Perfluorinated chemicals in relation to other persistent organic pollutants in human blood

Transcription

1 Chemosphere 64 (2006) Perfluorinated chemicals in relation to other persistent organic pollutants in human blood Anna Kärrman a, *, Bert van Bavel a, Ulf Järnberg b, Lennart Hardell a, Gunilla Lindström a a Man-Technology-Environment Research Centre, Department of Natural Sciences, Örebro University, SE Örebro, Sweden b Department of Applied Environmental Research, Stockholm University, SE Stockholm, Sweden Received 23 June 2005; received in revised form 9 November 2005; accepted 10 November 2005 Available online 3 January 2006 Abstract In order to evaluate blood levels of some perfluorinated chemicals (PFCs) and compare them to current levels of classical persistent organic pollutants (POPs) whole blood samples from Sweden were analyzed with respect to 12 PFCs, 37 polychlorinated biphenyls (PCBs), p,p 0 -dichlorodiphenyl-dichloroethylene (DDE), hexachlorobenzene (HCB), six chlordanes and three polybrominated diphenyl ethers (PBDEs). The median concentration, on whole blood basis, of the sum of PFCs was times higher compared to the sum of PCBs and p,p 0 -DDE, times higher than HCB, sum of chlordanes and sum of PBDEs. Estimations of the total body amount of PFCs and lipophilic POPs point at similar body burdens. While levels of for example PCBs and PBDEs are normalized to the lipid content of blood, there is no such general procedure for PFCs in blood. The distributions of a number of perfluorinated compounds between whole blood and plasma were therefore studied. Plasma concentrations were higher than whole blood concentrations for four perfluoroalkylated acids with plasma/whole blood ratios between 1.1 and 1.4, whereas the ratio for perflurooctanesulfonamide (PFOSA) was considerably lower (0.2). This suggests that the comparison of levels of PFCs determined in plasma with levels determined in whole blood should be made with caution. We also conclude that Swedish residents are exposed to a large number of PFCs to the same extent as in USA, Japan, Colombia and the few other countries from which data is available today. Ó 2005 Elsevier Ltd. All rights reserved. Keywords: Fluorochemicals; Human blood; PFOS; PFOA 1. Introduction Recent studies have shown the global presence of perfluorinated chemicals (PFCs). In the Nordic environment PFCs have been found for instance in sewage, sewage sludge, fish and guillemot eggs, as well as in fresh and sea water (Holmström et al., 2004; Kallenborn et al., 2004; Posner and Järnberg, 2004). PFCs are fully fluorinated man-made chemicals with properties that make them useful in different industrial and household applications. PFCs can be manufactured in 3 ways; electrochemical fluorination, telomerisation and oligomerisation (Kissa, 2001). Production takes place mainly in USA, Japan and Europe * Corresponding author. Tel.: ; fax: address: anna.karrman@nat.oru.se (A. Kärrman). (OECD, 2002; EPA, 2003). PFCs in the environment originate from production or from use and disposal of products containing perfluorinated compounds. No fluorochemicals are produced in Sweden. Major applications of PFCs in Sweden are impregnation of textiles, leather and cleaning aids, and in addition they are also used as surfactants in paint- and spray industry and for metal surface treatment (KEMI, 2004). Perfluorinated acids, including perfluorooctane sulfonate (PFOS) and perfluorooctanoic acid (PFOA), are heat resistant and also water and oil repellent. The environmental distribution and behavior of this type of chemicals are largely unknown and therefore exposure routes are difficult to assess. Various volatile fluorinated chemicals have been shown to degrade to more persistent non-volatile PFCs like PFOS and PFOA (Tomy et al., 2003; Dinglasan et al., 2004; Ellis et al., 2004) /$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved. doi: /j.chemosphere

2 A. Kärrman et al. / Chemosphere 64 (2006) Human exposure to perfluorinated compounds is not well understood. Potential routes of exposure may be through air, water, food and household dust where PFCs have been found (Moriwaki et al., 2003; Schultz et al., 2004; Stock et al., 2004; Tittlemier et al., 2005). PFOS is the most studied PFC in this respect. The most significant absorption route of PFOS for rodents is oral absorption, but inhalation uptake may also be possible (OECD, 2002). Dermal exposure contributes to a lesser extent. Prenatal exposure has been shown for rodents and humans (Thibodeaux et al., 2003; Inoue et al., 2004). The half-life of PFOS in humans has been estimated to about 9 years (3M, 2000). Elimination of PFOS in rodents takes place mainly through urine (OECD, 2002). The toxicological effects so far studied of PFOS and PFOA are described elsewhere (OECD, 2002; Kennedy et al., 2004). Due to their persistence, high levels in top predators, toxicity and the recently established presence in remote areas, PFOS and similar PFCs can be classified as persistent organic pollutants (POPs) and are of environmental and health concern (Martin et al., 2004; Tomy et al., 2004). Unlike polychlorinated and polybrominated organic compounds, PFOS and PFOA have been shown to distribute mainly to liver and plasma (OECD, 2002). In the plasma, association with albumin takes place (Han et al., 2003; Jones et al., 2003). While levels of for example PCBs and PBDEs are correlated to the lipid content of blood, there is no general normalization procedure for PFCs in blood. When PFC serum and whole blood levels are expressed on volume basis, comparison of levels between these matrices becomes questionable. Previous studies on human levels of PFCs have mostly been based on serum or plasma samples. Whenever whole blood has been analyzed, multiplication of the whole blood level by a factor of 2 has been the most used way in order to make the results comparable to serum levels (Kannan et al., 2004). In this study, PFC distribution between whole blood and plasma was studied in blood from five adults. The concentration ratio between plasma and whole blood was calculated for PFOS, PFOA, perfluorohexane sulfonate (PFHxS), perfluorononanoic acid (PFNA) and perfluorooctane sulfonamide (PFOSA). Individual whole blood samples (n = 66) were then used to determine levels of 12 PFCs together with 37 polychlorinated biphenyls (PCBs), p,p 0 -dichlorodiphenyl-dichloroethylene (DDE), hexachlorobenzene (HCB), six chlordanes and three polybrominated diphenyl ethers (PBDEs). This cross-sectional study included blood from 40 men and 26 women from Sweden. Correlation between levels of fluorinated, chlorinated and brominated organic compounds in whole blood was also explored. 2. Materials and methods 2.1. Sample collection For comparing PFC levels with traditional POPs, whole blood samples used as control samples in a testicular cancer study were analyzed (Hardell et al., 2003). Male blood donors were drawn from the Swedish population registry (Stockholm, Sweden) which is a national registry that covers the whole population, in order to match testicular cancer cases with respect to age. Their respective mothers were then identified through the same registry. There were originally 61 male samples and 45 female samples in the control group. The samples in the present study are control samples that were available for further analysis and consists of whole blood from men (sons) in the age (n = 40) and women (mothers), age (n = 26). Donors originated from ten counties in the southern half and one county in the northern half of Sweden. Collection of the samples was performed during the period Study subjects were instructed to only have a light meal before blood donation. Blood was collected in heparin vacuum tubes and transferred to glass containers before storage at 20 C. The study on plasma and whole blood distribution was performed on blood drawn in 2004 from five adult volunteers, two men (27 and 41 years old) and 3 women (24, 29 and 36 years old). Donation was performed in the morning before breakfast. Also for this purpose whole blood was taken in heparin vacuum containers. A whole blood aliquot was taken out of the container after donation, and the remaining sample was centrifuged at 1000 g for 10 min to produce the plasma sample. Both the whole blood and plasma samples were extracted and analyzed in duplicates or triplicates Analysis of PFCs The method used for the PFC blood analysis has been reported in detail elsewhere (Kärrman et al., 2005). In short, 0.75 ml sample was transferred to a polypropylene centrifuge tube and perfluoroheptanoic acid (PFHpA, 750 pg) was added as internal standard. Formic acid/water was added to the sample which then was sonicated and centrifuged ( g, 30 min). Extraction was performed on a C 18 SPE sorbent (BondElute HF. Varian, Harbor City CA, USA) using methanol as elution solvent. The methanol extract was filtered through a polypropylene filter into a polypropylene vial, prior to which a performance standard (7H-PFHpA, 100 pg) had been added. Analysis was performed with a LC MS system (HP 1100 LC/MSD, Waldbronn, Germany). An atmospheric electrospray interface operating in negative ion mode was used. Selected ion monitoring (SIM) was used for the detection of PFOS, PFOA, PFHxS, PFNA, PFOSA, perfluorobutane sulfonate (PFBuS), perfluorohexanoic acid (PFHxA), perfluorodecanoic acid (PFDA), perfluorodecanesulfonate (PFDS), perfluoroundecanoic acid (PFUnDA), perfluorododecanoic acid (PFDoDA) and perfluorotetradecanoic acid (PFTDA) Method performance and quality control for PFCs Internal standard quantification was performed with PFHpA as internal standard since no labelled candidate

3 1584 A. Kärrman et al. / Chemosphere 64 (2006) was commercially available at the time for this study. 1H,1H,2H,2H-perfluorooctanesulfonic acid (THPFOS) was also added (750 pg) as a second internal standard. Concentrations were calculated with both PFHpA and THPFOS and compared in case a sample should contain high levels of PFHpA. A number of samples in this study were analyzed without the addition of PFHpA and THP- FOS in order to evaluate natural occurring levels in human blood. No THPFOS could be detected. Minor concentrations of PFHpA were detected at levels around 5 pg which was found acceptable since it is less than 0.5% of the added amount IS. In the present study no indication of high levels of PFHpA was found. Quantification was performed using non-extracted standards in methanol and by relating the area of the analyte to the area of PFHpA. Linearity expressed as correlation coefficients determined by leastsquares linear regression analysis was >0.98. The recovery of the internal standard was calculated for each sample by adding 7H-PFHpA before injection and comparing the PFHpA/7H-PFHpA ratio with the same ration in the external standards. Recoveries of the internal standard for all 66 samples analyzed were within %. Specificity of the single quadrupole MS quantification was evaluated by MS/MS analysis using a triple quadrupole instrument (Micromass QuattroII, Altrincham, UK). Details about the instrument settings and quantification transitions are described elsewhere (Kärrman et al., 2005). The regression coefficient (r 2 ) for whole blood concentrations (range given in brackets) in 20 sample extracts analyzed with both single quadrupole MS and triple quadrupole MS was 0.88 for PFOS (5 44 ng/ml), 0.91 for PFOA ( ng/ml), 0.99 for PFHxS ( ng/ml), and 0.99 for PFOSA ( ng/ml). For each sample set (14 22 samples) extracted and analyzed, a quality control (QC) sample was included to verify extraction and instrument stability. The QC sample was prepared from a human whole blood sample. Various, known amounts of PFCs ( ng/ml) were added except for PFOS that was already present in the blood at a suitable concentration (8 ng/ml). The day-to-day variation during the analysis of the samples was <15 RSD% for PFHxA, PFHxS, PFOA, PFOS and RSD% for the remaining PFCs. The instrumental limit of detection (LOD) defined as the concentration needed to produce a signal to noise ratio of 3:1 was ng/ml for the 12 analytes except PFBuS (2 ng/ml). Stability or losses to the glass surface were evaluated by storing an authentic whole blood sample in a glass container at 20 C, at 20 C and through three freeze and thaw cycles. Coefficient of variation (CV) for PFOS, PFOA, PFHxS and PFNA in a sample stored at 20 C and analyzed three times during a 4 months period were 12%, 11%, 12% and 2%, respectively. Storage for nine days at 20 C between three analyses resulted in CV of 1%, 5%, 5% and 6% for PFOS, PFOA, PFHxS and PFNA. Compared to the method interday reproducibility (n = 13) for PFOS (10%), PFOA (6%), PFHxS (6%) and PFNA (12%) conclusions were drawn that concentrations of mentioned PFCs were unchanged when stored in glass containers for at least four months in 20 C and at least nine days in 20 C. Concentrations after three freeze and thaw cycles deviated from replicate determinations on a single occasion with 10% for PFHxS, 5% for PFOA, 3% for PFOS and 4% for PFNA. Further quality assurance was taken by successful participation in the 1st worldwide interlaboratory study on PFCs in a methanol study standard solution as well as a plasma and whole blood sample. Carry-over and potential instrument contamination were monitored by methanol injections during the sample sequence. In order to prevent carry-over contamination, the outside of the injector needle was cleaned in methanol between each injection. To monitor possible contamination during the sample pre-treatment a blank sample containing only water was processed in the same way as the blood samples. The PFCs included in this study were not detected in blank samples or methanol injections Analysis of polychlorinated and polybrominated compounds The analysis of the samples with respect to 37 polychlorinated biphenyls (PCBs, #28, #47, #52, #66, #74, #99, #101, #105, #110, #114, #118, #128/167, #138, #141, #153, #156, #157, #170/190, #172/192, #174, #177, #178, #180, #182/187, #183, #189, #194, #195, #196/ 203, #199, #200, #201, #202, #206, #207, #208, #209), p,p 0 -dichlorodiphenyl-dichloroethylene (DDE), hexachlorobenzene (HCB) and six chlordanes (cis-heptachlordane, cis-chlordane, oxychlordane, MC6, trans-nonachlordane and cis-nonachlordane) has been described elsewhere (Hardell et al., 2003). Approximately 20 ml of blood was used for analyses of the chlorinated and brominated compounds. The samples were fortified with 13 C-labelled internal standards. The blood lipids and analytes were removed from the sample by use of a Hydromatrix column (Varian, Palo Alto, CA, USA) and lipids were determined gravimetrically. Interferences and lipids were then removed by multilayer silica chromatography. The congener specific determinations were performed by GC MS in electron impact (EI) and selected ion monitoring (SIM) mode. The extraction procedure for the three polybrominated diphenyl ethers (PBDE) homologues (tetra-, penta- and hexabromodiphenyl ether) was conducted in the same way as the chlorinated compounds. Determination of PBDEs was by GC MS, negative chemical ionization (NCI) in SIM mode by monitoring m/z 79/ Quality control for polychlorinated and polybrominated compounds A total of C labelled chlorinated internal standards and in addition three 13 C labelled performance standards were used. For the brominated diphenylethers three 13 C labelled internal standards and one 13 C labelled PCB

4 A. Kärrman et al. / Chemosphere 64 (2006) performance standard was used. Recoveries of the internal standards were between 50% and 120%. One laboratory blank sample was included in every batch of ten samples processed. All blank levels were <10% of any analyte and the method detection limit was ng/g lipid. In those few cases when some of the congeners were under detection limit half the detection level was used in calculation of the sum levels. External quality control was assured by successful participation in interlaboratory studies coordinated by the Arctic Monitoring and Assessment Program (AMAP) and the International Union for Pure and Applied Chemistry (IUPAC). 3. Results 3.1. PFC distribution between plasma and whole blood Mean concentrations of detected PFCs in plasma and whole blood from five adult donors can be seen in Table 1 together with the variability of multiple determinations for each sample and the plasma/whole blood (P/WB) ratio. The overall mean and individual ratios between plasma and whole blood concentrations of five PFCs are illustrated in Fig. 1. The ratios for PFOS, PFOA and PFHxS are the mean obtained for five individuals. Due to low concentrations the ratios for PFNA and PFOSA are the mean obtained for four and three individuals, respectively. Mean P/WB ratios for PFOS, PFOA, PFHxS and PFNA were 1.2, 1.4, 1.2 and 1.1. This shows that the levels of detected PFCs in plasma are higher when compared to the levels in whole blood. The distribution behavior of PFOSA was significantly different from the other PFCs, with average P/WB ratio of Levels of PFCs in whole blood A total of 11 PFCs were detected in the Swedish human samples and among them PFOS was found at the highest level (geometric mean 16 ng/ml, Table 2). The second highest level was found for PFOSA (3.0 ng/ml) followed by PFOA (2.4 ng/ml) and PFHxS (1.5 ng/ml). PFNA, PFDA and PFUnDA were also frequently detected, but at concentrations near the detection limit ( ng/ml). PFHxA was found to be present above the detection limit in five samples ( ng/ml) and PFDS in only two samples (2.4 and 4.5 ng/ml). PFDoDA and PFTDA were detected in a few samples but the identity could not be verified with MS/MS due to higher detection limits in MS/MS. The only compound that was not detected (<2 ng/ml) in this study was PFBuS. Median and geometric mean were calculated for the compounds that were present in more than 50% of the blood samples. Concentrations lower than detection limits were replaced with half the detection limit. In Table 2 the result is also presented separately for men (median age 31) and women (median age 55). Any difference seen in Table 1 Mean concentrations (ng/ml) of PFCs in whole blood (WB) and plasma (P) samples collected in 2004 from five individuals (gender and age provided). The relative standard deviation (RSD%) for triplicate or duplicate samples is given together with the mean P/WB ratio for each compound and all individuals Woman (36 years) Man (27 years) Man (41 years) Woman (29 years) Woman (24 years) Mean P/WB ng/ml RSD% Ratio ng/ml RSD% Ratio ng/ml RSD% Ratio ng/ml RSD% Ratio ng/ml RSD% Ratio P/WB P/WB P/WB P/WB P/WB P/WB P/WB P/WB P/WB P/WB P/WB P/WB P/WB P/WB P/WB ratio PFHxS 2.4/2.2 8/ /2.6 15/ a /2.9 3/ a /2.0 5/ a /1.2 10/ PFOA 2.6/2.0 3/ /1.7 11/ a /3.8 8/ a /3.9 2/ a /1.2 39/ PFOS 17.3/14.2 6/ / / a /27.8 3/ a /15.0 2/ a /11.3 6/ PFNA 0.55/ / nd/0.26 na/22 na 0.69 a / / a /0.50 9/ a /0.25 7/ PFOSA 0.11/0.86 na/ nd/0.27 na/4 na 0.22 a /1.3 6/ nd/0.37 na/24 na 0.10 a /0.37 na/ nd = not detected. na = not applicable, e.g. only one or zero detectable concentrations. a Mean of duplicate samples.

5 1586 A. Kärrman et al. / Chemosphere 64 (2006) Ratio P/WB PFHxS PFOA PFOS PFNA PFOSA Fig. 1. Mean and individual ratios of plasma/whole blood (P/WB) concentrations for five PFCs in blood from five adults. Ratios are given for the PFCs detected in the samples. The ratio for each individual is the mean value of triplicate or duplicate sample determinations. concentrations in this study can not be directly related to gender or age factors. Frequency distributions of PFOS, PFOA, PFHxS and PFOSA concentrations in the blood samples are illustrated in Fig. 2. Levels of PFOS, PFOA and PFHxS in whole blood from women are consistent with a Gaussion distribution based on the Shapiro Wilk normality test. For the male donor samples only PFOS met such criteria Levels of polychlorinated and polybrominated compounds in whole blood Levels of polychlorinated and polybrominated compounds on whole blood weight basis for 66 individuals from Sweden are given in Table 3. Included are also levels on lipid weight basis for these compounds. Levels in whole blood were calculated from the lipid weight concentrations by using the actual lipid content and the sample volume of 20 ml. The lipid content ranged from 0.2% to 0.6%. The median PCB sum on whole blood basis was 1.5 ng/g, the sum of the six chlordanes was 0.08 ng/g, HCB and p,p 0 - DDE were 0.09 and 0.6 ng/g, respectively. The median of the three PBDEs was 0.06 ng/g. 4. Discussion 4.1. PFC distribution between whole blood and plasma The relatively poorly understood distribution behavior of different PFCs between human whole blood and plasma needs to be further elucidated. PFOS and PFOA supposedly bind to proteins (Han et al., 2003; Jones et al., 2003). PFOSA on the other hand can be expected to distribute differently in biological fluids since it is a neutral compound. Furthermore, PFCs with chains shorter or longer than 8 carbon atoms may also behave differently due to dissimilar non-polar interactions. The normal percentage of the blood volume that is occupied by red blood cells is approximately 45% in men and 42% in women (Vander et al., 2001). Hence the plasma content of whole blood should in theory be approximately 55 58%. Consequently, if all PFCs distribute only to plasma proteins the concentration ratio factor between plasma and whole blood should be 1.7 for women and 1.8 for men. In the present study on the PFC distribution between plasma and whole blood the ratios of PFOA and PFOS are lower than this theoretical ratio (Fig. 1 and Table 1). Average concentrations of PFOS and PFOA in plasma were only 1.2 and 1.4 times higher than in whole blood. Similar results were obtained for the 6-carbon sulfonate (PFHxS, 1.2) and the 9-carbon carboxylate (PFNA, 1.1). No gender trend could be seen. Information on longer or shorter Table 2 Mean and median are given when the compound occurred above the detection limit in more than 50% of the samples PFHxA PFHxS PFOA PFOS PFNA PFDA PFDS PFOSA PFUnDA All, n = 66 Range a < (0.1) (0.5) (0.1) < < < (0.1) < Number detected Median (Q1 Q3) b 1.5 ( ) 2.5 ( ) 17.1 (13 23) 0.3 ( ) 0.2 (< ) 2.7 ( ) 0.2 (< ) Geometric mean % CI c Men age 19 46, n = 40 Range < < < < < Median (Q1 Q3) b 1.7 ( ) 2.7 ( ) 17.7 (14 23) 0.3 ( ) 0.1 (< ) 2.7 ( ) 0.2 (< ) Geometric mean % CI c Women age 46 75, n = 26 Range < < < < < Median (Q1 Q3) b 1.2 ( ) 2.1 ( ) 16.9 (11 24) 0.3 ( ) 0.2 (< ) 2.7 ( ) 0.1 (< ) Geometric mean % CI c a Figures in parenthesis are the limit of detection. b 25th percentile 75th percentile. c Confidence interval for geometric mean.

6 A. Kärrman et al. / Chemosphere 64 (2006) No. of adults (A) < PFOS (ng/ml) Men Women No. of adults (B) < PFOA (ng/ml) Men Women No. of adults (C) < PFHxS (ng/ml) Men Women More No. of adults (D) < More PFOSA (ng/ml) Fig. 2. Frequency distribution of concentrations (ng/ml) in blood samples from 66 Swedish adults for (A) PFOS, (B) PFOA, (C) PFHxS and (D) PFOSA. Men Women Table 3 Range and median for perfluorinated, polychlorinated and polybrominated compounds in whole blood from Sweden. The concentrations are given in ng/ml whole blood and in (ng/g lipid weight) Sum PFC a Sum PCB b HCB p,p 0 DDE Sum chlordane c Sum BDE d All, n = 66 Range 6 70 ( ) ( ) (9 81) (29 895) (6 70) (4 110) Median 28 ( ) 1.5 (431) 0.09 (26) 0.6 (145) 0.08 (24) 0.06 (17) Men age 19 46, n = 40 Range 6 70 ( ) ( ) (9 47) (29 446) (8 70) (4 97) Median 29 ( ) 1.2 (361) 0.08 (24) 0.3 (109) 0.07 (22) 0.08 (22) Women age 46 75, n = 26 Range 6 41 ( ) ( ) (9 81) (51 895) (6 68) (4 110) Median 24 ( ) 2.4 (609) 0.12 (32) 1.6 (341) 0.13 (31) 0.05 (13) a Sum of 9 compounds. b Sum of 37 congeners. c Sum of 6 congeners. d Sum of 3 homologues. carbon chain acids could not be obtained since their levels were lower than the LOD. All ratios are lower than the expected theoretical value and although plasma content in whole blood can fluctuate a partial distribution to cellular elements can not be ruled out. PFOSA distributed mainly to the cellular volume as indicated by the low plasma/whole blood ratio (0.2). This suggests a different pharmacokinetic distribution behavior than for the acids. This fact should be considered in other situations as well, for instance in analytical method development of this type of compounds. As the observation in this study illustrates, using a factor of 2 to convert whole blood levels to plasma levels would result in an overestimation of the PFOS concentration by about 60% and PFOSA would be 10 times overestimated. The relatively low level in human blood taken together with the low distribution to plasma makes it uncertain to quantify PFOSA in plasma. Considering these results and the absence of a normalization parameter we suggest that the best choice of matrix for PFC determination in humans is whole blood since whole blood and plasma concentrations differed for several PFCs in this study. Further studies need to be performed considering the small sample set and relatively low blood concentrations of PFCs in this study. Individual variations in plasma/whole blood, as well as how sensitive the

7 1588 A. Kärrman et al. / Chemosphere 64 (2006) distribution is to dilution, day-to-day variation and sample handling should be addressed Comparisons between fluorinated, chlorinated and brominated compounds In recent years perfluorinated compounds have gained more and more interest. Production and usage of perfluorinated compounds increased in the 80s. A concentration increase between 1977 and 2003 with a factor 3 for PFOS and 10 for PFOA was observed in human serum from Japan (Harada et al., 2004). The same trend although to a lesser extent was observed in USA for PFOS, PFOA, PFHxS and perfluorooctanesulfonamidoacetate (PFO- SAA) (Olsen et al., 2005). Levels, trends and effects of polychlorinated and polybrominated organic compounds in the environment and humans are fairly well studied. Levels of most of the classical POPs have declined in the environment since the 80s due to bans and restrictions. In the present study on current levels of persistent halogenated compounds in Swedish blood, levels of perfluorinated compounds are higher than blood levels of the polychlorinated and polybrominated compounds (Table 3). On a whole blood basis, the concentration of perfluorinated compounds (sum PFCs) was times higher than the sum of PCBs and p,p 0 -DDE and about times higher compared to HCB, chlordanes and sum of PBDEs. Due to differences in tissue distribution between these compounds, such a comparison of levels on whole blood basis levels does not reflect the comparison of the total body burdens. Fluorinated compounds distribute mainly to the liver and blood while chlorinated and brominated POPs distribute mainly to fatty organs including adipose tissue and blood lipids. Since blood only contains approximately 0.5% lipid the amount of for instance PCBs in the blood compartment is only a fraction of the total body burden of PCB. The total body burden of PFOS in an average adult can be estimated on the assumptions of a total of 5 kg blood, a liver weight of 1.5 kg and that the liver to serum ratio of PFOS is 2.1:1 (Olsen et al., 2003b). Using the mean PFOS value in this study to calculate the body burden results in a burden of 1.6 mg PFOS (blood and liver). The calculated body burden of PCB #153 in a Swedish adult, with a total fat weight of 15 kg and 0.5% blood lipid content, is 1.7 mg. Although the absolute body burdens of PFOS and PCB #153 are similar, the dose of each compound at target organs or tissues may not be comparable neither the health risk associated with a certain body burden. However, since it is uncertain if PFC exposure is about to decline or continues increasing this group of chemicals demands attention from a health perspective in the near future. Presented individual information on blood levels of both PFCs and classical POPs allows for a comparison of the exposure and also behavior of the different compound classes. The scatter plot for PFOS and the sum of PCBs given in Fig. 3 illustrates an example of the non-existent association between perfluorinated compounds and the chlorinated or brominated compounds. Also given in Fig. 3 is the scatter plot of PCB sum and HCB, which shows an association typical for these lipophilic POPs due to a similar exposure and bioaccumulation behavior. No co-exposure seems to exist between studied PFCs and the other POPs. The PCB, HCB and p,p 0 -DDE concentrations in the older women were clearly higher compared to the younger men (Table 2). The same pattern can not be seen for PBDE and PFCs. Mean levels of PFHxS and PFOSA were higher in the blood of the young men. However, the median levels for the same compounds were similar for both men and women, which implies that the difference is due to a few male (young) donors with elevated levels. Since the PFC and POP blood levels assessed are for young men and older women, no conclusion can be drawn regarding gender or age differences PFC levels The purpose of this study was to compare human levels of perfluorinated chemicals and traditional persistent organic pollutants. The samples used are from sons and their mothers included as controls in a testicular cancer PFOS (ng/ml) r 2 < (A) sum PCB (ng/ml) HCB (ng/ml) r (B) sum PCB (ng/ml) Fig. 3. Scatter plots and regression analysis of association between POP concentrations in blood samples of 66 individuals from Sweden. (A) PFOS and sum PCB (ng/ml whole blood), (B) HCB and sum PCB (ng/ml whole blood).

8 A. Kärrman et al. / Chemosphere 64 (2006) study. This sample set is not representative for the whole Swedish population due to skewed gender and age distribution together with poor geographical distribution. However, since very little information about PFC levels in Swedish residents has been published prior to this study the sample set of the 66 individuals was used to indicate the background human levels of several PFCs in Sweden. Our results show that the Swedish population is exposed to a large number of PFCs. PFOS concentrations in whole blood from both men and women were shown to be normal distributed, which suggests a relatively uniform exposure or that a physiological equilibrium takes place. PFOSA for both sexes together with PFOA and PFHxS for men were found in high concentrations in a few samples, making the distribution skew (Fig. 2). Two men with high blood concentration of PFHxS (22 and 28 ng/ml) also showed high concentration of PFOS (32 and 35 ng/ml) but linear regression analysis of PFOS and PFHxS for all men did not result in a strong association (r 2 = 0.3). Other studies have shown a significant relationship between PFOS and PFHxS in human serum samples (Kannan et al., 2004). PFHxS can be a contaminant in PFOS-based products manufactured by electrochemical fluorination (3M, 1999). However, possible co-exposure of PFHxS as a contaminant in a PFOS product seems unlikely here since the concentration of PFHxS is nearly as high as the concentration of PFOS. More likely is the presence of a different PFHxS source. PFHxS has been used in fire-fighting foam and carpet treatment products (Olsen et al., 2003a). Higher levels of PFOSA ( ng/ml) were found in the blood of three men. PFOSA has not been connected to specific consumer applications but is a degradation product from other PFCs used for paper protection and surface treatment (Olsen et al., 2003a). The strongest association between all PFCs in this study was found for the concentrations of PFOS and PFOA in women (r 2 = 0.5), an association that has been shown in other studies as well (Olsen et al., 2003a, 2004; Kannan et al., 2004). This might indicate a common exposure for PFOS and PFOA although they are not known to be used together in products or to have the same bioaccumulation. For comparison, human levels of PFOS and PFOA so far reported in different countries are illustrated in Fig. 4. Both studies using serum and whole blood samples are given in the figure and the concentration difference between the two matrices can clearly be seen. The serum values for Sweden were obtained from the whole blood results by applying the mean ratios between plasma and whole blood presented in this study. Generally, North America has the highest levels and India, Columbia and Italy are countries with the lowest levels of PFOS. Levels of PFOA do not PFOA PFOS PFOS** USA (a) USA (b) USA (c) USA (d) Canada (e) USA (b) Sweden (f*) Belgium (b) Brazil (b) Japan (b) Italy (b) India (b) USA (b) Poland (b) Korea (b) Sweden (f) EU parliament (g) Malaysia (b) Columbia (b) serum whole blood Conc. (ng/ml) Fig. 4. Comparison of median levels (ng/ml) of PFOS and PFOA in serum and whole blood reported from different countries. * Concentrations transformed from whole blood through multiplication with (a) factors obtained in present study (PFOS, PFOA) and (b) factor of 2 (only PFOS ** ). (a) n =20 (Kuklenyik et al., 2004), (b) n =75(Kannan et al., 2004), (c) n = 645 (Olsen et al., 2003a), (d) n = 238 (Olsen et al., 2004), (e) n =56(Kubwabo et al., 2004), (f) n =70(Kannan et al., 2004) (g) n = 66 (this study), (h) n =20(Kannan et al., 2004), (i) n =27(Kannan et al., 2004), (j) n =38(Kannan et al., 2004), (k) n =50(Kannan et al., 2004), (l) n =45(Kannan et al., 2004), (m) n =30(Kannan et al., 2004), (n) n =25(Kannan et al., 2004), (o) n =50 (Kannan et al., 2004), (p) n =45(WWF, 2004), (q) n =23(Kannan et al., 2004), (r) n =56(Kannan et al., 2004).

9 1590 A. Kärrman et al. / Chemosphere 64 (2006) follow the levels of PFOS and in two studies (India and Korea) PFOA exceed the level of PFOS. Swedish PFOS levels are in the middle of the world-range in the whole blood group as well as in the serum group after correction using the plasma/whole blood ratio. This is an indication that the used ratios are correct. If a factor of 2 had been applied instead the Swedish concentrations in the serum group would be at the same level as the ones reported from USA (Fig. 4). This emphasizes the difficulties in comparing PFC concentrations and care must be taken when assessing human levels and body burdens of non-lipophilic compounds as long as there is no normalization procedure. Acknowledgment This research was supported by grants from the Cancer and Allergy Foundation (Cancer-och Allergifonden, Stockholm, Sweden) and the Swedish Environmental Protection Agency. References Dinglasan, M.J.A., Ye, Y., Edwards, E.A., Mabury, S.A., Fluorotelomer alcohol biodegradation yields poly- and perfluorinated acids. Environ. Sci. Technol. 38, Ellis, D.A., Martin, J.W., De Silva, A.O., Mabury, S.A., Hurley, M.D., Sulbaek Andersen, M.P., et al., Degradation of fluorotelomer alcohols: a likely atmospheric source of perfluorinated carboxylic acids. Environ. Sci. Technol. 38, EPA, US Environmental Protection Agency Public Docket AR 226. Washington, DC. Han, X., Snow, T.A., Kemper, R.A., Jepson, G.W., Binding of perfluorooctanoic acid to rat and human plasma proteins. Chem. Res. Toxicol. 16, Harada, K., Saito, N., Inoue, K., Yoshinaga, T., Watanabe, T., Sasaki, S., et al., The influence of time, sex and geographic factors on levels of perfluorooctane sulfonate and perfluorooctanoate in human serum over the last 25 years. J. Occup. Health 46, Hardell, L., van Bavel, B., Lindström, G., Carlberg, M., Dreifaldt, C., Wijkström, H., et al., Increased concentrations of polychlorinated biphenyls, hexachlorobenzene, and chlordanes in mothers of men with testicular cancer. Environ. Health Perspect. 111, Holmström, K.E., Järnberg, U., Bignert, A., Temporal trends of PFOS and PFOA in guillemot eggs from the Baltic Sea, Environ. Sci. Technol. 39, Inoue, K., Okada, F., Ito, R., Kato, S., Sasaki, S., Nakajima, A.U., et al., Perfluorooctane sulfonate (PFOS) and related perfluorinated compounds in human maternal and cord blood samples: assessment of PFOS exposure in susceptible population during pregnancy. Environ. Health Perspect. 112, Jones, P.D., Hu, W., de Coen, W., Newsted, J.L., Giesy, J.P., Binding pf perfluorinated fatty acids to serum proteins. Environ. Toxicol. Chem. 22, Kallenborn, R., Berger, U., Järnberg, U., Perfluorinated alkylated substances (PFAS) in the nordic environment. TemaNord 2004:552. Nordic Council, available at < Copenhagen, Denmark. Kannan, K., Corsolini, S., Falandysz, J., Fillman, G., Kumar, K.S., Loganathan, B.G., et al., Perfluorooctanesulfonate and related fluorochemicals in human blood from several countries. Environ. Sci. Technol. 38, Kärrman, A., van Bavel, B., Järnberg, U., Hardell, L., Lindström, G., Development of a solid-phase extraction-hplc/single quadrupole MS method for quantification of perfluorochemicals in whole blood. Anal. Chem. 77, KEMI, PFOS-relaterade ämnen; strategi för utfasning. Swedish Chemical Inspectorate. Available at < upload/trycksaker/pdf/rapporter/rapport3_04.pdf>. Stockholm, Sweden. Kennedy, G.L., Butenhoff, J.L., Olsen, G.W., O Connor, J.C., Seacat, A.M., Perkins, R.G., et al., The toxicology of perfluorooctanoate. Crit. Rev. Toxicol. 34, Kissa, E., Fluorinated Surfactants and Repellents, second ed. Marcel Dekker Inc., New York. Kubwabo, C., Vais, N., Benoit, F.M., A pilot study on the determination of perfluorooctanesulfonate and other perfluorinated compounds in blood of Canadians. J. Environ. Monit. 6, Kuklenyik, Z., Reich, J.A., Tully, L.L., Needham, L.L., Calafat, A.M., Automated solid-phase extraction and measurement of perfluorinated organic acids and amides in human serum and milk. Environ. Sci. Technol. 38, Martin, J.W., Smithwick, M.M., Braune, B.M., Hoekstra, P.F., Muir, D.C., Mabury, S.A., Identification of long-chain perfluorinated acids in biota from the Canadian Arctic. Environ. Sci. Technol. 38, Moriwaki, H., Takata, Y., Arakawa, R., Concentrations of perfluorooctane (PFOS) and perfluorooctanoic acid (PFOA) in vacuum cleaner dust collected in Japanese homes. J. Environ. Monit. 5, M, Fluorochemical Use, Distribution and Release Overview. US Environmental Protection Agency Public Docket AR M, Determination of Serum Half-lifes of Several Fluorochemicals. US Environmental Protection Agency Public Docket AR OECD, Draft assessment of perfluorooctane sulfonate (PFOS) and its salts: complete assessment. ENV/JM/RD(2002)17/FINAL. Organisation for Economic Co-operation and Development. Available at < Olsen, G.W., Church, T.R., Miller, J.P., Burris, J.M., Hansen, K.J., Lundberg, J.K., et al., 2003a. Perfluorooctanesulfonate and other fluorochemicals in the serum of American Red Cross blood donors. Environ. Health Perspect. 111, Olsen, G.W., Hansen, K.J., Stevenson, L.A., Burris, J.M., Mandel, J.H., 2003b. Human donor liver and serum concentrations of perfluorooctanesulfonate and other perfluorochemicals. Environ. Sci. Technol. 37, Olsen, G.W., Church, T.R., Larson, E.B., van Belle, G., Lundberg, J.K., Hansen, K.J., et al., Serum concentrations of perfluorooctanesulfonate and other fluorochemicals in an elderly population from Seattle, Washington. Chemosphere 54, Olsen, G.W., Huang, H.-Y., Helzlsouer, K.J., Hansen, K.J., Butenhoff, J.L., Mandel, J.H., Historical comparison of perfluorooctanesulfonate, perfluorooctanoate and other fluorochemicals in human blood. Environ. Health Perspect. 113, Posner, S., Järnberg, U., Fluorkarboner i slam och avloppsvatten från industritvätt och textila beredningsverk. VA Forsk VA- Forsk, Svenskt vatten AB. Stockholm, Sweden. Available at < Schultz, M., Barofsky, D.F., Field, J., Quantitative determination of fluorotelomer sulfonates in groundwater by LC MS/MS. Environ. Sci. Technol. 38, Stock, N.L., Lau, F.K., Ellis, D.A., Martin, J.W., Muir, D.C., Mabury, S.A., Polyfluorinated telomer alcohols and sulfonamides in the North American troposphere. Environ. Sci. Technol. 38, Thibodeaux, J.R., Hanson, R.G., Rogers, J.M., Grey, B.E., Barbee, B.D., Richards, J.H., et al., Exposure to perfluorooctane sulfonate during pregnancy in rat and mouse. I: maternal and prenatal evaluations. Toxicol. Sci. 74, Tittlemier, S.A., Pepper, K., Edwards, L., Tomy, G., Development and characterization of a solvent extraction gas chromatographic/ mass spectrometric method for the analysis of perfluorooctanesulfonamide compounds in soil matrices. J. Chromatogr. A 1066,

10 A. Kärrman et al. / Chemosphere 64 (2006) Tomy, G.T., Tittlemier, S.A., Palace, V.P., Budakowski, W.R., Braekevelt, E., Brinkworth, L., et al., Biotransformation of N-ethyl perfluorooctanesulfonamide by Rainbow Trout (Onchorhynchus mykiss) liver microsomes. Environ. Sci. Technol. 38, Tomy, G.T., Budakowski, W.R., Halldorson, T., Helm, P.A., Stern, G.A., Friesen, K., et al., Fluorinated organic compounds in an eastern arctic marine food web. Environ. Sci. Technol. 38, Vander, A.J., Sherman, J.H., Luciano, D.S., Human Physiology: The Mechanism of Body Function, eighth ed. McGraw-Hill Book Co., Singapore. WWF, Chemical check up-an analysis of chemicals in the blood of members of the European Parliament. WWFs European Policy Office. Available at < checkupmain.pdf>.