Direct Determination of Selenium and other Trace Elements in Serum Samples by ICP-MS

Journal of

J. Trace Elements Med. BioI. Vol. 12, pp. 240-247 (1998)

Trace Elements in Medicine and Biology

Applied Methodology

© 1999 by Urban & Fischer

Direct Determination of Selenium and other Trace Elements in Serum Samples by ICP-MS R. FORRER l , K. GAUTSCHI*, A. STROH** and H. LUTZ Clinical Laboratory, Dep. of Internal Yet. Med., Univ. of Zurich, Winterthurerstrasse 260, CH-8057 Zurich, Switzerland *Centre for Laboratory Medicine, State Hospital, Aarau, AG 5001, Swizerland **Yarian International, ChollerstraBe 38, 6303 Zug, Switzerland (Received July 1997/April1998)

Summary

Selenium belongs to a group of trace elements of special interest in biological samples for clinical diagnosis. Selenium has antioxidizing functions and is essential for providing the organism with triiodothyronine produced from thyroxine. Among several analytical techniques used to determine the Se concentration in serum, Inductively Coupled Plasma Mass Spectrometry (ICP-MS) has been used in the past because of its high sensitivity. Interference problems originating from different ions on the major Se isotopes have been described to be a limiting factor for the direct determination of Se in these matrices. Standard addition calibration or isotope dilution is often required to overcome carbon-enhanced ionisation effects in biological sample matrices. In most cases, the typical serum sample volume which is available for the analysis is limited to 0.5 ml or less, making multiple sample preparation for standard addition calibration impractical. Isotope dilution requires enriched isotopes and substantial sample preparation. Furthermore, the approximate Se concentration in every sample has to be known to adjust the appropriate amount of spike to each sample. Matrix matching with methanol has been described to overcome ionisation effects but we found limiting factors of this application when other trace elements are also determined within one sample run. This paper describes an effective sample preparation method which allows the direct determination of Se in serum without limiting the analytical capabilities for the additional determination of AI, Cu, Ni, Co, Cd, Mn and Zn in a single sample run by ICP-MS. Optimization procedures are presented and results of the analysis of reference samples are discussed, with a comparison of more than 150 serum data with those obtained by the GFAAS method. Keywords: Selenium, ICP-MS, serum, matrix modification, ICP-MS/GF-AAS Comparison.

Introduction

The determination of trace elements in biological/ clinical samples is gaining increasing importance in the study of chronic diseases (1). Blood and urine are comITo whom correspondence should be addressed

monly used for the direct analysis of trace elements in humans or animals because they are easily available and may represent the health and nutritional status of the total system (2). Selenium acts as an antioxidizing agent through the Se-enzyme glutathioneperoxidase, and catalyzes triiodothyronine production from thyroxine

Determination of selenium and other trace elements in serum by ICP-MS

through deiodease (3). Both selenium deficiency and excess can lead to symptomatic disturbances in humans and animals(4-8). Se is furnished by the human diet oranimal feeds which, in turn, reflect the concentration of selenium in the soil in which the plants are grown. In most areas of Europe, the natural concentration of Se in diet and soils is very low with typical concentrations of less than 0.003 Ilmol/g. and can be insufficient to meet the requirements of both the humans and animals. Serum is commonly used for analysis of the selenium status. Ultra trace level determinations of element concentrations in body fluids require analytical techniques with a high sensitivity. In most cases, only a limited sample volume is available for measurements. Graphite furnace atomic absorption spectroscopy (GF-AAS) has been widely used for trace element determinations in biological samples such as whole blood and serum (9-11). The high sensitivity of this technique is accompanied by several disadvantages. Multi-element analysis in these types of matrices suffers from interference caused by the sample matrix. As the determination is performed directly without matrix separation special temperature programs and background correction methods have to be used for the different elements for accurate determination. Multi-element techniques for trace element determinations in body fluids are therefore limited to inductively coupled plasma atomic emission spectroscopy (ICPAES) or inductively coupled plasma mass spectrometry (ICP-MS). Several applications using ICP-AES have been described in the clinical field (12). There has been marked improvement in the sensitivity of modern axial ICP-AES but detection limits are still not low enough for a direct determination of many trace elements in body fluids. Since the introduction of the first commercially available ICP-MS system in 1983 this technique has also made great progress and is currently used for a wide range of applications. Rapid, quasi-simultaneous, multielement detection capabilities, high sensitivity and detection power have made this technique an important tool for the analysis of trace elements in biological samples. Several papers have dealt with applications of ICP-MS in clinical chemistry (13-19). The determination of trace elements in biological samples requires careful selection of the operating parameters to reduce and'overcome polyatomic interferences, matrix-induced signal suppression and carbon-enhanced ionization effects in the plasma. The sample matrix is very complex, with high salt and protein contents. Stable selenium isotopes (17, 20) and isotope dilution for

241

the deternlination of Se in serum have also been described (21). This procedure requires sample digestion and emiched Se isotopes for spikes, which is time-consuming, expensive and has a limited sample throughput. Lyon et aL (15) and Vanhoe et aL (22) have discussed in detail the potential intelference problems for the determination of Se in body fluids and have made reference to inconsistencies of results for Se in serum. Larsen and StUrup (23) have discussed the ionization enhancement of several elements caused by organic matter contents of the sample matrix. Special sample preparation and optimization procedures are required to avoid interference problems and matrix-induced ionization effects during the determination of Se in biological samples. The organic sample matrix can be mineralized using microwave oven digestion (14) or other ashing techniques (25-27) Lyon et. aL and Sheppard et. al. (15, 27) used gelchromatographic procedures to eliminate the high chloride content in biological samples to overcome Cl-related interferences such as ClO+ and ArCl+. All these procedures require extended sample pretreatment and carry the risk of contamination of the samples during preparation. In many cases, the trace element concentration of the analyzed samples needs to be speedily available to provide a correct basis for medical treatment of the patient. The purpose of this paper is to describe a simple sample preparation method designed to overcome the matrix effects and most of the problems for the direct determination of Se and other trace elements in serum (22), without the need of standard addition or sample digestion. This method keeps the sample volume small so that it can be used in neonatology and with small animals. Part of the procedure includes the optimization of the instrument settings such as nebulizer flow rate and sampling depth.

Material and Methods In.strumentation

The ICP-MS instrumentation used throughout all measurements was a Varian UltraMass™ ICP-MS system (Varian Optical Spectroscopy Instruments, Melbourne Australia). As plasma source for the ICP-MS argon (99.998 % from Carbagas, 3097 Liebefeld; Bern; Switzerland) was used. For the preparation of standards and samples, only Teflon- or PVC- flasks and tubes were used. All operating parameters were under computer con-

242

R. Forrer, K. Gautschi, A. Stroh and H. Lutz

tro1, which allowed simple and fast optimization routines for different matrices. The instrument settings used for all experiments is listed in Table 1. Reagents, Standards, Sample Preparation and Statistics De-ionized 18 Mohm water was prepared by Elgastat Maxima (Elga Ltd. High Street, Lane End, Wycombe, Bucks, England HPl4 3JH) and was used for sample dilution and standard preparation. Acetic acid (Fluka Microselect, Art. # 45726) was used to acidify standards and samples, Merck ICP-MS multi-element standard solution (Merck VI, Darmstadt, Germany) was used for external calibration. Rhodium (Johnson Matthey GmbH, Karlsruhe, Germany) was used as internal standard throughout all measurements. Triton X-lOa (Sigma, Alt. # 90002-93-1) was added as detergent to all measured solutions at a concentration of 0.1 % vivo Freeze dried reference serum, normal and elevated level, from UTAK Laboratories Inc. (Valencia, CA 91355, USA) and Nycomed trace elements in serum (Nycomed Pharma, Oslo, Norway) were used for method validation. A total of 148 serum samples from human and animals were analyzed with both GF-AAS and ICP-MS methods. The samples were collected over a period of 6 weeks and stored at -200 C. All serum samples were prepared as described above in a lO-fold d,ilution in 4% viv acetic acid. TritonX-100 (0.1 % viv) was added to protect the sample introduction system from particle deposition and for better nebulization efficiency and stability of the aerosol production. Blank and standards for the external calibration procedure were prepared in a similar way. Rhodium was added as internal standard to all solutions at a concentration of 0.64 IlmoVl. Four point (blank and 3 standards) calibration was performed in a simple linear mode. The standard concentrations were 0.03, 0.06 and

0.13 Ilmolll for AI, Cd, Cu, Mn, Mo, Ni, Pb and 0.25, 0.64 and 1.27 Ilmolll for Se and Zn. The agreement between the two methods of selenium measurements in clinical samples was evaluetet by applying the statistic procedure described by Bland and Altman (24).

Results and Discussion

Several interferences caused by molecular ions (CIO+, CIOW, ArC!, CCl 2+) have been described (22) to be limiting factors for the direct analysis of Se and other trace elements like Cr and As. The sampling depth can be optimized to achieve minimum interference from polyatomic species (Figure 1). The z-axis position (i.e. sampling depth as distance from the torch to the interface) was varied and the normalized signal intensities for the internal standard element Rh and the polyatomic species ArCl+, ArC+ and CIO+ was monitored. The best signal-to-noise ratio for all parameters investigated was found at a distance of 5 mm from the torch to the interface. The loss of signal intensity at the cost of interference reduction was only marginal. About 10% of the Rh internal standard signal was lost whereby the interference caused by ArCl+, ArC+ and ClO+was greatly reduced (20 to 60 %). Using a sampling depth of 5 mm for the analysis of Se did not result in any significant interference from polyatomic species on Se isotopes at mass 78 and 82. For explanation of the supposed loss of selenium at low levels, a stock standard solution of organically bound selenium of 103 IlmoVI was prepared with selenomethio120 100 ~

';;

5 80

;:

~

>

Table I. Instrument settings Plasmagas flow CVmin) Auxiliary gas flow (l/min) Nebulizer gas flow CVmin) Sample flow rate (ml/min) HF-Power (kW) Nebulizer Type Sampling Depth (mm) Integration time (ms) Scans Scanning mode Total analysis time

60

~

"2 40 15

1.3

...... CIO 20

+ArC

0.88 0.75

1.3 Meinhard Type "C" 6

10 100 peak hopping, I point / peak I sec/isotope, 3 replicates

5.5

6.5 sampling depth (0101)

7.5

8

8.5

Figure 1. Influence of sampling depth (z-axis, Distance torch- sampler cone) on signal and polyatomic species intensities. Data given in percent of maximum value

Determination of selenium and other trace elements in serum by ICP-MS

243

id id

3

J

.1

lVmin]

6

o ani c mpound 1'/\ C{ I Figure 2. Effect of ionization efficiency for Se and Mo signals. Carbon enhanced ionization efficiency matched by acetic acid shoes a maximum at about 4 % v/v acetic acid. The effect is not depending on the concentration of Se up to 3.18 Jlmolll

nine in an isotonic sodium chloride solution at a pH of 9 (ammonia). This sta~1dard was diluted down to 0.07 Ilmol/I and measured with both ICP-MS and GF-AAS methods. Carbon-induced signal enhancement for high e V elements like Se and As has been described in detail by Larsen and Sturup (23). Methanol was used to increase sensitivity for As in speciation work. On the other hand, reduction of polyatomic interference, such as that due to ArCI+, was found.This was explained to come from an increase in CCI+ formation, which competes with the ArCI+ population. The method of standard addition calibration is most commonly used in ICP-MS to overcome matrix-induced signal enhancement or suppression in ICP-MS. This procedure requires a large amount of sample preparation and is, therefore, time-consuming. The

1.1

I

Figure 3. Optimization of nebulizer gas flow settings for nitric acid (l % v/v) and acetic acid (4% v/v)

method of standard addition is impractical for the analysis of samples with limited sample volume, as is usually the case with most clinical sample material. Extemal calibration with matrix matching of the blank and extemal calibration solutions for the organic matter is the method of choice for overcoming matrix- induced effects on signal intensities. The first attempt was made by addition of methanol to the blank and standard solutions. The results obtained were not satisfactory as many additional problems were observed. The major drawback was that the interference of CN+ on Al increased strongly and also background levels for most trace elements were substantially higher because of impurities in the methanol. Acetic acid was found to be an ideal solution which could be used for matrix matching in clinical samples. The acetic acid used was ultrapure grade (Fluka Microselect, Art. # 45726) and was almost free of metal impurities at concentration levels analyzed in this study. Several series of synthetic solution with variable Se and acetic acid content were prepared and analyzed under

.. . •

,

..

•

-c. ~-~ ., • .;~~

c ·.· '" . ~--~---+--~~--+---~---.----~--~--~

•

••

I C

.)

11th (.1·

\ \ 11m".' II

Figure 4. Comparison of selenium data of ICP-MS with the results obtained with GF-AAS in Jlmolll

·11..

·n.

....

II _

·11.1

c.t \ \ I./'

fI.J

O.

o.

•I

Figure 5. Comparison of the logarithmic transfonnation of the ICPMS data with the results obtained with GF-AAS in JlmoVI

244

R.. Forrer, K. Gautschi, A. Stroh and H. Lutz Table 2. Comparison of measured ICP-MS data with expected values of Nycomed Seronoml trace elements control serum (batch 010017) with ranges given in the certificate. Results are shown as IlmolJl. The coefficient of variation (CY) of 5 individual prepared sample aliquots are given.

.5

2

1.5 ' \\I!

hI

\

15

3

3.

Element

ICP-MS X (IlmoUI)

CY( %) (n=5)

Accep. Range Il moUI

Al Cu Fe Se Zn

2.5 17.5 27.5 1.3 23.7

2.6 3.1 3.4 2.5 2.7

2.1-2.5 18.9-19.5 25.3-26.0 1.2-1.3 24.5-26.9

II

the instrument settings described above, to find the optimum concentration of acetic acid to match the organic compounds in the blank and external standard solutions (matrix). The concentration of Se was varied between 0.32 and 3.18/1mol/l, which represents the normal to elevated concentration ranges usually 'expected in human and some animal serum samples, and Se was measured at the masses 78 and 82:'The acetic acid was varied from 0.3 to 7 % vlv for matrix matching. In addition, a mid-eV element like Mo was added to all spiked solutions at a fixed concentration of 0.13 /1m01l1. This allows monitoring unwanted effects on other trace elements caused by the acetic acid. The appropriate blanks for all solutions were prepared with the same amount of acetic acid but without Se spiking. These blanks were analyzed prior to the samples and the intensities obtained were subtracted from the spiked solutions. The results are presented in Figure 2. The normalized data show a distinct maximum of Se sensitivity at an acetic acid concentration of 4% v/v. The Mo sensitivity scatters around the same normalized intensity value for all solutions up to a 5% vlv acetic acid concentration. Further addition of acetic acid caused a loss of

signal intensity for both Mo and Se with the lost being more pronounced for Se. A suggested possible reason for this is the amount of total matrix brought into the plasma and which causes signal suppression by decrease of plasma temperature and/or reduction of ion transmission. No influence of the effects described above, up to 3.18 ,.unol!l, was observed on the concentration of Se in these solutions. The optimum amount of acetic acid for matrix matching was therefore set at 4% v/v. It is well known that the optimum nebulizer flow rate is strongly dependent on the nebulizer type used and the sample matrix aspirated. The most pronounced difference can be seen between aqueous and organic solutions. It was expected that the nebulizer flow rate would have to be optimized for the 4 % acetic acid solution to achieve maximum signal intensity and stability in the serum samples. Signal intensities for aqueous solution and 4% vlv acetic acid solution were monitored for different nebulizer flow rates. The results are shown in Figure 3. The nebulizer gas flow was set at 0.85 IImin for all measurements. The linearity of the analysis was checked with a spiked serum sample with a concentration of 6.35 /1m01l1, which was then diluted down to sub-llmo1/1 levels. The linearity was petfect.

Table 3. Comparison of measured ICP-MS data with expected values ofUtak control serum (normal level, batch 9430) with the ranges given in the certificate. Results are shown as IlmoUl. The coefficient of variation (CY) of 5 individual prepared sample aliquots are given

Table 4. Comparison of measured ICP-MS data with expected values of Utak control serum (elevated level, batch 9429) with raIlges given in the certificate. Results are shown as IlmolJl. The coefficient of variaton (CY) of 5 individual prepared sample aliquots are given

Figure 6. Comparison of the concentration of selenium in selenomethionin with ICP-MS and GF-AAS

Element

ICP-MS X (J..lInoUI)

CY (%) (n = 5)

Accep. Range Il mol/l

Element

ICP-MS X CI..lIll0lJl)

CY(%) (n=5)

Accep. Range IlmolJl

Al Cd Cu Mn Se Zn

0.7 0.002 15.7 0.009 1.5 13.2

7.3 42.6 5.9 20.2 1.6 6.0

0.5-0.8 0-0.002 14.2-18.1 0.013-0.016 1.3-1.6 7.7-10.7

Al Cd Cu Mn Se Zn

7.2 0.07 46.0 0.06 4.1 40.8

2.9 2.3 3.2 3.0 0.8 \.8

6.7-8.5 0.07-0.09 40.9-50.4 0.05-0.06 3.6-4.6 35.2-44.4

Detennination of selenium and other trace elements in semm by ICP-MS

245

I

7

.= 0.4

•

•

t

••

I~.ln

•

•

-

'J

i n

~ ·0.2

.In

'J

• e

• .6

• ••

0.4 \\~f3 '~

~ ·0.

• n 'In

2.4

3.4 .4 ~ k1) .1 \\ ;and 1< P 1 11m III

1.4

..,

~ 0.2 '" 0

1

~

= ;;!

S

-0.

.

" -0

~

--r.r---------

m.:an -

• •

• ·OJ \".

~

.2 (01 .\

0.1 II I' \,

Figure 7. Comparison of the differences between the two methods against their averages in Ilmol/l

Figure 8. Logarithmic transfonnation of the differences between the ICP-MS and the GF-AAS and their averages in Ilmolll

Validation (accuracy) of the method was carried out by means of analysis of a control serum solution from UTAK (normal and elevated level) and Nycomed Seronorm. The results obtained are summarized in Tab. 2 to 4 and show very close agreement with the data expected for these materials. The data presented in the tables (2-4) represent the mean values of 5 individually prepared sample aliquots from different freeze dried material. In addition, a comparison of the measurements of up to 148 individual serum samples from animals (cat, dog, horse, cow, sheep and goat) and from humans was carried out by ICP-MS and GF-AAS. Figure 4 shows the comparison of the selenium results from both ICP-MS and GF-AAS techniques. It is obvious that for low concentrations, <0.25 /lmol/l, the ICP-MS results tend to be on the higher side. This is better demonstrated in the logarithmic transfomlation of the compared results (Figure 5). It is well known that organic material is burned away during the ashing step in GFAAS. This suggests that a possible mechanism for this slight elevation with the ICP-MS-method might be the loss of organically- bound Se during the ashing steps of the GF-AAS temperature pre-treatment program used for these determinations.

Different concentrations of selenium in a selenomethionine solution were measured with the two methods, to clarify the role of this mechanism as the cause for the elevation of the results by ICP-MS (see Figure 6). For selenium concentrations lower then 0.12 /lmolll the ICP-MS method yielded a well-defined count number or concentration, while the GF-AAS method showed no measurable absorption. The coefficient of correlation of 0.96 shows a very good correlation between the two analytical techniques. But the coefficient of correlation does not garantee the absence of a systematic deviation; this has to be evaluated in a separate step. To determine the degree of agreement the differences between the values obtained by the GFAAS- and the ICP-MS-methods for each sample were plotted against the average of their means according to the procedure of Bland and Altman (24). The mean difference is -0.06 /lmolll with a 95 % confidence interval of -0.04 to -0.09 /lmolii. As shown in Figure 7 the limits of agreement for the two compared methods were -0.40 to 0.27 /lmol/l .These limits were small enough to confirm that the new method can be used for clinical interpretation of the selenium results. The 95 % confidence interval for the lower limit is 0.44 to -0.35 /lmol selenium 11. For the upper limit, the interval is 0.22 to 0.32/lmol selenium II. Using logarithmic transformation of these data (Figure 8), stronger deviations for the two methods at the low levels can be better seen. The precision of the method described above has been tested with a horse serum at a concentration of 1.87 /lmol/l and was analyzed 26 times from individually prepared aliquots. The standard deviation was found to be 0.03 /lmol/l (CV =0.016), which can be considered to be excellent for this type of matrix.

.~

Table 5. Lower limit of detection (LLD) in Ilmol/l for the trace element detennination in 4% v/v acetic acid. Calculation was perfonned with the following equation: the 3 standard deviations of the counts of the blank divided by the number of counts of the highest standard Element

LLD Il molll

Element

LLD Ilmolll

Al Zn Cd Cu Fe

0.0019 0.0015 0.000045 0.00016 0.027

Mn Mo Ni Pb Se

0.00036 0.000021 0.00017 0.000024 0.0013

246

R. Forrer, K. Gautschi, A. Stroh and H. Lutz

The detection limits listed in Table 5 have been caIculated using 3 standard deviations of the counts of the blank divided by the number of counts of highest standard,

Burkhardt Seifert from the Institute for Social" and Preventive Medicine (Biostatistics) from the University of Zurich for his careful revision of the statistics.

References Conclusions

Multi-element determinations in clinical samples such as serum can be performed directly and rapidly with ICP-MS. Careful sample preparation methods have to be established especially for Se, AI, and Zn if external calibration procedures are to be used. Matrix matching is essential when dealing with high e V elements such as Se and As in biological samples or any sample matrix containing significant amounts of organic matter. Acetic acid has proven to be advantageous for this purpose. Optimization of plasma parameters leads to reduced interference from polyatomic species. Nebulizer flow rate settings are different from those obtained with conventional aqueous solution settings. The achieved accuracy and precision for the method described allow routine analysis of these matrices with the required performance characteristics. Detection limits achieved by the multi-element ICP-MS technique are superior to those obtained by GF-AAS. It has been observed that during continuous sample analysis, the sample introduction system (nebulizer, spray chamber, injector, cones) has to be cleaned frequently to achieve stability of the analytical results. The torch, especially, must be cleaned more frequently depending on the number of samples analyzed (once after about 25-30 semm samples). Hence, the speed and multi-element character of the method described speakes for use ofICP-MS for this application, as both detection limits as well as sample throughput are far better as compared to any other technique, including GF-AAS and ICP-AES.

Acknowledgements

The authors thank Prof Dr. Walter Mertz, who was Director of the Human Nutrition Research Center of Beltsville (USA) till 1986, for reading the manusclipt and his valuable advice. We would also like to thank Prof. Dr. D. J. Vonderschmitt, Director, Department of Clinical Chemistry, University Hospital of Zurich for help with the acquisition, financial support and maintenance of the ICP-MS equipment, Mr. A. Ogilvie from the same department for his support and his reading of this paper, Dr.

I. SCHMIDING, T. ,SCHRAM, 1., POHL, B., BOMBELKA, R & SCHAD, H. (1996) Spurenanalytik von Elementen in Humanfllissigkeiten mittels Festphasenextraktion und Detektion durch ICP-OES. Fachzeitschrift fUr das Laboratorium, 40, 197-203, GIT Verlag GMBH 2. BRATTER P., FORTH w., FRESENIUS w., HOLTMEIER H. J., HOYER S., KRUSE-JARRES J.D., LIESE H., MOHN L., NEGRETTI DE BRATTER V.E., REICHLMAYR-LAIS A.M .. SITZER G., TOLG G. (1992) Mineralstoffe und Spurenelemente - Leitfaden fUr die arztliche Praxis, Verlag Berrtelsmann , Stiftung, Glitersloh 3. ROTRUCK, J.T., POPE. A.L., GANTHER, H.A., SWANSON, A.B., HAFEMAN, D.G. & HOEKSTRA, w.G. (1973) Selenium: biochemical role as a component of glutathione peroxidase, Science 179. 588-590 4. SCHRAUZER, G.N., WHITE, D.A. & SCHNEIDER, C.J. (1977) Cancer mortality correlation studies. III: Statistical associations with dietary selenium intakes. Bio Inorg. Chern. 7, 2334 5. JOHNSON, R.A., BAKER, S.S., FALLON, J.T., MAYNARD, E.P., RUSKIN, H.N., WEN, Z., GE, K. & COHEN, H.J. (1981) An occidental case of cardiomyopathy and selenium deficiency. N. Engl J. Med. 304,1210-1212 6. LEVANDER, O.A. (1982) Clinical consequences of low selenium intake and its relationship to vitamin E. Ann. N.Y. Acad. Sci. 393,70-82 7. SALONEN, J.T., ALFfHAN, G., HUTUNEN, J.K. & PUSKA, P. (1984) Association between serum selenium and the risk of cancer. Am. J. Epidemiol. 120, 342-349 8. MERTZ, w., MORRIS, E.R., SMITH, JR. J.C., UDOMKESSMALEE, E., FIELDS, M., LEVANDER, O.A. & ANDERSON, R.A. (1989) Trace Elements in Elderly Metabolism, Requirements, and Recommendations for Intakes. From: Nutrition, aging, and the elderly, Edited by Hamish N. Munro and Darla E. Danford (Plenum Publihing Corporation), 195-244 9. FORRER, R., GAUTSCHI, K. & LUTZ, H. (1991) Comparative Determination of Selenium in the Serum of Various Animal Species and Humans by means of electrothermal atomic absorption spectrometry. J. Trace Elem. Electrolytes Health Dis., 5, 101-113 10, VAN CAUWENBERGH, R., ROBBERECHT, H. & DEELSTRA H. (1990) Selenium Concentration Levels in Whole Blood of Belgian Blood Bank Donors, as Determined by Direct Graphite Furnance Atomic Absorption Spectrometry. 1. Trace Elem. Electrolytes Heath Dis. 4, 215-224 11. V AN CAUWENBERGH, R., ROBBERECHT, H. & DEELSTRA H. (1990) Direct Selenium Determination in Human Whole Blood by Graphite Furnance Atomic Absorption Spec-

Determination of selenium and other trace elements in serum by ICP-MS trometry with Deuterium Correction Using a L'vovplatform, a PdlMg Matrix Modification and Appropriate Dilution. 1. Trace Elem.,Electrolytes Health Dis., 4,127-131

247

20. VEILLON, C.( 1984) GCIMS Measurement of stable isotops of selenium for use in metabolic tracer studies. Reprint from American Chemical Symposium Series., Nr. 258 Stable Isotops in Nutrition, Turnlund, J.R. & Johnson, P.E. Editors

12. LYON, T.D.B., FELL, G.S., HUTTON, R.C. & EATON, A.N. (1988) Evaluation of inductiively Coupled Argon Plasma Mass Spectrometry (lCP-MS) for Simultaneous Multi-Element Trace Analysis in Clinical Chemistry. 1. Anal. At. Spectrom., 3, 265271

21. REAMER, D.C. & VEILLON, C. (1983) A Double Isotope Dilution Method for Using Stable Selenium Isotopes in Metabolic Tracer Studies: Analysis by Gas Chromatography/Mass Spectrometry (GC/MS). J. Nutr. Jl3, 786-792

13. YOSHINAGA,l., NAKAZAWA, M., SUZUKI, T. & MORITA, M. (1989) (1989) Determination of Trace Elements in Human Liver and Kidney by Inductively Coupled Plasma Mass Spectrometry. Analytical Science,S, 355-358

22. VANHOE, H., GOOSSENS, J., MOENS, L. & DAMS, R. (1994) Spectral Interferences Encountered in the Analysis of Biological Materials by Inductively Coupled Plasma Mass Spectrometry. J. Anal. At. Spectrometry 9, 177-185

14. FRIEL, J.K., CRAIG, S.K., JACHSON, S.E. & LONGERIC1H, H.P. (1990) Analysis of Biological Reference Material, Prepared by Microwave Dissolution, Using Inductively Coupled Plasma Mass Spectrometry, Analyst, 115,269-273

23. LARSEN, E.H. & STURUP, S. (1994) Carbon-enhanced Inductively Coupled Plasma Mass Spectrometric Detennination of Arsenic, Selenium and Its Application to Arsenic Speciation. J. Anal. At. Spectrometry, 9, 1099-1105

15. LYON, T.D.B., FELL, G.S., HUTTON, R.C. & EATON, A.N. (1988) Elimination of Chloride Interference on the Determination of Selenium in Serum by Inductively Coupled Plasma Mass Spectrometry. J. Anal. At. Spectrom., 3, 601-603

24. BLAND, J. M. & ALTMAN, D.G. (1986) Statistical methods for assessing agreement between two methods of clinical measurement. The Lancet, Vol. I, February 8, 307-310

16. MULLIGAN, K.J., DAVIDSPON, T.M. & CARUSO, lA. (1990) Feasibility of the Direct Analysis of Urine by Inductively Coupled Argon Plasma Mass Spectrometry for Biological Monitoring of Exposure to Metals,. J. Anal. At. Spectrom., 5, 303-306 17. TING, B.G., MOOERS, CH. S. & JANGHORBANI, M. (1989) Isotopic Detenmnation of Selenium in Biological Materials With Inductively Coupled Plasma Mass Spectrometry, Analyst, 114,667-674 18. BEAUCHEMIN, D., MCLAREN, J.w. & BERMAN, S.S. (1988) Useof External Calibration for the Determination of Trace Metals in Biological Materials by Inductively Coupled Plasma Mass Spectrometry. J. Anal. At. Spectrom.,3,775-780 19. VANHOE, H., VANDECASTEELE, VERSIEK, J. & DAMS R. (1989) Detennination of Iron, Cobalt, Copper, Zinc, Rubidium, Molybdenum, and Cesium in Human Semm by Inductively Coupled Plasma Mass Spectrometry. Anal. Chern. 61, 18511857

25. RIDOUT, P.S., JONES, H.R. & WILLIAMS, J.G. (1985) Determination of Trace Elements in a Marine Reference Material of Lobster Hepatopancreas (TORT-I) Using Inductively Coupled Plasma Mass Spectrometr. Analyst Jl3, 1383-1988 26. GRAY, A.L. (1985)The ICP as an ion source -origins, achivements and prospets. Spectrochim. Acta. Part B, 40, 1525-1537 27. SHEPPARD, B. S., SHEN, W-L., CARUSO, J.A., HEITKEMPER, D.T. & FRICKE FL. (1990) Elimination of the Argon Chloride Interference on Arsenic Speciation in Inductively Coupled Plasma Mass Spectrometry Using Ion Chromatography. J. Anal. At. Spectrom., 5, 431-435