Download An isotonic potassium chloride liquid junction minimises

Ann Clin Biochem 1988; 25: 228-232
An isotonic potassium chloride liquid junction
minimises the effects of ionic strength, protein and
haematocrit on ionised calcium measurement
RBPAYNE
From the Department of Chemical Pathology, St James's University Hospital, Leeds LS9 7TF, UK
When the reference electrode liquid junction of a Nova 2 analyser was
changed to isotonic potassium chloride, increasing the ionic strength of aqueous
solutions containing a constant total calcium concentration had a negligible effect
on measured ionised calcium. In contrast, measurements using hypertonic
potassium chloride, hypertonic sodium formate and isotonic sodium chloride liquid
junctions showed significant sample ionic strength effects. Interferences by sample
protein concentration and haematocrit were marked with hypertonic, but negligible
with isotonic junctions. Ionised calcium values in samples containing 25 mmoVL
acetate, bicarbonate, ~-hydroxybutyrate, lactate or pyruvate were all lower by
fr.7% with an isotonic than a hypertonic potassium chloride junction. Thus, anions
that replace bicarbonate during metabolic acidosis have a similar effect on residual
liquid junction potential. The clinical usefulness of an isotonic potassium chloride
liquid junction needs to be evaluated.
SUMMARY.
The clinically significant positive interference
with the measurement of ionised calcium by
sample protein concentration I. 2 was suggested
by others to be a Donnan effect.? However,
subsequent experiments showed that it disappeared when the reference electrodes liquid
junction was changed from 2·0 moVL KCI or
4·6 moUKg HCOONa to 150 mmoVL NaCI; in
addition, measurements became significantly
more reproducible." Isotonic NaCI bridges are
not used in clinical analysers because their
residual liquid junction potential varies with the
ionic strength of the sample and is said to vary
with its anionic composition. Preliminary experiments suggested that variation in ionic
strength had little effect when the brid~e was
150 mmoVL KCI rather than NaCI. The
purpose of the work described here was to
determine the effects of sample ionic strength,
protein concentration, haematocrit and anion
composition on ionised calcium measured using
a range of reference electrode liquid junctions.
Materials and methods
Bicarbonate-free solutions with high and ,w
protein concentrations were prepared trom
reconstituted Qualify I human control serum lot
QN21A (American Monitor UK Ltd, Storrington, Sussex, UK) with a total protein concentration of 62 giL. It was diluted with a solution :
containing 1·0 mmoVL CaCI2 and 150 mmoVL
NaCI to achieve a low total protein concentration of 32 giL. An aliquot was then
ultrafiltered" to achieve a high total protein
concentration of 82 giL in the retentate. Protein
concentration was measured on : Monitor
Parallel analyser by a biuret method with an
imprecision of 1·3% at 70 giL.
A bicarbonate-free 'whole blood' preparation with a high packed cell volume and its
'plasma' were prepared by washing the cells
from 10 ml of heparinised blood three times
with a solution containing 1·0 mmoVL CaCl2
and 150 mmoVL NaCI and then resuspending
them in reconstituted Qualify I control serum;
after centrifugation, an aliquot of the 'plasma'
was removed, leaving a 'whole blood' preparation with an haematocrit of 75%.
Bicarbonate-containing aqueous solutions
were equilibrated with humidified 5% CO 2 at
37°C immediately after preparation to avoid the
formation of carbonate complexes."
Ionised calcium concentrations were mea-
228
Downloaded from acb.sagepub.com at PENNSYLVANIA STATE UNIV on May 11, 2016
Isotonic liquid junctions
sured with a Nova 2 ionised calcium analyser.
The reference electrode liquid junction geometry was T-shaped with the manufacturer's
reference solution of 2·0 moVL KCI flowing
from the side into an upward-moving sample
stream; there was no liquid movement during
measurement. The liquid junction solution was
changed by removing the reference solution
straw from the fluids pack, inserting it in the
new solution and purging the machine twice. It
was then recalibrated after the displayed potential (mY) had stabilised. A 4·6 mol/Kg
HCOONa salt-bridge solution was purchased
from Radiometer (Copenhagen, Denmark).
The Nova's internal standards were ISO mmoVL
NaCI containing 1·0 and 2·0 mmoVLCaCI2 • All
measurements were made in duplicate after a
single calibration except when otherwise stated.
Results were recorded to three decimal places
and are reported as means to two places.
Results
229
1'04
~
0
1'02
.s
1'00
J!
8
0'98
]c
0·96
e
~
.2
1
j
0·94
0'92
0'90
,
,
,
,
,
,
100
110
120
130
140
150
Sample Noel concentration (mmoVLI
FIG. I. Ionised calcium values in aqueous solutions
containing 1·0 mmol/L CaCI2 with NaCI concentrations ranging from )(K) to 150 mmol/L, measured
using reference electrode liquid junctions of 2·0
mol/L KCI (.). ISO mmol/L KCI (0) and ISO
mmol/L NaCI (e). The data points using 4·6 mol/Kg
HCOONa were close to those for 2·0 mol/L KCI and
are not shown.
IMPRECISION
Coefficients of variation calculated from duplicate measurements made between calibrations
were less than 0·5% when aqueous solutions,
protein solutions or whole blood were measured with reference electrode liquid junctions
of either ISO mmoVL NaCI or KCI, and also
when aqueous solutions were measured with
the manufacturer's 2·0 moVL KCI junction.
However, within-calibration imprecisions were
significantlygreater (P«HH) when a 2·0 moVL
KCI bridge was used to measure either protein
solutions (CY 0,77%) or whole blood (CY
O·84'ra).
SAMPLE IONIC STRENGTH EFFECTS
The ionised calcium concentrations of aqueous
solutions containing 1·0 mmoVL CaCI2 and
having NaCIconcentrations ranging from to() to
ISO mmol/L were measured using four different
reference electrode liquid junctions (Fig. I).
With the standard 2·0 moVL KCI bridge the
values fell from 1·03 to HlO mmoVL (r=-O·994,
P<()·OO5); with a 4·6 mol/Kg I1COONa bridge
the values fell from 1·()4 to I·(K) mmoVL
(r=-()'978, P«H)()5); with a ISO mmoVLNaCI
bridge the values rose from ()·91 mmoVL to
1·00 mmoVL (r=()·996, P«HK)()5); and with a
150 mmoVL KCI bridge the values did not
change significantly from I·()() mmoVL
(r=0·594, P=()·2).
The ionised calcium concentrations of three
aqueous solutions each containing I·() mmoVL
CaCh and either 100, 150 or 185 mmoVLNaCI
were measured using reference liquid junctions of 2·0 mmol/L, 1·5 mol/L, 1·0 mol/L,
500 mmoVLand 150 mmol/L KCI. The calcium
values of the solution containing 100 mmoVL
NaCI fell significantly from 1·04 to ()·99mmollL
as the junction molarity was reduced from 2·0
mollL to ISO mmoVL (r=-0·964, P<O·OOI).
The calcium values in the solution containing
ISO mmoVL NaCI did not change significantly
across the range of bridge compositions from
0·99 mmoVL (r=-0·108, P=0·77). In the
solution containing 185 mmoVL NaCl, the
calcium values rose significantly from 0·96
mmoVL with the 2·0 mollL KCI bridge to 0·99
mmoVL with the 150 mmollL bridge (r=0'933,
P<O·OOI). Thus, values that ranged from 0·96
in a hyperosmolar solution to 1·04 mmollL in a
hypo-osmolar solution when measured using
the conventional liquid junction all became 0·99
mmollL when an isotonic KCI junction was
used.
PROTEIN CONCENTRATION EFFECTS
The ionised calcium concentrations of the
control serum preparations with total protein
concentrations ef 32 and 82 giL were 1·16 and
1·2() mmol/L, respectively, when measured with
the 2·() mollL KCI liquid junction. The values
became closer as the molarities of the bridge
solutions were reduced, until both became 1·05
mmoVL with the isotonic KCI junction (Fig. 2).
Downloaded from acb.sagepub.com at PENNSYLVANIA STATE UNIV on May 11, 2016
230
Payne
1'25
1'25
~e
.se
$.
"0
e
.!
e
1'20
.2
.~
..\!
..\!
0
u
1.§
0
u
1015
l
1015
.~
c:
.2
"0
;
2
~
1·20
"0
..2
I!
:>
1010
1010
~
1·05
1·05
2'0
1'5
1·0
0'5
2·0
0'15
Liquid junction KCI (mol/L)
1·5
1·0
0·5
0'15
Liquid junction KCI (mol/L)
FIG. 2. Ionised calcium values in a diluted
bicarbonate-free control serum and its own retentatc
after Ultrafiltration. measured using a range of liquid
junction KCI molarities.
FIG. 3. Ionised calcium values in a bicarbonate-free
control serum with and without the addition of red
blood cells to give a hacrnatocrit of 75%. measured
using a range of liquid Junction KCI molarities.
HAEMATOCRIT EFFECTS
for each anion heing lowered to a similar
extent.
When measured with the conventional bridge.
.the ionised calcium of the whole blood preparation with an haematocrit of 75'Yo was
1·23 mmoUL while that of its plasma was
1·19 mrnol/L, Both values were 1·05 mmoVL
when measurements were made using an isotonic KCI junction (Fig. 3).
EFFECTS OF ANIONS
The ionised calcium concentrations of isotonic aqueous solutions containing sodium
150 mmol/L, calcium 1·67 mmoVL, chloride
127 mmol/L and either acetate, bicarbonate, ~­
hydroxybutrate, lactate or pyruvate 25 mmol/L
were determined with 2·0 mol/L and
150 mmol/L KCI liquid junctions (Table I). In
each case the values were significantly lower
with the ISO mmol/L bridge (P<O·OOI), values
Discussion
Hypertonic reference electrode liquid junctions
have practical disadvantages for ion-selective
electrode measurements in patients' samples. A
clinically significant positive effect of protein
concentration has heen confirmed hy several
workers. I. 2. 7-'} The interference is more apparent with ionised calcium and sodium measurements than with pH because the reference
ranges for calcium and sodium arc only about
1·5 mY wide compared with 5-6 mY for pH.'} It
has been shown previously that this interference, which is greater with a 4·6 mol/Kg
HCOONa than with a 2·0 mol/L KCI bridge, is
abolished by the use of an isotonic NaCI bridge"
I. Ionised calcium concentrations in aqueous solutions containing ISO mmol/L sodium. 1·67 mmol/L
calcium, 128 mmoIlL chloride and 25 mmoIlL acetate. bicarbonate. ~-hydroxybutyrate. lactate or pyruvate
measured with different reference electrode liquid junctions
TABLE
Mean ionised calcium (mmoIlL)
2·0 moIlL KCI
150 mmoIlL KCL
Difference
Acetate
Bicarbonate
Butyrate
Lactate
Pyruvate
1·57
1-47
6·4%
1·50
1-41
1·54
1·43
7·I'Yc,
1·27
1·19
6·3%
1·17
I· J()
6-()%
6-()'}t,.
Downloaded from acb.sagepub.com at PENNSYLVANIA STATE UNIV on May 11, 2016
Isotonic liquid junctions
231
2. Summary of the relationships between reference electrode liquid junction composition and
interference by hacmatocrit. protein and ionic strength
TABLE
Liquid junction
Interference
4·6 mol/Kg
HCOONa
_-_..-
..
Haematocrit
Protein
Ionic strength
++
+
2·() mol/L
KCI
().IS mol/L
NaCI
+
+
+
++
..' - _..
presumably because it abolishes the precipitation of protein (and of red celIs) that is known
to occur at hypertonic junctions.9-13 An additional advantage of an isotonic junction is that
the instability of readings is reduced," 12 an
observation confirmed here.
The use of an isotonic NaCI bridge for pH
measurements to lessen irreproducibility due to
protein precipitation was proposed as long ago
as 1959;14. 15 it was noted that it would have
lowered normal blood pH values by about 0·1
units. 16 The proposal was not adopted because
when an isotonic NaCI bridge was used the
junction potential became more susceptible to
the ionic strength of the sample. 12 This observation has been confirmed in the present work,
but it was also found that even when 2·0 mol/L
KCI or 4·6 moVKg HCOONa junctions were
used the measured values changed with sample
ionic strength (Fig. 1). In contrast, with isotonic
KCI there was no measurable effect of changing ionic strength. It might be argued that the
falling values with the hypertonic junctions are
a reflection of the decreasing 'activity' of the
calcium ion as ionic strength increases. If that
were the case, the absence of the effect with
isotonic KCI must be due to a fortuitous
equivalence between faIling ionic activity and
increasing residual liquid junction potential.
With the hypertonic junction, the effect of ionic
strength extended across the range of clinicalIy
encountered sodium concentrations, from IOO
to 180 mmol/L, and there were measurable
differences in apparent ionised calcium at alI
KCI junction molarities tested until a concentration of ISO mmoVL was reached.
The present work has shown that interference by sample protein concentration is negligible with an isotonic KCI junction (just as it is
with an isotonic NaCI junction") and that the
haematocrit effect is also negligible (Fig. 2).
The advantages of this junction are summarised
in Table 2.
0·15 mol/L
KCI
The major problem with the adoption of an
isotonic KCI junction for clinical use is that
sample anions such as bicarbonate that diffuse
more slowly than chloride alter the residual
liquid junction potential, and therefore the
measured values, and this effect is more
marked the
more dilute the liquid
junction.'?' III However, the magnitude of the
observed effect was similar for bicarbonate and
for each of the organic anions tested (Table 1),
so the effect may prove to be acceptable in
practice because anions that replace bicarbonate during metabolic acidosis will have a similar
effect on residual liquid junction potential.
The clinical usefulness of an isotonic KCI
junction needs to be evaluated to determine
whether the disadvantage of any residual effect
of anion composition is outweighed by the
advantages of decreased imprecision, insensitivity to change in sample ionic strength and
negligible interference by protein concentration
and haematocrit.
References
1 Payne RB. ClinieaIly significant effect of protein
concentration on ion-selective electrode measurements of ionised calcium. Ann Clin Biochem 1982;
19: 233-7.
2 Butler SJ. Payne RB, Gunn IR et al. Correlation
between serum ionised calcium and serum albumin in two hospital populations. Br Med J 1984;
289: 948-50.
3 Thode J. Fogh-Andersen N, Sigaard-Andersen O.
Donnan effect or protein interference in ionised
calcium measurements'! Ann Clin Biochem 1983;
20: 271-3.
4 Payne RB, Jones DP. Protein interferes with
ionised calcium measurement at the reference
electrode liquid junction. Ann Clin Biochem 1987;
24: 400-7.
5 Payne RB. Isotonic liquid junctions minimise
protein interference with ionized calcium
measurements. In: Mass AHJ. Buckley B, Marsoner H et al. cds. Methodology and Clinical
Applications of Ion-Selective Electrodes. Proceedings of the 8th meeting of the wee European
Downloaded from acb.sagepub.com at PENNSYLVANIA STATE UNIV on May 11, 2016
232
Payne
Working Group on Ion-Selective Electrodes 1986
(ISBN 90-9001722-4). Copenhagen: IFCC Technical Secretariat, 1987; 287-8.
6 Boink FBTJ, Bijster P, Vink KI.J, Maas AHJ.
Direct potentiometric determination of sodium
ion in blood. III. Influence of (bi)carbonate. Clin
Chem 1985; 31: 523-6.
7 Buckley BM, Smith SCH, Hcath DA, Bold AM.
Clinical studies on ionized calcium using the
Radiometer ICAI analyzer. Scan J Clin Lab
Invest 1983; 43 (Suppl 165): 87-92.
8 Freaney R, Egan T, McKcnna MJ et al. Determination of ionised calcium is not independent of
albumin concentration. Clin Chim Acta 1986; 158:
129-38.
9 Pedersen KO. The effect of protein concentration
on potentiometric determination of ions in serum
and blood. In: Maas AHJ, Boink FBTJ, Saris NEL et al., eds. Methodology and Clinical
Applications of Ion-Selective Electrodes. Proceedings of the 7th meeting of the IFCC European
Working Group on Ion-Selective Electrodes 1985
(ISBN 87-88138-10-0). Copenhagen: Radiometer,
1986; 77-98.
10 Siggaard-Anderscn O. Factors affecting thc
liquid-junction potential in c1ectrometric blood
pH measurement. Scan J Clin Lab Invest 1961; 13:
205-11.
11 Maas AHJ. pH determination in body Ilnids using
a cell with an isotonic sodium chloride bridge. J
Appl Physiol 1971; 30: 248-50.
12 Salling N, Siggaard-Andersen O. Liquid-junction
potentials between plasma or erythrolysatc and
KCI solutions. Scand J Clin Lab Invest 1971; 28:
33-40.
13 Siggaard-Anderscn O. Fogh-Anderscn N. Thode
J. Christiansen TF. Elimination of the erythrocyte
effect on the liquid junction potential in potentiometric measurements on whole blood using
unconventional salt bridge solutions. c.g. sodium
formate. In: Maas AHJ. Kolstad J. SiggaardAndersen 0, Kokholm G. cds. Ionized Calcium,
Sodium and Potassium by Ion-selective Electrodes.
Proceedings of the 5th meeting of the IFCC
European Working Group on Ion-Selective Elcctrodes (ISBN 87-88 I38-()4-6). Copenhagen: Private Press. 1984; 149-51.
14 Wright MP. pH measurements with glass electrode. In: Woolmer RF, cd. A Symposium on pH
and Blood Gas Measurements. London: Churchill, 1959; 5-15.
15 Mattock G. Electrochemical aspects of blood pH
measurements. In: Wool mer RF, cd. A Symposium on pH and Blood Gas Measurements.
London: Churchill, 1959; 19-29.
16 Semplc SGH. Observed pH differences of blood
and plasma with different bridge solutions. J Appl
Physiol 1961; 16: 576-7.
17 Czaban JD. Cormier AD, Lcgg KD. Establishing the direct-potentiometric 'normal' range for
NalK: residual liquid junction potential and
activity coefficient effects. Clin Chem 1982; 28:
1936-45.
18 Winkelman JW. Merritt C, Scott WJ et al. Effects
of residual liquid junction potential in direct
potentiometry of potassium. Clin Chem 1984; 30:
482-4.
Accepted for publication 27 November 1987
Downloaded from acb.sagepub.com at PENNSYLVANIA STATE UNIV on May 11, 2016