| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Automation and Analytical Techniques |
1
Department of Physiology, University of Texas Health Science Center, San Antonio, Texas 78284-7756.
2
Avox Systems, Inc., San Antonio, Texas 78015.
a Author for correspondence. Fax 210-567–4410; e-mail shepherd{at}uthscsa.edu.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
instruments evaluated
In this study we assessed the effects of perflubron emulsion on
the performance of eight widely used oximeters and CO-oximeters: the
AVL Omni 6, AVL Scientific Corporation; the AVOXimeters 1000 and 4000,
Avox Systems, Inc.; the CC270 CO-Oximeter, Ciba Corning Diagnostics
Corp (now Chiron); the IL482 and IL682 CO-Oximeters and the IL
Synthesis 35, Instrumentation Laboratory; and the OSM3 Hemoximeter,
Radiometer America, Inc.
The AVL Omni 6, CC270, IL482, IL682, and OSM3 are conventional CO-oximeters that first hemolyze the blood sample before analyzing it to eliminate the light scattering caused by red blood cells. Therefore, we shall refer to these particular instruments as hemolyzing CO-oximeters. By contrast, the AVOXimeters 1000 and 4000 and the IL Synthesis 35 are relatively new devices that make their measurements directly in unaltered blood without hemolysis. Hence, we shall refer to this group of instruments as nonhemolyzing CO-oximeters. The IL Synthesis aspirates each blood sample in a manner similar to conventional CO-oximeters, whereas the AVOXimeters 1000 and 4000 use disposable optical cuvettes that are first filled with the blood sample and then inserted into the instrument. A more detailed comparison of the designs, specifications, and operating characteristics of all eight instruments can be found elsewhere (7).
other materials
The human blood used in this study was taken from hospital
patients for routine diagnostic purposes and was provided to us in
accordance with a protocol approved by the Institutional Review Board
of the University of Texas Health Science Center. Standardized
hemoglobin solutions (Multi-4TM CO-Oximetry Controls)
were obtained from Instrumentation Laboratory.
sample preparation
We sought to assess the effects of perflubron on the performance
of the CO-oximeters under conditions that simulated the anticipated
clinical use of this blood substitute. Because perflubron will be
administered most frequently to surgical patients who have experienced
substantial blood loss, have received perflubron emulsion and other
fluids for volume replacement, and thus are in a hemodiluted state, we
standardized our blood samples at a total hemoglobin concentration of
70 g/L (7 g/dL), i.e., at an hematocrit of ~20%. To do so, we
centrifuged the original heparinized, arterial blood samples from
hospital patients to separate red blood cells from plasma and then
re-mixed red cells and plasma in appropriate proportions to obtain
stock blood samples with a total hemoglobin concentration adjusted to
70 g/L (7 g/dL). If their oxyhemoglobin fractions were not already
~97%, we tonometered these standardized blood samples briefly with
room air to bring them to that percentage. Oxygenating the blood in
this manner also ensured that subsequent handling and the addition of
the perflubron emulsion did not inadvertently alter the oxyhemoglobin
concentration. After initially obtaining somewhat erratic readings, we
found that more than usual effort was required to keep the blood
samples sufficiently well mixed, and we assumed that this precaution
was necessary because the fluorocarbon droplets, which are nearly twice
as dense as water, tend to precipitate rapidly without continuous
agitation to keep them in suspension.
experimental protocol
After the blood samples were prepared in the previously described
manner, they were then analyzed on each of the instruments being
tested. Subsequently, starting with 1 part of 60% emulsion in 20 parts
of blood, perflubron emulsion was added in increasing amounts to the
blood samples, and the measurements were repeated on each instrument
until the final concentration of perflubron in blood exceeded the
maximum anticipated therapeutic concentration. In this study, the
concentration of perflubron ranged from 0 to 81.7 g/L (8.17 g/dL).
data analysis
To analyze the data, we plotted the measured values of each
analyte vs the known concentrations of perflubron. As
Figs. 1–4
show,
our graphs are similar to the "interferographs" described by Glick
et al. (8) except that the values on the ordinate were not
normalized to the initial value because the fractions of the various
hemoglobin species are already percentages. Because we had limited
access to the IL682 and the AVL Omni 6, each data point on the graphs
for these two instruments is the mean of two measurements. However, all
of the other data points on the graphs represent the mean of five
measurements on each of the other instruments. In
Figs. 1–4
, the
results from the hemolyzing CO-oximeters are plotted in the upper panel
and the results from the nonhemolyzing CO-oximeters are plotted in the
lower panel. In the upper and lower panels of
Figs. 1–4
, the scales on
each pair of graphs are the same.
|
|
|
|
instrument operation
The calibration of each of the eight instruments under test was
confirmed by taking readings on appropriate control material, and each
instrument was operated in accordance with the manufacturer's
instructions. The hemolyzing instruments were operated in the routine
sampling mode rather than the "micro" mode. When the blood samples
were analyzed on the AVOXimeters 1000 and 4000, the disposable optical
cuvettes were filled, inspected for air bubbles and external
blood or debris, and then inserted promptly (within a few seconds) into
the instrument.
At the two highest perflubron concentrations, the OSM3 and the IL682 reported an error message indicating high turbidity; however, we recorded all readings regardless of whether an error condition was indicated. Only a limited number of measurements could be obtained from the AVL Omni 6 because it displayed multiple question marks as error messages and did not report the hemoglobin fractions at any of the perflubron concentrations except the lowest one.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
. As Fig. 1
(upper panel) shows, the Radiometer OSM3 and the AVL
Omni 6 experienced the greatest interference in their oxyhemoglobin
measurements. At the lowest perflubron concentration, the
readings of the AVL fell spuriously from a control value of 96.0% to
83.7% saturation. At higher perflubron concentrations, the AVL showed
question marks as error messages but gave no readings of oxyhemoglobin.
Measurements of oxyhemoglobin on the OSM3 fell spuriously from a
control value of 96.8% to 79.1% at the highest concentration of
perflubron. The IL482 was the instrument next most severely affected.
The percent saturation on the IL482 fell from 96.7% to 90.4% at the
highest perflubron concentration. As Fig. 1
(lower panel) shows, the
analyzers least affected by perflubron were the nonhemolyzing
instruments: the AVOXimeters 1000 and 4000 and the IL Synthesis. Among
the hemolyzing CO-oximeters, the IL682 and the CC270 were also
relatively unaffected by perflubron.
carboxyhemoglobin fraction
The measurements of the carboxyhemoglobin fraction made by seven
of the eight test instruments are shown in Fig. 2
. In the case of this analyte, the instruments most severely
affected by perflubron were the OSM3 and the AVL, as the upper panel of
Fig. 2
shows. Carboxyhemoglobin readings on the OSM3 rose spuriously
from 0.5% to 13.2% at the highest concentration of perflubron. The
AVL gave no carboxyhemoglobin readings except at the lowest perflubron
concentration. In contrast, the nonhemolyzing instruments, the
AVOXimeter 4000 and the IL Synthesis, were virtually unaffected by even
the highest concentration of perflubron. To be more specific,
measurements on the AVOXimeter 4000 fell from 1.5% to 1.3%. Among the
hemolyzing CO-oximeters, the IL482, IL682, and the CC270 were also not
appreciably affected by perflubron. For example, measurements on the
IL482 increased from 1.2% to 1.6%. Readings on the CC270 fell
appreciably only at the highest concentration of perflubron. No data
are shown for the AVOXimeter 1000 because it does not report the
fractions of carboxy- or methemoglobin.
methemoglobin fraction
The measurements of the methemoglobin fraction made by seven of
the eight test instruments are shown in Fig. 3
. In the case of methemoglobin measurements, the instrument most
severely affected by perflubron was the OSM3; its methemoglobin
readings rose in a concentration-dependent manner from a control value
of 1.2% to 8.4% at the highest concentration of perflubron. The AVL
showed question marks as error messages but gave no readings of
methemoglobin at perflubron concentrations above 28 g/L (2.8 g/dL). The
IL482 was less severely affected; its methemoglobin measurements rose
from 0.3% to 4.8%. The other instruments were even less affected by
perflubron. For example, methemoglobin measurements on the AVOXimeter
4000 rose from a control value of 0.9% to 1.4% at the highest
concentration of perflubron. Similarly, values on the IL Synthesis rose
by only 1.3%.
total hemoglobin concentration
The effect of perflubron on the total hemoglobin measurements made
by the test instruments is shown in Fig. 4
. Neither the hemolyzing (upper panel) nor the nonhemolyzing
instruments (lower panel) were seriously affected by even the highest
concentration of perflubron. To be specific, hemoglobin readings on the
OSM3 fell in a concentration-dependent manner from a control value of
71 g/L (7.1 g/dL) to 66 g/L (6.6 g/dL) at the highest concentration of
perflubron, and those on the IL682 increased from 70 g/L (7.0 g/dL) to
77 g/L (7.7 g/dL). Readings on the nonhemolyzing IL Synthesis rose in a
concentration-dependent manner by 6 g/L (0.6 g/dL). Readings on the
AVOXimeters 1000 and 4000 and CC270 remained within 4 g/L (0.4 g/dL) of
the control value.
As Fig. 4
shows, all eight instruments were slightly affected by
perflubron; however, none of the perflubron-induced errors was likely
to be clinically significant. For example, the reading of the total
hemoglobin concentration on the OSM3 was only 5 g/L (0.5 g/dL) lower
than control at the highest concentration of perflubron. Of course, an
absolute error of 5 g/L (0.5 g/dL) is only a 3.3% error at a total
hemoglobin concentration of 150 g/L (15 g/dL), and in a hemodiluted
patient with a total hemoglobin concentration of 70 g/L (7 g/dL) is
still only a 7.1% error.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
show, standardizing
the blood samples in this manner also greatly facilitated comparing one
instrument with another. However, the interference of perflubron on the
measurements of the hemoglobin fractions was inversely related to
hematocrit in one study (1). Thus, the effects of perflubron
reported here may be greater than they would be at a normal hematocrit.
This caveat does not apply to the measurements of total hemoglobin.
Because this study was designed to simulate the arterial blood of
surgical patients receiving perflubron and oxygen, the concentrations
of carboxy- and methemoglobin were low by design, i.e., in the 0–3%
range. Although our data demonstrate the directional effects of
perflubron interference on each of the analytes shown in
Figs. 1–4
,
one should not extrapolate the present findings to markedly different
conditions. For example, Figs. 1
and 2
show that perflubron causes the
OSM3 to give spuriously low oxyhemoglobin readings and spuriously high
carboxyhemoglobin readings when the true carboxyhemoglobin
concentration is in the 0–3% range. It is not known, however, if
perflubron causes the OSM3 to give spuriously high carboxyhemoglobin
concentrations when the carboxyhemoglobin concentration is actually
50%. A similar caveat applies to increased methemoglobin
concentrations. Nevertheless, the data shown here should be useful
under the intraoperative conditions in which perflubron emulsion will
be most frequently be administered.
A third caveat regards the perflubron concentration range we studied. Perflubron is expected to be administered at a dose of 0.9–2.7 grams of fluorocarbon per kilogram of body weight. Because the actual pharmacokinetics and blood volume of a patient are seldom known, it is difficult to predict the concentration that perflubron will actually reach even if the exact dose is known; however, by assuming an normal blood volume and typical ratios of blood volume to body weight, e.g., 7.5%, one can estimate that the therapeutic perflubron concentration range will be ~10–40 grams of fluorocarbon per liter (~1–4 grams of fluorocarbon per deciliter) of blood. Of course, it could be slightly higher if administered to an obese or hypovolemic patient. Thus, it is likely that the perflubron concentrations we studied [0–80 g/L (0–8 g/dL)] include and exceed the therapeutic range.
hemolyzing co-oximeters
In this study we examined the effects of perflubron on the
performance of five of the most widely used conventional hemolyzing
CO-oximeters (AVL, CC270, IL482, IL682, and OSM3) and on three
relatively new instruments, the AVOXimeters 1000 and 4000 and the IL
Synthesis. Conventional CO-oximeters like the OSM3 and the IL482
operate by aspirating a blood sample, hemolyzing the sample either
ultrasonically (OSM3) or chemically (IL482) to eliminate the light
scattering caused by red blood cells, and then subjecting the hemolyzed
sample to multiwavelength spectrophotometry. Because hemolyzing
CO-oximeters rely on hemolysis to eliminate light scattering in the
sample, it was not surprising that several of them performed poorly in
our tests. In fact, previous studies have shown that the measurements
of conventional CO-oximeters are often seriously affected if
emulsified, light-scattering particles like fluorocarbons
(2)(3)(4) or lipids (9)(10)(11) persist after the blood
samples are hemolyzed.
In our tests the OSM3 and the AVL were the instruments most severely
affected by the fluorocarbon emulsion. This result was not anticipated
because the OSM3 measures the residual turbidity that persists in the
hemolyzed sample. Although the turbidity corrections of the OSM3 were
not sufficient to compensate for the additional light scattering caused
by the perflubron emulsion, the OSM3 often indicated turbidity was high
(see Materials and Methods). The performance of the AVL was
more difficult to assess because it did not give readings at all
perflubron concentrations; however, the sharp fall in its oxyhemoglobin
reading at the lowest perflubron concentration (Fig. 1
, upper panel) indicates that this instrument is not likely to be useful
for clinical assessments of oxygen transport in patients receiving
perflubron. Among the conventional hemolyzing CO-oximeters we
tested, the CC270 was least affected by perflubron; only at the highest
perflubron concentration did its readings of oxy-, carboxy-, and
methemoglobin depart appreciably from the control reading (
Figs. 1–3
,
upper panels).
nonhemolyzing co-oximeters
As a group, the nonhemolyzing instruments were less affected by
perflubron-containing blood samples than the conventional hemolyzing
instruments. At all of the perflubron concentrations we studied, the
AVOXimeters 1000 and 4000 and the IL Synthesis consistently gave
clinically acceptable results for all of the analytes they measure
(
Figs. 1–4
, lower panels). These instruments are produced under
license from the University of Texas, and they are the first to make
multiple spectrophotometric measurements directly in unaltered whole
blood without first hemolyzing the sample (patents pending). The
physical optics of these instruments were selected to maximize the true
optical absorbance of intact blood and to minimize the contribution
that light scattering makes to the total optical attenuation of
nonhemolyzed blood. In addition, they use a sufficient number of
wavelengths to assess the magnitude of light scattering in each sample
and thus to correct for it using a series of effective algorithms.
Moreover, it should be noted that the magnitude of light scattering in
nonhemolyzed blood (12)(13) exceeds the
relatively mild turbidity of hemolyzed blood by a factor of 20–50.
Therefore, we anticipated that the additional light scattering caused
by perflubron emulsion would not appreciably affect the measurements
made by the nonhemolyzing instruments. Because the nonhemolyzing
instruments gave clinically acceptable measurements, particularly of
oxyhemoglobin and total hemoglobin concentrations and thus of the
hemoglobin-bound oxygen, these instruments are well-suited to assess
oxygen transport in patients receiving perflubron therapeutically as a
temporary oxygen carrier. Of course, when oxygen and therapeutic
concentrations of perflubron are administered, measurements of
PO2 are also necessary to assess the
oxygen in plasma and the substantial amount dissolved in the
fluorocarbon. The precision, accuracy, and long-term stability of the
AVOXimeters, as well as their ability to function without
anticoagulants, have been reported in a series of recent publications
(14)(15)(16). The IL Synthesis is so new that only a preliminary
publication has reported its performance (17).
Blood samples containing perflubron will present a challenge to laboratorians and clinical chemists. In fact, even using the present findings to correct the readings of a given instrument will be difficult because there is presently no convenient way to measure the concentration of perflubron in a patient's blood. Therefore, the soundest approach to assessing oxygen transport in these patients is to use one of the instruments shown here to be unaffected by even the highest concentrations of perflubron. Our findings also indicate that the nonhemolyzing CO-oximeters provide an effective means for determining whether a hemolyzing CO-oximeter is experiencing clinically significant interference in blood samples from patients receiving fluorocarbon emulsions.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  A. A. Ali, G. S. Ali, J. M. Steinke, and A. P. Shepherd Co-Oximetry Interference by Hemoglobin-Based Blood Substitutes Anesth. Analg., April 1, 2001; 92(4): 863 - 869. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. G. Toffaletti and R. F. Wildermann Use of Turbidity-Correction Algorithm Eliminates the Effect of Perflubron Emulsion on CO-Oximeter Results Clin. Chem., January 1, 2000; 46(1): 136 - 137. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
), OSM3; (
), IL 482; (
),
IL682; (
), CC270; (
), AVL; (
), AVOX4000; (
), AVOX1000;
(
), Synthesis.
