| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Laboratory Management |
1
EURO-TROL b.v., NL-6702 EA Wageningen, The Netherlands.
2
Department of Medical Physiology, University of
Groningen, NL-9712 KZ Groningen, The Netherlands.
3
Department of Chemical Technology, Technical University
Eindhoven, NL-5600 MB Eindhoven,The Netherlands.
4
Department of Pediatrics, University Hospital, NL-9713
GZ Groningen, The Netherlands.
a Author for correspondence. Fax 31-503-611671; e-mail w.a.kamps{at}med.rug.nl.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Comparative measurements have shown that most of these instruments are quite reliable when used within the specifications of the manufacturers (5)(6)(7)(8). Quality control is rather difficult. Of course, the instruments can be tested with stable dye solutions. However, these solutions show non-identical spectral behavior in comparison with blood. The ideal test material would be a solution of hemoglobin derivatives of known concentration. Such solutions are not stable for sufficient time periods. Therefore, we investigated whether a lyophilized bovine hemoglobin (LBH) preparation containing various fractions of O2Hb, COHb, and MetHb may be suitable for quality control in MCA of hemoglobin derivatives.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
SFH solutions were equilibrated with O2 and CO to obtain O2Hb and COHb, respectively, and treated with an oxidizing agent (sodium nitrite) to induce formation of MetHb. The pH of these solutions was adjusted to 7.26 at 25 °C by the addition of diluted sodium hydroxide or hydrochloric acid solution. Solutions of pure O2Hb, COHb, or MetHb and a series of four mixtures of these hemoglobin derivatives covering the pathophysiologically interesting ranges (FO2Hb, 90–40%; FCOHb, 5–20%; and FMetHb, 0–40%) were lyophilized in 3-mL vacuum-sealed vials, each vial containing 0.5 mL of material, and stored at 2–8 °C in the dark.
reconstitution, measurement, and stability of lbh
For use, the contents of a vial were reconstituted by adding 0.5
mL of demineralized water with a PO2
of ~20 kPa and proper mixing until a homogeneous solution was
obtained, which takes ~2 min. The intervial precision was determined
by measuring five vials of each mixture with an IL282 multiwavelength
hemoglobin photometer. For practical reasons we decided not to use HHb
because of its affinity to oxygen, which leads to unpredictable results
after reconstitution. To ascertain that each mixture is free of HHb,
five vials of each mixture were reconstituted with demineralized water
with PO2 >100 kPa and measured on
an IL282 multiwavelength hemoglobin photometer.
The absorptivities of pure O2Hb, COHb, and MetHb of LBH were established by measuring the absorbance of reconstituted vials of each derivative with an HP8450A reversed-optics spectrophotometer at room temperature (480–650 nm with intervals of 2 nm). Calibration of the spectrophotometer was verified with a Corning HT yellow filter calibrated by NIST. The absorptivities of LBH derivatives were compared with those of bovine and human hemoglobin, which had been published previously (11)(12). Because of the dependence of the absorption spectrum of MetHb on the pH of the solution, we recorded the absorptivity as a function of wavelength for LBH-MetHb at various pH values. The pH of the solutions was measured at 25 °C with an ABL500 blood gas analyzer (Radiometer).
The stability of the LBH solutions after reconstitution of mixtures of LBH by adding 0.5 mL of water to the vials was established after storage of the solutions at 25 °C for various periods of time. The O2Hb, COHb, and MetHb fractions of the solutions were measured at 10-min intervals with a Radiometer ABL520 multiwavelength hemoglobin photometer.
The shelf life was determined by storage of vials with mixtures of LBH at 2–8 °C. Measurements of reconstituted samples were performed at 15-min intervals with an IL282 multiwavelength hemoglobin photometer.
computer program simulating mca
A computer program has been developed to calculate the composition
of mixtures of hemoglobin derivatives by simulating MCA, which is based
on the Lambert-Beer law:
 | (1) |
where A, the absorbance, is log Io/I,
is the millimolar absorptivity
(L · mmol-1 · cm-1) at wavelength
, c is the concentration (mmol/L), and l is
the lightpath length in centimeters. The equation can be solved for
c when A has been measured at a fixed wavelength
() where is known, provided that the
lightpath length is known exactly. For a solution containing several
components, Eq. 1
becomes:
 | (2) |
Writing A' for A/l, this becomes
for a solution of O2Hb, COHb, and MetHb:
 | (3) |
Measuring A' of the solution at three different
wavelengths yields a set of three linear equations with three unknown
concentrations. For an overdetermined system of n >3, this set of
equations can be written according to Eq. 4
for each wavelength at
which e is the error:
 | (4) |
Eq. 4
for all wavelengths used can be summarized as Eq. 5 , in
which 
e (a measure for total error) is
minimal when
(e)/cO2Hb
is 0, (e)/cCOHb
is 0, and
(e)/ cMetHb
is 0:
 | (5) |
This yields a determined system of three equations with three unknown concentrations of O2Hb, COHb, and MetHb that can be solved with standard mathematical procedures.
The program has been used with a set of absorbance data for wavelengths from 480 to 650 nm with intervals of 2 nm (general model) and specific sets of absorbance data for wavelengths as applied in the various types of multiwavelength hemoglobin photometers (specific models).
multicomponent analysis
We carried out a series of experiments with the four mixtures of
LBH derivatives to determine the following: (a), the
influence of the use of different absorptivities (LBH, bovine, and
human); (b), the influence of different sets of wavelengths
(general model or various specific models); and (c), the
combined influence of different absorptivities and different sets of
wavelengths. Each of the specific models when used in combination with
the absorptivities of human hemoglobin was the closest simulation of a
particular multiwavelength hemoglobin photometer we could make. The
results obtained by these simulations were then compared with those
obtained by measurements with the corresponding multiwavelength
hemoglobin photometers in various hospital laboratories.
The absorbance of the four mixtures of LBH derivatives was measured by a similar procedure on the same HP8450A spectrophotometer as had been used for the determination of the O2Hb, COHb, and MetHb absorptivities of LBH.
The influence of differences in the absorptivities on the results of MCA of mixtures of LBH derivatives was established by calculations with the general model. The O2Hb, COHb, and MetHb fractions of the four different mixtures of LBH derivatives were determined with the general model and the absorptivities of LBH, bovine hemoglobin, and human hemoglobin, respectively. The results obtained with the LBH absorptivities were designated the true values of the LBH mixtures.
To determine the influence of the specific models on the results of MCA of mixtures of LBH derivatives, we compared the true values of the LBH mixtures with the results of calculations with the specific models based on LBH absorptivities. The following specific models, corresponding as far as the wavelengths were concerned with several common multiwavelength hemoglobin photometers (7), were used: a 4-wavelength model (corresponding to the CO-Oximeter IL282/482, Instrumentation Laboratory), a 6-wavelength model (corresponding to the Hemoximeter OSM3/ABL520, Radiometer), an 8-wavelength model (corresponding to the CO-Oximeter CCD270, Chiron Diagnostics), and a 17-wavelength model (corresponding to the CO-Oxylite AVL912).
The combined influence of specific models and the absorptivities of human hemoglobin was determined by calculating the O2Hb, COHb, and MetHb fractions of the four different mixtures of LBH derivatives with the specific models based on human absorptivities. Thus, we applied to the LBH mixtures the methods as implemented in the common multiwavelength hemoglobin photometers. The results were compared with the true values of the LBH mixtures.
Finally, a comparison was made of the true values of the LBH mixtures with the results obtained with multiwavelength hemoglobin photometers at various hospital laboratories. To this end, a set of vials containing the four mixtures of LBH derivatives was sent to these laboratories as unknown test material in a quality assessment trial. The mean values of the results obtained for each LBH derivative were grouped for the various types of instruments: 14 IL282/482, 13 OSM3/ABL520, 17 CCD270, and 4 AVL912 multiwavelength hemoglobin photometers.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
20 kPa, is shown in Table 1
. The SD of each hemoglobin derivative of each LBH mixture of
all samples was ~0.2%, which corresponds to a coefficient of
variation (CV) of 0.1–0.2% in FO2Hb,
0.2–2.3% in FCOHb, and 0.2–2.4% in
FMetHb.2
Some minute differences were obtained when the reconstitution
was carried out with demineralized water with
PO2 >100 kPa (Table 1
), showing the
presence of traces of HHb in some of the LBH mixtures.
|
The left panels of Fig. 1
show the absorptivities of bovine, human, and lyophilized
bovine hemoglobin as a function of wavelength for O2Hb
and COHb. The right panels of Fig. 1
show the deviation the
absorptivities of bovine and human hemoglobin from the LBH. The
absorptivities of MetHb as a function of wavelengths are shown in Fig. 2
. In the upper panel, the spectrum of LBH MetHb at pH 7.06 is
compared with that of native bovine MetHb at pH 7.28, which is within
the range of pH values of freshly hemolyzed bovine blood. The lower
panel shows a comparison of the absorption spectrum of LBH MetHb at pH
7.26 with that of human MetHb at pH 7.21, which is within the range of
pH values of freshly hemolyzed human blood.
|
|
In Table 2
, the absorptivities of the various derivatives of LBH and of
bovine and human hemoglobin at the wavelengths used in common
multiwavelength hemoglobin photometers and in the corresponding
specific models are consolidated.
|
Fig. 3
shows the stability after reconstitution of a mixture of LBH
derivatives when stored for 1 h at 25 °C. When vials containing
lyophilized LBH derivatives were stored at 2–8 °C and measured
after reconstitution several times over 1.5 years, no changes in the
composition of the different LBH mixtures were observed.
|
The second column of Fig. 4
shows the true values of the LBH mixtures, i.e., the results of
calculating the fractions of O2Hb, COHb, and MetHb of
the four LBH mixtures with the help of the general model and the
absorptivities of LBH. The results of the same calculations with the
absorptivities of either bovine or human hemoglobin are shown as the
differences between these values and the corresponding true values.
When the absorptivities of bovine hemoglobin are used, the differences
from the true values are slight (from -0.8% to 1.3%), which is in
line with the expectation based on Figs. 1
and 2
. The values obtained
with the absorptivities of human hemoglobin deviate slightly more (from
-2.3% to 1.9%).
|
The open bars in Fig. 5
represent the results obtained when the specific models and the
absorptivities of LBH were used. The bars represent the differences
with respect to the true values of the four LBH mixtures in the same
manner as those in Fig. 4
. It seems that the more overdetermined
systems give a somewhat better agreement. The shaded bars in Fig. 5
show the results obtained with the specific models based on the
absorptivities of human hemoglobin, again with respect to the true
values. In addition, the results obtained with various multiwavelength
hemoglobin photometers at different laboratories are presented by the
solid bars.
|
It should be noted that in the calculations underlying the open bars, the only difference with the determination of the true values is the use of a limited set of wavelengths. The shaded bars show a greater deviation from the true values because in these calculations the use of the absorptivities of human hemoglobin instead of LBH absorptivities comes in addition to the selection of a limited set of wavelengths. These bars are the result of the closest simulation of the multiwavelength hemoglobin photometers, for which the actual outcome is presented by the solid bars. The differences between the simulated multiwavelength hemoglobin photometers, as given by the shaded bars, and the true values are often, but not in all cases, in agreement with the differences between the results of the actual multiwavelength photometers and the true values, as presented by the solid bars.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
For practical reasons we preferred bovine blood as starting material. It can be obtained easily, can be processed without the precautionary measures necessary in handling human blood, and its spectrophotometric properties are only slightly different from those of human blood (12). Because we chose a dry material, which must be redissolved by adding water, we had to restrict the mixture to the three components O2Hb, COHb, and MetHb. Ideally, a four-component mixture including HHb, which would have allowed us to check the oxygen saturation measurement simultaneously, would have been preferable; however, the strict control of PO2 necessary for the presence of a constant fraction of HHb has, to date, proven to be an insurmountable problem.
As is shown in Table 1
, there is hardly any difference between the
solutions obtained after reconstitution of different vials containing
the same mixture of LBH derivatives. Reconstitution with water with a
very high PO2 shows that there is
indeed practically no HHb present in the mixtures. Although part of the
differences found between the two reconstitution procedures is
statistically significant, the amount of HHb actually present in the
mixtures is negligible for all practical purposes. The long-term
stability of the LBH preparations warrants a shelf-life of >1 year
when the vials are kept at 2–8 °C. After reconstitution, there is
ample time for the measurements, as shown by Fig. 3
.
Comparison of the absorptivities of LBH with those of bovine and human
hemoglobin, as presented in Figs. 1
and 2
and in Table 2
, shows that
processing of bovine hemoglobin, i.e., preparation of the stroma-free
solution and lyophilization, storage, and reconstitution of the
solution, has little influence on the absorption spectra of the
hemoglobin derivatives. However, it appears that the affinity of LBH
MetHb for the OH- ion is different from that of
bovine hemoglobin, as demonstrated by the upper panel of Fig. 2
; the
best coincidence between the absorptivity spectra of LBH MetHb and
bovine MetHb is obtained at different pH values. The pH that ensues
after reconstitution of the LBH solution is 7.26, which is within the
range of pH values of freshly hemolyzed bovine blood, i.e., the pH at
which the absorptivities of bovine hemoglobin have been determined
(12). The pH of freshly hemolyzed human blood is ~7.21;
the absorptivities of human blood have therefore been measured at this
pH value (11). The fairly good agreement between the
absorptivities of LBH MetHb at pH 7.26 and human MetHb at pH 7.21, as
shown in the lower panel of Fig. 2
, was a good reason to keep the pH of
reconstituted LBH at 7.26, instead of lowering it to the value that
gave best agreement of the absorption spectrum with that of bovine
hemoglobin.
We termed the O2Hb, COHb, and MetHb fractions,
determined with the general model and with the absorptivities of the
pure LBH solutions, the true values. These values have been used as
baseline values in Figs. 4
and 5
. We realize that this term is only
justified within the perspective of what is presently attainable with
spectrophotometric methods. Because the method used to obtain the true
values is not different from the one that had been checked thoroughly
earlier by means of other, as far as possible, nonspectrophotometric
methods (2), the uncertainty in the true O2Hb,
COHb, and MetHb fractions may be assumed not to exceed ±
2%.2
As shown in Fig. 4
, the influence of using the absorptivities of bovine
hemoglobin instead of those determined for LBH is negligible for all
practical purposes, as could be expected on the basis of the
absorptivity spectra shown in Figs. 1
and 2
. This confirms that the
reconstituted LBH solution is spectrophotometrically equivalent with
fresh bovine hemoglobin. Fig. 4
also shows that the error introduced by
using the absorptivities of human hemoglobin is slight enough to
consider LBH a fair substitute for human hemoglobin in
spectrophotometric MCA.
As with the present methods, the results of MCA of hemoglobin
derivatives are not completely independent of the set of wavelengths
used; we therefore investigated the relevance of these differences with
the help of our computer model, by varying the set of wavelengths while
keeping absorbances and absorptivities constant. The difference in the
results obtained with the specific models in comparison with the
general model, as illustrated by the open bars in Fig. 5
, seems to
confirm the theoretical expectation that measuring at more wavelengths
allows better analysis of a multicomponent mixture. However, it also
appears that the selection of particular wavelengths has a distinct
influence on the results.
Each of the specific models used in combination with the absorptivities
of human hemoglobin was the closest simulation of a particular
multiwavelength hemoglobin photometer we could make. As shown by the
shaded bars in Fig. 5
, the results thus calculated for the four LBH
mixtures are in fairly good agreement with the true values, with the
exception of the four-wavelength model, in which the results sometimes
differ >5% from the true values. Fig. 5
also shows that the deviation
of the results obtained with the various multiwavelength hemoglobin
photometers from the true values (solid bars) is in quite a few cases
not in agreement with the deviation of the results of the corresponding
specific model from the true values. Several properties of the
multiwavelength hemoglobin photometers are obviously not taken into
account in the specific models or are considered in a different way.
There may be differences in spectral band width, in the absorptivities
used, and in the way the fractions of the hemoglobin derivatives are
calculated from the absorbances. Moreover, the difference in
temperature between the calculations, which were based on absorbances
and absorptivities measured at room temperature, and the
multiwavelength hemoglobin photometers, which usually are at 37 °C,
may have played a part. As shown by Steinke and Shepherd
(15), temperature has a small but appreciable influence on
the absorption spectra of O2Hb, COHb, and HHb. This
influence, however, can hardly account for much of the difference
between calculated and measured values, because the measured
absorbances and the absorptivities used in the calculations will be
similarly influenced by the differences in temperature. For MetHb the
situation is more complex, because the higher temperature in the
hemoglobin photometers will lower the pH, and any difference in
OH- affinity between LBH MetHb and human MetHb might have
yielded a less perfect agreement at 37 °C between the spectra of
human and LBH MetHb than the one shown in the lower panel of Fig. 2
.
Possible differences in the mathematical processing of the absorbance data even within the same type of multiwavelength hemoglobin photometer have been suggested by the work of Gourlain et al. (16) who, in their investigation of the interference of methylene blue in MCA of hemoglobin by various instruments, found that methylene blue affected the OSM3 and the Radiometer ABL520 differently. The two instruments contain the same photometric system; therefore, differences in data processing must be responsible for the difference in methylene blue interference. In the present investigation we could not analyze any differences between these instruments because we did not receive the results of the field trial with LBH separately for the OSM3 and the ABL520.
Our attempt to establish more accurate target values for the LBH preparations in the various multiwavelength hemoglobin photometers by computer simulation of the measuring systems implemented in these instruments was only partly successful. However, this approach may be useful for studying the influence of factors such as wavelength selection or differences in absorptivities on MCA of hemoglobin derivatives. The possible interference by contaminating dyes, as described by Gourlain et al. (16), probably can be studied more easily in this way.
The data for Fig. 5
suggest that our simulation is best for the OSM3
and the CCD270, although certainly not perfect. In these cases, when
the results obtained with the corresponding model had been taken as the
target values instead of our true values for the four LBH preparations,
the agreement with the results obtained with the actual instruments
would have been better.
Apart from these considerations, the potential of LBH as a reference material is shown by the fact that the fractions as measured with the multiwavelength hemoglobin photometers are generally within 6% of the true values and that often the agreement is much better. This is a promising result in light of the present limitations of spectrophotometric MCA, which include a high sensitivity to small differences in wavelength, the varying influence of turbidity as a result of different procedures used for hemolysis and turbidity control, and the uncertainty in the true values.
In summary, it has been demonstrated that stable LBH can be prepared containing constant mixtures of O2Hb, COHb, and MetHb, which after reconstitution yields a hemoglobin solution that is spectrophotometrically equivalent with a solution of fresh bovine hemoglobin. The results obtained thus far with an LBH preparation show its suitability as a quality-control material for photometric MCA of hemoglobin. Using a computer model of a particular multiwavelength hemoglobin photometer may become a suitable method for more accurately assigning target values to a particular LBH mixture.
|
Footnotes |
|---|
2 Fractions of the hemoglobin derivatives are given in percentages: F = 0.3 = 30%. When the spread in this value is given as 2%, this means that F is between 28% and 32%. The percentage is not used as a relative measure unless it is explicitly stated, as in the case of CVs.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  J. Toffaletti and W. G. Zijlstra Misconceptions in Reporting Oxygen Saturation Anesth. Analg., December 1, 2007; 105(6S_Suppl): S5 - S9. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
), LBH; (———), bovine hemoglobin; (- - - - -), human
hemoglobin.
