Science of the Total Environment 443 (2013) 104–111

Contents lists available at SciVerse ScienceDirect

Science of the Total Environment

j ourna l homepage: www.e lsev ie r .com/ locate /sc i totenv

Polychlorinated dibenzo-p-dioxins and dibenzofurans and their association withcancer mortality among workers in one automobile foundry factory

Lihua Wang a,b, Shaofan Weng a,b, Sheng Wen c, Tingming Shi c, Gangtao Sun a,b, Yuyu Zeng d,Cheng Qi d, Weihong Chen a,b,⁎a Department of Occupational & Environmental Health, School of Public Health, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, Hubei, 430030, Chinab Ministry of Education Key Lab of Environment and Health, School of Public Health, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, Hubei, 430030, Chinac Hubei Provincial Center for Disease Control and Prevention, Wuhan, Hubei, 430074, Chinad Dongfeng Prevention and Treatment Center for Occupational Disease, Shiyan, Hubei, 442001, China

H I G H L I G H T S

► PCDD/Fs generated from the automobile foundry process polluted the environment.► The PCDD/F concentrations in melting furnace areas were the highest in workshops.► Lung, liver and all cancer mortality in foundry workers increased significantly.► Relative risk of cancer mortality in melting workers to assistant workers was 2.55.► Dose–response relationship between PCDD/F exposure and cancer mortality was found.

⁎ Corresponding author at: Department of OccupationSchool of Public Health, Tongji Medical College, HuazhTechnology, Wuhan, Hubei, 430030, China. Tel.: +8683692333.

E-mail address: wchen@mails.tjmu.edu.cn (W. Chen

0048-9697/$ – see front matter © 2012 Elsevier B.V. Allhttp://dx.doi.org/10.1016/j.scitotenv.2012.10.073

a b s t r a c t

a r t i c l e i n f o

Article history:Received 7 June 2012Received in revised form 17 October 2012Accepted 18 October 2012Available online 23 November 2012

Keywords:PCDD/FsAutomobile foundryAirDustCancer mortality

Polychlorinated dibenzo-p-dioxins and dibenzofurans (PCDD/Fs) have been reported as possible carcinogenichazards to humans. However, epidemiological studies on their carcinogenic roles are limited. The currentstudy was designed to determine the concentrations and characteristics of PCDD/Fs and evaluate their associa-tion with cancer mortality in exposed workers in one automobile foundry factory.PCDD/F levels in factory and surrounding environment were analyzed through air and settling dust sampling.The cancer mortalities among workers in this foundry factory were calculated using data from a cohort study.The results showed that the PCDD/F concentrations of air in workplace ranged 0.36–2.25 pg World HealthOrganization-Toxic Equivalent (WHO-TEQ) Nm−3 (average 1.01 pg WHO-TEQ Nm−3), which were 1.16–7.26 times higher than those outside the factory. The PCDD/F concentrations of settling dust in the workplaceranged 3.34–18.64 pg WHO-TEQ g−1 (average 8.25 pg WHO-TEQ g−1), which were lower than those justoutside the factory (average 16.13 pg WHO-TEQ g−1). Furthermore, a cohort study of workers in this factorywith average follow-up of 24.52 years showed that cancer was the leading cause of death, with significantelevated mortality (standardized mortality ratio (SMR)=1.70, 95% confidence interval (CI): 1.35–2.13)among workers, when compared with Chinese national mortality. The cancer mortality among front-lineworkers was increased significantly (adjusted relative risk (RR)=1.73, 95% CI: 1.14–2.60), particularlyamong melting and casting workers, when compared with that among assistant workers. Our results indicat-ed that there was a dose–response relationship between PCDD/F exposure and cancer mortality amongfoundry workers.

© 2012 Elsevier B.V. All rights reserved.

1. Introduction

Polychlorinateddibenzo-p-dioxins anddibenzofurans (PCDD/Fs) are abroad class of compounds which consist of two benzene rings connected

al and Environmental Health,ong University of Science and27 83691677; fax: +86 27

).

rights reserved.

by 1–2 oxygen atoms and contain 1–8 chlorines. PCDD/Fs involve 75congeners in PCDDs and 135 congeners in PCDFs, but of these, only 17congeners (chlorinated in at least the 2, 3, 7 and 8 positions) are consid-ered to be biologically active. Because these compounds vary widely intoxicity, the concept of toxic equivalency factors (TEFs) has been devel-oped to conduct risk assessment and regulatory control. The referencecongener is the most toxic dioxin, 2,3,7,8-tetrachlorodibenzo-p-dioxin(TCDD), which is defined as a TEF of 1. Other congeners are assignedtoxicity ratings relative to this compound. So far, only seven dioxin

105L. Wang et al. / Science of the Total Environment 443 (2013) 104–111

congeners and ten furan congeners have been assigned TEF values. Thissystem was newly revised by the World Health Organization (WHO) in2005, and termedWHO-TEQs (Van den Berg et al., 2006).

PCDD/Fs are by-products of numerous industrial activities andcombustion processes, and have caused public concern worldwide(Chang et al., 2004; Fiedler, 2003; Oberg, 2007). In China, the ferrousand non-ferrous metal industries were reported to account for themajority of the estimated total PCDD/F emissions (Zhu et al., 2008).PCDD/Fs had been detected in the stack gas and fly ash releasedfrom thermal sources in some Chinese metallurgical industries (Baet al., 2009a, b; Chang et al., 2006; Lin et al., 2007; Liu et al., 2009;Lv et al., 2011; Wang et al., 2003). For example, Ba et al. (2009b)observed that the mean WHO-TEQ values of PCDD/Fs were 0.3–2.84 ng Nm−3 in stack gas and 4.42–29.3 ng g−1 in fly ash, respec-tively, released from secondary aluminum and copper metallurgy.Foundry processes are also considered to be a potential source ofPCDD/Fs (UNEP, 2005). However, the reports on PCDD/F levels inthe foundry industry and surrounding environment are very scarce.Furthermore, to our knowledge, neither the information of PCDD/Flevels in air or settling dust, nor personal exposure of workers toPCDD/Fs, was reported in foundries.

PCDD/Fs are commonly thought of as highly toxic compounds andmay produce a wide spectrum of adverse health effects, includingtoxicity and carcinogenicity to multiple organs and tissues (Schecteret al., 2006). The International Agency for Research on Cancer (IARC)and the United States Environmental Protection Agency (US EPA)have categorized PCDD/Fs as “possible human carcinogens” based onanimal experiments. Scientists indicated that the reports on adversehealth effects from PCDD/Fs on human beings were very limited andneed further evaluation (IARC, 1997).

In this study, we determined the PCDD/F levels and profiles in bothair and settling dust of several worksites in an automobile foundryfactory and surrounding environment. Based on a cohort study ofworkers in the same factory, we calculated the cancer mortality amongworkers in different job categories. The objectives of this study wereto: (1) determine PCDD/F levels and profiles in and out of the foundryfactory; (2) estimate workers' personal exposure to PCDD/Fs; and (3)evaluate the relationship between PCDD/F exposure and cancer mortal-ity among workers.

2. Materials and methods

2.1. The selected foundry and production process

A large automobile foundry factory located in Hubei province inChina was selected for this study. This factory mainly produces spareparts for large commercial vehicles. It is a subsidiary of the largest

Fig. 1. Sampling sites inside an

commercial vehicle company in China. Both workshops (A and B) ofthe factory were chosen for this study. In each workshop, mediumfrequency induction furnace is used for melting the iron. The furnacecapacity is 5 tons. The feedstock mainly include pig iron, steel scrapsand waste castings. In a typical 10 hour shift, the castings are 80 tons,to 100 tons. It includes natural and mechanical ventilation in the work-shop. The extracted gases were discharged to atmosphere after theyacross several condensing purification devices. All workers were ontwo rotating 10-h shifts per day, and 4 h were used for cleaning andservicing. The workers could stay about 2 to 3 h in control rooms (thedoors are not closed) during their operation. A brief description of theproduction process is as follows: the charging stocks including pig iron,steel scraps andwaste castings aremelted in furnaces, and then dumpedinto the casting furnaces on the bench. Next, molten metals are pouredinto the mode cavities, and then cooled down and the new castingsthen polished. In the melting process of raw materials in furnaces,smoke and gas often escape over the rooftops of workshops. In thewhole casting link, the maximum temperature in the furnace canreach >1600 °C and decline drastically, reaching 200–500 °C for about2 min prior to circulation.

2.2. Sampling sites and sample collection

Air samples were collected using high volume air samplers for atleast 8 h. The sampling process was conducted according to therevised US EPA method TO 9A (US EPA, 1999). Briefly, the samplingflow rate was set at 0.238 m3 min−1. Glass fiber filters (GFF,10.16 cm diameter) were used to collect particle-bound PCDD/Fsand polyurethane foam (PUF, 6.3 cm diameter, 7.6 cm length) mate-rials were used to absorb vapor-phase PCDD/Fs. Prior to sampling, theGFFs were baked at 400 °C for at least 6 h to remove organic contam-inants and the PUFs were extracted with 250 ml acetone in Soxhletextractors.

We set four sampling points (S1 to S4) in workshop A, one point(S5) in workshop B, and one point (S6) outside of the factory. S1 wasbetween two melting furnaces, S2 outside of the control room, S3near the casting table, and S4 located in the molding area. Because thetechnology and process in workshop B are similar to that in workshopA, we only set S5 near the melting furnace in workshop B. Taking intoaccount the local prevailing wind direction, S6 was set at the entranceof the factory which is about 250 m downwind from workshop A and500 m downwind from workshop B. S6 was used to determine PCDD/Fs emitted from the factory, and represented the surrounding environ-ment pollution. The layout of sampling sites was illustrated in Fig. 1.Two parallel air samples were taken at one sampling site. Meanwhile,five settling dust samples were collected near each air sampling site.The settling dusts were collected into pre-cleaned vials with small,

d outside of the foundry.

106 L. Wang et al. / Science of the Total Environment 443 (2013) 104–111

solvent-rinsed brushes. All samples were wrapped in aluminum foil toavoid contamination and photolysis, and were preserved at−20 °C.

2.3. Sample cleanup processing

Sample preparation was performed using the method adapted fromUS EPA Method 1613B (US EPA, 1997) and previously reported byWen et al. (2008). All solvents were pesticide residue grade andpurchased from J. T. Baker (J.T. Baker, Inc., NJ, USA). Silica gel (100–200mesh) and basic aluminum (Alumina B Super I) were purchased fromICN (Eschwege, Germany). Florisil was obtained from LGC Promochem(Hadfield, UK). Standard solutions of PCDD/Fs (1613-LCS, 1613-IS)were obtained fromCambridge Isotope Laboratories Inc. (Massachusetts,USA). Each sample was spiked with a cocktail of 13C12-labeled PCDD/Finternal standards and extracted for 24 h by using 250 ml 1:1 mixture(n-hexane:dichloromethane) in a Soxhlet extractor.Moreover, 50 ml ac-etone and some copper wires were also added to each sample for dehy-dration anddesulfurization, respectively. Sample cleanupwas conductedby successively using a multi-layer silica gel column, a florisil column,and an alumina column, and then gently concentrated with nitrogenflow. Then, 13C12-labeled injection standards (13C12-1,2,3,4-TCDD and13C12-1,2,3,7,8,9-HxCDD) were spiked before mass analysis. The finalvolume was adjusted to 10 μl.

2.4. Sample analysis by high resolution gas chromatography/highresolution mass spectrometer

The quantification was performed on a high resolution gas chroma-tography (HRGC) coupled with a high resolution mass spectrometer(HRMS) (Finnigan MAT 95xp, Thermo Electron) with an electronimpact (EI) ion source. Exactly 1 μl of sample solution was injected byan autosampler (AS2000, Thermo Electron) in splitless mode. Theexact instrumental condition for the analysis of PCDD/Fswas conductedaccording to the method reported by Wen et al. (2008).

2.5. Quality assurance (QA) and quality control (QC)

Field and laboratory blank samples covering the whole analyticalprocedures were analyzed every six samples and values were belowdetection limits (LOD). Sampling, extraction and injection internalstandards were used, and recoveries of all 13C12-labeled surrogateswere in the range of 50%–130%, which was in the acceptable rangeestablished by the USEPA 1613B. Meanwhile, the laboratory is quali-fied to determine the PCDD/Fs.

2.6. Study population and occupational exposure to other hazards

A cohort of 3529 workers, who had worked in this automobilefoundry factory at least 1 year between January 1, 1980 and December31, 1985, was identified previously. The subjects, including front-lineworkers (melting, casting and modeling workers) and assistantworkers, were followed up from January 1, 1980 through December31, 2005. Demographic datawas collected from register andworkhisto-ry information taken from occupational records at the factory. Allsubjects were tracked for their health status by local hygienists. For alldeceased subjects, the cause of death was collected through medicalrecords from local or regional hospitals. The study was approved bythe Tongji Medical College Institutional Review Board.

We used data from standardized monitoring program in workplacesto track potential environmental hazards, including crystalline silica,polycyclic aromatic hydrocarbons (PAHs), phenol, formaldehyde andmetal elements. The results showed that respirable crystalline silica con-centrations ranged from 0.29 mg m−3 to 2.76 mg m−3 in different jobcategories. The concentrations of PAHs ranged between 0.08 μg m−3

and 1.30 μg m−3, and carcinogenic benzo (a) pyrene ranged between0.001 μg m−3 and 0.016 μg m−3. The concentration range for

phenol and formaldehyde were 0.43 mg m−3 to 6.57 mg m−3 and0.05 mg m−3 to 0.39 mg m−3 separately. Both are lower than the ex-posure limit for phenol (10 mg m−3) and formaldehyde (0.5 mg m−3)exposure in the workplace specified by Chinese Occupational Healthand Standards Committee. Findings indicated very low exposure tolead, nickel, manganese, and cadmium in the studied workplaces.

2.7. Statistical analysis

Continuous variableswere described usingmeans and standard devi-ations, and compared using Analysis of Variance (ANOVA). Categoricalvariableswere comparedusing the Chi-square test. Standardizedmortal-ity ratio (SMR) was defined as the ratio of observed-to-expected deaths,and was calculated by using Chinese national mortality rates(1980–2005) as a reference. Relative risks (RRs) and the corresponding95% confidence intervals (CIs)were estimated usingmultivariate uncon-ditional logistic regression analysis adjusted for age, gender, smokingstatus, crystalline silica exposure, duration of employment and educationlevels. All analyses were performed using SAS version 9.1 (SAS institute,Cary, NC, USA). All given P values are two-sided.

3. Results and discussion

3.1. Concentrations of PCDD/Fs in air samples

The concentrations andWHO-TEQs of PCDD/F congeners in air sam-ples are shown in Table 1. For congeners with concentrations below thelimit of detection (LOD), a half value of LODwas assigned for the calcu-lation of total TEQ. For sites S1–S5 located in the workshops of the fac-tory, the air concentrations and WHO-TEQs of PCDD/Fs ranged 16.15–84.88 pg Nm−3 (average 50.06 pg Nm−3) and 0.36–2.25 pg Nm−3

(average 1.01 pg Nm−3), respectively. The PCDD/F concentrations inS1 and S5 were higher than those at other sampling sites. Both S1 andS5 were located near melting furnaces, although they were in differentworkshops. The concentration of PCDD/Fs in the casting table site waslower than S1 and S5, but higher than those at other sites. The concen-tration of PCDD/Fs outside of the control room (S2)was the lowest in allof the workshops. The concentration and WHO-TEQ of PCDD/Fs in S6were 19.26 pg Nm−3 and 0.31 pg Nm−3, respectively, which werelower than the PCDD/F concentrations inside the workshops.

In this study, the air concentrations of PCDD/Fs in the furnace andcasting areas were 1.73 and 0.69 pg WHO-TEQ Nm−3, respectively,somewhat lower than those in a steelmaking plant reported by Arieset al. (2008),whichwere 2.46 and0.94 pg WHO-TEQ Nm−3, respective-ly. The average air concentration of PCDD/Fs in the current studywas 1.01 pg WHO-TEQ Nm−3 lower than that in a secondary coppersmelter (12.4 pg WHO-TEQ Nm−3) and secondary aluminum smelter(7.2 pg WHO-TEQ Nm−3) similar to that in an electric arc furnace in-dustry (1.8 pg WHO-TEQ Nm−3) (Lee et al., 2009) but higher thanthat in a cement manufacturing plant (b0.6 pg WHO-TEQ Nm−3)(Sweetman et al., 2004). The other studies assessed the dioxin emissionsin foundry factories. The study by Lv's research group (2011) reportedthat the PCDD/F levels in the stack gas and fly ash were 0.06–0.23 ng m−3 and 1.83–112 ng kg−1, respectively and Grochowalskiet al. (2007) reported that the PCDD/F levels in the stack gas were0.02–1.18 ng m−3. The PCDD/F levels in stack gas are higher than thatin environmental air in the current study.

Several studies evaluated the PCDD/F levels both in workplaces andsurrounding environments. Chen et al. (2004) reported that the PCDD/Fconcentration of outdoor air was 0.37 pg I-TEQ Nm−3 during theoperation period of a secondary aluminum smelter, which wasslightly higher than what was obtained in the current study(0.31 pg WHO-TEQ Nm−3). The mean total PCDD/F concentration inoutdoor air of the municipal solid waste incinerators was previouslyreported as 0.078 pg I-TEQ Nm−3 (Shih et al., 2006), which was

Table 1Summary of PCDD/Fs and their special congeners' concentrations in air samples collected from six sites in the foundry factory (pg Nm−3) (n=2).

S1 S2 S3 S4 S5 S6

2378-TCDD 0.07±0.02 0.05±0.00 0.07±0.04 0.15±0.10 0.31±0.14 0.02±0.0112378-PeCDD 0.14±0.00 0.01±0.02 0.06±0.04 0.02±0.00 0.27±0.09 0.02±0.00123478-HxCDD 0.12±0.01 0.03±0.00 0.06±0.04 0.03±0.01 0.16±0.02 0.02±0.00123678-HxCDD 0.16±0.01 0.06±0.00 0.09±0.04 0.05±0.01 0.21±0.04 0.04±0.01123789-HxCDD 0.19±0.06 0.05±0.01 0.11±0.01 0.04±0.00 0.20±0.00 0.03±0.001234678-HpCDD 0.54±0.05 0.29±0.01 0.33±0.09 0.29±0.04 0.65±0.04 0.25±0.02OCDD 0.66±0.16 0.97±0.16 0.47±0.10 0.95±0.16 1.29±0.41 0.78±0.14Total PCDDs 6.04±0.05 3.64±0.13 4.29±0.93 2.89±0.15 12.79±2.84 3.31±0.572378-TCDF 0.59±0.01 0.30±0.08 0.43±0.06 0.19±0.02 1.27±0.83 0.15±0.0212378-PeCDF 0.83±0.00 0.18±0.06 0.41±0.14 0.19±0.05 1.87±1.26 0.18±0.0123478-PeCDF 1.15±0.01 0.30±0.04 0.56±0.28 0.65±0.61 2.35±1.31 0.21±0.00123478-HxCDF 1.58±0.24 0.44±0.09 0.80±0.45 0.40±0.12 2.07±0.86 0.42±0.06123678-HxCDF 1.38±0.02 0.39±0.05 0.92±0.19 0.35±0.08 2.09±0.87 0.45±0.08234678-HxCDF 1.18±0.08 0.31±0.01 0.65±0.37 0.32±0.04 1.71±0.46 0.39±0.07123789-HxCDF 0.41±0.04 0.11±0.01 0.25±0.14 0.11±0.02 0.64±0.13 0.12±0.021234678-HpCDF 4.39±0.65 1.77±0.27 3.20±1.67 1.78±0.36 5.22±0.31 2.61±0.511234789-HpCDF 0.81±0.17 0.29±0.04 0.56±0.33 0.28±0.09 0.86±0.01 0.38±0.09OCDF 3.71±1.80 1.91±0.20 3.19±1.76 1.95±0.34 2.68±0.60 2.40±0.53Total PCDFs 70.20±35.07 15.94±0.61 49.20±19.71 13.27±3.72 72.09±26.83 15.95±3.13Total PCDD/Fsa 76.23±35.11 19.58±0.48 53.49±18.77 16.15±3.86 84.88±29.68 19.26±3.69PCDFs/PCDDs ratio 0.10±0.05 0.23±0.01 0.10±0.06 0.23±0.05 0.18±0.03 0.21±0.01Total PCDD/Fs WHO-TEQ (1998)b 1.44±0.03 0.42±0.03 0.81±0.27 0.68±0.18 2.75±1.27 0.35±0.02Total PCDD/Fs WHO-TEQ (2005)c 1.20±0.03 0.36±0.02 0.69±0.21 0.56±0.06 2.25±0.98 0.31±0.02

Note: S1— 6 stand for these sampling sites: S1—melting furnace, S2— control room, S3— casting table, S4—modeling, S5—melting furnace in another workshop, S6— entrance offactory. Two parallel samples were taken at one sampling site.

a Sum of total PCDD/Fs (tetra- to octa-CDD/Fs).b TEQ was calculated by WHO-TEQ 1998 (Van den Berg et al., 1998).c TEQ was calculated by WHO-TEQ 2005 (Van den Berg et al., 2006).

107L. Wang et al. / Science of the Total Environment 443 (2013) 104–111

much lower than that in the current study andmainly due to differencesin the thermal process and materials.

3.2. Congener profiles of PCDD/Fs in air samples

The 17 PCDD/F congener profiles of air samples are shown in Fig. 2A.In this study, five congeners, comprising OCDF, 1,2,3,4,6,7,8-HpCDF,1,2,3,4,7,8-HxCDF, 1,2,3,6,7,8-HxCDF and OCDDwere the dominant spe-cies, which accounted for almost 60% of the total 17 PCDD/Fs. Withrespect to TEQ congener profiles, the concentration of 2,3,4,7,8-PeCDFwas the highest, followed by 2,3,7,8-TCDD, 1,2,3,4,7,8-HxCDF,1,2,3,6,7,8-HxCDF, 1,2,3,7,8-PeCDD and 2,3,4,6,7,8-HxCDF. These six con-geners contributed more than 80% to the total TEQs in all air samples.

Congener profiles of PCDD/Fs are correlated with different pathwaysinvolved in dioxin formation. Different techniques or different types ofraw material in the thermal processes may vary the concentrations andcharacteristics of PCDD/Fs, while similar industrial processes couldproduce similar profiles. Generally, the 17 PCDD/F congeners in the sixair samples were very similar, although the PCDD/F concentrations atdifferent sites varied over a wide range. The PCDD/Fs in S6 in this studywas thought to be derived from the foundry factory because its congenerprofileswere the same as other samples inside theworkshops. The ratiosof PCDFs/PCDDs at all sampling sites were greater than 1 (ranged 4.35–10.53), and the congener profiles described in this studywere quite sim-ilar to those in steel plants (Aries et al., 2008; Li et al., 2010), inwhich themain contributors to the TEQ were PCDFs, particularly 2,3,4,7,8-PeCDF.

3.3. Concentrations of PCDD/Fs in dust samples

The concentrations andWHO-TEQs of PCDD/Fs in dust samples weredescribed in Table 2. For sites S1–S5 located inworkshops, the concentra-tions andWHO-TEQs of PCDD/Fs ranged 171.55–812.62 pg g−1 (average391.89 pg g−1) and 3.34–18.64 pg g−1 (average 8.25 pg g−1), respec-tively. The PCDD/F concentrations in dust samples around melting fur-naces S1 and S5 (average 14.51 pg g−1) were much higher than those

in other sites inside the workshops. The PCDD/F concentration of dustsamples in S6 was higher than the samples inside the workshops, exceptS5.

Settling dust samples can provide information on the concentration,distribution and fate of contaminants in the surface environment. In thisstudy, the mean PCDD/F concentration of dust samples in workshopswas about 1/7 of those in the two workshops of an electronic-wasteprocessing site in China which was reported by Leung et al. (2011),whereas the PCDD/F concentration of dust outside factory in the currentstudy was five times higher than that of a reference site in Leung'spaper. Low PCDD/F concentrations inside the factory in this study maybe due to the technical ventilation system and daily floor cleaningwith water in workshops. Those operations resulted in PCDD/Fs beingdischarged to the outside of the factory and then deposited on theground. This may be the reason why the concentration of PCDD/Fsoutside the factory was much higher than those in the reference sitereported by Leung et al. (2011), which was an area far removed fromthe PCDD/F emission source (approximately 8 km).

Once PCDD/Fs enter into the environment, they are very difficult todegrade and could be spread a significant distance by rain or wind.Thus, higher PCDD/F concentrations in dust samples outside the factoryin this study call for more attention to PCDD/F control in the workshop.

3.4. Congener profiles of PCDD/Fs in dust samples

The congener profiles of 17 PCDD/Fs in dust samples are shown inFig. 2B. Differences existed in congener profiles among the samplesfrom different sites. The highly chlorinated congeners, OCDD and1,2,3,4,6,7,8-HpCDD, prevailed in PCDDs. In addition, 2,3,7,8-TCDD com-prised a significant proportion in some samples. The profile variedwide-ly for PCDFs, with 2,3,7,8-TCDF and 1,2,3,4,6,7,8-HpCDF being two maincontributors to the total concentration. In the case of TEQ congener pro-files, the concentrations of 2,3,4,7,8-PeCDF, 2,3,7,8-TCDD, 2,3,7,8-TCDF,1,2,3,7,8-PeCDD, 1,2,3,4,7,8-HxCDF and 1,2,3,6,7,8-HxCDF contributedmore than 80% to the total TEQs in all dust samples.

Fig. 2. Congener profiles of air samples (A) and dust samples (B) at sampling sites.

108 L. Wang et al. / Science of the Total Environment 443 (2013) 104–111

Table 2Summary of PCDD/Fs and their special congeners' concentrations in dust samples collected from six sites in the foundry factory (pg g−1) (n=5).

S1 S2 S3 S4 S5 S6

2378-TCDD 1.80±0.05 0.68±0.08 1.23±0.09 0.55±0.08 5.38±0.73 2.98±0.3312378-PeCDD 1.29±0.03 0.42±0.04 0.14±0.03 0.26±0.01 1.95±0.05 1.62±0.22123478-HxCDD 0.79±0.01 0.24±0.06 0.50±0.06 0.43±0.01 1.44±0.03 1.13±0.06123678-HxCDD 0.63±0.09 0.20±0.01 0.64±0.04 0.41±0.04 1.59±0.07 1.77±0.07123789-HxCDD 0.89±0.07 0.33±0.02 0.58±0.10 0.34±0.07 1.87±0.01 2.23±0.311234678-HpCDD 2.14±0.13 1.20±0.10 1.21±0.07 3.25±1.06 4.12±0.51 8.59±1.06OCDD 4.35±1.05 3.42±1.08 5.91±0.26 8.12±2.02 5.00±1.00 27.14±2.39Total PCDDs 60.93±4.58 34.25±2.14 13.85±0.59 45.19±2.17 57.33±9.85 121.99±11.322378-TCDF 15.72±2.10 10.37±2.05 8.56±1.37 3.57±0.53 13.95±1.26 13.82±1.5712378-PeCDF 9.95±0.56 5.39±1.16 2.28±0.59 3.25±0.39 14.32±2.21 15.08±1.9923478-PeCDF 10.53±1.10 5.61±0.95 1.78±0.60 2.96±1.00 17.21±1.51 12.16±0.86123478-HxCDF 6.07±0.80 3.50±0.23 1.91±0.62 3.79±1.03 9.31±1.00 17.03±1.56123678-HxCDF 6.39±0.61 3.54±0.57 1.92±0.85 2.99±0.51 10.03±0.67 17.21±2.89234678-HxCDF 4.72±0.33 3.11±0.82 1.37±0.45 1.97±0.67 11.12±1.00 11.11±0.96123789-HxCDF 2.05±0.22 1.01±0.06 0.49±0.06 0.70±0.05 4.62±0.55 2.76±0.061234678-HpCDF 7.29±0.11 6.02±1.04 3.80±0.71 7.35±1.02 22.78±2.38 57.53±3.851234789-HpCDF 1.45±0.06 0.99±0.03 0.74±0.06 1.32±0.03 4.55±0.57 6.14±1.12OCDF 4.59±0.91 3.73±0.92 2.39±0.19 5.38±1.01 11.65±1.13 29.30±2.58Total PCDFs 408.72±20.17 253.88±12.54 203.66±16.67 126.36±13.33 755.29±22.78 661.94±18.52Total PCDD/Fs a 469.65±19.76 288.13±14.61 217.50±17.78 171.55±15.11 812.62±23.92 783.93±22.67PCDFs/PCDDs ratio 0.15±0.03 0.13±0.02 0.07±0.01 0.36±0.05 0.08±0.01 0.18±0.03Total PCDD/Fs WHO-TEQ (1998) b 12.69±0.10 6.48±0.09 4.04±0.45 4.00±0.03 22.37±0.21 18.85±0.06Total PCDD/Fs WHO-TEQ (2005) c 10.38±0.09 5.25±0.07 3.63±0.42 3.34±0.03 18.64±0.19 16.13±0.05

Note: S1–6 stand for these sampling sites: S1—melting furnace, S2 — control room, S3— casting table, S4—modeling, S5 —melting furnace in another workshop, S6— entrance offactory. Five parallel samples were taken at one sampling site.

a Sum of total PCDD/Fs (tetra- to octa-CDD/Fs).b TEQ was calculated by WHO-TEQ 1998 (Van den Berg et al., 1998).c TEQ was calculated by WHO-TEQ 2005 (Van den Berg et al., 2006).

Fig. 3. Congener profiles of 17 PCDD/Fs for total air and dust samples.

109L. Wang et al. / Science of the Total Environment 443 (2013) 104–111

3.5. Relationships between air and dust samples

A high correlation was observed between the PCDD/F concentra-tions in air samples and those in dust samples collected at the samesites in the workshops (R=0.90, Pb0.05) (Tables 1 and 2). The conge-ner profiles of 17 PCDD/Fs for all air and settling dust samples areshown in Fig. 3. The congener profiles among dust samples were notas consistent as those among air samples. By combining these twotypes of samples, the congener profiles of 17 PCDD/Fs in dust sampleswere weakly correlated to those in air samples (R=0.63, Pb0.05).The variations of composition characteristics of the 17 PCDD/Fs indust were mainly due to differences in degradation rates for congeners.It suggested that the congener profiles of PCDD/Fs in dust samples maynot be suitable to be used as an indicator for the characteristics of PCDD/Fs generated from production process.

3.6. Personal exposure estimation for workers

PCDD/F exposure can occur through the digestive system, respirato-ry system, or skin contact. Most workplace PCDD/F exposure forworkers is through respiratory system. Personal exposure to PCDD/Fsfor workers in workshop Awas estimated using the following equation,adopted from Nouwen et al. (2001):

Inhm=f ¼Vrm=f � Cair � t� f r

Wm=fð1Þ

where Inhm/f is inhalation exposure for male/female workers inpg TEQ kg−1 d−1; Vrm/f is the ventilation rate classified by physicallabor intensities in L min−1 according to the results reported by Liuet al. (1990); Cair is the air concentration of different job categoriesexpressed in pg TEQ m−3; t is the time that male/female workersspend at the worksites, which is 10 h per day; fr is the alveolar fractionretained in lung (0.75 for adults); and body weight Wm/f is set at 70 kgfor males and 60 kg for females.

Table 3 indicates that personal exposure of PCDD/Fs varied withdifferent labor intensities and job categories. More PCDD/Fs were in-haled when physical labor intensity increased. In this study, personal

exposure of PCDD/Fs for melting furnace workers with heavy physicallabor intensity was more than 0.6 pg WHO TEQ kg−1 d−1. Bycomparing PCDD/F exposures in different job categories, we foundthat PCDD/F exposure levels of front-line workers were 1.63–8.25fold higher than those of assistant workers. Here we only took intoaccount the direct exposure to PCDD/Fs from inhalation during

Table 3Estimated personal exposure to PCDD/Fs for workers classified by physical labor inten-sities and job categories (pg TEQ kg−1 d−1).

Male (70 kg)/laborintensities

Female (60 kg)/laborintensities

Light Moderate Heavy Light Moderate Heavy

Pulmonary ventilation(L/min)

33.4 43.9 72.4 29.5 45.3 68.4

Job categoryMelting 0.28 0.36 0.6 0.29 0.44 0.66Casting 0.16 0.21 0.35 0.17 0.25 0.38Modeling 0.13 0.17 0.28 0.14 0.2 0.31Assistant work a 0.08 – – 0.08 – –

a Assistant work here include control room work and entrance guard.

Table 5Mortalities and SMRs from all causes and various cancers for workers in the foundryfactory.

Cause ofdeath

No. of deathsobserved

No. of deathsexpected

Mortality(‰)

SMR a 95% CI

All deaths 265 275.93 3.12 0.96 0.85–1.08Cancer 121 71.34 1.42 1.70b 1.35–2.13Lung cancer 43 20.16 0.56 2.13b 1.58–2.88Liver cancer 32 18.73 0.38 1.71b 1.21–2.42Gastric cancer 15 11.72 0.18 1.28 0.60–2.74Other cancer 31 20.73 0.36 1.50 0.86–2.62

a Standardized mortality ratios (SMRs) were calculated by using Chinese nationalmortality rates as reference.

b Pb0.05.

110 L. Wang et al. / Science of the Total Environment 443 (2013) 104–111

working time without considering other routes of absorption, such asdietary intake per day.

3.7. Cohort mortality study of workers in this foundry factory

The cohort included 3529 subjects (2543 males, 72.06%); baselineinformation is summarized in Table 4. The average age was 29.07 yearsfor subjects entering into the cohort, with no significant differenceamong groups. The average duration of exposure for assistant workers(23.20 years) was longer than that for front-line workers (19.50 years).The smoking prevalence in modeling and casting groups was higherthan that in melting and assistant groups.

By the end of 2005, there were 108 subjects (3.06%) lost to follow up,and 265 subjects died during 85,253.50 person-years of follow up. Can-cer was the leading cause of death, which accounted for 45.66% of alldeaths in this cohort. Although mortality for all causes of death in thiscohort was very close to the Chinese national mortality (SMR=0.96,95% CI: 0.85–1.08), the cancer mortality increased significantly (SMR=1.70, 95% CI: 1.35–2.13) (Table 5). Of all cancers, SMRs from lung cancerand liver cancer were 2.13 (95% CI: 1.58–2.88) and 1.71 (95% CI: 1.21–2.42), respectively. Gastric cancer mortality was slightly increased, butnot statistically significant (SMR=1.28, 95% CI: 0.60–2.74).

As presented in Table 6, logistic regression analysis revealed thatmortality from all cancers was significantly higher in front-lineworkerscompared with assistant workers (adjusted RR=1.73, 95% CI: 1.14–2.60). After adjustment for age, gender, smoking status, crystalline silicaexposure, duration of exposure and education levels, the melting andcastingworkers had significantly increased the risks of cancermortality,and the adjusted RRs were 2.55 (95% CI: 1.43–4.56) and 1.82 (95% CI:1.13–2.93) respectively, while the modeling workers had an increasedrisk of cancer mortality without statistical significance (adjusted RR=1.29, 95% CI: 0.76–2.19), when compared with assistant workers. To-gether with the air PCDD/F concentrations in different job categories,there was a dose–response relationship between PCDD/F exposureand cancer mortality among workers (ptrend=0.0005).

Some epidemiological studies provide evidence that certain ex-posures in iron and steel foundries are carcinogenic to humans, giv-ing rise to lung cancer (IARC, 1984). To the best of our knowledge, nocancer mortality study on PCDD/F exposed workers in a foundry hasbeen reported. As an important constituent of PCDD/Fs, TCDD is clas-sified as a known human carcinogen on the basis of sufficient animal

Table 4Description of the cohort (n=3529) based on different job categories.

Assistant work Modeling

Number (male) 1594 (1148) 752 (563)Age at entering cohort (years) 28.79±8.23 29.67±9.2Duration of exposure (years) 23.20±7.37 19.10±8.1Smoking prevalence (%) 38.90 48.14

a Continuous variables and categorical variables were compared by ANOVA and Chi-squa

experiments, extensive mechanistic studies and limited epidemio-logical data. TCDD acts through a mechanism involving the aryl hy-drocarbon receptor (AhR), which is present in both humans andanimals (Steenland et al., 2004). The carcinogenicity of TCDD hasbeen demonstrated in experiments on rats, mice, and hamsters; his-tiocytic lymphomas, fibrosarcomas, and tumors of the liver, skin,lung, thyroid, tongue, hard palate, and nasal turbinates have beenfound (Della Porta et al., 1987; NTP, 1982; Rao et al., 1988). However,epidemiological data were limited and not consistent in all studies(Bertazzi et al., 2001; Boers et al., 2010; Collins et al., 2009;Consonni et al., 2008; Manz et al., 1991; McBride et al., 2009;Steenland et al., 1999). One cohort study performed by Bertazzi etal. (2001) revealed that risks of all cancers mortality (SMR=1.3,95% CI: 1.0–1.7), lung cancer, lymphatic and hematopoietic tissuecancers, for residents in a dioxin-contaminated area were increasedsignificantly compared with those in a reference area. Conversely,Collins et al. (2009) reported that an excess of all cancer (SMR=1.0, 95%CI: 0.8–1.1) or lung cancer mortality wasn't observed forworkers exposed to dioxins in trichlorophenol production, with theexception of increased soft tissue sarcoma. In the current study, weobserved elevated lung and liver cancer mortality in workers ex-posed to PCDD/Fs. Moreover, increased all cancer mortality amongfront-line workers, when compared with assistant workers, indicat-ed that a dose–response relationship exists between PCDD/F expo-sure and all cancer mortality among foundry workers.

In this study, we used current PCDD/F concentrations to estimatesubjects' historical exposure levels in different job categories, basedon the fact that production processes in this foundry factory have notchanged greatly in the last 20 years. Because the air circulation inwork-shops in recent years has improved over previous years, our data mayunderestimate exposure for subjects.

Possible confounding factors contribute to the increasedmortality ofcancer in this factory include crystalline silica, carcinogenic PAHs,phenol and formaldehyde. The role of carcinogenic benzo(a)pyrene,phenol and formaldehyde was not further evaluated because monitor-ing data suggested that the levels of these hazards were lower thanthe limits of occupational exposure standards in China. We adjustedexposure to crystalline silica when calculated relative risk of cancerbased on the data of silica concentrations in various job categories.Increased mortality from liver cancer in this study could not beexplained by silica exposure. However, the role of crystalline silica

Casting Melting P value a

849 (640) 334 (192) 0.01300 29.07±8.60 29.09±8.23 0.14096 20.05±7.63 18.99±7.31 b0.0001

47.00 38.62 b0.0001

re test respectively.

Table 6Cancer mortalities among workers in different job categories.

Jobcategories

Person-years No. ofdeathsobserved

Mortality(‰)

RR (95%CI)b p trend

Assistant work 38,471.00 36 0.94 1.00Modeling 18,251.00 27 1.48 1.29 (0.76–2.19)Casting 20,577.00 38 1.85 1.82 (1.13–2.93)Melting 7954.50 20 2.51 2.55 (1.43–4.56) 0.0005Front-line work a 46,782.50 85 1.82 1.73 (1.14–2.60)

a Front-line work here include melting, casting and modeling work.b Data were calculated by unconditional logistic regression, adjusted for age, gender,

smoking status, crystalline silica exposure, duration of employment and educationlevels.

111L. Wang et al. / Science of the Total Environment 443 (2013) 104–111

induced lung cancer and possible joint effect of dioxin and silicaexposure in relation to lung cancer should be further analyzed in future.

4. Conclusion

PCDD/Fs generated in production processes in a foundry factorywere equal to, or lower than, those in other metallurgical industries.Industrial sources of PCDD/Fs resulted in contaminations to the sur-rounding environments. A dose–response relationship between PCDD/F exposure and cancer mortality was found, which added further evi-dence in support of PCDD/Fs' carcinogenic roles in humans, suggestingthat enhanced protective measures should be taken for workersexposed to PCDD/Fs.

Acknowledgment

The authors gratefully acknowledge the financial support provided bythe National Basic Research Program of China (2011CB503804), NationalNature Science Foundation of China (20907048), National PostdoctoralScience Fund (20090450459) and the foundation itemof China'sMinistryof Health for standards development projects (2009-03-10).

References

Aries E, Anderson DR, Fisher R. Exposure assessment of workers to airborne PCDD/Fs,PCBs and PAHs at an electric arc furnace steelmaking plant in the UK. Ann OccupHyg 2008;52:213–25.

Ba T, Zheng M, Zhang B, Liu W, Su G, Xiao K. Estimation and characterization ofPCDD/Fs and dioxin-like PCB emission from secondary zinc and lead metallurgiesin China. J Environ Monit 2009a;11:867–72.

Ba T, Zheng M, Zhang B, Liu W, Xiao K, Zhang L. Estimation and characterization ofPCDD/Fs and dioxin-like PCBs from secondary copper and aluminum metallurgiesin China. Chemosphere 2009b;75:1173–8.

Bertazzi PA, Consonni D, Bachetti S, RubagottiM, Baccarelli A, Zocchetti C, et al. Health effectsof dioxin exposure: a 20-year mortality study. Am J Epidemiol 2001;153:1031–44.

Boers D, Portengen L, Bueno-de-Mesquita HB, Heederik D, Vermeulen R. Cause-specificmortality of Dutch chlorophenoxy herbicide manufacturing workers. Occup EnvironMed 2010;67:24–31.

Chang MB, Chang SH, Chen YW, Hsu HC. Dioxin emission factors for automobiles fromtunnel air sampling in Northern Taiwan. Sci Total Environ 2004;325:129–38.

Chang MB, Huang HC, Tsai SS, Chi KH, Chang-Chien GP. Evaluation of the emission char-acteristics of PCDD/Fs from electric arc furnaces. Chemosphere 2006;62:1761–73.

Chen SJ, Lee WS, Chang-Chien GP, Wang LC, Lee WJ, Kao JH, et al. Characterizingpolychlorinated dibenzo-p-dioxins and dibenzofurans in the surrounding environmentand workplace of a secondary aluminum smelter. Atmos Environ 2004;38:3729–32.

Collins JJ, Bodner K, Aylward LL, Wilken M, Bodnar CM. Mortality rates amongtrichlorophenol workers with exposure to 2,3,7,8-Tetrachlorodibenzo-p-dioxin.Am J Epidemiol 2009;170:501–6.

Consonni D, Pesatori AC, Zocchetti C, Sindaco R, D'Oro LC, Rubagotti M, et al. Mortalityin a population exposed to dioxin after the Seveso, Italy, accident in 1976: 25 yearsof follow-up. Am J Epidemiol 2008;167:847–58.

Della Porta G, Dragani TA, Sozzi G. Carcinogenic effects of infantile and long-term2,3,7,8-tetrachlorodibenzo-p-dioxin treatment in the mouse. Tumori 1987;73:99-107.

Fiedler H. Dioxins and Furans(PCDD/PCDF). The handbook of environmental chemistry,3. ; 2003. p. 123–201.

Grochowalski A, Lassen C, Holtzer M, Sadowski M, Hudyma T. Determination of PCDDs,PCDFs, PCBs and HCB emissions from the metallurgical sector in Poland. EnvironSci Pollut Res Int 2007;14:326–32.

IARC. Monograph on the evaluation of the carcinogenic risk to humans. Polynuclear aro-matic compoundsPart 3 Industrial exposures in aluminium production, coal gasifica-tion, coke production, and iron and steel founding, vol. 34. Lyon: IARC; 1984.

IARC. Polychlorinated dibenzo-para-dioxins and polychlorinated dibenzofurans. IARCMonogr Eval Carcinog Risks Hum 1997;69:1-631.

Lee CC, Shih TS, Chen HL. Distribution of air and serum PCDD/F levels of electric arc furnacesand secondary aluminum and copper smelters. J Hazard Mater 2009;172:1351–6.

Leung AO, Zheng J, Yu CK, Liu WK, Wong CK, Cai Z, et al. Polybrominated diphenyl ethersand polychlorinated dibenzo-p-dioxins and dibenzofurans in surface dust at anE-waste processing site in Southeast China. Environ Sci Technol 2011;45:5775–82.

Li Y, Wang P, Ding L, Li X, Wang T, Zhang Q, et al. Atmospheric distribution ofpolychlorinated dibenzo-p-dioxins, dibenzofurans and dioxin-like polychlorinatedbiphenyls around a steel plant area, northeast China. Chemosphere 2010;79:253–8.

Lin LF, Lee WJ, Li HW, Wang MS, Chang-Chien GP. Characterization and inventory ofPCDD/F emissions from coal-fired power plants and other sources in Taiwan.Chemosphere 2007;68:1642–9.

Liu SF, Qi TS, Zhang D, Zhang BZ, Cui GS, Li S, et al. Research about the relationshipsamong labor load, heart rate, oxygen consumption and pulmonary ventilation vol-ume of vegetable growers. China Occup Med 1990;17:207–9.

Liu G, Zheng M, Liu W, Wang C, Zhang B, Gao L, et al. Atmospheric emission of PCDD/Fs,PCBs, hexachlorobenzene, and pentachlorobenzene from the coking industry. EnvironSci Technol 2009;43:9196–201.

Lv P, Zheng M, Liu G, Liu W, Xiao K. Estimation and characterization of PCDD/Fs anddioxin-like PCBs from Chinese iron foundries. Chemosphere 2011;82:759–63.

Manz A, Berger J, Dwyer JH, Flesch-Janys D, Nagel S,Waltsgott H. Cancer mortality amongworkers in chemical plant contaminated with dioxin. Lancet 1991;338:959–64.

McBride DI, Collins JJ, Humphry NF, Herbison P, Bodner KM, Aylward LL, et al. Mortalityin workers exposed to 2,3,7,8-tetrachlorodibenzo-p-dioxin at a trichlorophenolplant in New Zealand. J Occup Environ Med 2009;51:1049–56.

Nouwen J, Cornelis C, De Fre R, Wevers M, Viaene P, Mensink C, et al. Health risk assessmentof dioxin emissions from municipal waste incinerators: the Neerlandquarter (Wilrijk,Belgium). Chemosphere 2001;43:909–23.

NTP. Carcinogenesis bioassay of 2,3,7,8-Tetrachlorodibenzo-p-dioxin (CAS No.1746-01-6) in Osborne–Mendel rats and B6C3F1 mice (gavage study). Natl ToxicolProgram Tech Rep Ser 1982;209:1-195.

Oberg T. Low-temperature formation and degradation of chlorinated benzenes, PCDDand PCDF in dust from steel production. Sci Total Environ 2007;382:153–8.

Rao MS, Subbarao V, Prasad JD, Scarpelli DG. Carcinogenicity of 2,3,7,8-tetrachlorodibenzo-p-dioxin in the Syrian golden hamster. Carcinogenesis 1988;9:1677–9.

Schecter A, Birnbaum L, Ryan JJ, Constable JD. Dioxins: an overview. Environ Res2006;101:419–28.

Shih SI, Wang YF, Chang JE, Jang JS, Kuo FL, Wang LC, et al. Comparisons of levels ofpolychlorinated dibenzo-p-dioxins/dibenzofurans in the surrounding environ-ment and workplace of two municipal solid waste incinerators. J Hazard Mater2006;137:1817–30.

Steenland K, Piacitelli L, Deddens J, Fingerhut M, Chang LI. Cancer, heart disease, anddiabetes in workers exposed to 2,3,7,8-tetrachlorodibenzo-p-dioxin. J Natl CancerInst 1999;91:779–86.

Steenland K, Bertazzi P, Baccarelli A, Kogevinas M. Dioxin revisited: developmentssince the 1997 IARC classification of dioxin as a human carcinogen. Environ HealthPerspect 2004;112:1265–8.

Sweetman A, Keen C, Healy J, Ball E, Davy C. Occupational exposure to dioxins at UKworksites. Ann Occup Hyg 2004;48:425–37.

UNEP. Standardized toolkit for identification and quantification of dioxin and furan release.second ed. Geneva: UNEP Chemicals; 2005.

US EPA, Method 1613 (Revision B): tetra-through octa-chlorinated dioxins and furansby isotope dilution HRGC/HRMS. Guidelines Establishing Test Procedure for theAnalysis of Pollutants, Final Rule, 40 CFR Part 136; 1997.

US EPA, Method TO-9A: Determination of Polychlorinated, Polybrominated andBrominated Dibenzo-p-dioxins and Dibenzofurans in the Ambient AirEPA/625/R-96/010b; 1999.

Van den Berg M, Birnbaum L, Bosveld AT, Brunström B, Cook P, Feeley M, et al. Toxicequivalency factors (TEFs) for PCBs PCDDs PCDFs for humans and wildlife. EnvironHealth Perspect 1998;106:775–92.

Van den BergM, Birnbaum LS, DenisonM, De Vito M, FarlandW, Feeley M, et al. The 2005World Health Organization reevaluation of human and mammalian toxic equivalen-cy factors for dioxins and dioxin-like compounds. Toxicol Sci 2006;93:223–41.

Wang LC, Lee WJ, Lee WS, Chang-Chien GP, Tsai PJ. Characterizing the emissions ofpolychlorinated dibenzo-p-dioxins and dibenzofurans from crematories and theirimpacts to the surrounding environment. Environ Sci Technol 2003;37:62–7.

Wen S, Yang FX, Gong Y, Zhang XL, Hui Y, Li JG, et al. Elevated levels of urinary8-hydroxy-2′-deoxyguanosine in male electrical and electronic equipmentdismantling workers exposed to high concentrations of polychlorinated dibenzo-p-dioxins and dibenzofurans, polybrominated diphenyl ethers, and polychlorinatedbiphenyls. Environ Sci Technol 2008;42:4202–7.

Zhu J, Hirai Y, Yu G, Sakai S. Levels of polychlorinated dibenzo-p-dioxins and dibenzofuransin China and chemometric analysis of potential emission sources. Chemosphere2008;70:703–11.