Dioxin exposure in a residential community

Transcription

1 Journal of Exposure Analysis and Environmental Epidemiology (2001) 11, # 2001 Nature Publishing Group All rights reserved /01/$ KENNETH G. ORLOFF, a,1 DAVID HEWITT, a,1 SUSAN METCALF, a,1 STEVEN KATHMAN, a,1 MICHAEL LEWIN a,1 AND WAYMAN TURNER b,2 a Agency for Toxic Substances and Disease Registry, 1600 Clifton Road, MS-E32, Atlanta, Georgia b Centers for Disease Control and Prevention, 1600 Clifton Road, MS-F17, Atlanta, Georgia The Agency for Toxic Substances and Disease Registry ( ATSDR) conducted biological testing to assess dioxin exposure in residents of a community who lived in an area with heavy chemical industry. Dioxin concentrations were measured in blood serum samples from 28 adult residents of the community. Fourteen of those tested had blood dioxin concentrations that exceeded the 95th percentile prediction level of an age- matched comparison population. Specific congener analyses indicated that the elevated dioxin concentrations were primarily due to high concentrations of 2,3,7,8 tetrachlorodibenzo- p - dioxin ( TCDD), 1,2,3,7,8 pentachlorodibenzo- p - dioxin ( PeCDD), and hexachlorodibenzo - p- dioxins ( HxCDs). Principal components analysis ( PCA) indicated that the profiles of dioxin congeners were different in people with elevated blood dioxin concentrations compared to those with background concentrations. Elevated blood dioxin concentrations were detected only in older members of the population, which suggests that dioxin exposures were higher in the past. The sources of the dioxin exposure have not been identified. Journal of Exposure Analysis and Environmental Epidemiology ( 2001) 11, Keywords: blood serum, dioxin, dioxin congeners, principal components analysis. Introduction Residents of a community in Calcasieu Parish, Louisiana, expressed concern to health officials over possible exposures to hazardous substances from nearby chemical industries. The community is located across the road from a large vinyl chloride monomer ( VCM) plant. In addition, several other large chemical manufacturing plants that produce petroleum- based chemicals, chlorinated hydrocarbon solvents, and other organic chemicals are located in Calcasieu Parish. Chemical wastes from some of these operations are burned in permitted hazardous waste incinerators in the area. Although dioxins are not produced commercially, they are sometimes formed as by- products of chemical processes. Small quantities of dioxins can be formed during VCM production, although the contribution of VCM production to the overall environmental burden of dioxin is uncertain (EPA, 1994a). Dioxins can also be formed during the incineration of chlorine- containing wastes, during petroleum refining catalyst regeneration, and during 1. Address all correspondence to: Dr. Kenneth G. Orloff, Express mail address: 3119 Executive Park, Atlanta, GA 30329; U.S. Mail address: 1600 Clifton Road, MS-E32, Atlanta, GA Tel.: Fax: keo1@cdc.gov 1 Tel.: Fax: Tel.: Fax: Received 19 June the production of some halogenated organic chemicals (EPA, 1994a). In Calcasieu Parish in 1999, there were more than 300 cases of accidental chemical releases due to industrial spills, flares, fires, and leaks (CPOEP, 1999). Residents of the area have complained of irritating or noxious odors and numerous health symptoms that they attribute to flares and other sources of chemical releases. In response to citizen requests, the Agency for Toxic Substances and Disease Registry ( ATSDR) conducted an exposure investigation ( EI). Blood samples were collected from residents and analyzed for chlorinated dibenzodioxins ( CDDs), chlorinated dibenzofurans ( CDFs), and coplanar polychlorinated biphenyls ( PCBs). Materials and methods Target Population The target community occupies about 1/ 4 square mile, and the residents are predominantly black. ATSDR staff met with residents of the community to recruit participants who lived within four blocks of the VCM plant. Participation was restricted to participants who were 18 years of age or older and who had lived in the designated area for at least 5 years. Preference was given to older people who had been long- term residents in the neighborhood. Prior to biological testing, each participant signed an informed consent form.

2 Orloff et al. Biological Sampling and Analyses Participants were told to avoid eating a fatty meal prior to donating the blood sample. A licensed phlebotomist, under the supervision of a physician, collected a 70-ml blood sample by venipuncture from 28 participants. Blood samples were collected in 10-ml glass Vacutainer 1 tubes using a standardized protocol ( Patterson et al., 1991). Following collection, the blood samples were allowed to clot for 1 to 2 h at room temperature to facilitate clot retraction. The samples were then stored on ice and handdelivered the following day to the laboratory at the National Center for Environmental Health (NCEH) at the Centers for Disease Control and Prevention in Atlanta, Georgia, for analyses. In the laboratory, the Vacutainer 1 tubes were centrifuged at 1000g for 10 min, and the supernatant blood serum was removed for analysis. The entire blood serum sample from one person, consisting of 20 to 25 ml of serum, was analyzed for CDDs, CDFs, and coplanar PCBs using high-resolution gas chromatography/isotope dilution high resolution mass spectroscopy ( Patterson et al., 1991; Turner et al., 1997). The serum samples were spiked with a mixture of 13 C-labeled CDDs and CDFs, and the analytes were isolated using a C 18 solid-phase extraction procedure followed by a multicolumn automated cleanup and enrichment procedure. The analytes were chromatographed on a DB-5 ms capillary column (30 m0.25 mm0.25 m film thickness) using a Hewlett- Packard 6890 gas chromatograph and quantified by ID- HRMS using selected ion monitoring at 10,000 resolving power using a Micromass AutoSpec ULTIMA mass spectrometer in the EI mode. The concentration of each analyte was calculated from an individual standard linear calibration. Each analytical run was conducted blinded and consisted of three unknown serum samples, a method blank, and a quality control sample. In these analyses, concentrations of octachlorodibenzofuran ( OCDF) were not reported because of analytical interference from background contamination in the laboratory. The concentrations of 3,3 0,4,4 0 -PCB were reported as not detected, since this compound was not detected above background contamination levels in blank samples. The blood serum samples were also analyzed for total lipid content using a enzymatic summation method ( Akins et al., 1989), so the dioxin results could be expressed as a serum lipid concentration. Data Analyses A multivariate linear regression approach was used to model concentrations of dioxin toxicity equivalents ( TEQs) in blood serum. The independent variables were age and length- of- residency, modeled as linear variables. A log transformation of the dependent variable ( natural log of serum dioxin) was used to better satisfy the assumptions, and to allow the parameters to be estimated by ordinary least squares methods. The regression assumptions were verified prior to the use of statistical tests and prior to the construction of prediction intervals. Prediction intervals were calculated using standard methods ( Neter et al., 1990). Predicted values and prediction intervals for the original nonlinear model ( without a log transformation) were determined by taking the exponential of the corresponding parameters in the linear model ( using the log transformation). To compare the profiles of dioxin congeners in the EI participants, principal components analysis ( PCA) was carried out using SAS V.8 statistical software ( Rencher, 1995). The goal of PCA is to linearly transform possibly correlated variables ( dioxin congeners) into a smaller number of uncorrelated variables called principal components (PCs). The PC scores are then plotted to graphically reveal patterns in a complex data set that are not readily seen by visual inspection. Blood dioxin analytical data were first normalized by expressing the concentration of the individual dioxin congeners in a sample as a percentage of the combined sum of all CDD and CDF congeners. The factor loadings and scores were rotated according to the varimax rotation to assist in the interpretation of the factors ( Rencher, 1995). Results Biological Sampling In order to assess the health impact of a mixture of dioxinlike compounds, the toxicity of individual congeners of CDDs and CDFs was converted to 2,3,7,8- tetrachlorodibenzo-p-dioxin (TCDD) TEQs using toxicity equivalent factors (TEFs). The toxicity of the mixture is assumed to be equal to the sum of the individual components. In this investigation, the blood dioxin TEQ concentrations were calculated using the TEFs recommended by the World Health Organization ( Van den Berg et al., 1998). Blood serum samples from 28 adults were analyzed for dioxin- like compounds. The ages of the participants ranged from 21 to 83 years old, and the average age was 53. The test population contained 16 females and 12 males; the racial composition was 27 black and 1 white. Dioxin TEQs in the blood serum samples ranged from 3.8 to 186 pg/g lipid or parts per trillion (ppt). The median dioxin TEQ concentration was 54.8 ppt, and the mean dioxin TEQ concentration was 68.3 ppt. Table 1 contains the median, mean, and 95th percentile concentrations of dioxin- like compounds for participants in the investigation and a comparison population. The derivation of background dioxin levels in the comparison population is summarized below. Journal of Exposure Analysis and Environmental Epidemiology (2001) 11(5) 353

3 Orloff et al. Table 1. Concentrations of dioxin- like compounds ( picograms per gram lipid or parts per trillion) in blood serum samples from EI participants. Analyte Residents, median Comparison population, median Residents, mean Comparison population, mean Residents, 95th percentile Comparison population, 95th percentile 2378D D D/ D D D D OCDD F F F F/ F F F F F OCDF NR 11.0 NR 7.7 NR PCB ND 38.9 ND 34.7 ND PCB PCB PCB TCDD TEQs a NR, not reported due to analytical interference. ND, not detected at concentrations above background contamination levels. a Calculated using TEFs recommended by the World Health Organization ( Van den Berg et al., 1998). None of the study participants reported working for a chemical manufacturer except for one individual who had worked for a chemical company 4 years prior to testing. The blood dioxin TEQ concentration in this individual was 23 ppt, which is close to the mean of the comparison population. Dioxin Comparison Range During the past 20 years, environmental releases of dioxin- like compounds have decreased because of the switch to unleaded automobile fuels, process changes at pulp and paper mills, improved emission controls on incinerators, and reductions in the manufacture and use of chlorinated phenolic intermediates and products. This has been accompanied by decreases in body burdens of dioxin- like compounds ( EPA, 1994b; Furst et al., 1994; Papke, 1997). The most extensive national study of body burdens of dioxin in United States residents was the National Human Adipose Tissue Survey (EPA, 1994b). Based on an analyses of adipose tissue samples, which were collected in 1987, it was concluded that dioxin concentrations increased with age (EPA, 1994b). It was further concluded that dioxin concentrations appeared to be fairly uniform geographically, and there were no significant race or sex differences (EPA, 1994b). To date, no large- scale study of the concentrations of dioxin- like compounds in blood samples from a statistically based sample of the United States population has been conducted. Therefore, researchers have often cited reference ranges tabulated from studies conducted in the 1980s (DeVito et al., 1995; Needham et al., 1996). Since dioxin concentrations in the environment and in human populations have been decreasing ( Alcock and Jones, 1996), the use of such studies to derive comparison ranges for present day populations is not appropriate. Therefore, ATSDR, in collaboration with NCEH, developed comparison levels for blood dioxin concentrations based on the results of several recent studies conducted in the United States ( R. Canady, unpublished data, 2000). These comparison concentrations were derived from data obtained from five studies conducted by NCEH in five states ( Arkansas, Missouri, Wisconsin, North Carolina, and Oregon) between 1995 and The studies included about 400 individuals with an average age of 45 years old who were not believed to have had any unusual dioxin exposures. The gender distribution of the participants was 54% female and 46% male, and they were predominantly 354 Journal of Exposure Analysis and Environmental Epidemiology (2001) 11(5)

4 Orloff et al. Table 2. Loading coefficients for the first two principal components ( PCs) of congener relative abundances. Congener PC1 PC2 2378D D / 678D D D D OCDD F F F / 678F F F F F white. The mean dioxin TEQ concentration calculated for this population was 21 ppt (Table 1). This TEQ total consists of 18.7 ppt TEQs from dioxins and 2.3 ppt TEQs from four coplanar PCBs. This dioxin TEQ value is similar to the EPA s estimate of the current body burden of dioxins (16 ppt), which is based on current estimates of dioxin intake and steady- state pharmacokinetic modeling ( EPA, 2000). Principal Components Analyses To further examine the pattern of dioxin congeners in the EI participants, we conducted a PCA of the normalized serum dioxin concentrations. The relative abundance of each of the 15 most abundant congeners, expressed as a percentage of the sum of all congeners, was analyzed using SAS V.8. The PCs were calculated from the covariance matrix. Table 2 lists the loading coefficients for the first two PCs. PC1 accounted for 36% of the variance, and PC2 accounted for 26% of the variance; therefore, together, the two PCs accounted for 62% of the variance. Figure 1. Plot of the first two principal components ( PCs) for dioxin congener data from blood serum samples from 28 EI participants. Journal of Exposure Analysis and Environmental Epidemiology (2001) 11(5) 355

5 Orloff et al. A plot of the first two PCs for the EI participants is presented in Figure 1. This plot of the PCs derived from the blood dioxin data graphically depicts the similarities and differences in the dioxin congener profiles among the participants. The significance of the dioxin congener profiles is discussed below. Discussion As indicated in Table 1, blood serum dioxin TEQ concentrations were elevated in the EI participants. The median (54.8 ppt) and mean (68.3 ppt) concentrations of dioxin TEQs in the participants exceeded the 95th percentile concentration ( 37.5 ppt) of the comparison population. Furthermore, the concentrations of several, but not all, of the individual CDDs and CDFs in the EI population were elevated compared to the comparison population. The blood serum concentrations of 1,2,3,7,8 pentachlorodibenzo- p- dioxin (PeCDD) in the residents were particularly elevated. The mean concentration of PeCDD in the participants (28.8 ppt) was more than 3-fold higher than the 95th percentile concentration ( 9.1 ppt) of the comparison population. The mean concentrations of TCDD and the hexachlorodibenzop- dioxin ( HxCD) congeners in the participants also exceeded the 95th percentile concentrations of the comparison population. These results indicate that the concentrations of the total dioxin TEQs and several individual dioxin congeners in the participants are significantly elevated compared to the comparison population. Because dioxins are resistant to metabolism in the body, they bioaccumulate in the body, and blood dioxin concentrations tend to increase with age. In selecting the EI participants, we gave preference to older individuals who were long- term residents of the neighborhood, because they were at greater risk for cumulative exposures to dioxin. The Figure 2. Concentration of dioxin TEQs in blood serum ( picograms per gram lipid ) vs. age. The upper and lower 95th percentile prediction levels are indicated. 356 Journal of Exposure Analysis and Environmental Epidemiology (2001) 11(5)

6 Orloff et al. age of the residents who participated in blood testing ranged from 20 to 83, and the average age was 53. Most of the participants were long- term residents, and the average length of residency in the neighborhood was 32 years. To assess the relationship between age and blood dioxin concentration, we plotted blood dioxin TEQ concentrations vs. age for the comparison population and the EI participants. Figure 2 shows that the blood dioxin concentrations increased in the comparison population as a function of age. Blood dioxin concentrations in the EI participants also increased with age, but the age- related increase was much greater in the participants than in the comparison population. As shown in Figure 2, the elevations in blood dioxin concentrations were most pronounced in the older participants. In order to determine if an individual blood dioxin concentration was significantly elevated, the 95th percentile upper and lower prediction intervals for the comparison population as a function of age were determined. As indicated in Figure 2, 14 of the 28 EI participants had blood dioxin concentrations that exceeded the age- matched 95th percentile upper prediction level. All of the individuals with elevated blood dioxin concentrations were aged 47 or older. Regression analyses were conducted to evaluate the relationships between log transformed blood dioxin concentrations, age, and length of residency in the neighborhood. Age was strongly correlated with blood dioxin concentration (R 2 =0.639, p<0.001). Length of residency in the neighborhood was also correlated with blood dioxin concentration (R 2 =0.252, p<0.0065), but the correlation was weaker than with age. Using multiple regression analyses, it was determined that if age were controlled for, the correlation between blood dioxin concentration and length of residency was not significant. The reason for the selective elevation in blood dioxin concentrations in the older EI participants is not known. One possibility is that older residents have elevated dioxin concentrations because of exposure to dioxins that occurred in the past. Dioxin- like compounds are resistant to metabolic degradation, so after being absorbed into the body, they persist for long periods of time. In occupationally exposed workers, the biological half- lives of dioxin congeners range from 3.0 to 19.6 years (Flesch-Janys et al., 1996). The half- life of PeCDD, which is a major contributor to the TEQ total in the participants, was reported to be 15.7 years (Flesch-Janys et al., 1996). Therefore, because of the long biological half- lives of dioxin congeners, the elevated blood dioxin concentrations in the participants could be the result of exposures that occurred many years ago. As indicated in Table 1, the mean concentrations of many of the dioxin- like compounds, particularly the CDDs, were elevated in the participants compared to the comparison population. However, for some congeners, particularly some of the CDFs, the mean concentrations in residents were similar to or less than those in the comparison population. Therefore, based on the collective data, the pattern of dioxin- like compounds in the EI participants appeared to differ from the comparison population. To further examine the profile of dioxin congeners in the individual participants, we conducted a PCA as depicted in Figure 1. As indicated, the participants with normal blood dioxin concentrations plotted in the lower left hand corner of the PC plot. ( Normal blood dioxin concentrations were defined as being within the 95th percentile prediction levels Figure 2.) The finding that the participants with normal dioxin concentrations plot close to each other on the PC plot indicates that they have similar profiles of dioxin congeners. By contrast, most of the individuals with elevated blood dioxin concentrations plotted in distinctly different areas of the PC plot. ( Elevated blood dioxin concentrations were defined as being above the 95th percentile prediction level Figure 2.) This indicates that their dioxin congener profiles were different from those of individuals with normal dioxin levels. Moreover, within the population of individuals with elevated blood dioxin concentrations, there were differences in congener profiles, indicating possible variations in exposure sources. The source of dioxin exposures in the residents is not known. Pentachlorodibenzodioxins are produced during combustion processes and have been identified in emissions from municipal waste incinerators, cigarette smoke, wood and coal combustion, waste oil furnaces, and fires of polyvinylchloride ( PVC) materials ( Cruczwa and Hites, 1986; HSDB, 2001). Hexachlorodibenzodioxins ( HxCDDs) have been identified in emissions from municipal incinerators and industrial waste incinerators ( Cruczwa and Hites, 1986). Known sources of TCDD include incineration of municipal refuse and certain chemical wastes, chlorine bleaching of wood pulp, and as a byproduct of certain chemical manufacturing processes. In the general population, low- level dioxin contamination of foodstuffs is the major contributor to the body burden of dioxins (EPA, 1994c). Most of the participants in this investigation reported eating locally caught fish and shellfish. The State of Louisiana issued an advisory in 1992 to limit eating of fish from Bayou d Inde in Calcasieu Parish because of contamination with polychlorinated biphenyls, hexachlorobenzene, and hexachlorobutadiene. However, to date, fish from the area have not been tested for dioxins. Many of the participants also reported eating wild game, or locally produced fruits, vegetables, or chicken eggs. However, variations in blood dioxin concentrations could not be explained by dietary habits. The finding that age- adjusted, blood dioxin concentrations were elevated only in older participants suggests that dioxin exposures were higher in the past. However, in Journal of Exposure Analysis and Environmental Epidemiology (2001) 11(5) 357

7 Orloff et al. order to be proactive and protective of public health, the U.S. Environmental Protection Agency, in conjunction with the Louisiana Department of Environmental Quality, is conducting further sampling and analyses to test for potential dioxin contamination in air, water, sediment, and biota from Calcasieu Parish. The results of those studies may help to determine if there are any current sources of exposure to dioxin in the area. Acknowledgments The authors thank Dr. Richard A. Canady for developing the dioxin comparison values cited in this report. The authors also thank the residents of the community for their generous assistance and participation in this investigation. References Akins J.R., Waldrep K., and Bernert J.T. Jr. The estimation of total serum lipids by a completely enzymatic summation method. Clin Chim Acta 1989: 184: Alcock R.E., and Jones K.C. Dioxins in the environment: a review of trend data. Environ Sci Technol 1996: 30: Calcasieu Parish Office of Emergency Preparedness: Emergency Release Notification Records, Cruczwa J., and Hites R. Airborne dioxins and dibenzofurans: sources and fates; Environ Sci Technol 1986: 20 ( 2 ): DeVito M.J., Birnbaum L.S., Farland W.H., and Gasiewicz T.Al. Comparisons of estimated human body burdens of dioxinlike chemicals and TCDD body burdens in experimentally exposed animals. Environ Health Perspect 1995: 103: U.S. Environmental Protection Agency. Estimating exposure to dioxinlike compounds: Volume II. Properties, sources, occurrence and background exposures. Washington, DC, EPA / 600 / 6-88/ 005Cb, June 1994a: 3: U.S. Environmental Protection Agency. Cited in U.S. Environmental Protection Agency. Estimating exposure to dioxin- like compounds: Volume II. Properties, sources, occurrence and background exposures. Washington, DC, EPA /600/6-88/005Cb, June 1994b: 5: U.S. Environmental Protection Agency. Estimating exposure to dioxinlike compounds: Volume II. Properties, sources, occurrence and background exposures. Washington, DC, EPA / 600 / 6-88/ 005Cb, June 1994c: 5: U.S. Environmental Protection Agency. Exposure and human health reassessment of 2,3,7,8- tetrachlorodibenzo- p- dioxin ( TCDD) and related compounds: Part III. Integrated summary and risk characterization for 2,3,7,8- tetrachlorodibenzo- p- dioxin ( TCDD ) and related compounds. Washington, DC, EPA / 600/ P- 00/ 001Ag, June 2000: Flesch-Janys D., Becher H., Gurn P., Jung D., Konietzko J., Manz A., and Papke O. Elimination of polychlorinated dibenzo- p- dioxins and dibenzofurans in occupationally exposed persons. J Toxicol Environ Health 1996: 47: Furst P., Furst C., and Wilmers K. Human milk as a bioindicator for body burden of PCDDs, PCDFs, organochlorine pesticides, and PCBs. Environ Health Perspect Suppl 1994: 102 ( suppl 1): Hazardous Substances Data Bank [ database online]. Bethesda, MD: National Library of Medicine, Needham L.L., Patterson D.G., Burse V.W., Pascal D.C., Turner W.E., and Hill R.H. Jr. Reference range data for assessing exposure to selected environmental toxicants. Toxicol Ind Health 1996: 12: Neter J., Wasserman W., and Kutner M. Applied Linear Statistical Models: Regression, Analysis of Variance, and Experimental Designs. 3rd ed. Irwin, Boston, Papke O. Cited in International Agency for Research on Cancer. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans: Polychlorinated Dibenzo- para- dioxins and Polychlorinated Dibenzofurans, 1997: 69: 114. Patterson D.G., Isaacs S.G., Alexander L.R., Turner W.E., Hampton L., Bernert J.T., and Needham L.L. Method 6: Determination of specific polychlorinated dibenzo- p- dioxins and dibenzofurans in blood and adipose tissue by isotope dilution high-resolution mass spectroscopy. In: Rappe C., Buser H.R. ( Eds. ), Environmental Carcinogens - Methods of Analysis and Exposure Measurement, Vol. 11. WHO, International Association for Research on Cancer, Lyon, France, 1991, pp Rencher A.C. Methods of Multivariate Analysis. Wiley, New York, Turner W., DiPietro E., Lapenza C., Green V., Gill J., and Patterson D.J. A fast universal automated cleanup system for the isotope-dilution highresolution mass spectrometric analysis of PCDDs, PCDFs, coplanar PCBs, and persistent pesticides from the same serum sample. Organohalogen Compd 1997: 31: Van den Berg M., Birnbaum L., Bosveld A., Brunstrom B., and Cook P. Toxic equivalency factors ( TEFs) for PCBs, PCDDs, PCDFs for humans and wildlife. Environ Health Perspect 1998: 106 ( 12): Journal of Exposure Analysis and Environmental Epidemiology (2001) 11(5)