Biomonitoring of Chromium for Residents of Areas with a High Density of Electroplating Factories

www.nature.com/jes

Biomonitoring of chromium for residents of areas with a high density of electroplating factories

FENG-HSIANG CHANG,a,b SHU-LI WANG,a,c YEOU-LIH HUANG,d MING-HSIEN TSAI,e SHENG-TSUNG YUa AND LOUIS W. CHANGa aDivision of Environmental Health and Occupational Medicine, National Health Research Institute, Zhunan, Taiwan bDepartment of Information Management, Tzu Hui Institute of Technology, Pingtung, Taiwan cGraduate Institute of Occupational Safetyand Health, Kaohsiung Medical University,Kaohsiung, Taiwan dFacultyof Biomedical LaboratoryScience, Kaohsiung Medical University,Kaohsiung, Taiwan eDivision of Basic Medicine, Graduate Institute of Medicine, Kaohsiung Medical University, Kaohsiung, Taiwan

The high densityof electroplating factories in the geographic middle of Taiwan has prompted concern over the potential for exposure to harmful metal s. The present studyaimed to determine the levels of chromium in whole blood (B-Cr) of residents living in the high vs. low factory-densityareas, and to examine the relations to gender and age. A total of 660 residents who had not moved within the 5 years preceding the study were sampled according to the stratified random sampling approach, at ages 35–44, 45–54, and 55–64 years, for both genders. Chromium determinations (n ¼ 641) were made using a graphite furnace atomic absorption spectrometer. The geometric mean (95% C.I.) of B-Cr was 0.357 (0.34–0.38) mg/l. The International Federation of Clinical Chemistry(IFCC) nonparametric 0.95 reference limits of B-Cr was estimated to be o0.905 mg/l. B-Cr levels decreased with increasing age. Subjects in the areas with a high density(0.38 mg/l, 95% C.I.: 0.36–0.40) of electroplating factories had significantlyhigher B-Cr levels, compared to residents of the low-density(0.27, 0.25–0.30) areas and to the general population from western countries. The health significance of the elevated B- Cr remains to be determined. Journal of Exposure Science and Environmental Epidemiology (2006) 16, 138–146. doi:10.1038/sj.jea.7500445; published online 17 August 2005

Keywords: biomonitoring, chromium, whole blood, graphite furnace atomic absorption spectrometer, reference values.

Introduction Among the major oxidative states of chromium encountered in occupational and environmental settings, chromium (VI) Metal contamination of agriculture soil has been documented is a procarcinogen that is reduced intracellularlyto form in Changhua County, a rural area of approximately DNA-damaging species while chromium (III) is considered 2 1074 km situated in the middle of the island of Taiwan to be essential in nutrition and for the maintenance of normal (ROCEPA, 2002). While the origin of the contaminants is glucose tolerance (Versieck and Cornelis, 1989; Chi, 1997; not known absolutely, the metal-work-related factories in the Zhitkovich, 2002). Chromium (VI) does not itself bind to region are suspect. Most of the industrial plants in Changhua DNA but is reduced to chromium (III) which does. The Countyare located in the northern region (e.g. Changhua binding of chromium (III) is insufﬁcient to damage DNA in City, Hemei, Lugang, Sioushuei and Huatan) (Figure 1) and vitro and it is supposed that damage arises from intermediates served metal work, electroplating, and other metal surface in the reduction process, perhaps intermediate valence states treatment industries. These plants have long been suspected of Cr itself (Wetterhahn and Hamilton, 1989; Aiyar et al., of discharging wastewater containing large amounts of heavy 1991; Standeven and Wetterhahn, 1991). metals including chromium into irrigation channels and rivers Exposure to chromium (VI) compounds has been (IDBMOEA, 2002; Lin et al., 2002; ROCEPA, 2002). consistentlyfound to be associated with an elevated incidence Chromium is found in several oxidative and physical forms of respiratorycancers and other adverse health effects that differ substantiallyin their toxicological potency. (IARC, 1990; Langardt, 1990; Sorahan et al., 1998). The genotoxic potential of chromium (VI) has been conﬁrmed in animal experiments and in several cell-based assays (Bieder- 1. Address all correspondence to: Dr. S-L. Wang, Division of Environ- mann and Landolph, 1990; Snow, 1992). Chromium (VI) is mental Health and Occupational Medicine, National Health Research the second most potent allergen after nickel and chromium- Institutes, 100 Shih-Chuan 1st Road, Kaohsiung 807, Taiwan, ROC. contact allergies are frequentlyfound among occupationally Tel.: þ 886-7-3126772 ext. 4015. Fax: þ 886-7-3221912. E-mail: [email protected] exposed workers. The highest incidence of chromium Received 31 December 2004; accepted 30 May2005; published online 17 sensitivityis found in chromium-plating industry,manufac- August 2005 turing of mineral pigments, shipbuilding, textile industry, Chromium biomonitoring in exposed populations Chang et al.

High Factory-density Areas: Changhua A Changhua City County B Hemei C Lugang D Sioushuei E Huatan

Taiwan

Control Areas:

F Hsicou G Erhshui

Figure 1. Studyarea of the chromium biomonitoring in middle Taiwan.

and cement-exposed builders (Baruthio, 1992). A previous nated soil and both high and low concentrations of studyconducted for the electroplating workers exposed to electroplating factories to be identified. chromicacidinmiddleTaiwanhasshowntheassociated Subjects were recruited from seven different townships of health effects on immunological parameters (Kuo and Wu, Changhua County; Changhua City (A), Hemei (B), Lugang 2002). (C), Sioushuei (D), Huatan (E), Hsicou (F), and Erhshui The present studyaimed to determine the levels of (G) (Figure 1). Townships A–E were considered representa- chromium in whole blood (B-Cr) of randomlyselected tive of the environment encountered bythe largest proportion Changhua residents, assess the significance of the metal of the Changhua population. These five townships containing exposure and parameters including physical location, gender most of the metal-work-related factories in Changhua were and age, and to compare the determined chromium reference considered as the high factory-density areas. The other two values of Changhua residents with those from other countries. townships, F and G, were the control areas (the low factory- densityareas); few metal-work-related factories were located in these townships. Methods The selection process also utilized population census data of the Ministryof the Interior of the Republic of China. To StudyArea and Subjects be eligible, subjects needed to be 35–64 years of age and to Data of soil metal pollution was acquired from the have resided in their present household for at least 5 Environmental Protection Administration of the Republic consecutive years at the time of recruitment. By means of of China (ROCEPA, 1989, 1992, 2002) and data of factory stratified random sampling approach for each studyarea, we registration downloaded from an online database of In- further stratified the studypopulation bygender into three dustrial Development Bureau Ministryof Economic Affairs age groups (35–44, 45–54, and 55–64 years) and 22 villages. (IDBMOEA) at website: http://www.moeaidb.gov.tw/ Fiveresidentswereselectedineachstratum(2Â 3 Â 22=132 Fidbweb/index.jsp. This allowed areas with both contami- strata), thus a total number of 660 subjects were recruited.

Journal of Exposure Science and Environmental Epidemiology (2006) 16(2) 139 Chang et al. Chromium biomonitoring in exposed populations

Home interviews were conducted with the candidates by Table 1 . Distribution of age and consecutive resident duration by local public health nurses. Those who agreed to participate gender and township. were given an information sheet outlining the details of procedures and related outcomes, and were asked to provide Township n Age (years) Consecutive resident formal written consents. Ethics approval for this studywas duration (years) obtained from the Human Ethics Committee of the National Mean (SD) Mean (SD) Health Research Institutes, Taiwan (approval no. EC9109002). After obtaining informed consent, a question- Changhua City171 49.4 (8.3) 20.2 (12.7) Male 86 49.6 (8.3) 21.7 (14.2) naire was administered bythe interviewing nurse for the Female 85 49.1 (8.4) 18.7 (10.9) purpose of collecting information on residential and occupational history, personal habits, lifestyle, and medical history. Hemei 177 49.2 (8.5) 26.6 (16.5) All participants also provided a blood sample (see below), Male 88 49.6 (8.8) 29.7 (18.2) which was collected bynurses accompanied bya physicianat Female 89 48.9 (8.2) 23.5 (14.0) a local cooperative health center. Lugang 117 49.5 (9.0) 31.3 (18.0) A total of 649 people completed the questionnaire and Male 57 49.8 (8.9) 36.2 (18.6) provided a blood sample. However, the sample volume was Female 60 49.3 (9.2) 26.7 (16.3) insufﬁcient for eight subjects. Thus, the studyresults were based on 641 subjects. The general characteristics of the Sioushuei 27 49.8 (8.6) 28.1 (14.9) Male 12 50.0 (8.9) 35.6 (14.7) studypopulation are summarized in Table 1. Female 15 49.7 (8.6) 22.1 (12.5)

Sample Collection Huatan 29 52.1 (8.7) 15.8 (8.7) Blood samples were collected as previouslydetailed (Cornelis Male 10 51.7 (7.8) 16.3 (9.3) et al., 1996) to minimize contamination, deterioration, and Female 19 52.3 (9.4) 15.5 (8.6) overestimation of chromium levels. The conventional use of a Hsicou 60 49.7 (8.1) 32.4 (16.3) needle and syringe and heparin anticoagulant was avoided, to Male 30 50.3 (8.0) 39.0 (16.8) minimize their potential as contributors of chromium Female 30 49.0 (8.2) 25.7 (12.9) (Versieck et al., 1982; Minoia et al., 1992; Christensen et al., 1993). Instead, the BD Vacutainers Evacuated Blood Erhshui 60 51.3 (8.4) 23.3 (14.5) Male 28 52.2 (8.5) 26.0 (17.5) Collection System (Becton Dickinson, Franklin Lakes, NJ, Female 32 50.5 (8.4) 20.9 (11.0) USA) was used to collect venous blood samples in 7-ml BD s Vacutainer tubes (for trace element tests) (REF 367735, Total 641 49.7 (8.5) 25.6 (16.0) Belliver Industrial Estate, Plymouth, UK). All participants Male 311 50.0 (8.5) 29.0 (17.8) were notified not to eat seafood for 3 days and to have fasted Female 310 49.4 (8.5) 22.3 (13.4) for 10 h prior to specimen collection. n ¼ number of subjects; SD ¼ standard deviation. Blood samples were kept on ice during transport to the laboratory (within 2 days). At the laboratory, samples were aliquoted and frozen at À201C until analysis was carried out. of chromium to the diluted samples were highlylinear, at Trace metal analysis was conducted within 1 month of blood least up to 10 mg/l. The slope of the standard addition collection. method for the five-fold diluted whole blood was 0.0153, which was veryclose to that of the aqueous standard. It Chemical Analysis indicated that the standard addition method is not required A method for the graphite furnace atomic absorption for the determination of chromium for the diluted whole spectroscopyanalysis of diluted human whole blood (Huang blood when using 5 mg Mg(NO3)2 as the matrix modifier. et al., 2000) was modified as detailed in Table 2. Chromium In this study, chromium standard solutions were prepared analyses of the 641 whole blood samples were performed using bydiluting the 1-g/l chromic acid stock solution in a PerkinElmer SIMAA 6000 Graphite Furnace Atomic deionized water to give final concentrations of 0, 2, 4, 6, 8, Absorption Spectrometer (GFAAS) with transverselyheated and 10 mg/l. graphite furnace and Zeeman background correction (Perki- Table 3 summarizes the performance of the analytical nElmer, Bodenseewerk, U¨ berlingen, Germany). Automated procedure in terms of recovery, precision, and method dilutions and injections were made with a PerkinElmer AS-90 detection limit (MDL) (see Appendix A). autosampler. The spectrometer and autosampler were con- trolled byAA Winlab software (PerkinElmer). QualityControl Our experiments showed that the calibration curve In addition to the rigorous screening and clean sample obtained byadding the standard in the different amounts collection protocols that were observed, blood-based quality

140 Journal of Exposure Science and Environmental Epidemiology (2006) 16(2) Chromium biomonitoring in exposed populations Chang et al.

Ta bl e 2 . Operating conditions for chromium determination byGFAAS.

Wavelength 357.9 nm Slit width 0.7 nm Current 25 mA Sample volume injected 20 ml Lamp type Hollow cathode lamp Signal measurement Peak area Background Zeeman background correction Tube type Pyro/Platform Gas Argon and oxygen

Furnace program for chromium determination of whole blood (1 : 5)a

Step Temperature (1C) Ramp (s) Hold (s) Gas ﬂow (ml/min) Gas type

Drying 110 1 30 250 Argon 130 5 45 250 Argon Charring 700 5 10 50 Oxygen Cooling 20 5 5 250 Argon Ashing 1600 10 30 250 Argon Atomization 2400 0 5 0 Clean-out 2600 1 5 250 Argon 20 1 5 250 Argon 2500 1 5 250 Argon

a Diluent ¼ 0.2% HNO3, matrix modiﬁer ¼ 5 mg Mg(NO3)2.

Ta bl e 3 . Precision, accuracyand determination power of the GFAAS Table 4 . Criteria for the qualitycontrol analysesbyGFAAS. determination of chromium in whole blood samples. Qualitycontrol item Criteria Item C.V. (%) Calibration check Correlation coefﬁcient, r40.995 Precision Blank sample analysis Signal o 2LOD Series 1 (2.5-mg/l chromium standard solution used), 9.2 Duplicate sample analysis Relative percent difference, n ¼ 6 RPD(%) o25% Series 2 (5.0-mg/l chromium standard solution used), 2.1 Qualitysample analysis Relative error, RE(%): 725% n ¼ 6 Spiked sample analysis Recovery, R(%): 75–125% Series 3 (10.0-mg/l chromium standard solution used), 1.7 n ¼ 6 LOD ¼ limit of detection. Pooled, n ¼ 18 5.5 (except for calibration check) throughout the analysis. If any Concentration QC analysis did not meet the requirements (Table 4), analysis (mg/l) was halted and both the equipment and all aspects of the Accuracy analytical procedure were checked. Subsequently, the failed Certiﬁed value 11.0a QC analysis would be repeated once. If performance Measured value (mean7SD), n ¼ 9 11.671.0 standards were met, analysis continued. If standards were Method detection limit (MDL) 0.133 not achieved, all results that had been obtained between the n ¼ numbers of aliquots. successful and failed QC runs were discarded and reanalysis aSeronormt trace elements whole blood (lot no. Ml1256, Billingstad, of the affected samples was performed. In addition, during Norway). the period of sample analysis, control charts were established to monitor if anyof the determinations exceeded the QC limits based on the statistical probability. The run was control (QC) materials were included to evaluate the rejected when anyof the following conditions occurred. First, statistical control and accuracyof the experimental and one control observation exceeded the limit of mean73SD analytical procedures. Standard Reference Material (SRM), (standard deviation). Second, two consecutive different SeronormTM Trace Elements Whole Blood (Lot Ml1256, controls exceeded the limit of mean72 SD. Third, 10 Billingstad, Norway) was used for QC evaluation. QC consecutive control observations felled on the same side of analyses were performed once every 10 samples analyzed the mean.

Journal of Exposure Science and Environmental Epidemiology (2006) 16(2) 141 Chang et al. Chromium biomonitoring in exposed populations

Statistical Analyses 5.5% of measured values for B-Cr were below the MDL Data were calculated and analyzed by using EXCEL (0.133 mg/l). The estimated GM and the 95th percentile of B- (Microsofts EXCEL 2002, Redmond, WA, USA) and Cr were 0.357 and 0.790 mg/l, respectively. The median value SPSS software (Version 10.0;. SPSS, Chicago, IL, USA). and the GM were preferable because of the observed good Descriptive statistics including arithmetic mean (AM), SD, agreement between them. The B-Cr levels apparent in the geometric mean (GM), median and percentiles were calcu- blood samples collected from the high factory-density area lated. Concentrations below MDL were set to MDL/2 for were significantlyhigher than that of the control area. further data treatment. The normalityand log-normalityof However, B-Cr levels tended to increase with decreasing age. the results were evaluated byusing the Kolmorogov– Smirnov statistics (K–S test). One-wayanalysis of variance Calculated Reference Limits (ANOVA) was used to test the hypotheses that there were no The nonparametric 0.95 IFCC reference limit of B-Cr was differences in chromium levels between genders, age groups, estimatedtobeo0.905 mg/l (Table 5). and residential areas. A Po0.05 (two-tailed) was considered statisticallysignificant when testing the hypotheses. Comparison between Data from the Literature The reference limits were estimated according to the Compared with other general populations, the AM, GM and procedures recommended byInternational Federation of IFCC reference limit of B-Cr were approximatelytwo times Clinical Chemistry(IFCC) (Solberg, 1987). An established higher than the values of healthypopulation samples from protocol was used to reject outliers (Dixon, 1953). In the Italy, the United Kingdom and Spain (Table 6). latter test, if the observed value of D (the absolute difference between an extreme observation and the next largest Influencing Factors observation) was equal to or larger than one-third of the The influences of age group and residential area on B-Cr range R, the extreme observation would be deleted (Reed levels were highlysignificant ( Po0.05), but there was no et al., 1971). Nonparametric methods were preferred because significant effect from the gender (Figure 3). their estimation was robust to the distribution of the measurement results. No lower reference limit was reported if the 0.025 fractile fell below the MDL. Discussion

Biological monitoring is an efficient tool of monitoring Results individual exposure for manytoxic elements. Unfortunately, unequivocal reference levels cannot always be established. Chromium Levels in Whole Blood Reference levels maydiffer between countries and regions, None of the distributions of B-Cr were found to be normal or and exposure to the particular noxious compound can change log-normal (Figure 2). Therefore, it was unsuitable to present over time due to fluctuations in environmental exposure and the results as the mean7SD due to the skewed distribution of lifestyle changes. Thus, reference intervals must be estab- the value obtained. The skewed distribution of the measure- lished at regular intervals and with respect to the appropriate ment results suggested that nonparametric methods should be influence factors (Kristiansen et al., 1997). In addition, since adopted for the latter estimations of IFCC reference limits. manyfactors influence the biomonitoring results, character- Table 5 summarizes the concentrations of B-Cr of the ization of the reference population and strict quality Changhua population aged 35–64 years. Approximately assurance during sampling and chemical analysis are

300 120 S. D. = 0.39 S. D. = 0.29 Mean = 0.44 Mean = -0.45 100 N = 641 N = 641

200 80

60 Frequency 100 40

0 0 -1.3 0.2 1.7 3.2 4.7 6.2 -1.4 -0.9 -0.4 0.1 0.6 1.1 B-Cr (µg/l) Log (B-Cr) (µg/l) Figure 2. Percentage distribution of chromium in whole blood (B-Cr). None of the distribution of B-Cr was found to be normal or log-normal.

142 Journal of Exposure Science and Environmental Epidemiology (2006) 16(2) Chromium biomonitoring in exposed populations Chang et al.

Ta bl e 5 . Chromium in whole blood of the studypopulation from Changhua, Taiwan (35–64 years).

Nno MDL 10th 50th 90th 95th Range AM (SD) GM (GSD) CI GM IFCC reference limit

B-Cr (mg/l) 641 35 0.16 0.38 0.714 0.79 0.067–6.47 0.441 (0.392) 0.357 (1.930) 0.34–0.38 o0.905

Years of age* 35–44 years 205 5 0.198 0.415 0.76 0.905 0.067–6.47 0.498 (0.522) 0.405 (1.835) 0.37–0.44 o0.930 45–54 years 213 15 0.15 0.38 0.713 0.778 0.067–2.60 0.425 (0.312) 0.342 (1.997) 0.31–0.38 o0.808 55–64 years 223 15 0.157 0.35 0.694 0.75 0.067–3.71 0.403 (0.308) 0.330 (1.925) 0.30–0.36 o0.775

Area* High factory-density 521 8 0.17 0.405 0.73 0.815 0.067–6.47 0.470 (0.423) 0.379 (1.938) 0.36–0.40 o0.930 Control 120 27 0.15 0.29 0.525 0.564 0.067–0.97 0.316 (0.161) 0.274 (1.771) 0.25–0.30 o0.614

B-Cr ¼ chromium in whole blood; N ¼ sample size; MDL ¼ method detection limit (0.133 mg/l for B-Cr); n o MDL ¼ number of values below MDL (values below MDL were set to MDL/2); 10th, 50th, 90th, 95th ¼ percentiles; AM ¼ arithmetic mean; SD ¼ standard deviation; GM ¼ geometric mean; GSD ¼ geometric standard deviation; CI GM ¼ 95%-conﬁdence interval for GM. *: Signiﬁcant parameter (Po0.05). IFCC reference limit: nonparametric, 0.975 fractile is upper limit, 0.025 fractile is lower limit. If the lower limit is below MDL onlythe upper limit is presented.

Ta bl e 6 . Reference values for whole blood chromium concentration the easier specimen collection than blood (Lewis, 1997; obtained from the literature (mg/l) Rosenberg and Harrison, 1997; Huang et al., 1999). However, in recent studies whole blood and plasma were Literature Analytical AM GM Reference value found to be a more definitive gage of environmental method monitoring than urine chromium levels (Paustenbach et al., Italy( n ¼ 519) GFAAS 0.23 0.09–0.75 (range) 1997). In this work, we discussed blood chromium mainly (Minoia et al., 1990) based on three reasons: firstly, our discussion focused on general population’s exposure, not workers’. Theoretically, Spain (n ¼ 144) ICPMS 0.2 0.1–1.1 (range) for the general population investigated there should not be as (Llobet et al., 1998) high-level exposures as workplace. Also, easiness of specimen United Kingdom GFAAS 0.19 0.1–0.45 (IFCC) collection was not the onlyconsideration we concerned; (n ¼ 134) (White and secondly, blood analyses could provide information on not Sabbioni, 1998) onlyblood chromium concentration but also other biological parameters (Huang et al., 1999). These parameters might be Taiwan ( n ¼ 641) useful for other researches; and thirdly, blood chromium was (this study) GFAAS 0.44 0.36 o 0.905 (IFCC) o 0.133–6.47 (range) recommended as a biomarker byinstitutes of occupational health and safetyin some countries as well as urinary AM ¼ arithmetic mean; GM ¼ geometric mean; GFAAS ¼ graphite furnace atomic absorption spectrometer with Zeeman effect background chromium (Harada, 1990; DFG, 1992; HSE, 1992). correction; ICPMS ¼ inductivelycoupled plasma F mass spectrometry; The present studyis important and valuable because of the IFCC ¼ reference limit. large number of people sampled, their precise characterization and the strict QC measurements enacted. In Taiwan, there have been few comparable studies, except for those extremelyimportant (Minoia et al., 1990, 1992; Alessio, dealing with blood lead levels of Taiwanese adults and school 1993; Vesterberg et al., 1993; Christensen, 1995). A recent children (Liou et al., 1996a, b; Wang et al., 2002; Yang et al., review on trace element reference values in blood, serum, and 2002). urine in the Danish population (Poulsen et al., 1994) showed Despite the aforementioned care in sample collection and several shortcomings in some studies, mainlydue to a low analysis, the reference values of B-Cr estimated in this study number of reference persons or insufficient characterization are not representative of the entire Changhua population. of the reference group. This is because we selected subjects almost exclusivelyfrom Some researchers suggested that urinarychromium metal-contaminated and industrial areas, which mayhave determination was a suitable approach for the biomonitoring produced an overestimation of the reference values. Never- of workers exposed to high-level chromium and its theless, in the absence of alternative sources in Taiwan, it is compounds, mainlydue to the established relationships still valuable to compare these chromium reference values between chromium exposure and chromium excretion, and with the intervals determined in other studies to clarifythe

Journal of Exposure Science and Environmental Epidemiology (2006) 16(2) 143 Chang et al. Chromium biomonitoring in exposed populations

0.7

0.6

0.5

0.4

0.3

0.2 Whole blood Cr (µg/l)

0.1

0 Male Female 35-44 years 45-54 years 55-64 years High Control factory- area density area Influence factor Figure 3. Geometric means of chromium in whole blood (mg/l) according to gender, age group, and residential area. Mean values795% confidence interval for mean. Age group and residential area significant at Po0.05. Gender not significant (ANOVA after logarithmic transformation).

differences between different populations and to identify several European studies (Minoia et al., 1990; Llobet et al., trends in the increase or diminution of environmental 1998; White and Sabbioni, 1998), as listed in Table 6. Even exposure to chromium of toxicologicallysignificant levels of these comparisons are limited, however, given the variations chromium. in populations, analytical methods and data treatment. Regarding chromium measurement results below detection Presently, a statistically significant correlation was found limit, manyresearchers used half of the detection limit to between age group and B-Cr concentration, with an age- replace measurements below the detection limit (Christensen related decrease evident. Until further information is et al., 1993; Llobet et al., 1998; Paschal et al., 1998; Seifert available on occupational historyor exposure routes, the et al., 2000). In Paschal’s study, calculation of the medians basis of this decreasing trend remains unexplained. In and distribution statistics using DL/2 (half of the detection addition, statistical analysis showed B-Cr levels to be limit) and DL/(21/2) as substitute data did not result in significantlyhigher in the high factory-densityareas than in estimates that were statisticallysignificant different, so the the control areas. This observation was expected and seems substitution of DL/2 was used for all ‘‘missing (less than reasonable. However, further environmental risk assessments DL)’’ data. Regarding the other researchers, White and are necessaryto yield the scientific evidence linking the Sabbioni considered that if a large proportion of samples factories to the observed environmental chromium levels. gave values below the MDL, the median value could be In conclusion, this studyprovides valuable data for reported (White and Sabbioni, 1998). In addition, Ba´ ra´ ny environmentallyrelated health monitoring and reporting. et al. (2002) did not reject or change anyresults below the The data serves as a basis to establish chromium reference detection limit to avoid distorting the distributions, mean or values to characterize the population’s internal exposure to median values. In the present study, we did not exclude data environmental chromium contamination. It is anticipated less than DL to avoid overestimating the chromium levels of that the present results will lead to a revision of these the Changhua population but used MDL/2 to replace reference values and to the creation of companion reference measurements below the MDL and made further statistical values for other substances of concern. Furthermore, despite analyses. However, only 5.4% of measurements in this work our observation of higher B-Cr levels in the Changhua were below the MDL (Table 5) and neither significantly residents as compared to the general population from western distorted the distributions and descriptive statistics nor countries, it is still unknown if the B-Cr levels will be higher changed the results of hypothesis testing. than those of residents residing in other areas of Taiwan. A direct comparison of the situation in Changhua, Taiwan Selection of other appropriate counties as control areas to be with that in other areas of Taiwan or in other Asian countries compared with Changhua mayhelp distinguish whether or is difficult due to the limited number of large-scale and not the higher B-Cr levels in Changhua is a local general population studies. With regard to the human phenomenon. Lastly, the question of whether the elevated biomonitoring data and the year of sampling, the results of chromium levels indeed pose a health risk warrants further this present studycan most fruitfullybe compared with investigation.

144 Journal of Exposure Science and Environmental Epidemiology (2006) 16(2) Chromium biomonitoring in exposed populations Chang et al.

Acknowledgements Harada A. Biological monitoring in medical surveillance in Japan. In: Fiserova- Bergerova V., Ogata M. (Eds.). Biological Monitoring of Exposure to The financial support (No. DOH91-TD-B04) of Department Industrial Chemicals. ACGIH, Cincinnati, 1990, pp. 45–54. HSE (Health and SafetyExecutive). Biological Monitoring for Chemical of Health of the Republic of China is gratefullyacknowl- Exposures in the Workplace. UK HSE Guidance Note EH 56. London 1992. edged. We thank the Changhua population for support and Huang Y.L., Chen C.Y., Sheu J.Y., Chuang I.C., Pan J.H., and Lin T.H. Lipid participation in this study. We thank Dr. S. H. Liou and peroxidation in workers exposed to hexavalent chromium. J Toxicol Environ Health 1999: 56: 235–247. Dennis P. H. Hsieh of the Division of Environmental Health Huang Y.L., Chuang I.C., Pan C.H., Hsiech C., Shi T.S., and Lin T.H. and Occupational Medicine, National Health Research Determination of chromium in whole blood and urine bygraphite furnace Institutes (Taiwan), for their valuable suggestions. We are AAS. Atom Spectrosc 2000: 21: 10–16. IARC (International Agencyfor Research on Cancer). IARC Monograph on the highlyindebted to the Changhua Christian Hospital, the Evaluation of Carcinogenic Risk to Humans: Chromium, Nickel and Welding. Changhua Public Health Bureau and the public health Vol. 49, IARC, Lyon, 1990, p. 677. nurses, and the Changhua Environmental Protection Bureau IDBMOEA (Industrial Development Bureau Ministryof Economic Affairs). Handbook of the Integrated Pollution Prevention Techniques for Metal for their support. We thank Y.C. Wang, C.Y. Peng, C.J. Li Surface Treatment Industries F Electroplating Industry, Industrial Develop- and H.Y. Yu for their valuable assistance in collecting the ment Bureau Ministryof Economic Affairs, Executive Yuan, Republic of samples, and to H.J. Wang for her technical assistance in China, Taipei, 2002. Kristiansen J., Christensen J.M., Iversen B.S., and Sabbioni E. Toxic trace element analytical work. reference levels in blood and urine: influence of gender and lifestyle factors. Sci Total Environ 1997: 204: 147–160. Disclaimer Kuo H.W., and Wu M.L. Effects of chromic acid exposure on immunological parameters among electroplating workers. Int Arch Occup Environ Health 2002: 75: 186–190. The scientific content of this manuscript has been reviewed Langardt S. One hundred years of chromium and cancer: a review of andapprovedforpublicationbytheDivisionofEnviron- epidemiological evidence and selected case reports. Am J Ind Med 1990: 17: mental Health and Occupational Medicine of the National 189–215. Lewis R. Metals. In: LaDou J. (Ed.). Occupational and Environmental Medicine. Health Research Institutes. Approval for publication does McGraw-Hill Companies, Inc., OH, USA, 1997, Chapter 27. not necessarilysignifythat the content reflects the view and Lin Y.P., Teng T.P., and Chang T.K. Multivariate analysis of soil heavy metal policies of the DEHOM/NHRI, or condemnation or pollution and landscape pattern in Changhua Countyin Taiwan. Landscape Urban Plan 2002: 62: 19–35. endorsement and recommendation for use on this issue Liou S.H., Wu T.N., Chiang H.C., Yang T., Yang G.Y., Wu Y.Q., Lai J.S., Ho presented. S.T., Guo Y.L., Ko Y.C., Ko K.N., and Chang P.Y. Three–year survey of blood lead levels in 8828 Taiwanese adults. Int Arch Occup Environ Health 1996a: 68: 80–87. Liou S.H., Wu T.N., Chiang H.C., Yang G.Y., Yang T., Wu Y.Q., Lai J.S., Ho References S.T., Lee C.C., Ko Y.C., Ko K.N., and Chang P.Y. Blood lead levels in Taiwanese adults: distribution and influencing factors. Sci Total Environ 1996b: Aiyar J., Berkovits J., Floyd R.A., and Wetterhahn K.E. Reaction of chromium 180: 211–219. (VI) with glutathione or with hydrogen peroxide: identification of reactive Llobet J.M., Granero S., Torres A., Schuhmacher M., and Domingo J.L. intermediates and their role in chromium (VI)-induced DNA damage. Environ Biological monitoring of environmental pollution and human exposure to Health Perspect 1991: 92: 53–62. metals in Tarragona, Spain. Trace Elem Electrol 1998: 15: 76–80. Alessio L. Reference value for the studyof low doses of metals. Int Arch Occup Minoia C., Sabbioni E., Apostoli P., Pietra R., Pozzoli L., Gallorini M., Environ Health 1993: 65: S23–S27. Nocolaou G., Alessio L., and Capodaglio E. Trace element reference values in Ba´ ra´ nyE., Bergdahl I.A., BrattebyL.E., Lundh T., Samuelson G., Schutz A., tissues from inhabitants of the European communityI. A studyof 46 elements Skerfving S., and Oskarsson A. Trace element levels in whole blood and serum in urine, blood and serum of Italian subjects. Sci Total Environ 1990: 95: from Swedish adolescents. Sci Total Environ 2002: 286: 129–141. 89–105. Baruthio F. Toxic effects of chromium and its compounds. Biol Trace Elem Res Minoia C., Pietra R., Sabbioni E., Ronchi A., Gatti A., Cavallieri A., and 1992: 32: 145–153. Manzo L. Trace element reference values in tissues from inhabitants of the Biedermann K.A., and Landolph J.R. Role of valence state and solubilityof European community. III. The control of preanalytical factors in the chromium compounds on induction of cytotoxicity, mutagenesis and biomonitoring of trace elements in biological fluids. Sci Total Environ 1992: anchorage independence in diploid human fibroblasts. Cancer Res 1990: 50: 120: 63–79. 7835–7842. Paschal D.C., Ting B.G., Morrow J.C., Pirkle J.L., Jackson R.J., Sampson E.J., Chi X.Z. Microelement and Human Health. Chemical IndustryPress, Beijing, 1997 Miller D.T., and Caldwell K.L. Trace metals in urine of United States p. 87. residents: reference range concentrations. Environ Res 1998: 76: 53–59. Christensen J.M. Human exposure to toxic metals: factors influencing interpreta- Paustenbach D.J., Panko J.M., Fredrick M.M., FinleyB.L., and Proctor D.M. tion of biomonitoring results. Sci Total Environ 1995: 166: 89–135. Urinarychromium as a biological marker of environmental exposures: what Christensen J.M., Holst E., Bonde J.P., and Knudsen L. Determination of are the limitations? Regul Toxicol Pharmacol 1997: 26: S23–S34. chromium in blood and serum. Evaluation of qualitycontrol procedures and Poulsen O.M., Christensen J.M., Sabbioni E., and Van der Venne M.T. Trace estimation of reference values in Danish subjects. Sci Total Environ 1993: 132: element reference values in tissues from inhabitants of the European 11–25. Community: V. Review of trace elements in blood, serum and urine and Cornelis R., Heinzow B., Herber R.F., Christensen J.M., Poulsen O.M., critical evaluation of reference values for the Danish population. Sci Total Sabbioni E., Templeton D.M., Thomassen Y., Vahter M., and Vesterberg Environ 1994: 141: 197–215. O. Sample collection guidelines for trace elements in blood and urine. Reed A.H., HenryR.J., and Mason W.B. Influence of statistical method used on IUPAC Commission of Toxicology. J Trace Elem Med Biol 1996: 10: the resulting estimate of normal range. Clin Chem 1971: 17: 275–284. 103–127. ROCEPA. Surveyof heavymetals in the soil samples. In: Statistics Office of DFG (Deutsche Forschungsgemeinschaft). MAK- and BAT-Values. Report No. Environmental Protection Administration (Ed.). Yearbook of Environmental 28.DFG,VCH,Nurnberg,1992. Statistics Taiwan Area, the Republic of China. Environmental Protection Dixon W.J. Processing data for outliers. Biometrics 1953: 9: 74–89. Administration of the Republic of China, Taipei, 1989.

Journal of Exposure Science and Environmental Epidemiology (2006) 16(2) 145 Chang et al. Chromium biomonitoring in exposed populations

ROCEPA. Surveyof heavymetals in the soil samples. In: Statistic Office of Appendix A Environmental Protection Administration (Ed.). Yearbook of Environmental Statistics Taiwan Area, the Republic of China. Environmental Protection Administration of the Republic of China, Taipei, 1992. 1. According to our sample collection protocol, collect whole ROCEPA (Environmental Protection Administration of the Republic of China). blood from 5 to 10 nonexposed healthypersons and mix Farmland Soil HeavyMetal Investigation and Contaminated Site Control the bloods collected. Plan, Environmental Protection Administration of the Republic of China (Project No.: EPA-90-GA13-03-90A285) Taipei, 2002. 2. Take seven aliquots from the above whole blood mixture. Rosenberg J., and Harrison R.J. Biological Monitoring. In: LaDou J. (Ed.). Add chromium stock solution into each aliquot to prepare Occupational and Environmental Medicine. McGraw-Hill Companies, Inc., seven samples at the same concentration near one-fifth of OH, USA, 1997, Chapter 38. Seifert B., Becker K., Helm D., Krause C., Schulz C., and Seiwert M. The the lowest-concentration calibration standard. German environmental survey1990/1992 (GerES II): reference concentrations 3. Pretreat the seven samples and analyze them according to of selected environmental pollutants in blood, urine, hair, house dust, drinking the standard operation procedure, which is the same as water and indoor air. J Expos Anal Environ Epidemiol 2000: 10: 552–565. Snow E. Metal carcinogenesis: mechanistic implications. Pharamacol Ther 1992: used for subject’s specimens. 53: 31–65. 4. Calculate the standard deviation (Sa)oftheseven Solberg H.E. Approved recommendation on the theoryof reference values Part 5: readings. Statistical treatment of collected reference values: determination of reference limits. J Clin Chem Clin Biochem 1987: 25: 645–656. 5. Take another seven aliquots from the above whole blood Sorahan T., Burges D.C., Hamilton L., and Harrington J.M. Lung cancer mixture. Add chromium stock solution into each aliquot mortalityin nickel/chromium platers, 1946-95. Occup Environ Med 1998: 55: to prepare seven samples at the same concentration of half 236–242. Standeven A.M., and Wetterhahn K.E. Possible role of glutathione in chromium of the last additive concentration. Repeat Steps 3 and 4 (VI) metabolism and toxicityin rats. Pharmacol Toxicol 1991: 68: 469–476. and calculate another SD (Sb). Versieck J., Barbier F., Cornelis R., and Holst J. Sample contamination as a source of 6. Use F-test to evaluate the two SDs (let S and S be the error in trace element analysis of biological samples. Talanta 1982: 29: 973–984. 1 2 Versieck J., and Cornelis R. Trace Elements in Human Plasma or Serum, CRC larger and smaller ones between Sa and Sb). If Press, Orlando, FL, USA, 1989. 2 2 Fð¼ S1=S2Þ is less than 3.05, then use the following Vesterberg O., Alessio L., and Brune D. International project for producing equation to calculate the pooled standard deviation, reference values for concentrations of trace elements in human blood and urine S . F TRACY. Scand J Work Environ Health 1993: 19: 19–26. pooled sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Wang C.L., Chuang H.Y., Ho C.K., Yang C.Y., Tsai J.L., Wu T.S., and Wu T.N. Relationship between blood lead concentrations and learning achievement 6S2 þ 6S2 S 1 1 among primaryschool children in Taiwan. Environ Res 2002: 89: 12–18. pooled 12 Wetterhahn K.E., and Hamilton J.W. Molecular basis of hexavalent chromium carcinogenicityeffect on gene expression. Sci Total Environ 1989: 86: 113–129. White M.A., and Sabbioni E. Trace element reference values in tissues from inhabitants of the European Union. X. A studyof 13 elements in blood and 7. If in Step 6 F is not less than 3.05, then repeat Steps 5 and urine of a United Kingdom population. Sci Total Environ 1998: 216: 253–270. 6untilF is less than 3.05. Yang T., Wu T.N., Hsu S.W., Lai C.H., Ko K.N., and Liou S.H. Blood lead 8. Method detection limit (MDL) ¼ 2.681 Spooled/a.‘‘a’’ is levels of primary-school children in Penghu County, Taiwan: distribution and influencing factors. Int Arch Occup Environ Health 2002: 75: 528–534. the slope of the calibration curve. Zhitkovich A. Chromium: exposure, toxicity, and biomonitoring approaches. In: Wilson S.H., Suk W.A. (Eds.). Biomarkers of Environmentally Associated Disease: Technologies, Concepts, and Perspectives. Lewis Publishers, Boca Raton, 2002, Chapter 19.

146 Journal of Exposure Science and Environmental Epidemiology (2006) 16(2)