CLIN. CHEM.
33/8, 1326-1330
(1987)
A Method for Simultaneously Estimating Plasma, Erythrocyte, and Leukocyte Sodium,
Potassium, and Magnesium: Reference Values and Considerations from Biological Variation
Data
Ruth E. Gallacher,’ Margaret
C.
K. Browning,’
CaIlum (3. Fraser,’ Stephen P. WIlkinson,2 and William J. MacLennan2
We describe a method for measuring plasma, erythrocyte,
and Ieukocyte sodium (Na), potassium (K), and magnesium (Mg2) concentrations in 10-mL blood specimens. After
separating cells with Ficoll-Hypaque and washing with isotonic choline chloride, erythrocytes and leukocytes are counted, so that results can be expressed as amount of substance
per cell and also to monitor cell integrity and possible
contamination. Plasma and cell lysates were analyzed (CV
7.O%) with flame photometry and atomic absorption spectrometry.
Reference
intervals
for an elderly
population
with
values for plasma electrolytes within reference intervals are
similar to those for younger healthy subjects. From data on
biological variation, reference values for erythrocyte cations
are not of much use, and analytical goals for precision are not
met, but the results might be useful for monitoring disease
progression
in individual
patients.
In contrast,
reference
values for leukocyte cations are theoretically of use and
goals are achieved, but large changes are required before
consecutive results can confidently be said to be significantly
different.
Addftlonal Keyphrases:
analytical variation
electrolytes
reference infer.’al for elderly subjects . usefulness in diagnosis
Analysis
for electrolyte
concentrations
in serum or plasma may not give a true assessment
of whole-body
electrolyte
status, particularly
because K and Mg2 are predominantly intracellular.
In consequence,
disorders
of electrolyte
metabolism,
due either to disease (1-6) or drug therapy (710), may be better assessed by assay of blood-cell electrolyte
content (11-17). Analysis
of muscle cells obtained by biopsy
has been advocated,
but this imposes discomfort
and some
risk to the patient, and results are complicated
by contamination of the sample with blood and connective
tissue (18).
There have been several
reports of studies on blood-cell
electrolyte
content. Although
several methods
can be applied to erythrocytes
(2,3, 11-14) or leukocytes/lymphocytes
(6, 10, 15-17),
few have been applied to analysis
of both
types of cell in the same blood specimen (19), and only one
allows simultaneous
assay of Na
K and Mg2 concentrations (20). Moreover, the plethora of units that have been
Departments
of ‘Biochemical
Medicine
and 2 Medicine,
Ninewells Hospital and Medical School, Dundee, DD1 9SY, Scotland.
Received June 20, 1986; accepted April 29, 1987.
1326
CLINICALCHEMISTRY, Vol. 33, No. 8, 1987
used in reporting results makes it difficult to compare
the
results
from different studies. The failure of certain previously described
procedures
to remove plasma trapped in the
cell pellets has led to the use of correction
factors that may
not be truly applicable to all types of specimens
(21).
We therefore established a relatively rapid method that
involves separating plasma and cells by a modification of
the technique
of Boyum (22), washing the cells with isotonic
choline chloride to eliminate
trapped plasma (2,23), determining the number of cells present, assessing
any possible
contamination
of one cell type with another by inspection of
proffles of cell sizes, and measuring
Na and K concentrations by flame photometry
and Mg2
concentration
by
atomic absorption
spectrometry.
The method allows estimation of plasma, erythrocyte,
and leukocyte
Nat, K, and
Mg2 concentrations
in 10-mL specimens
of heparinized
blood.
Reference
values are needed, to facilitate interpretation
of
test results and, because disorders of electrolyte
homeostasis
are more common in the elderly, we selected for our reference population
hospital
patients older than 65 years. We
then derived reference values according to the recent recommendations (24) of the Expert Panel on Theory of Reference
Values of the International
Federation
of Clinical Chemistry (IFCC). Moreover, because we firmly believe (25) that
the components of biological variation
should be defined
early in the development of all assays so as to assess the
usefulness
of traditional reference values (26), delineate
objective analytical goals for precision (27), and assess what
degree
of change is required
in a series of results
for
statistical significance,
we investigated the biological variation in erythrocyte
and leukocyte
Na,
K, and Mg2
concentrations
in a cohort of 10 ostensibly healthy individuals chosen for ethical reasons instead of an elderly popula-
tion.
Materials and Methods
Methods
Ficoll-Hypaque
(relative density 1.077), choline chloride,
and lithium heparin were from Sigma Chemical
Co., St.
Louis, MO 63178; ‘251-labeled serum albumin (specific activity 4 .LCi/L) was from Amersham
International
Ltd., Amersham, Bucks., U.K.; lanthanum
chloride solution and magnesium nitrate (both “Spectrosol”
grade) were from BDH
Chemicals Ltd., Poole, Dorset, U.K.; and lithium nitrate and
NafK
standard solutions for flame photometry were from
Instrumentation
Laboratory,
Inc., Lexington,
MA 02173.
Na and K concentrations were measured with an IL
243 flame photometer
(Instrumentation
Laboratory
Inc.),
Mg2
concentrations
with a SP9 atomic absorption
spectrometer
(Pye Unicani
Ltd., Cambridge,
U.K.). For cell
counting we used a Coulter Counter ZF (Coulter Electronics
Ltd., Luton, Beds., U.K.). Gamma radioactivity was counted
in a NE 1612 scintillation
counter
(Nuclear Enterprises,
Sighthill,
Edinburgh,
Scotland).
Procedures
Sample preparation.
Collect
a 10-mL
specimen
of venous
blood, with minimal
stasis, into a 12-mL glass tube containing 250 i.
units of heparin.
Mix gently by inversion.
Commence
analysis
within 2 h. Take a 2-mL aliquot and
centrifl.ige
(3000 x g, 15 mm). Separate
the plasma. Split
the remaining 8 mL into two 4-mL aliquots. Carefully layer
each onto 3 mL of Ficoll-Hypaque
in 105 x 15 mm plastic
tubes, centrifuge
(500 x g, 20 mm) at room temperature,
and harvest and pool the separate
erythrocyte
and leukocyte
layers.
Wash the erythrocytes
twice with 6 mL of 150 mmolfL
choline chloride, centrifuging
at 500 x g for 10 mm.
Wash the leukocytes
twice with 6-mL portions of 150
mmol/L aqueous choline chloride, centrifuging
at 270 x g
for 15 mm (first wash) and 10 mm (second wash). Then
resuspend 200 zL of packed erythrocytes
in 1.8 mL of 150
aqueous
mmol/L
aqueous
choline
chloride,
and count
the number
of
cells present with the Coulter
Counter.
Resuspend
the
entire leukocyte
pellet in 1 mL of 150 mmol/L
aqueous
choline chloride, and determine the number of cells present.
Adjust the volume to ensure that the number
of leukocytes
is within the range 5 to 10 x 106 cells per milliliter.
Examination
of cell-size profiles will reveal the presence of
contamination
with other cell types. The leukocyte
pellet
will contain predominantly
lymphocytes with a small and
variable
number
of polymorphonuclear
leukocytes.
of plasma,
erythrocytes,
and leukocytes
immediately,
or after storage at -20 #{176}C.
The specimens
are analyzed
Analytical
procedures.
For determination
of Na and K
concentrations
in plasma we used the IL 243 flame photometer, utilizing the integral diluter and a standard of 140
mmol/L Na and 5.0 mmol/L K. For measuring
plasma
Mg2
concentration
we used the SP9 spectrometer,
with
100-jL samples diluted in 5.0 mL of 720 mmollL aqueous
lanthanum
chloride, and a series of five standards
of aqueous magnesium
nitrate with concentrations ranging from
0.40 to 2.10 mmol of Mg2 per liter.
We diluted 500 L of erythrocyte
lysate with 4.5 mL of 15
mmol/L aqueous lithium nitrate and 300 L of leukocyte
lysate with 2.7 mL of aqueous lithium nitrate before analysis for Na and K. We measured erythrocyte and leukocyte
Na’ and K with the IL 243 flame photometer, using direct
aspiration. Erythrocyte
Na and K concentrations were
measured
separately, with use of 0.1 mmol/L
and 1.0
mmolJL aqueous standards,
respectively, which were used to
set the response
to full scale. Leukocyte
Na and K
concentrations
were determined
simultaneously
with the
same 0.1 mmol/L aqueous standard.
We diluted 500 1zL of erythrocyte
lysate with 5.0 mL of
720 minol/L aqueous lanthanum
chloride, and analyzed
for
Mg2 exactly as described for the plasma specimens.
To measure leukocyte
Mg2 concentration,
we added 600
jL of leukocyte
lysate to 5.0 mL of 720 mmollL aqueous
lanthanum
chloride, and used the SP9 spectrometer
and five
standards
with Mg2 concentrations
ranging
from 0.08 to
0.40 mmolJL to construct
a calibration
curve.
Analytical Variables
Efficiency
To assess whether signifitrapped
in the cell pellets,
we added 20 p.L of I-labeled
human albumin
to twelve 4mL blood specimens,
and subjected
them to the entire
analytical
procedure.
We measured
the radioactivity
in the
plasma, the first and second washes, and in the erythrocyte
and leukocyte pellets-which
were respectively resuspended
cant
amounts
of wash procedures.
of plasma remain
in 2 mL and 1 mL of aqueous choline chloride.
Precision. The precision of the entire assay was assessed
by duplicate analyses of 15 specimens.
Standard
deviations
(SD) were calculated
from the difference
between the duplicate results.
Reference values. Following IFCC recommendations
(28),
we selected a reference population of 60 patients (27 women
and 33 men, ages :65 years) admitted
to hospital with
various medical and social problems but with plasma Nat,
K, and Mg2 concentrations
within the usual adult reference intervals. The specimens
assayed
were taken at the
time of admission.
Biological Variation
We recruited apparently
healthy
laboratory
staff members (five men and five women, ages 23-48 years) for this
study. Generally,
we collected
ten 10-mL specimens of
venous blood from each subject at regular intervals during
20 weeks. To minimize pre-analytical variance, the same
phlebotomist
collected the blood at the same time each day
(between
08:30 and 10:00 h) from the seated subjects. The
analytical
procedure
was followed until cell counting had
been performed,
and then all specimens of erythrocytes
and
leukocytes
were stored at -20 #{176}C
until analyses. All specimens from the same individual
were then thawed at room
temperature,
appropriately
diluted when necessary,
and
analyzed in singlicate-as
far as practicable in a single
analytical
precision
variance
calculate
(SD?) by
batch. We had previously
(CVA) from duplicate assays
estimated
analytical
of 15 specimens.
The
of the set of results for each subject was used to
the average intra-individual
biological
variance
simple subtraction of the derived SD. from the
observed dispersion (equal to SD? + SD). After calculating
the overall variance
for the set of results for all subjects, we
subtracted
SD? and SD. to determine
the interindividual
variance
(SD).
Results and Discussion
Analytical Method
Practicability.
The method described
here is practical.
From 10-mL specimens of blood we can harvest
enough
leukocytes
for assay. Harvesting,
washing,
and counting of
cells can all be done within 2 h, and as many as eight
specimens
can be processed
in a single batch.
Advantages of the Method
The use of Ficoll-Hypaque
for cell separation
is well
documented,
and has proved very satisfactory
in our method. We confirmed
that isotonic choline chloride, used as a
wash solution, satisfactorily
removes plasma trapped in the
cell interstices. The mean radioactivity of the erythrocyte
and leukocyte pellets after two washes was not significantly
CLINICALCHEMISTRY, Vol. 33, No. 8, 1987 1327
different from background (Table 1). Choline chloride solution contains
no Na, K, or Mg2, which facilitates the
simultaneous assay of all three cations in plasma and cells.
30
We believe that such washing of cells is preferable to the
widely used correction for trapped plasma, which is based
upon limited analytical
data but nonetheless
is universally
applied to all types of specimen (21).
L0.kOyt.
E,ythro,yt.
protein, or dry mass.
Precision. The precision of the assay, assessed from duplicate analyses
of specimens from 15 patients (Table 2), is
comparable with that obtained in previously published
methods.
Reference Values
tions were within appropriate reference values. Frequency
distribution
histograms are shown in Figure 1. Since, by
visual inspection, the distributions were not truly gaussian
or log-gaussian,
we adopted the recommendation
of the
Expert Panel on Theory of Reference Values of IFCC that a
nonparametric central 0.95 fractile be used as the reference
interval (24). Table 3 shows reference intervals derived by
using
.,
.
this approach.
Table 1. Mean (and SEM) Radioactive Content of
Washes, Erythrocytes, and Leukocytes, Expressed as
Percentages of Plasma Radioactivity
Fraction
0
0.10
Fig. 1. Frequency of erythrocyteand leukocyteNa, K, and Mg2
concentrations(nmol/1o6 cells) in specimens from 60 hospitalized
elderly patients withplasma electrolyte concentrations withinreference
intervals
Table 3. Reference intervals5 for 60 Elderly Patients
with Plasma Electrolyte Concentrations within
Reference intervals on Admission
Concn in
erythrocytes
Concn in
ieukocytes
nmol/1O’ cells
Na
K
Mg2
0.33-0.89
7.3-11.0
0.13-0.26
15-66
50-105
3.9-10.0
6Nonparametnc0.95 interfractiles.
Biological Variation
The means and absolute ranges for each of the 10 healthy
subjects are shown for erythrocyte
Nat, K, and Mg2
concentrations in Figure 2, and for leukocyte Na, K, and
Mg2
concentrations in Figure 3. We used these data to
calculate the overall mean values, the analytical (CVA),
intra-individual
(CV1), and interindividual
variation
(CVG),
as coefficients
of variation,
the indices of individuality
Radioactivecounts
Plasma
Erythrocytes
First wash
Second wash
.,
,.,,ol/10’
Analyte
Reference
values
were generated
from the results of
analyses
of specimens
from 60 hospitalized
patients, ages 65
years or over, whose plasma Nat, K, and Mg2 concentra-
2
100
l-.------I
3
F-0---l
4
I-.---l
1.64 (0.29)
0.26 (0.05)
Cells
0.13 (0.01)
Leukocytes
First wash
7.26 (1.16)
0.29 (0.03)
Secondwash
Cells
0.12 (0.01)
0.10 (0.03)
Background
=
12.
‘,
..,.1
______
_____
,,,oI/10.
0
100
200
20
N.’
00
150
2
and Mg2
K’
1-.-l
Mg2
3
I-a---l
0.50
0.035
8.67
0.51
0.19
0.011
CV, %
7.0
5.9
5.8
Leukocytes
35.7
SD, nmol/106cells
2.10
CV, %
5.9
F-a-.l
I-.-l
I-.--l
_e-i
8
7
58.1
3.48
6.0
fromduplicateanalysesotiS patients’ specimens.
1328 CLINICALCHEMISTRY, Vol. 33, No. 8, 1987
5.14
0.19
3.7
-.-I
l-.l
1.4
l-.-.-.---l
l-e--4
8
I-tI
I-.4
1*1
1.1
9
I-.-4
10
Mean, nmol/106 cells
1-.-----I
5
Erythrocytes
Mean, nmof/106cells
SD, nmol/106cells
0.II
10
F-ti
2
4
Na
6
Fig. 2. Parametric means and absolute ranges for erythrocyte Nat K,
and Mg2 concentrations in 10 healthysubjects
1
Table 2. Preclsionaof Analysis for Na4, K,
In Erythrocytes and Leukocytes
aCallated
Me”
N.’
A
The cell-counting step in our procedure allows assessment
of both the number and size distribution
of the harvested
cells. This gives a guide as to whether cellular integrity has
been maintained
and whether the cells have the usual size
distribution.
It also provides an index of whether the cells
are contaminated
with other cell types, and it allows results
to be expressed
as amount of substance per cell, which we
believe is more appropriate than a derived unit, particularly
units that depend upon a further assay of (e.g.) DNA,
K’
Eytho0yt
-.4
I-.--I
I-.-I
I-a-I
I-a-I
,,,,,nmolIlO’
0.4
0.6
N.’
0.8
6.0
8.5
K’
11.0
0.10
0.15
0.20
0.25
M92
Fig. 3. Parametric means and absolute ranges for leukocyte Nat Kt
and Mg2 concentrations in 10 healthy subjects
Table 4. Overall Mean Concentrations, Analytical
Variation (CVJ, lntra-indivldual Variation (CV1),
interlndlvldual Variation (CVG), Indices of Individuality
(CVI/CVG), Analytical Goals (#{189}CV,),
and Critical
DIfferences (2.77 VCV’
+ CV1)
Erythrocytes
Na
K
Leukocytes
______________________
Mg
Mg2
Na2
Mean, nmol/106 cells
0.61
8.84
0.18
48.1
61.7
5.9
4.82
5.6
5.9
6.0
3.4
51.0
13.6
18.5
10.8
36.4
13.4
12.4
CVA, %
7.0
CV, %
1.8
CV0, %
12.4
Index
0.14
6.5
of individuality
8
0.32
1.40
1.02
1.49
Analytical goal, %
1.7
25.5
6.8
9.2
18.6
142.2
41.2
54.0
0.9
Critical difference, %
20.0
15.7
8Not significantly different from zero,
according to the formula (CVI/CVG) of Harris (26), analytical goals (as ‘/2 CV1; 27), and the critical difference required
for two results from an individual
to be significantly (P
<0.05) different (2.77 VCV. + CVI); the results are shown
in Table 4.
Overall Means and Ranges
The overall ranges for the cohort of 10 younger healthy
subjects are very similar to the reference intervals we
derived from our reference population of 60 elderly patients.
We believe that this tends to validate the widely held view
(28) that it is not necessary
to adhere to very stringent
selection of young healthy reference individuals such as
those who fulfill the demanding criteria of the provisional
recommendations
of the Scandanavian
Committee
on Reference Values (29), but merely to select individuals who have
no disorders or other pre-analytical factors that would affect
the analytes being studied.
Indices of Individuality
Conventional population-based
reference values such as
those shown in Table 3 are made up of analytical,
intraindividual, and interindividual variation. The magnitude of
the ratio of the intra- to interindividual variation casts light
on the utility
of such values (26); if the ratio is >1.4,
reference values are useful but, if the ratio is <0.6, then
reference values are of limited utility. The indices (Table 4)
are low for erythrocyte
and high for leukocyte
Nat, K, and
Mg2 concentrations.
In consequence,
determinations
of
erythrocyte
Na, K, and Mg2 concentrations
would be of
little value in the initial assessment of patients for whom
previous results were unavailable, because a single result
might well be very unusual for an individual but still lie
within
the conventional reference limits. In contrast,
because of the large CV1, conventional reference values would
theoretically be of use in the interpretation of results of
assays of leukocyte
Na, K, and Mg2 concentrations.
The
leukocyte
population in blood is heterogeneous,
however,
and variation
in this population
is likely to contribute
significantly
to the large CV1.
Analytical Goals
It is widely stated that analytical
error (as SD or CV)
should not exceed one-half of the intra-individual
biological
variation (26); that is CVA
VsCV1. Objectively set analytical goals for the precision of assays of intracellular Nat, K,
and Mg2 concentrations
are shown in Table 4. The goals for
erythrocyte
Na4, K, and Mg2 concentrations
are very
stringent
because
of the small intra-individual
variation.
We think that it is highly unlikely that these goals would be
met with currently
available techniques,
especially
in view
of the fact that any method has many intrinsic steps-such
as cell separation,
washing,
cell counting, and analysiseach of which entails error. In contrast, goals for assays of
leukocyte Na, K4, and Mg2 concentrations are less stringent, and are attained
with our method.
Critical Differences
The greater the analytical and intra-individual variation,
the greater
must be the difference between
results from
serial specimens before they can be said to be significantly
different;
for P <0.05,
the difference
required
is
2.77 VCV.
+ CV. Small differences
are required
for
erythrocyte
Na, KI-, and Mg2 concentrations to be significantly different, and these are mainly ascribable to analytical variance. This suggests that these assays would be of
value in monitoring
changes in individual
patients. In
contrast, very large changes are required for leukocyte Nat,
K, and Mg2 concentrations
to differ significantly,
mainly
owing to inherent intra-individual
variation.
Thus we believe that these assays would be of limited value in following
particular patients over time.
Overall
Conclusions
This method for simultaneous
assay of plasma, erythrocyte, and leukocyte
Nat, K, and Mg2 concentrations
incorporates
the best features of some of the many previously described
assays. The method is practicable. Reference
values derived according
to IFCC recommendations
from
hospitalized
elderly patients with plasma electrolytes within reference intervals were wide, but similar to those
derived from a smaller cohort of healthy subjects. Biological
variation data for erythrocyte
Na, K, and Mg2 concentrations show the following: reference values are of very
limited utility and thus such assays are not suitable for
diagnostic purposes; it is unlikely that assays can meet
objectively set analytical goals; and it may be that serial
results would be of value in monitoring individual patients.
In contrast, although reference values for leukocyte
Na,
K, and Mg2 concentrations are useful and goals can be
easily attained,
the large intra-individual variations mean
that these assays would be of little value in following the
courses of patients, and because of variation in the composition of the cell population may also be of limited value in
diagnosis. In view of these difficulties,
it may be that the
results of assay of electrolytes
in serum or plasma, which are
simple and cheap to obtain, relatively easy to interpret, and
widely available will continue to be generally used to assess
electrolyte
status.
We thank Smith, Kline & French
generous funding of this project.
Laboratories
Ltd. for their
References
1. Cumberbatch
M, Morgan DB. Relations between sodium transport and sodium concentrations
in human erythrocytes in health
and disease. Clin Sd 1981;60:555-64.
CLINICALCHEMISTRY, Vol. 33, No. 8, 1987
1329
2. Afifi F. Some red blood cell abnormalities
in patients with liver
cirrhosis and ascites. I. Red blood cell sodium, potassium and water
content. Tohoku J Exp Med 1985;145:7-14.
3. Cumberbatch
M, Morgan DB. Erythrocyte sodium and potassium in patients with hypokalaemia.
Clin Sci 1983;64:167-76.
4. Ambrosini E, Costa FV, Montebugnoli L, et al. Increased intralymphocytic sodium content in essential hypertension: an index of
impaired Na cellular metabolism. Clin Sci 1981;61:181-6.
5. Glass KH, Sutterfin U, Schaz K, et al. Intracellular
sodium
concentration and transport in red cells in essential hypertension,
hyperthyroidism,
pregnancy and hypokalaemia.
Clin Physiol Biochem 1986;4:199-209.
6. Edmondson RPS, Thomas RD, Hilton PJ, et al. Abnormal
leucocyte composition and sodium transport in essential hypertension. Lancet 1975;i:1003-5.
7. Edmondson RPS, Thomas ED, Hilton PJ, et al. Leucocyte
electrolytes in cardiac and non-cardiac patients receiving diuretics.
Lancet 1974;i:12-4.
8. Zidek W, Karoff CL, Baumgort P, et al. Intracellular sodium and
calcium during antihypertensive
treatment. Klin Wochenschr
1985;63(Suppl 111):147-9.
9. Herlita H, Jonsson 0, Elmfeldt D, et al. Intraerythrocyte
sodium
content and tranamembrane
sodium flux in hypertensive
patients
treated with diuretics or beta-adrenoreceptor
blockers. Acta Pharmacol Toxicol 1984;54(Suppl 1):37-42.
10. Ryan MP, Ryan MF, Counihan TB. The effect of diuretics on
lymphocyte magnesium and potassium. Acta Med Scand
1981;(Suppl 647):153-61.
11. Lijnen P, Groeseneken D, Laermans M, et a]. Methodological
assessment of assays for intracellular
concentration and transmembrane fluxes of sodium and potassium in erythrocytes
of man.
Methods Findings Exp Clin Pharmacol 1984;6:293-301.
12. Nelson D, Henningen NC. Erythrocyte contents of electrolytes
(Na, K, Mg, Zn) in healthy male controls and offspring to established hypertensive patients: a follow-up study. Scand J Clin Lab
Invest 1983;43:317-22.
13. Fortes-Mayer KD, Starkey BJ. Simpler flame photometric
determination
of erythrocyte sodium and potassium: the reference
range for apparently healthy adults. Clin Chem 1977;23:275-8.
14. Weissberg PL, West MJ, Woods KL. An improved method for
measuring intracellular electrolytes in erythrocytes and the effect
of cold storage. Clin Chim Acts 1983;129:85-9.
15. Schlebusch H, Sorger M, Hock A, Knick F. An improved
method for the determination
of potassium in leucocytes. Clin Sci
1983;64:505-1O.
1330
CLINICALCHEMISTRY, Vol. 33, No. 8, 1987
16. Elm RJ, Johnson E. A method for the determination of the
magnesium
content of blood mononuclear
cells. Magnesium
1982;1:115-21.
17. Elm RJ, Hosseim JM. Magnesium
blood cells. Clin Chem 1985;31:377-80.
content
of mononuclear
18. Edwards RHT, Harris RC, Jones DA. The biochemistry
of
muscle biopsy in man: clinical applications [Review]. Rec Adv Clin
Biochem 1981;2:243-69.
19. Abraham AS, Eylath U, Rosenman D, et al. Lymphocyte and
erythrocyte concentrations
of potassium, magnesium
in normal controls. Magnesium 1985;4:102-5.
and calcium
20. Girardin E, Paunier L. Relationship between magnesium,
potassium and sodium concentrations
in lymphocytes and erythrocytes from normal subjects. Magnesium 1985;4:188-92.
21. Baron DN, Ahmed SA. Intracellular
concentrations of water
and of the principal electrolytes determined by analysis of isolated
human leucocytes. Clin Sci 1969;37:205-19.
22. Boyum A. Separation of leucocytes from blood and bone marrow. Scand J Clin Lab Invest 1968;21(Suppl 97):9-29,31-49,5176,77-89,91-109.
23. Suzuki K, Kobayashi J, Sekimizu M. Simple micromethod for
determination
of erythrocyte
electrolyte concentration [Letter].
Clin Chem 1985;31:156-7.
24. Dybkaer R. The theory of reference values. Part 6. Presentation
of observed values related to reference values. J Clin Chem Clin
Biochem 1982;20:841-.5.
25. Howey JEA, Browning MCK, Fraser CG. Assay of serum
fructosamine that minimizes standardization and matrix problems:
use to assess components of biological variation. Clin Chem
1987;33:269-72.
26. Harris EK. Statistical aspects of reference values in clinical
pathology (Review]. Prog Clin Pathol 1981;8:45-66.
27. Fraser CG. Desirable performance standards for clinical chemistry tests [Review]. Adv Clin Chem 1983;23:299-339.
28. PetitClerc C, Wilding P. Provisional recommendation
on the
theory of reference values. Part 2. Selection of individuals for the
production of reference values. J Clin Chem Clin Biochem
1984;22:203-8.
29. Alstrom T, Grasbeck R, Fjjelm M, Skandsen S. Recommendations concerning the collection of reference values in clinical diemistry and activity report by the Committee on Reference Values of
the Scandinavian
Society for Clinical Chemistry and Clinical
Physiology. Scand J Cliii Lab Invest 1975;35(Suppl 144):1-44.