Vol. 45, No. 4
Printed in U.S.A.
T H E AMERICAN JOURNAL OF CLINICAL PATHOLOGY
Copyright © 1966 by The Williams & Wilkins Co.
THE MEASUREMENT OF C0 2 CONTENT WITH THE AUTOANALYZER
A COMPARISON WJTH 3 STANDARD METHODS AND A DESCRIPTION OF A N E W METHOD
(ALKALINIZATION) FOR PREVENTING LOSS OF C0 2 FROM OPEN CUPS
S. R A Y M O N D G A M B I N O , M . D . , AND H A R R Y S C H R E I B E R
Department
of Pathology, Englewood Hospital,
The AutoAnalyzer method for determining total CO2 content11 is widely used,
but no detailed validation of this method has
been reported since the original publication
by Skeggs.11 We have compared the AutoAnalyzer method with 3 other standard
methods: Microgasometer,7 Astrup apparatus, 1 and calculation from directly measured
pH and Pc02-6, 6 The comparisons were
made by sampling directly from freshly
opened Vacutainer tubes, and our findings
are reported in this paper.
In the standard AutoAnalyzer method,
however, plasma or serum is transferred to
open plastic cups prior to sequential
sampling. Plasma or serum exposed to air
in open cups or tubes loses C0 2 . Loss of
C0 2 may be decreased or prevented with
mineral oil,4 plastic disks,2 or stoppers,8 but
none of these is ideally suitable for use with
the AutoAnalyzer. Mineral oil gets into the
tubing and coil system. Plastic disks do not
prevent loss; they only decrease it, and they
are difficult to position. Stoppers can not
be penetrated by the sampler probe. Therefore, a new method is needed for preventing
loss of C 0 2 from open AutoAnalyzer cups.
The rate and amount of loss of C 0 2 is
proportional to PCO2, and if plasma PCO2
equals room air PCO2, no gas exchange
occurs. Plasma PCO2 can be lowered by
alkalinization. Plasma was alkalinized to a
pH between S.6 and S.S. At this pH, the
calculated Pco 2 approximates that of room,
air (not standard atmospheric air) and no
significant gain or loss of C0 2 should occur.
CO2 loss from samples with and without
added alkali, and from samples covered by
plastic disks, was measured with the
AutoAnalyzer and the results are reported in
this paper.
Received, August 23, 1965.
Engleiuood, New
Jersey
MATERIALS AND EQUIPMENT
1. Heparinized Vacutainer tubes for collection
of
plasma
samples
(Becton
Dickinson No. 3204 K).*
2. Ammonium hydroxide, 1 N.
3. Distilled water.
4. diSPo transfer pipets (S.P. No. P
5205-1). f
5. Technicon AutoAnalyzer.!
6. Microgasometer. §
7. Astrup tonometer (Radiometer No.
AMT-1).«,I
8. pH/Pc02 electrode and meter (Instrumentation Laboratory No. 113-Sl).||
9. Calibrated gases containing 5.00 and
10.00 per cent CO2 in oxygen. ||
METHODS
Blood specimens from hospital patients
were drawn into heparinized Vacutainer
tubes,3 and all measurements were made on
plasma. The standard AutoAnalyzer method
for CO2 content11 was validated by being
compared with 3 other standard methods for
CO2 content: Microgasometer,7 Astrup, 1, 9
and calculation from pH and directly measured Pco2.8, 6 ' 1 0 In each comparison, measurements were made no later than 1 min.
after the stopper was removed from the full
Vacutainer tube.
Similar specimens were used to study C 0 2
loss from AutoAnalyzer cups. The C 0 2
* Becton Dickinson and Co., Rutherford, New
Jersey.
t American Hospital Supply Corp., E v a n s t o n ,
Illinois.
X Technicon Instruments Co., Chauncey, New
York.
§ Scientific Industries Co., Queens Village, New
York.
*,[ Radiometer, Copenhagen, Denmark.
|| Instrumentation Laboratory, Boston, Massachusetts.
406
April 1966
4.07
M E A S U R E M E N T O F CO2 C O N T E N T
content of plasma samples in alkalinized,
non-alkalinized, and disk-covered AutoAnalyzer cups was measured before and
after exposure to air.
Plasma samples were alkalinized by 1 drop
(0.035 ml.) of 1 N ammonium hydroxide
added to an empty AutoAnalyzer cup with a
disposable transfer pipet. One milliliter of
plasma was then transferred from a full
Vacutainer tube to the cup with a similar
disposable transfer pipet. A drop of alkali
was also added to all standards. All aspiration and expulsion of plasma was performed
gently. No bubble formation or foaming was
permitted and the alkali and plasma were
mixed by gentle swirling with the tip of the
transfer pipet. Cups containing no alkali and
cups Avhich were to be covered with plastic
disks were filled in a similar manner. The
transfer of plasma from the Vacutainer tube
to the open cup was validated by comparison
of measurements of CO2 content of plasma
aspirated directly from the Vacutainer tube
before transfer with values obtained from the
cup immediately after transfer.
HESULTS
Validation of AutoAnalyzer CO2 content
values by comparison luith 3 standard methods.
Heparinized plasma in full Vacutainer tubes
was obtained from patients and the C0 2
content was measured at the moment the
stopper was removed. Values obtained with
the AutoAnalyzer were compared with
values obtained with a Microgasometer, an
Astrup apparatus, and a pH/Pco 2 electrode.
Table 1 is a comparison of the results of the
AutoAnalyzer and the Microgasometer,
Table 2 the AutoAnalyzer and the Astrup
apparatus, and Table 3 the AutoAnalyzer
and a pH/Pco 2 electrode. No significant
differences were found. In each comparison,
the AutoAnalyzer probe was placed inside
the Vacutainer tube as soon as the stopper
was removed.
Validation of transfer step from the Vacutainer tube to an open AutoAnalyzer cup,
with and without alkali. The CO2 content of
plasma was measured with the AutoAnalyzer and samples were aspirated in 3
different ways: (1) direct aspiration from the
TABLE 1
COMPARISON OF T O T A L C 0 2 C O N T E N T M E A S U R E D
WITH TUB AUTOANALYZER AND TUB
MlCnOGASOMETEH*
Mean
AutoAnalyzer
Microgasometer
mM /liter
mM/liter
10.5
17.0
21.5
21.5
25.0
25.5
27.0
31.0
34.5
41.0
11.0
17.0
21.9
21.5
25.1
25.5
27.5
30.5
35.0
41.5
25.45
25.05
* Fifty-four additional paired comparisons
were made 4 years previously. The mean value for
the AutoAnalyzer was 20.80 mM per liter, and
for the Microgasometer 26.84 mM per liter. T h e
maximal difference (AutoAnalyzer minus Microgasometer) ranged from —1.4 to + 1.5 mM per
liter, and the S.D. of the difference was ±0.47 mM
per liter. T h e S.D. of the AutoAnalyzer method
was ± 0.28 mM per liter, and the S.D. of the M i crogasometer method was ± 0.71 mM per liter.
Vacutainer tube, (2) aspiration from an open
AutoAnalyzer cup immediately after transfer
of 1 ml. of plasma to a cup containing 1 drop
of 1 N ammonium hydroxide, and (3)
aspiration from an open AutoAnalyzer cup
immediately after transfer of 1 ml. of plasma
to a cup containing 1 drop of normal saline
solution. The results of this comparison are
listed in Table 4. There is no or only slight
C 0 2 loss when plasma is transferred from the
Vacutainer tube to a cup containing alkali.
A larger, but still clinically insignificant,
loss of C 0 2 occurs when plasma is transferred
to a cup containing 1 drop of saline solution.
Storage stability of CO 2 content in open
AutoAnalyzer cups, with and luitlwut alkali.
With the AutoAnalyzer, 15 plasma samples
were assayed for C 0 2 content. Measurements
were made under 3 conditions: (1) immediately, (2) after 1 hr. in an open cup containing 1 drop of 1 N ammonium hydroxide,
and (3) after 1 hr. in an open cup without
408
TABLE 2
COMPARISON OP T O T A L C 0 2 C O N T E N T M E A S U R E D
WITH T H E AUTOANALYZER AND T H E
ASTRUP APPARATUS
AutoAnalyzer
Astrup
mil/liter
mM/liicr
Mean
18.5
20.0
23.0
23.5
25.5
28.0
3.1.5
31.5
35.5
42.0
18.8
21.0
23.0
23.0
26.3
28.5
30.4
31.6
35.0
42.0
27.90
27.96
TABLE 3
COMPARISON QF T O T A L C 0 2 C O N T E N T M E A S U R E D
WITH
T H E A U T O A N A L Y Z E R AND BY CALCULA-
TION
FROM
DIRECTLY
MEASURED
PH
AND
Pco,
Mean
Vol. 45
GAMBINO AND SCHREIBEH
AutoAnalyzer
pH/PcOj Electrode
mM/liUr
mM/liter
20.5
20.5
26.0
27.0
28.0
29.0
29.0
29.5
30.5
32.0
32.0
33.0
33.5
33.5
41.0
20.0
20.5
25.0
27.0
27.5
28.0
29.0
29.5
31.5
30.0
31.0
33.5
32.0
33.5
39.0
29.67
29.13
alkali. The comparison of these results is
summarized in Table 5.
The alkalinized samples, after 1 hr. of air
exposure, had not lost or gained a clinically
significant amount of C02- The maximal
difference (immediate minus 1 hr.) ranged
from — 1 to + 1 mM per liter, and the mean
difference was only —0.1 mM per liter.
Seven of the 15 samples showed no difference
whatsoever. On the other hand, all samples
without alkali showed a clinically significant
loss of C0 2 . The range of the loss was from
— 1.5 to —6.5 mM per liter, and the mean
loss was —3.53 mM per liter.
The effectiveness of 1 N NH£)H in controlling loss of C02 for 4 hr. One milliliter
of plasma was added to 1 drop of 1 N ammonium hydroxide in an open AutoAnalyzer
cup and the initial reading of 10 different
samples was compared with the reading
obtained 4 hr. later. The results in Table 6
confirm the effectiveness of alkalinization in
the prevention of loss of C0 2 .
The time course of loss of COz in open
AutoAnalyzer cups without alkali. One-milliliter aliquots of a plasma with a high PCO2
(75 mm. Hg) were placed in each of 7
AutoAnalyzer cups. An immediate measurement of C 0 2 content was made directly from
the Vacutainer tube and subsequent
measurements were made on samples
exposed to air in open cups for 10, 30, 60,
120, 240, and 360 min. Only 1 cup was used
TABLE 4
VALIDATION O F T R A N S F E R S T E P FROM VACUTAINER
TO
AUTOANALYZER
C U P BY COMPARISON
OF
T O T A L C 0 2 C O N T E N T IN THE VACUTAINER T U B E
AND IN C U P S WITH AND W I T H O U T A L K A L I I M MEDIATELY AFTER F I L L I N G
AutoAnalyzer
AutoAnalyzer
Directly from Vacuum Tube Cup with Alkali Cup, No Alkali
mM/liter
mil/liter
21.0
25.5
25.5
26.0
28.0
28.5
28.5
29.5
29.5
30.5
31.0
32.0
35.0
40.0
20.0
25.5
25.5
25.5
27.5
28.5
28.0
29.0
29.5
30.5
31.0
31.5
34.5
39.5
19.5
25.0
25.0
23.5
27.0
28.0
28.0
28.5
29.0
30.0
29.5
31.5
33.0
39.5
29.32
29.00
28.36
mM/liter
Mean
April 1966
409
M E A S U R E M E N T O F CO2 C O N T E N T
TABLE 5
COMPARISON OF STORAGE STABILITY OF PLASMA
TOTAL
C02
CUPS
CONTENT
WITH
IN
AUTOANALYZER
AND WITHOUT
ALKALI
1 hr. Later,
Alkali
Initial
1 hr. Later,
Mo Alkali
mM/litcr
mM/litcr
13.5
23.0
24.0
25.0
20.0
20.0
27.0
27.0
27.5
28.0
28.5
31.0
31.0
30.0
37.5
14.5
24.0
24.5
25.0
20.0
25.0
27.5
27.0
27.5
28.5
28.0
31.0
31.0
30.0
37.0
11.0
18.0
22.5
22.5
23.0
22.0
24.0
24.0
25.0
25.0
25.0
27.0
20.0
32.0
31.0
27.40
27.50
23.87
mAt/liter
60
Moan
Initial
4 hr. Later
mil/liter
mM/litcr
23.0
25.0
25.5
20.0
20.0
27.5
28.0
29.0
30.0
32.0
24.0
25.0
23.5
20.0
20.5
27.5
29.5
29.0
29.5
32.0
27.20
27.25
for each time period. Figure 1 illustrates the
results of this experiment.
The loss is greatest at the start, and after
60 min. of air exposure, almost all of the loss
that will occur has taken place. No additional loss occurs after 3 hr. of air exposure,
because the pH of the sample has risen to
360
' 360
TABLE 7
GAIN
IN
TOTAL
C02
CONTENT
WHEN
S A M P L E S A R E E X P O S E D TO ROOM A I R
CO2
AUTOANALYZER C U P S
'
T H E RELATION OF PLASMA P H AND PCO2 TO L O S S
pll
C O N T E N T FOR 4 H O U R S IN ALKALINIZED
240
MINUTES
Immediate COi
COi after 1-hr.
Air Exposure
Estimated PcOi
mM/liter
mM/litcr
mm. Ug
25.0
25.0
25.0
25.0
25.0
25.0
21.5
23.3
24.7
20.0
20.S
20.8
TABLE 0
STORAGE STABILITY OF PLASMA T O T A L
~ISo
F I G . 1. The time course of loss of CO2 in an open
AutoAnalyzer cup without alkali. T h e sample had
a high Pco» (75 mm. Pig) and it was purposely
selected to exaggerate the pattern of CO2 loss.
OR
Mean
120
7.33
7.95
8.50
9.00
9.20
9.30
50
12
3.5
1
0.7
0.5
TABLE 8
STORAGE STABILITY OF PLASMA T O T A L
C02
C O N T E N T WHEN C O V E R E D BY A PLASTIC
DISK
Mean
Initial
1 hr. with Disk
mM/litcr
mM/litcr
21.0
24.5
25.0
25.0
25.0
25.0
20.5
41.0
18.0
22.0
23.5
22.5
23.5
23.5
24.5
3S.5
20.03
24.50
410
Vol. 45
GAMBINO AND SCHREIBER
the range of approximately S.5 and the Pco 2
is near 1 mm. Hg.
The effect of varying pH on the gain or
loss of COi in exposed AutoAnalyzer cups.
Sample pH was varied by adding a single
drop of alkali of varying concentration to
1 ml. of plasma. The resultant pH was
measured at 37 C. and the CO2 content was
measured immediately and 1 hr. later. In
addition, PCO2 was estimated by the
Henderson-Hasselbalch equation. The results are listed in Table 7. These data
confirm the hypothesis that high Pco 2 is
associated with C 0 2 loss, and very low Pco 2
with C 0 2 gain. A Pco 2 between 1 and 2
mm. Hg yields the greatest stability.
Comparison of the relative effectiveness of
alkali versus disks in the prevention of loss of
C02. One-milliliter aliquots of plasma from
8 patients were placed in AutoAnalyzer cups
and a thin plastic disk2 was floated carefully
on the surface. These samples were exposed
to room air for 1 hr. and the C 0 2 content was
then measured with the AutoAnalyzer and
compared with assays made directly from the
Vacutainer tube before transfer. The results
are tabulated in Table S. The disk inhibits
but does not prevent C 0 2 loss.
DISCUSSION
The addition of 1 drop of 1 N ammonium
hydroxide to 1 ml. of plasma or serum
prevents clinically significant loss of C 0 2
from open cups for up to 4 hr. Loss of C 0 2 is
proportional to Pco 2 and alkalinization
lowers Pco 2 . One drop of 1 N ammonium
hydroxide was selected because this brings
normal plasma Pco 2 to the 1- to 2-mm. Hg
range. If Pco 2 is higher, C 0 2 will be lost to
room air, and if Pco 2 is lower, the sample
may absorb C0 2 from room air.
The concentration of C0 2 gas in room air
is higher than in standard atmospheric air,
which has a C0 2 concentration of 0.03 per
cent and a Pco 2 of 0.2 mm. Hg at sea level.
Room air, on the other hand, contains
expired air which raises the C 0 2 concentration to between 0.15 and 0.30 per cent and
the Pco 2 to 1 to 2 mm. Hg. When the Pco 2
of normal plasma was brought to 0.2 mm.
Hg, the C 0 2 content rose from 25 to 2S
mM per liter in 1 hr., because the Pco 2 of
room air is significantly higher than 0.2
mm. Hg.
Ammonium hydroxide, rather than sodium or potassium hydroxide, was chosen
because it does not interfere with most other
assays. We tested and found no interference
in assays for sodium, potassium, chloride,
urea, sugar, and bilirubin. Other assays were
not tested, but there is no theoretical reason
for interference with any other measurement
except pH, ammonia, and urea nitrogen with
urease.
Although mineral oil or plastic disks will
inhibit C 0 2 loss, mineral oil may contaminate the flow system of the AutoAnalyzer,
and plastic disks are difficult to place.
Furthermore, plastic disks do not afford
complete protection. Our findings with
plastic disks are not in agreement with the
recent report by Friedner and Philipson.2
Their study revealed no loss of C0 2 , whereas
our study showed a loss of between 1 and
3 mM per 1 liter in 1 hr. Friedner and Philipson did not study the C 0 2 loss occurring
during transfer to the AutoAnalyzer cup.
Inasmuch as 1 to 2 mM per liter can be lost
during transfer without alkali, and another
1 to 2 mM per liter can be lost during the
first 10 min. of air exposure, significant loss
of C0 2 could have occurred before the disk
was used, thereby decreasing the tendency
to lose C 0 2 after the disk was put in place.
Additional studies in our laboratory have
demonstrated that the disk works best when
the AutoAnalyzer cup is completely full. The
1-ml. aliquot of plasma used in our study
only fills the cup to the halfway mark.
Significant losses of CO2 from plasma or
serum exposed to air will occur unless
precaution is taken. If samples are not
measured within 10 min. of being placed in
an open cup, then one must take 1 of 3 steps:
(1) alkalinize, (2) inhibit diffusion with
plastic disks, or (3) inhibit diffusion with
mineral oil. We believe that alkalinization is
the preferable and superior precaution.
SUMMARY
The AutoAnalyzer method for the
measurement of C 0 2 content has been
validated by being compared with 3 other
standard methods. If proper precautions
April 1966
•
MEASUREMENT OF CO2 CONTENT
against loss of C0 2 are taken, the AutoAnalyzer C 0 2 method can be used for
precise and accurate assays of CO2 content.
Alkalinization of plasma by the addition of
1 drop (0.035 ml.) of 1 N ammonium
hydroxide to 1 ml. of plasma prevents
significant loss of C0 2 from plasma or
serum in open AutoAnalyzer cups.
Acknowledgment.
Dr. Ronald Maenza, senior
resident in clinical pathology, made the measurements utilizing the Astrup apparatus.
REFERENCES
1. Astrup, P . , J0rgensen, K., Siggaard-Andersen,
0 . , and Engel, K . : T h e acid-base metabolism. A new approach. Lancet, 1: 10351039, 1960.
2. Friedner, S., and Philipson, A.: A method for
retaining CO2 in serum in the AutoAnalyzer.
Scandinav. J. Clin. & L a b . Invest., 17: 185188, 1965.
3. Gambino, S. R.: Heparinized vacuum tubes
for determination of plasma p l l , plasma CO2
content, and blood oxygen saturation. With
an extensive discussion of pll methodology.
Am. J. Clin. P a t h . , 32: 285-293, 1959.
411
4. Gambino, S. R.: Mineral oil and carbon dioxide. Am. J. Clin. P a t h . , 35: 26S-269, 1961.
5. Gambino, S. R.: Determination of blood pCO».
Clin. Chem., 7: 336-345, 1961.
6. Gambino, S. R.: pIL and Pco 2 . In Meites, S.:
Standard Methods of Clinical Chemistry,
Vol. 5. New York: Academic Press, Inc.,
1965, p p . 169-198.
7. Natelson, S.: Microtechniques of Clinical
Chemistry, E d . 2. Springfield, 111.: Charles
C Thomas, Publisher, 1961, p. 152.
8. Paulsen, L.: Comparison between total COj
content (total CO2) in plasma/serum from
blood collected with or without parafline oil.
Scandinav. J . Clin. & L a b . Invest., 9: 402405, 1957.
9. Siggaard-Andersen, O.: The p l l , log p C 0 2
blood acid-base nomogram revised. Scandinav. J. Clin. & L a b . Invest., 1J,: 59S-604,
1962.
10. Siggaard-Andersen, O.: Blood acid-base alignment nomogram. Scales for pIT, PCO2, base
excess of whole blood of different hemoglobin concentrations, plasma bicarbonate,
and
plasma total-COj. Scandinav.
J.
Clin. & L a b . Invest., 15: 211-217, 1963.
11. Skeggs, L.: An automatic method for the determination of carbon dioxide in blood
plasma. Am. J. Clin. P a t h . , S3: 181-185,
1960.