linear regression profile

(r2= 0.998). The average recovery efficiency of PAH was 86%. The total 21 PAHs concentration at Taichung Industrial Park and Tunghai university were measured to be 1560.1 F 1429.4 (mean F sd) and 734.1 F 541.0 ng m 3, respectively.Fang et al. (2004b)measured the PAHs concentration at three sampling sites (n = 12) situated at Taichung Industry Park, rural area (Tunghai University), and urban site from August 2002 to De-cember 2002 in that PAH concentration, size distribu-tion, and estimated PAHs dry deposition fluxes and health risk were also determined. The mean concentra-tions of total PAHs in ambient air at industrial, urban, and rural in central Taiwan, Taichung were measured to be 1652, 1185, and 834.1 ng m 3, respectively.

We have also adopted the data from Tsai et al. (2004)for PAHs concentrations measured at highway toll station to account for the traffic pollution sources in that they collected 48, 35, and 33 personal PAH sam-ples from booth attendants of the dayshift (08:00 AM– 16:00 PM), nightshift (16:00 PM–00:00 AM), and late-night (00:00 AM–08:00 AM), respectively. Their study showed the average recovery efficiency of PAHs was 85%. The limit of detections of the 21 PAH compounds were determined based on the analysis of serial dilu-tions of PAH standards.

A useful for estimation of the health risk posed by multi-component PAH exposure is based on the use of the individual compound’s potency equivalency factor (PEF) relative to B[a]P based on a PEF scheme devel-oped byCollins et al. (1998). In a first step, a B[a]P equivalent concentration (B[a]Peq) is calculated by the

multiplication of the individual PAH concentration by its PEF. The carcinogenic potency of all considered PAHs can then be estimated as the sum of each indi-vidual B[a]Peq. Table 1 lists the PAHs and the PEFs

used in the calculation associated with the cancer evi-dence of individual PAH compound.Nisbet and LaGoy (1992)and Collins et al. (1998) have summarized the assumptions underlying the use of the PEF approach to the calculation of exposure to PAHs.

2.2. Occupancy probability and daily exposure level Here we define an occupancy probability (OP) to quantify the likelihood of the time a person spends in a specific setting (Fig. 1). The OP is a practical assump-tion according to daily routine activities or daily work-ing logs for ordinary adults, children, and infants. In the future study, however, the detailed case investigation is essential to provide the reliable data required.

We assumed that a cohort lived at rural and their workplace or children’s school nearby an industrial

park and they were vulnerably exposed to PAHs pol-lution from traffic when they went to workplace or school. We divided a 24-h period into five time inter-vals as j = 1, 2, 3, 4, and 5 representing the time periods from 00:00–07:00, 07:00–09:00, 09:00– 17:00, 17:00–19:00, and 19:00–24:00, respectively. We further denote j = 1 and 5 are the time intervals spent at rural (home), j = 2 and 4 are the time intervals spent in commuting, and j = 3 is the time interval and spent in working or at school (Fig. 1). We assumed the OP within each period followed a uniform distri-bution and the summation of the daily basis OP at 24-h for rural, traffic, and industrial settings was equal to 1.

The indoor PAHs levels in rural (home) can be estimated by knowing the PAHs concentration in rural associated with the average indoor / outdoor (I / O) ratio. Here we adopted an average PAHs I / O ratio of 1.2 measured byLi and Ro (2000)in the rural areas of the northern Taiwan region. Fig. 1 demonstrates that the infant OP has the highest value at rural setting, whereas children have the higher than those of others at traffic setting. At industry setting, adult has higher OP than those of others.

The B[a]Peq-based PAH concentration in rural,

traf-fic, and industrial areas associated with age group-specific OP values at different settings are used to calculate daily exposure level through inhalation and

Table 1

Potency equivalency factors (PEF) and molecular weight for individ-ual PAHs relative to B[a]P (Nisbet and LaGoy, 1992) used in this study

Compound Molecular weight (g mol 1) PEF

Naphthalene (Nap) 128 0.001 Acenaphthylene (AcPy) 152 0.001 Acenaphthene (AcP) 154 0.001 Fluorene (Flu) 165 0.001 Phenanthrene (PA) 178 0.001 Anthracene (Ant) 178 0.01 Fluoranthene (FL) 202 0.001 Pyrene (Pyr) 202 0.001 Cyclopenta[c,d]pyrene (CYC) 228 0.1a Benzo[a]anthracene (BaA) 228 0.1 Chrysene (CHR) 228 0.01 Benzo[b]fluoranthene (BbF) 252 0.1 Benzo[k]fluoranthene (BkF) 252 0.1 Perylene (PER) 252 0.001a Benz[e]pyrene (BeP) 252 0.01a Benzo[a]pyrene (BaP) 252 1 Indeno[1,2,3-c,d]pyrene (IND) 276 0.1 Dibenz[a,h]anthracene (DBA) 278 1a Benzo[b]chrycene (BbC) 278 NA

Benzo[ g,h,i]perylene (BghiP) 276 0.01

Coronene (COR) 300 0.001a

a

dermal contact. Daily exposure level for a specific age group can be calculated as,

EI ¼ X3 i¼1 X5 j¼1 OPijd að iCiÞd 1 24OTj   dIRj " # ; ð1Þ

where EIis the daily exposure level for inhalation for

age group in three environmental settings (ng day 1), ai= 1.2 is the average PAHs I / O ratio for rural setting

(i = 1) otherwise ai= 1, OPij is the OP value in the jth

time interval ( j = 1, 2, 3, 4, and 5) in environmental setting i (i = 1, 2, and 3 represent rural, traffic, and industrial setting, respectively), Ciis the PAHs

concen-tration in environmental setting i (ng m 3), OTj is the

occupancy time for the jthtime interval (hr), IRj is the

inhalation rate in the jthtime interval (m3day 1) when j = 1 and 5 are the resting activity and j = 2, 3, and 4 are the active activity.

We can also adopt Eq. (1) to calculate the daily exposure level for dermal contact as

ED¼ X3 i¼1 X5 j¼1 OPijd að iCdÞd 1 24OTj   " # ; ð2Þ

where EDis the daily exposure for dermal contact (Ag

g 1) and Cd is the particle-bound PAH concentration

(Ag g 1). Here, adult are defined as individuals from ages 20 to 70 year, children are defined as individuals from ages 5 to 19 year, whereas infant are defined as individuals from ages 0 to 1 year.

2.3. Incremental lifetime cancer risk (ILCR) model A probabilistic risk model integrates the potency equivalence factors (PEFs), human OP and the ILCR approach to quantitatively estimate the exposure risk for three age groups of adult, children, and infant. The ILCR model for human inhalation is defined as,

R¼ EId CSFi BW 70  1=3! dEFdED BWdAT  cf ; ð3Þ

where R is the incremental individual lifetime cancer risk, EI is the daily exposure level for inhalation (ng

day 1), CSFiis the inhalation carcinogenic slope factor

(mg kg 1day 1) 1, EF is the exposure frequency (day year 1), ED is the exposure duration (year), AT is the 0 0.2 0.4 0.6 0.8 1 Probability B. Traffic 0 0.2 0.4 0.6 0.8 1 0 6 12 18 24 Time (hr) C. Industry 6 12 18 24 6 12 18 24 0 6 12 18 24 6 12 18 24 6 12 18 24 0 6 12 18 24 6 12 18 24 6 12 18 24 A. Rural 0 0.2 0.4 0.6 0.8

1 Adult Children Infant

averaging time for carcinogenic (day), BW is the body weight (kg), and cf is the conversion factor (10 6). We treated EI, CSFi, EF and BW in Eq. (3) probabilistically.

The ILCR model for dermal contact pathway is defined as, R¼ EDd CSFd BW 70  1=3! ABdSAdEVdAFddEFdED BWdAT  cf ; ð4Þ

where EDis the daily exposure level for dermal contact

(Ag g 1) (particle-bound PAHs concentration), CSFdis

the cancer slope factor for dermal contact (mg kg 1 day 1) 1, AB is the dermal adsorption fraction (dimen-sionless), SA is the dermal surface area exposed (cm2), EV is the event frequency (event day 1), AFd is the

particle-to-skin adherence factor (mg cm 2 event 1). We treated ED, AB, SA, AFd and BW in Eq. (4)

probabilistically. The total incremental lifetime cancer risk is the sum of risks associated with each exposure route.

The ILCRs for adults, children, and infants could be calculated respectively as follows. For adults: ILCR¼P70_i¼20Ri=51, for children: ILCR¼P19i¼5Ri=15, and for infants: ILCR¼ P1_i¼0Ri.

The potency of B[a]P is based on the 95 % upper confidence limit of the linearized multistage model (OEHHA, 1993) applied to the incidence of respiratory tumors after inhalation exposure in hamsters (Thyssen et al., 1981). Collins et al. (1991) have estimated the cancer slope factor for B[a]P inhalation exposure based on three different hamster inhalation rates of 0.037, 0.063, and 0.158 m3 day 1 and resulted in the CSF values of 6.1, 3.8, and 1.3 (mg kg 1day 1) 1, respec-tively. We averaged those three CSF values and log-transformed appropriately to a lognormal distribution with a geometric mean 3.14 (mg kg 1day 1) 1and a geometric standard deviation 1.80. For exposure to B[a]P through dermal contact pathway, the potency of 37.47 (mg kg 1 day 1) 1 based on incidence of skin tumors in mice (Schmahl et al., 1977) could be used (Hussain et al., 1998). The cancer slope factors are normalized to account for extrapolation to a different body weight from standard of 70 kg.

Table 2 shows the selected types of probability distribution for random variables including the inhala-tion risk and dermal contact risk parameters. The SA describes the amount of skin exposed to the contami-nated media. The amount of skin exposed depends upon the exposure scenarios. The adult in commer-cial/industrial workplace was assumed to wear a

Table 2

Risk parameters considered as random variables (lognormal distribution with geometric mean and geometric stand deviation: LN (gm, gsd)) for different age groups

Definition units Infant Children Adult

Population parameter

Age year 0–1 5–19 20–70

Body weight (BW)a kg LN(6.79, 1.27) LN(36.24, 1.05) LN(59.78, 1.07)

Inhalation risk parameter

Daily exposure level (EI) ng day 1 LN(257, 1.96) LN(1724, 1.83) LN(1687, 1.70) Inhalation rate(IRair)b Resting m3day 1 LN(2.75, 1.38) LN(7.71, 1.27) LN(9.01, 1.26)

Active LN(5.96, 1.49) LN(24.87, 1.38) LN(32.74, 1.14)

Dermal risk parameter

Daily exposure level (ED) Ag g 1 LN(4.8  10 4, 1.91) LN(8.8  10 4, 1.85) LN(7.1 10 4, 1.77) Dermal adsorption fraction (AB)c _{– LN(0.13, 1.26) LN(0.13, 1.26) LN(0.13, 1.26)}

Dermal surface exposure (SA)d _cm2 _{LN(719, 1.19) LN(2196, 1.08) LN(3067, 1.06)}

Events frequency (EV)c event year 1 1 1 1

Dermal adherence rate (AFd) c

mg cm 2event 1 LN(0.04, 3.414) LN(0.04, 3.404) LN(0.02, 2.668) Risk model parameter

Exposure frequency (EF)e day year 1 LN(252, 1.01) LN(252, 1.01) LN(252, 1.01)

Averaging time (AT)c day 25550 25550 25550

a

Adapted from Department of Health, ROC (http://www.doh.gov.tw/cht/index.aspx#). b

Adapted from ICRP 66 (ICRP, 1994). c

Adapted fromUSEPA (2001). d

Adapted formUSEPA (1992). e

short-sleeved shirt, long pants, and shoes; therefore, the exposed skin surface is limited to the head, hands, and forearms. We can calculate the age group SA according to the recommended value from USEPA (1992). Recommended dermal absorption fraction from soil for B[a]P and other PAHs followed a lognormal distri-bution with a geometric mean 0.13 and geometric standard deviation 1.26 (USEPA, 2001). An averaging time of 365 day year 1for 70 year (i.e., AT = 25,550 day) was used to characterize lifetime exposure for cancer risk calculation.

2.4. Uncertainty analysis

Uncertainty arises from estimation of both exposure and effects. Incorporating variability and uncertainty into risk calculation results a more realistic view about the real distribution of risk. In order to quantify this uncertainty and its impact on the estimation of expected risk, a Monte Carlo simulation was implemen-ted. We performed independent runs at 1000, 4000, 5000, and 10,000 iterations with each parameter sam-pled independently from the appropriate distribution at the start to test the convergence and the stability of the numerical output. The result shows that 5000 iterations are sufficient to ensure the stability of results. A con-fidence interval for expected risk was determined on the basis of the 2.5th and 97.5th quartiles of the simulation outcomes.

Sensitivity analysis identified the most significant parameters that were included in the uncertainty and variability analysis. The sensitivity of each variable relative to one another was assessed by calculating rank correlation coefficients between each input and output during simulations and then estimating each input contribution to the output variance by squaring the output variance and normalizing to 100%. The Monte Carlo simulation and sensitivity analysis were implemented using Crystal Ball software (Version 2000.2, Decisioneering, Inc., Denver, CO, USA). 3. Results

3.1. Total PAHs concentration for environmental PAHs pollution sources

The median total PAHs and the median B[a]Peq

concentrations of inhalation exposure at rural, indus-trial, and traffic settings are estimated to be 645 and 18.7; 1200 and 25.67; and 10611.28 and 158.5 ng m 3, respectively (Fig. 2A, B), whereas median parti-cle-bound PAHs and median B[a]Peq concentrations

for dermal contact exposure are estimated to be 68.48 and 6.20; 107.8 and 5.48; and 1103.97 and 127.84 ng m3, respectively, at rural, industrial and traffic settings (Fig. 2C, D). Data reanalysis also demonstrates that B[a]Peq levels in the traffic setting

were much higher than those measured in rural and industrial settings where B[a]Peq levels were slightly

closed (Fig. 2B, D).

3.2. Daily exposure B[a]Peq concentration

The daily exposure B[a]Peq concentration can be

estimated from Eqs. (1) and (2) associated with age group-specific OP values at different environmental settings (Fig. 1) through the potential exposures of the normal daily activities or working logs. The pre-dicted probability density functions of daily exposure B[a]Peqconcentration based on the inhalation and

der-mal contact are shown in Fig. 3A, C in that box and whisker plots are used to represent the uncertainty in that boxes show 25th and 75th percentiles and whiskers are 2.5th and 97.5th percentiles (Fig. 3B, D). We prac-tically used the average outdoor particle concentration of 60.29 Ag m 3measured from Atmospheric Monitor Station of Taiwan Environmental Potency Agency during 2002 to 2003 to carefully convert the PAHs concentration in unit of ng m 3 (Fig. 2D) to Ag g 1 (Fig. 3C, D).

The median B[a]Peq concentration of inhalation

exposure are estimated to be 1628, 1590, and 252 ng day 1, respectively, whereas the median B[a]Peq

concentration of dermal contact exposure are calcu-lated to be 7.0  10 4, 8.7  10 4, and 4.8  10 4Ag g1, respectively, for adults, children, and infants (Fig. 3B, D). In inhalation exposure, the B[a]Peq

levels for adults and children are nearly closed and higher than that of infants, whereas in dermal expo-sure, the B[a]Peq level of children is slightly higher

than that of adults and both are higher than that of infants (Fig. 3B, D).

3.3. Inhalation, dermal, and total incremental lifetime cancer risk (ILCR)

According to the daily exposure level of dermal contact and inhalation exposure, we can estimate the inhalation and dermal ILCRs (Fig. 4). Fig. 4A1–A3 indicate that the inhalation ILCRs for adults, children, and infants age group follow lognormal (LN) distribu-tions of LN(1.04  10 4, 2.10), LN(6.47  10 6, 2.21), and LN(5.50  10 8, 2.36), respectively.Fig. 4B1–B3 indicate that the dermal contact ILCRs for adults,

chil-dren, and infants age group have LN distributions of LN(3.85  10 5, 2.75), LN(5.16  10 6, 3.33), and LN(9.90  10 8, 4.83), respectively.

For adults and children age groups, inhalation ILCR values are higher than that of dermal contact, whereas for infants age group, dermal contact ILCR value is higher than that of inhalation. Percentile predictions of total ILCR personal exposure of three age groups could be determined from cumulative density functions (cdfs) corresponding to pdfs shown inFig. 4C1, D1, and E1. The total ILCR that combines dermal contact and in-halation exposure pathways are estimated to be

LN(1.66  10 4, 1.89), LN(1.44  10 5, 2.13), and LN(2.10  10 7, 2.93), for adults, children, and infants, respectively.

Under most regulatory program, an ILCR between 10 6 and 10 4 indicates potential risk. Our results indicate that the 97.5% probability inhalation and dermal contact ILCRs for adults and children exposed to PAH have orders of magnitude around 10 5 and 10 4, indicating high potential carcinogenic risk (Fig. 4A1, A2, B1, B2). All 97.5% probabilities of B[a]Peq-based total ILCRs (TILCRs) are large than

10 6, indicating unacceptable probability distributions Inhalation exposure 0 40 80 120 160 200 240 280

B

Rural Industry Traffic Rural Industry Traffic

Rural Industry Traffic Rural Industry Traffic

0 2000 4000 6000 8000 10000 12000 14000

A

Total PAHs concentration (ng m

-3) Median 25% - 75% 2.5% - 97.5% Dermal exposure 0 300 600 900 1200 1500 0 40 80 120 160 200

C

Particle-bound PAHs concentration (ng m

-3) 0 40 80 120 160 200 240 0 7 14 21 28 35

D

Fig. 2. Box and whisker plots of exposure levels of (A) total PAHs and (B) B[a]Peqfor inhalation, (C) particle-bound PAHs and (D) B[a]Peqfor dermal at rural, industry, and traffic settings, respectively.

for three age groups (Fig. 4C2, D2, E2). For adult personal exposure to carcinogenic PAH, 97.5% prob-ability TILCR (5.89  10 4) is much fall outside the range of 10 6–10 4, indicating high potential health risk; whereas for children and infants age groups, 97.5% probability TILCRs range from 10 6–10 5 (children: 7.05  10 5 and infant: 2.41 10 6), re-spectively (Fig. 4C2, D2, E2). The infants and chil-dren have the lower values of AFd, ED and BW than

those of adults (Table 2) that may provide an expla-nation for the infants and children who are at low risks.

3.4. Sensitivity analysis

We conducted a quantitative sensitivity analysis to evaluate the variability and uncertainty of parameters in the exposure pathway that contributed most significant-ly to the risk estimates. The results of the sensitivity analyses on inhalation and dermal contact risk models are shown in the form of tornado plots illustrating the Spearman rank order correlation coefficients (Fig. 5). For inhalation exposure, the inhalation cancer slope factors (CSFi) and daily inhalation exposure level (EI)

are the most influential variables for all three age 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.0000 0.0006 0.0012 0.0017 0.0023 0.0029 Dermal exposure B[a]Peq concentration (µg g-1)

C

Adult Children Infant

B[ a ]P eq concentration ( µ g g -1) 0.000 0.001 0.002 0.003

D

A

B[a]Peq concentration (ng day-1)

Probability Probability 0 1000 2000 3000 4000 5000 6000 Median 25% - 75% 2.5% - 97.5%

Adult Children Infant

B[ a ]P eq concentration (ng day -1)

B

Fig. 3. Predicted probability density functions of B[a]Peqconcentration (A, C) and the box and whisker plots (B, D) describing the three age groups for inhalation and dermal exposure to environmental PAHs sources, respectively.

Fig. 4. Box and whisker plots for (A1) adult, (A2) children, and (A3) infant of inhalation incremental lifetime cancer risk and for (B1) adult, (B2) children, and (B3) infant of dermal contact incremental lifetime cancer risk. Predicted probability density functions of total incremental lifetime cancer risk for (C1) adult, (D1) children, and (E1) infant and the box and whisker plots (C2, D2, and E2) are also shown.

groups, which average contributions to output variance are 67%, 69% and 69%, respectively, for adults, chil-dren, and infants. For dermal exposure to particle-bound PAH, the particle-to-skin adherence factor (AFd) and daily dermal exposure level (ED) is the

most influential variable for all three age groups, which the contributions to output variance range from 59–76% and 46–56%, respectively (Fig. 5).

4. Discussion

The study of human exposure to environmental PAH pollution sources identified the age-specific potential cancer risks. Recognition of the importance of inhala-tion cancer slope factor and particle-bound PAH-to-skin adherence factor in PAH exposure led to the efforts to estimate the significance of inhalation and dermal ILCRs. Sensitivity analysis indicates that to increase the accuracy of the results efforts should focus on a better definition of probability distributions for inhala-tion cancer slope factor and particle-bound PAH-to-skin

adherence factor. Given the scarcity of data, most of the probability distributions were based on US EPA data, and this may be a limit to the validity of the case presented.

The CSF values (upper 95% CI estimate of initial dose–response curve) for continuous inhalation expo-sure to B[a]P obtained from several data sets (respira-tory tract tumors in hamsters and gastric tumors in mice) byCollins et al. (1991)range between 1.3–16.9 (mg kg 1day 1) 1of B[a]P. The differences are due in part to the different treatment procedures and to assumptions adopted on animal physiological para-meters (e.g., surface area scaling factor). Ingestion ex-posure data were used for inhalation exposure assessment by considering the intake dose with route-to-route extrapolation.

It must also be noted that uncertainties are inherent in quantitative risk assessment because of assumptions required to extrapolate from one species to another, from high to low situations and because of the statistical modeling techniques required to fit data points.

Extrap-A

B

Fig. 5. Sensitivity analysis for (A) inhalation incremental lifetime cancer risk model and (B) dermal contact incremental lifetime cancer risk model for three age groups.

olation to low doses and the transferring of results from test animals to humans is not evident at all. Therefore, rodent-derived risk factors may have different orders of magnitude comparing with empirical data derived risk estimates.

Tsai et al. (2001) conducted a health risk assess-ment for palletizing and packaging workers exposed to PAHs in a carbon black manufacturing industry, indi-cating B[a]Peq-based lifetime lung cancer risks

esti-mated for both exposure groups were 4.35  10 2and 5.42  10 2, respectively. For dermal exposure, B[a]Peq-based lifetime skin cancer risks for both

ex-posure groups were estimated to be 1.13  10 3 and 1.56  10 3, respectively. Vyskocil et al. (2004) assessed the lung cancer in six localities with alumi-num smelting activities exposed to PAHs, indicating the upper bound B[a]Peq-based lifetime cancer risk

estimates ranged from 0.94 – 4.7  10 5. Our risk analysis indicates that average upper bound B[a]Peq

-based inhalation ILCR for three age groups was 1.59  10 4 (3.10  10 7–4.46  10 4), whereas for dermal ILCR was 1.19  10 4(2.19  10 6–2.99  10 4).

The B[a]Peq-based approach is necessarily limited

to a few PAHs that have monitored in ambient air and dose not account for the toxicity of all PAHs to which the general population is exposed. Moreover, the ad-ditivity assumption of this approach is quite uncertain and may lead to inaccurate assessments. Despite this limitation, this approach has the merit of taking into account the actual PAH profiles encountered in envi-ronmental settings reported in this study.

The carcinogenic risk assessment of PAH remains difficult, particularly due to the very high number of these compounds (in the hundreds) present in mixtures to which the general population may be exposed, as well as due to the possible contemporary presence of other risk factors and to possible synergistic and/or antago-nistic effects. Due to the limited number of dose-re-sponse data on carcinogenicity, and depending on the exposure route (intratracheal administration, intrapul-monary injection, and so on), different PEFs can be obtained (Goldstein, 2001). For example, the PEF value used in this study for DB[a,h]A is 1.0, as given byMalcom and Dobson (1994). This value may under-estimate the relevance of this compound, because other authors claim a PEF of 5.0 (Nisbet and LaGoy, 1992). It has also to be remembered that the B[a]Peq

concentra-tions used for this calculation represent an external exposure estimation of carcinogenic compounds and not the effective active concentration at the lung level. There are a number of areas in which further research could reduce the uncertainties and limit the variabilities

in this study. Among these are three areas that offer an opportunity for the most useful research. First, there is a need to conduct a more extensive characterization of the distribution of exposures within given population groups. This would require the collection of more de-tailed information on the characterization of occupation probabilities, PAH uptake in the lung and skin, and daily working logs. It would be useful to characterize better the distribution of exposures by age of individuals exposed. Second, there is a need for global sensitivity analysis using the Monte Carlo simulation model with the more detailed data sets as inputs. The ranges and distributions of parameters can then be combined by use of the Monte Carlo simulation model to produce a response surface. Relationships between the input ranges and model out-put should then be assessed with stepwise regression in order to identify the relationship between output vari-ability and input uncertainties and variabilities. Finally, on the basis of the results of the sensitivity analysis, research should be directed to those parameters that, if better characterized, could most effectively reduce var-iability in the results.

References

Alexandrov K, Cascorbi I, Rojas M, Bouvier G, Kriek E, Bartsch H. CYP1A1 and GSTM1 genotypes affect benzo[a]pyrene DNA adducts in smokers’ lung: comparison with aromatic/hydrophobic adduct formation. Carcinogenesis 2002;23(12):1969 – 77. Ando M, Katagiri K, Tamura K, Yamamoto S, Matsumoto M, Li YF,

et al. Indoor and outdoor air pollution in Tokyo and Beijing supercities. Atmos Environ 1996;30:695 – 702.

Armstrong B, Hutchinson E, Unwin J, Fletcher T. Lung cancer risk after exposure to polycyclic aromatic hydrocarbons: a review and meta-analysis. Environ Health Perspect 2004;112:970 – 8. ATSDR. Toxicological profile for benzo[a]pyrene. Agency for toxic

substance and disease registry. Atlanta, GA7 United States Public Health Service; 1990.

Boffetta P, Jourenkova N, Gustavsson P. Cancer risk from occupa-tional and environmental exposure to polycyclic aromatic hydro-carbons. Cancer Causes Control 1997;8:444 – 72.

Collins JF, Brown JP, Dawson SV, Marty MA. Risk assessment for benzo[a]pyrene. Regul Toxicol Pharm 1991;13:170 – 84. Collins JF, Brown JP, Alexeeff GV, Salmon AG. Potency equivalency

factors for some polycyclic aromatic hydrocarbons and polycyclic aromatic hydrocarbons derivatives. Regul Toxicol Pharm 1998; 28:45 – 54.

Deutsch-Wenzel RP, Brune H, Grimmer G, Dettbarn G, Misfeld J. Experimental studies in rat lungs on the carcinogenicity and dose– response relationships of eight frequently occurring environmental polycyclic aromatic hydrocarbons. J Natl Cancer Inst 1983;71: 539 – 44.

Fang GC, Wu YS, Chen MH, Ho TT, Huang SH, Rau JY. Polycyclic aromatic hydrocarbons study in Taichung, Taiwan, during 2002– 2003. Atmos Environ 2004a;38:3385 – 91.

Fang GC, Chang KF, Lu C, Bai H. Estimation of PAHs dry deposition and BaP toxic equivalency factors (TEFs) study at Urban,

Indus-try Park and rural sampling sites in central Taiwan, Taichung. Chemosphere 2004b;55:787 – 96.

Fang GC, Wu YS, Fu PC, Yang IL, Chen MH. Polycyclic aromatic hydrocarbons in the ambient air of suburban and industrial regions of central Taiwan. Chemosphere 2004c;54:443 – 52.

Fisher PH, Hoek G, van Reeuwijk H, Briggs DJ, Lebret E, van Wijnen JH, et al. Traffic-related differences in outdoor and indoor concentrations of particles and volatile organic compounds in Amsterdam. Atmos Environ 2000;34:3713 – 22.

Gaspari L, Chang SS, Santella RM, Garte S, Pedotti P, Taioli E. Polycyclic aromatic hydrocarbon-DNA adducts in human sperm as a marker of DNA damage and infertility. Mutat Res 2003; 535:155 – 60.

Goldstein LS. To BaP or not to BaP? That is the question. Environ Health Perspect 2001;109:A356 – 7.

ICRP (The international Commission on Radiological Protection). Human respiratory tract model for radiological protection. ICRP Publication. New York, NY7 Elsevier; 1994. ICRP.

Hussain M, Rae J, Gilman A, Kauss P. Lifetime health risk assessment from exposure of recreational users to polycyclic aromatic hydro-carbons. Arch Environ Contam Toxicol 1998;35:527 – 31. Kingham S, Briggs D, Elliott P, Fischer P, Lebret E. Spatial variations

in the concentrations of traffic-related pollutants in indoor and outdoor air in Huddersfield, England. Atmos Environ 2000; 34:905 – 16.

Li CS, Ro YS. Indoor characteristics of polycyclic aromatic hydro-carbons in the urban atmosphere of Taipei. Atmos Environ 2000; 34:611 – 20.

Malcom HM, Dobson S, 1994. The calculation of an environmental assessment level (EAL) for atmospheric PAHs using relative potencies. London, UK7 Department of the Environment; 1994. p. 34 – 46.

Naumova YY, Eisenreich SJ, Turpin BJ, Weisel CP, Morandi MT, Colome SD, et al. Polycyclic aromatic hydrocarbons in the indoor and outdoor air of three cities in the US. Environ Sci Technol 2002;36:2552 – 9.

Nisbet ICT, LaGoy PK. Toxic equivalency factors (TEFs) for poly-cyclic aromatic hydrocarbons (PAHs). Regul Toxicol Pharm 1992;16:290 – 300.

OEHHA (The international Commission on Radiological Protection). Benzo[a]pyrene as a toxic air contaminant, Part B, health effects of Benzo[a]pyrene. Berkeley, CA7 Air Toxicology and Epidemi-ology Section; 1993.

Perera F, Hemminke K, Jedrychowski W, Whyatt R, Campbell U, Hsu Y, et al. In utero DNA damage from environmental pollution is associated with somatic gene mutation in newborns. Cancer Epi-demiol Biomark Prev 2002;11:1134 – 7.

Schmahl D, Schmidt KG, Habs M. Syncarcinogenic action of poly-cyclic aromatic hydrocarbons in automobile exhaust gas conden-sates. IARC Scientific Publication. Lyon, France7 WHO; 1977. Somers CM, Yauk CL, White PA, Parfett CLJ, Quinn JS. Air pollu-tion induces heritable DNA mutapollu-tions. Proc Natl Acad Sci U S A 2002;99:15904 – 7.

Somers CM, McCarry BE, Malek F, Quinn JS. Reduction of partic-ulate air pollution lowers the risk of heritable mutations in mice. Science 2004;304:1008 – 10.

Thyssen J, Althoff J, Kimmerle G, Mohr U. Inhalation studies with benzo[a]pyrene in Syrian golden hamsters. J Natl Cancer Inst 1981;66:575 – 7.

Tsai PJ, Shieh HY, Lee WJ, Lai SO. Health risk assessment for workers exposed to polycyclic aromatic hydrocarbons (PAHs) in carbon black manufacturing industry. Sci Total Environ 2001; 278:137 – 50.

Tsai PJ, Shih TS, Chen HL, Lee WJ, Lai CH, Liou SH. Assessing and predicting the exposures of polycyclic aromatic hydrocarbons (PAHs) and their carcinogenic potencies from vehicle engine exhausts to highway toll station workers. Atmos Environ 2004; 38:333 – 43.

USEPA. Dermal exposure assessment: principles and applications. Office of Health and Environmental Assessment; 1992. EPA/ 600/6-88/005Cc.

USEPA. Risk assessment Guidance for superfund, volume I: human health evaluation manual (Part E, Supplemetal Guidance for dermal risk assessment), EPA/540/R/99/005. Washington DC, USA7 Office of Emerage and Remedial Response; 2001. Venkataraman C, Raymond J. Estimating the lung deposition of

particulate polycyclic aromatic hydrocarbons associated with mul-timodal urban aerosols. Inhal Toxicol 1998;10:183 – 204. Vyskocil A, Viau C, Camus M. Risk assessment of lung cancer related

to environmental PAH pollution sources. Human Exp Toxicol 2004;23:115 – 27.

Yang HH, Lai SO, Hsieh LT, Hsueh HJ, Chi TW. Profiles of PAH emission from steel and iron industries. Chemosphere 2002; 48:1061 – 74.

Yang HH, Hsieh LT, Liu HS, Mi HH. Polycyclic aromatic hydrocar-bon emissions from motorcycles. Atmos Environ 2005;39:17 – 25.