paraproteinemia, and myelomatosis, among others. In practice, this would seldom present any clinical problem. The NOVA! and the supplies for it used in these ...
mens (Figure 1). However, clinically
important
differences
between results by the two techniques appear only in very rare situations, such as hyperlipidemia, paraproteinemia, and myelomatosis, among others. In practice, this would seldom
present any clinical problem.
The NOVA! and the supplies for it used in these studies were a loan NOVA Biomedical, Newton, MA 02164. We thank Ole Faergeman, M.D., Medical Department B, Rigshospitalet, for laboratory assistance with lipemic plasma samples.
of
References 1. Ladenson, J. H., Evaluation of an instrument (Nova-i) for direct potentiometric analysis of sodium and potassium in blood and their indirect potentiometric determination in urine. Clin. Chem. 25, 757-763 (1979).
Ladenson, J. H., Direct potentiometric analysis of sodium and potassium in human plasma: Evidence for electrolyte interaction with a nonprotein, protein-associated substance(s). J. Lab. Clin. Med. 90, 654-665 (1977). 3. Dahms, H., Rock, R., and Seligson, D., Ionic activities of sodium, potassium, and chloride in human serum. Clin. Chem. 14, 859-870 (1968). 4. Mohan, M. S., Bates, R. G., Hiller, J. M., and Brand, M. J., Measurements of sodium in albumin solutions with ion-selective electrodes. Clin. Chem. 24,580-584 (1978). 5. Friedman, S. M., Wong, S.-L., and Walton, J. H., Glass electrode measurements of blood sodium and potassium in man. J. Appi. Physiol. 18,950-954(1963). 6. Moore, E. W., and Wilson, D. W., The determination of sodium in body fluids by the glass electrode. J. Clin. Invest. 42, 293-304 (1963). 7. Friedman, S. M., and Nakashima, M., Single sample analysis with thesodium electrode. Anal. Biochem. 2,568-575 (1961). 2.
CLIN. CHEM.28/1, 172-174 (1982)
Fluorometryof Seleniumin Serum or Urine L. Lalonde, V. Jean, K. D. Roberts, A. Chapdelaine, and C. Bleau procedure for determining selenium in human serum or urine is sensitive (requiring only 0.4 mL of sample), accurate, simple, and can be performed on several samples concurrently. Using this technique, we found a mean selenium concentration in the serum of normal Canadian men of 142.9 (SD 16.1) tg/L. The mean urinary excretion rate was 124.5 (SD 76.0) tg/day. This fluorometric
AddftlonalKeyphrasos: trace elements
.
reference interval
methods for the small laboratory The toxic effects of selenium are well described, particularly in livestock, where acute and chronic poisoning have been reported (1). Selenium is also recognized as an essential micronutrient, and this element can protect rats against experimental liver necrosis (2). Since this early observation, several pathological conditions in animals have been shown to result from selenium deficiency (3). It has also been suggested that a selenium deficiency could be associated with human diseases such as kwashiorkor (4,5), sudden-death syndrome in infants (6), cancer (7), and multiple sclerosis (8). This prompted us to develop a rapid, reliable procedure for measuring trace amounts of this element in human body fluids. Fluorometry offers many advantages because it can be adapted to the assay of nanogram amounts with a minimum of labor and equipment. Here we present a fluorometric technique that allows measurement of selenium in many samples with the required sensitivity and accuracy. Department of Obstetrics and Gynecology, University of Montreal; Maisonneuve-Rosemont Research Center, 5415 Boulevard de l’Assomption, MontrSal, Qu#{233}bec, Canada, HiT 2M4. Address correspondence to G.B., at the Center. Received April 10, 1981; accepted Oct. 17, 1981.
and
172
CLINICAL CHEMISTRY, Vol. 28, No. 1, 1982
Materials and Methods Apparatus We digested samples in disposable 18 X 150 mm borosilicate glass tubes, heated in a sand bath. This bath wascustom-made
from a hot plate (Lindberg Co., Watertown WI 53094; Model 53014; 2800 W, 240 V) around which a metal border, 7.5 cm in height, was soldered directly to the plate; into this container was placed a 2.5-cm layer of fine sand. Fluorescence was measured with an Aminco Bowman spectrophotofluorometer equipped with a xenon lamp.
Reagents All aqueous solutions were prepared in doubly distilled, doubly demineralized water. High-quality nitric acid (Ultrex, J.T. Baker) was used for preparing the digestion mixture. Diaminonaphthalenedihydrochloride(DAN) was purchased from AldrichChemical Co. and solutions were prepared immediately before use. DAN (99% pure) did not require further purification because blank values were consistently lower than 5 arbitrary units of relative fluorescence intensity, i.e., equal to cyclohexane itself.
Standard Solutions Standard A waspreparedby dissolvingin HC1 (0.1 mol/L) known amounts of selenium oxide (Puratronic; Johnson Matthey ChemicalsLimited); standard B was purchased as selenous acid solution (Ventron Alfa Division); standard C was purchased as a selenium oxide reference solution. All other reagents were of “Fisher Certified” grade.
Procedure Digestion of samples. To 0.4 mL of standard solutions or urine or serum, add 1.0 mL of an equivolume mixture of HNO3
and HC1O4.This volume
of acid is more than sufficient
to
allow complete digestion. Slowly heat for 30 mm at 150 #{176}C, Table then increase the temperature to 190 #{176}C and maintain it for
2 h. Increase the temperature to 210 #{176}C and maintain it for 1 h. Then cool the tubes. Add 0.2 mL of 6 molfL HC1 and heat at 150 #{176}C for 5 mm. Repeat this step if fumes of NO2 are detected (reddish vapors). Cool the tubes. Reduction of selenate to selenite. Add 2 mL of a solution containing, per liter, 20 mmol of ethylenediaminetetraacetate, 10 mg of bromcresol purple, and 7 mol of NH.OH. Heat at 140 #{176}C until a distinct yellow color is observed (pH 1 to 2). Cool the tubes. Add 5 mL of HC1 (0.1 mol/L) and adjust the volume to 10 mL with water. Let stand overnight at room temperature. This interval allows complete reduction and is convenient. Formation of the piazselenol light, add 0.5 mL of DAN (4 g/L
Under
complex.
subdued solution in 0.1 mol/L HC1).
Incubate for 30 mm in a water bath at 40 #{176}C. Extract the complexwith 5 mL of cyclohexane,and measurethe fluorescenceof the extract with excitationwavelengthsetat 360 nm and emission wavelength at 520 nm.
a
Std. B
Std. C
fluorsscence Int.n.Ity
0 100 200 300 400
5.16 ± 076b 32.33 ± 0.57 59.33 ± 0.57 87.33 ± 1.52 119.33±1.15
5.16 ± 0.76 34.00 ± 1.00 64.00 92.33 ± 2.08 123.00
S
0.28
0.29
5.16 ± 18.00 30.83 ± 43.33 ± 56.33±
0.76 1.04 0.57 1.15
0.12
Arbitraiy unfts. #{176} Standarddeviation. C Slope ofthecurve.
Analytical-recovery studies demonstrated that a mean of 99.6% of the selenium added (50 or 100 ig/L) to a control pooled serum was accounted for after digestion. The mean within-assay variation (CV) of this technique was 2.93% (SD 1.45%); the between-assay variation was 2.67% (SD 2.16%).
Several techniques for measuring selenium have been decurve of fluorescence
concentration (Figure 1) is linear (coefficient of correlation: 0.999) up to a concentration of 400 igfL. We testedthree different selenium standard solutions (Table 1). All three gave this linear relation, with high correlation coefficients, but the slopes (S) were different (SA = 0.28; SB = 0.29; Sc = 0.12). Standard C, purchased as a prepared solution, was rejected because it differed significantly from standard A, which we prepared in our laboratory from pure solid selenium oxide. Standard B was satisfactory because its selenium concentration was accurate. The mean concentration of selenium in the serum of 15 normal men was 142.9 (SD 16.1) ig/L. Selenium was also measured in the 24-h urine samples obtained from 10 men on three consecutive days. The mean daily urinary excretion was 124.5 (SD 76.0) ag/day. tLAT)VE
Std. A R.IatIve
S., gig/L
DiscussIon
Results The standard
1. RelatIve Fluorescence Intensity of Three Different Standard Preparations
vs selenium
tUORI$CNCE
INTINSITY AISITSAIT UNIT)
80.
80.
scribed. However, these techniques could not be applied in a clinical laboratory where many samples had to be assayed on a routine basis. The fluorometric technique described here is an adaptation of the method of Watkinson (10). It is simple, sensitive, reproducible, precise, and requires relatively small sample volumes. Thus, the selenium content of urine can effectively be measured in 0.4mL of samplewith this technique, as compared with 10 mL required with neutron activation analysis (9). Moreover, many samples can be digested concurrently,
a characteristic
shared only by the semi-automated
fluorometric method (10). The digestion unit is readily made, and the technique requires little instrumentation. The simplicity of the procedure renders this technique accessible to many
laboratories
where sophisticated
apparatus
is not
Blank values are very low, representing about 5 arbitrary units of relative fluorescence intensity. The accuracy of this method, as reflected in the high recoveries, compares well with that of any other technique. All steps in the procedure should be considered critical. The digestion must be carefully controlled to avoid changes of temperature by more than approximately 10 #{176}C. Also, as pointed out by Watkinson (10), all residual nitric acid must be removed if reduction of selenate to selenite is to be complete. The pH adjustment with bromcresol purple as indicator allows rapid adjustment of acidity in the range of pH 1 to 2. Levesque and Vendette (11) have commented upon the importanceofa proper system toreduceselenatetoselenite and to provide suitable conditions for the formationofthe selenium complex. The presentsystem, which allowsan overnight reduction step in hydrochloric acid, fulfills this available.
prerequisite.
It is noteworthy that selenium reference standards can vary
40
considerablyfrom onesupplier to another. Also, it is advisable r
to have periodic quality-control assays of these standard solutions and compare the results with those for freshly prepared solutions of pure solid selenium oxide. The mean value for selenium in serum, 142.9 tg/L, that we obtained for Canadian men with this method corresponds well with reported values. For example, Dickson and Tomlinson (12) reported a mean plasma value of 144 igfL in 254 normal
.0.999
20
individuals
o 200
) 5ELINIUM
30 ‘
FIg. 1. RepresentatIve standard curve
L)
400
in a Canadian
population.
The selenium in 24-h human urine specimens was reported to be 79.3 (SD 38.7) sgfL, with a range of 21.5 to 203.0 ig/L (13), but the daily excretion in micrograms was not reported. For comparison purposes, one needs such data. In our study this value is 95.5 (SD 47.2) g/L,
range 29.1-198
4ugfL. This
CLINICAL CHEMISTRY, Vol. 28, No. 1, 1982
173
falls within therangereported by Valentine et al., who used atomic absorption spectrophotometry (13). Othershave reported values that differ somewhatfrom ours: 25 g/L (9); 42 (SD 18) zg/L, range0-69 (14); and 34 (SD 24) g/L, range 0-150(15). However, the interindividual variations are great, probably reflecting dietary intake of the element or individual
environmental exposure. Our simplified technique measurement
of selenium
represents a valuable tool for in body fluids with precision,ac-
curacy, and reproducibility.Usingthis technique,we recently reported on the percutaneous absorption of selenium from a commercialshampoothat contains selenium suifide (16).
References 1. Wilber, C. G., Toxicology of selenium: A review. Clin. Toxicol. 17, 171-230 (1980). 2. Schwarz, K., and Foltz, C. M., Selenium as an integral part of factor 3 against dietary necrotic liver degeneration. J. Am. Chem. Soc. 79, 3292-3293 (1957). 3. Schwarz, K., Essentiality and metabolic functions of selenium. Med. Clin. North Am. 60, 745-758 (1976). 4. Schwarz, K., Selenium and kwashiorkor. Lancet i, 1335-1336 (1965). 5. Burk, R. F., Jr., Pearson, W. N., Wood, R. P., and Viteki, F., Blood selenium levels and in vitro red blood cell uptake of 75Se in kwashiorkor. Am. J. Clin. Nutr. 20,723-733 (1967). 6. Money, D. F. L., Vitamin E and selenium deficiencies and their
CLIN.
CI#{128}M. 28/1, 174-177 (1982)
possible aetiological roleinthesudden death ininfantssyndrome. N.Z. Med. J. 71, 32-34 (1970). 7. Shamberger, R. J., and Willis, C. E., Selenium distribution and human cancer mortality. CRC Crit. Rev. Clin. Lab. Sd. 2,211-221 (1971).
8. Wikstrom, J. T., and Westermarck, P. J., Selenium, vitamin E and copper in multiple sclerosis. Acta Neurol. Scand. 54, 287-290 (1976). 9. Weingarten,R.,Shamai,Y.,and Schlesinger, T., Determination of seleniumin urineby neutron activation analysis. mt. J. APP!. Radiat. Isot. 30, 585-587 (1979). 10.Watkinson,J.H.,Semi-automatedfluorimetric determination of nanogram quantities of seleniumin biological material. Anal. Chim. Acta 105,319-325(1979). 11. Levesque, M., and Vendette, E. D., Selenium determination in soil and plant materials. Can. J. Soil Sci. 51, 85-93 (1971). 12.Dickson,R. C.,and Tomlinson, R. H., Selenium in blood and human tissues. Clin. Chim. Acta 16,311-321(1967). 13. Valentine, J.L.,Kang, H. K.,and Spivey,G. H.,Seleniumlevels inhuman blood,urine, and hair inresponse to exposure via drinking water. Environ. Res. 17, 347-355 (1978). 14.Sterner, J.H.,and Lidfeldt, V., The seleniumcontentof“normal urine.”J. Pharmacol. Exp. Ther. 73,205-211(1941). 15.Glover,J. R., Selenium in human urine: A tentative maximum allowable concentration for industrial and rural populations. Ann. Occup. Hyg. 10,3-14 (1967). 16. Jean, Y., Rochefort, J. G., Lanct#{244}t, C., et al., Plasma levels and urinaryexcretion ofseleniumafterapplication ofa 1% seleniumsalfideshampoo to patientswith seborrheicdermatitis. Selenium in Biology and Medicine, J. E. Spallholz, J. L. Martin, and H. E. Gasther, Eds., AVI PubI. Co., Westport, CT, 1981, pp 422-426.
-
RadioimmunoassayandChemicalIonization/MassSpectrometryCompared for PlasmaCortisol Determination Claes Lindberg,’ Sven JOnsson,1Pavo Hedner,2 and Asa Gustafsson2 We describe a method for determination of cortisol in plasma and urine, based on chemical ionization/mass spectrometry with deuterium-labeled cortisol as the internal standard. The within-run precision (CV) was 2.55.7%, the between-run precision 4.6%. Results by this method were compared with those by a radioimmunological method (RIANEN Cortisol, New England Nuclear)
for
395 plasma samples. The latter method gave significantly higher (approx. 25%) cortisol values. Because of its speed and the sensitivity with which a large of samples can be analyzed, radioimmunoassay (RIA) is commonly used in routine laboratory work for measuring cortisol in plasma. As compared with other methods, such as number
fluorometry, RIA also has a relatively high selectivity; however, there is some cross reaction with other steroids-e.g., cortisol metabolitesand syntheticcorticosteroids, which may become
important
in the diseased
state and during
drug
(1 ).3 Gas chromatography/mass spectrometry (GCMS) withselected-ion monitoringisrecognized as one of the most selective analytical techniques available today and has been suggested as a “definitive” method in clinical chemistry (2). Bj#{246}rkhem et al. (3) described a GCMS method with electron-impact ionization for the determination of plasma cortisol. Cortisol labeled with 14Cwas originally used treatment
as internal standard, but has recently been replaced with a deuterium-labeled analog (4). We present here a modified method based on chemical-ionization (CI) GCMS, which we used to measure cortisol in plasma and urine during an investigation of the systemic effects of glucocorticoid ointments. We used RIA to rapidly check if plasma cortisol suppression was severe, in which case the steroid treatment was discontinued. For the fmal evaluation of cortisol concentrations, the same plasma samples were analyzed at a later date by the more laborious GCMS method. The data compared are those for 395 plasma samples from 16 healthy men.
Materials and Methods 1 Pharmacokinetic Laboratory, AB Draco (subsidiary of AB Astra, Sweden), Box 1707, S-221 01 Lund, Sweden. 2 Department of Medicine, University Hospital, S-221 85 Lund, Sweden. Received Aug. 14, 1981; accepted Oct. 5, 1981.
174
CLINICAL CHEMISTRY.
Vol. 28, No. 1, 1982
Subjects and blood collection. Nonstandard
abbreviations:
Sixteen
healthy men gave
RIA, radioimmunoassay;
GC, gas
chromatography; MS, mass spectrometry; CI, chemical ionization.
