CLIN.CHEM.

30/4,553-556 (1984)

Effects of Storage Temperature and Time before Centrifugation on Ionized

Calcium in Blood Collected in Plain Vacutainer Tubes and Silicone-Separator
(SST) Tubes
John Toffaietti,2 Nancy Biosser,2and Kathryn Klrvan3
We studied the stability of ionized calcium and pH in samples
stored at either room temperature or 4 #{176}C,
in centrifuged and
uncentrifuged blood-collection tubes and in centrifuged tubes
containing a silicone-separator
gel (SST tubes).At room
temperature,inuncentrifuged blood from healthy individuals,
mean ionized calcium usually increased no more than 10
mol/L per hour; at 4 #{176}C
it did not change detectably for 70 h.
This stability
was fortuitous,
however: the concentrations
of
both hydrogen and lactate ions in these samples increased,
apparently with offsetting effects on the concentration of
ionized calcium. Blood stored for 70 h at 4 #{176}C
in centrifuged
SST tubes, although showing a slightly greater change in
ionized calcium, had less change of pH and no change in the
ionized calcium corrected to pH 7.4. In 11 heparinized wholeblood samples from eight patients in intensive care, the mean
change per hour in ionized calcium and pH after storage at
room temperature was + 10 mol/L and -0.04 units, respectively.
AdditionalKeyphrases:pH
sample handling

lactate
variation, source of
ion-selective electrodes
ionized calci-

um vs total calcium

One of the subcommittees of the AACC Ionized Calcium

Working Group was formed to gather information and
formulate guidelines for collection of samples for measuring
concentrations
of ionized calcium; a preliminary
report has
recently appeared (1). Here, we report our study of the
effects of temperature
of storage, delay in cell separation,
and type of blood-collection
tube on results for ionized
calcium, which may be useful in developing such guidelines.
We have observed little change in dialyzable calcium in
serum from blood left to clot for as long as 6 h at room
temperature
(2), which is longer than usual for processing
samples for determinations of ionized calcium (3-5). FoghAndersen et al. (6) have also reported that clotted blood may
be stored at 0-4 #{176}C
for as long as 4 h without affecting
results for ionized calcium. The most rapid analyses require
the use of heparinized whole blood or plasma, but calcium
binding by heparin and problems with protein adherence to
electrodes are drawbacks to procedures involving heparin
(7).

By measuring the changes in ionized calcium, pH, and

related constituents,
we wanted to determine whether lessrapid collection and handling procedures would still give
satisfactory results for ionized calcium in serum. We measured ionized calcium, pH, and in some cases dialyzable
calcium, lactate, and phosphate in blood samples stored
without separation from the clot for as long as 24 h at room
temperature,
and in samples stored at 4#{176}C
for as long as 70

Department of Pathology, 2 Clinical Chemistry Laboratory, and

Blood Gas Laboratory, Duke University Medical Center, Durham,

NC 27710.

Received July 25, 1983; accepted January 18, 1984.

h with or without separation from the clot in both plain and

silicone-separator
tubes. We then evaluated the potential
value of, and precautions
for, using routinely
collected
samples for measurement of ionized calcium.

Materials and Methods

Ionized calcium, pH, and ionized calcium corrected to pH
7.4 were measured simultaneously with a Radiometer ICA 1
ionized calcium analyzer (Radiometer America Inc., Westlake, OH 44145). Dialyzable calcium was measured
by
continuous-flow analysis (1), and lactate (8) and phosphate
(9) by centrifugal
analysis.
Measurements
of ionized calcium, pH, and dialyzable calcium were done in duplicate.
Blood samples from apparently healthy volunteers, ages
23-57 years, were collected in large syringes, then injected
into evacuated blood-collection
tubes containing either no
additive (plain Vacutainer Tubes) or a silicone gel for
maintaining
separation of serum from erythrocytes after
centrifugation
(SST), both from Becton Dickinson, Rutherford, NJ 07070. For anticoagulants
we used either sodium
heparin (Becton Dickinson) or a calcium-titrated
heparin
solution (no. S 4500, Radiometer). Samples were stored at
room temperature
(22-25 #{176}C)
or 4 #{176}C,
as noted.
Eleven whole-blood samples, collected in heparinized syringes (no. MQ6O1 or MQ6O5LD; Marquest Medical Products, Inc., Englewood, CO 80112), containing 30 or 40 mt.
units of lithium heparin per milliliter, were selected without
conscious bias from eight patients who were either in
intensive care or undergoing surgery. These samples were
kept on ice before the initial analysis of the whole blood,
then stored at room temperature for 2 h before re-analysis.

Results
Table 1 shows the mean, SD, and range of changes in
ionized calcium, corrected ionized calcium, and pH in plasma and serum from blood samples that were left uncentrifuged for as long as 3 h. In both plasma and serum the mean
change and most of the individual changes of ionized
calcium were within the imprecision of the method, about 20
molfL.
One sample consistently
had a greater change,
increasing by 40 to 55 mol/L in 1 to 3 h.
With longer storage before centrifugation,
as shown in
Table 2, ionized calcium in blood from four of the five donors
changed by no more than 20 molJL after 4 h, although
blood from the other donor changed by 80 mo1!L at 4 h. By
6 h, the ionized calcium in all the samples increased by 50
mo1JL or more. The corrected ionized calcium, however, did
not change detectably in plasma or serum.
In general, the pH of the serum or plasma samples
changed by about 0.01 unit per hour of storage at room
temperature; lactate, in the study shown in Table 2, increased by two- to threefold during 6 h.
Although not shown in Table 1, 28 mt. units of sodium
heparin per milliliter decreased both ionized and dialyzable
calcium by about 100 &moI/L from their concentrations in
serum. This indicates that the effect of heparin was not an
CLINICAL CHEMISTRY, Vol. 30, No. 4, 1984

553

Table 1. Effect of Time before Centrifugation on Changes in Ca2, Ca27.4, and pH in Blood Samples
Stored at Room Temperature
aCa2,

h before

centrlf.

Mean

iimol,L

4tCa2

SD

Range

Plasma containing sodium heparin (28

umol/L

oH

Mean

SD

Range

Mean

2
23
5
-5

22

-0.01

15
27

-20 to40
0 to 40
-15 to 25
-40 to 20

8
-15
-19

11
30
30

-5 to20
-50 to 25
-60to15

-0.02
-0.03
-0.05

SD

mt. units/mL)a

0.5
1.0

5
27

23
26

-15to45
-5to55

2.0
3.0
SerumL
1.0

18
20

16
25

Oto4O
-lOtoSO

33

2.0
3.0

8
20

18
33
25

10to45
-35 to 45
-45 to 40

18

-0.01

0.005
0.003
0.004

-0.02
-0.04

0.010
0.007

0.017

a Changesare relative
toconcentrations
inplasma

from blood centrifugedwithin5 mm after collection.

bchangesare relativeto concentrations in serum from blood centrifuged within30 mm aftercollection.

5 donors.

Table 2. Effect of Time before Centrifugation on Changes in Ca2, Ca27.4, pH, and Lactate in Blood
Samples Stored at Room Temperature
aCa2, pmol/L
Ca2,4,
h before
centrif.
Mean
SD
Range
Mean
SD
Plasma containing calcium-titrated heparin (22 mt. units/L)#{176}

2
4
6

24

20
35

19
30

65

20

157

64

5to45
20 to 80
50 to 95
110to230

8
45

-130

imoI/L

Mactate,
mmol/L

apH

Range

Mean

SD

Mean

SD

-5tolO
-5 to 10
-10 to10
-90 to -180

-0.04

0.024

1.5

0.1

-0.06

0.049

-0.10
-0.37

0.041
0.150

2.5
3.6
10.5

0.5
0.3
2.2

0 to20
-15 to -10
-15 to 20

-0.02
-0.04
-0.06

0.011
0.010
0.015

1.5
2.2
2.7

0.6
0.8
0.7

-4Oto-65

-0.14

0.039

5.4

1.1

Serumb

2
23
12
4
19
13
6
36
16
24
72
11
Footnotes and n as in Table 1.

15 to40
5 to 35
25 to 60
65to85

9
-11
-1
-56

artifact of methodology. The use of calcium-titrated sodium

heparin, 22 mt. units/mL, eliminated any detectable difference in ionized or dialyzable calcium between serum and
plasma.
We also measured dialyzable calcium in the samples
represented in Tables 1 and 2. Up to 4 h, dialyzable calcium
changed by no more than 30 JLmolJL in any sample; the
change after 6 h was no more than 50 mol/L, consistent
with an earlier report (2).
To determine the stability of ionized calcium and pH at
4#{176}C,
we collected blood samples from four volunteers into
large syringes, injected the samples into both plain Vacutamer Tubes and SST tubes, let them clot for 0.5 h,
centrifuged, then stored the unopened tubes at 4#{176}C
for 6 h.
As Table 3 shows, serum separated from cells in either plain
tubes or SST tubes underwent no detectable changes in
ionized calcium, corrected ionized calcium, pH, or lactate for
6 h. However, there was consistently a bias between similar
samples from plain tubes and SST tubes.
Of 12 blood samples simultaneously
injected from a large
syringe into a plain tube and an SST tube, then allowed to
clot for 30 mm, ionized calcium was always slightly higher
in samples from SST tubes, as shown in Table 4. The higher
concentrations of lactate in samples from SST tubes appear
to account for these changes. The corrected ionized calcium
was the same in samples from plain tubes or SST tubes.
After the 6-h storage study, we wanted to determine how
long samples
could be stored at 4#{176}C
before changes in
ionized calcium or pH became significant. Sets of samples
were collected from each of six healthy volunteers and left to
clot for 20 mm at 4 #{176}C.
Some were then analyzed for baseline
values of ionized calcium and pH; others, kept at 4#{176}C
until
554 CLINICALCHEMISTRY, Vol. 30, No. 4, 1984

9
3
17
14

Table 3. Changes in Serum Stored for 6 h at 4 #{176}C

after Centrifugatlon a
Mean
Plain tubes
ACa2, imol/L

pH
lactate, mmol/L
iCa274, mol/L

SSTtubes

-2
-0.01

0.0
-8

SD

Range

7
0.003

-lOto5
-0.005 to -0.01
-0.1 to 0.1

0.1
10

-2OtoO

Ca2, .mol/L
2
4
Oto7
-0.01
0.003
-0.01 to -0.015
ApH
Alactate,mmol/L
0.0
0.2
-0.2 to 0.1
Ca274, Lmol/L
-3
4
-5toO
a Results are from four volunteers. Blood was allowed toclotfor30 mm at
room temperaturebeforecentrifugation.
Samples were analyzedimmediately
and 6 h aftercentrifugation.

analysis, were divided into three groups: (a) uncentrifuged

blood left in plain tubes for 24 or 70 h, (b) centrifuged blood
left in plain tubes for 24 or 70 h, and (c) centrifuged blood
left in SST tubes for 24 or 70 h before analysis. As Table 5
shows, ionized calcium was relatively stable at 4#{176}C
under
all these conditions but most stable in the uncentrifuged
tubes. The pH was altered most in the uncentrifuged
tubes,
and lactate, measured in one set of uncentrifuged samples,
increased by almost threefold (data not shown). Values for
pH and ionized calcium corrected to pH 7.4 varied least for
the blood stored in the SST tubes.
In addition to samples from healthy persons (Tables 1-5),
we also studied 11 heparinized
whole-blood samples from
eight patients in intensive care or undergoing surgical
procedures. As shown in Table 6, the changes in ionized

Table 4. Differences in Ca2, pH, and Lactate between Serum from Plain Tubes and from SST Tubes1
Plain tubes
Mean
1250
1270
7.44

Ca2, hmol/L
Ca274, rnol/L
pH
Lactate,mmol/L

SST tubes
SD

40
40
0.04

1.76

Mean

DIfference

SD
40
30
0.03

1270
1270
7.41
1.92

Mean

SD

23
6
-0.025
0.21

9
7
0.012
0.16

0.43
0.4.3
Blood allowed to clot for 30 mm at room temperature before centritugationand analysis.
n = 12 each.

Range

10 to 40
0 to 20
-0.01 to 0.04
0 to 0.4

Table 5. Effect of Stora ge at 4#{176}C

on Values for Ca2, pH, and Ca2+74a
PlaIn tubes, uncentrifuged

Storage,

Mean

tCa,

SD

Plain tubes, centrifuged

Range

Mean

SD

SST tubes, centrIfuged

Range

Mean

SD

Range

pmolL

24
70

2
0

ispH
24
70

8
16

-0.04
-0.08

iCa74,

mo1/L

24
70

-21
-53

0.02
0.04
15
18

-10 to10
-20 to20
0 to -0.06
-0.01 to -0.13
Oto-40

-20 to-70

16
21

18
12

-0.03
-0.08

-20 to 30
10 to 40

0.02
0.03

-32

0.01 to -0.06
-0.03 to -0.12

-lOtolO

18

-10 to

25
30

8
12

-0.04
-0.05

0.01
0.02

0
-60

-2

20 to 40
10 to 40
-0.02 to 0.06
-0.04 to -0.08

-lOtoO

-10 toO

Results are from six different healthy individuals.

Table 6. Effect of Storage at Room Temperature on Ca2 and pH in Uncentrifuged Heparinized Blood
from Patients In intensive Care1
niolIL

Moan
SD
Range
=

pH

O.2h

2.2h

O.2h

2.2 h

820

840

24

7.32

7.28

340

330

19

0.17

0.16

0 to 50

7.00-7.55

6.96-7.50

0.29-1.16

1.34-1.16

-0.039
0.021

0 to -0.08

11 samples from eight patients.

calcium in any sample after 2 h of storage at room temperature ranged from 0 to 50 (mean 24) pmol/L, and the changes
in pH ranged from 0 to -0.08 (mean 0.04) pH unit, both
ranges slightly greater than those seen in the healthy
individuals.

Discussion
Among the clinical
monitoring
of patients

uses of calcium measurements

are
in surgery or intensive care, daily
monitoring of patients to maintain appropriate concentrations of calcium, and determinations
for purposes of diagnosis. Although special collection and handling procedures for
ionized calcium may be economically justified for the relatively few patients in intensive care, they may be impractical for measurements on large numbers of other patients.
Given the large number of total calcium measurements
requested, more efficient collection and handling procedures
are necessary if ionized calcium becomes the predominant
clinical measurement for calcium.
The use of ordinary sodium heparin in plasma or whole
blood should be discouraged, because it decreases the values
for ionized calcium. Calcium-titrated
sodium heparin would
be a better choice when fast analyses are needed; furthermore, results obtained on samples so treated should agree
with those from serum, the sample type used for most other
chemistry assays and which reportedly gives fewer problems with maintenance
of the calcium electrodes (7) than
does plasma or whole blood.
Blood may remain uncentrifuged
at room temperature
for
as long as 2 h or at 4#{176}C
for 24 to 70 h without appreciable
change in the concentration of ionized calcium; use of SST
tubes to keep serum and cells separated after centrifugation

gives the least change in pH and corrected ionized calcium

after 70 h of storage at 4#{176}C.
Nonetheless, we consistently
observed a 20 mo1/L bias in ionized calcium between serum
from plain tubes and that from SST tubes (Table 1), apparently because of the lower initial pH in samples from the
SST tubes. As shown in Table 3, some heparinized wholeblood samples from patients in intensive care had slightly
greater changes in calcium and pH after storage at room
temperature
than did samples from healthy individuals,
although the maximum change in ionized calcium for any
sample was only 25 umo1/L per hour.
Our data from healthy volunteers
and hospitalized
patients indicate that ionized calcium in serum is stable,
especially if stored at 4#{176}C,
whether the blood is centrifuged
or not. However, we emphasize several cautions: First,
aerobic handling is not acceptable unless other means of pH
correction,
perhaps such as that incorporated into the Radiometer ICA 1 display, are used. Second, because most of our
studies involved healthy donors, we do not rule out the
possibility that samples from some patients may undergo
changes
of pH that are not compatible with delays in
processing. Finally, this apparent stability of ionizedcalcium seems to be a dynamic process, such that the hydrogen
cation and the lactate anion from the lactic acid offset the
effect of the other, respectively releasing and chelating
ionized calcium. Although doubling or tripling the concentration of lactate in samples stored at 4#{176}C
did not appreciably affect our measurements
of ionized calcium, this may
have been fortuitous for the samples we studied.
If total calcium remains the calcium measurement
in
greatest use, then special handling of a few samples for
determinations
of ionized calcium should present few probCLINICAL CHEMISTRY, Vol. 30, No.4, 1984

555

lems. However, as instrumentation

for ionized calcium
continues to improve, measurements of ionized calcium may
eventually displace total calcium, and more efficient procedures for blood collection and handling
will be needed.
Further work is needed to confirm conclusively that storage
of blood at room temperature for 2 h or at 4 #{176}C
for longer
periods does not alter ionized calcium values by more than
about 2%.

References
1. Graham G, Burritt M. Preliminary report: AACC Ionized Calcium Working Group on Reference Intervals. Clin Chem 29, 1187
(1980). Abstract.
2. Toffaletti J, Kirvan K. Spectrophotometric
micro method for

measurement of dialyzable calcium by use of cresolphthalein complexone and continuous-flow analysis. Clin Chem 26,1562-1565
(1980).
3. Ladenson J, Bowers GN. Free calciumin serum. I. Determina-

tion with the ion-specific electrode, and factors affecting the results.
Clin Chem 19, 565-574 (1973).
4. Husdan H, L.eung M, Oreopoulos D, Rapaport A. Measurement
of serum and plasma ionic calcium with the Space-Stat 20 Ionized
Calcium Analyzer. Clin Chem 23, 1175-1777 (1977).
5. Fogh-Andersen N. Ionized calcium analyzer with built-in pH
correction. Clin Chem 27, 1264-1267 (1981).
6. Fogh-Andersen N, Christiansen TF, Komarmy L, SiggaardAndersen 0. Measurement of free calcium ion in capillary blood and
serum. Clin Chem 24, 1545-1552 (1978).
7. Larsson L, Finnstrom U, Nilsson B, Ohman S. Evaluation of
Radiometer ICA 1 as a routine instrument
for serum ionized
calcium and its application for whole blood capillary samples from
newborn infants. Scand J Clin Lab Invest 43 (Suppl 165),21-26

(1983).
8. Pesce MA, Bodourian SH, Nicholson JF. Rapid kinetic measurement of lactate in plasma with a centrifugal analyzer. Gun Chem
21,1932-1934 (1975).
9. Wentz PE, Savory J, Cross RE. Improved method of measurement of inorganic phosphate in serum with a centrifugal analyzer.
Clin Chem 22, 257-260 (1976).

CLIN.CHEM. 30/4,556-559(1984)

Urea, Creatinine, and Glucose Determined in Plasma and Whole Blood by a

Differential pH Technique
M. Ripamonti,

A. Mosca, E. Rovida, M. Luzzana, L Luzi, F. Ceriotti, F. Cottini, and L. Rossi-Bernardi

We report the conditions (buffer composition and enzyme

activity) required for estimating three frequently determined
analytes-urea, glucose, and creatinine-by use of an improved version of the differential pH apparatus previously
described (C/in Chem 29: 80-85, 1983). For each analyte,
we used only one specific enzyme, thus avoiding a chain of
auxiliary and indicator reactions. The method requires about
a minute for each determination in undiluted plasma or whole
blood.
We previously described (1,2) a new instrument, based on
the differential measurement
of pH, and illustrated
its
application by the determination of glucose in plasma. We
believe that this pH technique can find interesting applications in clinical chemistry. Indeed, the two more interesting
advantages are that (a) the system automatically performs a
sample blank and (b) turbid solutions can be analyzed
directly. Moreover, many reactions lead to a change of pH in
the solution. The most relevant among them are: the
oxidoreductase-catalyzed
reactions involving NAD/NADH
or NADP/NADPH
interconversion; the reactions involving
transfer of phosphate residue; reactions producing CO2 and
NH3; and reactions involving ester bond synthesis or cleavage. Clearly, a potentially wide range of analytes and
enzyme activities
can be detected by this technique.
We now describe our further refinement of the technique
and its extension to determination of urea, creatinine, and
glucose in plasma and whole blood.

Dipartimento
di Scienze e Tecnologie Biomediche, Centro di
Fisiologia del Lavoro Muscolare del CNR, do Ospedale S. Raffaele,
University of Milan, Via Olgettina 60, 20132 Milan, Italy.
Received September 12, 1983; accepted January 5, 1984.

556 CLINICALCHEMISTRY, Vol.30, No. 4, 1984

Materials and Methods

Instrumentation.
pH was measured with a commercially
available differential pH analyzer (Delpas CL; Kontron AG,
Milan, Italy), an instrument similar in design to the homemade instrument
we previously described (2). Improved
features consist of a thermostated
stainless-steel
block to
ensure a better temperature control of the two electrodes, a
16-character alphanumeric display, a printer, and a keyboard for parameter entry and instrument control. All
measurements here reported were obtained by use of this
apparatus at 23 #{176}C
except for creatinine, which was measured at 37 #{176}C.
An IL919 analyzer (Instrumentation Laboratory, Milan, Italy) and IL-associated
chemical procedures
were used as comparison methods for the determination of
urea and creatinine. Computations were performed by an
Altos ACS 8000/2 (Altos Computer Systems, Cuppertino,
CA). The software was written in PASCAL high-level language.
Chemicals
and solutions.
Lyophilized jack-bean urease
(EC 3.5.1.5) was from Ames/Miles (Milan). A solution of it
was prepared by dissolving a suitable amount in 10 mL of
0.1 mol/L KC1 and adjusting the pH to 7.50 with 0.1 mol/L
K2C03; 15 mL of glycerol was added as a stabilizing agent.
The final activity was 1440 kUIL. For each urea determination, we routinely used 11.2 U of urease. Typically, the
measurement was taken 20 s after the enzyme was added.
No significant
loss of urease activity in solution was detected after six months at 4 #{176}C,
but a 12% loss of activity was
observed after four days at room temperature.
Creatinine (99% pure) and creatinine iminohydrolase (EC
3.5.4.21, microbial source) were from Farmitalia Carlo
Erba, Milan. The enzyme suspension was dialyzed against
0.1 molIL KC1 solution containing NaN3, 1 g/L. The pH was
then adjusted to 7.54 with 10 mmolIL NaOH, with continu-