Printer Friendly

Effects of carboxyhemoglobin on hemoglobinometers.

To the Editor:

During a recent comparison of two portable hemoglobinometers designed for near-patient testing, we noted disparate readings on a quality-control material derived from human hemoglobin (Level 1, Multi-4 [R] CO-Oximeter Controls; Instrumentation Laboratory). The insert sheet listed an acceptable range of total hemoglobin readings of 159-185 g/L (15.9-18.5 g/dL) for the OSM3 co-oximeter (Radiometer A/S). Our results on the OSM3 were at the midpoint of this range, as were results with one of the hemoglobinometers, the Hb-Quick [R] hemoglobinometer from Avox Systems [mean (SD), 175 (0.5) g/L; n = 5]. By contrast, readings on a HemoCue hemoglobinometer (HemoCue AB) were 16 g/L higher than the mid-range value and averaged 188 [+ or -] 1.5 g/L. Both hemoglobinometers had recently been calibrated by factory-trained employees of the respective companies.

Because Level 1 controls contained ~60% carboxyhemoglobin, we made additional measurements on fresh whole blood from a nonsmoking donor and on other controls with low carboxyhemoglobin (Level 2, Multi-4 CO-Oximeter Controls). The same blood samples and controls were then equilibrated with carbon monoxide by rotating a syringe containing a small volume of blood or control and a much larger volume of carbon monoxide gas. With CO equilibration, the carboxyhemoglobin fraction (percentage of HbCO), as measured by the OSM3, increased from 2.9% to 94.9% in Level 2 controls, from 0.2% to 96.8% in whole blood, and from 57.3% to 93.1% in Level 1 controls, respectively.

The Hb-Quick and HemoCue both initially gave readings on Level 2 controls that agreed with the insert sheet (Table 1). With increased HbCO, the HemoCue's readings on Level 2 controls increased spuriously by 17 g/L, whereas readings on the Hb-Quick were not affected (P >0.05). Equilibrating whole blood with CO increased the Hb-Quick's average reading by a small although statistically significant 2.6 g/L, but the HemoCue's reading increased by 12.4 g/L. The effect of CO on the HemoCue was not consistent: equilibrating Level 2 control with CO increased readings on the HemoCue, but equilibrating Level 1 control with CO reduced the HemoCue's average reading by 3.0 g/L. With CO equilibration, the Hb-Quick's readings on Level 1 control rose slightly, by 2.2 g/L.

The two hemoglobinometers operate on different principles. To measure the total hemoglobin concentration spectrophotometrically in a typical clinical blood sample, these instruments must solve two problems: the intense light scattering caused by red blood cells, and the fact that the four principal species of hemoglobin (oxy-, deoxy-, carboxy-, and methemoglobin) each have distinct but overlapping absorbance spectra. The Hb-Quick hemoglobinometer (1) measures directly in unaltered whole blood by a proprietary optical design and mathematical algorithms to compensate for light scattering, and it uses a sufficient number of wavelengths to include all four principal species of hemoglobin in its measurements. By contrast, the HemoCue system (2,3) uses disposable cuvettes preloaded with dry reagents that first hemolyze the sample to eliminate light scattering and then convert the various hemoglobin types into a single species (azide-methemoglobin) so that the total hemoglobin concentration can be measured by the optical absorbance at a single wavelength.

In view of the principle of operation used by the Hb-Quick, it is understandable that this hemoglobinometer was relatively unaffected by carboxyhemoglobin because it includes that species in its measurement of total hemoglobin. The most likely explanation for the HemoCue's aberrant readings is that carboxyhemoglobin is relatively resistant to conversion into azide-methemoglobin. In fact, carboxyhemoglobin is a well-known source of error in the cyanmethemoglobin method and others that depend on chemical conversion (4-6). In further support of this explanation, when HemoCue cuvettes are filled with blood or hemoglobin-based controls, the sample usually turns brown within 1-2 min as the azide-methemoglobin is formed; however, we noticed that samples with high carboxyhemoglobin concentrations remained bright red, indicating that some or all of the carboxyhemoglobin was not converted into azide-methemoglobin.

As Table 1 shows, the effects of carboxyhemoglobin on the HemoCue were inconsistent and seemed to depend on the initial percentage of HbCO. Therefore, the HemoCue should be used with caution if quality-control solutions, proficiency-testing materials, or clinical blood samples contain significant fractions of carboxyhemoglobin. By contrast, the Hb-Quick gives accurate readings in the presence of high concentrations of carboxyhemoglobin both in whole blood and in hemoglobin-based controls.

We thank Olle Hagstrom for calibrating the HemoCue and providing a generous supply of cuvettes.


(1.) Gong, AK, Backenstose B. Evaluation of the Hb-Quick [R]: a portable hemoglobinometer. J Clin Monit 1999;15:171-7.

(2.) Vanzetti G. An azide-methemoglobin method for hemoglobin determination in blood. J Lab Clin Med 1966;67:116-26.

(3.) Granata S, Vanzetti G. Evaluation of HemoCue, an instrument for the assay of hemoglobin in undiluted blood specimens by the azidemethemoglobin method. Biochim Clin 1986;10:944.

(4.) Taylor JD, Miller JDM. A source of error in the cyanmethemoglobin method of determination of hemoglobin concentration in blood containing carbon monoxide. Am J Clin Pathol 1965; 43:265-71.

(5.) Chilcote ME, O'Dea AE. Lyophilized carbonylhemoglobin as a colorimetric hemoglobin standard. J Biol Chem 1953;200:117-24.

(6.) van Kampen EJ, Zijlstra WG. Spectrophotometry of hemoglobin and hemoglobin derivatives. Adv Clin Chem 1983;23:199-257.

J.M. Steinke [1,2]

A.K. Gong [1]

[1] Departments of Physiology and Pediatrics

University of Texas

Health Science Center

San Antonio, TX 78229-3900

[2] Avox Systems, Inc.

28267 Ruffian Drive

* Address correspondence to this author at: Department of Physiology-7756, University of Texas Health Science Center, San Antonio, TX 78229-3900. Fax 210567-4410; e-mail

A representative of HemoCue AB responds:

To the Editor:

It is well known that increased concentrations of carboxyhemoglobin (COHb) will lead to overestimation of hemoglobin concentrations as measured photometrically at a single wavelength (1). In the study above, at COHb concentrations of 60-90%, the HemoCue instrument read <3 g/L (0.3 g/dL) outside the acceptable range.

The relevance of these findings to clinical settings is tenuous. The normal COHb concentration in blood is 0.4-0.8%. In urban environments, nonsmokers have COHb concentrations typically in the 1-2% range. In smokers, COHb is typically 4-5% and as high as 9%. Values >10% are abnormal. At 20%, dizziness, nausea, and syncope develop, at 30%, visual disturbances develop, and at 50%, seizures and coma develop; 60% is associated with death (2).

The conversion of COHb to azidemethemoglobin in the HemoCue system or to cyanmethemoglobin in the International Reference Method takes longer than conversion of other hemoglobin derivatives. Because the absorbance of COHb at 570 nm is ~14.3 L x [mmol.sup.-1] x [cm.sup.-1], a total non-conversion of COHb produces higher values for the hemoglobin concentration. The upper limit of this error can be calculated. At 5% and 10% COHb, the maximum errors are 1.5% and 3%, respectively.

The authors of the above letter conclude that the effect of COHb is inconsistent on the HemoCue system. Only one fresh whole blood from a single nonsmoking donor and two control materials were assayed. When control material is assayed, matrix effects have to be considered. Careful study of additional samples is needed before this conclusion can be supported.


(1.) van Kampen EJ, Zijlstra WG. Spectrometry of hemoglobin and hemoglobin derivatives. Adv Clin Chem, 1983;23:199-257.

(2.) Kaplan LA, Pesce AJ. Clinical chemistry. Theory, analysis and correlation, 2nd ed. St. Louis, Mosby, 1989:523.

Olle Hagstrom

HemoCue AB

Box 1204

SE-262-23 Angelholm, Sweden

Fax 46-431-458225

Table 1. Effects of carboxyhemoglobin on two portable
hemoglobinometers. (a)

 Level 2 control

 HemoCue Hb-Quick

Before CO 144.0 [+ or -] 1.2 144.0 [+ or -] 1.2
After CO 160.6 [+ or -] 2.7 145.6 [+ or -] 0.5
Difference 16.6 1.6
P <0.0001 NS (b)

 Whole blood

 HemoCue Hb-Quick

Before CO 163.0 [+ or -] 2.5 160.4 [+ or -] 0.9
After CO 175.4 [+ or -] 1.8 163.0 [+ or -] 1.6
Difference 12.4 2.6
P 0.0002 0.03

 Level 1 control

 HemoCue Hb-Quick

Before CO 187.8 [+ or -] 1.5 174.6 [+ or -] 0.5
After CO 184.8 [+ or -] 1.3 176.8 [+ or -] 0.8
Difference -3.0 2.2
P 0.0026 0.01

(a) Readings of total hemoglobin (g/L) before and after
equilibration with CO gas were compared by a paired t-test.

(b) NS, not significant.
COPYRIGHT 2000 American Association for Clinical Chemistry, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2000 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:Letters
Author:Shepherd, A.P.; Steinke, J.M.; Gong, A.K.; Hagstrom, Olle
Publication:Clinical Chemistry
Article Type:Letter to the editor
Date:Apr 1, 2000
Previous Article:Increased carbohydrate-deficient transferrin concentration and abnormal protein glycosylation of unknown etiology in a patient with achondroplasia.
Next Article:AACC 2000 Award winners.

Related Articles
The Other Face of Botox; The Medical Letter Reports Wrinkles in the Botox(R) Cosmetic Panacea.
Fitch Upgrs $250MM Massachusetts Water Resources Authority, Series 1994 CP to 'F1+'.
Carbon monoxide exposure and carboxyhemoglobin: response.
Harvard Mental Health Letter Cites Lack of Testing as Reason for Caution on Antidepressants For Children.
Carbon monoxide test can be used to identify smokers.
Inappropriate reference intervals for carboxyhemoglobin at some Florida hospitals.

Terms of use | Copyright © 2017 Farlex, Inc. | Feedback | For webmasters