Science of the Total Environment 529 (2015) 198–212
Contents lists available at ScienceDirect
Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv
Food intake and serum persistent organic pollutants in the Greenlandic pregnant women: The ACCEPT sub-study Manhai Long a,⁎, Ane-Kersti Skaarup Knudsen a, Henning Sloth Pedersen b, Eva Cecilie Bonefeld-Jørgensen a a b
Centre for Arctic Health & Unit of Cellular and Molecular Toxicology, Department of Public Health, Aarhus University, Aarhus, Denmark Primary Health Care Center, Nuuk, Greenland
H I G H L I G H T S • • • • •
Pregnant women in East and North Greenland have higher intake of marine mammals Higher serum levels of PCBs, OCPs, PFASs and mercury in North and East Greenland Serum levels of PFASs were significantly associated with levels PCBs and OCPs Legacy POPs, PFOS and PFOA levels have decreased but other PFASs have sustained No significant regional differences for serum levels of PBDEs
a r t i c l e
i n f o
Article history: Received 22 December 2014 Received in revised form 6 May 2015 Accepted 6 May 2015 Available online xxxx Editor: Adrian Covaci Keywords: Contaminants Pregnant Diet Lifestyle Greenland
a b s t r a c t The Greenlandic Inuit have high blood concentrations of environmental persistent organic pollutants (POPs). High POP concentrations have been associated with age, smoking and consumption of marine mammals. Studies have indicated that exposure to POPs during pregnancy may adversely affect fetal and child development. To assess geographical differences in diet, lifestyle and environmental contaminant exposure among pregnant women in Greenland, blood samples and questionnaire data were collected from 207 pregnant women in five Greenlandic regions (North, Disco Bay, West, South and East). Blood samples were analyzed for 11 organochlorine pesticides (OCPs), 14 polychlorinated biphenyls (PCBs), 5 polybrominated diphenyl ethers (PBDEs), 15 perfluoroalkylated substances (PFASs) and 63 metals. A trend of higher intake of marine mammals in the East and North regions was reflected by a higher n−3/n−6 fatty acid ratio. Participants in the East region tended also to have higher intake of terrestrial species. A significant higher seabird intake was seen for pregnant women in the West region. Significant regional differences were found for blood concentrations of PCBs, OCPs, PFASs and mercury, with higher levels in the North and East regions. PFASs were significantly associated with PCBs and OCPs in most of the regions. In the North region, PFASs were associated with both selenium and mercury. No significant regional difference was observed for PBDEs. The regional differences of blood levels of POPs and mercury were related to differences in intake of the traditional food. Compared to earlier reports, decreased levels of legacy POPs, Hg and Pb and perfluorooctane sulfonate and perfluorooctanoic acid were observed, but the levels of PFAS congeners perfluorohexane sulfonate and perfluorononanoic acid were sustained. The detection of POPs and heavy metals in maternal blood indicates fetal exposure to these compounds possibly influencing fetal development. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Lipophilic persistent organic pollutants (POPs) such as organochlorine pesticides (OCPs), polychlorinated biphenyls (PCBs), and polybrominated diphenyl ethers (PBDEs) are ubiquitous and found ⁎ Corresponding author at: Centre for Arctic Health & Unit of Cellular and Molecular Toxicology, Department of Public Health, Aarhus University, Bartholins Allé 2, Building 1260, 8000 Aarhus C, Denmark. E-mail addresses: [email protected] (M. Long), [email protected] (A.-K.S. Knudsen), [email protected] (H.S. Pedersen), [email protected] (E.C. Bonefeld-Jørgensen).
http://dx.doi.org/10.1016/j.scitotenv.2015.05.022 0048-9697/© 2015 Elsevier B.V. All rights reserved.
globally due to long-range transport by atmospheric and oceanic currents. Thus, these environmental contaminants are also brought to the Arctic where they are trapped by the cold climate (Barrie et al., 1992; Burkow and Kallenborn, 2000; Macdonal et al., 2000). The lipophilic POPs are accumulated and biomagnified in lipid-based Arctic food webs (Borga et al., 2004; Kelly et al., 2008; Laender et al., 2010) causing international scientific concern. In 1994, all circumpolar countries agreed to address these concerns by establishing a health monitoring program for contaminants called AMAP (Arctic Monitoring and Assessment Program). The systematic studies concerning human exposure to
M. Long et al. / Science of the Total Environment 529 (2015) 198–212
POPs and possible health effects in the Arctic population have provided knowledge and new perspectives to the environmental agenda in the last 20 years. Arctic populations such as the Greenlandic Inuit display a higher body burden of POPs than people living in industrialized regions close to major emission sources. This is largely due to Arctic populations that rely on their traditional food, including species at the top of the marine food chain (AMAP, 2009; Bjerregaard et al., 2001, 2013; Bonefeld-Jorgensen, 2010). The levels of these lipophilic POPs in Arctic populations correlate with age, smoking and the plasma level of n −3 polyunsaturated fatty acids; the latter is a strong indicator of the main source of POP contamination in their traditional marine food (Bonefeld-Jorgensen, 2010; Carriere et al., 2012; Deutch et al., 2003, 2007a; Krüger et al., 2007). In general, the levels of many legacy-lipophilic POPs such as PCBs and OCP are decreasing in Arctic populations; this is supposedly due to a combination of regulation of the legacy POP causing reduced concentrations of the chemicals in the marine food web and a reduction in the consumption of traditional food from marine mammals such as whales, walrus and seals (Bjerregaard et al., 2013; Deutch et al., 2007a). Even though the significant downward trend of legacy-lipophilic POPs shows that the legislation works, it must be kept in mind that some POPs like perfluoroalkylated substances (PFASs) and PBDEs are still used and bioaccumulate in marine mammals and humans. PFASs are a large group of chemicals used since the 1950s in industry and in commercial products (e.g. Teflon, carpets, furniture, foodstuff packing). PFASs are persistent with half-lives in humans between 4 and 10 years (Olsen et al., 2008). Unlike the legacy-lipophilic POPs (PCBs and OCPs), PFASs bind to blood proteins and are stored mainly in the liver, kidneys, and biliary system (Bonefeld-Jorgensen et al., 2014; Long et al., 2012). PFASs are found globally, and governmental regulations on use and production of specific compounds such as perfluorooctane sulfonate (PFOS) and perfluorooctanoic acid (PFOA) have been made in both USA and Europe. The use of PFOS was restricted in USA in 2001 (Paul et al., 2009) and Europe from June 2008 (European Parliament, 2006) and PFOS has been added to Annex B of the Stockholm Convention on POPs (Stockholm Convention, 2009). USA launched the “PFOA Stewardship Program” where eight of the major PFOA producing companies committed to reduce emissions of PFOA and related chemicals by 95% by 2010 (US EPA, 2006). The classification and labeling of PFOA are currently discussed in the EU (Posner et al., 2013). Even though the concentrations of legacy-lipophilic POPs are decreasing, the long-term effects of POPs on general public health status in Greenland are of concern (Bonefeld-Jorgensen, 2010; Bonefeld-Jorgensen et al., 2014). The adverse effects of these chemicals on human health have been connected to disruption of the immune system, reduced reproductive ability, developmental and behavioral disabilities, cancer, and metabolic disorders (Bonefeld-Jorgensen et al., 2014). Biomarker studies have indicated that exposure to POPs can cause genetic alterations, decrease defense against oxidative stress, and disrupt endocrine systems (AMAP, 2009; Bonefeld-Jorgensen et al., 2014). Human studies found neurological impairment and immunological effects in the developing fetus and infants exposed to POPs (Chen et al., 2014; Heilmann et al., 2006; Van Oostdam et al., 2004), PFASs (Fei et al., 2007; Grandjean et al., 2012) and heavy metals (Claus Henn et al., 2014; Taylor et al., 2014). Since fetal development is the most sensitive period in life, lifestyle, diet and environmental exposure to POPs during pregnancy are important factors when examining and monitoring the health outcome of the next generation. The data presented in this study is a sub-study of the current establishment of a new Greenlandic birth cohort ACCEPT (Adaptation to Climate Change, Environmental Pollution, and Dietary Transition). The aim of the present study was to investigate lifestyle characteristics, dietary habits and levels of environmental contaminants in a geographically representative population of pregnant women in Greenland between 2010–11 and 2013.
199
2. Materials & methods 2.1. Study population We conducted this population based cross-sectional study in two steps: Between August 2010 and September 2011 a doctor employed at the Queen Ingrid Health Care Center in Nuuk and between June 2013 and November 2013 midwives at the Greenlandic hospitals recruited participants for this study. Data collection was carried out in six different towns at the West coast in Greenland (Fig. 1) and included a study population of pregnant women (N18 year); 212 pregnant women were enrolled in the present study. The enrolment depended on medical coast visit and workload in the hospitals. Among the asked pregnant women there were only few refusals caused by movement, fear of needles and being abroad. Five women were defined as nonInuit and excluded from the study. The final study population consisted of 207 pregnant women who all had lived more than 50% of their life in Greenland. The participant percentage of the total births and the participant percentage of possible inclusion of the population during 2010–2011 was 11.2 and 22.2%, respectively. Among the participants, approximately 84% had completely Inuit descent defined as both parents being born in Greenland. The remaining 16% were partly-Inuit and had one parent from Greenland and/or had Greenlandic as their native language. Each included pregnant women filled in a culturally-appropriate lifestyle questionnaire and a specially designed food frequency questionnaire (FFQ) for Inuit. Both questionnaires were available in Danish and Greenlandic. Data on pre-pregnancy height (in meters) and weight (in kilograms), reproductive factors and smoking status (current/noncurrent) of the participants were obtained from medical records. We applied the information from the lifestyle questionnaire about where the women were born or grew up (town) and for how long they had lived there (years). The women belonged to the region where they had lived N 50% of their lives. The FFQ included approximately 40 traditional Greenlandic food items and 23 imported food items. Traditional food items were grouped into four main groups (marine mammals, seabirds, fish and terrestrial species). The marine mammal group included polar bear, walrus, seals, and whales. The seabird group included guillemot, eider, and kittiwake. The fish group included trout, cod, Greenlandic halibut, Atlantic halibut, redfish, Atlantic salmon, and capelin. The terrestrial species included caribou, muskox, Greenlandic lamb, hare, and ptarmigan. The imported food group included 23 imported food items e.g. meat products including non-Greenlandic fish, fast food, sauce, pasta, rice, vegetables, fruit, chocolate, sweets, and snacks. For each food item, the frequency of intake was reported in three categories: Never; Rare (less than once a month, approximately once a month, 2–3 times a month) and Frequently (once a week, 2–3– 4 times a week, every day or almost every day, several times a day). The number of answers in the three categories of each food item was calculated as percentages, then each percentage was added up and divided by the number of food items in the corresponding main traditional food groups and imported food group. A detailed calculation of food intake frequency of the main food groups is shown in Appendix A. Samples of venous blood were collected from the pregnant women at inclusion into the study (mean gestational age 25.1 week, range 7–40 weeks); 7.2% of blood samples were taken in the first trimester (gestational weeks 7–12), 57% during the second trimester (gestational weeks 13–28) and 36.2% in the third trimester (gestational weeks 29–40). The blood samples were stored at − 80 °C until analyses. Informed consent was obtained from all participants prior to data collection. The study was approved by the Ethical Committee for Scientific Investigations in Greenland (KVUG) and conducted in accordance with the Helsinki Declaration.
200
M. Long et al. / Science of the Total Environment 529 (2015) 198–212
Fig. 1. Map of Greenland with the collection sites and regions where the participants lived more than 50% of their life.
2.2. Determination of contaminants 2.2.1. Lipophilic POPs Serum samples were analyzed for 30 lipophilic POPs and lipid content at Le Centre de Toxicologie du Québec (Sainte-Foy, Québec, Canada), a certified laboratory by the Canadian Association for Environmental Analytical Laboratories (Butler Walker et al., 2003). Fourteen polychlorinated biphenyl (PCB) congeners [PCB 28, 52, 99, 101, 105, 118, 128, 138, 153, 156, 170, 180, 183, 187], five flame retardants including one polybrominated biphenyl [PBB153] and four polybrominated diphenyl ethers (PBDEs) [PBDE 47, 99, 100, 153], and 11 organochlorine pesticides (OCPs) [p,p′-DDT (dichlorodiphenyltrichloroethane) and its major metabolite p,p′-DDE (dichlorodiphenyldichloroethylene), aldrin, mirex, β-hexachlorocyclohexane (β-HCH), hexachlorobenzene (HCB), cis- and trans-nonachlor, and α-, γ- and oxy-chlordane] were analyzed in purified extracts by high-resolution gas chromatography with electron capture detection. All determined PCBs, OCPs and PBDEs were adjusted to the plasma lipid content analyzed in the same sample (Butler Walker et al., 2003; Deutch et al., 2007b) and reported as μg/kg plasma lipid. If the value was below the detection limit, we assigned the value given by the detection limit divided by two.
2.2.2. Perfluoroalkylated substances (PFASs) Fifteen PFASs were analyzed in serum samples using liquid chromatography–tandem mass spectrometry (LC–MS–MS) with electrospray ionization (ESI) (Long et al., 2012): perfluorooctane sulfonate (PFOS), perfluorohexane sulfonate (PFHxS), perfluoroheptanesulfonate (PFHpS), perfluorooctanoic acid (PFOA), perfluorononanoic acid (PFNA), perfluoroheptanoic acid (PFHpA), perfluorodecanoic acid (PFDA), perfluoroundecanoic acid (PFUnA), perfluorododecanoic acid (PFDoA), perfluorotridecanoic acid (PFTrA), perfluorobutanesulfonic acid (PFBS), perfuoro-1-decanesulfonate (PFDS), perfluorooctane sulfonamide (PFOSA), perfluoropentanoic acid (PFPeA) and perfluoroheptanoic acid (PFHpA). The PFAS method performance is continuously controlled and tested in a Quality Assurance/Quality Control (QA/QC) by participating in inter-laboratory comparison studies performed by Institute Nationale de Santé Publique du Québec for AMAP (AMAP, 2000–2010). If the value was below the detection limit, we assigned the value given by the detection limit divided by two. 2.2.3. Blood metals Sixty-three blood metals including essential elements selenium (Se), zinc (Zn), iron (Fe) and copper (Cu) and heavy metals including lead (Pb), mercury (Hg), arsenic (As) and cadmium (Cd) were measured using Inductively Coupled Plasma Mass Spectrometry (ICP-MS) after
M. Long et al. / Science of the Total Environment 529 (2015) 198–212
digesting blood with nitric acid in the microwave at the accredited element laboratory, Institute for Bioscience-Arctic Research Centre, Aarhus University, Denmark. The quality was ensured by repeated analyses and by frequent analysis of certified reference material (Seronorm), as well as by participation in the Quality Assurance of Information in Marine Environmental monitoring (QUASIMEME) inter-laboratory comparison program (Asmund et al., 2004). In this paper, we only reported the relevant toxic metals, Hg, Pb, Cd, As, Al, and Cr and trace element Selenium (Se). 2.2.4. Plasma fatty acids The ratio between n−3 polyunsaturated fatty acids and n−6 fatty acids is known to be a strong indicator of marine food intake and thus a good indicator of the relative consumption of traditional versus imported food (Deutch et al., 2004; Tjonneland et al., 1993). Plasma fatty acids were determined by capillary gas–liquid chromatography at the Biology Department, University of Guelph, Canada (Cote et al., 2004). The fatty acid composition of plasma phospholipids was expressed as a percentage of the total area of all fatty acid peaks from 14:0 to 24:0. Plasma phospholipids of fatty acids correspond to relative percentage of total fatty acids by weight. We reported n−3 polyunsaturated fatty acids on the sum of C18:3, n−3; C18:4, n−3; C20:3, n−3; C20:4, n−3; C20:5, n−3; C22:5, n−3; and C22:6, n−3, and the n−6 fatty acids as the sum of C18:2, n−6; C18:3, n−6; C20:2, n−6; C20:3, n−6; C20:4, n−6; C22:2, n−6; C22:4, n−6 and C22:5, n−6. 2.2.5. Determination of plasma cotinine Cotinine is a metabolite of nicotine. The level of cotinine in the blood is proportionate to the amount of exposure to tobacco smoke and is thus a valuable indicator of recent exposure to tobacco smoke. The Calbiotech Cotinine Direct ELISA Kit was used to measure plasma cotinine (Calbiotech Inc., CA, USA) at the Unit of Cellular & Molecular Toxicology, Centre of Arctic Health, Department of Public Health, Aarhus University, Denmark. This assay is a solid phase competitive ELISA and the absorbance was read on ELISA Reader at 450 nm. The plasma cotinine concentration was expressed as ng/mL and the detection limit was 1 ng/mL. If the value was below the detection limit, we assigned a value given by the detection limit divided by two. 2.3. Statistical analysis Lifestyle and FFQ data were double-entered in the validation program EpiData file version 3.1.2701.2008. Data on lifestyle, diet and chemicals were analyzed in SPSS Statistics version 19.0 & 21 (SPSS Inc, Chicago, IL, USA) with a significance level of 5%. Based on the chemical structure, we grouped the analyzed PFASs in perfluorosulfonated acids (∑ PFSA, sum of PFBS, PFHxS, PFHpS, PFOS, PFDS and PFOSA) and perfluorocarboxylated acids (∑ PFCA, sum of PFPeA, PFHxA, PFHpA, PFOA, PFNA, PFDA, PFUnA, PFDoA, PFTrA, and PFTeA). Because of high legacy POP inter-correlation, PCB congeners were also grouped as the sum of 14 PCB congeners (Σ 14 PCB) and the sum of 11 pesticides (Σ OCP). Among tested PCBs, PCB 105, PCB118 and PCB156 are dioxin-like PCB (DL-PCB) (Van den Berg et al., 2006) and their summation was grouped as Σ DL-PCB. ∑ PBDE was the sum of four PBDE congeners and PBB153. In addition, analyses were also performed for the single compounds detected in the samples. The distribution of data was checked by Q–Q plots. The natural logarithmic transformed variables made the distribution more symmetrical and improved the normality of data; thus, the analysis was performed on the ln-transformed data. The data were treated as continuous and categorical variables. We used One-way ANOVA to compare the graphical continuous variables and chemical levels among the regions. When ANOVA showed statistical significance, complementary multiple comparison ad hoc tests was performed. Test for equal variances of variables was performed using Levene's test. The least-significant difference (LSD) test was used if equal variance was shown, otherwise Dunett T3 was used. Previous studies reported a correlation between POPs and age and the
201
n−3/n−6 ratio (Deutch et al., 2007b; Fromme et al., 2007; Haug et al., 2009). Therefore comparisons of means for POPs among regions were also performed by the General Linear Model procedure adjusted for age and n−3/n−6 ratio. Pearson's chi-square test was used to check the difference among regions for category variables. Due to a large number of analytes, we used data reduction technique Principal Component Analysis (PCA) to identify potential underlying components of POPs and metal levels in the blood. The dimension of the data was reduced with PCA by forming a few linear combinations of the original variables. By PCA the correlated variables were grouped together. The coefficients defining these linear combinations, called “factor loadings”, are the correlations of each input variable with the component. The POPs and metals which were detectable in more than 1% of all blood samples were included into the model: 9 OCPs, 11 PCBs, 4 PBDEs, 10 PFASs, six toxic metals (Hg, Pb, Cd, As, Al and Cr) and trace element selenium (Se). The trace element Selenium (Se) was included together with the POPs and heavy metals because Greenlandic traditional food has a high content of Se; Greenlanders thus had a high burden of both POPs, Hg and Se (Hansen et al., 2004). The number of components was extracted on the basis of eigenvalues (N 1), and scree plot. A varimax rotation was used to obtain a set of independent and best interpretable components (Newby and Tucker, 2004). The Kaiser–Mayer–Olkin (KMO) test (0.901) and the Bartlett Test of Sphericity (p b 0.0001) were used to test suitability of the analysis. All factor loadings N|0.3| were used to identify the variables comprising a component. Factor scores for each participant were created by multiplying factor loadings with the corresponding standardized value for each variable and summing across variables. Therefore the factor scores indicate the extent to which the chemical concentrations conformed to the corresponding component. We performed a Spearman's correlation analysis to assess the bivariate correlation of the identified components, single POP congeners, blood level of Hg, Pb, Cd, As, Al, Cr and Se with characteristics and lifestyle factors such as age, body mass index (BMI), smoking marker plasma cotinine and seafood intake biomarker n−3/n−6 ratio. The nonparametric Mann–Whitney U test was used to test the differences between smoking status for identified components, chemicals and cotinine level, while the difference of factor scores and chemicals among parity were tested using the Kruskal–Wallis and Jonckheere–Terpstra tests. We used multiple linear regression analysis to assess the relationship between serum PFAS levels and serum levels of legacy lipophilic POPs as well as blood Se and Hg levels adjusting for age, n − 3/n − 6 ratio in each region. For the pooled regional data, the association was assessed adjusting for age, n−3/n−6 and region. 3. Results 3.1. Lifestyle and diet The study population included 207 pregnant women (18–44 years) with a median age of 27.0 years and median pre-pregnancy BMI of 24.2 kg/m2; 60% of the participants lived in the West region (Table 1). We observed similar population characteristics across the regions. Almost half of the women (44.2%) were current smokers at the beginning of their pregnancy (approx. at gestational week 12). We did not find a significant difference in self-reported smoking status across the regions. Current smokers had a much higher plasma cotinine level than non-smokers (median: 75.7 ng/ml vs. 0.5 ng/ml, p b 0.0001), supporting plasma cotinine being a surrogate marker of current smoking. No significant difference of plasma cotinine level among regions was observed. A tendency of higher smoking rate and plasma cotinine level was found in participants in the North (Table 1). Parity, defined as a full-term pregnancy, was evenly distributed in the regions West, Disko Bay and South concerning one birth or equal to or more than two births. The three women from the East region had two or more births. In the regions North, Disko Bay, West and
202
M. Long et al. / Science of the Total Environment 529 (2015) 198–212
Table 1 Characteristics of the study population: Pregnant women in Greenland.
Age (years)
BMI b
Smokingc P-cotinine (ng/ml)d
Alcohol consumption Before pregnancy
During pregnancy Reproductive factors Parity
Regiona
North
Disko Bay
West
South
East
p value
N (%) n Mean ± SD Median (min–max) n Mean ± SD Median (min–max) n Yes (%) n Mean ± SD Median (min–max)
15 (7.2) 15 28.3 ± 3.9 27.0 (21–35) 13 24.1 ± 2.9 23.8 (20.1–29.1) 14 9 (64.3) 15 48.99 ± 55.57 15.2 (0.50–160.0)
50 (24.2) 50 27.7 ± 5.3 27.0 (19–40) 49 25.9 ± 5.2 24.8 (17.7–39.3) 50 25 (50.0) 50 47.61 ± 61.13 0.74 (0.50–184.0)
124 (59.9) 124 27.1 ± 5.3 27.0 (18–44) 118 25.3 ± 5.0 24.4 (16.4–46.6) 124 51 (41.1) 124 30.52 ± 47.41 0.50 (0.50–190.5)
15 (7.2) 15 28.3 ± 5.1 28.0 (22–41) 15 23.0 ± 3.6 22.6 (17.5–33.7) 15 6 (40.0) 15 33.02 ± 48.80 0.50 (0.50–126.1)
3 (1.4) 3 33.7 ± 2.9 32.0 (32–37) 3 27.5 ± 3.3 28.9 (23.7–29.7) 3 0 (0.0) 3 2.83 ± 4.04 0.50 (0.50–7.50)
b0.0001# 207 (100) 0.230* 207 27.5 ± 5.2 27.0 (18–44) 0.226* 198 25.2 ± 4.9 24.2 (16.4–46.6) 0.204# 206 91 (44.2) 0.565* 207 35.78 ± 51.80 0.50 (0.50–190.5)
All GRL towns
0.256# n ≤1 a month (%) N2–3 times a month (%) n ≤1 a month (%) N2–3 times a month (%)
15 6 (40.0) 9 (60.0) 14 13 (92.8) 1 (7.1)
47 28 (59.6) 19 (40.4) 44 44 (100) 0 (0.0)
121 83 (68.6) 38 (31.4) 117 117 (100) 0 (0.0)
15 10 (66.7) 5 (33.3) 12 12 (100) 0 (0.0)
3 2 (66.7) 1 (33.3) 3 0 (0.0) 0 (0.0)
n Mean ± SD Median (min–max) 0 (%) 1 (%) ≥2 (%)
15 1.7 ± 1.5 1.0 (0–5) 4 (26.7) 4 (26.7) 7 (46.7)
50 1.2 ± 1.2 1.0 (0–4) 17 (34.0) 15 (30.0) 18 (36.0)
124 0.98 ± 1.2 1.0 (0–6) 52 (41.9) 42 (33.9) 30 (24.2)
15 0.93 ± 1.1 1.0 (0–4) 6 (40.0) 6 (40.0) 3 (20.0)
3 2.3 ± 0.6 2.0 (2–3) 0 (0.0) 0 (0.0) 3 (100)
0.013#
0.117#
201 129 (64.2) 72 (35.8) 190 189 (99,45) 1 (0.5) 207 1.1 ± 1.2 1.0 (0–6) 79 (38.2) 67 (32.4) 61 (29.5)
N: total number of participants in the region. n: total number of respondents having information for the corresponding parameters. GRL: Greenland, 15 towns are included. p value calculated with *One-way ANOVA analysis on ln-transformed data, and #Chi square test. a Region: Where the women have lived N50% of their life, given as following regions: North: Upernavik (8), Uummannaq (4), Qaanaaq (3); Disko Bay: Ilulissat (26), Aasiaat (18), Qeqertarsuaq (3), Qasigiannguit (3); West: Nuuk (59), Maniitsoq (49), Sisimiut (11), Paamiut (5); South: Qaqortoq (8), Nanortalik (5), Narsaq (2); East: Tasiilaq (3). b BMI: Body mass index, calculated pre-pregnancy as kg/m2. c Current smoking: the smoking status is obtained from the pregnancy journal. The women were asked about current and non-current smoking during their early pregnancy. The Yespercent is calculated from the total answers in the region-group. d P-cotinine: plasma cotinine is a metabolite of nicotine and thus a primary biomarker for the determination of tobacco exposure.
South, the first-time mothers accounted for 26.7%, 34.0%, 41.9% and 40.0%, respectively (Table 1). Table 2 shows the food intake of the pregnant women in the different regions. Across the regions the only significant difference observed was for seabirds (p = 0.011) where the highest frequent intake was found in the West region and the lowest in the Disko Bay region and East (Table 2). There was a tendency, though not statistically significant, to higher intake of marine mammals in the North and East regions; the frequency of intake of fish and terrestrial species was higher for the pregnant women in the East region, indicating a preference for traditional Inuit diet. The n−3/n−6 ratio, indicator of seafood intake, also supported a higher marine food intake for the women in the East and North regions (Table 2). 3.2. Serum concentrations of lipophilic POPs Most lipophlic POPs were detectable in the blood samples. However, PCB 28, PCB52, PBB 153, aldrin and γ-chlordane were only detected in less than 1% of the samples (Table A.1a). Among 14 tested PCB congeners, three PCBs (PCB 138, 153, 180) showed higher concentration ranges (Table A.2, Fig. 2A). As shown in Fig. 2A, the median of PCB153 level was 100, 64, 51, 57 and 220 μg/kg plasma lipid for the pregnant women in the North, Disko Bay, West, South and East regions, respectively. The median of PCB 153 for all the included Greenlandic towns was 57 μg/kg plasma lipid. We found a significant difference among regions for the serum levels of indicator PCBs, PCB 138, PCB 153 and PCB 180 and the dioxin-like PCB (DL-PCB) such as 105 and PCB 118, although the PCB156 level did not differ among the regions (Fig. 2A). Pregnant women living in the East (Tasiilaq, n = 3) and the North regions (n = 15) had higher levels of PCBs compared to the other regions
(Fig. 2A). A similar pattern was observed for the sum of DL-PCB (PCB 105, PCB118 and PCB156) and the sum of 14 PCBs (Table A.2). Among the tested organochlorine pesticides (OCPs), p,p′-DDE showed the highest concentration in participants from all regions with a median of 130 μg/kg plasma lipid for all the included Greenland towns (Fig. 2B, Table A.2). Similarly to the PCB levels, higher concentrations of OCPs such as p,p′-DDE, oxychlordane, trans-nonachlor, HCB and β-HCH were found in pregnant women living for the longest period in the East and North regions (Fig. 2B, Table A.2). After adjustment for potential confounders such as age and seafood intake marker (n − 3/ n−6), significant regional differences still remained (data not shown). PBDE 47, PBDE 153, PBDE99 and PBDE 100 were detected in 4.5%, 3.5%, 2% and 1% of participants, respectively while PBB153 was not detected at all (Table A.1a). The median levels of the four tested PBDE congeners ranged between 1.0 and 1.5 μg/kg of plasma lipid. We found no significant difference among different regions for all tested PBDEs (Table A.2). 3.3. Serum concentrations of PFASs Among the measured PFASs, only six (PFOS, PFHxS, PFOA, PFNA, PFDA and PFUnA) were detected in all samples while PFBS, PFDS, PFOSA, PFPeA and PFHxA were below the detection limit in all samples (Table A.1a). There was significant regional difference for the perfluorosulfonated acids (PFSAs) such as PFOS, PFHxS and PFHpS with median levels of 10.15, 0.70 and 0.19 ng/mL, respectively in the pregnant women from all five regions (Fig. 3A). For perfluorocarboxylated acids (PFCAs), PFUnA, PFNA and PFDA differed significantly among regions (Fig. 3B). We observed significantly higher levels in the pregnant Inuit women from North and East compared to South and West regions for most
M. Long et al. / Science of the Total Environment 529 (2015) 198–212
203
Table 2 Food intake for pregnant women in Greenland.
Traditional fooda Marine mammals
Seabirds
Fish
Terrestrial species
Imported fooda
n−3/n−6
Region
North
Disko Bay
West
South
East
N
15
50
124
15
3
n Never Rare Frequently n Never Rare Frequently n Never Rare Frequently n Never Rare Frequently n Never Rare Frequently n Mean ± SD Median (min–max)
15 27.1% 63.8% 9.2% 14 41.9% 55.7% 2.4% 15 21.0% 71.2% 7.8% 14 31.8% 57.1% 11.1% 15 7.0% 49.4% 43.6% 15 0.26 ± 0.10 0.23 (0.17–0.55)
50 37.9% 56.8% 5.3% 47 39.3% 59.9% 0.7% 50 23.7% 69.0% 7.3% 50 38.3% 56.6% 5.0% 50 9.2% 43.5% 47.2% 50 0.24 ± 0.09 0.21 (0.11–0.51)
123 34.4% 62.5% 3.1% 124 21.9% 72.2% 5.9% 123 16.9% 74.3% 8.7% 123 29.5% 61.0% 9.5% 124 6.2% 47.5% 46.4% 124 0.24 ± 0.11 0.23 (0.08–0.87)
14 37.5% 59.2% 3.2% 14 31.5% 65.9% 2.6% 14 14.3% 80.6% 5.1% 14 18.9% 76.8% 4.3% 14 5.8% 39.8% 54.3% 15 0.23 ± 0.10 0.21 (0.09–0.55)
3 0.0% 87.5% 12.5% 3 0.0% 100% 0.0% 3 0.0% 73.8% 26.2% 3 0.0% 66.7% 33.3% 3 0.0% 40.6% 59.4% 3 0.30 ± 0.09 0.35 (0.20–0.35)
p value
All GRL town 207
0.799#
0.011#
0.559#
0.563#
0.617#
0.576*
205 34.3% 61.5% 4.2% 202 27.8% 68.1% 4.1% 205 18.3% 73.4% 8.3% 205 30.8% 60.8% 8.4% 206 6.8% 46.1% 47.1% 207 0.24 ± 0.10 0.22 (0.08–0.87)
N: total number of study participants in the region. n: total number of respondents having information for the corresponding parameters. n−3: n−3 polyunsaturated fatty acids; n−6: n−6 fatty acids. p value calculated with: *One-way ANOVA on ln-transformed data and # Chi square test. a Calculated as mean percent (%) from the FFQ for each food item and summed in categories (see Materials & methods and Supplementary material). For food items included in the Marine mammals, Seabirds, Fish, Terrestrial species and Imported food see Materials & methods.
PFASs (Fig. 3A, B). However, we did not observe significant regional difference for PFOA (Fig. 3B), PFHpA, PFDoA and PFTrA (Table A.3).
Heavy metals, Pb, Hg, As, Cd, Al and Cr were detected in 39%, 85%, 5%, 9%, 11.6% and 15.5% of the Greenlandic pregnant women, respectively, and the trace element Se was detected in all participants (Table A.1b). The overall median level of Hg in the pregnant women from all of the included Greenland towns was 4.23 μg/L and there were significant differences in blood Hg levels among the regions; women from the North and East regions had a significantly higher Hg level with a median level of 6.97 and 7.74 μg/L, respectively (Fig. 4A). It should be noted that in the North region the Hg level in Qaanaaq was approximately four times higher than in the rest of the included towns. We did not observe significant difference among the regions for Pb (Fig. 4A), As, Cd, Al and Cr (Table A. 4). The Se levels in the blood and plasma samples from all of the included Greenlandic pregnant women were 109 μg/L and 75.2 μg/L, respectively. A tendency, though not statistically significant, of higher levels of Se in the North and the East regions was observed (Fig. 4B, Table A.4). Again it should be noted that the whole blood Se level was much higher in Qaanaaq (North) than in the other towns (Fig. 4B).
explained 12.98% of the total variation in the original chemical concentrations. The third component mainly had high factor loading for PBDEs and was thus termed PBDEs component explaining 7.85% of the total variation in the original chemical concentrations. The three components together accounted for 62% of the total variance of chemical variables (Table 3). Other toxic heavy metals Pb, Cd, As, Al and Cr had very low factor loadings (b0.19) in the three principle components and their correlation with characteristics and lifestyle factors were thus assessed separately. As shown in Table 4, lipophilic POPs/Hg/Se component and blood Pb correlated significantly and positively with age and seafood intake biomarker (n − 3/n − 6 fatty acid ratio). The PFASs component correlated significantly and positively with n − 3/n − 6 fatty acid ratio. The PBDEs component did not correlate any of the analyzed lifestyle factors. Blood Cd and Pb correlated positively with plasma cotinine level, the smoking biomarker, with a correlation coefficient of 0.47 and 0.18, respectively (Table 4). Single POP congeners mostly had positive correlations with age and n − 3/n − 6 (Table A.5). We observed no significant differences between the self-reported smoking status and parity for all identified components. Women self-reporting as current smokers had significantly higher levels of blood Cd and Pb compared with non-smokers (p b 0.05, data not shown).
3.5. Searching for principle components of chemicals and correlation with lifestyle factors
3.6. Associations between PFASs and lipophilic POPs, and metals
When nine OCPs, 11 PCBs, four PBDEs, 10 PFAS and six toxic metals and trace element Se were used as input variables in PCA, three distinct components were identified (Table 3). The first and most distinct component mainly had high factor loadings for PCBs and OCPs and Hg and Se and thus represented lipophilic legacy POPs/Hg/Se (termed lipophilic legacy POPs/Hg/Se component). The lipophilic legacy POPs/Hg/Se component explained 40.82% of the total variation in the original chemical concentrations. The second identified component had a high factor loading for most of the PFASs and was termed PFASs component that
After adjustment for age and n−3/n−6, ∑PFAS was positively associated with ∑14 PCBs and ∑OCPs in the North, Disko Bay and West regions, while such an association was not observed for the South region except for ∑DL-PCB. We observed an inverse association between ∑PFAS and ∑PBDE in Disko Bay (Table 5a). The association between ∑PFSA and the heavy metal Hg and trace element Se was significantly positive in the North region. The trace element Se was positively associated with ∑PFAS in the West region and the data pooled for all regions (Table 5b).
3.4. Blood concentrations of trace element and heavy metals
204
M. Long et al. / Science of the Total Environment 529 (2015) 198–212
A. Polychlorinated biphenyls (PCBs) Indicator PCBs (µg/kg lipid) 210 180 150 120 90 60 30 0
Disko Bay
West
Sout
Qaanaaq Upernavik Uummannaq Illullisat Qasigisnnguit Qeqertarsuaq Aasiaat Sisimiut Maniitoq Nuuk Paamiut Nanortalik Narsaq Qaqortoq Tasiilaq
North
30 27 24 East 21 18 15 12 9 6 3 0
CB153
PCB180
North
PCB138
PCBs (µg/kg lipid) Indicator PCBs
PCB153 (median) PCB180 (median) PCB138 (median) Dioxin-like PCBs PCB105 (median) PCB118 (median) PCB156 (median)
B.
54 µg/kg lipid
Disko Bay
West
Sout
East
Qaanaaq Upernavik Uummannaq Illullisat Qasigisnnguit Qeqertarsuaq Aasiaat Sisimiut Maniitoq Nuuk Paamiut Nanortalik Narsaq Qaqortoq Tasiilaq
340 µg/kg lipid
Dioxin-like PCBs (µg/kg lipid)
PCB105
PCB118
PCB156
Region North Disko Bay West South East p value* All GRL towns n=15 n=45 n=118 n=13 n=3 100 64.0 50.5 57.0 220
Contents lists available at ScienceDirect
Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv
Food intake and serum persistent organic pollutants in the Greenlandic pregnant women: The ACCEPT sub-study Manhai Long a,⁎, Ane-Kersti Skaarup Knudsen a, Henning Sloth Pedersen b, Eva Cecilie Bonefeld-Jørgensen a a b
Centre for Arctic Health & Unit of Cellular and Molecular Toxicology, Department of Public Health, Aarhus University, Aarhus, Denmark Primary Health Care Center, Nuuk, Greenland
H I G H L I G H T S • • • • •
Pregnant women in East and North Greenland have higher intake of marine mammals Higher serum levels of PCBs, OCPs, PFASs and mercury in North and East Greenland Serum levels of PFASs were significantly associated with levels PCBs and OCPs Legacy POPs, PFOS and PFOA levels have decreased but other PFASs have sustained No significant regional differences for serum levels of PBDEs
a r t i c l e
i n f o
Article history: Received 22 December 2014 Received in revised form 6 May 2015 Accepted 6 May 2015 Available online xxxx Editor: Adrian Covaci Keywords: Contaminants Pregnant Diet Lifestyle Greenland
a b s t r a c t The Greenlandic Inuit have high blood concentrations of environmental persistent organic pollutants (POPs). High POP concentrations have been associated with age, smoking and consumption of marine mammals. Studies have indicated that exposure to POPs during pregnancy may adversely affect fetal and child development. To assess geographical differences in diet, lifestyle and environmental contaminant exposure among pregnant women in Greenland, blood samples and questionnaire data were collected from 207 pregnant women in five Greenlandic regions (North, Disco Bay, West, South and East). Blood samples were analyzed for 11 organochlorine pesticides (OCPs), 14 polychlorinated biphenyls (PCBs), 5 polybrominated diphenyl ethers (PBDEs), 15 perfluoroalkylated substances (PFASs) and 63 metals. A trend of higher intake of marine mammals in the East and North regions was reflected by a higher n−3/n−6 fatty acid ratio. Participants in the East region tended also to have higher intake of terrestrial species. A significant higher seabird intake was seen for pregnant women in the West region. Significant regional differences were found for blood concentrations of PCBs, OCPs, PFASs and mercury, with higher levels in the North and East regions. PFASs were significantly associated with PCBs and OCPs in most of the regions. In the North region, PFASs were associated with both selenium and mercury. No significant regional difference was observed for PBDEs. The regional differences of blood levels of POPs and mercury were related to differences in intake of the traditional food. Compared to earlier reports, decreased levels of legacy POPs, Hg and Pb and perfluorooctane sulfonate and perfluorooctanoic acid were observed, but the levels of PFAS congeners perfluorohexane sulfonate and perfluorononanoic acid were sustained. The detection of POPs and heavy metals in maternal blood indicates fetal exposure to these compounds possibly influencing fetal development. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Lipophilic persistent organic pollutants (POPs) such as organochlorine pesticides (OCPs), polychlorinated biphenyls (PCBs), and polybrominated diphenyl ethers (PBDEs) are ubiquitous and found ⁎ Corresponding author at: Centre for Arctic Health & Unit of Cellular and Molecular Toxicology, Department of Public Health, Aarhus University, Bartholins Allé 2, Building 1260, 8000 Aarhus C, Denmark. E-mail addresses: [email protected] (M. Long), [email protected] (A.-K.S. Knudsen), [email protected] (H.S. Pedersen), [email protected] (E.C. Bonefeld-Jørgensen).
http://dx.doi.org/10.1016/j.scitotenv.2015.05.022 0048-9697/© 2015 Elsevier B.V. All rights reserved.
globally due to long-range transport by atmospheric and oceanic currents. Thus, these environmental contaminants are also brought to the Arctic where they are trapped by the cold climate (Barrie et al., 1992; Burkow and Kallenborn, 2000; Macdonal et al., 2000). The lipophilic POPs are accumulated and biomagnified in lipid-based Arctic food webs (Borga et al., 2004; Kelly et al., 2008; Laender et al., 2010) causing international scientific concern. In 1994, all circumpolar countries agreed to address these concerns by establishing a health monitoring program for contaminants called AMAP (Arctic Monitoring and Assessment Program). The systematic studies concerning human exposure to
M. Long et al. / Science of the Total Environment 529 (2015) 198–212
POPs and possible health effects in the Arctic population have provided knowledge and new perspectives to the environmental agenda in the last 20 years. Arctic populations such as the Greenlandic Inuit display a higher body burden of POPs than people living in industrialized regions close to major emission sources. This is largely due to Arctic populations that rely on their traditional food, including species at the top of the marine food chain (AMAP, 2009; Bjerregaard et al., 2001, 2013; Bonefeld-Jorgensen, 2010). The levels of these lipophilic POPs in Arctic populations correlate with age, smoking and the plasma level of n −3 polyunsaturated fatty acids; the latter is a strong indicator of the main source of POP contamination in their traditional marine food (Bonefeld-Jorgensen, 2010; Carriere et al., 2012; Deutch et al., 2003, 2007a; Krüger et al., 2007). In general, the levels of many legacy-lipophilic POPs such as PCBs and OCP are decreasing in Arctic populations; this is supposedly due to a combination of regulation of the legacy POP causing reduced concentrations of the chemicals in the marine food web and a reduction in the consumption of traditional food from marine mammals such as whales, walrus and seals (Bjerregaard et al., 2013; Deutch et al., 2007a). Even though the significant downward trend of legacy-lipophilic POPs shows that the legislation works, it must be kept in mind that some POPs like perfluoroalkylated substances (PFASs) and PBDEs are still used and bioaccumulate in marine mammals and humans. PFASs are a large group of chemicals used since the 1950s in industry and in commercial products (e.g. Teflon, carpets, furniture, foodstuff packing). PFASs are persistent with half-lives in humans between 4 and 10 years (Olsen et al., 2008). Unlike the legacy-lipophilic POPs (PCBs and OCPs), PFASs bind to blood proteins and are stored mainly in the liver, kidneys, and biliary system (Bonefeld-Jorgensen et al., 2014; Long et al., 2012). PFASs are found globally, and governmental regulations on use and production of specific compounds such as perfluorooctane sulfonate (PFOS) and perfluorooctanoic acid (PFOA) have been made in both USA and Europe. The use of PFOS was restricted in USA in 2001 (Paul et al., 2009) and Europe from June 2008 (European Parliament, 2006) and PFOS has been added to Annex B of the Stockholm Convention on POPs (Stockholm Convention, 2009). USA launched the “PFOA Stewardship Program” where eight of the major PFOA producing companies committed to reduce emissions of PFOA and related chemicals by 95% by 2010 (US EPA, 2006). The classification and labeling of PFOA are currently discussed in the EU (Posner et al., 2013). Even though the concentrations of legacy-lipophilic POPs are decreasing, the long-term effects of POPs on general public health status in Greenland are of concern (Bonefeld-Jorgensen, 2010; Bonefeld-Jorgensen et al., 2014). The adverse effects of these chemicals on human health have been connected to disruption of the immune system, reduced reproductive ability, developmental and behavioral disabilities, cancer, and metabolic disorders (Bonefeld-Jorgensen et al., 2014). Biomarker studies have indicated that exposure to POPs can cause genetic alterations, decrease defense against oxidative stress, and disrupt endocrine systems (AMAP, 2009; Bonefeld-Jorgensen et al., 2014). Human studies found neurological impairment and immunological effects in the developing fetus and infants exposed to POPs (Chen et al., 2014; Heilmann et al., 2006; Van Oostdam et al., 2004), PFASs (Fei et al., 2007; Grandjean et al., 2012) and heavy metals (Claus Henn et al., 2014; Taylor et al., 2014). Since fetal development is the most sensitive period in life, lifestyle, diet and environmental exposure to POPs during pregnancy are important factors when examining and monitoring the health outcome of the next generation. The data presented in this study is a sub-study of the current establishment of a new Greenlandic birth cohort ACCEPT (Adaptation to Climate Change, Environmental Pollution, and Dietary Transition). The aim of the present study was to investigate lifestyle characteristics, dietary habits and levels of environmental contaminants in a geographically representative population of pregnant women in Greenland between 2010–11 and 2013.
199
2. Materials & methods 2.1. Study population We conducted this population based cross-sectional study in two steps: Between August 2010 and September 2011 a doctor employed at the Queen Ingrid Health Care Center in Nuuk and between June 2013 and November 2013 midwives at the Greenlandic hospitals recruited participants for this study. Data collection was carried out in six different towns at the West coast in Greenland (Fig. 1) and included a study population of pregnant women (N18 year); 212 pregnant women were enrolled in the present study. The enrolment depended on medical coast visit and workload in the hospitals. Among the asked pregnant women there were only few refusals caused by movement, fear of needles and being abroad. Five women were defined as nonInuit and excluded from the study. The final study population consisted of 207 pregnant women who all had lived more than 50% of their life in Greenland. The participant percentage of the total births and the participant percentage of possible inclusion of the population during 2010–2011 was 11.2 and 22.2%, respectively. Among the participants, approximately 84% had completely Inuit descent defined as both parents being born in Greenland. The remaining 16% were partly-Inuit and had one parent from Greenland and/or had Greenlandic as their native language. Each included pregnant women filled in a culturally-appropriate lifestyle questionnaire and a specially designed food frequency questionnaire (FFQ) for Inuit. Both questionnaires were available in Danish and Greenlandic. Data on pre-pregnancy height (in meters) and weight (in kilograms), reproductive factors and smoking status (current/noncurrent) of the participants were obtained from medical records. We applied the information from the lifestyle questionnaire about where the women were born or grew up (town) and for how long they had lived there (years). The women belonged to the region where they had lived N 50% of their lives. The FFQ included approximately 40 traditional Greenlandic food items and 23 imported food items. Traditional food items were grouped into four main groups (marine mammals, seabirds, fish and terrestrial species). The marine mammal group included polar bear, walrus, seals, and whales. The seabird group included guillemot, eider, and kittiwake. The fish group included trout, cod, Greenlandic halibut, Atlantic halibut, redfish, Atlantic salmon, and capelin. The terrestrial species included caribou, muskox, Greenlandic lamb, hare, and ptarmigan. The imported food group included 23 imported food items e.g. meat products including non-Greenlandic fish, fast food, sauce, pasta, rice, vegetables, fruit, chocolate, sweets, and snacks. For each food item, the frequency of intake was reported in three categories: Never; Rare (less than once a month, approximately once a month, 2–3 times a month) and Frequently (once a week, 2–3– 4 times a week, every day or almost every day, several times a day). The number of answers in the three categories of each food item was calculated as percentages, then each percentage was added up and divided by the number of food items in the corresponding main traditional food groups and imported food group. A detailed calculation of food intake frequency of the main food groups is shown in Appendix A. Samples of venous blood were collected from the pregnant women at inclusion into the study (mean gestational age 25.1 week, range 7–40 weeks); 7.2% of blood samples were taken in the first trimester (gestational weeks 7–12), 57% during the second trimester (gestational weeks 13–28) and 36.2% in the third trimester (gestational weeks 29–40). The blood samples were stored at − 80 °C until analyses. Informed consent was obtained from all participants prior to data collection. The study was approved by the Ethical Committee for Scientific Investigations in Greenland (KVUG) and conducted in accordance with the Helsinki Declaration.
200
M. Long et al. / Science of the Total Environment 529 (2015) 198–212
Fig. 1. Map of Greenland with the collection sites and regions where the participants lived more than 50% of their life.
2.2. Determination of contaminants 2.2.1. Lipophilic POPs Serum samples were analyzed for 30 lipophilic POPs and lipid content at Le Centre de Toxicologie du Québec (Sainte-Foy, Québec, Canada), a certified laboratory by the Canadian Association for Environmental Analytical Laboratories (Butler Walker et al., 2003). Fourteen polychlorinated biphenyl (PCB) congeners [PCB 28, 52, 99, 101, 105, 118, 128, 138, 153, 156, 170, 180, 183, 187], five flame retardants including one polybrominated biphenyl [PBB153] and four polybrominated diphenyl ethers (PBDEs) [PBDE 47, 99, 100, 153], and 11 organochlorine pesticides (OCPs) [p,p′-DDT (dichlorodiphenyltrichloroethane) and its major metabolite p,p′-DDE (dichlorodiphenyldichloroethylene), aldrin, mirex, β-hexachlorocyclohexane (β-HCH), hexachlorobenzene (HCB), cis- and trans-nonachlor, and α-, γ- and oxy-chlordane] were analyzed in purified extracts by high-resolution gas chromatography with electron capture detection. All determined PCBs, OCPs and PBDEs were adjusted to the plasma lipid content analyzed in the same sample (Butler Walker et al., 2003; Deutch et al., 2007b) and reported as μg/kg plasma lipid. If the value was below the detection limit, we assigned the value given by the detection limit divided by two.
2.2.2. Perfluoroalkylated substances (PFASs) Fifteen PFASs were analyzed in serum samples using liquid chromatography–tandem mass spectrometry (LC–MS–MS) with electrospray ionization (ESI) (Long et al., 2012): perfluorooctane sulfonate (PFOS), perfluorohexane sulfonate (PFHxS), perfluoroheptanesulfonate (PFHpS), perfluorooctanoic acid (PFOA), perfluorononanoic acid (PFNA), perfluoroheptanoic acid (PFHpA), perfluorodecanoic acid (PFDA), perfluoroundecanoic acid (PFUnA), perfluorododecanoic acid (PFDoA), perfluorotridecanoic acid (PFTrA), perfluorobutanesulfonic acid (PFBS), perfuoro-1-decanesulfonate (PFDS), perfluorooctane sulfonamide (PFOSA), perfluoropentanoic acid (PFPeA) and perfluoroheptanoic acid (PFHpA). The PFAS method performance is continuously controlled and tested in a Quality Assurance/Quality Control (QA/QC) by participating in inter-laboratory comparison studies performed by Institute Nationale de Santé Publique du Québec for AMAP (AMAP, 2000–2010). If the value was below the detection limit, we assigned the value given by the detection limit divided by two. 2.2.3. Blood metals Sixty-three blood metals including essential elements selenium (Se), zinc (Zn), iron (Fe) and copper (Cu) and heavy metals including lead (Pb), mercury (Hg), arsenic (As) and cadmium (Cd) were measured using Inductively Coupled Plasma Mass Spectrometry (ICP-MS) after
M. Long et al. / Science of the Total Environment 529 (2015) 198–212
digesting blood with nitric acid in the microwave at the accredited element laboratory, Institute for Bioscience-Arctic Research Centre, Aarhus University, Denmark. The quality was ensured by repeated analyses and by frequent analysis of certified reference material (Seronorm), as well as by participation in the Quality Assurance of Information in Marine Environmental monitoring (QUASIMEME) inter-laboratory comparison program (Asmund et al., 2004). In this paper, we only reported the relevant toxic metals, Hg, Pb, Cd, As, Al, and Cr and trace element Selenium (Se). 2.2.4. Plasma fatty acids The ratio between n−3 polyunsaturated fatty acids and n−6 fatty acids is known to be a strong indicator of marine food intake and thus a good indicator of the relative consumption of traditional versus imported food (Deutch et al., 2004; Tjonneland et al., 1993). Plasma fatty acids were determined by capillary gas–liquid chromatography at the Biology Department, University of Guelph, Canada (Cote et al., 2004). The fatty acid composition of plasma phospholipids was expressed as a percentage of the total area of all fatty acid peaks from 14:0 to 24:0. Plasma phospholipids of fatty acids correspond to relative percentage of total fatty acids by weight. We reported n−3 polyunsaturated fatty acids on the sum of C18:3, n−3; C18:4, n−3; C20:3, n−3; C20:4, n−3; C20:5, n−3; C22:5, n−3; and C22:6, n−3, and the n−6 fatty acids as the sum of C18:2, n−6; C18:3, n−6; C20:2, n−6; C20:3, n−6; C20:4, n−6; C22:2, n−6; C22:4, n−6 and C22:5, n−6. 2.2.5. Determination of plasma cotinine Cotinine is a metabolite of nicotine. The level of cotinine in the blood is proportionate to the amount of exposure to tobacco smoke and is thus a valuable indicator of recent exposure to tobacco smoke. The Calbiotech Cotinine Direct ELISA Kit was used to measure plasma cotinine (Calbiotech Inc., CA, USA) at the Unit of Cellular & Molecular Toxicology, Centre of Arctic Health, Department of Public Health, Aarhus University, Denmark. This assay is a solid phase competitive ELISA and the absorbance was read on ELISA Reader at 450 nm. The plasma cotinine concentration was expressed as ng/mL and the detection limit was 1 ng/mL. If the value was below the detection limit, we assigned a value given by the detection limit divided by two. 2.3. Statistical analysis Lifestyle and FFQ data were double-entered in the validation program EpiData file version 3.1.2701.2008. Data on lifestyle, diet and chemicals were analyzed in SPSS Statistics version 19.0 & 21 (SPSS Inc, Chicago, IL, USA) with a significance level of 5%. Based on the chemical structure, we grouped the analyzed PFASs in perfluorosulfonated acids (∑ PFSA, sum of PFBS, PFHxS, PFHpS, PFOS, PFDS and PFOSA) and perfluorocarboxylated acids (∑ PFCA, sum of PFPeA, PFHxA, PFHpA, PFOA, PFNA, PFDA, PFUnA, PFDoA, PFTrA, and PFTeA). Because of high legacy POP inter-correlation, PCB congeners were also grouped as the sum of 14 PCB congeners (Σ 14 PCB) and the sum of 11 pesticides (Σ OCP). Among tested PCBs, PCB 105, PCB118 and PCB156 are dioxin-like PCB (DL-PCB) (Van den Berg et al., 2006) and their summation was grouped as Σ DL-PCB. ∑ PBDE was the sum of four PBDE congeners and PBB153. In addition, analyses were also performed for the single compounds detected in the samples. The distribution of data was checked by Q–Q plots. The natural logarithmic transformed variables made the distribution more symmetrical and improved the normality of data; thus, the analysis was performed on the ln-transformed data. The data were treated as continuous and categorical variables. We used One-way ANOVA to compare the graphical continuous variables and chemical levels among the regions. When ANOVA showed statistical significance, complementary multiple comparison ad hoc tests was performed. Test for equal variances of variables was performed using Levene's test. The least-significant difference (LSD) test was used if equal variance was shown, otherwise Dunett T3 was used. Previous studies reported a correlation between POPs and age and the
201
n−3/n−6 ratio (Deutch et al., 2007b; Fromme et al., 2007; Haug et al., 2009). Therefore comparisons of means for POPs among regions were also performed by the General Linear Model procedure adjusted for age and n−3/n−6 ratio. Pearson's chi-square test was used to check the difference among regions for category variables. Due to a large number of analytes, we used data reduction technique Principal Component Analysis (PCA) to identify potential underlying components of POPs and metal levels in the blood. The dimension of the data was reduced with PCA by forming a few linear combinations of the original variables. By PCA the correlated variables were grouped together. The coefficients defining these linear combinations, called “factor loadings”, are the correlations of each input variable with the component. The POPs and metals which were detectable in more than 1% of all blood samples were included into the model: 9 OCPs, 11 PCBs, 4 PBDEs, 10 PFASs, six toxic metals (Hg, Pb, Cd, As, Al and Cr) and trace element selenium (Se). The trace element Selenium (Se) was included together with the POPs and heavy metals because Greenlandic traditional food has a high content of Se; Greenlanders thus had a high burden of both POPs, Hg and Se (Hansen et al., 2004). The number of components was extracted on the basis of eigenvalues (N 1), and scree plot. A varimax rotation was used to obtain a set of independent and best interpretable components (Newby and Tucker, 2004). The Kaiser–Mayer–Olkin (KMO) test (0.901) and the Bartlett Test of Sphericity (p b 0.0001) were used to test suitability of the analysis. All factor loadings N|0.3| were used to identify the variables comprising a component. Factor scores for each participant were created by multiplying factor loadings with the corresponding standardized value for each variable and summing across variables. Therefore the factor scores indicate the extent to which the chemical concentrations conformed to the corresponding component. We performed a Spearman's correlation analysis to assess the bivariate correlation of the identified components, single POP congeners, blood level of Hg, Pb, Cd, As, Al, Cr and Se with characteristics and lifestyle factors such as age, body mass index (BMI), smoking marker plasma cotinine and seafood intake biomarker n−3/n−6 ratio. The nonparametric Mann–Whitney U test was used to test the differences between smoking status for identified components, chemicals and cotinine level, while the difference of factor scores and chemicals among parity were tested using the Kruskal–Wallis and Jonckheere–Terpstra tests. We used multiple linear regression analysis to assess the relationship between serum PFAS levels and serum levels of legacy lipophilic POPs as well as blood Se and Hg levels adjusting for age, n − 3/n − 6 ratio in each region. For the pooled regional data, the association was assessed adjusting for age, n−3/n−6 and region. 3. Results 3.1. Lifestyle and diet The study population included 207 pregnant women (18–44 years) with a median age of 27.0 years and median pre-pregnancy BMI of 24.2 kg/m2; 60% of the participants lived in the West region (Table 1). We observed similar population characteristics across the regions. Almost half of the women (44.2%) were current smokers at the beginning of their pregnancy (approx. at gestational week 12). We did not find a significant difference in self-reported smoking status across the regions. Current smokers had a much higher plasma cotinine level than non-smokers (median: 75.7 ng/ml vs. 0.5 ng/ml, p b 0.0001), supporting plasma cotinine being a surrogate marker of current smoking. No significant difference of plasma cotinine level among regions was observed. A tendency of higher smoking rate and plasma cotinine level was found in participants in the North (Table 1). Parity, defined as a full-term pregnancy, was evenly distributed in the regions West, Disko Bay and South concerning one birth or equal to or more than two births. The three women from the East region had two or more births. In the regions North, Disko Bay, West and
202
M. Long et al. / Science of the Total Environment 529 (2015) 198–212
Table 1 Characteristics of the study population: Pregnant women in Greenland.
Age (years)
BMI b
Smokingc P-cotinine (ng/ml)d
Alcohol consumption Before pregnancy
During pregnancy Reproductive factors Parity
Regiona
North
Disko Bay
West
South
East
p value
N (%) n Mean ± SD Median (min–max) n Mean ± SD Median (min–max) n Yes (%) n Mean ± SD Median (min–max)
15 (7.2) 15 28.3 ± 3.9 27.0 (21–35) 13 24.1 ± 2.9 23.8 (20.1–29.1) 14 9 (64.3) 15 48.99 ± 55.57 15.2 (0.50–160.0)
50 (24.2) 50 27.7 ± 5.3 27.0 (19–40) 49 25.9 ± 5.2 24.8 (17.7–39.3) 50 25 (50.0) 50 47.61 ± 61.13 0.74 (0.50–184.0)
124 (59.9) 124 27.1 ± 5.3 27.0 (18–44) 118 25.3 ± 5.0 24.4 (16.4–46.6) 124 51 (41.1) 124 30.52 ± 47.41 0.50 (0.50–190.5)
15 (7.2) 15 28.3 ± 5.1 28.0 (22–41) 15 23.0 ± 3.6 22.6 (17.5–33.7) 15 6 (40.0) 15 33.02 ± 48.80 0.50 (0.50–126.1)
3 (1.4) 3 33.7 ± 2.9 32.0 (32–37) 3 27.5 ± 3.3 28.9 (23.7–29.7) 3 0 (0.0) 3 2.83 ± 4.04 0.50 (0.50–7.50)
b0.0001# 207 (100) 0.230* 207 27.5 ± 5.2 27.0 (18–44) 0.226* 198 25.2 ± 4.9 24.2 (16.4–46.6) 0.204# 206 91 (44.2) 0.565* 207 35.78 ± 51.80 0.50 (0.50–190.5)
All GRL towns
0.256# n ≤1 a month (%) N2–3 times a month (%) n ≤1 a month (%) N2–3 times a month (%)
15 6 (40.0) 9 (60.0) 14 13 (92.8) 1 (7.1)
47 28 (59.6) 19 (40.4) 44 44 (100) 0 (0.0)
121 83 (68.6) 38 (31.4) 117 117 (100) 0 (0.0)
15 10 (66.7) 5 (33.3) 12 12 (100) 0 (0.0)
3 2 (66.7) 1 (33.3) 3 0 (0.0) 0 (0.0)
n Mean ± SD Median (min–max) 0 (%) 1 (%) ≥2 (%)
15 1.7 ± 1.5 1.0 (0–5) 4 (26.7) 4 (26.7) 7 (46.7)
50 1.2 ± 1.2 1.0 (0–4) 17 (34.0) 15 (30.0) 18 (36.0)
124 0.98 ± 1.2 1.0 (0–6) 52 (41.9) 42 (33.9) 30 (24.2)
15 0.93 ± 1.1 1.0 (0–4) 6 (40.0) 6 (40.0) 3 (20.0)
3 2.3 ± 0.6 2.0 (2–3) 0 (0.0) 0 (0.0) 3 (100)
0.013#
0.117#
201 129 (64.2) 72 (35.8) 190 189 (99,45) 1 (0.5) 207 1.1 ± 1.2 1.0 (0–6) 79 (38.2) 67 (32.4) 61 (29.5)
N: total number of participants in the region. n: total number of respondents having information for the corresponding parameters. GRL: Greenland, 15 towns are included. p value calculated with *One-way ANOVA analysis on ln-transformed data, and #Chi square test. a Region: Where the women have lived N50% of their life, given as following regions: North: Upernavik (8), Uummannaq (4), Qaanaaq (3); Disko Bay: Ilulissat (26), Aasiaat (18), Qeqertarsuaq (3), Qasigiannguit (3); West: Nuuk (59), Maniitsoq (49), Sisimiut (11), Paamiut (5); South: Qaqortoq (8), Nanortalik (5), Narsaq (2); East: Tasiilaq (3). b BMI: Body mass index, calculated pre-pregnancy as kg/m2. c Current smoking: the smoking status is obtained from the pregnancy journal. The women were asked about current and non-current smoking during their early pregnancy. The Yespercent is calculated from the total answers in the region-group. d P-cotinine: plasma cotinine is a metabolite of nicotine and thus a primary biomarker for the determination of tobacco exposure.
South, the first-time mothers accounted for 26.7%, 34.0%, 41.9% and 40.0%, respectively (Table 1). Table 2 shows the food intake of the pregnant women in the different regions. Across the regions the only significant difference observed was for seabirds (p = 0.011) where the highest frequent intake was found in the West region and the lowest in the Disko Bay region and East (Table 2). There was a tendency, though not statistically significant, to higher intake of marine mammals in the North and East regions; the frequency of intake of fish and terrestrial species was higher for the pregnant women in the East region, indicating a preference for traditional Inuit diet. The n−3/n−6 ratio, indicator of seafood intake, also supported a higher marine food intake for the women in the East and North regions (Table 2). 3.2. Serum concentrations of lipophilic POPs Most lipophlic POPs were detectable in the blood samples. However, PCB 28, PCB52, PBB 153, aldrin and γ-chlordane were only detected in less than 1% of the samples (Table A.1a). Among 14 tested PCB congeners, three PCBs (PCB 138, 153, 180) showed higher concentration ranges (Table A.2, Fig. 2A). As shown in Fig. 2A, the median of PCB153 level was 100, 64, 51, 57 and 220 μg/kg plasma lipid for the pregnant women in the North, Disko Bay, West, South and East regions, respectively. The median of PCB 153 for all the included Greenlandic towns was 57 μg/kg plasma lipid. We found a significant difference among regions for the serum levels of indicator PCBs, PCB 138, PCB 153 and PCB 180 and the dioxin-like PCB (DL-PCB) such as 105 and PCB 118, although the PCB156 level did not differ among the regions (Fig. 2A). Pregnant women living in the East (Tasiilaq, n = 3) and the North regions (n = 15) had higher levels of PCBs compared to the other regions
(Fig. 2A). A similar pattern was observed for the sum of DL-PCB (PCB 105, PCB118 and PCB156) and the sum of 14 PCBs (Table A.2). Among the tested organochlorine pesticides (OCPs), p,p′-DDE showed the highest concentration in participants from all regions with a median of 130 μg/kg plasma lipid for all the included Greenland towns (Fig. 2B, Table A.2). Similarly to the PCB levels, higher concentrations of OCPs such as p,p′-DDE, oxychlordane, trans-nonachlor, HCB and β-HCH were found in pregnant women living for the longest period in the East and North regions (Fig. 2B, Table A.2). After adjustment for potential confounders such as age and seafood intake marker (n − 3/ n−6), significant regional differences still remained (data not shown). PBDE 47, PBDE 153, PBDE99 and PBDE 100 were detected in 4.5%, 3.5%, 2% and 1% of participants, respectively while PBB153 was not detected at all (Table A.1a). The median levels of the four tested PBDE congeners ranged between 1.0 and 1.5 μg/kg of plasma lipid. We found no significant difference among different regions for all tested PBDEs (Table A.2). 3.3. Serum concentrations of PFASs Among the measured PFASs, only six (PFOS, PFHxS, PFOA, PFNA, PFDA and PFUnA) were detected in all samples while PFBS, PFDS, PFOSA, PFPeA and PFHxA were below the detection limit in all samples (Table A.1a). There was significant regional difference for the perfluorosulfonated acids (PFSAs) such as PFOS, PFHxS and PFHpS with median levels of 10.15, 0.70 and 0.19 ng/mL, respectively in the pregnant women from all five regions (Fig. 3A). For perfluorocarboxylated acids (PFCAs), PFUnA, PFNA and PFDA differed significantly among regions (Fig. 3B). We observed significantly higher levels in the pregnant Inuit women from North and East compared to South and West regions for most
M. Long et al. / Science of the Total Environment 529 (2015) 198–212
203
Table 2 Food intake for pregnant women in Greenland.
Traditional fooda Marine mammals
Seabirds
Fish
Terrestrial species
Imported fooda
n−3/n−6
Region
North
Disko Bay
West
South
East
N
15
50
124
15
3
n Never Rare Frequently n Never Rare Frequently n Never Rare Frequently n Never Rare Frequently n Never Rare Frequently n Mean ± SD Median (min–max)
15 27.1% 63.8% 9.2% 14 41.9% 55.7% 2.4% 15 21.0% 71.2% 7.8% 14 31.8% 57.1% 11.1% 15 7.0% 49.4% 43.6% 15 0.26 ± 0.10 0.23 (0.17–0.55)
50 37.9% 56.8% 5.3% 47 39.3% 59.9% 0.7% 50 23.7% 69.0% 7.3% 50 38.3% 56.6% 5.0% 50 9.2% 43.5% 47.2% 50 0.24 ± 0.09 0.21 (0.11–0.51)
123 34.4% 62.5% 3.1% 124 21.9% 72.2% 5.9% 123 16.9% 74.3% 8.7% 123 29.5% 61.0% 9.5% 124 6.2% 47.5% 46.4% 124 0.24 ± 0.11 0.23 (0.08–0.87)
14 37.5% 59.2% 3.2% 14 31.5% 65.9% 2.6% 14 14.3% 80.6% 5.1% 14 18.9% 76.8% 4.3% 14 5.8% 39.8% 54.3% 15 0.23 ± 0.10 0.21 (0.09–0.55)
3 0.0% 87.5% 12.5% 3 0.0% 100% 0.0% 3 0.0% 73.8% 26.2% 3 0.0% 66.7% 33.3% 3 0.0% 40.6% 59.4% 3 0.30 ± 0.09 0.35 (0.20–0.35)
p value
All GRL town 207
0.799#
0.011#
0.559#
0.563#
0.617#
0.576*
205 34.3% 61.5% 4.2% 202 27.8% 68.1% 4.1% 205 18.3% 73.4% 8.3% 205 30.8% 60.8% 8.4% 206 6.8% 46.1% 47.1% 207 0.24 ± 0.10 0.22 (0.08–0.87)
N: total number of study participants in the region. n: total number of respondents having information for the corresponding parameters. n−3: n−3 polyunsaturated fatty acids; n−6: n−6 fatty acids. p value calculated with: *One-way ANOVA on ln-transformed data and # Chi square test. a Calculated as mean percent (%) from the FFQ for each food item and summed in categories (see Materials & methods and Supplementary material). For food items included in the Marine mammals, Seabirds, Fish, Terrestrial species and Imported food see Materials & methods.
PFASs (Fig. 3A, B). However, we did not observe significant regional difference for PFOA (Fig. 3B), PFHpA, PFDoA and PFTrA (Table A.3).
Heavy metals, Pb, Hg, As, Cd, Al and Cr were detected in 39%, 85%, 5%, 9%, 11.6% and 15.5% of the Greenlandic pregnant women, respectively, and the trace element Se was detected in all participants (Table A.1b). The overall median level of Hg in the pregnant women from all of the included Greenland towns was 4.23 μg/L and there were significant differences in blood Hg levels among the regions; women from the North and East regions had a significantly higher Hg level with a median level of 6.97 and 7.74 μg/L, respectively (Fig. 4A). It should be noted that in the North region the Hg level in Qaanaaq was approximately four times higher than in the rest of the included towns. We did not observe significant difference among the regions for Pb (Fig. 4A), As, Cd, Al and Cr (Table A. 4). The Se levels in the blood and plasma samples from all of the included Greenlandic pregnant women were 109 μg/L and 75.2 μg/L, respectively. A tendency, though not statistically significant, of higher levels of Se in the North and the East regions was observed (Fig. 4B, Table A.4). Again it should be noted that the whole blood Se level was much higher in Qaanaaq (North) than in the other towns (Fig. 4B).
explained 12.98% of the total variation in the original chemical concentrations. The third component mainly had high factor loading for PBDEs and was thus termed PBDEs component explaining 7.85% of the total variation in the original chemical concentrations. The three components together accounted for 62% of the total variance of chemical variables (Table 3). Other toxic heavy metals Pb, Cd, As, Al and Cr had very low factor loadings (b0.19) in the three principle components and their correlation with characteristics and lifestyle factors were thus assessed separately. As shown in Table 4, lipophilic POPs/Hg/Se component and blood Pb correlated significantly and positively with age and seafood intake biomarker (n − 3/n − 6 fatty acid ratio). The PFASs component correlated significantly and positively with n − 3/n − 6 fatty acid ratio. The PBDEs component did not correlate any of the analyzed lifestyle factors. Blood Cd and Pb correlated positively with plasma cotinine level, the smoking biomarker, with a correlation coefficient of 0.47 and 0.18, respectively (Table 4). Single POP congeners mostly had positive correlations with age and n − 3/n − 6 (Table A.5). We observed no significant differences between the self-reported smoking status and parity for all identified components. Women self-reporting as current smokers had significantly higher levels of blood Cd and Pb compared with non-smokers (p b 0.05, data not shown).
3.5. Searching for principle components of chemicals and correlation with lifestyle factors
3.6. Associations between PFASs and lipophilic POPs, and metals
When nine OCPs, 11 PCBs, four PBDEs, 10 PFAS and six toxic metals and trace element Se were used as input variables in PCA, three distinct components were identified (Table 3). The first and most distinct component mainly had high factor loadings for PCBs and OCPs and Hg and Se and thus represented lipophilic legacy POPs/Hg/Se (termed lipophilic legacy POPs/Hg/Se component). The lipophilic legacy POPs/Hg/Se component explained 40.82% of the total variation in the original chemical concentrations. The second identified component had a high factor loading for most of the PFASs and was termed PFASs component that
After adjustment for age and n−3/n−6, ∑PFAS was positively associated with ∑14 PCBs and ∑OCPs in the North, Disko Bay and West regions, while such an association was not observed for the South region except for ∑DL-PCB. We observed an inverse association between ∑PFAS and ∑PBDE in Disko Bay (Table 5a). The association between ∑PFSA and the heavy metal Hg and trace element Se was significantly positive in the North region. The trace element Se was positively associated with ∑PFAS in the West region and the data pooled for all regions (Table 5b).
3.4. Blood concentrations of trace element and heavy metals
204
M. Long et al. / Science of the Total Environment 529 (2015) 198–212
A. Polychlorinated biphenyls (PCBs) Indicator PCBs (µg/kg lipid) 210 180 150 120 90 60 30 0
Disko Bay
West
Sout
Qaanaaq Upernavik Uummannaq Illullisat Qasigisnnguit Qeqertarsuaq Aasiaat Sisimiut Maniitoq Nuuk Paamiut Nanortalik Narsaq Qaqortoq Tasiilaq
North
30 27 24 East 21 18 15 12 9 6 3 0
CB153
PCB180
North
PCB138
PCBs (µg/kg lipid) Indicator PCBs
PCB153 (median) PCB180 (median) PCB138 (median) Dioxin-like PCBs PCB105 (median) PCB118 (median) PCB156 (median)
B.
54 µg/kg lipid
Disko Bay
West
Sout
East
Qaanaaq Upernavik Uummannaq Illullisat Qasigisnnguit Qeqertarsuaq Aasiaat Sisimiut Maniitoq Nuuk Paamiut Nanortalik Narsaq Qaqortoq Tasiilaq
340 µg/kg lipid
Dioxin-like PCBs (µg/kg lipid)
PCB105
PCB118
PCB156
Region North Disko Bay West South East p value* All GRL towns n=15 n=45 n=118 n=13 n=3 100 64.0 50.5 57.0 220
