Scandinavian Journal of Clinical & Laboratory Investigation

Vol. 69, No. 5, September 2009, 585–591

ORIGINAL ARTICLE

Potential preanalytical errors in whole-blood analysis: Effect of

syringe sample volume on blood gas, electrolyte and lactate values

PIRJO HEDBERG1,∗, AILA MAJAVA1,∗, KAI KIVILUOMA2 & PASI OHTONEN3

1Laboratory, Oulu University Hospital, Department of Clinical Chemistry, University of Oulu, 2Department of
Anesthesiology, Oulu University Hospital, and 3Department of Anesthesiology and Surgery, Oulu University Hospital,
Scand J Clin Lab Invest Downloaded from informahealthcare.com by CDL-UC Santa Cruz on 10/26/14

Oulu, Finland

Abstract
Background: Arterial blood samples are sensitive to bias because of the physiological properties of blood. Several errors
can occur in the preanalytical phase leading to incorrect diagnosis and improper treatment of patients. Collection of a blood
specimen, as well as its handling and transport, belong to the key factors to affect the accuracy and good quality of clinical
laboratory analysis. Methods: The aim of this study was to validate the effect of different sample volumes on the blood
gas, electrolyte and lactate values using 3 mL Rapidlyte plastic syringes with filter cap and Rapidlab 865 blood gas analyser.
Also, the stability of blood gas analyser parameters with different sample volume was studied. Results: No substantial
change in blood gas, electrolyte and lactate parameters was found when the results of 3 mL, 1.8 mL sample volumes in the
For personal use only.

3 mL syringes were compared. The sample volume of 1.0 mL or 1.5 mL in the 3 mL syringe is not suitable for the measure-
ment of oxygen tension, especially when accurate results of pO2 and arterial blood is needed for patient’s diagnosis.
Conclusions: The minimum sample volume when blood gases, electrolytes and lactate are all measured with the
Rapidlab system should be 1.8 mL using 3 mL Rapidlyte plastic syringe with filtercap. According to this study  1.8 mL
sample volumes can provide inaccurate results and can impose biases on measurements.

Key Words: Blood gas analyser, preanalytical, sample volume, syringe, whole-blood analysis

Introduction
Blood-gas analysis is widely used in the diagnosis of
diseases and clinical follow-up. Alterations in values
give a clue to the nature of the disturbance, if the
primary problem is in gas changes or metabolism.
Successful treatment can be seen as returning the
values towards normal.
Blood-gas analysis has pronounced value in
diagnostics and follow-up in intensive care treatment.
Successful ventilator treatment is based on reliable
blood-gas analysis. Trends in blood-gas analysis give
the basis for the follow-up of life-threatening situations
like sepsis.
In diagnostics it is usually important to know
the level, where the values are. In follow-up it is
obligatory to see reliable changes in absolute values.
It is the responsibility of health organizations to guar-
antee a high level of healthcare by using adequate
methodologies and instruments, creating secure
conditions for treatment, and preventing adverse
events to human or system errors [1].
The preanalytical phase plays a major role in
the source of errors in whole-blood analysis.
Arterial blood sample is sensitive to bias because
blood physiologically changes after it leaves
the body [2,3].
The arterial specimen collection and analysis
guidelines published by the Clinical and Laboratory
Standards Institute (CLSI; Previous NCCLS; National
Committee for Clinical Laboratory Standards) has
provided specific recommendations regarding speci-
men collection devices, sample handling, specimen
transport and storage conditions [4,5]. Arterial
specimens should be collected in a plastic syringe, left
at room temperature and analysed within 30 minutes.
Blood collected for special analyses, like shunt analy-
ses, should be analysed within 5 minutes. Actually,
over the past several years the practice has led to a

∗P.Hedberg and A. Majava contributed equally to this work.

Correspondence: Pirjo Hedberg, PhD, Laboratory, Oulu University Hospital, Oulu, PO Box 500, FI-90029 OYS, Finland. Tel: 358 8 3155453.
Fax: 358 8 5134409. E-mail: pirjo.hedberg@ppshp.fi

(Received 22 August 2008; accepted 4 March 2009)

ISSN 0036-5513 print/ISSN 1502-7686 online © 2009 Informa UK Ltd. (Informa Healthcare, Taylor & Francis AS)
DOI: 10.1080/00365510902878716
 586 P. Hedberg et al.
change of collection device from glass syringes to
plastic ones. Using the right sampling equipment
and improving the instructions to medical staff on
handling blood samples including the possible sources
of errors are essential for the sampling quality.
To our knowledge, no exact studies of the effect
of syringe sample volume to the blood gas, electro-
lyte and lactate values using Siemens Rapidlab
865 analyser and Siemens Rapidlyte plastic syringes
have been published.Very often the blood gas syringes
sent to the core laboratory from the hospital wards are
insufficiently filled with patient blood. The aim of
this study was to examine this bias and to give more
information and instructions to medical staff on
handling blood samples which should be applied in
Scand J Clin Lab Invest Downloaded from informahealthcare.com by CDL-UC Santa Cruz on 10/26/14
daily practice in the hospital laboratories, critical care
units and emergency rooms.
Materials and methods
Samples
Arterial venting 3 mL syringe with filter cap (04180028,
Bayer Corporation, East Walpole, MA 02032-1597,
USA) with total balanced heparin concentration of
about 70 IU/3 mL for blood gas, electrolyte and lactate
For personal use only.
analysis were collected from patients in the critical care
unit in Oulu University Hospital. The arterial samples
were obtained from an arterial cannula and the mixed
venous samples from a pulmonary artery catheter
(PAC). The total amount of sample serials was 127, of
which 71 were arterial and 56 pulmonary. The samples
were analysed as soon as possible after collection. The
samples were kept at 4C to slow down the metabo-
lism keeping the maximum storage time under 15 min-
utes. The samples were mixed very thoroughly
according to the manufacturer of the syringes before
they were analysed to ensure that a homogeneous part
of the samples are introduced into the analyser. Bub-
bles in the samples were avoided as much as possible,
but if bubbles had formed in the syringe, they were
removed immediately after sampling. To avoid air bub-
bles, the Rapidlyte syringes have a filter cap to remove
them after sample collection. The filter cap device is
also used as a stopper tip cap during transport to the
analyser. In the stability test the samples in the first
group were stored at room temperature (232C)
and mixed thoroughly before analysis and in the
second group they were stored at cooled conditions
(4C) before mixing and analysing.
Patients
The effect of sample volume on the results was studied
with a total of 127 patients. Four samples were
collected from each patient in plastic heparin bal-
anced syringes. No sample selection criteria were
used. Data exclusions were only made when there
was clear evidence of incomplete or suboptimal
sample aspiration or processing faults. All results
were processed anonymously.
All the patients granted written permission for
blood samples to be collected, and the local hospital
ethical committee approved the study.
Analytical methods
All analyses were done using the Rapidlab 865
analyser (Siemens Medical Solutions Diagnostics,
Tarrytown, NY 10591-5097, USA).
The analytical performance of the analyser was
followed with the protocol used in the blood gas
analysers in Oulu University Hospital laboratory.
The performance of the analyser was checked with
the three quality controls with different levels of
parameters (RapidQC Complete 1, 2 and 3, Bayer
Health Care). It was performed by analysing one
level in turn in the morning and afternoon and a
patient pool in the morning.
The study protocol
Before sample collection the full grant for permissions
were asked from the patients. After permission four
specimens were collected from each patient; 3.0 mL,
1.8 mL, 1.5 mL and 1.0 mL in 3.0 mL syringe samples.
The sample volumes were chosen as following; the
3.0 mL is a fully filled syringe and is a reference sam-
ple volume. The 1.8 mL is a volume, which has been
orally quoted by the manufacturer and is used in the
Oulu University Hospital, the last two volumes were
chosen to find out the exact limit of volume where
the possible changes in the results would appear.
The sample volume 1.0 mL in the 3 mL syringe
would have been insufficient for the analyser. Samples
were mixed thoroughly and analysed as soon as
possible. The results from 1.0 mL, 1.5 mL and 1.8 mL
samples were compared to the results from full-filled
(3 mL) samples.The data was processed in two groups,
venous and arterial blood samples.
Statistical methods
All data was expressed as mean with standard devia-
tion (SD), unless otherwise indicated. Statistical
analyses were performed with SPSS for Windows
software package (SPSS 15.0.1, SPSS Inc., Chicago,
Illinois, USA). Since the 3.0 mL, 1.8 mL, 1.5 mL and
1.0 mL blood samples were collected from the same
person the results were analysed using analysis of
variance for repeated measurements (R-ANOVA), if
the sphericity assumption did not hold the p-value
according to Greenhouse-Geisser is presented. When
the R-ANOVA showed a p-value  0.05 a pair wise
comparison between 3.0 mL (reference volume) and
other volumes was done, mean differences with 95%
confidence intervals and p-values with Sidak’s
multiple comparison corrections were presented.

Results
Changes observed in the blood gas, electrolyte
and lactate results of blood syringe samples
with different sample volumes are summarized in
Tables I and II.
When comparing the 1.8 mL sample volume
results to the baseline values established using a
3.0 mL sample volume, the bias % between the groups
ranged between  7.8% and  8.7%. The standard
deviations of biases for the base excess in blood
(BEvt) parameter in all groups were high, which may
be due to the low absolute values (negative vs posi-
tive). Also, the standard deviations for the fraction of
carboxyhemoglobin in total hemoglobin (COHb) and
Scand J Clin Lab Invest Downloaded from informahealthcare.com by CDL-UC Santa Cruz on 10/26/14
the fraction of methemoglobin in total hemoglobin
(MetHb) parameters were high in all groups and the
amounts of COHb and MetHb were very low.
All the analytes in the group 1.5 mL arterial samples
were found to change less than 10%, except COHb in
arterial samples, which showed a decrease of 14.8%
from the 3 mL sample value. In other parameters the
mean percentage changes varied between  8.7% and
7.3%. The bias % for pO2 was  8.6%.
Among the arterial blood samples over 10%
changes between the results of 3 mL and 1.0 mL
For personal use only.
samples were observed for pO2 (13.5%), BEvt
(11.3%) and COHb ( 19.8%). BEvt increased
14.8% and COHb decreased 13.5% with 1.0 mL
sample volume compared to 3.0 mL sample volume.
In other parameters the bias % varied between  8.9
and  8.2% in both of the groups.
The R-ANOVA showed a p-value  0.05 only for
hemoglobin, hematocrit, and Ca2 (7.4) among the
arterial samples, but the pair wise comparison was
also made for those. Although there were changes in
the bias % with a few parameters, the statistically
significant changes were observed more often. The
significant changes (p  0.05) were found for pO2,
O2Sat and Na in both of groups and with all the
sample volumes compared to 3 mL sample volume.
Among the arterial samples p-values were  0.05 for
Ca2(7.4) between the results of 3 mL and 1.5 mL
samples. In the comparison of 3 mL and 1 mL sam-
ples the p-values were  0.05 for COHb, MetHb, K
and Ca2 (7.4) and lactate.
The R-ANOVA showed a p-value 0.05 only for
hemoglobin, hematocrit and pH among the venous
samples, but the pair wise comparison was also made
for those. In our investigation, significant changes
were found with several parameters in the venous
sample group when the results of 3.0 mL and 1.0 mL
samples were compared. The changes were significant
for carbon dioxide tension (pCO2), HCO3 act, BEvt,
COHb, Kand Ca2 (7.4).The last parameter changed
significantly also with the 1.8 mL sample volume.
Additionally, when comparisons were made between
3 mL and all the other sample volumes, the significant
changes were found for lactate.
Sample volume needed for blood gas parameters 587
The change was especially seen in higher pO2
values. To clarify whether this difference is due to the
stability of the 1 mL sample, we investigated the
effect of storage of samples at  4C and 23  2C
with different time points (Table III). This study was
performed with 25 1 mL specimens taken from
5 different patients. The first sample of each patient
was analysed at a time point of one to five minutes
(baseline value) and the other samples were analysed
at the time points of 15, 30, 60 and 90 minutes. The
mean changes were less than 10% for pO2 parameter
(in the tested range of 4.1–20.6 kPa). The mean
changes for hemoglobin derivatives (COHb and
MetHb) and BEvt were over  10% at both storage
temperatures over the time inconsistently. The stan-
dard deviations of bias%s were high and the tested
values of these parameters in the samples were very
low. Lactate values increased also over the time
producing mean percentage changes over 10% within
30 minutes at room temperature and within 60 min
at  4C. The mean change for pCO2 was 10.8% at
23  2C during 90 minutes storage.
Discussion
The typical preanalytical errors are caused by the
air bubbles in the sample, inhomogeneous samples,
hemolysis in samples, improper storage of samples,
clots in the specimen, dilution of sample due to saline
in arterial lines and liquid heparin, heparin inter-
ference, arterial samples mixed with venous blood
and patient in an unsteady state of ventilation. The
impact of each of these causes of error must be
clearly acknowledged by those who are collecting
the specimens, analysing them and interpreting
the results to avoid making false decisions in
patient managements.
An air bubble left in the sample for a few minutes
can affect the pO2 values significantly [6,7]. An air
bubble with a relative volume of 0.5–1% of the sample
may seriously affect the pO2 value of the sample. In
our study the bubbles in the samples were avoided,
but if bubbles had formed in the syringe, they were
removed immediately after sampling. To avoid air
bubbles, the Rapidlyte syringes have a filter cap to
remove them after sample collection.
Significant differences have been found between
the type of collection device, storage conditions and
time to analysis in several studies. Mahoney et al. [8]
examined the changes in oxygen measurements in
blood using plastic and glass syringe samples stored ice.
The authors concluded that samples collected in a
plastic syringe should not be stored in ice. The glass
syringes may still be more appropriate in these
specific conditions. Wu et al. [9] according to
their studies recommended storage at room temper-
ature when specimens are collected in plastic syringes
and immediate analysis of the specimen. If the
 Scand J Clin Lab Invest Downloaded from informahealthcare.com by CDL-UC Santa Cruz on 10/26/14
For personal use only.

588

Table I. Changes of arterial blood gas, electrolyte and lactate values induced by different syringe sample volume. Data are presented as mean with standard deviation (SD), differences between the mean
values (in absolute values and bias %) and statistical significances (a pair wise comparison; a p value  0.05 was considered statistically significant and is marked with asterisks∗). The 95% confidence
interval (95%CI) was presented, when a p value was  0.05.
P. Hedberg et al.

Difference between Difference between Difference between

3.0 mL and 1.8 mL 3.0 mL and 1.5 mL 1.0 mL 3.0 mL and 1.0 mL
3.0 mL (n  71) 1.8 mL (n 71) 1.5 mL (n  71) (n 71)
Arterial sample Mean (SD) Mean (SD) Bias % p-value Mean (SD) Bias % p-value Mean (SD) Bias % p-value

pH 7.37 (0.05) 7.37 (0.05) 0.1 0.03∗, 0.000 7.37 (0.05) 0.02 0.90 7.37 (0.05)  0.01  0.9
to 0.007
pCO2 (kPa) 5.4 (0.6) 5.4 (0.6)  0.4 0.93 5.4 (0.5) 0.5 0.79 5.3 (0.5) 1.3 0.07
pO2 (kPa) 14.4 (4.1) 15.0 (4.2)  3.9  0.001∗,  0.883 15.7 (4.3)  8.6  0.001∗,  1.690 16.4 (4.5)  13.5  0.001∗,  2.516
to  0.253 to  0.779 to  1.386
HCO3 act (mmol/L) 23 (3) 23 (3) 0.4 0.63 23 (3) 0.8 0.22 23 (3) 1.2 0.03∗, 0.014
to 0.524
BEvt (mmol/L)  2.0 (3.1)  2.1 (3.1)  7.8 0.08  2.1 (3.1)  8.7 0.16  2.2 (3.0)  11.3 0.06
tHb (g/L) 103 (15) 102 (15) 0.3 0.64 103 (15) 0.04  0.9 102 (16) 0.4 0.58
Hct (g/dL) 30 (5) 30 (4) 0.2 0.98 30 (4)  0.03  0.9 30 (5) 0.4 0.79
O2SAT (%) 97.0 (1.9) 97.3 (1.7)  0.3 0.01∗,  0.4 97.5 (1.7)  0.5  0.001∗,  0.698 97.7 (1.7)  0.7  0.001∗,  0.927
55 to  0.038 to  0.249 to  0.428
COHb (%) 0.51 (0.31) 0.46 (0.29) 8.7 0.10 0.43 (0.27) 14.8 0.004∗, 0.018 0.41 (0.28) 19.8  0.001∗, 0.039
to 0.131 to 0.161
MetHb (%) 0.73 (0.20) 0.74 (0.20)  2.1 0.63 0.77 (0.20)  5.2 0.07 0.79 (0.19)  8.9 0.002∗,  0.112
to  0.018
Na (mmol/L) 138 (4) 138 (4) 0.2  0.001∗, 0.162 138 (4) 0.3  0.001∗, 0.283 138 (4) 0.6  0.001∗, 0.601
to 0.466 to 0.554 to 0.934
K (mmol/L) 4.2 (0.5) 4.1 (0.5) 0.2 0.54 4.1 (0.5) 0.2 0.54 4.1 (0.5) 0.8 0.002∗, 0.010
to 0.060
Ca2 (7.4) (mmol/L) 1.08 (0.06) 1.10 (0.10)  1.3 0.54 1.09 (0.06)  0.9  0.001∗,  0.013 1.10 (0.06)  1.6  0.001∗,  0.022
to  0.006 to  0.012
Lactate (mmol/L) 1.49 (0.54) 1.50 (0.54)  0.6 0.21 1.51 (0.54)  1.4 0.003∗,  0.038 1.52 (0.54)  1.8 0.002∗,  0.045
to  0.006 to  0.007

pO2, Oxygen tension; pCO2, Carbon dioxide tension; tHb, Total haemoglobin; metHb or FmetHb, Fraction of methemoglobin in total hemoglobin; COHB or FCOHb, Fraction of carboxyhemoglobin
in total hemoglobin; Hct, Hematocrit; HCO3 act, Actual concentration of bicarbonate; K, Potassium; Na, Sodium; Ca2, Ionized calcium; BEvt, Base excess in blood; O2Sat, Oxygen saturation.
 Scand J Clin Lab Invest Downloaded from informahealthcare.com by CDL-UC Santa Cruz on 10/26/14
For personal use only.

Table II. Changes of venous blood gas, electrolyte and lactate values induced by different syringe sample volumes. Data are presented as mean with standard deviation (SD), differences between the
mean values (in absolute values and bias %) and statistical significances (a pair wise comparison; a p value  0.05 was considered statistically significant and is marked with asterisks∗). The 95% confidence
interval (95%CI) was presented, when a p value was  0.05.

Difference between Difference between Difference between

3.0 mL and 1.8 mL 3.0 mL and 1.5 mL 1.0 mL 3.0 mL and 1.0 mL
3.0 mL (n 56) 1.8 mL (n 56) 1.5 mL (n 56) (n 56)
Venous Sample Mean (SD) Mean (SD) Bias % p-value Mean (SD) Bias % p-value Mean (SD) Bias % p-value

pH 7.33 (0.05) 7.33 (0.05)  0.02 0.54 7.33 (0.05)  0.01 0.99 7.33 (0.05)  0.01 0.99
pCO2 (kPa) 6.3 (0.6) 6.3 (0.6) 0.3 0.85 6.2 (0.6) 0.7 0.27 6.2 (0.6) 1.3 0.001∗, 0.026
to 0.130
pO2 (kPa) 4.5 (0.6) 4.7 (0.7)  2.3  0.001∗,  0.166 4.7 (0.6)  3.0  0.001∗,  0.192 4.8 (0.7)  5.2  0.001∗,  0.329
to  0.046 to  0.079 to  0.143
HCO3 act 24 (3) 24 (3) 0.04  0.9 24 (3) 0–7 0.39 24 (3) 1.2 0.003∗, 0.078
(mmol/L) to 0.511
BEvt  1.6 (3.2)  1.6 (3.2) 1.3  0.9  1.7 (3.0)  8.6 0.58  1.8 (3.2)  14.8 0.02∗, 0.021
(mmol/L) to 0.443
tHb (g/L) 101 (14) 101 (14) 0.2 0.99 100 (14) 0.6 0.47 100 (14) 0.6 0.36

Hct (g/dl) 30 (4) 30 (4) 0.5 0.71 30 (4) 0.8 0.32 29 (4) 0.9 0.10

O2SAT (%) 60.3 (8.6) 61.8 (9.1)  2.4  0.001∗,  2.238 62.2 (8.8)  3.1  0.001∗, 2.670 63.5 (8.9)  5.3  0.001∗,  4.253
to  0.676 to  1.076 to  2.083
COHb (%) 0.55 (0.30) 0.53 (0.28) 4.3 0.76 0.51 (0.30) 8.5 0.29 0.48 (0.29) 13.5 0.01∗, 0.011
to 0.139
MetHb (%) 0.74 (0.19) 0.76 (0.16)  2.6 0.75 0.78 (0.17)  5.6 0.08 0.78 (0.17)  6.0 0.11

Na (mmol/L) 139 (3) 139 (3) 0.1 0.03∗, 0.013 138 (3) 0.3  0.001∗, 0.266 138 (3) 0.6  0.001∗, 0.631
to 0.291 to 0.606 to 1.008
K (mmol/L) 4.2 (0.6) 4.2 (0.6) 0.1 0.99 4.2 (0.6) 0.6 0.31 4.2 (0.6) 1.0 0.001∗, 0.013
to 0.071
Ca2(7.4) 1.06 (0.05) 1.07 (0.05)  0.6  0.001∗,  0.011 1.06 (0.05)  0.4 0.51 1.07 (0.05)  1.2  0.001∗,  0.020
(mmol/L) to  0.003 to  0.006
Lactate 1.42 (0.43) 1.44 (0.43)  1.2 0.01∗,  0.030 1.44 (0.43)  1.6 0.01∗,  0.040 1.44 (0.42)  1.7 0.002∗,  0.040
(mmol/L) to  0.004 to  0.005 to  0.007

pO2, Oxygen tension; pCO2, Carbon dioxide tension; tHb, Total haemoglobin; metHb or FmetHb, Fraction of methemoglobin in total hemoglobin; COHB or FCOHb, Fraction of carboxyhemoglobin
in total hemoglobin; Hct, Hematocrit; HCO3–act, Actual concentration of bicarbonate; K, Potassium; Na, Sodium; Ca2, Ionized calcium; BEvt, Base excess in blood; O2Sat, Oxygen saturation.
Sample volume needed for blood gas parameters
589
 590 P. Hedberg et al.

Table III. The effect of storage of 1 mL samples at  4C and 23  2C with different time points. Stability of blood gas, electrolyte and
lactate values. The results were compared to the 0-5 min results. Data are presented as mean of bias % with standard deviation (SD).

23 2C  4C

n 5 15 min 30 min 60 min 90 min 15 min 30 min 60 min 90 min

pH 0.1 (0.1) 0.1 (0.1) 0.3(0.2) 0.6 (0.1) 0.1 (0.1)  0.1 (0.2) 0.3 (0.1) 0.2 (0.2)
pCO2 (kPa)  1.8 (1.4)  1.6 (5.1)  6.6 (3.7)  10.8 (3.0)  1.1 (2.2) 1.9 (4.6)  5.1 (1.5)  4.4 (2.3)
pO2 (kPa)  0.1 (3.4)  0.9 (0.7)  2.2 (8.0)  0.7 (6.6)  1.2 (6.4)  2.8 (5.8)  3.8 (8.6)  2.9 (5.7)
HCO3 act  0.7 (1.7)  0.4 (4.7)  0.7 (0.9)  0.1 (1.1)  0.1 (1.9) 0.9 (2.5)  0.4 (1.4)  0.6 (1.3)
(mmol/L)
BEvt  23.5  117.0 7.3 154.7  15.2  6.6 (11.6)  7.1 (13.6)  3.4
(mmol/L) (102.4) (271.0) (15.5) (224.7) (17.7) (16.2)
tHb (g/L) 0.6 (4.0) 3.5 (5.4) 1.4 (4.0)  0.1 (2.1)  1.7 (2.1)  1.7 (2.8)  1.5 (3.3)  1.0 (2.0)
Hct (g/dl)  0.4 (6.0) 2.2 (6.4) 1.2 (5.4)  0.8 (3.2)  1.5 (3.1)  0.7 (2.9)  1.5 (3.1)  0.7 (1.6)
O2SAT (%)  0.1 (3.0)  0.8 (0.8)  0.4 (1.6) 0.6 (1.4) 0.4 (4.8)  3.7 (6.9)  0.7 (1.1)  1.1 (5.0)
COHb (%)  90.0  120.0  85.0  105.0  13.3  66.7  51.7 5.0 (71.1)
(175.5) (164.3) (151.7) (94.2) (29.8) (188.2) (101.4)
Scand J Clin Lab Invest Downloaded from informahealthcare.com by CDL-UC Santa Cruz on 10/26/14

MetHb (%)  1.8  7.3 28.2 27.5 (18.9)  5.0 (25.9) 14.7 (24.3) 8.5 (35.6) 20.0 (27.7)
(43.3) (54.4) (28.1)
Na 0.7 (0.9) 0.6 (0.6) 0.7 (0.7) 0.6 (0.6)  0.1 (0.2)  0.1 (0.4) 0.2 (0.3) 0.2 (0.2)
(mmol/L)
K (mmol/L) 1.1 (0.8) 3.5 (3.5) 2.7 (0.6) 3.4 (0.7) 0.2 (0.3)  0.4 (0.7)  0.8 (0.9)  1.0 (1.3)
Ca2 (7.4)  0.2 (0.8) 1.4 (2.8) 0.9 (1.1) 1.6 (1.2)  0.2 (0.8) 0.5 (1.4) 0.6 (0.8) 1.3 (0.8)
(mmol/L)
Lactate  4.0 (2.5)  13.6  34.5  51.7 (21.8)  3.7 (2.3)  6.1 (9.1)  13.7  20.5
(mmol/L) (7.5) (14.8) (11.9) (12.3)

pO2, Oxygen tension; pCO2, Carbon dioxide tension; tHb, Total haemoglobin; metHb or FmetHb, Fraction of methemoglobin in total
hemoglobin; COHB or FCOHb, Fraction of carboxyhemoglobin in total hemoglobin; Hct, Hematocrit; HCO3 act, Actual concentration
of bicarbonate; K, Potassium; Na, Sodium; Ca2, Ionized calcium; BEvt, Base excess in blood; O2, Oxygen saturation.
For personal use only.

analysis is going to be delayed, the samples should
be drawn and stored in glass [10]. According to the
studies of Pretto and Rochford in 1994 [11] for
the blood gas samples with high arterial pO2 should
be analysed immediately for accurate results and if a
longer storage time is required, then the use of glass
syringes and adequate specimen cooling may be
required. According to Wiwanitkit in 2006 glass
syringes are better than plastic for preserving
arterial blood gas for oxygen partial pressure deter-
mination: an explanation based on nanomaterial
composition [12]. In the 100% oxygen test, the
glass (not plastic) syringes have been recommended
to be used. The syringe should be cooled immedi-
ately, even when the sample is analyzed as soon as
possible [13]. In the present study the main analyses
were done without delay thus there was no reason
for use of glass syringes. Additionally, in the stability
study no substantial changes were observed in high
pO2 values (bias% 3.8).
Excess sodium heparin solution as an anti-
coagulant, the type of heparin or dilution of blood
samples can significantly change the measured
values [7,14–19]. According to the manufacturer
Rapidlyte syringes used in this study are preheparinized
with dry heparin covering the whole syringe inner to
avoid the possible dilution of the sample, and avoid too
high an anticoagulant/blood volume ratio. Thus, the
results of our study can not be explained by the pos-
sible effect of the ratio of anticoagulant/blood volume.
The reason for the differences in the pO2, BEvt and
hemoglobin derivative parameter values using different
sample volumes remain still unclear. It may be due
to the analyser used; the plunger of the syringe is very
close to the sample needle of the analyser when 1 mL
sample volume is used. Also, the possibility of lower
numbers of blood cells in the syringe may cause
variations of blood gases. The stability study showed
that pO2 is stabile in 1 mL samples at room
temperature and at  4C for 90 minutes thus indi-
cating that the reason for differences was not the
stability. The stability study also showed that the dif-
ferences in the hemoglobin derivate parameters may
be due to the very low levels of these parameters.
The precision performance of each parameter has
also to be considered when finalizing the results. Our
experience with the Rapidlab 865 analysers are that
the between run precisions of the parameters with
different level of concentrations and values differ. For
example the precision for pO2 is typically 4–17%
(CV) or even higher depending of the level of
RapidQC complete control. The highest CV% is
usually observed for low level of pO2. The CV%s for
COHb are normally 15, 12.5 and 5% at the levels of
3, 5 and 20%, respectively. Among the metabolites,
lactate has the highest CV% at the very low
concentration level. The CV% can reach up to 16%
at the level of under 1.0 mmol/L, or even higher at
lower concentration values. The CV%s for MetHb
are usually 15, 12.5 and 5% at the levels 2.4, 7 and
16%. The typical CV%s for all the other parameters
are 4.5% and for pH 0.17. The CV%s is

presumably higher in the very low levels of COHb
and MetHb. The very high CV% may have had an
affect the results of this study. The precisions of the
calculated parameters (Bevt, HCO3act and O2Sat)
have not been followed in our laboratory. One
weakness of our study was the constricted sample
material. During the study the samples with patho-
logical hemoglobin derivative (MetHb and COHb)
and base excess values were not available. Thus
the possible 10% difference in these parameters at
the pathological level remains uncertain. The SD
for these three analytes was also very large in the
stability experiment. Further studies are needed to
clarify what are the exact relative differences at
higher values.
Scand J Clin Lab Invest Downloaded from informahealthcare.com by CDL-UC Santa Cruz on 10/26/14
Ricós et al. [20] have published the databases on
biological variation and derived from these data the
desirable specifications for total error, imprecision
and bias has been calculated [21]. Among the param-
eters in this list the desirable biases for hemoglobin,
lactate, pCO2 and pH are 1.8, 8.0, 1.8, 1.0%, respec-
tively. Thus, the present study indicates that the
biases with all sample volumes tested for these four
parameters are lower than these thresholds.
From a clinical point of view the errors in arterial
pO2 in 1.5 and 1.0 mL samples are critical. In bor-
For personal use only.
derline these errors can lead to misdiagnosis. The risk
for wrong clinical conclusion is high especially dur-
ing follow-up if the sample size varies. Measured
10% differences between 3.0 mL and 1.5 mL samples
in COHb have no clinical significance in normal
situations, but in carbon monoxide intoxication they
might be misleading.
In conclusion, using Rapidlyte plastic syringes
with filter cap, the minimum sample volume when
blood gases, electrolytes and lactate are all measured
should be 1.8 mL using a 3 mL syringe. No substantial
changes were found when 1.8 mL sample results were
compared to 3 mL sample results. According to our
results, 1.0 and 1.5 mL sample volumes, especially
when accurate results of pO2 and arterial blood is
needed for patient’s diagnosis, are not recommended.
Acknowledgements
In memoriam: We owe this study and our warmest
thanks to Dr Päivi Kaukoranta for her invaluable
collaboration during this work.
Declaration of interest: The authors report no
conflicts of interest. The authors alone are responsible
for the content and writing of the paper.
Abbreviations:
CLSI: Clinical and Laboratory Standards Institute;
NCCLS: National Committee for Clinical Laboratory
Standards.
Sample volume needed for blood gas parameters 591
References
[1] Rambaldi M, Baranzoni MT, Coppolecchia P, Moschello JN,
Novaco F. Blood gas and patient safety: considerations
based on experience developed in accordance with the
Risk Management perspective. Clin Chem Lab Med.
2007;45:774–80.
[2] Batchelder PB, Raley D. Maximizing the laboratory setting
for testing devices and understanding statistional output
in pulse oximetry. Anesth Analg. 2007;105(6 Suppl.):S85–94.
[3] Risch A, Biedler A, Mertzlufft F. Preanalytical errors on the
determination of arterial O(2)-partial pressure and their
impact on the AaDO(2). Anaesthesist. 2000;49:29–33.
[4] NCCLS. Blood gas and pH analysis and related measurements.
NCCLS document C46-A4 Wayne, PA. NCCLS, 2001.
[5] NCCLS. Procedures for the collection of arterial blood speci-
mens. NCCLS document H11-A4. Wayne, PA. NCCLS, 2004.
[6] Biswas CK, Ramos JM, Agroyannis B, Kerr DNS. Blood gas
analysis: effect of air bubbles in syringe and delay in
estimation. Br Med J. 1982;284:923–7.
[7] Wuillemin WA, Gerber AU. Sources of errors in the pre-
analytical phase of blood gas analysis. Scweiz Rundsch
Med Prax. 1995;84:200–3.
[8] Mahoney JJ, Harvey JA, Wong RJ, Van Kessel AL. Changes
in oxygen measurement when whole blood is stored
in iced plastic syringe or glass syringes. Clin Chem.
1991;37:1244–8.
[9] Wu EY, Baranzanji KW, Johnson RL. Source of error on
A-aDO2 calculated from blood stored in plastic and
glass syringes. J Appl Physiol. 1997;82:196–202.
[10] Knowles T, Mullin R, Hunter J, Douce FH. Effects of syringe
material, sample storage time and temperature on
blood gases and oxygen saturation in arterialized
human blood samples. Respir Care. 2006;51:732–6.
[11] Pretto JJ, Rochford PD. Effects of sample storage time,
temperature and syringe type on blood gas tensions in samples
with high oxygen partial pressures. Thorax. 1994;49:610–2.
[12] Wiwanitkit V. Glass syringes are better than plastic for
preserving arterial blood gas for oxygen partial pressure
determination: an explanation based on nanomaterial
composition. In J Nanomedicine. 2006;1:223–4.
[13] Smeenk F, Janssen J, Arends B, Harff J, van den Bosch J,
Schönberger J, Postmus P. Effects of four different methods
of sampling arterial blood and storage time on gas tensions
and shunt calculation in the 100% oxygen test. Eur Respir
J. 1997;10:910–3.
[14] Toffaletti J, Ernst P, Hunt P, Abrams B. Dry electrolyte-
balanced heparinised syringes evaluated for determining
ionized calcium and other electrolytes in whole blood.
Clin Chem. 1991;37:1730–3.
[15] Hopper K, Rezende Ml, Haskins SC. Assessment of the
dilution of blood samples with sodium heparin on blood gas,
electrolyte, and lactate measurements in dogs. Am J Vet Res.
2005;66:656–60.
[16] Börner U, Müller H, Höge R, Hempelmann G. The influence
of anticoagulant on acid-base status and blood-gas analysis.
Acta Anaesthesiol Scand. 1984;28:277–9.
[17] Goodwin NM, Schreiber MT. Effects of anticoagulants on acid-
base and blood gas estimations. Crit Care Med. 1979;7:473–4.
[18] Crawford AG. Liquid versus non-liquid heparinization of
blood samples. Resp Ther. 1983;5:103–5.
[19] Hutchison AS, Ralston SH, Dryburgh FJ, Small M,
Fogelmann I. Too much heparin: possible source of error in
blood gas analysis. Brit Med J. 1983;287:1131–2.
[20] Ricós C, Alvarez V, Cava F, García-Lario JV, Hernández A,
Jiménez CV, Minchinela J, Perich C, Simón M. Current
databases on biological variation: pros, cons and progress.
Scand J Clin Lab Invest. 1999;59:491–500.
[21] http://westgard.com/biodatabase1.htm