Biomagnification of organohalogens in Atlantic salmon (Salmo salar) from its main prey species in three areas of the Baltic Sea

Science of the Total Environment 421-422 (2012) 129–143

Contents lists available at SciVerse ScienceDirect

Science of the Total Environment

j ourna l homepage: www.e lsev ie r .com/ locate /sc i totenv

Biomagnification of organohalogens in Atlantic salmon (Salmo salar) from its mainprey species in three areas of the Baltic Sea

Pekka J. Vuorinen a,⁎, Marja Keinänen a, Hannu Kiviranta b, Jaana Koistinen b,1, Mikko Kiljunen c,Timo Myllylä d, Jukka Pönni e, Heikki Peltonen f, Matti Verta f, Juha Karjalainen c

a Finnish Game and Fisheries Research Institute, P.O. Box 2, FI-00791 Helsinki, Finlandb National Institute for Health and Welfare (THL), P.O. Box 95, FI-70701 Kuopio, Finlandc University of Jyväskylä, Department of Biological and Environmental Science, P. O. Box 35, FI-40014 University of Jyväskylä, Finlandd Finnish Game and Fisheries Research Institute, Itäinen Pitkäkatu 3, FI-20520 Turku, Finlande Finnish Game and Fisheries Research Institute, Sapokankatu 2, FI-48100 Kotka, Finlandf Finnish Environment Institute, P. O. Box 140, FI-00251 Helsinki, Finland

⁎ Corresponding author. Tel.: +358 205 751 277; faxE-mail addresses: [email protected] (P.J. Vuorin

(M. Keinänen), [email protected] (H. Kiviranta), jaa(J. Koistinen), [email protected] (M. Kiljunen), [email protected] (J. Pönni), [email protected]@ymparisto.fi (M. Verta), juha.s.karjalainen@

1 Present address: University of Helsinki, Tvärminne Ztie 260, FI-10900 Hanko, Finland.

0048-9697/$ – see front matter © 2012 Elsevier B.V. Alldoi:10.1016/j.scitotenv.2012.02.002

a b s t r a c t
a r t i c l e i n f o
Article history:Received 9 November 2011Received in revised form 1 February 2012Accepted 2 February 2012Available online 3 March 2012

Keywords:DioxinPCBOrganobromineHerring Clupea harengusSprat Sprattus sprattusSalmon Salmo salar

Factors affecting the biomagnification of organohalogens in Baltic salmon from sprat, herring and three-spinedstickleback were assessed in three feeding areas. Second sea-year salmon contained (in fresh weight of wholefish) 79–250 ng g−1 polychlorinated biphenyls (ΣPCB), 0.9–2.7 pg g−1 dibenzo-p-dioxins (ΣPCDD), 8–19 pg g−1

dibenzofurans (ΣPCDF), 96–246 pg g−1 coplanar PCBs, 2.4–3.6 ng g−1 polybrominated diphenylethers (ΣPBDE),and 39–136 ng g−1 Σindicator PCB6. The EU limits forWHO toxic equivalent concentrations in fish feed were alreadyexceeded in one-year-old sprat and herring and were exceeded many-fold in older age groups. The differences inthe biomagnification rates of organohalogens in salmon appeared to be related to the feeding area, principal preyspecies, and the fat content and growth rate of the prey species.

© 2012 Elsevier B.V. All rights reserved.

1. Introduction

The Baltic Sea has been and remains much more polluted thanthe North Sea or the north-eastern Atlantic Ocean with respect tothe concentrations of persistent organic pollutants (POPs), e.g., organo-halogens, in the biota of these areas (Burreau et al., 2006; de Roodeet al., 2002; Falandysz et al., 1994; Karl and Ruoff, 2007). Althoughthe concentrations of organochlorines (OCs) such as polychlorinatedbiphenyls (PCBs) and DDTs in birds and fish in the Baltic Sea areahave substantially decreased over the last thirty years (Järnberg et al.,1993; Jorundsdottir et al., 2006; Kannan et al., 1992; Pikkarainen andParmanne, 2006; Szlinder-Richert et al., 2008), more persistent OCshave not decreased over the last two decades (Bignert et al., 2008;Kiviranta et al., 2003).

: +358 205 751 201.en), [email protected]@[email protected] (T. Myllylä),to.fi (H. Peltonen),jyu.fi (J. Karjalainen).oological Station, J.A.Palménin

rights reserved.

In the Baltic Sea, the main prey species of Atlantic salmon (Salmosalar, hereafter referred to as Baltic salmon or salmon) are sprat(Sprattus sprattus) and herring (Clupea harengus), which togetherhave constituted over 90% of the salmon diet in the Baltic Proper[BPr, Subdivisions (SDs) 24–28, excluding the Gulf of Riga, of the In-ternational Council for the Exploration of the Sea (ICES)] (Hanssonet al., 2001; Karlsson et al., 1999). While the proportions of thesetwo fish stocks have varied over the decades, sprat has been the prin-cipal prey species since the latter half of the 1980s (Mikkonen et al.,2011; Vuorinen et al., 2002). The proportion of herring in the salmondiet increases towards the northern Baltic Sea, and herring becomesthe dominant prey species in the Bothnian Sea (BS), where the pro-portion of sprat decreases to less than 5–10% (Hansson et al., 2001;Karlsson et al., 1999; Mikkonen et al., 2011). In the Gulf of Finland(GoF), herring was once the main prey species, but since 1996,when the biomass of sprat reached its highest level since the begin-ning of the ICES stock assessment in 1974 (ICES, 2009), sprat hasbeen the most numerous prey item in the GoF (Keinänen et al.,2000). The proportions of prey species in the salmon diet are evident-ly related to their spatial abundances (see ICES, 2007). The three-spined stickleback (Gasterosteus aculeatus) has been included in thesalmon diet in all three areas (Hansson et al., 2001; Karlsson et al.,1999; Keinänen et al., 2000).

http://dx.doi.org/10.1016/j.scitotenv.2012.02.002

mailto:[email protected]










http://dx.doi.org/10.1016/j.scitotenv.2012.02.002

http://www.sciencedirect.com/science/journal/00489697

Fig. 1. The sampling areas of salmon, sprat, herring, and three-spined stickleback [dottedellipses in the Bothnian Sea (ICES Subdivision, SD 30: salmon, sprat, herring, andthree-spined stickleback), the Baltic Proper (SD 28: salmon; SD 29: sprat, herring, andthree-spined stickleback), and the Gulf of Finland (SD 32: salmon, sprat, herring, andthree-spined stickleback)].

130 P.J. Vuorinen et al. / Science of the Total Environment 421-422 (2012) 129–143

Organohalogen concentrations in the Baltic Sea are generally higherin salmon than in their prey fish (Isosaari et al., 2006; Lizak et al., 2007;Vuorinen et al., 2002). In addition to their position in the food chain, thehigh fat content of salmon affects the bioaccumulation rate of lipid-soluble organohalogens in salmon (Larsson et al., 1996). However, thequality of prey fish appears to determine the content and patterns of ac-cumulated POPs in salmon (Svendsen et al., 2008; Vuorinen et al.,2002). The increased concentrations of coplanar PCBs and polychlori-nated dibenzofurans (PCDFs) in salmon that ascended the SimojokiRiver for spawning in 1991 (Vuorinen et al., 1997) coincided with arapid increase in the sprat stock of the Baltic Sea and a subsequent de-crease in the proportion of herring in the BPr (ICES, 2009; Mikkonenet al., 2011). It was hypothesised that a change in the salmon dietfrom herring to sprat was the primary reason for the increase in copla-nar PCBs and PCDFs, and this hypothesis was supported by the fact thatthe concentrations of specifically those OCs that were higher in spratthan herring were high in salmon. Moreover, on the basis of the conge-ner profiles of PCBs and polychlorinated dibenzo-p-dioxins and diben-zofurans (PCDD/Fs) in sprat and herring in 1994–1995 and in salmon,it was concluded that OCs in salmon were largely accumulated fromsprat (Vuorinen et al., 2002).

The accumulation of organohalogens in fish depends on their age(Parmanne et al., 2006; Perttilä et al., 1982; Vuorinen et al., 2002).However, the fresh weight-based concentrations of coplanar PCBsand PCDFs of Baltic sprat were not associated with age, but the age-relationship became evident after fat-normalisation (Vuorinen et al.,2002). There have been considerable variations in the concentrationsof POPs in salmon, herring, and sprat caught in different parts of theBaltic Sea (Karl et al., 2010; Karl and Ruoff, 2007; Simm et al., 2006;Szlinder-Richert et al., 2008). For example, in the BS, the EC limit(EC, 2006b) of dioxins for human consumption (WHO dioxin toxicequivalent concentration, WHO-TEQPCDD/F, 4 pg g−1 in fresh weight)has been exceeded in edible portions of herring older than 5 yearswith an average length of 16.6 cm (Parmanne et al., 2006). Becauseof biomagnification, the settled maximum tolerable limits in sub-stances and products intended to be used in animal feed, like unpro-cessed fish, are lower than the limit for food intended for humanconsumption (EC, 2006a). Salmon prefer prey fish with a length ofless than 19 cm, and mostly prey on fish with a length less than15 cm (Hansson et al., 2001). Therefore, all age groups of sprat andthree-spined stickleback are appropriate as salmon prey, but onlyyoung age groups of herring, i.e., mainly 1–3 years old, are suitable(Keinänen et al., 2000). However, changes in the growth of herringdue to food availability and feeding area may substantially influencethe upper age of the herring preyed on by salmon (Mikkonen et al.,2011, Fig. 1) and, consequently, their toxicant concentrations. Spratand smaller herring feed on the same zooplankton species (Casini etal., 2004; Möllmann et al., 2004). The increase in the size of thesprat stock since the 1980s reduced the growth rate and conditionfactor (CF) of herring and sprat in the BPr (Mikkonen et al., 2011)and, evidently, in the GoF as well (Peltonen et al., 2004) as a conse-quence of increased food competition (Casini et al., 2004). Theconcentrations of OCs were higher in whole fish samples of spratthan in herring (b16 cm) caught in the northern BPr in 1994–1995(Vuorinen et al., 2002).

The aim of the current study was to examine the biomagnificationof different organohalogens in Baltic salmon from their main preyspecies, sprat and herring, in the context of factors such as age, sea-son, and fat content, as well as from the three-spined stickleback asa potential source, in three feeding areas of salmon in the Baltic Sea:the BPr, BS, and GoF. Increased knowledge of the bioaccumulationof different organohalogens in prey species and their biomagnifica-tion in salmon will facilitate the future reduction of toxicant accumu-lation in Baltic fish. This knowledge would aid in planning andadjusting fishing strategies and thereby reducing the exposure ofthe biota, including humans, to organohalogens from Baltic fish

while ensuring the efficient exploitation of Baltic fish resources andthe sustainable development of fisheries.

2. Materials and methods

2.1. Fish samples

Baltic salmon (S. salar L.)were caughtwith drift nets from theGoF be-tween 24 November and 2 December 2003 and the BPr in the GotlandDeep in October 2003; ten salmon were caught from each area (Fig. 1).In addition, five salmonwere caught from the BS in October 2004. Imme-diately after capture, the salmon specimens were sealed in polyethylenebags and frozen at −20 °C. After thawing in the laboratory, the totallength and body weight were measured. To determine age, scalesabove the lateral line below the dorsal fin were removed. Both malesand femaleswere included in the salmon catches from all three samplingareas, and the proportion of males ranged from 0.4 to 0.6. However, sexwas not expected to affect the results because the salmon were caughtin late autumn during their feeding migration and hence not in thespawning season.

Specimens of Baltic sprat [S. sprattus (L.)] and herring (C. harengus L.)were collected from commercial catches (by midwater trawls) fromthe BS, the GoF, and the northern BPr (Fig. 1) during the last quarter of2003 (autumn) and the first quarter of 2004 (spring). Specimens ofthree-spined stickleback (G. aculeatus L.) caught as trawl bycatch werealso sampled from the same areas. The total length and weight of eachsprat and herring were measured, and otoliths were removed for agedetermination. Herring, sprat, and three-spined stickleback were indi-vidually sealed in numbered Mini-Grip® bags and immediately frozenat −20 °C.

The ages of the salmon were determined from scales, and the agesof the sprat and herring were determined from thin slices of otoliths

131P.J. Vuorinen et al. / Science of the Total Environment 421-422 (2012) 129–143

that were stained with a modification of the neutral red staining de-scribed by Richter and McDermott (1990). In the current study, the‘age’ of salmon refers to sea years, i.e., the number of growing seasonsa fish has been in the sea before capture (Salminen et al., 1994). TheCF of salmon, sprat, and herring was calculated from the equation:CF=100 W L−3, where W=body weight (g) and L=total length(cm). The sample data are presented in Table S1.

Whole specimens of salmon (N=25) were individually homoge-nised. After the age determination, the sprat (n=327 in autumn andn=503 in spring) and herring (n=971 in autumn and n=1311 inspring) specimens were pooled into age groups of 1, 2, 3, 6, and 10(9–11) years for herring and 1, 3 (2–3), and 7 (6–8) years for sprat foreach sampling area and season, but no 1-year-old sprat were caughtfrom the BS in spring. The total number of sprat and herring samplesor homogenates was therefore 47. The number of sprat and herringspecimens varied from 17 to 198 in the age groups or pools/homoge-nates for analysis. The pooled samples of sticklebacks weighed15–180 g and comprised one sample from the BPr, two from the BS,and four from the GoF. Sticklebacks in the pools had a mean weight of3.4 g and a mean length of 7.5 cm.

Salmon were homogenised with an industrial meat mincer (Metos,Bear No. 5), whereas the pooled samples of prey fish were preparedwith a smaller laboratory mincer (Tecator 1094 homogeniser). Homog-enates were thoroughly mixed, and samples were taken for organoha-logen analysis. Between each homogenisation, the equipment wasthoroughly washed and rinsed with ethanol. The efficiency of homoge-nisation for the analysis of the organohalogen content was evaluated byanalysing the concentrations of congeners of polychlorinated biphenyls(PCBs) in five subsamples of a salmon homogenate. The coefficient ofvariation of the different congeners was 1.2–10.3%.

2.2. Analysis of organohalogens

The concentrations of polychlorinated dibenzo-p-dioxins anddibenzofurans (PCDD/Fs), polychlorinated biphenyls (PCBs), andpolybrominated diphenylethers (PBDEs) were measured from thehomogenates of salmon, sprat, herring, and stickleback. The deter-mined compounds included 2,3,7,8-substituted PCDD/Fs (17 conge-ners) and 37 PCB congeners (PCB 18, 28, 33, 47, 49, 51, 52, 60, 66,74, 77, 81, 99, 101, 105, 110, 114, 118, 122, 123, 126, 128, 138, 141,153, 156, 157, 167, 169, 170, 180, 183, 187, 189, 194, 206, and 209),including 4 non-ortho-PCBs (coplanar/CoPCBs, in bold in the list)and 8 mono-ortho-PCB congeners (in italics in the list), both ofwhich are referred to as dioxin-like PCB (dl-PCB) congeners. Thesum concentration for indicator PCBs (IndPCB6: PCB 28/31, 52, 101,138, 153, and 180) was additionally calculated. Subsequently in thetext, the abbreviation PCB or ∑PCB refers to the sum of all 37 PCBcongeners. Furthermore, 14 PBDE congeners were measured (BDE28, 47, 66, 71, 75, 77, 85, 99, 100, 119, 138, 153, 154, and 183).

The fish homogenate was freeze-dried and Soxhlet-extracted withtoluene, and the lipid content of the sample was determined gravimet-rically. The OC concentrations were analysed as previously described(Isosaari et al., 2005; Kiviranta et al., 2004). In brief, the lipidswere sep-arated on amultilayer silica gel column that included acidic and neutrallayers, and the analytes were eluted with 1:1 (v:v) dichloromethane/cyclohexane. After elution, PCDD/Fs were separated from PCBs andPBDEs on a carbon-Celite column by first eluting PCBs and PBDEs with1:1 (v:v) dichloromethane/cyclohexane, then back-eluting PCDD/Fswith toluene. Both fractions were further purified over an alumina col-umn. Impurities were removed from the PCDD/F fraction with hexaneand 2% dichloromethane in hexane, and PCDD/Fs were then elutedwith 20% dichloromethane in hexane. The fraction containing PCBsand PBDEswas purified on an alumina column by eluting the impuritieswith hexane and eluting the analytes with 2% dichloromethane in hex-ane. After the alumina column, the CoPCBs were separated from theother PCBs and PBDEs on an activated carbon column by eluting the

other PCBs and PBDEs with 1:1 (v:v) dichloromethane/hexane andback-eluting the CoPCBs with toluene.

The three sample fractions obtained, which contained (1) PCDD/Fs,(2) CoPCBs, and (3) other PCBs and PBDEs, were analysed by high-resolution gas chromatography/high-resolution electron ionisationmass spectrometry (HRGC/EI-HRMS). The HRMSwas operated in selec-tive ion monitoring (SIM) mode with a resolution of 10000. The inves-tigated compounds were quantified through the use of 13C-labelledinternal standards that were spiked into the sample extracts beforesample preparation.

2.3. Quality assurance and control

Chemical analyseswere carried out in the Chemical Exposure Unit ofthe National Institute for Health and Welfare, which has been accre-dited according to the ISO 17025 standard. The scope of accreditationincludes PCDD/F, PCB, and PBDE analyses from biological matrices.The QA/QC procedures included participation in interlaboratory com-parisons of fishmatrices and the analysis of procedural blanks and a ref-erence fish oil. The recoveries of the individual internal standards ofPCDD/F, PCB, and PBDE congeners were determined by adding the re-covery standards just before mass spectral analysis. The recoveries ofthe internal standards were all above 60%. The limits of quantification(LOQ) for individual congeners, which correspond to a signal-to-noiseratio of 3:1, were 0.1–1.5 pg/g for PCDD/Fs, 0.1–0.5 pg/g for CoPCBs,0.01–0.3 ng/g for other PCBs, and 0.01–2.0 ng/g for PBDEs, eachexpressed per lipid weight of the sample. The method uncertaintiesfor both WHO-TEQPCDD/F and WHO-TEQPCB were below 20%. An in-house laboratory control sample prepared from Baltic herring was runwith every analysis batch with an allowable variation of b10%. Thus,the analyses of the present samples, which were performed in fourbatches within one year and within a few months after the collectionof each of the samples, were comparable.

The concentrations of all of the compound groups are so-called lower-bound values, which indicate that congeners below LOQ are included inthe sum as zero on a freshweight basis (the difference from upperboundvalues was practically zero, on average −0.6%). The EC directive foranimal feed (2006a) is based on the toxic equivalency factors (TEFs)recommended by the WHO, i.e., from Van den Berg et al. (1998), andthey were used in calculations. However, WHO-TEQ concentrationswere also calculated using the new TEFs (Van den Berg et al., 2006),and the TEQ concentrations calculated by both TEFs are given in Table S2.

2.4. Calculations and statistics

The concentrations of organohalogens in salmon are calculated ona fresh weight basis as arithmetic means of each sea-age group withstandard errors. Those are also given in Table S2 as well as themean (with SEs and ranges) concentrations of POPs and TEQs in allsprat, 1–3-year-old herring, and stickleback by areas with seasonscombined. For congener profiles, the percentage proportions of differ-ent types of organohalogen congeners were calculated with autumnand spring samples combined for sprat and herring. The organohalo-gen results were tested for normality with the Kolmogorov–Smirnovtest, and log-transformation was applied when appropriate. Variancehomogeneity was tested with Bartlett's test. For the calculations, thearcsine transformation was applied for percentage values.

In the EC directive (2006a), the maximum allowable concentra-tions in unprocessed fish intended to be used as animal feed arethe WHO-TEQ concentrations, which are 1.25 ng kg−1 for PCDD/Fsand 4.5 ng kg−1 for PCDD/Fs+dl-PCBs. These values are determinedwith a moisture content of 12%. Fresh fish, however, contain morewater. In sprat and herring, the total mean water content was75.4%. Thus, the EC limit concentrations of WHO-TEQs in fresh sprator herring are approximately 0.35 ng kg−1 for PCDD/Fs and

Fig. 2. Mean (± SE) concentrations of organohalogens and WHO-TEQs for PCDD/Fs andPCDD/Fs plus dioxin-like-PCBs (in fresh weight) in 2nd-sea-year salmon from threeareas in the Baltic Sea: the Baltic proper (BPr), the Bothnian Sea (BS), and the Gulf ofFinland (GoF). A different letter above the column indicates a significant (Tukey's test,pb0.05) difference between the means. The percentage fat contents were 16.1±0.7,


approximately 1.26 ng kg−1 for PCDD/Fs+dl-PCBs; thus, thesevalues would apply if unrefined sprat or herring was used for foodin aquaculture.

Biomagnification factors (BMF) from sprat, herring, or three-spinedstickleback to 2nd-sea-year salmon in each areawere calculated by divid-ing the respective toxicant concentration in salmon by that in preyfish both on a fresh weight and fat content basis. The mean toxicant con-centrations were calculated for sprat of all age groups, herring of agegroups of 1–3 years, and three-spined stickleback by areas with seasonscombined.

A linear regression model was applied to test the dependence oforganohalogen concentrations on the age of sprat and herring (all agegroups included; i.e., 1–10-year-old herring) and salmon from the GoFwith the fresh weight concentrations untransformed for the best fit.Significant differences in the slope were tested between the areas andspecies. Pearson correlation coefficients were calculated to evaluatedependences between the concentrations of organohalogens (on afresh weight basis) and the total length, body weight, CF, and fat con-tent of sprat, herring, and salmon. Two-way ANOVA (LS means as apost-hoc test) was applied to test differences in the total length, bodyweight, CF, and fat content between prey species (sprat and herring)and areas. To evaluate species, spatial, and seasonal differences inweight, length, CF, fat content, and organohalogen concentrations ofsprat and herring, multi-way ANOVA with Tukey's test (pb0.05) as apost-hoc test was applied. The differences in organohalogen concentra-tions between the areas of 2nd-sea-year salmon and between 2nd-sea-year salmon and their prey species, sprat (all age groups), herring(1–3-year-olds), and stickleback, with autumn and spring samplescombined, were compared by one-way ANOVA plus Tukey's test at asignificance level of pb0.05. ANCOVAwith the fat content as a covariate,followed by the LS means test for significant differences between themeans in the three areas, was also applied to evaluate organohalogenconcentrations in 2nd-sea-year salmon. The statistical tests were per-formed with the Statistical Analysis System (SAS version 9.2) software(SAS Institute Inc., 2003). The figures were drawn by Origin (version8.1; OriginLab Co.).

16.7±0.3 and 12.6±1.3 for salmon from the BPr, BS, and GoF, respectively.

3. Results

3.1. Salmon

3.1.1. Spatial relationships of POPs in salmon

The mean concentrations of the organohalogens and TEQs for eachage group by area are given in Table S2. In 2nd-sea-year salmon, theconcentration of PCBs (in freshweight) was lowest in the BPr and high-est in the GoF, where it was almost two times as high as in the BPr(Fig. 2A). In the BS, the concentration of PCBs was also significantlyhigher than in the BPr but considerably and significantly lowerthan in the GoF. The concentrations of IndPCB6 followed a similar pat-tern (range): 39–56 ng g−1 in the BPr, 69–90 ng g−1 in the BS, and92–136 ng g−1 in the GoF. By contrast, the concentration of CoPCBswas highest in salmon from the BPr; the difference was significantwhen compared to salmon caught from the BS (Fig. 2B).

Like the concentrations of PCBs, the concentrations of PCDDs insalmon in the GoF were almost twice as high as in salmon feedingin the BPr and also significantly higher than in salmon caught fromthe BS (Fig. 2C). The concentrations of PCDFs and PCDD/Fs were sig-nificantly lower in salmon from the BPr than in those from theother two areas (Fig. 2D). There were no significant differences be-tween the areas in the concentrations of PBDEs (Fig. 2E).

When comparing the TEQ concentrations in 2nd-sea-year salmonin different areas, the WHO-TEQPCDD/F was considerably lower(Tukey's test, pb0.05) in the BPr than in the GoF and BS, and theWHO-TEQPCDD/F+PCB was lower (pb0.05) in salmon from either theBPr or the BS than from the GoF (Fig. 2F and Table S2).

Taking the fat content as a covariate in 2nd-sea-year salmon, theconcentrations of other POPs, except CoPCBs and PBDEs, were signif-icantly related to the area (least square means at the following signif-icance levels: PCBs, pb0.001; PCDDs, pb0.01; PCDFs, pb0.05; PCDD/Fs, pb0.05). The WHO-TEQ concentration of PCDD/Fs was highly sig-nificantly (pb0.001) related to the area and also related (pb0.05) tothe fat content, while the WHO-TEQPCDD/F+PCB was also related tothe area (pb0.01) but not the fat content (p>0.05).

3.1.1. The relationship between POPs and salmon sea age, fat content,and size

In the GoF, where salmon of three different sea years (1st, 2nd,and 3rd sea years; n=5, 4, and 1, respectively) were caught, theage-dependencies of all of the organohalogen concentrations aswell as the WHO-TEQ concentrations (WHO-TEQPCDD/F and WHO-TEQPCDD/F+PCB) were significant (Figs. 3–6 and Table 1). The age de-pendency was least clear for the concentration of CoPCBs, whichwas also evident in salmon data from the BPr (9 2nd-sea-year and1 3 rd-sea-year fish). The strongest age dependency was in theWHO-TEQPCDD/F. The 2nd-sea-year salmon had a WHO-TEQPCDD/F

concentration of 3.9–7.6 pg g−1 fresh weight (on a whole fish basis)and a WHO-TEQPCDD/F+PCB concentration of 11–17 pg g−1. The con-centrations depended on the area. In the GoF, the concentrations ofWHO-TEQPCDD/F and WHO-TEQPCDD/F+PCB in the 1st-sea-year salmon(n=5) were 4.1±0.5 pg g−1 and 10.0±1.3 pg g−1 and 9.4 pg g−1

and 21 pg g−1 in one 3rd-sea-year salmon, respectively. In the 3rd-sea-year salmon from the BPr, the WHO-TEQPCDD/F and WHO-

image of Fig.�2

Fig. 3. Dependence of ΣPCB and ΣCoPCB concentrations (in fresh weight) on age/sea years for sprat, herring and salmon with the linear models indicated and ΣPCB concentration inthree-spined stickleback from the three sea areas.


TEQPCDD/F+PCB concentrations were slightly lower, 8.0 pg g−1 and18 pg g−1, respectively.

Among the 2nd-sea-year salmon in the BPr (n=9), the concentra-tions of all of the organohalogens, except PCBs, correlated positivelywith the fat content (Fig. 7 and Table 2). The most pronounced correla-tion with fat content occurred with the concentrations of PBDEs andCoPCBs. Only the PCB concentrationwas positively related to theweightand length of salmon in the BPr, while theWHO-TEQPCDD/F+PCBwas alsopositively related to their length, but no significant positive correlationswere detected with the CF (Table 2). On the contrary, in the salmon ofthe BS (n=5), there was a positive correlation between the CF andthe concentrations of all of the POPs, excluding CoPCBs, which alonehad a positive sign in the correlation with the fat content (Table 2). Inthe BS, the WHO-TEQPCDD/F and WHO-TEQPCDD/F+PCB were both posi-tively related to the CF, and theWHO-TEQPCDD/F+PCB was positively re-lated to the weight of the salmon. In the salmon from the GoF (n=4),the only significant and positive correlations were detected betweenthe concentrations of the CoPCBs and PBDEs and the weight and lengthof the fish (Table 2).

Themean length,weight, CF, and fat content of all age groups of salm-on in the study areas are given in Table S1. There were no other signifi-cant differences in those parameters between areas, but the fat contentwas significantly lower in the GoF than in the BPr and BS: 12.6 ±1.3%,16.1±0.7%, and 16.7±0.3%, respectively. The fat content increased as afunction of age in salmon caught from the GoF (pb0.05).

3.2. Prey species of salmon

3.2.1. The relationship between POPs and prey species, area, and seasonAll of the individual POP concentrations are presented in Figs. 3–6,

and the mean values by area with seasons combined are presented in

Table S2. Among all age groups of sprat and 1–3-year-old herring (i.e.,the main prey items preferred by salmon), the concentrations ofCoPCBs and IndPCB6 differed (multi-way ANOVA, pb0.001 andpb0.05, respectively) between the prey species and were higher insprat than in herring (Tukey's test, pb0.05). Smaller interspecific dif-ferences (multi-way ANOVA, pb0.05) were found in the concentra-tions of PCBs and PBDEs; both were similarly higher in sprat than inherring, but no such differences were observed for PCDDs or PCDFs.Of the POPs in sprat and herring, only the concentration of PCDDswas related to the area and was highest in the GoF, where the levelsof PCDDs were significantly higher (Tukey's test, pb0.05) than inthe BS. A seasonal difference was observed only for CoPCB, whichwas higher (pb0.05) in autumn than in spring. There were no signif-icant differences in the WHO-TEQ concentrations between species,areas, or seasons.

There were no significant differences (ANOVA, p>0.05) in toxicantconcentrations between three-spined stickleback and herring or spratin the GoF; in the other two areas, the number of stickleback sampleswas too low for comparisons. The concentrations of organohalogens inthe stickleback tended to be higher in spring than in autumn (Figs. 3–6).

3.2.2. The relationship between POPs and ageThe age-related increase in the PCB concentration was significant in

sprat and herring in all three study areas (Fig. 3 and Table 1). Similar butless conspicuous age relationships were detected for CoPCBs; in sprat,the increase was only significant in the BS, and in herring the relation-ship in the GoF was the weakest among POPs and area (Fig. 3 andTable 1). The concentration of PCDDs in sprat and herring increased sig-nificantly as a function of age, except in sprat from the BPr (Fig. 4 andTable 1). The age-related increase in the PCDF concentrationwas highlysignificant in herring in all of the areas, but in sprat the age-related

image of Fig.�3

Fig. 4. Dependence of ΣPCDD and ΣPCDF concentrations (in fresh weight) on age/sea years for sprat, herring and salmon with the linear models indicated and ΣPCDD concentrationin three-spined stickleback from the three sea areas.


increase was only significant in the BS (Fig. 4 and Table 1). The positiverelationship between the concentration of PBDEs and the age of spratand herring from the three study areas was significant in all cases(Fig. 5 and Table 1).

The age-related increase in the concentration of POPs in sprat wassimilar or less clear than in herring (Table 1 and Figs. 3–6): In the BS,all the POPs increased significantly (pb0.05–0.001) with age moreclearly in herring than in sprat, and the difference was most clearfor PCBs (pb0.001), followed by CoPCBs and WHO-TEQPCDD/F+PCB

(pb0.01). In the BPr, only the accumulation of PCBs and PBDEs was

Fig. 5. Dependence of the ΣPBDE concentration (in fresh weight) on age/sea year for sprat,three-spined stickleback from the three sea areas.

more (pb0.05) age-dependent in herring than in sprat. In the GoF,PCDDs more clearly accumulated with age (pb0.001) in herringthan in sprat, and the differences were also significant for PCDFs(pb0.05) and WHO-TEQPCDD/F (pb0.01).

In herring in the BS, the age-related increase in the concentration ofCoPCBs and PCDF as well as WHO-TEQPCDD/F and WHO-TEQPCDD/F+PCB

wasmore pronounced than in the other two sea areas, and the increasein PCDD was greater in the BS than in the BPr (Table 1 and Figs. 3–6).There were no significant differences in the age-related accumulationof organohalogens in sprat among the sea areas.

herring and salmon with the linear models indicated and the ΣPBDE concentration in

image of Fig.�4

image of Fig.�5

Fig. 6. Dependence of WHO-TEQPCDD/F and WHO-TEQPCDD/F+PCB concentration (in fresh weight with TEFs from Van den Berg et al., 1998) on age/sea year of fish for sprat, herringand salmon from the three sea areas with the linear models indicated. Horizontal lines indicate the maximum allowable concentration of WHO-TEQPCDD/F (approximately1.1 pg g−1 in fresh weight) and WHO-TEQPCDD/F+PCB (approximately 3.8 pg g−1) set by the EC for unprocessed fish intended for animal feed, e.g., in aquaculture (EC, 2006a).


The concentration ofWHO-TEQPCDD/F was 0.8–4.1 pg g−1 in the 1–7-year-old sprat and 0.8–3.1 pg g−1 in the 1–3-year-old herring, and theconcentration of WHO-TEQPCDD/F+PCB was 1.6–8.5 pg g−1 in sprat and1.6–5.1 pg g−1 in herring (Fig. 6 and Table S2).Thus, the maximumallowable concentrations, corrected for fresh weight, of dioxins (WHO-TEQPCDD/F, 0.35 pg g−1) and dioxins plus dl-PCBs (WHO-TEQPCDD/F+PCB,1.26 pg g−1) in fish feed, as set by the EC, was already exceeded in1-year-old sprat and herring and was clearly exceeded in older agegroups (Fig. 6, with the EC limit values indicated by horizontal lines).

3.2.3. The relationship between POPs and body size and fat contentWhen the 1–3-year-old herring from the three study areas were con-

sidered separately, the concentrations of all OCs (but not those of organo-bromines PBDEs) in the BS and BPr correlated positively with the weightand length of the fish (Table 3). In the GoF, only the concentrations ofPCDFs, PCDD/Fs, and WHO-TEQPCDD/F+PCB displayed such a correlation.Among all POPs and areas, only the concentration of CoPCBs in the GoFhad a positive correlationwith the fat content and CF of herring (Table 3).

In sprat from the BS, the concentrations of all of the POPs correlat-ed positively with the length and negatively with the CF of the fish,and the concentration of PCBs also correlated positively with theweight of the fish (Table 4). In the BPr and GoF, the PCB and PBDE

concentrations correlated positively with the weight and length ofsprat. Moreover, the concentration of PCDDs in sprat correlated posi-tively with weight in the GoF and negatively with the CF in the BPr. Inaddition, in the BS there was a significant negative correlationbetween the fat content of sprat and the concentrations of all othertoxicants except CoPCBs (Table 4).

The total length and body weight of 1–3-year-old herring andall age groups of sprat are given in Table S1. Those parameters didnot differ between the sea areas (p>0.05), although there werespecies-specific differences in size (pb0.05, two-way ANOVA):among the appropriate-sized prey for salmon, the mean length andbody weight of herring were higher than those of sprat (LS meanstest, pb0.05). By contrast, the CF of the prey species (sprat and her-ring; Table S1) was related to the sea area (pb0.05): the CF washigher in both the BPr and the BS than in the GoF (pb0.05). Boththe sea area (pb0.01) and species (pb0.01) affected the fat contentof salmon prey fish (Table S1): the fat content was significantlyhigher in prey in the BS than in the GoF and higher in sprat than inherring (pb0.05). The fat content of stickleback (Table S1) was higherthan that of herring in the BS (pb0.05), but there were no other sig-nificant differences (p>0.05) in the fat content between sticklebackand herring or sprat in any areas.

image of Fig.�6

Table 1Parameters (intercept, a±SE, and slope, b±SE) of linear regression equations of organohalogen concentrations (in fresh weight) as a function of the age of sprat and herring andsea years of salmon from three areas in the Baltic Sea. A different letter as a superscript to the SE value of coefficient b indicates a significant difference in the slopes between theareas, and a different capital letter indicates a significant (pb0.05) difference in the slopes between species within the areas. The coefficients of determination (R2) are given, andthe p value indicates the statistical significance of the regression model (ns=nonsignificant).

a/b Baltic proper Bothnian Sea Gulf of Finland

R2 pb R2 pb R2 pb

SpratΣPCB a 8.21±4.19 0.888 0.01 4.88±5.20 0.994 0.001 10.94±4.19 0.825 0.05

b 5.35±0.94k A 5.16±1.08k A 5.36±0.94k AΣCoPCB a 44.0±20.2 0.237 ns 30.4±25.1 0.788 0.05 45.2±20.2 0.290 ns

b 4.13±4. 6k A 3.90±5.19k A 8.36±4.56k AΣPCDD a 0.91±0.24 0.084 ns 0.07±0.29 0.854 0.05 0.65±0.24 0.737 0.05

b 0.04±0.05k A 0.14±0.06k A 0.17±0.05k AΣPCDF a 3.63±1.35 0.347 ns 2.70±1.68 0.877 0.05 3.03±1.35 0.476 ns

b 0.39±0.30k A 0.53±0.35k A 0.80±0.30k AΣPBDE a 0.28±0.10 0.921 0.01 0.25±0.13 0.959 0.01 0.42±0.10 0.815 0.05

b 0.15±0.02k A 0.14±0.03k A 0.12±0.02k AWHO-TEQPCDD/F a 1.01±1.22 0.664 0.05 0.73±0.49 0.901 0.05 0.85±0.65 0.723 0.05

b 0.22±0.27k A 0.24±0.10k A 0.36±0.15k AWHO-TEQPCDD/F+PCB a 1.75±1.69 0.713 0.05 1.23±0.67 0.959 0.01 1.74±1.39 0.713 0.05

b 0.55±0.38k A 0.54±0.14k A 0.73±0.31k A

HerringΣPCB a −0.21±5.34 0.787 0.001 −3.60±5.34 0.978 0.001 3.80±5.34 0.927 0.01

b 7.39±0.98k B 10.1±1.0k B 8.48±1.0k AΣCoPCB a 18. 5±5.17 0.730 0.01 12.4±5.17 0.982 0.001 26.9±5.17 0.571 0.05

b 5.44±0.94k A 8.75±0.94m B 3.46±0.94k AΣPCDD a −0.21±0.28 0.722 0.01 −0.33±0.28 0.982 0.001 −0.10±0.28 0.947 0.001

b 0.34±0.05k A 0.53±0.05m B 0.47±0.05km BΣPCDF a −1.06±1.38 0.762 0.001 −0.57±1.38 0.989 0.001 0.46±1.38 0.948 0.001

b 2.01±0.25k A 2.83±0.25m B 1.76±0.25k BΣPBDE a 0.13±0.20 0.702 0.01 0.20±0.20 0.894 0.001 0.13±0.20 0.797 0.001

b 0.19±0.04k B 0.25±0.04k B 0.18±0.04k AWHO-TEQPCDD/F a −0.95±0.89 0.794 0.01 −0.91±0.29 0.984 0.001 −0.34±0.48 0.957 0.001

b 0.96±0.16k A 1.33±0.05m B 0.95±0.09k BWHO-TEQPCDD/F+PCB a −0.76±1.23 0.768 0.001 −0.86±0.40 0.985 0.001 0.39±1.02 0.965 0.001

b 1.37±0.22k A 1.89±0.07m B 1.36±0.19k A

SalmonΣPCB a 67.9±20.9 0.610 0.01

b 62.4±12. 1 BΣCoPCB a 63.9±24.2 0.532 0.05

b 45.7±114.0 BΣPCDD a 0.62±0.24 0.736 0.01

b 0.82±0.14 BΣPCDF a 3.05±1.84 0.758 0.01

b 6.21±1.06 CΣPBDE a 0.71±0.31 0.732 0.01

b 0.96±0.18 BWHO-TEQPCDD/F a 1.32±0.74 0.793 0.001

b 2.94±0.43 CWHO-TEQPCDD/F+PCB a 4.28±1.58 0.734 0.01

b 6.05±0.91 B


3.3. Salmon vs. its prey species

3.3.1. Spatial and age-related differencesIn salmon, taking into account only the most numerous year class,

the 2nd sea year, the organohalogen concentrations in all areas werehigher or tended to be higher than in the prey species (all age groupsof sprat, 1–3-year-old herring, or stickleback; Figs. 3–6, Table S2). Thefat content of salmon was higher (pb0.05) than in herring in all areas,but only in the BPr significantly higher than in sprat, and no signifi-cant (p>0.05) differences were detected with stickleback (Table S1).

In the BPr, the concentrations of WHO-TEQPCDD/F, WHO-TEQPCDD/F+PCB, and all other organohalogens (i.e., PCDFs, PCDD/Fs,PCBs, CoPCBs, and PBDEs), except PCDDs, were significantly higher(ANOVA, Tukey's test pb0.05) in 2nd-sea-year salmon than in sprat,herring, or three-spined stickleback (Figs. 3–6, Table S2). However,there were no significant differences (p>0.05) between sprat and her-ring. By contrast, the concentration of PCDDs in salmon in the BPr did

not differ significantly from that in sprat; the concentration of PCDDswas higher in both sprat and salmon than in herring (Fig. 4, Table S2).

In the BS, the concentrations of all of the organohalogens aswell as theWHO-TEQs were significantly higher in salmon than in herring or sprat(Figs. 3–6, Table S2). The concentrations of other POPs, except PCDDs,were also significantly higher in salmon than in three-spined stickleback.The differences between the prey species were not significant.

In the GoF, the concentrations of all the POPs andWHO-TEQs weresignificantly higher in salmon than in sprat, herring, or stickleback(Figs. 3–6, Table S2). The differences between the prey species werenot significant. The concentrations of almost all POPs (on a freshweight basis) increased more significantly with sea year in salmonthan with age in herring and sprat (Figs. 3–6 and Table 1): the differ-ence in the slope was highly significant (pb0.001) for both prey spe-cies for the concentrations of PCDFs, PCBs, PBDEs, WHO-TEQPCDD/F,and WHO-TEQPCDD/F+PCB. A less clear difference was detected forCoPCBs, particularly with sprat (pb0.05) but also with herring

Fig. 7. The concentrations of organohalogens in fresh weight in the 2nd-sea-year salmonfrom the three sea areas against the fat content (on whole fish basis). A regression linewas fitted to the Baltic Proper data (n=9), and the coefficient of determination and signifi-cance of the regression model is given.


(pb0.01), and for PCDDs with sprat (pb0.01). No significant differ-ence (p>0.05) was observed between salmon and herring in theage dependency of the accumulation of PCDDs (Table 1).

Table 2Pearson correlation coefficients in 2nd-sea-year salmon between toxicants (in fresh weightstudy areas.

ΣPCDD ΣPCDF ΣPCDD/F ΣCoPCB

Baltic ProperLength 0.572 0.566 0.568 0.636⁎⁎⁎

Weight 0.554 0.561 0.561 0.598CF 0.185 0.230 0.225 0.071Fat 0.728⁎ 0.711⁎ 0.714⁎ 0.861⁎⁎

Bothnian SeaLength 0.696 0.657 0.661 0.630Weight 0.868 0.842 0.845 0.576CF 0.942⁎ 0.944⁎ 0.944⁎ 0.379Fat −0.807 −0.812 −0.811 0.164

Gulf of FinlandLength 0.672 0.771 0.770 0.996⁎⁎

Weight 0.675 0.773 0.772 0.994⁎⁎

CF −0.283 −0.365 −0.366 −0.595Fat −0.482 −0.383 −0.385 0.106

⁎ pb0.05.⁎⁎ pb0.01.⁎⁎⁎ pb0.001.

3.3.2. BiomagnificationBiomagnification factors (BMFs), indicating accumulation from

the prey species to 2nd-sea-year salmon, varied from 1.23 to 12.4on a fresh weight basis and from 0.70 to 9.21 on a fat weight basis(Table 5). In both the GoF and the BS, the BMF values from all preyspecies were highest for PCBs on both a fat and fresh weight basis,while in the BPr the highest BMF values were for PCBs, PBDEs, orCoPCBs. In the BPr, the BMF values from all prey species were lowestfor PCDDs, while in the BS and the GoF, the lowest values from theprey species were either for PCDDs or CoPCBs on both a fat andfresh weight basis. On average, the BMF values tended to be lowestin the BPr and highest in the GoF, irrespective of the calculation meth-od (Table 5). The BMF value (on a fresh weight basis) for 2,3,7,8-tet-rachloro dibenzo-p-dioxin (2,3,7,8-TCDD) in the 2nd-sea-yearsalmon from sprat was 4.8 in the BS and 2.7 in the other two seaareas and 4.2 in the GoF, 3.4 in the BPr, and 3.0 in the BS from herring.

3.3.3. Congener proportionsThe two most common congeners of PCDD/Fs in salmon, sprat,

and herring were 2,3,7,8-tetrachloro-dibenzofuran (2,3,7,8-TCDF)and 2,3,4,7,8-pentachloro-dibenzofuran (2,3,4,7,8-PeCDF), with theproportion of the penta-congener increasing as a function of fishage (Figs. 8–10).

In 2nd-sea-year salmon, the proportion of 2,3,7,8-TCDF was signif-icantly (ANOVA, Tukey's test, pb0.05) higher in the BPr (42%) than inthe BS (30%) or GoF (26%). In turn, the proportion of 2,3,4,7,8-PeCDFin the 2nd-sea-year salmon was significantly (pb0.05) lower (37%) inthe BPr than in the BS (53%) and GoF (56%) (Fig. 8).

In one-year-old sprat in all areas, the proportion of 2,3,7,8-TCDFwas higher (34–48%) than that of 2,3,4,7,8-PeCDF (27–35%). In7-year-old sprat, the proportion of 2,3,7,8-TCDF was 18–29%, andthe proportion of 2,3,4,7,8-PeCDF was 45–47% (Fig. 9).

In one-year-old herring, the proportion of 2,3,7,8-TCDF was higherthan that of 2,3,4,7,8-PeCDF in the BPr and the BS, but in the GoF, theopposite relationship was observed (Fig. 10). In ten-year-old herring,the proportion of the penta-congener was 55% in the BPr and the BSand 58% in the GoF; the proportion of the tetra-congener was lowestin the GoF, 4%, and 15% in the other two areas.

The ratio of the penta-congener to the tetra-congener was highestin the Gulf of Finland in salmon, sprat, and herring of a similar age. In2nd-sea-year salmon, the ratio was close to those in 2-year-old spratand herring, except in the Bothnian Sea, where the value in salmonwas twice that in its prey species.

) and the body weight, length, condition factor (CF) and fat content of fish in the three

ΣPCB ΣPBDE WHO-TEQPCDD/F WHO-TEQPCDD/F+PCB

0.714⁎ 0.656 0.606 0.675⁎

0.686⁎ 0.662 0.600 0.6590.188 0.261 0.238 0.1900.659 0.883⁎⁎ 0.694⁎ 0.808⁎⁎

0.691 0.727 0.674 0.7320.847 0.860 0.856 0.890⁎

0.921⁎ 0.887⁎ 0.950⁎ 0.944⁎

−0.549 −0.545 −0.817 −0.701

−0.384 0.992⁎⁎ 0.192 0.276−0.375 0.976⁎ 0.202 0.286

0.415 −0.725 0.088 0.031−0.808 0.103 −0.739 −0.709

image of Fig.�7

Table 3Pearson correlation coefficients in 1–3-year-old herring between toxicants (in fresh weight) and the mean body weight, length, condition factor (CF) and fat content of fish in thethree study areas.

ΣPCDD ΣPCDF ΣPCDD/F ΣCoPCB ΣPCB ΣPBDE WHO-TEQPCDD/F WHO-TEQPCDD/F+PCB

Baltic properLength 0.963⁎⁎ 0.918⁎⁎ 0.931⁎⁎ 0.818⁎ 0.904⁎ 0.105 0.873⁎ 0.901⁎

Weight 0.909⁎ 0.907⁎ 0.914⁎ 0.847⁎ 0.911⁎ 0.067 0.828⁎ 0.865⁎

CF 0.254 0.172 0.185 0.120 0.211 −0.722 0.011 0.078Fat 0.284 0.310 0.309 0.325 0.334 −0.517 0.127 0.194

Bothnian SeaLength 0.897⁎ 0.898⁎ 0.902⁎ 0.861⁎ 0.823⁎ 0.556 0.870⁎ 0.864⁎

Weight 0.956⁎⁎ 0.936⁎⁎ 0.943⁎⁎ 0.917⁎⁎ 0.886⁎ 0.497 0.908⁎ 0.910⁎

CF 0.281⁎⁎⁎ 0.025 0.062 0.053 −0.035 −0.553 −0.050 −0.029Fat 0.375 0.236 0.257 0.204 0.098 0.045 0.178 0.166

Gulf of FinlandLength 0.643 0.877⁎ 0.855⁎ 0.486 0.763 0.520 0.704 0.946⁎⁎

Weight 0.610 0.853⁎ 0.829⁎ 0.485 0.736 0.497 0.673 0.902⁎

CF −0.293 0.099 0.028 0.877⁎ −0.079 −0.345 −0.218 0.317Fat −0.584 −0.179 −0.260 0.874⁎ −0.388 −0.674 −0.535 −0.025

⁎ pb0.05.⁎⁎ pb0.01.⁎⁎⁎ pb0.001.


Many of the PCDD/F congeners with low proportions in sprat andherring were not detectable in the congener profile of salmon(Fig. 10). The proportion of 2,3,7,8-TCDD in the PCDD/F congener pro-file of salmon was below 4%.

4. Discussion

4.1. Trends in POPs

Svendsen et al. (2008) reported somewhat lower values of PCBsin salmon from the BPr, i.e., 58 ng g−1 in fish caught in April 2004 and83 ng g−1 in February 2006, than those measured in the presentstudy, most likely because their samples were collected further southin the Baltic. On the basis of the congener profiles of the PCDD/Fs inthe present study, salmon feed on younger prey fish in the southernBaltic Sea than, for instance, in the BS. The age-dependent increase inthe concentrations of organohalogens in prey fish of salmon appear tobe more pronounced in the BS than in more southern areas, simplydue to the slower growth rate (Mikkonen et al., 2011; Vuorinen et al.,

Table 4Pearson correlation coefficients in sprat of all age groups between toxicants (in fresh weightstudy areas.

ΣPCDD ΣPCDF ΣPCDD/F ΣCoPCB

Baltic ProperLength −0.064 0.724 0.652 0.621⁎⁎⁎

Weight −0.089 0.750 0.671 0.656CF −0.860⁎ 0.021 −0.150 0.034Fat −0.766 −0.124 −0.265 −0.042

Bothnian SeaLength 0.946⁎ 0.891⁎ 0.913⁎ 0.888⁎

Weight 0.837 0.779 0.801 0.813CF −0.940⁎ −0.977⁎⁎ −0.981⁎⁎ −0.909⁎

Fat −0.968⁎⁎ −0.905⁎ −0.929⁎ −0.789

Gulf of FinlandLength 0.808 0.716 0.736 0.582Weight 0.825⁎ 0.791 0.803 0.682CF −0.356 −0.133 −0.167 −0.020Fat −0.116 0.191 0.148 0.318

⁎ pb0.05.⁎⁎ pb0.01.⁎⁎⁎ pb0.001.

2004). In addition, salmon in the present study were sampled inautumn. In sprat and herring, the toxicant concentrations tended to behigher in autumn than in spring, evidently due to the higher fat content,although this increase was only significant for CoPCBs.

In the present study, the concentrations of PCDDs and PCDFs insalmon from the BPr were, on average, similar to those reported byShelepchikov et al. (2008) in salmon caught from the eastern BPr(ICES SD 26), although they did not report the sampling time, fishsize, or age. In 2–6-year-old sprat caught from the western GoF in2002–2004, the concentrations of PCDDs and PCDFs were approxi-mately 1–2 and 4–13 pg g−1, respectively (Simm et al., 2006), i.e.,quite similar to those in sprat in the present study.

When the concentrations of POPs in sprat and herring in the northernBPr and the GoF are compared with those in sprat and herring collectedin late autumn–earlywinter in thewesternGoF in 1994–1995 (Vuorinenet al., 2002), the concentration of PCBs in sprat decreased by approxi-mately 62%, from 77 ng g−1 to 28–31 ng g−1, and by approximately37% in herring, from 31 ng g−1 to 17–22 ng g−1. The same trend wasobserved for PCDDs and PCDFs during this nearly ten-year period,

) and the mean weight, length, condition factor (CF) and fat content of fish in the three

ΣPCB ΣPBDE WHO-TEQPCDD/F WHO-TEQPCDD/F+PCB

0.861⁎ 0.893⁎ 0.857⁎ 0.871⁎

0.851⁎ 0.880⁎ 0.863⁎ 0.876⁎

−0.347 −0.318 −0.158 −0.177−0.555 −0.555 −0.367 −0.377

0.957⁎ 0.894⁎ 0.882⁎ 0.939⁎

0.901⁎ 0.789 0.758 0.844−0.893⁎ −0.960⁎ −0.987⁎⁎ −0.959⁎⁎

−0.903⁎ −0.941⁎ −0.944⁎ −0.944⁎

0.872⁎ 0.853⁎ 0.842⁎ 0.839⁎

0.875⁎ 0.846⁎ 0.873⁎ 0.874⁎

−0.422 −0.473 −0.304 −0.294−0.147 −0.177 −0.031 −0.012

Table 5Biomagnification factors (BMF) of organohalogens in 2nd-sea-year salmon from its preyspecies (sprat of all age groups, 1- to 3-year-old herring, and three-spined stickleback)calculated from the concentrations expressed on a fresh weight and fat content basis.

Fresh weight-based Fat-based

Toxicant Sprat Herring Stickleback Sprat Herring Stickleback

Baltic ProperΣPCB 4.75 6.13 4.77 3.11 2.07 2.72ΣCoPCB 3.79 6.67 3.60 2.27 2.24 2.04ΣPCDD 1.44 2.24 1.23 0.98 0.75 0.70ΣPCDF 2.41 3.19 1.86 1.47 1.07 1.06ΣPBDE 4.80 5.72 5.85 3.09 1.98 3.33

Bothnian SeaΣPCB 6.50 8.71 7.20 5.61 3.79 7.39ΣCoPCB 2.42 3.85 2.35 1.94 1.66 2.42ΣPCDD 3.91 3.04 2.45 3.43 1.25 2.63ΣPCDF 3.60 3.73 3.43 2.95 1.58 3.67ΣPBDE 4.81 5.53 6.63 4.12 2.41 6.89

Gulf of FinlandΣPCB 8.23 9.60 12.4 4.89 3.38 9.21ΣCoPCB 2.71 5.56 3.52 1.36 1.72 2.53ΣPCDD 2.20 2.89 5.64 1.24 1.01 4.37ΣPCDF 3.43 4.25 7.82 1.82 1.45 5.60ΣPBDE 3.80 5.92 8.91 2.16 2.03 5.56

Fig. 9. The proportions of different PCDD/F congeners (from the total concentration) insprat of different ages caught from the three sea areas.


corresponding to an annual decrease of 5–6%. However, the concentra-tion of CoPCBs was only somewhat lower in the present study than inthe earlier study, with an annual decrease of 2–5%.

In salmon in the GoF, the relationship between age and toxicant con-centrations (on a fresh weight basis) appeared to be clear for all POPs,and the age-dependence of the concentration increase was greater insalmon than in the prey species, excluding, however, PCDDs in herring.Probably due to their high fat content, the difference in age-dependency

Fig. 8. The mean (±SE) proportions of different PCDD/F congeners (from the totalconcentration) in salmon of different sea years caught from the three sea areas.

in sprat relative to salmon appeared least evident in the concentrationof CoPCBs. In a previous study by Vuorinen et al. (2002), the concentra-tions of PCDDs and PCBs also increased significantly with age in spratand herring caught in late 1994 and early 1995 in the western GoF.However, the age relationship of the total concentrations of CoPCBsand PCDFs was only apparent after fat normalisation. Consistent withthis finding, the age-related concentration increase on a fresh weightbasis in the present study was, in general, least pronounced for CoPCBsin sprat, herring, and salmon.

4.2. Spatial differences in biomagnification

Generally, in the 2nd-sea-year salmon, the concentrations of all ofthe OCs and TEQs were or tended to be highest in the GoF and lowestin the BPr, with the exception of the CoPCB concentration, which washighest in the BPr and lowest in the BS. The concentrations of PBDEswere similar in all areas.

Among the body parameters, only the fat content of the 2nd-sea-year salmon differed between the areas and was lowest in the GoF.In the GoF, the negative correlation (non-significant) between thefat content or CF and the concentration of POPs, excluding CoPCBsand PBDEs of 2nd-sea-year salmon, reflects the spatial dietary qualityof prey: from slow-growing herring and sprat (Peltonen et al., 2004),the biomagnification of most POPs to salmon is higher than the accu-mulation of fat. The lower percentage of fat in these salmon in com-parison to the other two areas reflects the lower fat content of theirprey fish, as dietary lipid levels directly affect the muscle lipid con-centration in Atlantic salmon (Hemre and Sandnes, 1999). The con-centrations of PCBs and PCDDs as well as those of WHO-TEQPCDD/F+PCB

were higher in salmon from the GoF than in the other two areas,most likely due to their feeding in the GoF on lean, slow-growing preyfish that contained higher concentrations of these OCs.

image of Fig.�8

image of Fig.�9

Fig. 10. The proportions of different PCDD/F congeners (from the total concentration)in herring of different ages caught from the three sea areas.


The significant positive association between the CF and the concen-tration of all of the POPs, except CoPCBs, in salmon from the BS indicatesthat a good nutritional state in the BS is related to the increase in orga-nohalogen concentrations. In the BS, where salmon mainly prey onherring, the prey fish had a clearly higher CF than in the GoF. Theconcentration of PCDFs in salmon was as high in the BS as in the GoF,because the age-dependent increase in the concentrations of organoha-logens in herring in the BS was, on average, more pronounced than inthe other sea areas.

In the BPr, the organohalogen concentrations, excluding PCBs,increased as a function of the increase in the fat content of salmon,and the clearest correlations with fat appeared in the concentrationsof PBDEs and CoPCBs. Feeding on sprat, the fat content of which ishigher than in herring and which is the main prey fish of salmonin the southern BPr (Hansson et al., 2001), seems to especially lead

to accumulation of CoPCBs. This is reflected in salmon in the ratioof CoPCBs to PCDD/Fs (in fresh weight), which was, on average, 6in the BS, 9 in the GoF, and 13 in the BPr, regardless of the high fatcontent of salmon from the BS. However, the concentrations ofCoPCBs and PBDEs could be higher in the southern parts of the BalticSea due to the vicinity of more industrialised and densely populatedregions. The proportion of dl-PCBs in the TEQPCDD/F+PCB of salmonwas consistently higher in the Baltic Proper than in the other areas,which is in accordance with the findings from herring (SCALE, 2004).Due to their high biomagnification rate also in the BPr, PBDEs werethe only toxicants that did not exhibit any spatial relationship.

Some distortion in the BMFs in the BPr was probably caused by thefact that the salmon were from a more southern area in the BPr thanthe prey fish. However, fish in the Baltic undertake micro- and macro-scale movements and migrations (Aro, 1989). Herring in the BS arethe most localised and in the GoF hardly migrate from the area. Con-versely, sprat migrate during the feeding period in the BPr within SDs29, 28, and 26, and salmon move together with their prey. Accordingto Aro (1989), sprat and herring are more stationary during winter,i.e., at times when the prey fish samples of the present study werecollected. Despite this, fish sampled in the BPr cannot be consideredas quite local, and moreover, toxicant concentrations reflect long-term exposure.

4.2.1. Biomagnification of organochlorinesThe most conspicuous spatial difference in the organohalogen

content of 2nd-sea-year salmon was that the concentrations ofPCDDs and PCBs were highest in salmon preying on lean herringand sprat from the GoF and lowest in salmon from the BPr feedingmainly on sprat. In the BPr, which is situated more to the souththan the other two areas, the growth rate of fish is evidently higherdue to more beneficial environmental conditions, which enablemore frequent feed intake throughout the year (Karlsson et al.,1999). Thus, the increase in the PCB concentration seems to be relatedto the slow growth and leanness of the prey fish. This is further sup-ported by the fact that the BMF of PCBs was highest in the GoF (withlow fat content prey fish) and, moreover, that the concentration ofPCBs among all the analysed organohalogens in the BPr had no signif-icant positive correlation with the fat content of salmon.

Among the whole prey species data, only the concentration ofPCDDs was related to area and was highest in the GoF. The KymijokiRiver, the sediments of which are highly polluted by PCDD/Fs accordingto Salo et al. (2008), contributes significantly to the loading of thesecompounds into the GoF, with a typical pattern of octa-, hepta-, andhexa-chlorinated dibenzofurans dominating over PCDDs. However,these highly chlorinated dibenzofurans are only marginally absorbedinto fish (Isosaari et al., 2004; Oppenhuizen and Sijm, 1990). Conse-quently, the proportions of these congeners were practically zero insalmon, although their proportions were slightly higher in sprat andherring in the GoF than in the two other areas. However, the propor-tions of the other congeners in salmon remained low because of theextensive accumulation of the 2,3,7,8-TCDF and 2,3,4,7,8-PeCDF conge-ners. As a consequence, the proportion of 2,3,7,8-TCDD, the most toxicof the PCDD/F congeners (Van den Berg et al., 1998; 2006), in salmonwas only slightly higher than in its prey species in spite of its relativelyhigh BMF. The high increase in the concentration of PCDFs withherring age is apparently the main reason for the high concentrationof PCDFs in salmon in the BS, where herring is the main prey fish(Hansson et al., 2001). Moreover, due to the slow growth rate of herringin the northern Baltic Sea, the average age of herring preyed upon bysalmon in the BS appeared to be higher than in the BPr (Mikkonenet al., 2011).

In this study, the two principal PCDD/Fs congeners in salmon aswell as in its prey fish were 2,3,7,8-TCDF and 2,3,4,7,8-PeCDF, aswas detected previously (Isosaari et al., 2006; Simm et al., 2006;Vuorinen et al., 2002). However, on a TEQ basis, the latter becomes

image of Fig.�10


more important due to the higher TEF value (Van den Berg et al.,1998, 2006). The age-related increase in the ratio of these particularcongeners appeared to be similar in sprat and herring in the BPr,but the increase was more pronounced in herring in the GoF, whichis evidently related to the slow growth rate due to the shortage of nu-trition in that sea area (Peltonen et al., 2004). 2,3,4,7,8-PeCDF is read-ily enriched in fish with age because it is apparently hardly detoxifiedin fish. In sediments of the GoF and the Kymijoki River, which is high-ly polluted by dioxins, only small amounts of 2,3,4,7,8-PeCDF weredetected (Salo et al., 2008). The 2,3,4,7,8-PeCDF to 2,3,7,8-TCDFratio in 2nd-sea-year salmon was lower in the BPr than in the BSand GoF, indicating that, as recently reported by Mikkonen et al.(2011), salmon in the BPr feed on younger, faster growing fish. Bycontrast, the penta-congener to tetra-congener ratio in salmon wastwice as high as in the prey fish in the BS, indicating that salmon inthe BS most likely also feed on herring older than 1–3 years.

The BMF values of CoPCBs from different salmon prey species tosalmon were, like the BMFs of PCDD/Fs, also rather low, which maybe related to their similarly long biological half-life in the prey spe-cies. The concentration of CoPCBs in salmon, unlike PCDDs andPCBs, tended to be higher in the BPr than in the GoF and was lowestin the BS, where the predominant prey species of salmon is herring,in contrast to the BPr (Hansson et al., 2001; Karlsson et al., 1999).This is reasonable because the high concentration of CoPCBs wasshown to be related to the prey species (fatty sprat). In an earlierstudy, the concentrations of CoPCBs were similarly higher in spratthan in herring caught from the western GoF (Vuorinen et al.,2002). In that study, the high CoPCB content in sprat was connectedwith both their higher fat content and the slower growth of olderage groups of sprat in comparison to 1–3-year-old herring. Therefore,no significant age-dependent relationship was detected in the con-centrations of CoPCBs and PCDFs in female sprat without fat-normalisation. The age-related increase in the concentrations ofCoPCBs and PCDFs in sprat in the present study was likewise not sig-nificant in the BPr or GoF but was significant in the BS, where theoverall age-dependent increase in the concentrations of organohalo-gens in herring was, on average, more pronounced than in the otherareas. In the present study, the fat content was higher in sprat thanin herring, similar to the earlier study, but was also higher in salmonprey fish species in the BS than in the GoF. According to Hemre andSandnes (1999), a high lipid concentration in feed increases the fatcontent of salmon, which in the present study was apparent in thehigher fat content in salmon from the BS compared to the GoF,where salmon prey fish were also lean.

4.2.2. Biomagnification of organobrominesIn contrast to the PCDD/Fs and PCBs, the concentration of PBDEs

on a fresh weight basis tended to be lowest in salmon in the GoF.The mean PBDE concentrations in salmon in the present study(1.6–6.2 ng g−1) were among the highest values that have been ob-served when compared to the world-wide sampling and analysis forPBDEs in different salmon species, in which PBDEs varied on a freshweight basis from approximately 0.1 to 4.2 ng g−1 (Hites et al.,2004). The sum of seven PBDE congeners, 2.5 ng g−1, reported bySzlinder-Richert et al. (2010) for salmon from the southern Baltic(SDs 25 and 26) in 2004–2006 was clearly exceeded in the 2nd-sea-year salmon of the present study at 3.0 for the BPr, 2.8 for the GoF,and 3.1 ng g−1 for the BS. However, based on the reported size data,the salmon analysed by Szlinder-Richert et al. (2010) were apparent-ly younger on average than those in the present study.

The PBDE concentration was dependent on the prey species andwas higher in sprat than in herring; PBDEs seemed to bioaccumulateas readily in fat as did the CoPCBs. Consistently, PBDE congeners accu-mulated as a function of the lipid content of Atlantic salmon in a feed-ing experiment (Isosaari et al., 2005). It is therefore reasonable thatthe concentrations of PBDEs tended to be lowest and varied most in

low-fat salmon in the GoF, where the PBDE levels in prey fish weresimilar to the levels in the other two areas.

The biomagnification rate of PBDEs equalled that of PCB in the BPr,but the pattern of BMFs for PBDEs did not vary much between thestudy basins, in contrast to PCBs. The BMF of PCBs was much higherthan that of PBDEs in the GoF and was also higher in the BS. The con-geners of these two compound groups also had the highest accumu-lation efficiencies in the feeding experiments with Atlantic salmon(Isosaari et al., 2004, 2005).

4.3. Suitability of Baltic fish for feed

Off the German coast in the Baltic Sea, both the WHO-TEQPCDD/F

and WHO-TEQPCDD/F+PCB concentrations in fillets of herring with atotal length of 24–29 cm caught in 2006 were considerably lower, ap-proximately 2.3 pg g−1 and 4.4 pg g−1 (Karl et al., 2010), than inwhole 1–10-year-old herring in the northern BPr in the presentstudy (on average 4.0 pg g−1 for WHO-TEQPCDD/F and 6.4 pg g−1 forWHO-TEQPCDD/F+PCB). In addition to age, the growth rate affects theaccumulation of POPs; in contrast to the northern Baltic Sea, herringin the southern areas grow much faster, as evident in the meanweight-at-age (ICES, 2009). Therefore, herring caught by trawl, forexample, are on average older in the Gulf of Bothnia than thosecaught further south in the Baltic Sea.

In sprat with all age groups included, the concentration ofWHO-TEQPCDD/F was lower (mean 1.9 pg g−1) than in 1994–1995(5 pg g−1) (Vuorinen et al., 2002), although that study includedolder age groups than those in the present study. Karl et al. (2010)reported higher concentrations of WHO-TEQPCDD/F (1.6–3.0 pg g−1)and WHO-TEQPCDD/F+PCB (3.1–6.3 pg g−1) in sprat (fillets) caughtin the Western Baltic than those recorded for the northern BPr inthe present study (means, 1.9 pg g−1 and 3.9 pg g−1, respectively);those authors did not, however, report the ages or sizes of all thesprat.

The maximum allowable TEQ concentrations of PCDD/Fs and ofPCDD/Fs plus dl-PCBs in fish feed set by the EC (EC, 2006a) and cor-rected for the water content of fresh fish would already be exceededif the youngest age groups of sprat and herring were used directly asfeed. The regulation ensures that dioxin concentrations in fishintended for animal feed in aquaculture do not attain excessivelyhigh levels. Thus, herring and sprat, particularly those caught fromthe northern areas of the Baltic Sea, would not be permissible asfood in aquaculture without refinement.

5. Conclusions

All of the organohalogens accumulated with age in sprat and her-ring and especially strongly in salmon. Thus, the age in sea yearsshould always be given in reporting organohalogen concentrationsin salmon. Among the organohalogens, the bioaccumulation ofCoPCBs was most distinctly affected by the fat content of the preyand predator fishes, while the accumulation of PCDD/Fs was mostclearly affected by the age of the fish. The accumulation of PBDEsappeared to be highly dependent on both the age and fat content.The slow growth rate of prey fish, due to northern feeding groundsor high food competition, increases the biomagnification of all orga-nohalogens in salmon. This was especially evident in the concentra-tions of PCBs and PCDDs, which were lowest in the 2nd-sea-yearsalmon from the BPr and highest in the GoF. If the cause of the slowgrowth is overcrowding of sprat and herring, as has apparentlybeen the case in the GoF when the sprat stock has been dense, thebiomagnification of POPs could be reduced by increasing the intensityof fishing while ensuring sufficient recruitment. This would lower theorganohalogen load in Baltic salmon and likewise in human Baltic fishconsumers. The increase in the ratio of 2,3,4,7,8-PeCF to 2,3,7,8-TeCFwith prey fish age provides insight into the feeding behaviour of


salmon. Differences in that ratio indicated that salmon mainly feedon younger prey in the southern BPr than in the north. In thesouthern Baltic, fish growth is evidently faster due to favourableenvironmental conditions (e.g., higher salinity and warmer waterduring the cold period), and the biomagnification of POPs, with theexception of CoPCBs, is consequently lower there than in the northernbasins. Conversely, the bioaccumulation of CoPCBs in salmon ishighest in the southern regions (in the BPr) because the salmonfeed on fatty sprat in the vicinity of high CoPCB emissions. As the EClimit values for both dioxins and dioxins plus dl-PCBs were exceededin all age groups of sprat and herring, these prey fish should not beused as feed in aquaculture without refinement to reduce the POPlevels.

Supplementary materials related to this article can be foundonline at doi:10.1016/j.scitotenv.2012.02.002.

Acknowledgements

The personnel of the Finnish Game and Fisheries Research Institutecollected the samples of prey species (Folke Halling, Hannu Harjunpää,Timo Jääskeläinen, Petteri Karttunen, Heikki Savolainen and PenttiVirtanen), aged them (Folke Halling, Jari Raitaniemi, PhD, and TarjaWiik) and also organised salmon samples (Folke Halling and HannuHarjunpää) and aged the salmon (Irmeli Torvi). This research wasfunded by the Academy of Finland (project DIOXMODE, no.102557, inthe Baltic Sea Research Programme BIREME) and the Nordic Councilof Ministers.

References

Aro E. A review of fish migration patterns in the Baltic. Rapp Pv Réun Cons int ExplorMer 1989;190:72–96.

Bignert A, Danielsson S, Strandmark A, Nyberg E, Asplund L, Eriksson U, et al. Com-ments concerning the national Swedish contaminant monitoring programme inmarine biota. (31.3.2008, 142 pp.)Swedish Museum of Natural History. Sakrapport,Metaller och organiska miljögifter i marin biota, trend- och områdesövervakning,Överenskommelse 212 0714, dnr 721-4235-07Mm; 2008.

Burreau S, Zebuhr Y, Broman D, Ishaq R. Biomagnification of PBDEs and PCBs in foodwebs from the Baltic Sea and the northern Atlantic Ocean. Sci Total Environ2006;366:659–72.

Casini M, Cardinale M, Arrhenius F. Feeding preferences of herring (Clupea harengus)and sprat (Sprattus sprattus) in the southern Baltic Sea. ICES J Mar Sci 2004;61:1267–77.

de Roode DF, Gustavsson MB, Rantalainen AL, Klomp AV, Koeman JH, Bosveld ATC.Embryotoxic potential of persistent organic pollutants extracted from tissues ofguillemots (Uria aalge) from the Baltic Sea and the Atlantic Ocean. Environ ToxicolChem 2002;21:2401–11.

EC. Commission Directive 2006/13/EC of 3 February 2006 amending Annexes I and II toDirective 2002/32/EC of the European Parliament and of the Council on undesirablesubstances in animal feed as regards dioxins and dioxin-like PCBs. OJL 2006a; 49(L 032): 44–53.

EC. Commission regulation (EC) No 1881/2006 of 19 December 2006 setting maximumlevels for certain contaminants in foodstuffs. OJL 2006b; 49 (L 364): 5–24.

Falandysz J, Kannan K, Tanabe S, Tatsukawa R. Organochlorine pesticides and poly-chlorinated biphenyls in cod-liver oils — North Atlantic, Norwegian Sea, NorthSea and Baltic Sea. Ambio 1994;23:288–93.

Hansson S, Karlsson L, Ikonen E, Christensen O, Mitans A, Uzars D, et al. Stomach ana-lyses of Baltic salmon from 1959–1962 and 1994–1997: possible relations betweendiet and yolk-sac-fry mortality (M74). J Fish Biol 2001;58:1730–45.

Hemre GI, Sandnes K. Effect of dietary lipid level on muscle composition in Atlanticsalmon Salmo salar. Aquacult Nutr 1999;5:9-16.

Hites RA, Foran JA, Schwager SJ, Knuth BA, Hamilton MC, Carpenter DO. Global assess-ment of polybrominated diphenyl ethers in farmed and wild salmon. Environ SciTechnol 2004;38:4945–9.

ICES. Report of the Baltic Fisheries AssessmentWorking Group (WGBFAS). ICES DocumentCM 2007/ACFM:15; 2007. (735 pp.).

ICES. Report of the Baltic Fisheries AssessmentWorking Group (WGBFAS). ICES DocumentCM 2009/ACOM:07; 2009. (635 pp.).

Isosaari P, Kiviranta H, Lie O, Lundebye AK, Ritchie G, Vartiainen T. Accumulation anddistribution of polychlorinated dibenzo-p-dioxin, dibenzofuran, and polychlori-nated biphenyl congeners in atlantic salmon (Salmo salar). Environ Toxicol Chem2004;23:1672–9.

Isosaari P, Lundebye AK, Ritchie G, Lie O, Kiviranta H, Vartiainen T. Dietary accumu-lation efficiencies and biotransformation of polybrominated diphenyl ethers infarmed Atlantic salmon (Salmo salar). Food Addit Contam 2005;22:829–37.

Isosaari P, Hallikainen A, Kiviranta H, Vuorinen PJ, Parmanne R, Koistinen J, et al.Polychlorinated dibenzo-p-dioxins, dibenzofurans, biphenyls, naphthalenes and

polybrominated diphenyl ethers in the edible fish caught from the Baltic Sea andlakes in Finland. Environ Pollut 2006;141:213–25.

Järnberg U, Asplund L, deWit C, Grafström A-K, Haglund P, Jansson B, et al. Polychlorinatedbiphenyls and polychlorinated naphthalenes in Swedish sediment and biota: levels,patterns, and time trends. Environ Sci Technol 1993;27:1364–74.

Jorundsdottir H, Norstrom K, Olsson M, Pham-Tuan H, Huhnerfuss H, Bignert A, et al.Temporal trend of bis(4-chlorophenyl) sulfone, methylsulfonyl-DDE and -PCBs inBaltic guillemot (Uria aalge) egg 1971–2001 — a comparison to 4,4′-DDE and PCBtrends. Environ Pollut 2006;141:226–37.

Kannan K, Falandysz J, Yamashita N, Tanabe S, Tatsukawa R. Temporal trends of organo-chlorine concentrations in cod-liver oil from the southern Baltic Proper, 1971–1989.Mar Pollut Bull 1992;24:358–63.

Karl H, Ruoff U. Dioxins, dioxin-like PCBs and chloroorganic contaminants in herring,Clupea harengus, from different fishing grounds of the Baltic Sea. Chemosphere2007;67:S90–5.

Karl H, Bladt A, Rottler H, Ludwigs R, Mathar W. Temporal trends of PCDD, PCDF and PCBlevels in muscle meat of herring from different fishing grounds of the Baltic Sea andactual data of different fish species from the Western Baltic Sea. Chemosphere2010;78:106–12.

Karlsson L, Ikonen E, Mitans A, Hansson S. The diet of salmon (Salmo salar) in the Balticsea and connections with the M74 syndrome. Ambio 1999;28:37–42.

Keinänen M, Tolonen T, Ikonen E, Parmanne R, Tigerstedt C, Rytilahti J, et al. Reproduc-tion disorder of Baltic salmon — M74. Riista- ja kalatalouden tutkimuslaitos.Kalatutkimuksia – Fiskundersökningar 2000;165. (38 pp. In Finnish with Englishabstract).

Kiviranta H, Vartiainen T, Parmanne R, Hallikainen A, Koistinen J. PCDD/Fs and PCBs inBaltic herring during the 1990s. Chemosphere 2003;50:1201–16.

Kiviranta H, Ovaskainen ML, Vartiainen T. Market basket study on dietary intake ofPCDD/Fs, PCBs, and PBDEs in Finland. Environ Int 2004;30:923–32.

Larsson P, Backe C, Bremle G, Eklov A, Okla L. Persistent pollutants in a salmon popula-tion (Salmo salar) of the southern Baltic Sea. Can J Fish Aquat Sci 1996;53:62–9.

Lizak R, Piskorska-Pliszczynska J, Kowalski B, Rachubik J, Warenik M, Wijaszka T. PCDD,PCDF, and dl-PCB in baltic herring, salmon and sprat from Western Gotland basinof the Baltic Sea. Bull Vet Inst Pulawy 2007;51:661–6.

Mikkonen J, Keinänen M, Casini M, Pönni J, Vuorinen PJ. Relationships between fishstock changes in the Baltic Sea and the reproductive syndrome, M74 of the Atlanticsalmon (Salmo salar L.). ICES J Mar Sci 2011;68:2134–44.

Möllmann C, Kornilovs G, Fetter M, Köster FW. Feeding ecology of central Baltic Seaherring and sprat. J Fish Biol 2004;65:1563–81.

Oppenhuizen A, Sijm DTHM. Bioaccumulation and biotransformation of polychlorinateddibenzo-p-dioxins and dibenzofurans in fish. Environ Toxicol Chem 1990;9:175–86.

Parmanne R, Hallikainen A, Isosaari P, Kiviranta H, Koistinen J, Laine O, et al. The depen-dence of organohalogen compound concentrations on herring age and size in theBothnian Sea, northern Baltic. Mar Pollut Bull 2006;52:149–61.

Peltonen H, Vinni M, Lappalainen A, Pönni J. Spatial feeding patterns of herring(Clupea harengus L.), sprat (Sprattus sprattus L.), and the three-spined stickleback(Gasterosteus aculeatus L.) in the Gulf of Finland, Baltic Sea. ICES J Mar Sci2004;61:966–71.

Perttilä M, Tervo V, Parmanne R. Age dependence of the concentration of harmful sub-stances in Baltic herring (Clupea harengus). Chemosphere 1982;11:1019–26.

Pikkarainen AL, Parmanne R. Polychlorinated biphenyls and organochlorine pesticidesin Baltic herring 1985–2002. Mar Pollut Bull 2006;52:1304–9.

Richter H, McDermott JG. The staining of fish otoliths for age determination. J Fish Biol1990;36:773–9.

Salminen M, Kuikka S, Erkamo E. Divergence in feeding migration of Baltic Salmon(Salmo salar L.); the significance of Smolt Size. Nord J Freshwat Res 1994;69:32–42.

Salo S, Verta M, Malve O, Korhonen M, Lehtoranta J, Kiviranta H, et al. Contamination ofRiver kymijoki sediments with polychlorinated dibenzo-p-dioxins, dibenzofuransand mercury and their transport to the Gulf of Finland in the Baltic Sea. Chemo-sphere 2008;73:1675–83.

SAS Institute Inc.. SAS/STAT Guide for Personal Computers, Release 9.1, WindowsVersion 5.1.2600. Cary, NC: SAS Institute Inc.; 2003.

SCALE. Baseline report on Integrated monitoring of dioxins and PCBs in the Balticregion, technical working group on integrated monitoring, 9th Jan 2004.COM(2003)338 final. The framework of European environment and health strate-gy; 2004.

Shelepchikov AA, Shenderyuk VV, Brodsky ES, Feshin D, Baholdina LP, Gorogankin SK.Contamination of Russian Baltic fish by polychlorinated dibenzo-p-dioxins, di-benzofurans and dioxin-like biphenyls. Environ Toxicol Pharmacol 2008;25:136–43.

Simm M, Roots O, Kotta J, Lankov A, Henkelmann B, Shen H, et al. PCDD/Fs in sprat(Sprattus sprattus balticus) from the Gulf of Finland, the Baltic Sea. Chemosphere2006;65:1570–5.

Svendsen TC, Vorkamp K, Ronsholdt B, Frier JO. Retrospective determination of primaryfeeding areas of Atlantic salmon (Salmo salar) using fingerprinting of chlorinatedorganic contaminants. ICES J Mar Sci 2008;65:921–9.

Szlinder-Richert J, Barska I, Mazerski J, Usydus Z. Organochlorine pesticides in fish fromthe southern Baltic Sea: Levels, bioaccumulation features and temporal trendsduring the 1995–2006 period. Mar Pollut Bull 2008;56:927–40.

Szlinder-Richert J, Barska I, Usydus Z, Grabic R. Polybrominated diphenyl ethers(PBDEs) in selected fish species from the southern Baltic Sea. Chemosphere2010;78:695–700.

Van den Berg M, Birnbaum L, Bosveld ATC, Brunström B, Cook P, Feeley M, et al. Toxicequivalency factors (TEFs) fos PCBs, PCDDs, PCDFs for humans and wildlife. EnvironHealth Perspect 1998;106:775–92.

doi:10.1016/j.scitotenv.2012.02.002


Van den Berg M, Birnbaum LS, Denison M, De Vito M, Farland W, Feeley M, et al. The2005 World Health Organization reevaluation of human and mammalian toxicequivalency factors for dioxins and dioxin-like compounds. Toxicol Sci 2006;93:223–41.

Vuorinen PJ, Paasivirta J, Keinänen M, Koistinen J, Rantio T, Hyötyläinen T, et al. TheM74 syndrome of Baltic salmon (Salmo salar) and organochlorine concentrationsin the muscle of female salmon. Chemosphere 1997;34:1151–66.

Vuorinen PJ, Parmanne R, Vartiainen T, Keinänen M, Kiviranta H, Kotovuori O, et al.PCDD, PCDF, PCB and thiamine in Baltic herring (Clupea harengus L.) and sprat[Sprattus sprattus (L.)] as a background to the M74 syndrome of Baltic salmon(Salmo salar L.). ICES J Mar Sci 2002;59:480–96.

Vuorinen PJ, Parmanne R, Kiviranta H, Isosaari P, Hallikainen A, Vartiainen T. Differ-ences in PCDD/F concentrations and patterns in herring (Clupea harengus) fromthe southern and northern Baltic Sea. Organohalogen Compd 2004;66:1858–63.

Species Area Sea years/Age N1 N2

Salmon Baltic Proper 2 3638 ± 242 711 ± 15 1.00 ± 0.020 9 16.1 ± 0.7 9

Salmon Baltic Proper 3 8317 ± 903 ± 1.13 ± 1 20.6 ± 1

Salmon Bothnian Sea 2 3718 ± 307 709 ± 13 1.03 ± 0.030 5 16.7 ± 0.3 5

Salmon Gulf of Finland 1 1302 ± 79 489 ± 9 1.11 ± 0.010 5 7.9 ± 1.1 5

Salmon Gulf of Finland 2 3985 ± 769 714 ± 49 1.07 ± 0.040 4 12.6 ± 1.3 4

Salmon Gulf of Finland 3 4098 ± 711 ± 1.14 ± 1 13.1 ± 1

Herring Baltic proper 1–3 16.2 ± 0.3 137 ± 1 0.59 ± 0.003 479 5.5 ± 0.4 6

Herring Bothnian Sea 1–3 14.7 ± 0.4 132 ± 1 0.59 ± 0.003 342 7.3 ± 0.6 6

Herring Gulf of Finland 1–3 9.5 ± 0.3 119 ± 1 0.50 ± 0.004 348 4.1 ± 0.4 6

Sprat Baltic proper 1–7 7.5 ± 0.2 109 ± 1 0.56 ± 0.005 202 9.5 ± 1.6 6

Sprat Bothnian Sea 1–7 9.7 ± 0.2 117 ± 1 0.60 ± 0.007 137 12.5 ± 2.1 5

Sprat Gulf of Finland 1–7 7.4 ± 0.2 111 ± 1 0.52 ± 0.005 206 6.5 ± 1.5 6

Stickleback Baltic proper - - - - - 9.1 1

Stickleback Bothnian Sea - - - - - 16.0 ± 4.1 2

Stickleback Gulf of Finland - - - - - 7.4 ± 1.9 4

Fat, %Weight, g Length, mm CF

[Vuorinen et al., Biomagnification of organohalogens in Atlantic salmon (Salmo salar) from its main prey species in three areas of the Baltic Sea]

Supplement Table 1. Mean (± SE) fresh weight, total length, condition factor (CF), and fat content of salmon and its prey species, sprat, herring, and

three-spined-stickleback, in three sea areas. The sea years of salmon, age of herring and sprat, number of fish in homogenates (N1), and number of

homogenates (N2) are also given.

Fish in homogenates Homogenate

Mean SE Min Max N Mean SE Min Max N Mean SE Min Max N

Sprat PCDD, pg/g 1.07 0.16 0.41 1.52 6 1.26 0.22 0.71 2.22 6 0.65 0.18 0.21 1.23 5

1- to 7-year-old PCDF, pg/g 5.07 0.74 3.24 7.58 6 5.96 1.29 3.19 11.40 6 4.92 0.68 2.62 6.62 5

PCDD/F, pg/g 6.14 0.81 3.66 8.79 6 7.22 1.49 3.90 13.62 6 5.56 0.84 2.83 7.85 5

CoPCB, pg/g 59.16 9.45 39.85 91.45 6 75.89 17.31 42.79 147.35 6 46.73 5.27 27.45 56.80 5

PCB, ng/g 27.84 6.34 12.38 49.67 6 30.61 6.59 12.65 57.17 6 26.58 6.22 9.89 41.43 5

IndPCB6, ng/g 15.75 3.80 6.81 29.02 6 16.58 3.67 6.77 30.96 6 15.81 3.83 5.68 24.82 5

PBDE, ng/g 0.82 0.17 0.39 1.37 6 0.87 0.15 0.43 1.48 6 0.83 0.17 0.32 1.28 5

PCDD/F-TEQ98, pg/g 1.80 0.29 1.00 2.60 6 2.16 0.47 0.99 4.12 6 1.75 0.31 0.78 2.60 5

TOT-TEQ98, pg/g 3.75 0.72 1.84 5.99 6 4.43 0.97 2.04 8.46 6 3.49 0.66 1.55 5.28 5

PCDD/FTEQ06, pg/g 1.34 0.21 0.76 1.90 6 1.62 0.35 0.76 3.11 6 1.28 0.21 0.58 1.86 5

TOT-TEQ06, pg/g 2.90 0.55 1.43 4.62 6 3.41 0.76 1.64 6.59 6 2.72 0.49 1.25 4.02 5

Herring PCDD, pg/g 0.63 0.09 0.36 0.89 6 0.91 0.11 0.53 1.34 6 0.72 0.15 0.30 1.39 6

1- to 3-year-old PCDF, pg/g 3.83 0.55 2.35 5.70 6 4.20 0.48 2.80 5.38 6 5.06 0.92 2.64 8.69 6

PCDD/F, pg/g 4.46 0.63 2.86 6.59 6 5.11 0.58 3.33 6.72 6 5.79 1.07 2.94 10.08 6

CoPCB, pg/g 31.46 2.96 21.62 42.41 6 33.07 4.48 20.88 47.70 6 29.74 4.00 20.40 46.64 6

PCB, ng/g 17.26 1.94 12.74 24.65 6 21.96 2.33 13.16 27.37 6 17.03 2.79 9.87 28.59 6

IndPCB6, ng/g 9.52 1.07 7.16 13.97 6 11.41 1.24 7.16 14.82 6 9.64 1.52 5.66 16.12 6

PBDE, ng/g 0.58 0.06 0.34 0.76 6 0.50 0.05 0.34 0.68 6 0.66 0.10 0.36 0.92 6

PCDD/F-TEQ98, pg/g 1.37 0.21 0.76 2.00 6 1.68 0.21 0.93 2.44 6 1.77 0.35 0.93 3.13 6

TOT-TEQ98, pg/g 2.52 0.36 1.56 3.64 6 3.24 0.26 2.40 4.13 6 2.94 0.54 1.67 5.11 6

PCDD/FTEQ06, pg/g 1.03 0.16 0.58 1.50 6 1.24 0.15 0.71 1.78 6 1.33 0.26 0.70 2.33 6

TOT-TEQ06, pg/g 1.95 0.28 1.22 2.82 6 2.45 0.21 1.68 2.97 6 2.29 0.41 1.32 3.95 6

Three-spined PCDD, pg/g 1.04 - - - 1 1.08 0.48 0.19 2.13 4 0.91 0.41 0.50 1.31 2

stickleback PCDF, pg/g 5.92 - - - 1 3.03 0.95 1.14 5.21 4 5.76 2.53 3.24 8.29 2

PCDD/F, pg/g 6.96 - - - 1 4.10 1.42 1.33 7.33 4 6.67 2.94 3.73 9.60 2

CoPCB, pg/g 55.62 - - - 1 118.05 77.66 24.67 350.40 4 49.29 14.59 34.70 63.88 2

PCB, ng/g 20.97 - - - 1 38.21 24.48 8.24 111.39 4 19.66 5.33 14.33 25.00 2

IndPCB6, ng/g 11.61 - - - 1 20.78 13.50 4.58 61.17 4 11.11 2.67 8.44 13.78 2

PBDE, ng/g 0.52 - - - 1 0.52 0.25 0.20 1.25 4 0.54 0.18 0.36 0.72 2

PCDD/F-TEQ98, pg/g 1.59 - - - 1 0.74 0.22 0.31 1.19 4 1.59 0.64 0.95 2.24 2

TOT-TEQ98, pg/g 3.21 - - - 1 2.52 0.91 0.94 5.08 4 3.08 1.09 1.99 4.17 2

Baltic Proper (BPr) Gulf of Finland (GoF) Bothnian Sea (BS)


Supplement Table 2. Mean, standard error, minimum and maximum concentrations of organohalogens and WHO-TEQs in fresh weight (calculated according toTEFs from Van den

Berg et al. 1998 and 2006). Values are for all ages of sprat, 1- to 3-year-old herring, and 1-, 2- and 3-year-old salmon from the three study areas. N refers to number of homogenates

/ pools for sprat and herring and for individuals of salmon.

1







PCDD/FTEQ06, pg/g 1.32 - - - 1 0.64 0.18 0.28 1.01 4 1.31 0.50 0.81 1.82 2

TOT-TEQ06, pg/g 2.63 - - - 1 1.91 0.63 0.77 3.66 4 2.53 0.86 1.67 3.39 2

Salmon PCDD, pg/g - - - - - 1.39 0.14 1.15 1.90 5 - - - - -

1st sea-year PCDF, pg/g - - - - - 8.79 0.99 7.07 11.71 5 - - - - -

PCDD/F, pg/g - - - - - 10.18 1.12 8.21 13.61 5 - - - - -

CoPCB, pg/g - - - - - 105.23 12.49 73.31 141.30 5 - - - - -

PCB, ng/g - - - - - 128.74 17.87 83.00 184.39 5 - - - - -

IndPCB6, ng/g - - - - - 71.14 10.00 43.74 100.78 5 - - - - -

PBDE, ng/g - - - - - 1.59 0.13 1.30 1.94 5 - - - - -

PCDD/F-TEQ98, pg/g - - - - - 4.12 0.50 2.82 5.44 5 - - - - -

TOT-TEQ98, pg/g - - - - - 10.03 1.28 7.34 14.23 5 - - - - -

PCDD/FTEQ06, pg/g - - - - - 2.99 0.34 2.21 3.96 5 - - - - -

TOT-TEQ06, pg/g - - - - - 6.56 0.77 5.11 9.12 5 - - - - -

Salmon PCDD, pg/g 1.27 0.07 0.93 1.54 9 2.39 0.21 1.93 2.74 4 1.77 0.13 1.47 2.14 5

2nd sea-year PCDF, pg/g 10.95 0.58 7.89 13.13 9 16.63 1.37 13.68 19.39 4 15.99 1.02 13.50 18.87 5

PCDD/F, pg/g 12.23 0.65 8.82 14.67 9 19.06 1.60 15.61 22.27 4 17.75 1.15 14.97 21.01 5

CoPCB, pg/g 197.35 11.16 151.18 246.33 9 166.25 18.06 142.21 218.79 4 105.60 2.66 96.49 110.93 5

PCB, ng/g 99.43 4.13 78.56 111.23 9 196.39 19.21 167.04 249.97 4 130.88 5.73 111.87 143.80 5

IndPCB6, ng/g 49.55 2.04 38.55 55.77 9 108.32 10.11 92.30 135.76 4 81.76 3.74 69.06 89.95 5

PBDE, ng/g 3.06 0.12 2.43 3.50 9 2.83 0.26 2.48 3.59 4 3.18 0.11 2.83 3.50 5

PCDD/F-TEQ98, pg/g 3.90 0.22 2.78 4.68 9 7.56 0.56 6.39 9.05 4 6.68 0.49 5.50 8.05 5

TOT-TEQ98, pg/g 10.83 0.53 8.37 12.39 9 17.14 1.30 14.33 20.48 4 12.81 0.69 11.00 14.73 5

PCDD/FTEQ06, pg/g 2.97 0.16 2.14 3.55 9 5.45 0.40 4.57 6.45 4 4.77 0.33 3.96 5.70 5

TOT-TEQ06, pg/g 8.60 0.43 6.65 10.00 9 11.51 0.86 9.62 13.04 4 9.25 0.45 8.03 10.50 5

Salmon PCDD, pg/g 2.60 2.60 2.60 1 2.82 2.82 2.82 1 - - - - -

3rd sea-year PCDF, pg/g 22.14 22.14 22.14 1 19.34 19.34 19.34 1 - - - - -

PCDD/F, pg/g 24.79 24.79 24.79 1 22.16 22.16 22.16 1 - - - - -

CoPCB, pg/g 247.62 247.62 247.62 1 179.11 179.11 179.11 1 - - - - -

PCB, ng/g 188.61 188.61 188.61 1 247.39 247.39 247.39 1 - - - - -

IndPCB6, ng/g 95.05 95.05 95.05 1 138.56 138.56 138.56 1 - - - - -

2







PBDE, ng/g 6.27 6.27 6.27 1 3.19 3.19 3.19 1 - - - - -

PCDD/F-TEQ98, pg/g 7.98 7.98 7.98 1 9.44 9.44 9.44 1 - - - - -

TOT-TEQ98, pg/g 17.66 17.66 17.66 1 20.93 20.93 20.93 1 - - - - -

PCDD/FTEQ06, pg/g 6.06 6.06 6.06 1 6.71 6.71 6.71 1 - - - - -

TOT-TEQ06, pg/g 13.22 13.22 13.22 1 13.95 13.95 13.95 1 - - - - -

3

Biomagnification of organohalogens in Atlantic salmon (Salmo salar) from its main prey species in three areas of the Baltic Sea

Documents