The Science of the Total Environment, 99 (1990) 37 47 Elsevier Science Publishers B.V., Amsterdam

37

L E V E L S OF S E L E N I U M A N D A N T I O X I D A T I V E E N Z Y M E S F O L L O W I N G O C C U P A T I O N A L E X P O S U R E TO INORGANIC MERCURY

LARS BARREG/~RD*

Department of Occupational Medicine, Sahlgren's University Hospital, S:t Sigfridsgatan 85, S.41266 GSteborg (Sweden) YNGVAR THOMASSEN

National Institute of Occupational Health, Pb 8149 Dep, N-0033 Oslo 1 (Norway) ANDREJS SCHt~TZ

Department of Occupational Medicine, University Hospital, S-221 85 Lund (Sweden) STEFAN L. MARKLUND

Department of Clinical Chemistry, University Hospital, S-901 85 Ume~ (Sweden) (Received December 15th, 1989; accepted January 16th, 1990)

ABSTRACT Levels of selenium and mercury in blood and urine were analysed in 37 male workers exposed to elemental mercury vapour in a chloralkali plant and in 39 unexposed controls of the same age. Mean urinary Hg was 223 nmol 1 1 (15 nmol/mmol creatinine) in the exposed group and 26 nmol l (2.0nmol/mmol creatinine) in the controls. Mean blood and plasma Hg levels were 46 and 36 nmol 1 1, respectively, in the exposed group, as compared with 17 and 7 nmol 1-1 in the controls. The concentrations of Se in plasma and erythrocytes did not differ between the two groups. Urinary Se levels were, however, slightly but significantly lower in the exposed group (median values 23 vs 29 nmol/mmol creatinine), and there was a negative correlation between urinary Se and plasma Hg in the exposed group. This may be due to a retention of Se in the kidneys. In a subgroup of exposed workers and controls, glutathione peroxidase, superoxide dismutase and catalase were also analysed. No differences were found between the groups with respect to these antioxidative enzymes. The effect on Se status of moderate Hg exposure seems to be of minor clinical importance. INTRODUCTION

Inorganic mercury is present in the working environment in the chloralkali industry and in mercury mines. It can also be found in the production of thermometers, fluorescent tubes and mercury batteries. Low level exposure is seen among dental personnel working with amalgam, a mixture of mercury and other metals. In many countries, dental amalgam is also the main non-occupa*Author to whom correspondence and reprint requests should be addressed.

0048-9697/90/$03.50

© 1990 - - Elsevier Science Publishers B.V.

38 tional source of exposure to inorganic mercury, whereas fish is the main source of methyl mercury exposure [1]. In occupational settings the predominant form of mercury is elemental mercury (Hg °) vapour. This is readily absorbed through the lungs and distributed to various organs before elimination by urinary and faecal excretion. Urinary and blood mercury levels are used for biological monitoring of occupational exposure. The main target organs following exposure to mercury are the central nervous system and the kidneys [1]. The interaction between mercury and other elements, in particular selenium, has been an area of interest in the last few decades. In animal experiments, administration of selenium can protect against methyl mercury toxicity [2-4]. Furthermore, it reduces the nephrotoxic effect of mercuric mercury [5]. After simultaneous administration of Se and Hg salts an increased whole-body retention of both elements is seen [2, 6]. After Se administration in rats and pigs the distribution of Hg is also affected, resulting in higher levels in blood and lower levels in kidneys [7]. However, little is known about the kinetic interaction between Se and Hg in man. In a study of autopsy samples from retired miners exposed to Hg ° vapour, concentrations of Hg and Se with a molar ratio close to unity were found in the brain and kidneys [8]. From a cross-sectional Norwegian study of workers exposed to Hg ° vapour, urinary excretion of Se was found to be higher when compared with that of unexposed controls [9]. However, the available information is contradictory: in another study the urinary Se levels were lower in the exposed group [10]. Selenium is an essential trace element in man [2, 3]. Selenium exists in the body mainly as selenocysteine and selenomethionine incorporated in various proteins [11-14]. At least three of these are glutathione peroxidase (GSHPX) isoenzymes: the first to be discovered, cytosolic GSHPX [12], the phospholipid hydroperoxide GSHPX [13] and the secretory GSHPX [14]. These enzymes reduce hydrogen peroxide and a variety of organic hydroperoxides, thereby contributing to the protection against oxygen toxicity [15]. Other antioxidant enzymes are the superoxide dismutases (SOD), which catalyse the disproportionation of the superoxide anion radical to hydrogen peroxide and oxygen [16], and catalase, which catalyses the disproportionation of hydrogen peroxide to water and oxygen [15]. Catalase is also of interest in connection with mercury exposure, as it is the enzyme repsonsible for the oxidation of Hg ° to Hg 2÷ in blood and tissues. The reaction is brought about by the catalase-hydrogen peroxide complex I [17]. The aim of the present study was to determine Se levels in workers exposed to mercury vapour and those in unexposed controls. Furthermore, in a subgroup of workers and controls the levels of three antioxidative enzymes, GSHPX, SOD and catalase, were determined. SUBJECTS We examined 41 male chloralkali workers exposed to mercury vapour at a

39 chloralkali plant and 41 age-matched, non-exposed controls in the same company. The exposed group included all available workers, except one diabetic. Six subjects took Se supplements and were thus excluded, leaving 37 exposed workers and 39 controls for the study of Hg and Se levels. The mean ages of the exposed workers and controls were 35 and 36 years, respectively. The chloralkali workers had been exposed to mercury for 1-31 years (2 = 10 years) (Table 1). The average Hg levels in air in the exposed part of the factory have been 25-50 pg m-3 during the last decade. The controls had no present or historical occupational exposure to mercury. In the exposed group, 16/37 were current smokers versus 19/39 among the controls. In randomly chosen subgroups of exposed workers and controls, the activities of GSHPX (20 exposed and 22 controls), and SOD and catalase (11 exposed and 12 controls) were determined. METHODS Blood and urine (morning) samples were obtained in metal-free tubes and bottles, respectively, during the same period for exposed and control subjects. Urine, whole blood, plasma and erythrocytes were frozen the same day and stored at - 25°C. Mercury in whole blood (B-Hg), plasma (P-Hg) and urine (U-Hg) was analysed using the cold vapour atomic absorption (CVAAS) technique. After the digestion of urine with KMnO4 and H2SO 4, and blood with HC104 and HNO3, Hg 2+ was reduced to Hg ° with stannous chloride [18, 19]. Each sample was analysed in duplicate. In the present range of U-Hg concentrations (5880 nmol 1-1) the coefficients of variation (CV) were 7 and 5% for the low and high values, respectively. The CV of P-Hg values was 4% and there was no significant difference in precision between high and low values. The accuracy has been tested and found acceptable when blood and urine samples were also analysed by another laboratory using radiochemical neutron activation analysis (RNAA). In the range 50-400nmoll-', our results averaged 99% of those determined by RNAA. In the range 10-50 nmol 1 1, our analyses showed, on average, 5 nmol higher values than RNAA. Selenium in plasma was determined using electrothermal atomic absorption spectrophotometry [20]. The plasma samples were first thawed and then diluted (1:5) with a chemical modification solution containing 0.25% nickel nitrate and 0.2% Triton X-100. A Perkin-Elmer Model 5000 atomic absorption spectrophotometer equipped with a deuterium background corrector, a HGA-500 graphite furnace, an As-40 automatic sampler and a selenium electrode discharge lamp was used for all determinations. Calibration was made against standards based on human serum with a known selenium content [21]. The quality assurance of the selenium plasma determinations was ensured using certified reference serum (Seronorm Trace Element 105 Human Serum, Nycomed, Oslo, Norway) [22]. The mean concentration of selenium determined in the STE 105 reference serum was 1.18#mol 1 1 (SD, 0.04), while the reported concentration is 1.15 (SD, 0.08) t~moll '. The within-day CV was 3%.

40 Selenium in urine and in erythrocytes was determined by hydride-generation atomic absorption spectrophotometry. Aliquots of the samples were digested according to the IUPAC recommended decomposition procedure using nitric, sulphuric and perchloric acids, and Se(VI) was subsequently reduced to Se(IV) using hydrochloric acid [23]. A Perkin-Elmer Model 5000 atomic absorption spectrophotometer equipped with a jMHS-20 hydride generation system and a selenium electrode discharge lamp was used for all hydride measurements. Quality assurance of the determinations was ensured using Seronorm Trace Element Human Urine 108 and Human Whole Blood 904. The average values for urine, 0.61 _+ 0.03#molSel ', and whole blood, 1.19 + 0.04#molSel-', were in good agreement with the reported reference values of 0.63 +_ 0.04~molSe1-1 [23] and 1.20 + 0.02~molSel-' [24], respectively. The typical CV of the method was 4%. For the analysis of the enzymes CuZnSOD, GSHPX and catalase in erythrocytes, 150 #l of packed erythrocytes were lysed in 2.5 ml 5 mM sodium N-2hydroxyethyl-piperazine-N-1,2-ethane-sulfonic acid (HEPES), pH 7.40, with 50#M EDTA. After 15min the hemolysate was centrifuged. All analytical results were related to the hemoglobin contents of the hemolysates, which were determined with a standard cyanomethemoglobin assay. For the GSHPX assay, the hemolysates were treated with 1 m M potassium ferrocyanide and 8.7 mM sodium cyanide to inhibit the peroxidase activity of hemoglobin [25]. Treated hemolysate (25pl) was then added to 500#1 0.1 M sodium HEPES, pH 7.40, with 1 m M EDTA, 2 mM reduced glutathione, 1 U ml-1 glutathione reductase, 0.16 m M NADPH and 0.88 m M tert-butylhydroperoxide. The reaction was followed at 340 nm at 37°C. The CuZnSOD activity of the hemolysates was determined with the direct spectrophotometric method employing potassium superoxide [26, 27]. For catalase analysis the rate of H202 disproportionation was studied at 240 nm in a cuvette containing 3 ml of 10 m M H2 O: in 25 m M sodium HEPES, pH 7.40, with 0.25 mM EDTA at 25°C. One unit was defined as the activity that brought about disproportionation at a rate of 10 3s-1. For the exposed group, a cumulative exposure index was calculated for each subject from previous B-Hg levels, which had been determined since 1970. The index was calculated by integrating B-Hg as a function of exposure time. Three of the workers had been exposed before 1970. Their B-Hg levels during the 1960s were estimated by an occupational hygienist. Smoking habits, fish meals per week and recent dental care were also registered. Student's t-test and Wilcoxon's rank sum test were used when comparing the exposed group with the controls. For correlations between levels of Hg, Se and enzymes, Pearson's and Kendall's correlation coefficients were used. Associations between more than two variables were analysed using the stepwise multiple linear regression technique of a SAS programme package.

41 TABLE 1 Present levels of mercury in whole blood (B-Hg), plasma (P-Hg) and urine (U-Hg) and time-integrated B-Hg among chloralkali workers and controls Exposed subjects (n = 37a)

Controls (n = 39b)

Mean

Mean

Median

Range

36

37

18411

Median

Range

Age (years)

35

34

19-58

Exposure time (years)

10

10

1-31

Time-integrated B-Hg (years. nmol 1- ~)

848

523

34 4300

B-Hg (nmoll 1)

46

35

8-160

17

15

8-60

P-Hg (nmol 1-1 )

36

30

14-119

7

6

2-12

U-Hg (nmol l- 1)

223

170

65-880

26

22

5~0

15

11

U-Hg (nmol/mmol creatinine)

4.5-53

2.0

1.9

0.5-5.0

a16 smokers, b19 smokers. RESULTS

As expected, the concentrations of mercury in blood, plasma and urine were higher in the exposed subjects than in the controls (Table 1). The levels of mercury in plasma, whole blood and urine were highly intercorrelated in the exposed group, and to some extent also in the controls (Table 2). In the controls, P-Hg increased with age. In the exposed group the tendency was the opposite, but some of the older chloralkali workers had become foremen and were now less exposed to Hg. There were no correlations between the present U-Hg, P-Hg or B-Hg on the one hand and the time integrated B-Hg in the exposed group on the other (not shown in the table). In the controls the Hg levels tended to be somewhat lower among the smokers. The difference was statistically significant only with respect to P-Hg (~ = 6.0 vs 8.0nmoll-1). Selenium levels were not normally distributed (Table 3). The main deviation from normality was seen in the U-Se of the exposed group with one extremely high value (Fig. 1). Urinary selenium was significantly lower in the exposed group, as compared with that of the controls (Fig. 1 and Table 3). The levels of P-Se were approximately the same in exposed and control groups, as were the levels of selenium in erythrocytes (Ery-Se). The plasma and urinary selenium levels were positively correlated in both exposed and control subjects (Table 2). There were no correlations between Ery-Se and P-Se or U-Se. Urinary selenium was negatively correlated with Hg levels in all media when all subjects were analysed. In the exposed group this tendency was statistically significant only for P-Hg. Erythrocyte Se in the exposed group was negatively correlated with U-Hg. The Se levels were not correlated with time-

42 TABLE 2 Correlation coefficients (Kendall's Tau) b e t w e e n m e r c u r y and s e l e n i u m levels. U r i n a r y Hg and Se levels were corrected for u r i n a r y c r e a t i n i n e

All subjects U-Hg P-Hg B-Hg U-Se P-Se Ery-Se Exposed subjects U-Hg P-Hg B-Hg U-Se P-Se Ery-Se Controls U-Hg P-Hg B-Hg U-Se PoSe Ery-Se

P-Hg

B-Hg

U-Se

P-Se

Ery-Se

Age

0.72****

0.62**** 0.68****

- 0.26*** - 0.33**** - 0.20*

-0.01 - 0.05 - 0.02 0.27****

- 0.05 0,03 0.05 - 0,08 0,09

- 0.02 0.00 0.01 0.11 0.06 0.17

0.53****

0.50**** 0.66****

-- 0.09 - 0.25*

- 0.01 - 0.02

- 0,26* - 0,07

- 0.15 - 0.28* - 0.19 0.29* 0.16 0.27*

-

0.35***

0.31"* 0.39***

0.10

0.01

- 0.08 - 0.19 0.06

-

0.19

0.27*

0.03 0.13

0.20 0.08 0.19 0.25*

- 0.06 0.09 0.15 - 0.22 - O.02

0.15 0.33*** 0.26* - 0.05 - O.O2 0.11

****p < 0.0001; ***p < 0.005; **p < 0.01; *p < 0.05.

TABLE 3 Selenium levels in p l a s m a (P-Se), e r y t h r o c y t e s (Ery-Se) and u r i n e (U-Se) of chloralkali w o r k e r s and controls Exposed subjects

P-Se 0~mol1-1) Ery-Se (nmol g - l) U-Se (pmol1-1) U-Se (nmol/mmol creatinine)

Controls

n

Median

Range

n

Median

Range

37 30 a 37 37

1.09 1.29 0.35 23 c

0.63-1.55 0.51-2.02 0.14-1.41 13-116

39 32 b 39 39

1.11 1.22 0.41 29

0.67-1.82 0.81-1.77 0.154).76 20-47

a12 smokers. b13 smokers. Cp < 0.005, Wilcoxon's r a n k s u m test.

43

U-Se (nmol/mmol creat) 5O T

40

3O |

•

•

!

a

!

:,, !

I0

1e

•

r

i

exposed

controls

Fig. 1. Urinary selenium (U-Se) in workers exposed to mercury (n - 37) and in unexposed controls (n = 39). Median values indicated. One exposed "outlier" (U-Se - ll6nmol/mmol creatinine) not shown in the diagram.

integrated B-Hg. Nor were there any correlations between P-Se and Hg levels in either group (Table 2). Selenium levels were slightly lower in smokers than in non-smokers, but the differences were not statistically significant. As U-Se and Ery-Se in the exposed group were significantly correlated with several variables, stepwise multiple linear regression analysis was conducted on this group. In these analyses, one subject - - an "outlier" - - was excluded because his high U-Se would have otherwise totally determined the result of the regression calculations. With U-Se as the dependent variable, only P-Hg and P-Se met the 0.15 significance level for entry into the model. With these two independent variables in the model, the coefficient of regression for P-Hg (fl = -0.12) differed significantly from zero (p < 0.05), but the coefficient of regression for P-Se did not. With Ery-Se as the dependent variable, only U-Hg met the criterion for entry into the model (fl = - 0.014, p < 0.05). There were no statistically significant differences between exposed subjects and controls for GSHPX, SOD and catalase activities (Table 4). In these

44 TABLE 4 Glutathione peroxidase (GSHPX), superoxide dismutase (SOD), and catalase in erythrocytes of chloralkali workers and controls Enzyme

Exposed subjects n

x (SD)

Controls

Range

n

x (SD)

Range

GSHPX (pkat/g Hb)

20a

0.76-1.65

22h

SOD (U/rag Hb)

11

54

(5.6)

47-66

12

54

(3.1)

46-58

Catalase (U/rag Hb)

11

155

(18)

123~186

12

154

(22)

115-189

115(0.24)

1.21(0.30)

0.86-1.96

aNine smokers. bFour smokers.

subgroups, blood and urinary Hg levels were comparable to those in their respective whole groups (Table 1). Glutathione peroxidase in the total population (n = 42) was not significantly correlated with Se or Hg levels. In a subgroup with Ery-Se < 1.10nmol g-1 (n -- 29), there was, however, a weak positive correlation between Ery-Se and GSHPX, r = 0.40 (Pearson); p < 0.05. DISCUSSION

Mercury and selenium levels In the present study, the Hg and Se levels of the unexposed subjects were consistent with those reported for other healthy Swedish subjects [28]. The Se levels in blood and urine were rather similar in the exposed and control groups. Urinary selenium was, however, slightly lower in the exposed group than in the controls. This is in agreement with one [10] of the two [9, 10] previous studies, and it seems reasonable to assume that this is an effect of exposure to inorganic mercury. If exposure to Hg affects Se levels, the impact of present and previous exposure to Hg is probably different. As long as Hg accumulates in the kidneys, retention of Se in the kidneys in the form of a Hg-Se-protein complex may lead to reduced U-Se levels. This initial effect was seen in mice which had been given Hg and Se [29]. The slight negative correlation between U-Se and P-Hg (high values indicating recent exposure) could lend some support to this hypothesis. When, after an unknown period of time, kidney Hg has reached a steady state, the intake and excretion of selenium might be in balance. After exposure to mercury has ceased, the kidney Hg levels, and thereby also kidney Se levels, should decrease. This might be observed as a temporary increase in U-Se. Such an effect on the turnover of Se may explain why similar studies in workers exposed to mercury have reported different results [9, 10]. Further, geographical differences in the amount and forms of dietary Se might have an influence.

45 Plasma selenium and Ery-Se levels were not seen to be affected in this study. The slight, but statistically significant, negative correlation between Ery-Se and U-Hg in the exposed group may well be a random finding. Such a correlation, although not statistically significant, was also found by Suzuki et al. [30]. Thus, our finding could alternatively indicate that increasing kidney Hg levels lead to a flow of Se from blood to the kidneys where Se is bound to Hg, as mentioned above. Plasma selenium levels change rather rapidly with Se intake. Erythrocyte Se is less susceptible to such changes. In moderate occupational exposure to mercury the metabolic interaction between selenium and mercury seems to be complex. Multiple comparisons were made for a limited number of subjects both in the present study and in the above-mentioned study [30] on workers in a thermometer factory. Thus, statistically significant regression coefficients between Se and Hg levels in various biological fluids should be interpreted with caution.

Enzymes No effect of the exposure to mercury was shown on the activities of antioxidative enzymes in erythrocytes. The number of subjects examined was, however, small. The present study permits a 90% chance of detecting a 20% decrease in GSHPX levels in the exposed group as compared with that of the controls (~ = 0.05, one-tailed). Generally, GSHPX is positively correlated with selenium in erythrocytes, provided that Ery-Se does not exceed a certain threshold value [3, 31]. In our subjects a positive correlation was seen only at rather low Ery-Se levels. As Ery-Se levels in the present study were the same in exposed and control subjects, it is not surprising that GSHPX activities were also similar. The result indicates, however, that the amount of Se "available" for the synthesis of GSHPX was not substantially decreased by the exposure to mercury. As there was no difference in catalase activity between the exposed subjects and the controls, it would seem that exposure to mercury does not lead to any "induction" of the enzyme. The reason for this could be that the predominant function of catalase is the disproportionation of hydrogen peroxide. In conclusion, this study indicates that long-term, but moderate, exposure to inorganic mercury could result in a limited decrease in urinary selenium excretion. Blood Se levels and the activities of GSHPX, SOD and catalase, major components in the protection against free radicals, seem to be unaffected. ACKNOWLEDGEMENTS We thank Gerd S~llsten, occupational hygienist, Bengt J~irvholm, M.D., and Professor Staffan Skerfving, M.D., for advice and assistance. The Swedish Society of Medicine and the Swedish National Environmental Protection Agency provided financial support.

46 REFERENCES 1 2

3 4

5 6

7 8 9 10 11 12

13 14

15 16 17 18 19 20 21

22

23

M. Berlin, Mercury, in L. Friberg, G.F. Bordberg and V.B. Vouk (Eds), Handbook on the Toxicology of Metals, Vol. 2, Elsevier, Amsterdam, 2nd edn, Chapt. 16, 1986, pp. 387445. J. HSgberg and J. Alexander, Selenium, in L. Friberg, G.F. Bordberg and V.B. Vouk (Eds), Handbook on the Toxicology of Metals, Vol. 2, Elsevier, Amsterdam, 2nd edn, 1986, pp 482-520. WHO Task Group on Selenium, in O. Levander (Ed.), Environmental Health Criteria 58, Selenium, World Health Organization, Geneva, 1987. H.E. Ganther, C. Goudie, M.L. Sunde, M.J. Kopecky, P. Wagner, S. Oh and W.G. Hoekstra, Selenium: relation to decreased toxicity of methyl mercury added to diets containing tuna, Science, 175 (1972) 1120-1124. J. Parizek and I. Ostadalova, The protective effect of small amounts of selenite in sublimate intoxication, Experientia, 23 (1967) 142-145. L. Magos and M. Webb, Differences in distribution and excretion of selenium and cadmium or mercury after their simultaneous administration subcutaneously in equimolar doses, Arch. Toxicol., 36 (1976) 63-69. J.C. Hansen, Has selenium a beneficial role in human exposure to inorganic mercury? Med. Hypotheses, 25 (1988) 45-53. L. Kosta, A.R. Byrne and V. Zelenko, Correlation between selenium and mercury in man following exposure to inorganic mercury, Nature, 254 (1975) 238-239. J. Alexander, Y. Thomassen and J. Aaseth, Increased urinary excretion of selenium among workers exposed to elemental mercury vapor, J. Appl. Toxicol., 3 (1983) 143-145. T. Hongo, T. Suzuki, S. Himeno and C. Watanabe, Does mercury vapor exposure increase urinary selenium excretion? Ind. Health, 23 (1985) 163-165. W.C. Hawkes, E.C. Wilhelmsson and A.L. Tappel, Abundance and tissue distribution of selenocysteine-containing proteins in the rat, J. Inorg. Biochem., 23 (1985) 77-92. J.T. Rotruck, A.L. Pope, H.E. Ganther, A.B. Swanson, D. Hafenan and W.G. Hoekstra, Selenium, biochemical role as a component of glutathione peroxidase, Science, 179 (1973) 588-590. F. Ursini, M. Maiorino and C. Gregolin, The Selenoenzyme phospholipid hydroperoxide glutathione peroxidase, Biochem. Biophys. Acta, 839 (1985) 62-70. K. Takahashi, N. Avissar, J. Whitin and H. Cohen, Purification and characterization of human plasma glutathione peroxidase: a selenoglycoprotein distinct from the known cellular enzyme, Arch. Biochem. Biophys., 256 (1987) 677-686. S. Marklund, Oxygen toxicity and protective systems, Clin. Toxicol., 23 (1985) 289-298. J.V. Bannister, W.H. Bannister and G. Rotilio, Aspects of the structure, function and applications of superoxide dismutase, CRC Crit. Rev. Biochem., 22 (1987) 1373-1377. L. Magos, S. Halbach and T.W. Clarkson, Role of catalase in the oxidation of mercury vapor, Biochem. Pharmacol., 27 (1978) 1373-1377. O. Einarsson, G. Lindstedt and T. BergstrSm, A computerized automatic apparatus for determination of mercury in biological samples, J. Autom. Chem., 6 (1984) 74-79. I. Skare, Microdetermination of mercury in biological samples. Part III: Automated determination of mercury in urine, fish and blood samples, Analyst, 97 (1972) 148-155. K. Saeed, Y. Thomassen and F.J. Langmyhr, Direct electrothermal atomic absorption spectrometric determination of selenium in serum, Anal. Chim. Acta, 110 (1979) 285-289. M. Ihnat, M.S. Wolynetz, Y. Thomassen and M. Verlinden, Interlaboratory trial on the determination of total selenium in lyophilized human blood serum, Pure Appl. Chem., 58 (1986) 1063-1076. M. Ihnat, Y. Thomassen, M.S. Wolynetz and C. Veillon, Trace element data reliability in clinical chemistry - - interlaboratory trials and reference materials, Acta Pharmacol. Toxicol., 59, Suppl. VII (1986) 566-572. B. Welz, M.S. Wolynetz and M. Verlinden, Interlaboratory trial on the determination of selenium in lyophilized human serum, blood and urine using hydride generation atomic absorption spectrometry, Pure Appl. Chem., 59 (1987) 927-936.

47 24 25 26

27 28 29

30 31

J. Neve, Y. Thomassen and M. Van Damme, Interlaboratory test for the determination of total selenium in lyophilized human whole blood, Pure Appl. Chem., in press. W.A. Giinzler, H. Kremers and L. Flohe, An improved coupled test procedure for glutathione peroxidase in blood, Z. Clin. Chem. Clin. Biochem., 12 (1974) 444-448. S.L. Marklund, A spectrophotometric study of the spontaneous disproportionation of the superoxide anion radical and a sensitive direct assay for superoxide dismutase, J. Biol. Chem., 251 (1976) 7504 7507. S.L. Marklund, Direct assay with potassium superoxide, in R. Greenwald (Ed.), Handbook of Methods for Oxygen Radical Research, CRC Press, Boca Raton, FL, 1986, pp. 249-255. B-G. Svensson, /~. BjSrnham, A. Schiitz, U. Lettevall, A. Nilsson and S. Skerfving, Acid deposition and human exposure to toxic metals, Sci. Total Environ., 67 (1987) 101-115. P. Kristensen and J.C. Hansen, Urinary and fecal excretion of selenium (Na~SSeO3) and mercury (~°3HgC12)administered separately and simultaneously to mice, Toxicology, 16 (1980) 39-47. T. Suzuki, S. Himeno, T. Hongo and C. Watanabe, Mercury-selenium interaction in workers exposed to elemental mercury vapor, J. Appl. Toxicol., 6 (1986) 149-153. H.M. Rea, C.D. Thomson, D.R. Campbell and M.F. Robinson, Relation between erythrocyte selenium concentrations and glutathione peroxidase (EC 1.11.1.9) activities of New Zealand residents and visitors to New Zealand, Br. J. Nutr., 42 (1979) 201-208.