Anesth Analg 2005;100:437-439
© 2005 International Anesthesia Research Society
doi: 10.1213/01.ANE.0000143337.05366.CD
TECHNOLOGY, COMPUTING, AND SIMULATION
Does Methemoglobin from Oxidized Hemoglobin-Based Oxygen Carrier (Hemoglobin Glutamer-200) Interfere with Lactate Measurement (YSI 2700 SELECTTM Biochemistry Analyzer)?
Stephen L. Osgood, MD*,
Jonathan S. Jahr, MD*
,
Poonam Desai*,
Jessica Tsukamoto*, and
Bernd Driessen, DVM, PhD*
*Department of Anesthesiology, David Geffen School of Medicine at University of California Los Angeles; Charles R. Drew University of Medicine and Science, Martin Luther King Jr./Drew Medical Center, Los Angeles, California; and Department of Clinical Studies, University of Pennsylvania School of Veterinary Medicine, Philadelphia
Address correspondence and reprint requests to Jonathan S. Jahr, MD, Department of Anesthesiology, David Geffen School of Medicine at UCLA, PO Box 951778, Los Angeles, CA 90095-1778. Address e-mail to jsjahr{at}mednet.ucla.edu.
 |
Abstract
|
|---|
In this study, we validated the accuracy of lactate measurements (YSI 2700 SELECTTM glucose/lactate analyzer) in the presence of methemoglobin from an oxidized bag of hemoglobin-based oxygen carrier (Met-HBOC), hemoglobin glutamer-200 (Oxyglobin®; Biopure Corp). Different combinations of concentrated l-lactate solution, pooled canine plasma, and Plasmalyte ATM were added to 4 sample groups (1%, 10%, 20%, and 40% Met-HBOC [1.3 g/dL]) to yield linear increases in lactate concentration in consecutive samples. The mean difference between measured and calculated lactate was –5.1 mg/dL (1% Met-HBOC), –5.8 mg/dL (10% Met-HBOC), –4.6 mg (20% Met-HBOC), and –8.5 mg/dL (40% Met-HBOC). The root mean square error was 6.5 mg/dL, 7.4 mg/dL, 6.8 mg/dL, and 10.3 mg/dL, respectively. The Bland-Altman correlation (r) was r = –0.94 (P = 0.01), r = –0.91 (P < 0.001), r = –0.90 (P < 0.001), and r = –0.94 (P < 0.001), respectively, where r = 0 for perfect agreement between measured and calculated values. Results indicate that true lactate levels in the presence of Met-HBOC are underestimated when measured by an YSI 2700 analyzer independent of the amount of Met-HBOC present. When interpreting lactate concentrations from a patient with a HBOC present in plasma, underestimation of true lactate levels may occur unrelated to methemoglobin concentrations.
|
Introduction
|
|---|
Hemoglobin-based oxygen carriers (HBOCs) are novel solutions of cell-free hemoglobin that effectively transport and deliver oxygen to the periphery, thereby serving as alternatives to allogenic red blood cells for use in blood transfusions (1,2). HBOCs have been targeted for use in cases of severe blood loss and hemorrhagic shock secondary to traumatic injury or to treat perioperative anemia during elective surgery. Other investigators have studied interference effects from HBOCs on routine chemistries, therapeutic drugs, coagulation, hematology, and blood bank analysis, but none have looked specifically at oxidized HBOCs in these settings (3–5).
Because cell-free hemoglobin is an avid oxygen scavenger in the absence of reducing enzymes contained within red blood cells, HBOCs exposed to air become rapidly oxidized, increasing the amount of methemoglobin compared with oxyhemoglobin (6). This study validated the accuracy of lactate measurements using a YSI 2700 SELECTTM Biochemistry Analyzer (YSI Inc, Yellow Springs, OH) in the presence of methemoglobin from oxidized HBOC (Met-HBOC) and hemoglobin glutamer-200 (Oxyglobin®; Biopure Corp, Cambridge, MA). The correlation between analyzer-measured lactate and calculated or actual lactate concentration was studied (reference range for lactate [venous] is 4.5–20 mg/dL). We hypothesized that the presence of Met-HBOC would interfere with the accuracy of measured lactate values because Met-HBOC undergoes a rapid redox reaction with hydrogen peroxide that the YSI 2700 uses, in part, to measure lactate. With less hydrogen peroxide reaching the electrode in the analyzer, a smaller current is generated, and underestimation of actual lactate concentrations could occur.
|
Methods
|
|---|
We used remainder canine plasma from the University of California at Davis, where Animal Use and Care Administrative Advisory Committee approval was obtained. Different combinations of 80% Met-HBOC and 1% Met-HBOC were mixed to yield 4 groups containing 1%, 10%, 20%, and 40% Met-HBOC. Different combinations of concentrated l-lactate solution (Sigma, St Louis, MO), pooled canine plasma, and Plasmalyte ATM (Baxter, Deerfield, IL), a balanced salt solution (Plasmalyte ATM does not contain any lactate), were added to the Met-HBOC groups to make a linear and constant increase in lactate concentration in consecutive samples (15 mg/dL to 99 mg/dL; 15 total samples per group except for the 1% Met-HBOC group, which had 8 samples because of a shortage of canine plasma). The final HBOC concentration in each sample was 1.3 g/dL (Table 1).
A previously opened bag and a new sealed bag of hemoglobin glutamer-200 were determined to be 80% and 1% methemoglobin, respectively, by co-oximetry (Radiometer ABL 700 Series, Copenhagen, Denmark) on the experiment day. The 80% Met-HBOC had a baseline lactate concentration of 97.8 mg/dL, and the 1% Met-HBOC contained 96.5 mg/dL. Baseline lactate present in canine plasma was 16.6 mg/dL. Duplicate lactate measurements were taken by the YSI 2700, and the average was used for statistical analysis. The data were analyzed using root mean square error of the differences (RMSE; RMSE = 
[(mean of calculated – measured difference)2 + [sd of the differences)2]) and Bland-Altman statistical analysis to compare the calculated values to the measured values (7,8).
|
Results
|
|---|
The mean differences between measured minus calculated lactate were –5.1 mg/dL, (9.5% of the overall mean of measured and calculated lactate values) for 1% Met-HBOC, –5.8 mg/dL (10.7% overall mean) for 10% Met-HBOC, –4.6 mg (8.5% overall mean) for 20% Met-HBOC, and –8.5 mg/dL (16.1% overall mean) for 40% Met-HBOC (Fig. 1). The RMSE was 6.5 mg/dL (12% of overall mean), 7.4 mg/dL (13.7% overall mean), 6.8 mg/dL (12.6% overall mean), and 10.3 mg/dL (19.5% overall mean), respectively. The Bland-Altman correlation (r) between measured minus calculated difference versus the average of the measured and calculated lactate was r = –0.94 (P = 0.01), r = –0.91 (P < 0.001), r = –0.90 (P < 0.001), and r = –0.94 (P < 0.001), respectively, where Bland-Altman r = 0 for perfect agreement between measured and calculated values.
View larger version (19K):
[in this window]
[in a new window]
|
Figure 1. Lactate interference by oxidized hemoglobin-based oxygen carriers (HBOCs). Bland-Altman plots for samples containing hemoglobin glutamer-200 (Oxyglobin®), added l-lactate, Plasmalyte ATM, and canine plasma tested on a YSI 2700 SELECTTM glucose/lactate analyzer. The difference between analyzer-measured and calculated lactate concentration versus the average of measured and calculated lactate for each different series of methemoglobin (MetHb) concentration (1%, 10%, 20%, and 40%) is plotted. The reference range for lactate (venous) is 4.5–20 mg/dL.
|
|
|
Discussion
|
|---|
Calculated lactate was the term we used for what the true lactate level should be, taking into consideration everything containing lactate that was added to the test samples. This included the baseline lactate levels contained in the initial preparation of hemoglobin glutamer-200, the lactate present in canine plasma, and different amounts of lactate added from a concentrated stock solution. Measured lactate was the value given by the YSI 2700 after analyzing a given test sample.
In comparing measured to calculated values for every test sample for each concentration of Met-HBOC, results indicate that actual lactate levels in the presence of Met-HBOC were consistently underestimated (mean differences were all negative) independent of the amount of Met-HBOC present. Although the 20% Met-HBOC group had the most accurate measured values (mean difference and Bland-Altman correlation coefficient statistic) and the 40% Met-HBOC group had the most negative mean difference and greatest measurement error (RMSE), neither was significantly different from the 1% and 10% Met-HBOC groups. However, all groups significantly underestimated actual lactate concentrations (P 
0.01).
All hemoglobins, including modified hemoglobins (i.e., HBOCs) in particular, have a propensity to react with hydrogen peroxide (i.e., oxidation of heme Fe) in an enzymatic fashion (6). Biopure Corp. double packs hemoglobin glutamer-200 in airtight containers to prevent oxidation. Usual methemoglobin concentration in an unopened bag is approximately 1%. However, some clinical trials and studies using a similar product have detected increased methemoglobin levels in new products (4).
In the YSI analyzers, lactate is bound by a substrate-specific enzyme, lactate oxidase, which oxidizes lactate, generating an amount of hydrogen peroxide that is directly proportional to the amount of lactate present in the sample. It seems that the HBOCs likely scavenged enough hydrogen peroxide to cause the machine to report a smaller lactate concentration than it should have. Diaspirin cross-linked hemoglobin has been shown to interfere with l-lactate measurement using the Dade ACA IV analyzer (Dade Behring Inc, Wilmington, DE) (5).
Therefore, when interpreting lactate concentrations from a patient with a HBOC present in plasma, underestimation of true lactate levels may occur unrelated to methemoglobin concentrations but as a result of the redox reaction between HBOC and hydrogen peroxide in the lactate analyzer.
The authors thank Mohammed Saad, MD, FRCP, for use of his laboratory and Rima Boyadjian for her superior technical assistance. We also thank Jeffrey Gornbein, Dr PH, UCLA Department of Biomathematics, for his insightful statistical support. For access to the clinical laboratory at UCLA, we thank Anthony Butch, PhD.
|
Footnotes
|
|---|
Accepted for publication August 10, 2004.
|
References
|
|---|
- Winslow RM. Current status of blood substitute research: towards a paradigm. J Intern Med 2003;253:508–17.[ISI][Medline]
- Winslow RM. Alternative oxygen therapeutics: products, status of clinical trials, and future prospects. Curr Hematol Rep 2003;2:503–10.[Medline]
- Jahr JS, Lurie F, Gosselin R, et al. Effects of hemoglobin glutamer-250 (bovine) (HBOC-201, Hemopure) on coagulation testing. Am J Ther 2002;9:431–6.[Medline]
- Ali AA, Ali GS, Steinke JM, Shepherd AP. Co-oximetry interference by hemoglobin-based blood substitutes. Anesth Analg 2001;92:863–9.[Abstract/Free Full Text]
- Kazmierczak SC, Cartrou PG, Best AE, et al. Multiple regression analysis of interference effects from a hemoglobin-based oxygen carrier solution. Clin Chem Lab Med 1999;37:453–64.[Medline]
- Yeh LH, Alayash AI. Redox side reactions of haemoglobin and cell signalling mechanisms. J Intern Med 2003;253:518–26.[ISI][Medline]
- Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1:307–10.[ISI][Medline]
- Mantha S, Roizen MF, Fleischer LA, et al. Comparing methods of clinical measurement: reporting standards for Bland-Altman analysis. Anesth Analg 2000;90:593–602.[Abstract/Free Full Text]
Top
Abstract
Introduction
