| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Oak Ridge Conference |
a Author for correspondence: fax 416-798-0152.
Abstract
The interferences of o-raffinose cross-linked hemoglobin
(HemolinkTM) were examined and compared in two
serum creatinine methods on the Hitachi 717 [Boehringer Mannheim
(BMC) and Synermed] and in an enzymatic creatinine method on the
Vitros 750 (Johnson & Johnson). Interference was considered significant
when the change in creatinine concentration from the control exceeded
the 95% confidence limits of each method. Significant interference was
observed for the BMC/Hitachi 717 method with Hemolink 
5 g/L. No
interference was observed for the Synermed/Hitachi 717 assay with
Hemolink up to 30 g/L in normal samples and up to 50 g/L in samples
with increased creatinine. No interference was observed for the Vitros
750 assay with Hemolink up to 50 g/L at both normal and increased
concentrations of creatinine. Although the BMC/Hitachi 717 method was
considered unacceptable, the Synermed/Hitachi 717 and the Vitros 750
methods allow accurate quantification of serum creatinine in the
presence of Hemolink.
Accurate measurement of clinical analytes in the presence of icterus, lipemia, and hemolysis has long been a challenge to clinical chemists (1)(2)(3). Before being introduced into the clinical setting, every new method and analytical instrument is routinely tested for possible interference from these factors as well as from commonly used drugs. However, new drugs are continually being brought to market, yet testing for interference with routine laboratory methods is rarely included as part of the development of these drugs. An exception to this has been the development of stabilized hemoglobin-based oxygen carriers (HBOCs).1 Because of the red color of these hemoglobin solutions, determination of the interference of HBOCs in routine laboratory tests has become a primary focus of the manufacturers of HBOCs, as evidenced by the formation of the Blood Substitutes Division of the AACC. Hospital-based clinical chemists are becoming sensitized to this iatrogenic cause of interference from hemoglobin now that several of these products are in late clinical trials throughout the US, Europe, and Canada.
The association between extracellular plasma hemoglobin and the development of renal damage is well known (4)(5)(6). In the early development of HBOCs, stroma-free hemoglobin preparations were associated with many of the physiological changes associated with hemolysis, including nephrotoxicity (7). The proposed mechanism of renal injury involves the release of iron from hemoglobin and the subsequent promotion of free radical formation, lipid peroxidation, and renal dysfunction (8)(9). The newer generation of HBOCs have been chemically modified to increase the molecular mass of the hemoglobin, thus minimizing renal clearance of hemoglobin while prolonging the life of the HBOC circulating in blood. Although renal damage from HBOCs is no longer a major safety concern, there remains the concern of a protein-overload type of phenomenon, given the potentially large doses of HBOCs that could be administered to trauma or surgical patients or a possible increase in the filtered load of HBOCs in patients with compromised glomerular permeability. Thus, accurate laboratory assessment of renal function is needed in patients receiving HBOCs.
In this study, we compared the interference of o-raffinose
cross-linked hemoglobin (HemolinkTM) with three methods for
serum creatinine that are based on different assay principles. Hemolink
is prepared from hemoglobin purified from outdated human erythrocytes
that were previously tested and released for use in transfusions. The
purified hemoglobin reacts with o-raffinose, a polyaldehyde
obtained through oxidation of the trisaccharide raffinose
(10)(11). The reductive alkylation process
covalently cross-links amino groups within the 2,3-diphosphoglycerate
binding pocket to form an irreversible linkage between two 
ß
dimers (forming a 64-kDa tetramer). In addition, o-raffinose
reacts with surface amino groups of the stabilized tetramers to create
intermolecular linkages producing stable hemoglobin polymers (128–600
kDa). The final Hemolink preparation is ~40% hemoglobin tetramer,
the remaining ~60% consisting of hemoglobin polymers. In this study,
our goal was to identify the method(s) with the least interference from
Hemolink added to sera from both healthy individuals and individuals
with increased concentrations of creatinine. The method(s) with the
least interference would then be used in future clinical trials as part
of the clinical assessment of renal function.
Materials and Methods
Creatinine assays.
The three methods for creatinine we
evaluated were the Boehringer Mannheim Corp. (BMC) method for the
Hitachi 717 analyzer; the Synermed Inc. (Montreal, Canada) method, also
for the Hitachi 717 analyzer; and the Johnson & Johnson Clinical
Diagnostics method for the Vitros 750 analyzer. The BMC method on the
Hitachi 717 (BMC/Hitachi 717) is a kinetic Jaffe method based on a
two-reagent system (reagent 1: 0.15 mol/L NaOH; reagent 2: 3.9 mmol/L
picric acid) monitored at 570 nm. The Synermed method on the Hitachi
717 (Synermed/Hitachi 717) is also a kinetic Jaffe method based on a
two-reagent system (reagent 1: 0.14 mol/L NaOH and a proprietary
mixture of nonreactive surfactants and solvents; reagent 2: 4.2
µmol/L picric acid) and is also monitored at 570 nm. The Johnson &
Johnson assay on the Vitros 750 (J&J/Vitros 750) is a kinetic enzymatic
assay in which the rate of formation of the final oxidized leuco dye is
monitored at 670 nm. All assays were performed according to the
procedures defined by the manufacturers.
Test samples.
Pools containing creatinine concentrations
within the reference interval were prepared from serum samples
collected from laboratory staff. Abnormal creatinine pools were
prepared from clinical specimens tested in a routine hospital
laboratory and found to have increased concentrations of creatinine.
All samples were collected in compliance with institutional ethical
standards. Samples were stored at 4 °C upon arrival in the
laboratory and tested within 7 days of collection.
Analysis of Interference.
Interference of Hemolink was
calculated as the recovery of creatinine in the Hemolink-supplemented
samples by use of the following formulas: absolute error =
CI - C0, and relative
error = CI/C0, where
CI is the creatinine concentration in a sample
containing Hemolink and C0 is the creatinine
concentration in the control sample without Hemolink (12).
Absolute and relative error were plotted as a function of the Hemolink
concentration in each sample. Any interference observed was judged to
be analytically significant when the relative interference exceeded the
95% confidence limits for each method.
Effect of Hemolink.
Hemolink was prepared as a 100 g/L
solution, measured as total hemoglobin with an IL-482 CO-oximeter
(Instrumentation Laboratory). Each serum sample was diluted
4:1 (by vol) with various amounts of Hemolink and Lactated Ringer's
Injection USP (LRS; excipient for Hemolink) to obtain approximate
Hemolink concentrations of 0.0 (control sample), 1.25, 2.5, 5.0, 7.5,
10, 15, and 20 g/L. Two production lots of Hemolink were tested with
serum pools containing normal and increased creatinine concentrations
to determine lot-to-lot and serum pool-to-pool consistency of the
potential interference. Samples were analyzed for creatinine by each of
the methods described above.
Those methods showing no interference from Hemolink up to 20 g/L underwent additional studies to examine the effect of Hemolink up to 50 g/L. This time, normal and abnormal serum pools were diluted 1:1 (by vol) with various amounts of Hemolink and LRS to obtain approximate Hemolink concentrations of 0.0 (control sample), 10, 20, 30, 40, and 50 g/L. Interference was analyzed as described above.
Interference of Hemolink vs hemoglobin.
For all three
methods, the interference from Hemolink was compared with that of a
purified hemoglobin preparation, to determine whether the interference
response was similar. Hemoglobin was prepared as an 80 g/L solution, as
measured with an IL-482 CO-oximeter. Pooled sera containing normal or
abnormal creatinine concentrations were diluted 4:1 (by vol) with
hemoglobin and LRS to obtain approximate hemoglobin concentrations of
0.0 (control sample), 4, 8, 12, and 16 g/L. Samples were analyzed for
creatinine by each of the methods described above, and results were
compared with the interference of Hemolink in each creatinine method.
Results
Several serum pools were used to complete the study. Creatinine concentrations ranged from 44 to 97 µmol/L in normal pools and from 147 to 548 µmol/L in abnormal pools. The 95% confidence limits for each method were calculated from the precision data for routine quality-control materials tested by each laboratory. The 95% confidence limits for the BMC/Hitachi 717 assay were 0.76–1.24 at a creatinine control concentration of 53 µmol/L and 0.91–1.09 at a control concentration of 141 µmol/L. The 95% confidence limits for the Synermed/Hitachi 717 assay were 0.89–1.11 at a creatinine control value of 159 µmol/L. A creatinine control at the low end of the reference range was not run for the Synermed/Hitachi 717 method. The 95% confidence limits for the J&J/Vitros 750 assay were 0.80–1.20 at a creatinine control concentration of 67 µmol/L and 0.90–1.10 at 203 µmol/L.
Hemolink had a positive interference with the BMC method on the Hitachi
717 (Fig. 1
). Recovery of creatinine in samples containing Hemolink at 20
g/L increased to 224% ± 14% of control values in samples with
creatinine values within the reference range (n = 6) and increased
to 143% ± 7% of control values in samples with increased creatinine
(n = 2). Hemolink interference with the BMC/Hitachi 717 method
exceeded the 95% confidence limits at Hemolink concentrations 5.0
g/L in both normal and abnormal serum pools. The absolute error in the
creatinine measurement (Fig. 1A
) showed a linear response to the
Hemolink concentration (normal pool, r = 0.984;
abnormal pool, r = 0.951). The error in the creatinine
measurement was 5.4 µmol/L per 1 g/L of Hemolink in the normal
samples and 7.9 µmol/L per 1 g/L of Hemolink in abnormal samples,
calculated as the slope of the regression line of absolute error vs
Hemolink concentration. The relative error depended on both the
creatinine concentration of the original sample and the Hemolink
concentration in each sample.
|
Both the Synermed/Hitachi 717 method and the J&J/Vitros 750 method
showed no significant interference from Hemolink at concentrations as
great as 20 g/L in the initial studies (Fig. 2
). The relative error observed for the Synermed/Hitachi 717
method varied from 0.89 to 1.12 in normal samples (n = 6) and from
0.95 to 1.06 in abnormal samples (n = 6). All results for normal
and abnormal pools supplemented with 1.25 to 15 g/L of Hemolink were
within the 95% confidence limits of the assay. For those samples
supplemented with 20 g/L of Hemolink, the relative errors of 4 of the 6
samples with normal creatinine concentrations were slightly above the
limit of 1.11 at relative errors of 1.12, 1.12, 1.14, and 1.22, whereas
the results for the samples containing increased creatinine
concentrations and Hemolink at 20 g/L were all within the 95%
confidence limits of the assay. The relative error observed for the
J&J/Vitros 750 method varied from 0.92 to 1.01 in normal samples
(n = 2) and from 0.97 to 1.00 in abnormal samples (n = 4) and
was within the 95% confidence limits of the assay.
|
When the Hemolink concentration was extended up to 50 g/L in later
studies (Fig. 3
), the 95% confidence limits of the Synermed/Hitachi 717 method
were exceeded at Hemolink concentrations 30 g/L in samples containing
normal concentrations of creatinine (n = 3) and were marginally
above the upper 95% confidence limit of 1.11 in samples with increased
creatinine concentrations (n = 3) (relative error of 1.12 ±
0.02 at 40 g/L Hemolink and 1.15 ± 0.03 at 50 g/L Hemolink). With
the J&J/Vitros 750 method, the recovery of creatinine was slightly
below the lower 95% confidence limit of 0.80 in the normal samples
(n = 6) containing 30 g/L Hemolink (relative errors of 0.79
± 0.07, 0.78 ± 0.06, and 0.79 ± 0.06 in samples containing
30, 40, and 50 g/L Hemolink, respectively). For samples containing
increased concentrations of creatinine (n = 6), the results were
marginally below the lower 95% confidence limit (0.90) of the
J&J/Vitros 750 method (relative error of 0.86 ± 0.04) for
Hemolink at 50 g/L.
|
The difference in Hemolink interference observed between samples from different individuals or between serum pools, and between lots of Hemolink was well within the 95% confidence limits of the assay and was not considered significant.
We also compared the interferences from Hemolink and a purified preparation of hemoglobin. Interestingly, hemoglobin up to 16 g/L showed no interference with any of the three creatinine methods tested. Recovery of creatinine was within the 95% confidence limits for samples with both normal (n = 3) and increased (n = 3) creatinine concentrations—in contrast to the results for Hemolink, which showed a positive interference with the BMC/Hitachi 717 method.
Discussion
This study was performed in preparation for a Phase I safety trial in healthy volunteers, and was part of an extensive evaluation of routine laboratory methods for interference from Hemolink (13). As mentioned previously, during the early development of HBOCs there were reports of renal toxicity following administration of HBOCs. In addition, administration of the current generation of HBOCs has been associated with increased iron in the renal tubules, the presence of red casts, and low concentrations of hemoglobin in the urine (14)(15)(16)(17), indicating that small amounts of these stabilized HBOCs may be filtered through the kidney. Thus, in early clinical trials of Hemolink, a method for serum creatinine was required that would give accurate results for an unambiguous clinical evaluation of renal function.
For each of the methods compared in this study, Hemolink interference that exceeded the 95% confidence limits of each assay was considered to be analytically significant. We also considered the within-subject biological variability of serum creatinine in determining the importance of the observed interference on the clinical interpretation of the reported result. The serum creatinine measured in an individual subject can change by ±10% by random variation alone (18). When this variability is taken in context with the analytical variability of the assay, a difference of 14%–18% is required for detection of a significant (P <0.05) change between two serial creatinine results from the same subject (19)(20)(21). Therefore, in addition to identifying the potential analytical interference by Hemolink, we defined a limit of ±14% from the control value as identifying an interference that would confound the clinical interpretation of serum creatinine measured in the presence of Hemolink.
The magnitude of the interference observed with the BMC/Hitachi 717
creatinine assay was considered unacceptable for reporting results in
the presence of Hemolink. Interference exceeded both analytical and
clinical limits at Hemolink concentrations 5.0 g/L in both normal and
abnormal samples.
The J&J/Vitros 750 analyzer, however, showed no analytically or clinically significant interference from Hemolink (at concentrations up to 50 g/L). A trend was observed of decreasing recovery of creatinine with increasing concentration of Hemolink, but this trend did not attain clinical significance. The J&J method involves the application of the sample to a spreading layer on the surface of the slide, and the sample diffuses through to the reagent layer, where the indicator reaction takes place. Possibly, the spreading layer on the slide acts as a molecular sieve to restrict the passage of hemoglobin and other large molecules such as Hemolink. As a result, neither hemoglobin nor Hemolink would diffuse into the reagent slide layer, thereby minimizing the potential of hemoglobin-based interference with this method. In addition, the rate of formation of the indicator dye is monitored at 670 nm (10/92 package insert for creatinine single-slide method, no. MP2–49; Eastman Kodak Clinical Diagnostics Division). The absorbance of hemoglobin is minimal at this wavelength, virtually eliminating the potential for interference from this source.
The Synermed wet-chemistry method performed on the Hitachi 717 showed significant positive interference from Hemolink at concentrations >30 g/L in samples with normal creatinine concentrations. There was also a positive trend in the recovery of creatinine in abnormal samples, but this trend was not considered clinically significant for Hemolink concentrations up to 50 g/L. The Synermed method is similar to the BMC method, both being based on the kinetic Jaffe reaction and measuring the rate of formation of the Janovsky complex at 570 nm. The traditional kinetic Jaffe method was subject to interference from hemolysis because the absorbance of hemoglobin changes progressively during the measurement phase of the reaction because of the oxidation of hemoglobin. Current methods have addressed this interference by optimizing the reading times of the creatinine assay to minimize the interference from hemoglobin and other factors (22)—as can be seen by the lack of interference from purified hemoglobin at up to 16 g/L in both the BMC/Hitachi 717 and Synermed/Hitachi 717 assays. In contrast to the lack of hemoglobin interference with the BMC/Hitachi 717 creatinine assay, Hemolink caused a large positive interference with this method. The observed interference probably results from the more rapid rate of oxidation of the o-raffinose cross-linked hemoglobin than that of purified hemoglobin. The addition of a proprietary mixture of solvents and surfactants into the NaOH reagent of the Synermed/Hitachi 717 assay is thought to pre-oxidize the hemoglobin before the start of the kinetic measurement, thus stabilizing the background absorbance of the test sample during the reading phase and minimizing the hemoglobin interference (M. Arentz, Synermed Inc., personal communication). Our results show that this reagent system is also able to minimize the interference from Hemolink.
During the interference studies performed in preparation for the Phase I study, the question arose as to whether the interference from Hemolink in routine assays could be predicted by published data or manufacturers' data on the interference of hemoglobin or hemolysis. However, the comparison of the interference of Hemolink and purified hemoglobin in the three creatinine methods described here showed that the effect of Hemolink and hemoglobin on creatinine methods can differ substantially. In fact, one cannot assume that if a method has no interference from hemoglobin, then Hemolink, or other HBOCs, will also have no effect. Although the Synermed and J&J methods performed well in the presence of both hemoglobin and Hemolink, the BMC method performed poorly, showing significant positive interference from Hemolink but not from hemoglobin, as discussed above. Thus the published data on the effects of hemoglobin on clinical assays cannot be used to predict the effect of HBOCs.
In the routine clinical setting, these results indicate a need for each laboratory to investigate the effects of HBOCs on their own analytical systems. Interference results on one analytical system cannot be generalized to other reagents or instruments. From the differences in interference results observed between Hemolink and hemoglobin, it is also likely that interference results for one HBOC cannot be extended to other HBOCs. The issue is further complicated if the laboratory does not know when or which HBOC the patient has received, or whether in vitro or in vivo hemolysis has occurred. The long-term solution is for the reagent and instrument manufacturers to develop new systems that either eliminate the interference of HBOCs and hemolysis, or at least warn of the presence of chromogens that produce inaccurate results (3). In the meantime, laboratories will need to assess the effect of these products on their analytical systems as the HBOCs are introduced into their hospitals, and, if necessary, introduce alternative methods (in-house or contract tests) to be able to report clinically meaningful results.
In conclusion, the BMC/Hitachi method showed significant interference from Hemolink and was considered unacceptable for evaluation of renal function in clinical trial subjects. The J&J/Vitros 750 dry-slide method showed no clinically significant interference from Hemolink up to 50 g/L. The Synermed method could, however, overestimate creatinine concentrations at the lower end of the reference range when Hemolink concentrations were >30 g/L, but showed no clinically significant interference from Hemolink up to 50 g/L at creatinine concentrations near or above the upper limit of the reference range. Thus both of these methods were deemed acceptable for use in the presence of plasma Hemolink concentrations as great as 50 g/L.
Acknowledgments
Testing was performed by ClinTrials BioResearch (Montreal, Canada) and MDS Laboratories (Etobicoke, Canada) under contract from Hemosol Inc. Some of the interference testing was coordinated courtesy of Synermed Inc. at a third-party clinical laboratory. We are grateful to Yves Deschamps and his staff at ClinTrials BioResearch for their technical expertise. We also thank Marcia Arentz (Synermed Inc.) and Jean Falagario (Johnson & Johnson Diagnostics) for valuable discussions.
Footnotes
Hemosol Inc., 115 Skyway Ave., Etobicoke, Ontario, Canada M9W 4Z4.
1 Nonstandard abbreviations: BMC, Boehringer Mannheim Corp.; HBOC, hemoglobin-based oxygen carrier; J&J, Johnson & Johnson Clinical Diagnostics; and LRS, lactated Ringer's injection. 
References
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  G. L. Myers, W. G. Miller, J. Coresh, J. Fleming, N. Greenberg, T. Greene, T. Hostetter, A. S. Levey, M. Panteghini, M. Welch, et al. Recommendations for Improving Serum Creatinine Measurement: A Report from the Laboratory Working Group of the National Kidney Disease Education Program Clin. Chem., January 1, 2006; 52(1): 5 - 18. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 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]
|
|
|
|
|
|
C. M. Balion, P. A. Champagne, and A. C. Y. Ali Evaluation of HemogloBindTM for Removal of o-Raffinose Cross-Linked Hemoglobin (HemolinkTM) from Serum Clin. Chem., September 1, 1997; 43(9): 1796 - 1798. [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
) Samples with creatinine concentrations within the reference range
(n = 6; mean ± SD); (•) samples with increased
creatinine concentrations (n = 2; individual values are
plotted).
