| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Oak Ridge Poster Sessions |
Hemosol Inc., 115 Skyway Ave., Etobicoke, ON M9W 4Z4, Canada
a author for correspondence: fax 416-798-0152
Hemoglobin interference in laboratory analyses due to hemolysis has been well characterized (1)(2)(3). Laboratory scientists are becoming sensitized to an additional iatrogenic cause of interference from hemoglobin. Stabilized hemoglobin-based oxygen carriers (HBOCs) have potential applications in trauma and emergency surgery, perioperative hemodilution, cardiopulmonary bypass, and hematopoiesis (4). Several of these products are in late clinical trials and are nearing marketing approvals. As a result, there has been increasing interest in the effects of these HBOCs on routine laboratory assays by clinical chemists and diagnostic reagent manufacturers.
This paper presents the results of a study that examined the effects of
an HBOC, o-raffinose cross-linked hemoglobin
(HemolinkTM) on selected routine assays on the Hitachi 717
with reagents from Boehringer Mannheim Corp. (BMC). Hemolink is
prepared from hemoglobin purified from outdated human red blood cells
previously tested and released for use in transfusion. The purified
hemoglobin reacts with o-raffinose, a polyaldehyde obtained
through oxidation of raffinose (5)(6). As a
result, two hemoglobin 
ß dimers are covalently cross-linked
between amino groups within the 2,3-diphosphoglycerate binding pocket
to form stable hemoglobin tetramers (64 kDa). In addition,
o-raffinose reacts with surface amino groups to form
intermolecular linkages producing stable hemoglobin polymers (128–600
kDa). The final Hemolink preparation contains ~40% of the hemoglobin
tetramer and ~60% of the hemoglobin polymers.
To study the effects of Hemolink on the measurement of routine analytes, three normal plasma pools and one plasma pool with increased analyte concentrations were prepared and tested with each of two lots of Hemolink to determine (a) whether there were pool-to-pool or (b) lot-to-lot differences in the interferences observed, and (c) whether the interference was dependent on analyte concentration. All samples were collected in compliance with institutional ethical standards. Samples were diluted 4:1 (by vol) with Hemolink and (or) lactated Ringer's injection USP (excipient for Hemolink) to obtain final Hemolink concentrations of 0.0 (control), 1.25, 2.5, 5.0, 7.5, 10, 15, and 20 g/L, measured as total hemoglobin.
Methods tested on the Hitachi 717 with BMC reagents included those for
sodium, potassium, chloride, urea (urease), creatinine (kinetic Jaffe),
total bilirubin (2,5-dichlorophenyldiazonium tetrafluoroborate),
albumin (bromcresol green), total protein (biuret), calcium
(o-cresolphthalein complexone), phosphorus (molybdate),
glucose (hexokinase, glucose-6-phosphate dehydrogenase), alkaline
phosphatase (ALP; p-nitrophenyl phosphate, AMP buffer),
alanine aminotransferase (ALT; IFCC), aspartate aminotransferase (AST;
IFCC), 
-glutamyltransferase (GGT; glutamyl-3-carboxy-4-nitroanilide,
glycylglycine), creatine kinase (CK; NAC-activated), lactate
dehydrogenase (LDH; lactate, NAD+), and iron (ferrozine).
Interference was calculated in terms of the absolute and relative error. Absolute error is defined as AI - A0, where AI is the analyte concentration in the sample containing the interferent and A0 is the analyte concentration in the control sample containing no interferent (7). Relative error is the ratio of these values and is calculated as AI/A0 (7). The interference of Hemolink was considered significant when the relative error exceeded a limit of 3 x CV of each method.
No interference was observed for Na, K, Cl, urea, glucose, total
protein, CK, AST, calcium, and phosphorus up to 20 g/L Hemolink.
Clinically significant interference was observed for albumin and ALT at
Hemolink concentrations >5.0 g/L; for creatinine, total bilirubin,
LDH, and GGT at Hemolink concentrations >1.25 g/L; and for ALP and
iron at all Hemolink concentrations tested. There were no significant
differences in the absolute or relative error between the normal and
abnormal sample pools tested and no significant differences between the
two lots of Hemolink tested. In addition, the interference was directly
proportional to the Hemolink concentration in the sample and was
independent of the analyte concentration in the range of values tested,
with the exception of bilirubin. The error in the total bilirubin
measurement varied with both the Hemolink and bilirubin concentration
in the sample, similar to results reported previously for the
interference of hemolysis with the BMC bilirubin method on the Hitachi
717 (7). For all analytes having linear interference from
Hemolink, the absolute error per g/L Hemolink and regression parameters
from analysis of the absolute error vs Hemolink concentration are
summarized in Table 1
. The error per g/L Hemolink differs significantly from the
reported effect of hemolysis on similar assays because of the presence
of intracellular contaminants from erythrocytes (e.g., LDH, AST, K, and
adenylate kinase) (7).
|
For those analytes showing significant interference from Hemolink, several variables were considered before reporting the results. First, the hemoglobin concentration of the sample was considered. Results for albumin, ALT, creatinine, total bilirubin, LDH, and GGT were within acceptable limits (±3 x CV) in the presence of low amounts of Hemolink, as described above. The intraindividual biological variability of each analyte was also considered. If the interference from Hemolink fell within the combined analytical and biological variability (8)(9)(10), Hemolink was not considered to interfere with clinical interpretation of the results. And finally, in conjunction with medical personnel, the interference was considered along with the magnitude and time course of changes ineach affected analyte associated with disease states. If the interference by Hemolink was not considered to obscure a clinically significant change in analyte concentration, the results were considered reportable. For example, at a plasma concentration of 20 g/L Hemolink, reported ALT values would be underestimated by 14 U/L. This interference could not be distinguished from the biological variability of ALT (8) and would not mask increases associated with hepatic disorders such as obstructive jaundice [2–10 times the upper limit of normal (ULN)] or viral hepatitis (>4 x ULN) (11)(12). Therefore, ALT values were considered reportable in the presence of Hemolink. In contrast, at 20 g/L plasma Hemolink, ALP values would be underestimated by ~120 U/L. This interference could mask increases in ALP after hepatitis or mild obstructive jaundice (2–5x ULN) (13)(14). Thus ALP values were not considered reportable in the presence of Hemolink.
Although the Hemolink concentration of the sample and the slope of the linear regression line can be used to predict the error in a given analyte measurement, correction factors should be used with caution. The laboratory should confirm that interference is caused only by the HBOC and not by hemolysis, since the correction factors for interference from hemolysis will differ from those of HBOCs because of the contribution of intracellular components of red cells (e.g., K, AST, and LDH) (1)(7). In addition, whether correction factors can be transferred between analyzers, reagents, or different types of HBOCs is not currently known.
In conclusion, BMC methods on the Hitachi 717 were acceptable for the clinical evaluation of electrolyte and calcium balance, urea, glucose, total protein, CK, and AST for plasma Hemolink concentrations up to 20 g/L. However, creatinine, LDH, GGT, total bilirubin, albumin, ALT, ALP, and iron had significant interference; thus, clinicalevaluation of renal and hepatic function in samples containing Hemolink is limited.
Acknowledgments
This work was performed by ClinTrials BioResearch (Montreal, Canada) under contract from Hemosol Inc. We are grateful to Yves Deschamps and his staff for their technical expertise.
References
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]
|
|
|
|
 |
|
 A. C. Y. Ali, C. C. Mihas, and J. A. Campbell Interferences of o-raffinose cross-linked hemoglobin in three methods for serum creatinine Clin. Chem., September 1, 1997; 43(9): 1738 - 1743. [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 |
