benzidine. Because there are relatively few methods for determining selenium in urine, there are few published results. In 1936, after testing six urine samples ...
CLIN.CHEM, 26/9,1272-1274(1980)
Fluorometryof Seleniumin Urine Azlz Geahchan’ and Paul Chambon We describe a fluorometric method for estimating selenium in urine. A piazselenol is formed between selenium and 2,3-diaminonaphthalene, extracted withcyclohexane, and selenium fluorometrically determined in the extract (Xex, 377 nm; Xem, 519 nm). Analytical recovery after di-
gestionwithnitric-perchloric acidmixturewas estimated forinorganicselenium and itsorganicurinarymetabolites. Within-dayprecisionwas 6.3% ( = 10 jzg/L, n = 6), 5% ( = 20 zg/L, n = 6), 2.6% ( = 40 jg/L, n = 6), 2.2% ( = 100 g/L, n = 6),1.8% ( = 200 tgIL, n = 6), and 1.2% ( = 400 tg/L, n = 6). Day-to-day precision was 11.2% ( = 10 ug/L, n = 12),4.4% ( = 40 tg/L, n = 12),and 3.2% ( = 400 .tg/L, n = 12).The detectionlimit is0.39 .tg/L.We determined the “normal” excretionof
seleniuminrelation tosome biological factors. AddItIonal Keyphrases:reference intervals #{149} toxicology occupational hazards
trace elements
seleniferous areas of Wyoming, South Dakota, and Nebraska. The range for selenium was 20 to 990 tg/L; 7.9% of the samples contained no detectable selenium, and only 3.1% contained more than 100 fig/L. Sterner and Lidfelt (6) gave results of 19 estimations of urinary selenium. The results ranged from 2 to 70 zgfL. All samples contained selenium, but none more than 100 ig/L. On the most recent report, Glover (1) published results obtained between 1956 and 1958 for 793 urinary selenium estimations in people presenting themselves for a pre-employment examination. A mean of 34 zgfL was reported, and a maximum allowable concentration of 100 tg/L for selenium in urine was recommended, both for individuals occupationally exposed to selenium and for a rural population. We describe a rapid, sensitive wet-digestion method involving a semi-closed condensation system that avoids losses of selenium, and show that selenium is measurable in the urine of a normal French population, but in very low amounts.
.
Although many workers have attempted to estimate selenium in urine, there are few published assays for submicrogram amounts of total selenium because of lack of sensitivity and of losses during digestion or charring procedures. Glover (1) used a sulfuric/nitric acid mixture with mercuric
oxide added to prevent
the “escape of volatile selenium
and
its compounds.” After distillation with HBr/Br2 solution, selenium, isolated as a red precipitate formed by adding SO2, is separated by filtration and estimated by iodine/thiosulfate titration. The detection limit is 20 ig/L, but 300 mL of urine is necessary for the titration. This method is unsuited for use in the routine laboratory because of the time involved in digestion, distillation, and precipitation of selenium before titrimetry. Taussky et al. (2) used the Shoniger flask for the combus-
tion of organic materials. urine was not determined.
Analytical recovery of selenium in Points deserving attention are the
retention of selenium in the ash in the case of samples with a high inorganic content, and the traces of fluorescent compounds produced during combustion. Roquebert and Truhaut (3) used the dry-ashing method in an open system with a nitric acid/magnesium nitrate digestion mixture that results in high losses of selenium by volatilization. After distillation with HBr/Br2 solution, selenium was determined colorimetrically with 3,3’-diamino-
benzidine. Because there are relatively few methods for determining selenium in urine, there are few published results. In 1936, after testing six urine samples from unexposed individuals, Dudley (4) concluded that “normal human urine shows no detectable quantities of selenium.” One year later, Smith and Westfall (5) published results of 127 urinary selenium estimations on non-occupationally exposed subjects living in the Laboratoire de Toxicologie et Hygiene Industrielle, Facult#{233} de Pharmacie, 8 avenue Rockefeller, 69373 Lyon Cedex 2, France. Present address (and address for correspondence): Institut Francais de Recherches et Essais Biologiques, Domaine des Oncins, 69210 Saint-Germain-sur-l’Arbresle, France. Received April 19, 1979; accepted April 22, 1980.
1272 CLINICALCHEMISTRY,Vol.26,No.9,1980
Materials and Methods Apparatus A Jobin Yvon JY3 (Jobin Yvon I.S.A., Longjumeau, France) fluorescence spectrophotometer equipped with holographic concave grating and a 150-W stabilized xenon source was used for the measurements. We also used microKjeldahl-type digestion flasks with a bowl capacity of 50 mL; Teflon aciddigestion bombs (Parr Instruments Co., Moline, IL 61265); a water bath, at 50 #{176}C and at boiling; and a pH meter with
general-purpose Paris,
glass electrode
(Model PW 9408; Philips,
France).
Reagents All reagents used were of analytical grade; doubly distilled water was used throughout. A stock 0.1 g/L solution of selenium was prepared by dissolving 100 mg of metallic selenium in 20 mL of nitric acid. This was boiled gently until brown fumes no longer appeared, cooled, and diluted to 100 mL with 0.1 mol/L hydrochloric
acid. Working-standard selenium solutions, 10-1000 ng/L, were prepared by diluting aliquots of the stock solution with an appropriate volume of 0.1 molfL hydrochloric acid. A stock 1 g/L solution of selenium (as selenite) was also prepared, by dissolving 3.3291 g of sodium selenite (Na2SeO3.5H20, minimum purity 99%; Merck, Darmstadt, F.R.G.) in 1 L of 0.1 mol/L hydrochloric acid. The ethylenediaminetetraacetate masking-agent solution was prepared by dissolving 1 g of disodium ethylenediaminetetraacetate and 2.5 g of hydroxylamine hydrochloride in 100 mL of de-ionized doubly distilled water. Other reagents include hydrochloric acid, 1 mol/L; cresol red indicator solution, 1 g/L; and diluted ammonium hydroxide, 500 mL/L. 2,3-Diamino naphthalene solution was prepared by dissolving 100 mg of it (Carlo Erba, France) in 100 mL of 0.1 mol/L hydrochloric acid. The solution was heated in a 50 #{176}C water bath for 40 mm, then purified by extracting it three
Table 1. DigestIon-Mixture Comparisons b Mlxtur.
Table 3. Analytical Recovery of Inorganic Selenium Added to Urine after Mineralization
SD
a
S#{149},ig/L
CV, %
31.92
1.10
B (5.0/2.5) C (7.5/2.5)
29.75
3.24
1.60
10.88 4.43
Urine
36.42
D (10.0/2.5)
36.17
1.25
3.45
n=
NUmberS in parentheses n = 6.
3.43
ng/s.mpl.
0
Selenium, ng/sample SD, ng/sample
723 60
939 34
mm 3O
6.
998 36
times with 10-mL portions is stored at 4 #{176}C, protected
of cyclohexane.
The aqueous phase
from light.
aliquot
containing
less than
1 j.ig of
selenium into a 50-mL microKjeldahl flask. Add 7.5 mL of concentrated 700 mLIL perchioric acid. Place on a digestion unit (hot plate or sand bath) under a fume hood and remove fumes with an aspirator. Continue heating until perchloric acid fumes appear. Digest for another 30 mm, removed, and cool. Add 1 mL of doubly distilled water, and boil until white fumes appear again. Remove and cool. Add 2.5 mL of 1 mol/L hydrochloric acid and heat for 30 mm in a boiling water bath. The addition of hydrochloric acid reduces all selenate to selenite and expels nitric acid in the form of nitrosyl chloride. After cooling, addS mL of the masking agent solution and one drop of the cresol red indicator solution. Add diluted ammonium hydroxide until there is no color, and then add 5 mL of 2,3-diaminonaphthalene solution, mix, and place in a 50 #{176}C water bath for 35 mm. Remove and cool to room temperature. Add 5 mL of cyclohexane and shake vigorously for 5 mm. Cool in ice water. When the layers have separated, draw off
layer with a Pasteur
pipette.
S.I.nlum,
ig/L
syst..i (Tsf Ion Parr bomb)
00
999 44
Procedure a 2-mL urine
Table 4. Analytical Recovery of Selenium from Organic Urinary Metabolites CI0.d
One microgram of selenium was added to 2 mL. of urine and digested as described under “Procedure.”
(upper)
10 21
501
indicate mL HNO3/mL HCIO4.
Tim..
the organic
28
Urine+ 500 ng ofSe
Table 2. Effect of Duration of Digestion after Disappearance of HCIO4 Fumes
Transfer
SD
M.*n
A (2.5/2.5)
(Adding
water to bring the organic layer up into the neck of the microKjeldahl flask will facilitate this operation.) Transfer the organic layer into a quartz cell, zero the fluorometer against the reagent blank (2.5 mL of 700 mL/L HC1O4), and measure the fluorescence of the sample. Calibrate with standards ranging from 10 to 1000 ng, prepared by treating workingstandard solutions with 2.5 mL of 700 mL/L HC1O4 in 50-mL test tubes, adding the 2.5 mL of 1 molfL HC1, and proceeding as described above.
Results and Discussion DigestionMethod Validation Of the many acid digestion mixtures proposed in the literature, the nitric/perchloric acid mixture is undoubtedly the best because of its capacity to oxidize relatively large samples per unit weight, its high oxidation potential, and a maximum assay temperature of 200 #{176}C, at which volatilization of selenite and selenate is negligible. Sulfuric acid and oxidizing agents are eliminated; however, 2,3-diaminonaphthalene, like 3,3’-diaminobenzidine, may
Urine 1 Urine 2 n=
3,
630 552 except n
=
2 for urine 1,
S.ml-clos.d sy.t.m (mlcrolcj.IdahI digestion flask)
585 548 closed system.
precipitate in the presence of sulfate and be oxidized by vanadate and molybdate, resulting in a loss of sensitivity. We added 2- mL urine aliquots to different nitric/perchloric acid mixtures (Table 1) and digested as described above.With mixtures C and D, the observed selenium concentrations were higher by 18%. Application of the t-test to the results indicated that the differences betweeji mixtures C and D, B and C, A and D, and B and D are all statistically significant (p = 0.05), but the differences between mixtures C and D are not.
Duration of Digestion Samples must be heated for 30 mm after perchioric acid fumes no longer appear, to assure removal of any objectionable amount of nitric acid and complete igestion of organic matter (Table 2), but further prolonging the digestion results in a partial oxidation of selenite into selenate. We estimated the extent of this oxidation under our conditions to be 8.8%. Selenate must be reduced to selenite by adding diluted hydrochloric acid and heating in a boiling water bath for 30 mm.
AnalyticalRecovery Inorganic selenium. Table 3 shows that no selenium was lost during digestion of urine to which was added 0.5 g of selenium as sodium selenite. Recovery averaged 99.4%. Organic selenium. Trimethyl selenium is the major selenium metabolite excreted in urine (7, 8). Two other urinary organic metabolites have not yet been identified. To study the recovery of selenium from the organic metabolites during the digestion procedure, we added selenium as sodium selenite to the drinking water of two male Sprague-Dawley lbino rats, and collected their urine for four weeks. Its selenium content was estimated fluorometrically, with a closed system (Teflon Parr bomb) that avoids any losses of selenium. Results were compared with the results obtained by the digestion method described above. Table 4 summarizes recovery data for the two methods; average values were in the 90-100% range (n = 6). Thus we conclude that no measurable selenium is lost during digestion of urine that contains trimethyl selenium and other
selenium
Detection
metabolites.
Limit and Precision
The mean blank selenium
concentration was about 6.18 ng (SD = 0.58 ng, n = 10). The detection limit of the method is 394 ng/L of urine at p = 0.05. Within-day precision was 6.3% ( = 10 gfL, n = 6), 5% ( = 20 fig/L, n = 6), 2.6% (1 = 40 tgfL, n = 6), 2.2% ( = 100 g/L, n = 6), 1.8% ( = 200 g/L,
CLINICALCHEMISTRY,Vol.26,No.9,1980 1273
sion. These subjects had not been occupationally exposed to selenium. Figure 1 shows the distribution of the urinary selenium concentrations. The average concentration found was 12.3 (SD 8.21) g/L. The range was 2.6 to 47 igfL. All samples contained selenium, but none more than 100 ig/L, which represents
the maximum
allowable
concentration
recommended
30
by Glover (1) for persons exposed to selenium and for a rural population. There were no significant sex-related differences in selenium excretion, average values for women and men being 11.8 and 11.74 .g/L, respectively (t = 0.29). Nor did we observe any correlation between selenium excretion and pH or volume of urine. The average concentration of selenium in urine we report is much lower than the values estimated by Glover (1), Smith and Westfall (5), and Sterner and Lidfelt (6). Presumably, this is because of the lack of seleniferous areas in France, and the low selenium concentrations in soils, plants, forage, and water (9).
10
References
0
0
5
10
15
zO
Selenium
25
30
35
40
45
concentration, g/L
Fig.1.DistributIon of selenium concentrations in urine of unexposed subjects n = 6), and 1.2% ( = 400 ig/L, n = 6). Day-to-day precision was 11.2% (x = 10 igfL, n = 12), 4.4% ( = 40 g/L, n = 12) and 3.2% (1 = 400 g/L, n = 12).
Selenium
in Samples of Urine
To determine the “normal” range for selenium excretion by the French population of Lyon, we assayed urine samples collected from 92 subjects (ages 14 to 77 years; 43 men, 33 women, 16 unreported) at the Institut Pasteur de Lyon and at Edouard Herriot Hospital immediately after their admis-
1274
CLINICAL
CHEMISTRY,Vol.26,No.9,1980
1. Glover, J. R., Selenium in human urine: A tentative maximum allowable concentration for industrial and rural population. Ann. Occup. Hyg. 10,3 (1967). 2. Taussky, H. H., Washington, A., Zubillaga, E., and Milhorat, A. T., Determination of trace selenium in biological fluids and tissues. Microchem. J. 10,470 (1966). 3. Roquebert, J. P., and Truhaut, R., Dosage colorim#{233}trique du selenium dana lea milieux biologiques. Bull. Soc.Pharm. Bordeaux 101,
143 (1962). 4. Dudley, H. C., Toxicology of selenium. II. The urinary excretion of selenium. Am. J. Hyg. 23, 181 (1936). 5. Smith, M. I., and Westfail, B. B., Further field studies on the selenium problem in relation to public health. U.S. Public Health Rep. 52, 1375 (1937). 6. Sterner, J. H., and Lidfelt, V. J., The selenium content of “normal” urine. J. Pharmacol. 73, 205 (1941). 7. Byard, J. L., Trimethyl selenide: A urinary metabolite of selenite. Arch. Biochem. Biophys. 130,555 (1969). 8. Palmer, I. S., Fischer, D. D., Halverson, A. W., and Olson, 0. E., Identification of a major selenium excretory product in rat urine. Biochim. Biophys. Acta 177, 336 (1969). 9. Geahchan, A., Contribution au dosage du selenium et de ses metabolites dana divers milieux (liquides biologiques, eaux, aliments). These Pharm., Universit#{233} Claude Bernard, no. 150(1978).
