Survey of consumption fish from swedish waters for chlorinated pesticides and polychlorinated biphenyls

Chemosphere, Vol. 33, No. 5, pp. 791-799, 1996

Pergmnon

Pll: S0045-6535(96)00235-4

Copyright © 1996 Elsevier Science Ltd Printed in Great Britain. All fights reserved 0045-6535/96 $15.00 + 0.00

Survey of Consumption Fish from Swedish waters for Chlorinated Pesticides and Polycldorinated Biphenyls S.S. Atuma, C-E Linder, A. Wicklund-Gtynn, O. Andersson and L. Larsson National Food Administration, Box 622, S-75126 Uppsala, Sweden (Received in Germany 24 April 1996; accepted 27 May 1996) ABSTRACT In this s~a'vey eighty-seven samples of consumption fish comprising mainly of salmon, pike, eel, herring, whitefish, sea-trout, perch, pike-perch, mackerel, cod, flounder, plaice and sole collected between 11992 - 1993 have been analysed for the levels of chlorinated pesticides and polychlorinated bipbenyls (PCBs). Considering the diversity in the sizes and assortment, locations, and time of catch, the samples are deemed to represent the normal fish variety available to the local people at various seasons of the year. V'~adly all the samples contained CB-153 levels below the new established maximum limit of 0.1 mg/kg flesh weight. Results are compared with those obtained between 1985 and 1993. Neither the total PCB (on fat weight basis) nor CB 153 as a marker showed any clear trend for most of the fish species analysed, particularly for the short period 1991- 1993. The pesticides, on the other hand, showed a rather good downward trend up till 1991, some of them seem to have virtually attained a steady state after 1991. Copyright © 1996 Elsevier Science Ltd Introduction Polychlovmted biphenyls and chlorinated pesticides are ubiquitously distributed in the environment and belong to a group of contaminants whose occurrence in the environment is of a serious concern to environmental chemists and toxicologists. This is due to their resistance to degradation in the environmcmt as well as their potential toxicity. It is now established that fish and fish products constitute an important route of human exposure to organochlorine compounds (OCCs) in Sweden 1. A number of studies have shown that fish samples from various waterways, including those lying in prestine a,-eas, are contaminated at various levels2"4 . Thus the dietary intake of OCCs can be governed to a large extent by the quantity of fish consumed by individuals. Since the ]Latesixties, levels of organochlorine pesticides and PCBs in fish from Swedish waters have been monitored 5-7 by the Swedish National Food Administration. Previously the PCB analyses were based mainly on "total PCB", which was equivalent to the sum of 14 PCB peaks measured on a packed column. Today, P.CB levels are evaluated on the basis of individual congeners which give more infolmafion with regard to exposure assessment. During the period 1992 - 1993 eighty-seven fish samples collected from grocery stores in several areas in Sweden (Fig 1) were analysed separately:, and in some cases pooled, for OCCs. The objeOives of the present investigation can be summarised as follows: 1. to monl'tor the levels of OCCs in a variety of fishes commonly found in the Swedish markets; 2. to check what effect the change in the currently accepted/permitted levels of PCBs based entirely

791

792 on CB 153 would have in monitoring programs; 3. to provide some of the data needed for the estimation of the dietary intake of OCCs in Sweden. For environmental assessment it has been considered adequate to monitor routinely CB 153 based entirely on abundance, response, resolution and prevalence just in the same way as "total PCB" had earlier been used8. This may not allow any conclusions on the contents of the toxicologically relevant PCB congeners in the residues present but it can allow some measure of assessment of the potential health risk3.

Sampling Samples of various sizes, age, sex and exposure were bought from grocery stores in Uppsala and other parts of Sweden during the period 1992-1993 (Fig. 1). The samples were bought at random at different times of the year, and were not aprotq suspectedof any contamination nor were they regarded as being representative of each species population with respect to contaminant levels. However, they represented a normal consumer's choice. Epaxial muscle was taken from each fish sample and stored frozen (-20oc) until analysis. For herring samples the skin was included. The number of each fish species analysed varied, depending e~tirely on availability at the time of sample coUection. Several fish of the same species do not necessarily imply that they came from the same sampling site (fig. 1).

Fig

1.

Sampling sites for the fishes (1992 - 1993)

Stm:im

nJa

Namber at'

aaalym

a b c

5m-~ SaLmo. Trout

t2 7 3

• f

Cod w, a a ~

5 2

g h i k !

~ ~ Bake ego~pe~ Bab'b~

2 z t 4

m

P~l~:e

4 4

$8 |

;mr

P

8

•

~

p

Ttut~

3

q

Pach

5

r

Pike

8

s

Mscka~

5

t

~1

5

793 Method

The ar~flytical methodology follows essentially the same extraction technique, clean-up and GC/ECD analysis employed in earlier studies2,9,10. Briefly, all samples were homogenized and extracted with acetom'Jbexane and hexane/ether. Clean-up consisted of pretreatment with H2SO 4 to remove the bulk of the lipids and then column chromatography on silica gel for the separation of the PCBs from other OCCs. In some cases it was necessary to further separate the PCB fraction on an HPLC using hypercarb colunm11. The fractions obtained were concentrated to appropriate volumes and analysed for PCB congeners (CBs 28, 52, 101, 118, 138, 153 and 180) and some chlorinated pesticides (HCB, or- and y- HCH, p,p'-DDD, -DDE and -DDT) using GC equipped with dual columns and dual detectors. The dual column system helps to eliminate the problems of coelution or overlapping of two or more congeners on any of the columns. Solvent blanks were run through the entire analytical procedure. Quantification was carried out by comparison of the gas chromatographic peak heights of the OCCs in the samples with those of standards.

Result.,; and Discussion

The practical working range for all the target CBs was 1 - 30 pg/ul solution. The results were calculated using peak-height measurements. Detection limits based on a signal-to-noise ratio of 2.5 were approximately 50 pg/g sample. Results are based on the average of at least three samples of the same fish species (Tables la and lb). Those from samples comprising of only one or two in number were cxmsidered statistically nonsignificantbut the levels from freshwater fshes are, however, shown in Table 2. Many of the species analysed are migratory and so would be exposed to considerable variation in OCC concentration depending on individual movement patterns 12. Similarly, due to the lipophilicity of these compounds and the variable lipid content of the samples, wet weight coneenl.~rationsusually show large variations. Tables la & lb show the residue levels of both the chlorinated biphenyl congeners and the pesticdes in a variety of fishes, found in this study. As expected and as many other studies have shown 13 the sample.,; from the Baltic are still by far more contaminated than those from any other watershed. For herring, for example, the pollution load is over two times higher than the levels in the west coast except :for the s-HCH.This is obviously due to the fact that the Baltic has always been more vulnerable than the west coast because it is an enclosed sea with a high load of OCCs from both the industrial and agricultural activities. As w o r d be expected the fat content in the edible part of the fish is a major determinant of the fresh weight OCC level (Tables la and lb). However, in spite of the high fat content of the mackerel from Skagen'ak (west coast), the levels of the contaminants were very low compared to those in herring sample.,; from both the west coast and the Baltic. This may, for instance, be due to food choice by or migration patterns of the different fish species. The importance of the fat level as a determinant of the fresh weight OCC level is illustrated by the low OCC levels in cod from the Baltic sea. The levels of the chlorinated pesticides in fish samples from the G-ulfof Bothnia and Baltic Proper indicated decreasing concentrations of OCCs from 1976 -19847,13. Table la shows that the levels of HCB, I-[CH and DDT in herring from both the west coast (Skaggerak) and the east coast (Baltic sea) are about 5-fold and 10-fold, respecticely, lower than those obtained in similar samples in 198413 It is difficult to make a definite comparison between these results and those from other studies since the prevailing factors affecting sampling, season, locations and sizes of the samples were not regulated or properly defined. This is illustrated by results obtained by Andersson et a114 where levels of OCCs in herring found in the Kattegat (July 1979) seemed to be of the same magnitude as those observed in the Baltic proper (September 1979).

794 TABLE la. Levels (range and average) of some chlorinated pesticides and CB-153 found in different fishes bought fi'om fish shops (1992/93) in ug/kg flesh weight

Fish Type (number)

Probable place of catch

Weight kg

Lipid %

HCB

s-HCH

s-DDT

CB 153

Mackerel (5)

Skaggerak

0.19-0.66 0.42

1.6-25 12

0.20-2.1 1.0

0.05- 10 4.5

1.3-11 6.0

0.44-4.1 2.1

Herring* (4)

west coast

0.10-0.19 0.16

6.4-12 9.1

1.5-2.5 2.1

3.8-6.6 5.2

11-51 24

4.0-9.0 6.4

Herring# (8)

Baltic sea

0.029-0.068 0.047

3.6-8.2 6.1

3.6-7.6 5.3

1.6-7.0 4.6

22-112 67

8.0-36 20

Whiter'rob (4)

Baltic sea

0.30-0.60 0.45

0.80-4.7 2.3

0.70-4.2 2.6

0.05-2.0 0.70

4.7-45 22

2.5-19 9.99

Cod

(4)

Southern Baltic sea

0.33-1.50 0.81

0.50-0.79 0.65

0.10-0.30 0.22

0.20-0.40 0.30

1.8-7.8 5.0

0.50-1.9 1.3

Plaice

west coast

0.41-0.47 0.45

0.62-1.8 1.01

0.10-0.20 0.13

0.05-0.50 0.22

0.50-1.4 0.87

0.30-1.7 0.90

Eel (5)

Baltic sea

0.48-1.7 1.0

14-26 20

3.1-6.9 4.8

5.2-12 8.4

77-135 98

43-112 69

Pike

Baltic sea

0.48-1.7 1.0

0.51-0.77 0.65

0.10-1.1 0.52

0.10-0.40 0.18

1.7-24 11

3.5-10 7.7

Perch (3)

Baltic sea

0.24-0.57 0.40

0.79-1.0 0.95

0.40-1.0 0.73

0.10-0.40 0.27

6.6-9.5 7.6

3.1-3.8 3.5

Halibut (5)

North sea

0.59-15 8.9

0.59-6.6 2.9

0.05-2.9 1.3

0.05-1.5 0.58

0.05-19 9.1

0.05-5.1 2.0

Sole (4)

West coast

0.28-0.45 0.33

0.66-0.84 0.78

nd

0.05-0.20 0.14

0.30-0.50 0.42

0.20-0.50 0.35

(3)

(4)

* Pooled samples comprising of 6-7 individual fish each # Pooled samples each of which comprises of 9 - 27 individual fish (include samples wen from Gulf of

Bothnia). nd = not detected

795 TABLE lb. Levels (average and range) of some chlorinated biphenyi congeners in differentfishes bought from the fish shops (1992/93) in ug/kg fresh weight (place of catch and weight as shown in Table la)

Fish Type (numlx;r)

Lipid %

CB 28

CB 52

CB 101

CB 118

CB 138

C-'B 153

CB 180

Mackerel a (5)

12 1.6-25

0.09 0.03-0.28

0.41 0.05-1.2

0.98 0.16-2.2

0.71 0.04-1.5

1.6 0.27-3.3

2.1 0.44-4.1

0.38 0.05-0.80

Herring *a (4)

9.2 6.4-12

0.30 0.05-0.99

0.86 0.05-2.5

2.9 1.4-4.7

1.9 0.5-3.4

3.8 1.8-7.4

6.4 4.0-9.0

0.79 0.05-1.7

Herrins# b (8)

6.1 3.6-8.2

0.38 0.05-0.80

2.3 0.80-3.6

7.9 3.2-13

6.6 2.2-13

13 4.9-25

20 8.0-36

4.8 0.05-13

Whitefu~h b

2.3 0.80-4.7

0.17 0.05-0.39

1.1 0.05-2.0

3.6 0.60-6.7

2.0 0.60-5.1

6.6 1.6-13

9.9 2.5-19

3.0 1.0-7.0

0.65 0.50-0.79

0.06 0.15 0.05-0.10 0.10-0.20

0.60 0.50 0.30-0.70 0.20-0.60

0.90 0.40-1.2

1.3 0.50-1.9

0.35 0.10-0.50

1.0 0.62-1.8

0.07 0.05-0.10

<0.05

0.13 0.30 0.10-0.20 0.20-0.50

0.63 0.20-1.2

0.90 0.30-1.7

0.15 0.10-0.30

Eel b

20 14-26

0.81 0.04-3.2

~4 2.9-10

19 5.3-43

30 15-64

S0 25-89

69 43-112

19 9.3-26

Pike b

(4)

0.65 0.51-0.77

0.07 0.05-0.10

0.39 0.05-0.6

2.5 0.60-4.1

1.7 0.60-2.8

4.2 1.4-8.4

7.7 3.5-13

2.6 1.4-4.1

Perch b (3)

0.95 0.79-1.0

<0.05

0.27 0.20-0.40

1.1 0.9-1.4

0.97 0.80-1.1

2.2 1.8-2.6

3.5 3.1-3.8

1.5 0.9-2.2

Halibut ¢ (5)

2.9 0.59-6.6

0.17 0.46 0.05-0.50 0.05-0.80

0.72 0.05-1.6

0.87 0.05-2.1

1.2 0.05-3.4

2.0 0.05-5.1

0.62 0.05-1.9

0.09

0.07

0.25

0.35

0.00

(4) Cod b

(4) Plaie@I

(3) (s)

Sole a

0.78

(4)

0.66-0.84

west coast Baltic sea North sea

<0.05

<0.05

0.05-0.1 0.05-0.10 0.10-0.40 0.20-0.50 0.05-0.I0

* #

same as m Table la same as in Table la

796 One of the eel samples from the Baltic (Table lb) had a CB 153 level of 112 ~tg/kg fresh weight which is about 12% higher than the acceptable level. Since this was just a single sample out of a total of eighty-seven fish samples analysed, it was regarded as an isolated case. Only two of the seven salmon samples analysed were from the Baltic sea, two from lake Vanern and the remaining ones were from fish farms. The levels of contaminants in the Baltic samples were on the average 199 ~tg/kg fresh weight for EDDT, 6.9 ~tg/kg HCB, 11 ~tg/kg ]~HCH and 37 ~tg/kg CB 153, and are higher than the levels in most fish species from other sources in this study. Table 2 shows the levels of some OCCs in freshwater fish species which are not taken up in Table 1. Apart from two samples, pike and salmon from lake Vanern, comprising of two subsarnples each, all the other samples were single species. These results should therefore be interpreted or viewed with caution. Nevertheless, the total polution load seems to be generally higher in lake V~inern than in the other smaller lakes (I-Ijalmaren, Vombsj6n and SOvdesjOn). Figs 2a and 2b show the mean concentrations of total PCB and CB 153 respectively(lipid weight basis) in different fish species from Gulf of Bothnia, Baltic sea and the west coast of Sweden, analysed between 1987 and 1993. The total PCB determination was based on a packed column system used in our laboratory until 1993. Although the samples were of various sizes and not necessarily collected at the same seasonal period of the year nor from the same location, some concentration trends were apparent but very minimal. Levels of s-DDT (not shown here) and ~PCB in herring from the Baltic sea (1991) were relatively high in comparison to herring samples from the west coast of Sweden. Samples of herring (east coast) from 1989 to 1993 were analysed for CB 153 while other samples were analysed only from 1991 to 1993 (fig.2b). In most of the cases, no clear trends could be discerned. This is actually in agreement with our earlier findings where we showed that both ~ETCB and ]E~DDT seemed to have attained fairly constant levels since 19902. It is therefore di~cult to discuss temporal trends in this paper since there were no defmed criteria for the sampling and choice of the fishes. These criteria have been emphasized as being necessary in studies of spatial distribution and temporal trend monitoring of concentrations of organochlorine compounds in various watersheds 13. Perttilit et all5 reported a correlation between age and concentrations of DDT and PCB in Baltic herring samples from different age classes. Jensen et al9 had earlier pointed out that the higher levels of contaminants in older fishes were not solely the effect of longer exposure time but also emanated from differences in migrational habits between various age classes. 2-4 year-old herring samples have shown far less variability in concentrations than the older species owing to less migratory tendencyl3,16 and can therefore be suitable for trend studies. In the light of the above, the results obtained in this study can be classified as results from investigative analyses sin~ they emanate from scattered or pooled samples, randomly collected to check for compliance with the Swedish National Food Administration ordinance for acceptable levels of these pollutants in consumption fishes offered for sale in Sweden. These results are therefore not really intended for temporal trend studies. Conclusion

The decrease of most of the chlorinated compounds studied in fish seems to have deaccelerated compared to studies carried out about 10 years ago. However, in view of the continuing unc~rtainty about the possible effects of low-level exposure (on humans) it is considered prudent to avoid excessive intake of fish with elevated OCC levels as contained in the dietary advisory8. The present results indicate that the cmrent Swedish limits for PCBs (based on CB 153) are far from being exceeded.

797

(a) 15000-

X~l

. 10000-

%

o/

5000-

0 -

/

"'"-.....

I

I

I

i

l

I

I

J

year

Co)

CB 153 s

1000--

11

-If

,~,*"

~..,

3 ~ , , , ~ / ~

~.,A

,"

r" 0

--

I

I

I

I

I

1~8

1000

lg~l

1~2

1~3

year

Fig. 2

Time trends of ~PL-'B (a) and CB 153 (b) in a limited number of fish species from the East coast (nos 1, 2, 3, 4, and 5) and West coast (no 6) of Sweden (Baltic and Kattegat/Skagerrak) between 1985 - 1993, expressed on lipid weight basis. 1. Sea trout 2. Salmon 3. Herring (east coast) 4. Whitefish 5. Eel 6. Herring (west coast)

798

TABLE 2. Levelsof some chlorinatedpesticidesand CB-I53 found i different fresh wate~ fishes (1992/93) in ~tg/kgfresh weight.

Probable FJsh Type Weight place of kg catch Lake Vinern

Lipid %

HCB

s.-HCH

s-DDT

CB-153

Pike (2) Perch Pike-perch Salmon (2) Trout Whitefmh

1.5 0.49 0.91 2.2 1.2 0.93

0.52 0.91 0.55 2.5 1.5 8.1

0.20 0.10 0.10 0.75 0.70 19

0.10 0.20 0.20 1.4 0.05 3.0

39 4.7 4.0 12 7.4 140

25 3.3 2.6 9.0 4.9 62

Pike Perch I'ike-perch

1.4 0.50 1.1

0.69 0.74 0.56

0.10 0.05 0.05

0.20 0.05 0.05

1.5 4.6 1.0

1.2 4.1 2.0

VombsjSn Pike

1.4

0.48

0.05

0.05

2.8

1.1

Lake SSvdesjSn

0.67

0.59

0.10

0.I0

2.1

2.1

Lake

~jv-

nHtren

T~mk~

Pike-perch

~)~2,-,,?la

799

Acknowledgement We would like to thank Lena Hansson for her valuable technical assistance and ~ his immense interest and valuable discussion in this work.

Johnsson for

References 1 Asphmd L, Svensson B-G, Nilsson A, Eriksson U, Jansson B, Jensen S, Wideqvist U, Skeffving S. PCB, p,p'-DDT and p,p'-DDE in human plasma related to fish consumption. Arch. Environ. Health 1995; 46:477-486 2 Atuma S.S, Andersson O, Linder C-E, Hansson L. Levels of some organochlorine compounds in sea trout (salmo trutta) and whitefish (coregonus lavaretus) ~om the Gulf of Bpthnia. Aqua Fennica 1993; 23: 221-226. 3 Falandysz J, Kannan K, Tanabe S, Tatsukawa R. Organochiorine pesticides and polychlorinated biphenyis in cod-liver oils: North Atlantic, Norwegian sea, North sea and Baltic sea. Ambio 1994; 23: 288-293. 4 Koistinen J, Paasivirta J, Vuorinen P.J. Dioxins and other planar polychioroaromatic compounds in Baltic, Finnish and Arctic fish samples. Chemoshere 1989, 18, 527-530. 5 West,s0 G, North K: Levels oforganochiorine pesticides and polychiorinated biphenyls in fish caught in Swedish water areas or kept for sale in Sweden 1967-1970. V~r F0da 1970; 9-10. 6 Noren K, Rosen G. Levels of organochlorine pesticides and PCB in fish fi'om Swedish waters. V& FOda, suppl Nr 1, 1976; 28. 7 Andersson O, Linder C-L, Vaz R. Levels of organochlorine pesticides, PCBs and certain other organohaiogen compounds in fishery products in Sweden, 1975-1982. V~zrFc~ta, awppl 1 1984; 36. 8 Wicldund Glynn A, Darnerud P.O, Andersson 0, Atuma S.S, Johnsson H, Linder C-E, Becket W Revised fish consumption advisory regarding PCBs and dioxins. Swedish National Food Aagninistration, Uppsala, Sweden - SLV Report 1996;No 4 : 4 6 pp. 9 Jensen S, Johnels A.G, Olsson M, Ottedind G. DDT and PCB in herring and cod from the Baltic, Kattegat and the Skagerrak. Ambio Special Report No 1. 1972; 1: 71-85. 10 Atun~t S.S, Andersson 0. Separation of PCB congeners using active coal columns. Chemo~phere 1993; 27: 1-8. 11 Atunut S.S, Hansson L. Using porous graphitic carbon column for HPLC separation and isolation of potentially toxic PCB congeners in fish. Dioxin "94 - Proceedings of the 14th international symposium on chlorinated dioxins, PCB and related compounds, Kyoto, Japan, Nov. 1994; 19" 81-84 12 GuUec M.C, Connell D.N. Bioaccumulation of chiorohydrocarbon pesticides by fish in natural environment. Chemosphere 1992; 25, 1579-1587. 13 Bignert A, G-OthbergA, Jensen S, Litzen K, Odsj6 T, Olsson M, Reuthergtrdh L. The need for "The need for adequate biological sampling in ecotoxicological investigations: a retrospective study of twenty years pollution monitoring", Sci. TotalEnvironm. 128, (1993), 121-139 14 Ande~Ison O, Linder C-E, Olsson M, Reuterglrdh L, Uvemo U-B, Wideqvist U. Spatial differences and temporal trends of organochlorine compounds in Biota from the northwestern hemisphere. Arch. Environ. Contain. Toxicoi 1988; 17: 755-765. 15 Perttil~t M, Tervo V, Parmanne R. Age dependence of the concentration of harmful substances in Baltic herring (Clupea harengus). Chemosphere 1982; 11:1019-1026 16 Pannmme P. Growth, morphological variation and migrations of herring (Clupea harengus L) in the northern Baltic Sea. Fish. Res. 1990; 10: 1-48.