Marine Pollution Bulletin 75 (2013) 250–258

Contents lists available at SciVerse ScienceDirect

Marine Pollution Bulletin

journal homepage: www.elsevier .com/locate /marpolbul

Baseline

Edited by Bruce J. Richardson

The objective of BASELINE is to publish short communications on different aspects of pollution of the marine environment. Only thosepapers which clearly identify the quality of the data will be considered for publication. Contributors to Baseline should refer to‘Baseline—The New Format and Content’ (Mar. Pollut. Bull. 60, 1–2).

A baseline study on levels of polychlorinated dibenzo-p-dioxins,polychlorinated dibenzofurans, non-ortho and mono-ortho PCBs,non-dioxin-like PCBs and polybrominated diphenyl ethers inNortheast Arctic cod (Gadus morhua) from different parts of theBarents Sea

0025-326X/$ - see front matter � 2013 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.marpolbul.2013.07.017

⇑ Corresponding author. Tel.: +47 99 48 77 01; fax: +47 55 90 52 99.E-mail address: kaare.julshamn@nifes.no (K. Julshamn).

Kaare Julshamn a,⇑, Arne Duinker a, Marc Berntssen a, Bente M. Nilsen a, Sylvia Frantzen a, Kjell Nedreaas b,Amund Maage a

a National Institute of Nutrition and Seafood Research (NIFES), P.O. Box 2029, Nordnes, N-5817 Bergen, Norwayb Institute of Marine Research, P.O. Box 1870, Nordnes, N-5817 Bergen, Norway

a r t i c l e i n f o

Keywords:DioxinsPCBsPBDEsBarents SeaNortheast Arctic codGadus morhua

a b s t r a c t

This study is one of several baseline studies on commercially important Norwegian wild fish species thatwill provide information concerning metals and persistent organic pollutants (POPs) and food safety. Thecod liver is a traditional food product in Norway and a potential source for POPs in the diet. The concen-trations of dioxins and furans (PCDD/Fs), dioxin-like PCBs (DL-PCBs), non-dioxin-like PCBs (NDL-PCBs,PCB6) and polybrominated flame retardants (PBDEs) were determined in the liver of 784 individualNortheast Arctic cod caught at 32 positions in the Barents Sea in the period from 2009–2010. In addition,muscle samples from 30 individual cod were analysed for the same substances. The mean concentrationof the sum of PCDD/Fs and DL-PCBs for all samples was 14.2 ng TEQwho-2005/kg ww with a variationbetween 1.0 and 151 ng TEQ/kg ww. The concentrations of POPs in the fillet samples were very low.

� 2013 Elsevier Ltd. All rights reserved.

Atlantic cod (Gadus morhua) is distributed in the North AtlanticOcean, and the Northeast Arctic cod stock is the population ofAtlantic cod that dominates the Barents Sea ecosystem. This is cur-rently, and possibly historically, the largest cod stock in the world(Yaragina et al., 2011) and the largest biomass (of fish aged threeyears and older) ever recorded was 4.2 million tonnes in 1946(Øiestad, 1994). Migration over long distances is a characteristicof this species. Adult fish can migrate up to 2000 km annually, fromfeeding grounds in the Barents Sea and in the waters around Sval-

bard, to the spawning areas along the Norwegian coast and back(Yaragina et al., 2011). Cod are well adapted to the environmentin the Barents Sea and, in 2011, cod was the commercially mostvaluable species of wild-caught fish in Norway (data from theDirectorate of Fisheries in January 2012). The catch of NortheastArctic cod has varied from 212,000 to 1,343,000 tonnes, with anaverage of 656,000 tonnes caught between 1946 and 2007 (Hylenet al., 2008; ICES, 2008).

In Norway, it is common practice to utilise not only the fillet butalso the liver from codfishes (family Gadidae), particularly Atlanticcod (Gadus morhua), saithe (Pollachius virens) and haddock(Melanogrammus aeglefinus). These fish species have a lean fillet,

K. Julshamn et al. / Marine Pollution Bulletin 75 (2013) 250–258 251

with a fat content of less than 1 g/100 g, and energy storage pri-marily in the liver. The liver is very oily, with a lipid content thatcan vary between 31 and 75 g/100 g (www.nifes.no/seafooddata),and with x-3 fatty acids accounting for up to 30% of the total fat.Fish liver is also rich in vitamins A and D (VKM, 2007). Extensivevolumes of fish liver from the three codfish species are producedas a side-product during fishing with about 13,000 tonnes fromcod, 23,000 tonnes from saithe and 5000 tonnes from haddock.Part of this volume is today used for production of cod liver oil, fol-lowing extensive clean-up to minimise the content of undesirablesubstances, while the major part is discarded. A minor part is alsoused directly for human consumption (VKM, 2007).

Since the cod musculature is very lean, only very small amountsof lipophilic compounds, including POPs, can be accumulated inthe fillet. However, the oily liver has the potential to accumulaterelatively large amounts of lipophilic organohalogen compounds,such as polychlorinated dibenzo-p-dioxins (PCDDs), polychlori-nated dibenzofurans (PCDFs) (collectively referred to as dioxinsor PCDD/Fs), polychlorinated biphenyls (PCBs) and the brominatedflame-retardants such as polybrominated diphenyl ethers (PBDEs).Of the PCBs, there are both dioxin-like PCBs (DL-PCBs), whichexhibit dioxin-like toxicity (i.e. non-ortho and mono-ortho PCBs),and the non-dioxin-like PCBs (i.e. NDL-PCBs, PCB6), which have adifferent toxicological action (EC, 2001). These organohalogencompounds are all defined as persistent organic pollutants (POPs)that are ubiquitous in the environment, and also present in lowconcentrations in food. To ensure that seafood products containingparticularly high levels of these substances are not sold in theEuropean market and hence to protect consumer safety, the EUhas set a maximum limit for the concentrations of the sum ofPCDD/Fs in fish muscle of 3.5 ng TEQ kg�1 wet weight (ww), andfor the the sum of PCDD/Fs and DL-PCBs in muscle and liver of fishof 6.5 and 20 ng TEQ kg�1 ww, respectively, using re-evaluatedToxic Equivalency Factors (TEF-2005) (EC, 2011; Van den Berget al., 2006). Recently, upper limits for the sum of six NDL-PCBs(PCB6) in fish muscle and liver have also been established at 75and 200 lg kg�1 ww, respectively (EC, 2011). To date, no upperlimit has been set for the PBDEs, even though risk assessmentshave been performed (JECFA, 2006).

The levels of undesirable substances such as PCB7 in fish andother seafood in Norway have been regularly monitored by spotchecks since 1994 (Julshamn et al., 2004; www.nifes.no/seafood-data). To establish more reliable surveillance data on undesirablesubstances, comprehensive baseline studies, including a largenumber of stations over a broad geographical area throughoutthe seasons, are needed for commercially important species suchas Arctic cod. Such baseline studies can also provide recommenda-tion to food safety authorities with recommendations on how fu-ture monitoring should be performed (e.g. frequency of samplingand number of sampling sites). Several fish stocks from Norwegianwaters (Greenland halibut (Reinhardtius hippoglossoides), Norwe-gian spring spawning (NSS) and North Sea herring (Clupea haren-gus), mackerel (Scomber scombrus), saithe and different stocks ofAtlantic cod) have been subjected to baseline studies where thelevels of undesirable substances, i.e. heavy metals and POPs, havebeen determined for a large number of fish. Results from thesestudies have earlier been reported for the baseline study of POPsin NSS herring (Frantzen et al., 2011) and for metals in NortheastArctic cod (Julshamn et al., 2013). This study reports on the base-line study of POPs in livers from 784 Northeast Arctic cod and afew muscle samples (30) of Northeast Arctic cod (the same sam-ples as reported for metals by Julshamn et al., 2013). Muscle andliver samples were analysed for PCDDs, PCDFs, DL-PCBs, NDL-PCBs(PCB6) and PBDEs (PBDE7). Biological parameters such as age,weight, length, gender, condition factor and liver index were re-corded in order to help explain the observed patterns.

The sampling of Northeast Arctic cod was carried out in 2009and 2010, covering an area in the Barents Sea from approximately70� to 75�N and from 16� to 41�E (Fig. 1). This area mainly coveredthe Norwegian part of the Barents Sea, although some sampleswere also taken in the Russian zone. Six commercial fishing vesselsbelonging to the reference fleet of the Institute of Marine Research(IMR) were involved in collecting the fish samples, and trawl wasthe major fishing gear in addition to nets, longlines and handlines.In total, samples were retrieved from 32 positions from February2009 to May 2010. At most positions, 25 cod were sampled accord-ing to a protocol given to the fishermen and, in total, 807 fish werecollected.

Whole fish were individually frozen before being shipped tothe laboratory. The fish were defrosted and the length, weightand gender for each individual fish were determined and the oto-liths were removed for subsequent age determination. To obtainas much information as possible, all analyses were performedon samples from individual fish. The liver was removed from eachfish and weighed, and then homogenised using a food processor.The homogenate was stored in closed glass bottles at �20 �C untilanalysis. Approximately 200 g muscle tissue without skin from 30individual fish (i.e. 20 fish from the southwest area and 10 fishfrom the east area with high dioxin contents in the liver) was re-moved for analysis. The wet subsamples were homogenised andfreeze-dried, and the dry matter content was calculated. Thedried muscle samples were ground to a fine powder, which wasagain homogenised and kept dry prior to analysis. A sample ofthe remaining wet homogenates was prepared for the determina-tion of the fat content. Fish age was determined by otolith read-ing. The otoliths were broken, and the annual growth zonesvisible on the broken sections were counted using a binocularmicroscope with light transmitted from the side and the brokensurface shadowed.

The fat was extracted from the wet sample homogenate byshaking with 30% isopropanol in ethyl acetate. The solution wasthen filtered before the solvent was evaporated and the fat residueweighed. Measurement uncertainties ranged between 5% and 10%and the limit of quantification (LOQ) was 0.1 g/100 g. The methodwas tested by performing three proficiency tests (PT) in 2010, ana-lysing three different seafood samples having fat contents between2 and 20 g/100 g. The z-scores obtained were between �0.3 and+2.0. This method is accredited in accordance with ISO-EN 17025and is registered as a Norwegian Standard, NS 9402.

The same method of sample clean-up and extraction was usedfor dioxins and furans (PCDD/Fs), DL-PCBs, NDL-PCBs and PBDEs.The applied method was slightly modified from the method de-scribed in detail by Berntssen et al. (2010), based on method num-bers 1613 and 1668 of the United States Environmental ProtectionAgency (USA EPA, 1994,1999). Briefly, wet and homogenised liversample or dried homogenised muscle sample was mixed withhydromatrix (Agilent Technologies) and 13C-labelled internal stan-dards were added (27 standards for dioxins, furans and dioxin-likePCBs, CIL, and one standard for PBDEs, Sigma–Aldrich). The mix-ture was transferred to an Accelerated Solvent Extractor 300(ASE, Dionex Corp.) or a Pressurised Liquid Extraction System(PLE, FMS Inc.), with a layer of acidic silica gel (Merck), and ex-tracted with hexane under elevated pressure and temperature(100 �C, 1500 psi). The extract was further purified chromato-graphically using PowerPrep (FMS Inc.) over three columns packedwith multilayer silica, basic alumina and carbon, respectively, andwas eluted with different solvents. Two fractions were collected:fraction 1 contained PBDEs, PCB6 and mono-ortho PCBs andfraction 2 contained dioxins, furans and non-ortho PCBs. Determi-nations of PCDD/Fs and non-ortho-PCBs were performed byhigh-resolution gas chromatography/high-resolution massspectrometry (HRGC/HRMS) using a DFS-MS, HRGC-HRMS, Ther-

Fig. 1. Sampling positions for Northeast Arctic cod (Gadus morhua) from February 2009 to May 2010, with an overview map showing the sampling area in the upper rightcorner.

252 K. Julshamn et al. / Marine Pollution Bulletin 75 (2013) 250–258

moScientific, equipped with a fused silica capillary column(RTX-5SILMS, Restek, Bellefonte, USA) according to USA EPA meth-ods (USA EPA, 1994,1999). Quantification was done using an inter-nal standard, a five-point calibration curve and the ratio RF.Determination of mono-ortho PCBs were performed by triple-quadchromatography using a GC-MSMS (7000B, Agilent), also equippedwith a fused silica capillary column (RTX-5SILMS, Restek, Belle-fonte, USA), and quantified using the internal standard and thefive-point calibration curve through the origin. The World HealthOrganization (WHO) has established TEFs for human risk assess-ment (Van den Berg et al., 2006). Values are expressed as theWHO-TEQ, with use of the WHO-TEFs from 2005 as is used in cur-rent EU legislation on dioxins and DL-PCBs (EC, 2011). The range oflimits of quantification for liver and muscle samples was 0.02–0.10 pg/g ww for PCDD/Fs, 0.07–1.50 pg/g ww for non-ortho PCBsand 0.6–166 pg/g ww for mono-ortho PCBs. Non-quantified indi-vidual congeners were set at the LOQ (upper-bound LOQ). Recov-ery values were calculated as described in the USEPA (1994) andwere found to be between 78% and 110%. Concentrations are ex-pressed as ng upper-bound WHO-2005-TEQ per kg ww, accordingto EU legislation (EC, 2011). The method was tested in a proficiencytest (PT) of cod liver oil arranged by the National Institute of PublicHealth, Oslo, Norway in 2012 with concentrations ranging from0.03 to 6.50 pg/g for PCDD/Fs and from 0.01 to 13,600 pg/g forDL-PCBs. The trueness achieved as a z-score for PCDD/Fs rangedfrom �0.35 to +3.60 (only four congeners out of 17 showed valuesabove +2, and all with low concentrations) and the z-score for DL-PCBs ranged from �0.41 to +0.29.

PCB6 is defined as the sum of the six NDL-PCB congeners (i.e.PCB-28, �52, �101, �138, �153, �180). PCB6 was also determinedby the triple-quad GC-MSMS (7000B, Agilent) with the sameinstrument method as for mono-ortho PCBs, and was quantifiedusing the internal standard and the five-point calibration curvethrough the origin. The LOQ in liver and muscle for the sum ofPCB6 was estimated as 0.3 ng/g ww and the LOQ for each of thecongeners varied between 0.06 and 0.15 lg/kg ww. Measurementuncertainty on liver samples was estimated as 20% for the PCB6

congeners. The NDL-PCB method was tested in a PT on cod liveroil from the National Institute of Public Health, Oslo, Norway in2012. The concentrations for the PCB congeners of the PT rangedfrom 2 to 24 ng/g. The z-scores were acceptable for the congenersand ranged from �0.07 to +0.96.

The PBDE congeners (PBDE-28, 47, 66, 99, 100, 119, 138, 153,154 and 183) were determined on GC–MS NCI (negative chemicalionising; GC: Thermo Quest Trace GC 2000 series, MS: Trace DSQ)and were quantified using the internal standard and a five-pointcalibration curve. The determinations were performed in SIM (Sin-gle Ion Monitoring) mode. The method quantifies ten differentcongeners of PBDEs, including seven congeners that add up to a‘‘standard sum PBDE’’ (i.e. the sum of PBDE-28, 47, 99, 100, 153,154 and 183). The LOQ for liver and muscle was estimated as0.05 ng/g ww for the congeners PBDE 28, 100, 153 and 154 and0.10 ng/g ww for the congeners PBDE 47, 99 and 183. The measure-ment uncertainties were highly related to the concentration of thecongeners. The measurement uncertainty ranged from 25% to 45%for the PBDE congeners. The trueness of the PBDE method was also

Table 1Physical and biological parameters of 807 cod captured in the Barents Sea during 2009 and 2010. Overall mean, standard deviation (SD), and minimum and maximum values areshown for each area (i.e. East, Northwest and Southwest) and overall. Letters a and b indicate significant differences.

Parameter East Northwest Southwest Overall

N 312 271 224 807Age (year) 6.0 ± 1.6a 5.4 ± 1.3b 6.3 ± 1.4a 5.9 ± 1.5

(3–13) (3–10) (3–11) (3–13)N 310 269 222 801Length (cm) 66 ± 14a 62 ± 11b 67 ± 12a 65 ± 12

(39–110) (38–98) (37–96) (37–110)Weight (kg) 2.7 ± 1.9a 2.2 ± 1.3b 2.7 ± 1.5a 2.5 ± 1.6

(0.5–14.3) (0.5–10.3) (0.4–9.0) (0.4–14.3)Liver weight (g) 69 ± 64a 108 ± 90b 127 ± 112b 102 ± 93

(8–372) (3.3–490) (9.6–630) (8–630)(99) (144) (98) (341)

(N) Condition (weight (g) length�3 (cm)) 0.83 ± 0.10a 0.86 ± 0.09b 0.82 ± 0.10a 0.84 ± 0.10(0.60–1.15) (0.63–1.38) (0.64–1.28) (0.60–1.38)

Fat content liver (g 100 g�1) 50.9 ± 10.9a 53.8 ± 11.2b 53.5 ± 16.7ab 53 ± 12.9(6.3–77.0) (11.2–76.8) (6.2–84.3) (6.2–84.3)(301) (265) (214) (780)

(N)% female 51 45 54 50

Fig. 2. Plot of fat content (g 100�1 g) against liver index for livers of NortheastArctic cod (Gadus morhua) sampled from February 2009 to May 2010.

K. Julshamn et al. / Marine Pollution Bulletin 75 (2013) 250–258 253

tested in a PT on cod liver oil from National Institute of PublicHealth, Oslo, Norway in 2012. The concentrations for the PBDEcongeners of the PT ranged from 0.005 to 4.6 ng/g. The z-scoresfor the congeners using this method ranged from �1.0 to +0.15,showing acceptable results for trueness.

All three analytical methods are accredited in accordance withISO-EN 17025 by the Norwegian Accreditation body.

To achieve an initial overview of the data, a principal compo-nent analysis (PCA) was performed and PCA biplots wereconstructed.

The sum of PCDD/Fs and DL-PCBs as well as the sum of NDL-PCBs were highly skewed, with long tails of high values, as wellas heteroscedasticity showing increasing variation with highermean values. Logarithmic and square root transformations werenot sufficient to remove these violations. Box–Cox transformations(Box and Cox, 1964) both assigned normal distribution to the dataand removed the problem of increasing variation with increasingmeans and were therefore used for transforming data before statis-tical analysis. With the high number of samples and a significancelevel of 0.01, however, plain and Box–Cox transformed data gavethe same results.

For the comparison of physical parameters and POPs concentra-tions among the different areas of the Barents Sea (Northwest,Southwest and East), one-way analysis of variance (ANOVA) andanalysis of covariance (ANCOVA) followed by Tukey multiple com-parison tests (Zar, 2010) were applied to both plain concentrationsand Box–Cox transformed data. Sirius 8.1 (Pattern Recognition Sys-tems AS, Norway) was used for the PCA analysis and Statistica 10(Statsoft Inc, USA) was used for all other analyses and graphs.

A significance level of 0.01 was used in all analyses.In total, 807 samples of Northeast Arctic cod were caught and

out of these, livers from 784 fish and muscle samples from 30 fishwere analysed for organohalogen compounds. The physical andbiological parameters of the fish are shown in Table 1. The fishshowed an age range from three to 13 years and had body lengthsbetween 37 and 110 cm and weights between 0.4 and 14.3 kg.Males and females were equally represented. The mean age of fishfrom the areas to the east, northwest and southwest were 6.0, 5.4and 6.3 years, respectively. Cod caught in the area to the northwestwere significantly smaller and younger compared to cod from theeast and southwest areas (Table 1; Tukey, p < 0.001). The conditionfactor (given as fish weight�length�3) was, however, significantlyhigher (p < 0.001) in cod sampled in the northwest area (mean0.86) compared to the two other areas (means 0.83 and 0.82).The fat content was determined in 780 livers of cod and varied

between 6.2 and 84.3 g/100 g and the liver index (liverweight � 100) � fish weight�1) determined for 340 fish varied be-tween 0.31 and 11.7. Both the lowest liver fat content and the low-est liver index were found in cod sampled in the east area (Table 1).Mean liver index in the east area was 2.8, while the northwest andthe southwest showed mean liver indices of 4.4 and 4.1, respec-tively. The liver index (also called hepato-somatic index) can beused as an indicator of gonad maturation. A low liver index isattributed to mature cod in the spawning period (Kjesbu et al.,1991). A low fat content and low liver index indicate that these fishwere spawning. The relationship between fat content and liver in-dex is shown in Fig. 2. The highest fat content recorded in liver inthe present study was 84.3 g 100 g�1, with an overall mean of 51 g100 g�1 ww. The liver fat contents found in this study were some-what lower than those found for previous Norwegian monitoringof fat content in the livers of Northeast Arctic cod in the years2006 and 2007 (averaged 56–64 g/100 g ww), but the number of li-ver samples in these monitoring reports were only 15 and 10,respectively (www.nifes.no/seafood). Horst and Wiederholt-Lahrs-sen (2009) reported data for the fat content in pooled livers of codfrom east Greenland in the same concentration range as found inthe present study. In the present study there was a positive linearcorrelation between the fish weight and liver weight (r = 0.69,p < 0.001). The condition factor was higher in younger fish, andmean age and condition showed opposite patterns with season(data not shown).

Table 2Concentrations of the sum of PCDDs, PCDFs, non-ortho PCBs, mono-ortho PCBs, PCDD/PCDFs, PCDD/Fs + DL-PCBs (ng TEQWHO-2005/kg ww), non-dioxin-like PCBs (PCB6) and PBDEsare given as means ± SD, range (minimum–maximum) and the median in livers of Northeast Arctic cod caught during 2009 and 2010. Concentrations are given for each area andall samples. Sums are based on the upper-bound LOQ for dioxins and PCBs and lower bound for PBDEs. The EU maximum value for consumption of fish livers is given.

Compound (mg/kg wet weight) Parameter East (N = 308) Northwest(N = 256)

Southwest(N = 220)

All samples(N = 784)

Range of means(N = 32)

EU maximumvalue

RPCDD (ng TEQ/kg ww) Mean ± SD 1.2 ± 0.6 0.79 ± 0.49 1.3 ± 0.67 1.1 ± 0.6Min–max 0.1–4.3 0.2–3.1 0.3–4.6 0.1–4.6Median 1.1 0.57 1.2 0.97

RPCDF (ng TEQ/kg ww) Mean ± SD 1.6 ± 0.8 0.85 ± 0.51 1.8 ± 1.1 1.4 ± 0.9Min–max 0.2–5.3 0.2–3.4 0.1–6.0 0.1–6.0Median 1.5 0.72 1.6 1.2

RPCDD + PCDF (ng TEQ/kg ww) Mean ± SD 2.8 ± 1.2 1.6 ± 0.8 3.1 ± 1.7 2.5 ± 1.4Min–max 0.3–9.2 0.6–4.7 0.5–8.7 0.3–9.2Median 2.6 1.3 2.7 2.2

Rnon-ortho PCB (ng TEQ/kg ww) Mean ± SD 14.4 ± 11.8 6.2 ± 5.1 11.4 ± 7.5 10.9 ± 9.6Min–max 1.6–132 1.7–32 0,23–50 0.23–132Median 11.3 4.2 9.4 8.5

Rmono-ortho PCB (ng TEQ/kg ww) Mean ± SD 1.1 ± 0.9 0.44 ± 0.36 0.84 ± 0.58 0.79 ± 0.71Min–max 0.1–9.9 0.1–2.1 0.2–3.6 0.1–9.9Median 0.84 0.30 0.69 0.61

RPCDD/F + DL-PCB (ng TEQ/kg ww) Mean ± SD 18.3 ± 13.5 8.2 ± 6.0 15.3 ± 9.4 14.2 ± 11.2 4.6–25.7 20Min–max 2.0–151 2.7–37.3 1.0–60.4 1.0–151Median 14.5 6.2 12.9 11.5

PCB6 (lg/kg ww) Mean ± SD 110 ± 73 60 ± 44 103 ± 66 92 ± 67 37–164 200Min–max 9.7–510 16–296 18–352 9.7–510Median 92 46 85 72(N) 305 255 221 781

PBDE7 (lg/kg ww) Mean ± SD 4.5 ± 2.6 3.0 ± 2.1 6.1 ± 5.0 4.5 ± 3.5 1.8–12.8Min–max 0.2–16 0.6–15 1.4–36 0.2–36Median 3.8 2.4 4.7 3.4(N) 305 254 215 774

Table 3Determinations of the sum of PCDDs, PCDFs, non-ortho PCBs, mono-ortho PCBs andPCDD/Fs + DL-PCB (ng TEQWHO-2005/kg ww and ng TEQWHO-1998/kg ww) in the liversof Northeast Arctic cod caught during 2009 and 2010. Means ± SD are given. The ratiobetween TEQ calculated with TEFwho-2005 and TEFWHO-1998 are given. The EUmaximum values for the consumption of fish livers are 20 ng TEQWHO-2005/kg wwand 25 ng TEQWHO-1998/kg ww.

Compound Mean ± SD ngTEQWHO-2005

kg�1 ww

Mean ± SD ngTEQWHO-1998

kg�1 ww

TEQWHO-2005/TEQWHO-1998

PPCDDs 1.1 ± 0.6 1.1 ± 0.7 1.0

PPCDFs 1.4 ± 0.9 1.6 ± 1.1 0.88

PNon-ortho PCBs 10.9 ± 9.6 10.3 ± 9.3 1.05

PMono-ortho PCBs 0.79 ± 0.71 3.5 ± 3.2 0.23

PPCDD/Fs + DL-PCB 14.2 ± 11.2 16.7 ± 13.4 0.85

254 K. Julshamn et al. / Marine Pollution Bulletin 75 (2013) 250–258

A summary of the concentrations of the sums of PCDDs, PCDFs,non-ortho PCBs, mono-ortho PCBs and PCDD/Fs and DL-PCBs is gi-ven in Table 2, showing results for the different areas and overall.The concentration of the sum of PCDDs and PCDFs (PCDD/Fs) in li-ver samples of 784 individual Northeast Arctic cod ranged from 0.3to 9.2 ng TEQ kg�1 ww, with a mean ± SD of 2.5 ± 1.4 ng TEQ kg�1 -ww (Table 2). The concentration of the sum of PCDD/Fs and DL-PCBs varied from 1.0 to 151 ng TEQ kg�1 ww, with a mean ± SDof 14.2 ± 11.2 ng TEQ kg�1 ww. The mean concentration of thesum of PCDD/Fs and DL-PCBs for each of the 32 sampling positionsranged from 4.6 to 25.7 ng TEQ kg�1 ww (Table 2).

In total, 166 samples out of 784 individual cod livers (21%) had aconcentration of PCDD/Fs + DL-PCBs higher than 20 ng TEQ kg�1 -ww, which is the maximum level set by the EU for fish liver in-tended for human consumption (EC, 2011). Eight positions hadmean concentrations higher than 20 ng TEQ/kg ww, and only sevenpositions out of 32 had no individual samples with concentrationsof the sum of PCDD/Fs and DL-PCBs exceeding 20 ng TEQ kg�1 ww.

A set of TEF values for the seven congeners of PCDDs, 10 cong-eners of PCDFs, four congeners of non-ortho PCBs and eight cong-eners of mono-ortho PCBs was introduced by the WHO in 1998 (EC,2006), and in 2005 a number of these TEF values were changed,

especially for mono-ortho PCBs (Van den Berg et al., 2006; EC,2011). In this study, TEQ has been calculated for each of the cong-eners of 784 liver samples with both WHO-TEF-2005 as well asWHO-TEF-1998 (Table 3). Small differences were recorded forTEQ concentrations using WHO-TEF-1998 and WHO-TEF-2005 forPCDDs, PCDFs and non-ortho PCBs, whereas the largest differencewas recorded for mono-ortho-PCBs as expected, with a concentra-tion level of 0.79 ng TEQWHO-2005 kg�1 ww and 3.2 ng TEQWHO-1998

kg�1 ww, respectively. The ratio between TEQ calculated withTEF-WHO-2005 and TEF-WHO-1998 for all samples was 0.85.The maximum level of the sum of PCDD/Fs and DL-PCBs in fish li-ver set by the EU was changed from 25 ng TEQWHO-1998 kg�1 ww to20 ng TEQWHO-2005 kg�1 ww from January 2012. As a consequence,several samples that did not exceed the 25 ng TEQWHO-1998 kg�1

ww limit now exceeded the new limit of 20 ng TEQWHO-

2005 kg�1 ww.The total concentration of PCDD/Fs and DL-PCBs was dominated

by the non-ortho-PCBs, which constituted 77% of the total concen-tration of PCDD/Fs and DL-PCB, whereas the sums of PCDD, PCDFand mono-ortho PCBs contributed only 7.7%, 9.9% and 5.6%, respec-tively, to this sum. In this study, the ratio between the mean con-centration of the sum of DL-PCBs and PCDD/Fs in cod liver was 4.7.Much lower ratios were found for the fillets of oily fishes such asthe Norwegian Spring Spawning herring (Clupea harengus), 1.0(Frantzen et al., 2011); mackerel (Scomber scombrus), 1.5 (Frantzenet al., 2010); Greenland halibut (Reinhardtius hippoglossoides), 1.5(Nilsen et al., 2010) and brown meat of crab (Cancer pagurus), 0.8(Julshamn et al., 2012).

Among the 29 different PCDD, PCDF and DL-PCB congenersdetermined in this study, the non-ortho PCB congener PCB-126was the dominating congener, constituting 95% of the sum ofnon-ortho PCBs and 73% of the total TEQ concentration of thesum of PCDD/F + dl-PCB. The TEQ concentration of the PCDDs,2378 TCDD and 12378 PeCDD were the dominating congeners,with 52% and 43% of the sum of PCDDs, respectively, and amongthe PCDFs, 2378 TCDF and 23478 PeCDF were the dominatingcongeners, with 64% and 19% of the sum of PCDFs, respectively.

Fig. 3. Principal component analysis plot with principal component 1 vs. 2 for the associations between concentrations of the sum of PCDD/Fs and DL-PCBs in liver andphysical parameters of Northeast Arctic cod (Gadus morhua), with A. lipid weight based concentrations and B. wet weight based concentrations. The variables arestandardised. The samples are marked with area codes (E for East, NW for Northwest and SW for Southwest).

Table 4Concentrations of the sum of PCDD/Fs, non-ortho PCBs and mono-ortho PCBs (DL-PCBs), sum of PCDD/Fs + DL-PCBs (ng TEQWHO-2005/kg ww), PCB6 and PBDE7 in muscle ofNortheast Arctic cod caught during 2009 and 2010 (N = 30). The EU maximum values are given for the sum of PCDD/Fs and PCDD/Fs + DL-PCB as well as for PCB6. N.D. (notdetermined).

Parameter N = 30 PCDD/Fs(ng TEQWHO-2005 kg�1 ww)

DL-PCBs(ng TEQWHO-2005 kg�1 ww)

PCDD/Fs + DL-PCBsa

(ng TEQWHO-2005 kg�1 ww)PCB6

a

(lg kg�1 ww)PBDE7

b

(lg kg�1 ww)

Mean 0.045 0.030 0.076 0.60 0.010SD 0.026 0.019 0.032 0.22 N.D.Median 0.039 0.025 0.063 0.60 0.007Minimum 0.026 0.013 0.048 0.60 0.002Maximum 0.13 0.10 0.14 1.3 0.030EU maximum level 3.5 6.5 75

a Sums are based on the upper-bound LOQ.b Sums are based on the lower-bound LOQ.

K. Julshamn et al. / Marine Pollution Bulletin 75 (2013) 250–258 255

PCB-118 and PCB-105 were the dominating mono-ortho PCB cong-eners, constituting 65% and 20% of the sum of mono-ortho PCBs,respectively (the congener profiles are not shown). The highestchlorinated dioxin compounds, which also have the lowest TEF,showed the lowest TEQ concentrations, which is in agreement withprevious findings (Rappe et al., 1987; Jimenez et al., 1991). The lowTEQ concentration for higher chlorinated is a product of their lowTEF values and/or low concentration in ng/g.

The concentrations of the sum of PCDD/Fs and DL-PCBs found inthis study were more or less in agreement with results previouslyreported for the livers of Northeast Arctic cod from the Barents Sea(www.nifes.no/seafooddata). The mean concentration for each ofthe 32 sampling positions ranged from 4.6 to 25.7 ng TEQ kg�1 wwfor the sum of PCDD/Fs and DL-PCBs, whereas cod livers analysedfor the Norwegian monitoring programme in 2006, 2007 and2008 showed mean concentrations of the sum of PCDD/Fs andDL-PCBs of 10.5 (N = 91), 19.1 (N = 25) and 15.2 ng TEQ kg�1 ww(N = 41), respectively, with a variation for the 3 years’ samplesranging from 1.5 to 92 ng TEQ kg�1 ww.

There are only a few papers reporting the sum of PCDD/Fs andDL-PCBs in livers of cod. However, Horst and Wiederholt-Lahrssen(2009) reported such data from different fishing grounds of theNorth Sea, Baltic Sea and the North Atlantic, whereas Green and

Knutzen (2003) reported data from different coastal and fjord areasof Norway. In the report from Horst and Wiederholt-Lahrssen(2009), one pooled sample from each position was analysed (fiveto 20 samples in each pooled sample). From the North Atlantic, fishfrom three positions east of Greenland were included, where theyfound concentrations of the sum of PCDD/Fs and DL-PCBs rangingfrom 2.5 to 3.9 ng TEQWHO-1998 kg�1 ww, with a mean concentra-tion of 3.4 ng TEQWHO-1998 kg�1 ww. In the present study, meanconcentration ± SD given as ng TEQWHO-1998 kg�1 ww was16.7 ± 13.1 ng TEQ kg�1 ww (Table 3), and only five out of 256 liv-ers of cod sampled in the northwest area had a concentration loweror equal to the mean value of 3.4 ng TEQWHO-1998 kg�1 ww. The rea-son why concentrations were higher in cod from the Barents Seathan in cod from Greenland waters could be that the cod caughteast of Greenland belonged to another cod stock with a generallylower exposure to dioxins and DL-PCBs than the Northeast Arcticcod stock. The Greenland cod stock does not migrate eastwardsinto the Barents Sea and Greenland is farther away from populatedand polluted areas than the Barents Sea. Samples from the BalticSea and North Sea were also included in the same study by Horstand Wiederholt-Lahrssen (2009). In those areas, the concentrationof the sum of PCDD/Fs and DL-PCBs ranged from 69 to 139 ngTEQ kg�1 ww and from 44 to 123 ng TEQ kg�1 ww, respectively,

Fig. 4. Plot of the median sums of PCDD/Fs and DL-PCBs concentrations (ng TEQkg-1 ww) against length-classes for the three areas of the Barents Sea. Vertical barsindicate 25% and 75% percentiles.

ig. 5. Plot of concentrations of the sum of PCDD/Fs and DL-PCBs (ng TEQ kg�1 ww)the livers against the sum of PCDD/Fs and DL-PCBs (ng TEQ kg�1 ww) in muscle

f Northeast Arctic cod (Gadus morhua).

Fig. 6. Plot of concentrations of the sum of PCDD/Fs and DL-PCBs (ng TEQ kg�1 lipidweight; lw) and fat content (g 100�1 g) in the livers of Northeast Arctic cod (Gadusmorhua).

Table 5Correlation coefficient (r) matrix of size, age, fat content, the sum of PCDD/Fs + DL-PCBs, PCB6 and PBDE7 in livers of Northeast Arctic cod (Gadus morhua) from the Barents Sea.Correlations are significant at p < 0.0001 (���), p < 0.001 (��) and p < 0.01 (�). N = 746.

Fish length Fish weight Age Fat content PBDD/Fs + DL-PCB PCB6 PBDE7

Fish length 1.00 0.93��� 0.84��� 0.11�� 0.40��� 0.43��� 0.35���

Fish weight 0.93��� 1.00 0.80��� 0.13��� 0.39��� 0.40��� 0.32���

Age 0.84��� 0.80��� 1.00 0.05 0.42��� 0.45��� 0.34���

Liver fat content 0.11�� 0.13��� 0.05 1.00 �0.24��� �0.18��� �0.09PBDD/Fs + DL-PCB 0.40��� 0.39��� 0.42��� �0.24��� 1.00 0.85��� 0.61���

PCB6 0.43��� 0.40��� 0.45��� �0.18��� 0.85��� 1.00 0.65���

PBDE7 0.35��� 0.32��� 0.34��� �0.09 0.61��� 0.65��� 1.00

256 K. Julshamn et al. / Marine Pollution Bulletin 75 (2013) 250–258

with mean concentrations of 93 and 83 ng TEQ kg�1 ww. The con-centrations of the sum of PCDD/Fs and DL-PCBs reported in cod li-ver from the North Sea were five to 10 times higher than theconcentrations found in this study. Green and Knutzen (2003) re-ported values of the sum of PCDDs and DL-PCBs in pooled samplesfrom coastal and fjords areas of Norway ranging from 27 to 138 ngTEQ kg�1 ww, which all exceeded the EU maximum level for thesum of PCDD/Fs and DL-PCBs of 25 ng TEQWHO-1998 kg�1 ww in fishliver intended for human consumption. It is not surprising that codfrom the North Sea, Baltic Sea and the Norwegian fjords and coastalareas, surrounded by much more densely populated and industria-lised areas, show higher concentrations of dioxins and PCBs com-pared with cod from the more pristine Barents Sea.

A principal component analyses (PCA) showed that the concen-trations of PCDD/Fs and DL-PCBs, PCB6 and PBDEs in liver co-variedwith cod length, weight and age, but not with the condition factoror the liver fat content of the fish (Fig. 3). The sum of PCDD/F andDL-PCB concentration was significantly correlated with age, weightand length, with correlation coefficients (r) of 0.43, 0.42 and 0.43,respectively (Table 5). The sum of PCDD/Fs and DL-PCBs concentra-tions in cod liver has also previously been shown to increase withincreasing age and size (Horst and Wiederholt-Lahrssen, 2009). Forother species, several reports have shown a positive significantcorrelation between the sum of PCDD/Fs and DL-PCBs and ageand size in fish species such as Norwegian Spring Spawning herring(Frantzen et al., 2010), mackerel (Frantzen et al., 2010) and bluewhiting (NORA, 2003), but no significant correlation was foundfor Greenland halibut (Nilsen et al., 2010).

In the present study, difference in PCDD/Fs and DL-PCB levels indifferent areas of the Barents Sea seemed to be explained by fishage, length and size. One sampling position out of 10 from thenorthwest area showed a mean concentration higher than 20 ng

Fino

TEQ kg�1 ww, whereas two sampling positions out of nine from thesouthwest area and five sampling positions out of 13 from the eastarea showed concentrations higher than 20 ng TEQ kg�1 ww (datanot shown). Because the fish sampled in the northwest were signif-icantly smaller and younger than the fish sampled in the other twoareas (Table 1), some of the reason for the differences in PCDD/Fsand DL-PCBs concentrations might possibly be attributed to thedifferences in size. However, the concentration of the sum ofPCDD/Fs and DL-PCBs differed significantly between the threeareas also when age, weight or length were used as covariates (AN-COVA followed by Tukey, p < 0.0001), and similar results were ob-

K. Julshamn et al. / Marine Pollution Bulletin 75 (2013) 250–258 257

tained with Box–Cox transformed concentrations. This is illus-trated in Fig. 4, showing that the sum of PCDD/Fs and DL-PCBs lev-els in liver were lower in cod from the northwest than in cod fromthe other two areas for almost all length classes.

As PCDD/Fs and PCBs are fat soluble, concentrations and levelstend to be higher in more oily fish and concentrations are often ex-pressed on a lipid basis. A negative correlation was found betweenfat content and the concentration based on lipid weight of the sumof PCDD/Fs and DL-PCBs in livers of Northeast Arctic cod (Fig. 6,r = �0.51; p < 0.001), a weaker correlation was found between fatcontent and the concentration based on wet weight (Table 5).These data show that the lipid-based concentration of the sum ofPCDD/Fs and DL-PCBs in cod liver is diluted when the fat contentincreases. The highest lipid-based concentration of the sum ofPCDD/Fs and DL-PCBs was found in livers with a very low fat con-tent (<10 g 100 g�1 ww, Fig. 6). Similarly, in other wild fish stockssuch as blue whiting (Micromesistius poutassou), the loss of storedfat during spring causes the PCDD/F concentration in the fat com-partment to increase, and when fat storage increases during winterthe PCDD/Fs lipid concentration decreases (NORA, 2003). Incontrast, in NSS herring, it was shown that when fat contents de-creased in spring due to spawning, lipid-based POP concentrationsincreased while wet weight concentrations decreased (Frantzenet al., 2011). In the cod livers of the present study, storage of fatincreased through the feeding season, lipid-based POP concentra-tions decreased while wet-weight-based concentrations increasedor remained constant. Through the winter starvation period, bothlipid-based and wet-weight-based concentrations of POPs in-creased as the fat contents decreased. Thus, in the material as awhole, the concentrations of POPs on a wet weight basis wereuncorrelated with fat contents because the relationship varieddepending on season, while there was a negative correlation be-tween lipid-based concentrations of POPs and fat contents.

Concentrations of PCDD/Fs and DL-PCBs, non-dioxin-like PCBs(PCB6) and PBDE7 in muscle of 30 Northeast Arctic cod are givenin Table 4. The concentration of PCDD/Fs ranged from 0.026 to0.13 ng TEQ kg�1 ww with a mean ± SD of 0.045 ± 0.026 ngTEQ kg�1 ww, and the concentrations of the sum of PCDD/Fs andDL-PCBs ranged from 0.048 to 0.14 ng TEQ kg�1 ww, with amean ± SD of 0.076 ± 0.032 ng TEQ kg�1 ww. From a food safetypoint of view, the levels were much lower than the concentrationsof PCDD/Fs of 3.5 ng TEQ kg�1 ww and the sum of PCDD/Fs and DL-PCBs of 6.5 ng TEQ kg�1 ww, which has been set by the EU as max-imum values for fish fillet for human consumption (EC, 2011). Theconcentrations of lipophilic organohalogen compounds are ex-pected to be low in a lean fish species such as cod compared to oilyfish species as very little of the POPs is stored in the muscle tissue.The TEQ concentrations of the PCDD/Fs, 12378-PeCDD, 2378-TCDDand 2378-PeCDF were the dominating congeners, with 46%, 24%and 8.5% of the sum of PCDD/Fs, respectively. Of the non-orthoPCBs, PCB-126 was the dominating congener, with 93% of thesum of non-ortho PCBs. Of the mono-ortho PCBs, PCB-118 wasthe dominating congener, with 56% of the sum of mono-orthoPCBs. The concentration of the sum of PCDDs, PCDFs, non-orthoPCBs and mono-ortho PCBs as a proportion of the sum of PCDD/Fs and DL-PCBs was 45%, 14%, 37% and 4%, respectively. The con-centrations of the congeners of PCDDs and non-ortho PCBs showedthe highest proportion of the sum of PCDD/Fs and DL-PCBs in mus-cle of cod, with 45% and 37%, respectively. The congener profiles insamples of cod muscle differed from those found for cod liver,where the congeners of PCDD in liver showed a low proportionof about 7.7%, whereas non-ortho PCBs accounted for 77% of thesum of PCDD/Fs and DL-PCBs. Horst and Wiederholt-Lahrssen(2009) found that the congener pattern of muscle was very similarto that found for the livers of cod caught in the Baltic Sea. The rea-son for the discrepancy might be attributed to the difference in

concentration between cod sampled in the Baltic Sea and theBarents Sea. The concentrations of the sum of PCDD/Fs and DL-PCBs in liver samples were positively linearly correlated with theconcentrations in the muscle samples (Fig. 5, r = 0.42). It shouldbe noted that the number of muscle samples was only 30.

The concentrations of PCB6 in liver samples of 781 individualNortheast Arctic cod ranged from 9.7 to 510 lg kg�1 ww, with amean ± SD of 92 ± 67 lg kg�1 ww (Table 2). The concentration ofthe sum of NDL-PCBs found in this study in livers of NortheastArctic cod from the Barents Sea did not greatly differ from previ-ously reported results from the same area (www.nifes.no/seafood,Julshamn et al., 2004), although the number of samples in the pre-vious study was rather limited. Mean concentrations for each ofthe 32 sampling positions ranged from 37 to 164 lg kg�1 wwfor PCB6 (Table 2), while cod liver analysed for the Norwegianmonitoring programme in the years 2006, 2007 and 2008 showedmean concentrations of the sum of PCB7 of 92 lg kg�1 ww (23–575 lg kg�1 ww, N = 91), 165 lg kg�1 ww (15–680 lg kg�1 ww,N = 25) and 113 lg kg�1 ww (11–389 lg kg�1 ww; N = 41),respectively.

Horst and Wiederholt-Lahrssen (2009) found that the concen-tration of the sum of five NDL-PCBs (PCB-28 is missing) rangedfrom 26 to 41 lg kg�1 ww in pooled samples of liver of cod fromeast Greenland. In the present study, the concentration of NDL-PCBs in the liver of cod caught in the northwest area of the BarentsSea showed values between 16 and 296 lg kg�1 ww, with a meanof 60 lg kg�1 ww (Table 2). There are smaller differences betweenconcentration data from east Greenland and our data consideringthe concentration of NDL-PCBs, compared with that of PCDD/Fsand DL-PCBs. Liver samples from the Baltic Sea and North Sea var-ied in concentration from 303 to 1055 lg kg�1 ww (Horst and Wie-derholt-Lahrssen, 2009). The concentration of the sum of NDL-PCBswas positively correlated with the concentration levels of the sumof PCDD/Fs and DL-PCBs as well as with the sum of PCDD/Fs, withlinear correlation coefficients of 0.85 and 0.68, respectively (DL-PCB vs. PCDD/Fs not shown in Table 5).

The congener profile of PCB6 showed that PCB-153 and PCB-138were the dominating congeners, with 34% and 32% of PCB6, respec-tively, followed by PCB-101 with 15%, and PCB-52 and PCB-180with 8.3%. Horst and Wiederholt-Lahrssen (2009) reported a some-what different congener profile in livers of cod from east Greenlandwith PCB-153 as the dominant congener of 40%, followed by PCB-101 and PCB-138, with 20% of each. Similar to the data on PCDD/Fsand DL-PCBs, the lowest concentrations of NDL-PCBs were found inthe livers of cod sampled in the northwest area, ranging from 16 to296 lg kg�1 ww, with a mean concentration of 60 ± 44 lg kg�1 ww(Table 2). As for DL-PCBs, the highest concentration was found inthe liver of cod from the east area, ranging from 9.7 to510 lg kg�1 ww, with a mean concentration of 110 ± 73 lg kg�1 -ww. In total, 62 samples of individual livers out of 781 had a con-centration of NDL-PCBs higher than 200 lg kg�1 ww, which is themaximum level set by the EU for NDL-PCBs for human consump-tion (EC, 2011). Five individual liver samples from the northwestarea, 24 liver samples from the southwest area and 33 liversamples from the east area exceeded a concentration of200 lg kg�1 ww (data not shown).

The concentrations of PCB6 in muscle samples of 30 individualNortheast Arctic cod were low and ranged from 0.6 to 1.3 lg kg�1 -ww, and 14 muscle samples showed concentrations of all the dif-ferent congeners less than or equal to the LOQ. For these samplesthe concentration of PCB6 was estimated to be 0.6 lg kg�1 ww,using ‘‘upper bound LOQ’’ (Table 4). The dominating congeners ofPCB6 in cod muscle were PCB-153 and PCB-138 (not shown). Theconcentration of the sum of PCB6 in muscle of Northeast Arcticcod was well below the maximum level of 75 lg kg�1 ww set bythe EU for human consumption. Cod muscle has a low fat content

258 K. Julshamn et al. / Marine Pollution Bulletin 75 (2013) 250–258

of approximately 1 g 100 g�1 ww (www.nifes.no/seafooddata), andin such samples low levels of fat-soluble compounds, such as PCBs,are expected (www.nifes.no/seafooddata).

The concentration of the sum of PBDE7 in liver samples of 774individual Northeast Arctic cod ranged from 0.20 to 36 lg kg�1 -ww, with a mean ± SD of 4.5 ± 3.5 lg kg�1 ww (Table 2). The meanconcentration of PBDE7 for the 32 positions ranged from 1.8 to12.8 lg kg�1 ww. The congener profile of PBDE7 showed thatPBDE-47 was the dominating congener with 75%, followed byPBDE-100 and PBDE-28 with 10% and 5%, respectively (data notshown). The proportions of each of the seven congeners were fairlyuniform among the samples (data not shown).

As for the other POPs, the lowest concentration of the sum ofPBDE7 was found in the livers of cod sampled in the northwestarea, ranging from 0.6 to 15 lg kg�1 ww, with a mean concentra-tion of 3.0 ± 2.1 lg kg�1 ww and the highest concentration wasfound in the livers of cod from the southwest area, ranging from1.4 to 36 lg kg�1 ww, with a mean concentration of6.1 ± 5.0 lg kg�1 ww (Table 2). There is currently no maximumlimit for the level of PBDEs in food.

The concentration of the sum of PBDEs found in this studyin the livers of Northeast Arctic cod was lower than that foundduring previous monitoring from 2005 to 2007, when the meanconcentration of PBDE7 in cod liver ranged from 6.1 to 9.4 lg kg�1 -ww (www.nifes.no/seafood). This might represent a decline in thelevel of the sum of PBDEs, since the production and use of the low-er brominated PBDEs were banned in the EU and Norway since2004 (de Wit et al., 2006). As for the other POPs, there were posi-tive correlations between the concentration levels of PBDE7 andthe parameters age, length and weight (Table 5). The concentrationof PBDE7 was strongly positively correlated to the sum of PCDD/Fsand DL-PCBs and also to the sum of NDL-PCBs, with an r-value of0.61 and 0.65, respectively (Table 5). Also the PCA plot showed thatall these substances covaried (Fig. 3). A likely explanation for thiscorrelation is that the contaminants are acquired from the samefood source and that net uptake occurs at comparable rates. Also,because these substances are all lipophilic, their seasonal varia-tions in concentration and dilution in the liver are regulated bythe same mechanisms. Frantzen et al. (2011) showed that PCDD/Fs, DL-PCBs, NDL-PCBs and PBDEs in NSS herring varied similarlyover the seasons, probably caused by maturation, spawning, feed-ing and starvation.

The concentration of the sum of PBDE7 in cod muscle was verylow. The only congener with concentrations higher than the LOQwas PBDE-47 with a mean of 0.009 lg kg�1 ww. The lower-boundLOQ was used for concentrations lower than the LOQ, thus concen-trations lower than the LOQ were set to zero.

Acknowledgements

The authors would like to thank the Norwegian Fisheries andAquaculture Industry Fund for funding this work. Annette Bjordaland co-workers are acknowledged for their skilful technical assis-tance, and we would also like to thank the personnel of the Insti-tute of Marine Research’s research vessels, who caught and agedthe fish.

References

Berntssen, M.H.G., Julshamn, K., Lundebye, A.K., 2010. Chemical contaminants inaquafeeds and Atlantic salmon (Salmo salar) following the use of traditional-versus alternative feed ingredients. Chemosphere 78, 637–646.

Box, G.E.P., Cox, D.R., 1964. An analysis of transformations. J. R. Stat. Soc. Ser. B –Stat. Meth. 26 (2), 211–252.

de Wit, C.A., Alaee, M., Muir, D.C.G., 2006. Levels and trends of brominated flameretardants in the Arctic. Chemosphere 64, 209–233.

European Commission, 2001. Opinion of the scientific committee on food on therisk assessment of dioxins and dioxin-like PCBs in Food. CS/CNTM/DIOXIN/20final.

European Commission, 2011. Commission regulation (EC) No 1259/2011of 2December 2011 amending Regulation (EC) No 1881/2006 as regardsmaximum levels for dioxins, dioxin-like PCBs and nondioxin-like PCBs infoodstuffs. Official Journal of the European Union L 320/18, 03.12.2011.

Frantzen, S., Maage, A., Julshamn, K., 2010. Baseline study of undesirable substancesin Northeast Atlantic mackerel (Scomber scombrus). Norw. Food Saf. Authority,33 (in Norwegian).

Frantzen, S., Måge, A., Iversen, S.A., Julshamn, K., 2011. Seasonal variation in thelevels of organohalogen compounds in herring (Clupea harengus) from theNorwegian Sea. Chemosphere 85, 179–187.

Green, N.W., Knutzen, J., 2003. Organohalogens and metals in marine fish andmussels and some relationships to biological variables at reference localities inNorway. Mar. Pollut. Bull. 46, 362–377.

Horst, K., Wiederholt-Lahrssen, M., 2009. Dioxin and dioxin-like PCB levels in cod-liver and -muscle from different fishing grounds of the North- and Baltic Seaand the North Atlantic. J. Consumer Prot. Food Saf. 4, 247–255.

Hylen, A., Nakken, O., Nedreaas, K., 2008. Northeast Arctic cod: fisheries, life history,stock fluctuations and management. In: Nakken, O. (Ed.), Norwegian Spring-spawning Herring and Northeast Arctic cod. 100 Years of Research andManagement. Tapir Academic Press, Trondheim, pp. 83–118.

ICES, 2008. Report of the Arctic Fisheries Working Group. ICES Document CM/ACOM:1, pp. 531.

JECFA, 2006. Polybrominated diphenyl ethers. Safety evaluation of certaincontaminants in food. Prepared by the 64th meeting of the Joint FAO/WHOExpert committee on Food Additives (JECFA).

Jimenez, B., Gonzalez, M.J., Hernandez, L.M., 1991. Presence of PCDDs and PCDFs infood for human consumption: a review. Food Chem. 39, 257–271.

Julshamn, K., Lundebye, A.K., Heggstad, K., Berntssen, M.H.G., Boe, B., 2004.Norwegian monitoring programme on the inorganic and organiccontaminants in fish caught in the Barents Sea, Norwegian Sea and North Sea,1994–2001. Food Addit. Contam. 21, 365–376.

Julshamn, K., Nilsen, B., Valdersnes, S., Frantzen, S., 2012. A survey of contaminantsin crabs (Cancer pagurus). Norw. Food Saf. Authority, 52 (in Norwegian).

Julshamn, K., Duinker, A., Nilsen, B.M., Frantzen, S., Maage, A., Valdersnes, S.,Nedreaas, K., 2013. A baseline study on levels of mercury, arsenic, cadmium andlead in Northeast Arctic cod (Gadus morhua) from different part of the BarentsSea. Mar. Pollut. Bull. 67, 187–195.

Kjesbu, O.S., Solemdal, P., Bratland, P., Fonn, M., 1991. Fecundity, atresia, and eggsize of captive Atlantic cod (Gadus morhua) in relation to proximate bodycomposition. Can. J. Fish. Aquat. Sci. 48, 2333–2343.

Nilsen, B.M., Frantzen, S., Nedreaas, K., Julshamn, K., 2010. A baseline study onundesirable substances in Greenland halibut (Reinhardtius hippoglossoides).Norw. Food Saf. Authority, 42 (in Norwegian).

NORA, 2003. Dioxin and dioxin like PCB in four commercially important pelagic fishstocks in the North East Atlantic Ocean. <http://www.nora.fo/docs/Dioxin_Final_report.pdf>.

Rappe, C., Anderson, R., Bergkvist, P.-A., Brohede, Ch., Hansson, M., Kjeller, L.O.,Lindstrøm, G., Marklund, S., Nygren, M., Swanson, S.E., Tysklind, M., Wiberg, K.,1987. Overview on environmental fate of chlorinated dioxins anddibenzofurans. Sources, levels and isomeric pattern in various matrices.Chemosphere 16, 1603–1618.

United States Environmental Protection Agency. USEPA, 1994. Method 1613: Tetra-through octa chlorinated dioxins and furans by isotope dilution HRGC/HRMS.EPA no 821-B-94-005.

United States Environmental Protection Agency, 1999. Method 1668 rev. A:Chlorinated biphenyl congeners in water, soil, sediment and tissue by HRGC/HRMS.: EPA no. 821-R-00-002.

Van den Berg, M., Birnbaum, L.S., Denison, M., De Vito, M., Farland, W., Feeley, M.,Fiedler, H., Hakansson, H., Hanberg, A., Haws, L., Rose, M., Safe, S., Schrenk, D.,Tohyama, C., Tritscher, A., Tuomisto, J., Tysklind, M., Walker, N., Peterson, R.E.,2006. The 2005 World Health Organization re-evaluation of human andmammalian Toxic Equivalency Factors for dioxins and dioxin-like compounds.Toxicol. Sci. 93, 223–224.

VKM, Norwegian Scientific Committee for Food safety, 2007. Opinion of the panelon contaminants of the Norwegian scientific committee for food safety. Riskassessment of dioxins and dioxin-like PCBs in fish liver. 22 November, p. 23.

Yaragina, N.A., Aglen, A., Sokolov, K.M., 2011. Cod. In: Jakobsen, T., Ozhigin, V.K.(Eds.), The Barents Sea – Ecosystem, Resources and Management. Half a Centuryof Russian–Norwegian Cooperation. Tapir Academic Press, Trondheim, pp. 223–270.

Zar, J.H., 2010. Pearson Education. New Jersey.Øiestad, V., 1994. Historical changes in cod stock and cod fisheries: Northeast Arctic

cod. ICES Mar. Sci. Symp. 198, 17–30.