CLIN.CHEM.23/7, 1301-1305 (1977)
Simultaneous Determination of Iron, Zinc, Selenium, Rubidium,and
Cesium in Serum and Packed Blood Cells by Neutron Activation
Analysis
Jacques Versieck,’ Julien Hoste,2 Fabrice Barbier,1 Hilde Michels,1 and Julien De Rudder2
We determined
the iron, zinc, selenium,
rubidium,
and
cesium concentrations in serum and packed blood cells
by
instrumental
neutron
activation
analysis
without
chemical separations. Lyophilized samples were irradiated
for 12 days at a flux of 1013neutrons#{149}cm2#{149}s1,
mineralized by wet digestion, and measured two times with a
high-resolution Ge(Li) detector-for
6 h about a month after
the irradiation and for 15 h two or three months after the
irradiation. The following values were obtained: 1.63 ±
0.43 mg/liter (serum iron), 1025 ± 136 mg/kg wet wt
(packed cells iron), 1.13 ± 0.20 mg/liter (serum zinc),
11.15 ± 1.83 mg/kg wet wt (packed cells zinc), 0.13 ±
0.02 mg/liter (serum selenium), 0.16 ± 0.03 mg/kg wet
wt (packed cells selenium), 0.17 ± 0.04 mg/liter (serum
rubidium), 4.28 ± 0.98 mg/kg wet wt (packed cells rubidium), 0.74 ± 0.20 jzg/liter (serum cesium), and 4.82 ±
2.10 tg/kg wet wt (packed cells cesium).
Additional Keyphrases:
trace elements
-
normal values
In view of the growing interest in the role of trace elements in physiology and pathology, reliable analytical
techniques are needed. In earlier work we developed a
method for the determination
of manganese, copper,
and zinc in serum and packed blood cells (1), and re-
ported in this journal abnormal serum concentrations
in disease (2, 3).
Instrumental neutron activation analysis has already
been applied for simultaneous
determination
of trace
elements in serum, for example, by Behne and Diel (4)
and Kasperek et al. (5). However, some of the reported
results are conflicting. Other discrepancies appear when
the data are compared with previously described values
(6-9). Moreover, the number of published trace-element
determinations
in packed blood cells is still small. Thus,
further research in normal individuals seemed justified.
Instead of analyzing whole blood as Zdankiewicz
1
Rijksuniversiteit
Gent, Department
and
of Internal Medicine, Division
of Gastroenterology,
Academisch
Ziekenhuis,
De Pinte!aan
135, B9000 Gent, Belgium.
2Rijksuniversiteit
Gent, Institute for Nuclear Sciences, Laboratory
for Analytical
Chemistry,
Proeftuinstraat
86, B-9000 Gent, Belgium.
Received Jan. 5, 1977; accepted May 4, 1977.
Fasching did (10), we determined
the trace-element
concentrations separately in serum and packed blood
cells because the values can be widely divergent. For
example, whereas the mean copper concentration
in
serum is 1.5-fold higher than that in packed blood cells,
the mean manganese and zinc concentrations
are, respectively,
26- and 11-fold lower. Furthermore,
we
showed that the manganese, copper, or zinc concentrations in serum can be definitely abnormal, in contrast
with the concentrations
in packed blood cells, as in
patients with acute hepatitis, chronic aggressive hepatitis, posthepatitic
cirrhosis,
or liver metastases (2,
11).
We applied a simple neutron activation
analysis
technique for the simultaneous
determination
of selenium, rubidium, and cesium in serum and packed blood
cells. Iron and zinc can easily be determined in the same
samples, though the method was primarily not designed
for these elements: iron can accurately be determined
by numerous other, less-expensive, nonradioanalytical
methods, and zinc can be determined more easily via the
68Zn (n,
69mZn reaction
(1) than via the #{176}4Zn
(n, -y)
.)
65Zn reaction. The present paper outlines the applied
procedure and mentions
of known disease.
the results in 36 subjects
free
Materials and Methods
Subjects.
We studied 36 apparently
normal individuals, 19 men and 17 women. Blood was sampled after
an overnight fast.
Samples
and standards.
In trace-element
research,
contamination-free
sampling and analysis are of paramount importance.
Ignorance of the real extent of the
problem leads to errors, as discussed elsewhere (3, 12,
13). Although the risk of errors is much smaller during
selenium, rubidium,
or cesium determinations
than
during chromium, manganese, or cobalt determinations
(14), we took the same strict precautions as in previous
work. Venous blood samples were taken with a plastic
cannula trocar (Intranule
110 16; Vygon) and collected
in high-purity
quartz
tubes (Spectrosil;
Thermal
CLINICAL CHEMISTRY,
Vol. 23, No. 7, 1977
1301
ENERGY
Fig. 1. Gamma spectrum of a serum sample irradiated at a flux of iO’ neutrons cm2
irradiation time
(t,) =
12 days. Decay time
(t.)
=
Quarz-Schmelze)
previously
cleaned with doubly distilled water, boiled in a mixture of equal volumes of
nitric and sulfuric acid (“Suprapur”;
Merck), rinsed
again, and finally steam-cleaned with quartz-distilled
water. Twenty to 25 mg of lyophilized serum or packed
cells was weighed into high-purity
quartz ampoules
(Spectrosil), which were sealed by fusion. Before the
irradiation the samples were handled in a dust-free
room. Twenty to 25 mg of a serum supplemented with
known amounts of the elements of interest (15) was
used as a standard.
Reactor irradiation
and nuclear data. During reactor
irradiation of serum and packed blood cells, long-lived
radioisotopes of the trace elements arise [59Fe, half-life
(T112): 45.1 days; 65Zn, T112: 245 days; 75Se, T112: 121
days; 96Rb, T112:18.66 days; and 134Cs,T112:2.07 years].
The lyophilized samples were irradiated for 12 days at
a flux of iO neutrons cm2 s1. Preliminary
studies
proved that after decay of the short-lived isotopes of the
matrix elements, the long-lived isotopes of the trace
elements could be determined by Ge(Li) gamma spectrometry without chemical separations. Figure 1 shows
the spectrum of 25 mg of lyophilized serum, 64 days
after the end of the irradiation. The 1173.1 and 1332.4
keV photopeaks of 60Co are also seen; however, the
signal-to-background
ratio is too small to allow a precise
-
cobalt determination.
Isotopes
of other trace elements
were not detected.
Instrumentation.
The gamma spectra were measured
with a Ge(Li) detector and associated equipment:
#{149}
coaxial Ge(Li) detector (Philips) and preamplifier,
size 70 cm3, resolution 1.9 keV, relative detection
efficiency15.6 %.
1302
CLINICAL CHEMISTRY,
-
s
64 days. Counting time (ta)= 15 h
Vol. 23, No. 7, 1977
#{149}
amplifier (Canberra) with integration and differentiation constant at 4 is.
#{149}
4000-channel analyzer (Didac. Intertechnique).
Data reduction was done with a PDP-9 computer
(Digital Equipment Corp.) by means of a program developed by Op de Beeck and Hoste (16). For each element, the per cent standard deviation of the number of
counts of the photopeak was calculated in order to have
an indication as to the uncertainty in the individual
measurement. The following values were obtained: For
iron: 2.3%-5.6% (serum) and 2.2%-4.5% (packed cells);
for zinc: 0.496-1.8% (serum) and 0.296-2.3% (packed
cells); for selenium: 0.596-2.5% (serum) and 1.596-4.8%
(packed cells); for rubidium: 1.996-4.9% (serum) and
1.296-3.2% (packed cells); and for cesium: 3.096-11.4%
(serum) and 2.396-9.6% (packed cells).
Post-irradiation
procedure. After a decay of about
20 days, the quartz ampoules were immersed for 10 mm
in a 2:1 mixture of HF 50% and HNO:t 14 mol/liter at
room temperature, to remove outside contaminations.
After cleaningwith water,they were immersed in liquid
nitrogen and the top was removed with a diamond saw.
As radiation damage makes a quantitative removal of
the samples from the ampoules impossible, they were
placed in a 25-ml spherical flask equipped with a reflux
condenser. Three milliliters
of an equivolume
mixture
of HC1O4 70% and HNO:i 14 mol/liter was added and the
mixture was heated on a hot plate with magnetic stirrer
until complete dissolution
of the sample. The solution
was transferred into a 20-ml vial and counted for 6 h
about a month after the irradiation and for 15 h two or
three months after the irradiation.
Statistical
methods.
Outlying observations according
to the criteria of Grubbs (17) were excluded from the
calculations; they are individually
mentioned in the
text. Standard statistical techniques were used to determine means, standard deviations (SD) and coefficients of variation
(CV)
(18). The two-sample
t-test
(comparison of two means, unpaired case) was used to
test the significance of differences in mean concentrations of men and women (18).
Recalculations.
For comparison of our values with
those of other authors, we had to recalculate some of
their values so as to express them in mg/liter or tg/liter
(serum) or in mg/kg or pg/kg wet weight (packed blood
cells). Serum concentrations
given in ppm (parts per
million) or ppb (parts per billion) were multiplied
by
1.026; data on packed blood cells given in mg/liter or
tg/liter
were divided by 1.096. In some of the original
papers, the concentrations
are expressed in mg/kg or
ag/kg dry wt. Conversion into the wet weight was made
by using the conversion factors reported by the authors.
For comparison of our values ‘yith those of Zdankiewicz
and Fasching (10), we estimated the mean whole-blood
trace-element contents of our subjects from their serum
and packed blood cells concentrations,
using a hematocrit of 44%.
The values we obtained
are normally
distributed.
We
did not find outliers for iron, zinc, or selenium concenin serum or packed blood cells.
serum, the mean ±1 SD is 1.63 ± 0.43
For
mg/liter. For
centration in
is significantly
SD
iron in packed blood cells, the mean conmen (1072 mg/kg wet wt; SD = 122 mg/kg)
higher than in women (973 mg/kg wet wt;
134 mg/kg) (t
2.312;0.02 <P
<0.05).
in men (1.22
mg/liter; SD = 0.20mg/liter)issignificantly
higherthan
in women (1.03 mg/liter; SD = 0.14 mg/liter) (t = 3.369;
0.001 <P <0.01). For packed cells, the mean ±1 SD is
11.15 ± 1.83 mg/kg wet wt.
Selenium. For serum, the mean ±1 SD is 0.13 ± 0.02
mg/liter. The dispersion of the results is very small (CV
= 15.4%; range, 0.09-0.18
mg/liter). For packed cells, the
mean ±1 SD is 0.16 ± 0.03 mg/kg wet wt, and again the
dispersion is small (CV = 18.8%; range, 0.09-0.21 mg/kg
=
slightly greater than for rubidium (CV = 27.0%; range,
0.45-1.18 pg/liter). For packed cells, the individual
having an outlying serum and packed cells rubidium
and serum cesium concentration
has also an outlying
packed cells cesium concentration,
namely 15.77 Lg/kg
wet wt. The mean concentration ±1 SD is 4.82 ± 2.10
pg/kg wet wt. The dispersion is thus markedly higher
than the serum value (CV = 43.6%; range, 2.19-11.04
tg/kg wet wt).
The mean selenium concentration in serum is thus
1.3-fold lower than the mean concentration in packed
blood cells, whereas the mean rubidium and cesium
concentrations in serum are respectively 26- and 6.7-fold
lower (the units for all values being mg/kg or ag/kg wet
wt).
Discussion
Zinc in serum. The significant difference between the
mean concentration
in men and women reported in an
earlier publication (2) is confirmed.
The selenium, rubidium,
Results
trations
Iron.
1.62 big/liter (a 22-year-old nurse who was taking oral
contraceptives).The mean concentration ±1 SD is 0.74
± 0.20 tg/liter. The dispersion of the results is thus
=
Zinc. In serum, the mean concentration
wet wt).
Rubidium.
For serum, we found one outlying value,
namely 0.34 mg/liter (a 29-year-old Moroccan male, who
had worked for nine years in Belgium and who underwent a cornea transplant because of keratoconus). The
mean ±1 SD is 0.17 ± 0.04 mg/liter. The dispersion of
the results is somewhat greater than for selenium (CV
= 23.5%; range, 0.09-0.27
mg/liter). For rubidium in
packed cells, the individual having an outlying serum
rubidium concentration also has an outlying value for
packed cells concentration, namely 7.38 mg/kg wet wt.
The mean ±1 SD is 4.28 ± 0.98 mg/kg wet wt. The dispersion is similar to that for the serum value (CV =
22.9%; range, 2.66-7.24 mg/kg wet wt).
Cesium.
For serum, we found two outliers: 2.57 tg/
liter (same subject having also outlier values for serum
and packed blood cells rubidium concentration) and
and cesium values we ob-
tained are compared with the previously reported results of other investigators
(Tables 1-3).
Selenium in serum. The values we found in this study
are higher than determined
by Bebne and Diel (4) and
Maria et al. (22), lower than reported by Wester (9), but
in agreement with the results of others (19-21, 23).
Selenium
in packed cells. The values we found are
higher than determined by Maxia et al. (22) but similar
to the results of Dickson and Tomlinson (20). Selenium
in plasma but not in packed blood cells has been found
to be markedly depressed in three patients with extensive burns (20) and in children with kwashiorkor and
marasmus (24, 25). The concentration in plasma or
serum seems to be a more sensitive indicator of the selenium status than that in whole blood (26).
Rubidium
in serum. Our values are lower than determined by Niedermeier et al. (7) and Wester (9), but
similar to the results of Kasperek et al. (5), Chechan et
al. (27), and Wood (28).
Rubidium in packed cells. Our values agree perfectly
with those of Chechan et al. (27) and Wood (28). Interest in rubidium
has been stimulated
by its close
physicochemical
relationship
to potassium
and its
unique neurophysiological
characteristics.
Studies have
demonstrated
that rubidium
and lithium
have contrasting behavioral, electroencephalographic,
and biochemical properties (29). Wester (9) mentions that the
serum rubidium concentration of untreated hypertensive patients does not differ significantly from that of
normotensive subjects, but decreases after five days of
treatment with chlorthalidone. As far as we know, the
rubidium concentration in serum and packed cells has
not yet been extensively
studied in other human diseases.
Cesium in serum. The concentration
mentioned
by
Niedermeier
and Griggs (8) differs from all others. The
CLINICAL
CHEMISTRY,
Vol.
23, No. 7,
1977
1303
Table 1. Reported Selenium Concentrations
COncentrations
Serum
Reference
This study
(4)
(19)
(20)
(21)
Mean ±1 SD
Range
Mean ±1 SD
Mean
Mean ±1 SD
Mean ± 1 SD
(23)
(9)
mg/kg
± 0.02
0.13
0.09-
wet wt
0.16 ± 0.03
0.09
0.21
0.18
-
0.061 ± 0.016
0.11
0.144
0.13 1
±
±
0.215 ± 0.055
0.029
0.007
0.113 ± 0.008
9
(22)
Packed blood cells
mg/liter
0.052
Mean
Range
Mean ±1 SD
Range
Mean ±1 SD
0.046
0.045
-
0.062
0.040
-
0.054
0.129 ± 0.034
0.055
-
0.169
0.19 ± 0.06
Table 2. Reported Rubidium Concentrations
Serum
Refer.nc.
Packed blood cells
mg/kg wet wt
mg/lIter
This study
Mean ± 1 SD
(27)
Range
Mean ±1 SD
(5)
(7)
Mean ± 1 SD
Mean
(9)
Range
Mean ±1 SD
Mean ±1 SD
0.17 ± 0.04
0.09
0.27
0.12 ± 0.038
0.10 ± 0.045
0.174 ± 0.092
0.39
0.06
0.96
0.27 ± 0.19
0.16 ± 0.03
-
9
(28)
Concentrations
4.28 ±
2.66 4.04±
3.98 ±
0.98
7.24
1.12
0.81
3.81 ±
0.57
-
Table 3. Repo rted Cesium Concentrations
Concentrations
Serum
ag/liter
Reference
Packed blood cells
pg/kg wet wi
This study
Mean ±1 SD
Range
0.74 ± 0.20
0.45-1.18
4.82 ± 2.10
2.19-11.04
(19)
Mean
Mean ±1 SD
3.1
2.6
(5)
(8)
(9)
Mean±1 SD
Mean ±1 SD
values we found are similar to the results of Kasperek
et al. (5) and Wester (9), but the dispersion
is much
smaller: CV = 27.0% vs. 51.1% (5) or 90.0% (9).
Cesium in packed cells. We are aware of only one
previously
published
value, namely 2.6 Lg/kg wet wt
(19), a result calculated from concentrations
in whole
blood and plasma, assuming a hematocrit of 42.8%. The
values we found are thus higher. It is noteworthy
that
the cesium values for packed cells are more dispersed
than the serum values. Niedermeier
and Griggs (8) report that cesium appears in higher concentrations
in
serum of patients
with rheumatoid
arthritis.
This
statement has to be re-evaluated in view of the existing
discrepancy
between their values for normal subjects
1304 CLINICALCHEMISTRY,Vol. 23, No. 7,
1977
1.33 ± 0.68
63.0 ± 42.0
1.0 ± 0.9
and those of other investigators.
Wester (9) mentions
that the serum cesium concentration
of untreated hy-
pertensive patients does not differ significantly from
that of normotensive subjects. He observed a decreased
urinary excretion during treatment with chlorthalidone.
Zdankiewicz and Fasching (10) found a 17% lower
concentration in whole blood of cancer patients as
compared with normal subjects. We are not aware of
other data in human pathology.
The recently reported whole-blood iron, zinc, and
selenium concentrations
(10) agree with the values of
this study, calculated as outlined above. Zdankiewicz
and Fasching (10) did not determine the concentration
of rubidium of which the Rb 1076.6 keV photopeak is
seen just before the 59Fe 1098.6 keV photopeak. The
mean cesium values are different: 16.3 sg/kg (10) vs. 2.9
sg/kg (this study).
We found normally distributed
values. None of the
36 subjects had outlier values for iron, zinc, or selenium
in serum or packed blood cells. One non-Belgian
male
had an outlying (high) serum and packed cells rubidium
and cesium concentration,
and one female an outlying
(high) serum cesium concentration.
The results of our
experiments
illustrate that iron, zinc, selenium, rubidium and cesium concentrations
in serum and packed
blood cells of normal individuals
vary only within rel-
atively narrow limits, as do the manganese, copper, and
zinc concentrations
(2, 12, 30-32). This finding is in
striking contrast with the data of Niedermeier et al. (7)
and Niedermeier
and Griggs (8). Collectively, our data
suggest that trace-element
concentrations
in serum and
packed blood cells are controlled
by as accurate hO-
meostatic mechanisms as other important biochemical
variables. Further studies should be directed to learning
whether the values change in disease, and how far this
may have diagnostic or therapeutic implications.
We acknowledge the careful technical assistance and secretarial aid
of Miss Lidia Vanballenberghe and Mrs. Yvette Odent. This work was
supported by the Nationaal Fonds voor Wetenschappelijk Onderzoek
and the Interuniversitair
Instituut voor Kernwetenschappen.
References
1. Versieck, J., Speecke, A., Hoste, J., and Barbier, F., Determination
of manganese, copper, and zinc in serum and packed blood cells by
neutron activation analysis. Z. Kim. Chem. Kim. Biochem.
11, 193
(1973).
2. Versieck, J., Barbier, F., Speecke, A., and Hoste, J., Manganese,
copper, and zinc concentrations
in serum and packed blood cells
during acute hepatitis, chronic hepatitis and posthepatitic cirrhosis.
Clin. Chem. 20, 1141 (1974).
3. Versieck, J., Barbier, F., Speecke, A., and Hoste, J., Influence of
myocardial infarction on serum manganese, copper, and zinc concentrations.
Clin. Chem. 21, 578 (1975).
4. Behne, D., and Diel, F., Relations between carbohydrate
and
trace-element metabolisms investigated by neutron activation analysis. In Nuclear Activation
Techniques
in the Life Sciences,
International Atomic Energy Agency, Vienna, 1972, pp 407-414.
5. Kasperek, K., Schicha, H., Siller, V., et a!., Trace-element
concentrations
in human serum: Diagnostic implications.
Ibid., pp
517-526.
6. Grafflage, B., Buttgereit, G., Kubler, W., and Mertens, H. -M., Die
Messung der Spurenelemente
Chrom und Mangan im Serum mittels
flammenloser
Atomabsorption.
Z. Kim. Chem. Kim. Biochem. 12,
287 (1974).
7. Niedermeier, W., Creitz, E. E., and Holley, H. L., Trace metal
composition of synovial fluid from patients with rheumatoid arthritis.
Arthritis
Rheum.
5,439 (1962).
8. Niedermeier, W., and Griggs, J. H., Trace metal composition of
synovial fluid and blood serum of patients with rheumatoid arthritis.
J. Chronic Dis. 23,527 (1971).
9. Wester, p. 0., Trace elements in serum and urine from hypertensive patients before and during treatment with chlorthalidone. Acta
Med. Scand.
194,505 (1973).
10. Zdankiewicz, D. D., and Fasching, J. L., Analysis of whole blood
by neutron
activation:
A search for a biochemical
plasia. Clin. Chem. 22, 1361 (1976).
indicator
of neo-
11. Versieck, J., Hoste, J., and Barbier, F., Determination
du manganbse, du cuivre et du zinc dans le serum chez des sujets normaux
et des malades atteints de m#{233}tastases
h#{233}patiques.
Acta Gastroenterol.
Belg. 39,340 (1976).
12. Versieck, J., Barbier, F., Speecke, A., and Hoste, J., Normal
manganese
concentrations
in human serum. Acta Endocrinol.
(Copenhagen)
76, 783 (1974).
13. Cornelis, R., Hoste, J., Speecke, A., et al., Activation
analysisPart 2. In Physical
Chemistry
Part 1, International
Review
Series Two, Analytical
Chemistryof Sciences,
Butterworths,
London,
Boston, 1976, pp 71-149.
14. Versieck, J., and Speecke,
A., Contaminations
induced
by collection of liver biopsies and human blood. In Nuclear
Activation
Techniques
in the Life Sciences,
International
Atomic
Energy
Agency, Vienna, 1972, pp 39-49.
15. Cornelis, R., Speecke, A., and Hoste, J., A multi-element
serum
standard for neutron activation analysis. Anal. Chim. Acta 68, 1
(1973).
16. op de Beeck, J. P., and Hoste, J., The application of computer
techniques
to instrumental
neutron
activation
analysis.
In Physical
Chemistry Series Two, Analytical
Chemistry-Part
1, International
Butterworths,
London, Boston, 1976, pp 151-
Review of Sciences,
189.
17. Grubbs, F. E., Sample criteria for testing outlying observations.
Ann. Math. Stat. 21, 27 (1950).
18. Armitage, P., Statistical Methods in Medical Research, Blackwell
Scientific Publ. Ltd., Oxford, England, 1971, pp 28-43, 118-112.
19. Bowen, H. J. M., Trace Elements in Biochemistry,
Academic
Press, London and New York, 1966, pp. 81-82.
20. Dickson,
R. C., and Tomlinson, R. H., Selenium in blood and
human tissues. Clin. Chim. Acta 16, 311 (1967).
21. Hahn, H. K. J., Williams, R. V., Burch, R. E., et a!., Determination
of serum selenium by means of solvent extraction combined with
activation analysis. J. Lab. Clin. Med. 80,718 (1972).
22. Maxia, V., Meloni, S., Rollier, M. A., et al., Selenium and chromium assay in Egyptian foods and in blood of Egyptian children by
activation analysis. In Nuclear Activation
Techniques
in the Life
Sciences,
International
Atomic Energy Agency, Vienna, 1972, pp.
527-550.
23. Morss, S. G., Ralston, H. R., and Olcott, H. S., Selenium determination in human serum lipoprotein fractions by neutron activation
analysis. Anal. Biochem.
49, 598 (1972).
24. Burk, R. F., Pearson, W. N., Wood, R. P., and Viteri, F., Bloodselenium levels and in vitro red cell uptake
Amer. J. Clin. Nutr. 20, 723 (1967).
of 75Se in kwashiorkor.
25. Levine, R. J., and Olson, R. E., Blood selenium in Thai children
with protein-calorie
malnutrition.
Proc. Soc. Exp. Biol. Med. 134,
1030 (1970).
26. Burk, R. F., Selenium in man. In Trace Elements
in Human
Health and Disease, II. Essential
and Toxic Elements.
A. S. Prasad
and D. Oberleas, Eds., Academic Press, New York, San Francisco,
London, 1976, pp 105-133.
27. Chechan, C., Marchandise,
X., and Lekieffre, J., Dosage du rubidium par spectrophotom#{233}trie
d’absorption atomique dans quelques
milieux biologiques. Ecueils a #{233}viter,
r#{233}sultats
pr#{233}liminaires.
C. R. Soc.
Bioi. 169,991 (1975).
28. Wood, 0. L., Determination
of rubidium in human erythrocytes,
plasma, and urine by atomic absorption spectrophotometry.
Riochem.
Med. 3,458 (1970).
29. Fieve, R. R., Meltzer, H., Dunner, D. L., et al., Rubidium: Biochemical,
behavioral,
and metabolic
studies in humans. Amer. J.
Psychiat.
130,55
(1973).
30. Cotzias, G. C., Miller, S. T., and Edwards, J., Neutron activation
analysis: The stability of manganese concentrations in human blood
serum. J. Lab. Ciin. Med. 67, 836 (1966).
31. Lahey, M. E., Gubler, C. J., Cartwright, G.E., and Wintrobe, M.
M., Studies on copper metabolism. VI. Blood copper in normal human
subjects. J. Clin. Invest.
32, 322 (1953).
32. Halsted, J. A., and Smith, J. C., Plasma zinc in health and disease.
Lancet
i, 322 (1970).
CLINICALCHEMISTRY,Vol. 23, No. 7, 1977
1305